Diss. ETH No. 12700

Reduction of Polyhalogenated Alkanes by an Iron Porphyrin as Electron Transfer Mediator: System and Product Analysis

A dissertation submitted to the SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH for the degree of DOCTOR OF NATURAL SCIENCES

presented by

JOHANNA BUSCHMANN Dipl. Chem. ETH born February 12, 1969 citizen of Arbon (TG)

accepted on the recommendation of Prof. Dr. Rene Schwarzenbach, examiner Dr. Werner Angst, co-examiner Prof. Dr. Bernhard Krautler, co-examiner

Zilrich 1998 Der Schauende

Ich sehe den Baumen die Sttirme an, die aus laugewordenen Tagen an meine i:ingstlichen Fenster schlagen, und hore die fernen Dinge sagen, die ich nicht ohne Freund ertragen, nicht ohne Schwester lieben kann.

Da geht der Sturm, ein Umgestalter, geht

Wie ist das klein, womit wir ringen, was mit uns ringt, wie ist das gross; liessen wir, ahnlicher den Dingen, uns so vom grossen Sturm bezwingen, wir wurden weit und namenlos.

Was wir besiegen, ist das Kleine, und der Erfolg selbst macht uns klein. Das Ewige und Ungemeine will nicht von uns gebogen sein. Das ist der Engel, der den Ringern des Alten Testaments erschien: wenn seiner Widersacher Sehnen im Kampfe sich metallen dehnen, fi.ihlt er sie unter seinen Fingern wie Saiten tiefer Melodien.

Wen dieser Engel uberwand, wekher so oft auf Kampf verzichtet, der geht gerecht und aufgerichtet und gross aus jener harten Hand, die sich, wie formend, an ihn schmiegte. Die Siege laden ihn nicht ein. Sein W achstum ist: der Tiefbesiegte von immer Crosserem zu sein. Rainer Maria Rilke Acknowledgements

I want to thank Prof. Rene Schwarzenbach for letting me work in his group and for advising this research. Special thanks to Dr. Werner Angst and Dr. Judith Perlinger for their discussions where many ideas were evaluated and thought over. I thank Dr. Angst and Prof. Krautler for being my co-examiners. Dr. Angst, Prof. Krautler, Janet Kessel- mann and Prof. Anderson are kindly acknowledged for reviewing this dissertation.

I would like to thank people in the ,,Dehalo"-meetings: Guy Glod, Wolfram Schu- macher, Christoph Holliger, Werner Angst, Judith Perlinger and Rene Schwarzenbach. For the interesting presentations and discussions the people from the ,,ET" -meetings are kindly acknowledged: Stephan Hug, Silvio Canonica, Iganz Bilrge, Thomas Hofstetter, Werner Angst and Rene Schwarzenbach.

Numerous people assisted me in parts of my work. I would like to acknowledge Adrian Amman and Thomas Ri.ittimann for their help with the IC-measurements, Roland Hany for the NMR-measurements, John Westall for the discussions about the right use of FITEQL, Vesna Klingel for the contributions in her diploma work, Martin Schwarz for his help in the syntheses of the cyclopropanes and Christian Saxer and Etienne Michel for performing a number of kinetic measurements.

I would like to express my thanks to the Schwaba-crew for creating a pleasant wor- king atmosphere: Werner Angst, Cedric Arnold, Michael Berg, Thomas Bucheli, Andrea Ciani, Rick Devlin, Dieter Diem, Urs Domman, Martin Elsner, Beate Escher, Claudia Pesch, Guy Glod, Kai-Uwe Goss, Andreas Gerecke, Franca Gri.iebler, Edi Hohn, Stefan Haderlein, Thomas Hofstetter, Rene Hunziker, Jorg Klausen, Markus Meier, Ji.irg Mi.ihle- mann, Markus Millier, Stephan Muller, Klaus Pecher, Judith Perlinger, Csaba Reisinger, Kenny Weissmahr, Markus Ulrich, Thomas Waxweiler, Andre Weidenhaupt and Hans- ruedi Zweifel.

I also like to thank Verena Cajochen, Sabine Hilger, Donald Tillman and many others for the common joggings at ,,a quarter to twelve". Warm thanks to my colleages outside EAWAG: Andrea, Annie, Bettina, Birgit, Edmee, Felix, Gabi H. and Gabi A., Marcel, Susi, Ulrich and Ursi.

Finally, I would like to express my gratitude to my family for their love and support during all those years. Table of Contents List of Figures V List of Tables VIII Abbreviations X Summary XII Z usammenfassung XV

1. Introduction 1 2. Evaluation of the Iron Porphyrin/ Cysteine System 9 2.1 Introduction 9 2.2 Meso-tetrakis-(N-methyl-pyridyl)iron porphin 9 2.3 Kinetics of PHA reduction: determination of rate constants 12

2.4 Effect of pH on the reduction rates of CC14 and C2Cl6 14 2.4.1 Effect of pH on iron porphyrin with respect to redox potential and spin state 15 2.4.2 Effect of pH on cysteine 16 2.4.3 Buffers 21 2.5 Effect of cysteine and phosphate concentration on the reduction

rates of CC14 and C2Cl6 22 2.5.1 Effect of cysteine concentration 22 2.5.2 Effect of phosphate concentration 25 2.6 Ionic strength 28 2.7 Rereduction of the iron porphyrin 31 2.8 Blank reactions 34 2.8.1 Porphyrin ring as electron transfer mediator 35 2.8.2 Reactions without iron porphyrin 36 2.9 Summary and conclusions 38 3. Reaction Intermediates and Reaction Products 39 3.1 Introduction 39

3.2 Degradation of CC14: product analysis 41 II

3.2.1 Product distribution between different phases 43 3.2.2 Ion chromatographic analysis 45 3.2.3 Product analysis by NMR and MS 47

3.3 Degradation of CC14: reaction intermediates 50 3.3.1 Trapping of CC4-radicals 50 3.3.2 Trapping of carbenes 52

3.4 Degradation of CC14: proposed reaction mechanism in the iron porphyrin/ cysteine system 54

3.5 Degradation of CC14: proposed reaction mechanism in the reaction with cysteine (blank system) 57

3.6 Degradation of CBr4, CBr2Cl2, CBrC13: comparison with CC14 58 3.7 Degradation of polyhalogenated methanes containing one

hydrogen: CHBr3, CHBr2Cl, CHBrC12 63

3.8 Degradation of fluorinated methanes: CFBr31 CF2Br2, CFC13 66 3.9 Degradation of polyhalogenated ethanes 77 3.10 Summary and conclusions 84 4. Kinetics and Reaction Mechanisms. Structure-Reactivity Considerations 86 4.1 Introduction 86 4.2 Competition 86 4.3 Activation parameters 94 4.4 Comparison of degradation mechanisms of polyhalogenated methanes in different model systems 100 4.5 Structure- Reactivity Considerations 111 4.5.l Polyhalogenated methanes 111 4.5.2 Polyhalogenated ethanes 113 5. Conclusions 114 6. References 117 III

Appendix A A-1 Al Chemicals A-1 A.2 Syntheses A-2 A.2.1 Meso-tetrakis(N-methyl-pyridyl)iron porphin A-2 A.2.2 N-formylcysteine A-3 A.2.3 3-methyl-3-butenoic acid A-4 A.2.4 Diazomethane A-4 A.2.5 2-(2,2-dichloro-1-methyl cyclopropyl)-ethanoic acid A-5 A.2.6 2-(2,2-dibromo-1-methyl cyclopropyl)-ethanoic acid A-5 A.2.7 2-(2-bromo-2-chloro-1-methyl cyclopropyl)-ethanoic acid A-6 A.3 Stock solutions and buffers A-6 A.4 Experimental procedures: Kinetic experiments A-7 A.4.1 Standard procedure for kinetic experiments A-7 A.4.2 Determination of rate constants A-10 A.4.3 pH variation A-10 A.4.4 Variation of cysteine concentration A-11 A.4.5 Variation of phosphate concentration A-12 A.4.6 Variation of ionic strength A-12 A.4.7 Variation of ligand concentration A-13 A.4.8 Reactions without porphyrin A-14

A.4.9 Variation of initial concentration of CBr2Cl2 A-15

A.4.10 Competition experiments: CBr2Cl2 versus CBrCl3 A-17

A.4.11 Competition experiments: CC14, CHBrClv CHBr2Cl A-18

A.4.12 Competition experiments: CC14 versus 4-chloro- nitrobenzene A-19 A.4.13 Determination of activation parameters A-20 IV

A.5 Experimental procedures: Analysis of products and reaction intermediates A-21

14 A.5.1 Distribution experiments with C-CC14 A-21 13 A.5.2 C-CC14 experiments A-22 A.5.3 D' abstraction experiments A-22 A.5.4 Carbene trap experiments A-23 A.5.5 Quantification of products: methods of analyses A-26 A.5.6 Product studies of polyhalogenated ethanes A-26 A.6 Analytical prodedures A-27 A.6.1 GC-ECD A-27 A.6.2 GC-MS A-27 A.6.3 IC A-28 A.6.4 HPLC A-28 A.6.5 Scintillation counter A-29 A.6.6 NMR A-29

Appendix B B-1 B.1 UV-VIS experiments B-1 B.1.1 Cysteine-ligated iron porphyrin at different pH B-1 B.1.2 Effect of oxygen: Formation of superoxide radical B-2 B.1.3 Azide and cyanide as ligands B-3 B.1.4 Effect of cysteine concentration B-4 B.1.5 Effect of substrates B-4 B.1.6 UV-VIS spectrum of porphyrin ligand B-5 B.2 Polarographic measurements B-6 v

List of Figures

Figure 1.1: Worldwide production of some polyhalogenated alkanes (Fisher and Midgley, 1993i Rippen, 1988). 2 Figure 1.2: Half-reaction reduction potentials of organic redox couples (left side), various electron transfer mediators (middle) and some biogeochemically important redox couples (right side). Standard reduction potentials are indicated at environmentally relevant conditions: 25°C, pH 7.0, [Ci-)= [HCO;] = 10-3 M, [Br'] = lff' M. (Figure adopted from Schwarzenbach et al., 1997). 3 Figure 1.3: A mediator catalyses the electron transfer from the bulk electron donor to the pollutant. Typical bulk electron donors are iron(O), iron carbonate, hydrogen sulfide, hydrogen or iron sulfides. Mediators include surface- complexed iron, reduced cobalamins, reduced quinones or iron porphyrins. Organic pollutants that can be reduced are for example azo-dyes, nitroaromatic compounds or polyhalogenated alkanes (Figure adorpted from Schwarzenbach et al., 1997). 4 Figure 2.1: Meso-tetrakis(N-methyl-pyridyl)iron porphyin, an artificial water-soluble iron porphyrin. 10 Figure 2.2: UV-VIS spectrum of iron (III) porphyrin with water as axial ligands at pH 7. 10 Figure 2.3: .Kinetics of the reduction of lµM hexachloroethane (HCA) in the presence of 30 µM FeP at pH 6. The slope of linear regression analysis is -ko.,,. 13 Figure 2.4: Second- order rate constants for CCI, (D) and C,Cl6( •) normalized to pH 6. Error bars indicate one standard deviation from three replicates (see also Appendix A). 14 Figure 2.5: Reduction of iron (III) porphyrin by cysteine includes coordination (step 1), deprotonation of cysteine (step 2) and release of a cysteine radical (step 3). The electron transfer occurs immediately after deprotonation. 17 Figure 2.6: Speciation of iron porphyrin with cysteine as axial ligand(s) as a function of pH and cysteine concentration. 18 Figure 2.7: Shift of Soret band at pH 6 after addition of cysteine to an anaerobic iron (III) porphyrin solution. Dark line: Fe(III)P, bright line: Fe(II)P. 20 Figure 2.8: Shift of Soret band at pH 8.7 after addition of cysteine to an anaerobic iron (III) porphyrin solution. Dark line: Fe(IIl)P, bright line: Fe(II)P. 20 Figure 2.9: Second-order rate constants for CCI, (D) and C2Cl6 ( •) normalized to SmM cysteine. Error bars indicate one standard deviation from three replicates. 23 Figure 2.10: Ll.E (= E11 , w1thcysw1"" -E112wlthoutcys1e1,,.)/0.059V as a function of log [cysteine] 24 Figure 2.11: Second-order rate constants for the degradation of CCI, (D) and C2Cl6 (•) normalized to 50mM phosphate. Error bars indicate one standard deviation from three replicates. 26 Figure 2.12: Ll.E (= E112 withcystein• -E, 12 withoutcy"•'n•)/0.059V as a function of log [phosphate] 27 Figure 2.13: Second-order rate constants for C,Cl6 as a function of ionic strength. Error bars indicate one standard deviation from three replicates. 29 Figure 2.14: Apparent second-order rate constants as a function of the initial concentration of CBr,Cl2 with [FeP] 310.. M for all experiments. 31 VI

Figure 2.15: Concentration of CBr2Cl, versus time for five different initial concentrations: fitting of degradation curves with AQUASIM (Reichert, 1994). For initial concentrations of experiments 1 to 5 see caption of Figure 2.19. 33 Figure 2.16: Concentration of iron (II) porphyrin during degradation experiments of CBr,Cl,: depending on the initial concentration of CBr,Cl, the concentration of Fe(II)P can be quite small compared to the total concentration of FeP. Initial concentrations of CBr,Cl, were: exp. 1: l 10-6 M, exp. 2: 310.o M, exp. 3: 1 lff5 M, exp. 4: 3.5 HT5 M and exp.5: 5 10.oM. The concentration of Fe(II)P versus time was simulated with AQUASIM (Reichert, 1994). 34 Figure 2.17: Observed degradation rate constants for C,C!, in systems without iron porphyrin: influence of pH and cysteine concentration. 3 7 Figure 2.18: Observed degradation rate constants for CCI, in systems without iron porphyrin: influence of pH and cysteine concentration. 37 Figure 3.1: Postulated abiotic and biotic transformations of CCl4 (Criddle and McCarty, 1991) 42 Figure 3.2: Iron porpyhrin mediated reaction of 14CCI. : after degradation of CCI. 90% of the acitivity was found in the aqueous phase. This phase contained 90% of non-volatiles, 5% of volatiles and 5% of CO,. 44 Figure 3.3: Blank system: the product distribution is -50% volatiles (CHCI,) and -50% non-volatiles. 45 Figure 3.4: Ion chromatogram of the reaction solution after complete degradation of CCI,. The filled area contains the whole "C-activity. 46 Figure 3.5: 2-dimensional NMR of the 13C-labeled isolated product in D,O. The NMR was taken at 40°C. 47 Figure 3.6: Figure a shows an IC-chromatogram of the reaction solution. The product peak is marked black. Figure b shows the IC-chromatogram after addition of N-formylcysteine. The product peak increased. 4 9 Figure 3.7: Proposed mechanism of the formation of N-formylic compounds from dichlorocarbene (Frankel et al., 1959). 52 Figure 3.8: Olefins used for trapping carbenes: cis-/trans-3-pentenoic acid and 3- methyl-3-butenoic acid. 53 Figure 3.9: In the iron porpyhrin system the second electron is not transferred to the CCI, radical. 54 Figure 3.10: Proposed reaction mechanism of the degradation of CCI, in the iron porpyhrin mediated system: compounds framed by solid lines have been identified, compounds framed by broken lines are proposed. 55 Figure 3.11: Proposed reaction mechanism of the degradation of CCI, in the blank system. Compounds framed are isolated products or trapped reaction intermediates. 57 Figure 3.12: Proposed reaction mechanism for the degradation of CHX, in the iron porphyrin/ cysteine system. Compounds framed by solid lines were detected, compounds framed by broken lines are suggested reaction intermediates. 64 Figure 3.13: H-atom abstraction of CCI, and CF, radical from isopropyl alcohol (Landolt and Bornstein) 69 Figure 3.14: Possible reaction mechanisms in the degradation of fluorinated methanes. Species framed with solid lines were detected, species framed with broken lines are proposed. 71 VII

Figure 3.15: Reactions of radicals with olefins (Landolt and Bornstein) 72 Figure 3.16: Reaction of CFBr2 radical with the porphyrin ring: Possible pathway in the degradation of CFBr3• 73

Figure 3.17: SN2·reactions in the blank systems of a) CFBr3 and b) CF,Br2• 75 Figure 3.18: Proposed reaction pathway for the reduction of hexachloroethane in the iron porpyhrin mediated system. Two electrons are transferred successively. 82 Figure 4.1: Effect of substituents on relative rates of reduction of 10 different monosubstituted NBs. (a) Juglone /H,S system where mercaptojuglone is postulated to be the reactive species and (b) iron porphyrin/ cysteine system (Schwarzenbach et al., 1997; Schwarzenbach and Gschwend, 1990). 88 Figure 4.2: Second-order rate constants for 4-chloro-nitrobenzene (•),CCI, (D) and C2CI,, (•)normalized to pH 6. 93 Figure 4.3: Degradation rates of CCl4 as a function of the inverse temperature: (a) with 30µM iron porphyrin, (b) without iron porphyrin. Error bars indicate one standard deviation from three replicates. 97 Figure 4.4: Degradation rates of C,Cl, as a function of the inverse temperature: (a) with lOµM iron porphyrin, (b) without iron porphyrin. Error bars indicate one standard deviation from three replicates. 97 Figure 4.5: Mercaptojuglone in its proposed double function as nucleophile (~2- reaction) and as electron transfer mediator (ET-reaction). 105 Figure 4.6: Second-order rate constants relative to CCl4 in the iron porphyrin/ cysteine system (left hand side) and in the juglone system (right hand side) for a series of PHAs. 107

Figure 4.7: Log kJugkmc (•)and log k.,r (•)versus log k(CoW120,t). The slope of the regression line in the FeP system is 1, in the juglone system -1.6. 108 Figure 4.8: Structure of Co(II)W110,o"(Keggin structure). The Co(II) ion is situated in the center of the tetrahedron and the tungsten atoms are situated in the octahedron centers. Each comer of an octahedron represents an oxygen atom (Eberson and Ekstrom, 1988). 109 Figure 4.9: Completely reduced mercaptojuglone can react as a nucleophile or electron transfer mediator. As an electron transfer mediator, the electron transfer is outer-sphere, because only the semiquinone can act as an inner-sphere electron transfer mediator. 110 Figure 5.1: The reduction of PHAs in an anoxic aquifer containing reduced iron adsorbed to mineral surfaces and iron reducing bacteria can be coupled with the oxidation of organic pollutants. 114 VIII

List of Tables

Table 1.1: List of polyhalogenated alkanes investigated by Perlinger (Perlinger, 1994) 6 Table 2.1: Goods buffers used in the pH experiments 21 Table 2.2: Observed and second-order rate constants for the degradation of C2Cl6 in the presence of porphyrin ligand 35 Table 3.1: Products of polyhalogenated alkanes in different degradation processes 40 Table 3.2: List of experiments performed with radiolabeled product(s) of CCI, 43 Table 3.3: Portion of deuterated chloroform in the iron porphyrin/cysteine system as a function of the amount of d7-isopropyl alcohol present in solution 51 Table 3.4: Product distribution and reaction intermediates of CCI., CBr,, CBrCl, and CBr2Cl2 in the iron porphyrin/ cysteine system and in the blank system 59 Table 3.5: Rate of decay I protonation rate of cxyz· (Hine and Ehrenson, 1958) 62 Table 3.6: Product distribution and reaction intermediates in the degradation of CHBr3, CHBr,Cl and CHBrCI, 65 Table 3.7: Product distribution and reaction intermediates of CFBr3, CF,Br, and CFC!, in the iron porphyrin/ cysteine system and in the blank system (cysteine only) 68 Table 3.8: Bond dissociation energies in kJ mo1" 1 (Slayden, 1983) 76 Table 3.9: Polyhalogenated ethanes: degradation rates and reaction products in the iron porphyrin system and in the blank system, data from (Perlinger, 1994). 77 Table 3.10: Anions found in the degradation of CH3-CCI., CF,Cl-CFCI,, CF,Cl-CCl,, CF3-CC1, 79 Table 4.1: Competition experiments between CBr,Cl, and CBrCI,: apparent second- order rate constants 89 Table 4.2: Competition experiments with CCI., CHBr,Cl and CHBrCl, 90 Table 4.3: Competition experiments between two compound classes: polyhalogenated alkanes (here: CCI,) versus nitrobenzenes (here: 4-chloronitrobenzene) 92 Table 4.4: Activation parameters of different types of reactions 95 Table 4.5: Activation parameters in the iron porphyrin/ cysteine system and in the blank system 98

Table 4.6: Second-order reaction rates (kJug) and first-order reaction rates (kMecl= k0 ., - k.1..,. and kb1""') for the juglone/hydrogen sulfide system (Hofstetter, 1995; Perlinger, 1994) 101 Table 4.7: Percentage of free carbenes trapped as cyclopropane in the juglone mediated system 102 Table 4.8: Products and reaction intermediates in the juglone system 104 Table 4.9: Second-order rate constants for the reaction of halogenated methanes and ethanes with iron porphyrin (kFeP) and without iron porphyrin (kt,,, ... ) at 20°C. For kFeP the ± 95 % C.l. are given (Hofstetter, 1995; Perlinger, 1994) 112

Table A.1: Henry's law constants of polyhalogenated alkanes at 25°C A-9 Table A.2: Conditions in pH experiment A-11 Table A.3: Conditions in the experiments in which cysteine concentration was varied A-11 Table A.4: Conditions in the experiments in which phosphate concentration was varied A-12 Table A.5: Conditions in the experiments in which the ionic strength was varied A-12 Table A.6: Conditions in experiments in which ligand concentration was varied A-13 IX

Table A.7: Conditions for experiments without iron porphyrin A-14 Table A.8: Conditions in the experiments in which the regeneration rate of iron porphyrin was evaluated A-15 Table A.9: Definition of variables, processes and compartments for the fitting of the degradation rates as a function of the initial concentration of CBr,Cl, A-16 Table A.10: Conditions in competition experiments between CBr,Cl, and CBrCI, A-17 Table A.11: Conditions in competition experiments between CCI,, CHBrCI, and CHBr,Cl A-18 Table A.12: Conditions in competition experiments between carbon tetrachloride and 4- chloro-nitrobenzene A-19 Table A.13: Conditions in the experiments where activation parameters were determined A-20 Table A.14: Conditions in the D abstraction experiments A-23 Table A.15: Conditions for series 1 of carbene trap experiments A-24 Table A.16: Conditions for series 2 of the carbene trap experiments A-25 Table A.17: Conditions of the carbene trap experiments in the juglone system A-25 Table A.18: Conditions in the experiments with polyhalogenated ethanes A-26 Table B.1: Conditions in the polarographic measurements B-6

Table B.2: Table of components and species for the fitting of K 1 with FITEQL B-8 Table B.3: Table of components and species for the fitting of ~' with FITEQL B-10 Table B.4: Association constants of free aqueous iron and of iron porphyrin with various ligands B-12 x

Abbreviations

A. Angstrom (10-10 m) P2 association constant for a cation complexed by two ligands

4 1 F Faraday constant (9.6485 10 C mol- ) FeP iron porphyrin AG' Gibbs free energy of activation GC gas chromatography h Planck constant (6.626 10-34 Js) AH' enthalpy of activation HPLC high performance liquid chromatography I ionic strength k rate constant 1 rate constant for blank reaction (s- ) XI

second-order rate constant for a PHA in the iron porphyrin

1 1 system (M- s- ) 1 1 second-order rate constant for a PHA in the ligand system (M- s- ) 1 pseudo first-order rate constant (s- ) association constant for a cation complexed by one ligand A. wavelength (nm) M mol L-1 MS mass spectrometry NADH reduced form of nicotinamide adenine dinucleotide NOM natural organic matter it* it* orbital pKa negative decadic logarithm of acidity constant PHA polyhalogenated alkane PHM polyhalogenated methane 1 1 R gas constant (8.314 J K moI- ) R-s- cysteine anion RSSR cystine SN2 binary nudeophilic substitution reaction

AS~ entropy of activation a* cr*-orbital SHE standard hydrogen electrode T absolute temperature (K) TCA tricarboxylic acid u potential (V) UV-VIS ultra violet-visible spectroscopy x halogen charge of ion j XII

Summary

Polyhalogenated alkanes (PHAs) are a widespread class of anthro- pogenic pollutants in the environment. They are used as solvents, refriger- ants, propellants and pesticides. Due to use, improper handling, accidents and dump sites they may enter the environment by various pathways. In the atmosphere they contribute to ozone depletion and the green house effect. In the subsurface they are primarily of concern with respect to groundwater contamination. PHAs are persistent under oxic conditions. Under anoxic conditions, however, they can be degraded both in biotic as well as abiotic reactions. The most important transformation reaction is reductive dehalogenation because it is commonly faster than other reactions including hydrolysis, particularly for the highly halogenated PHAs. Reductive dehalogenation reactions of PHAs are interesting from two points of view. Such reactions may lead to products that are of considerable toxicological concern. On the other hand, products may be formed that are less harmful and/ or more easily degraded. In this work, reductive dehalogenation of PHAs has been studied in a model system containing cysteine as bulk reductant, a water-soluble iron porphyrin as electron transfer mediator and the PHAs as electron acceptor. Cysteine, a proteinogenic amino acid, is part of many biological molecules, e. g. glutathione, an antioxidant in the liver. Iron porphyrin is a model for organically complexed iron species. The major goals of this work were (i) the systematic evaluation and characterization of the model system using spectrophotometric and polarographic methods, (ii) the characterization of products and reaction XIII intermediates of various PHAs and (iii) the interpretation of rate constants with respect to structure and mechanism. The thesis is devided in three parts. Part one gives a description of the model system and shows the influence of system constituents on degradation rates. It is shown that pH, concentration of cysteine or phosphate have only a small influence on the rate of the electron transfer. The difference can be explained primarily by different speciation of the iron porphyrin. Iron porphyrin can be complexed by two ligands in its axial positions. These axial ligands are varied when the composition of the solu- tion is changed resulting in a slightly different redox potential of the iron porphyrin and thus in different rates of electron transfer. In the second part, reaction intermediates and final products of ten polyhalogenated methanes (PHMs) are presented. PHMs were found to be transformed quantitatively into a formyl group in the model system except for PHMs exhibiting fluorine atoms. In all cases, complete dehalogenation took place. Radicals were found to be reactive intermediates, however, no free carbenes could be trapped. The reaction of PHMs with cysteine (no iron porphyrin) yielded quite different results: It is postulated that in a nucleophilic substitution reaction at the halogen the PHMs were transformed to anions that were protonated resulting in less halogenated PHMs. These products are much less reactive than the parent compounds. In the third part of the work additional kinetic aspects to those of part one were investigated. No competition was observed between PHAs in mixtures. Thus, the interaction between the iron porphyrin and the PHA is assumed to be weak. Activation energies supported the assumption of a dissociative electron transfer. The results were compared with those of other model systems. It is proposed that iron porphyrin acts as an outer-sphere XIV one-electron donor, whereas another environmentally relevant electron transfer mediator such as mercaptojuglone shows a double functionality: it reacts as an outer-sphere one-electron donor and as a nucleophile in an SN2 reaction at the halogen. From a practical engineering point of view, the results of this study offer an interesting perspective for a fast complete dehalogenation of PHMs, by using very reactive iron species in the presence of organic compounds exhibiting reduced sulfur and nitrogen groups. PHMs can thus be transformed to harmless products. xv

Zusammenfassung

Polyhalogenierte Alkane (PHA) stellen eine weit verbreitete Klasse von Schadstoffen in der Umwelt dar. Sie werden u. a. als Losemittel verwendet, auch als Kiihlmittel, als Treibmittel in Spraydosen und als Pestizide. Durch ihren Gebrauch,

Es wird vorgeschlagen, class Eisenporphyrin in einem outer-sphere Ein- Elektronentransfer reagiert, wahrend eine aromatische Mercaptoverbindung sowohl als outer-sphere Ein-Elektronentransfer-Mediator als auch als Nukleophil in einer SN2-Reaktion am Halogen wirkt. Diese Resultate sind von einem praktischen Gesichtspunkt her wichtig: Zur Sanierung von kontaminierten Aquiferen und auch in technischen Systemen lassen sich reaktive Eisenspezies verwenden. In Gegenwart organischer Molekiile mit Thiol- und/oder Aminogruppen konnen PHM vollstandig dehalogeniert und zu harmlosen Prod ukten transformiert werden. 1 . Introduction

Polyhalogenated alkanes (PHAs) are a widespread class of anthropogenic pollutants. They are used as solvents, refrigerants, pesticides, propellants and in metal degreasing. Although worldwide production has decreased over the last fifteen years, a considerable amount of PHAs is still produced (Fisher and Midgley, 1993; Rippen, 1988) (Figure 1.1). PHAs are introduced into the environment by several sources including spills, leaking waste disposal sites and effluents from municipal and industrial waste treatment facilities. They are potentially dangerous for the environment. In the atomosphere they contribute to ozone depletion and green house effect, as they are volatile and quite inert (Nowell and Hoigne, 1992; Penkett, 1982; Tsani-Bazaca et al., 1982; von Diiszeln and Thiemann, 1985). In the soil and groundwater they are mobile, persistent under oxic conditions (Jafvert and Wolfe, 1987; McCarty et al., 1981; Roberts and Gschwend, 1994) and some of them are toxic when they are incorporated by organisms. Three types of toxic effects are known: depression of the central nervous system; kidney, liver and brain damage; and, finally, mutagenesis and cancerogenesis (Renschler, 1994). 2

Figure 1.1: Worldwide production of some polyhalogenated alkanes (Fisher and Midgley, 1993; Rippen, 1988).

As a variety of other important environmental pollutants (e.g. nitroaromatic compounds, azo compounds, sulfoxides and sulfones), PHAs may be reduced in the environment under anoxic conditions. They have been found to undergo reductive dehalogenation (Curtis and Reinhard, 1994), that is, they are reduced while losing their halogen(s). These abiotic and/ or microbially catalysed reactions may yield a variety of products that may or may not be of environmental concern. Presently, there is a considerable interest in such reactions because of their potential applicability in the treatment of wastes and in remediation approaches. Thus an important goal of investigations of reductive dehalogenation processes is to develop methods for converting PHAs to more easily degradable products or even directly to harmless products. 3

Pollutants ET-Mediators Bulk Reductants r:g(wl ~--O_x__ ····~_R_e_d____ O_x _ _,____R_ed_____ O_x_-r-_R_e_d _ __,Ej;

CHBr3

cc~ CHCtJ Cl2C•CCl2 Cit:•CHCI CHCl3 C'"2C~ 0.5 MnOi(s) MnCO:i(s) 0.5 cr042· (1Q-5M) Cr(OHb{s) @-N02 @-NH2

Cob(lll)alamln Cob(ll)alamln felli>orph FellPorpl1 0.0 ~i,Qk>n• - 0.0 000 Qi so{ HS- a-FeOOH(s) Fe2•(10-5M) S(sJ- Hs- • u-Fe2()3(s) Fe2•(10-5M) a.feOOH(s) FeCQi(s) Cl2C•CCl2 C!iC-CCI +Cl- R-S·S·R (cystlne) 2 R-SH (cystoine) Fe,04(s) FeiSI04(s) -0.5 @-No2 @-NOi' - HCQi- CHi(l --0.5 Fe1• (1Q-5M) Fe (s) Cob(ll)a~min Cob(l)alamln •)pH• 7; [HC00}•(Cl']• 1

Figure 1.2: Half-reaction reduction potentials of organic redox couples (left side), various electron transfer mediators (middle) and some biogeochemically important redox couples (right side). Standard reduction potentials are indicated at environmentally relevant conditions: 25'C, pH 7.0, 3 5 [0-] = [HC03·] = 10- M, [Br}= 10· M. (Figure adopted from Schwarzenbach et al., 1997).

When one considers abiotic reductions of organic pollutants in the environment, the most important bulk reductants are reduced sulfur- and iron-species. As PHAs have a higher redox potential than the major reductants in the environment (Schwarzenbach et al., 1993; Stumm, 1992), these redox processes are thermodynamically favoured (see Figure 1.2). However, reactions are often slow. It could be shown, that natural organic matter (Dunnivant et al., 1992) as well as quinones (Perlinger, 1994) and transition metal complexes (Gantzer and Wackett, 1991) can accelerate such reductive processes by orders of magnitudes. A fast transfer of electrons from 4 a bulk electron donor to an electron transfer mediator and then to a pollutant takes place (see Figure 1.3).

e.g. Ar-N=N-Ar e.g.----- CCl4 FeOOH(s) Cl:ciC=CCl2 H+ ArN02 Cro~- C02 8042- S(s) Oxidized Oxidzed

Organic Bulk Pollutant Mediator Electrm Donor

RedL.Ced Red.Jced e.g.----~ e.g.----~ Ctc=CHO {Cl-lO} HS- CHCe ArNH2 FeCQ:3(s) H2 Cr(OH):Js) Fe(s} FeS

Figure 1.3: A mediator catalyses the electron transfer from the bulk electron donor to the pollutant. Typical bulk electron donors are iron(O), iron carbonate, hydrogen sulfide, hydrogen or iron sulfides. Mediators include surface-complexed iron, reduced cobalamins, reduced quinones or iron porphyrins. Organic pollutants that can be reduced are for example azo-dyes, nitroaromatic compounds or polyhalogenated alkanes (Figure adopted from Schwarzenbach et al., 1997).

In order to study reductive dehalogenation reactions of PHAs different approaches are used. In field studies, PHAs are injected into the ground and sampled after movement through various reducing zones where transformation takes place (Christensen et al., 1994; Criddle et al., 1986; Jafvert and Wolfe, 1987; Kuhn and Suflita, 1989; Riigge, 1997; Riigge et al., 1995). Another approach is to collect aquifer material and put it in columns in order to investigate transport, interaction with the aquifer material and degradation reactions in a more defined system (Elsner, 1998; Heijman et al., 1993). Often, however, the differentiation between biotic and abiotic processes and the evaluation of the the important species in the reduction process is rather 5 difficult in such complex systems. Therefore a common and very useful approach is to carry out experiments in model systems. Here, the composition and the conditions are exactly adjustable. Model systems to study abiotic degradation reactions usually contain a pH-buffer, an electron transfer mediator, a bulk reductant and the pollutant. For example, Curtis et al. examined the reduction of C2Cl6 by a chemically reduced hydroquinone

(Curtis, 1992). The degradation of CC14 was studied in a model system containing iron-bearing minerals like biotite, vermiculite, pyrite and hydrogen sulfide (Kriegmann-King and Reinhard, 1992; Kriegmann-King and Reinhard, 1994). Other important studies in model systems containing iron porphyrins as electron transfer mediator were carried out by Gantzer et al., Klecka et al. and Wade et al. (Gantzer and Wackett, 1991; Klecka and Gonsior, 1984; Wade and Castro, 1973). The focus of such studies was, however, often the determination of degradation rates of a few compounds only. Furthermore, in most cases, rather little information was gained about reaction mechanisms, rate-limiting steps and, particularly, about product formation in the presence of different water constituents. In a previous study in our group the kinetics of degradation were determined for a large number of C1- and C2- polyhalogenated alkanes (see Table Ll) in a model system containing cysteine as bulk reductant and a water-soluble iron porphyrin as electron transfer mediator (Perlinger, 1994). These kinetic data were the starting point for the work presented in this thesis. There are several reasons for choosing iron porphyrin and cysteine in this model system: as for the iron porphyrin, it serves as a model for organically complexed iron. Iron is very important for living organisms, as it is part of many vitally necessary molecules, like hemoglobin for the transport of oxygen (Ando and Moro-Oka, 1988). As iron (II) can be easily transformed to iron (III) and vice versa, it plays a pivotal role in electron transfer processes. For example the respiratory chain contains cytochrome c, a protein with an iron porphyrin as prosthetic group, which is an essential mediator in the electron transport from NADH, a product of the TCA-cycle, to the final acceptor oxygen (Kaim and Schwederski, 1991; Moore and Pettigrew, 1990). 6

Iron is also found in enzymes like peroxidases, catalase, cytochrome c-oxidase, cytochrome P450 (a flavoprotein-oxygenase) and myoglobine proteins. All these proteins contain heme-groups as active parts for the electron transfer (Kaim and Schwederski, 1991).

Table 1.1: List of polyhalogenated alkanes investigated by Perlinger (Perlinger, 1994)

compound name formula usage (Randall and Baselt, 1993)

fluorotrichloromethane CFC13 propellant bromodichloromethane CHBrCl2 chemical intermediate tetrachloromethane CCl4 dry-cleaning chemical, degreasing agent tribromomethane CHBr3 chemical intermediate chlorodibromomethane CHBr2Cl chemical intermediate fluorotribromomethane CFBr3 chemical intermediate bromotrichloromethane CBrC13 chemical intermediate dibromodichloromethane CBr2Cl 2 chemical intermediate tetrabromomethane CBr4 chemical intermediate chloroform CHC13 solvent, chemical intermediate 1,1,2,2-tetrachloroethane CHC1 2-CHC12 solvent

1,1,1-trichloroethane CCl3-CH3 solvent

1,1,2-trichlorotrifluoroethane CC12F-CCIF2 propellant

1,1,1,2-tetrachloroethane CC!a-CC1H2 solvent

1,1,1-trichlorotrifluoroethane CC13-CF3 propellant pentachloroethane CHCl2-CCl3 chemical intermediate

1,1-difluorotetrachloroethane CF2Cl-CC13 propellant hexachloroethane CC13-CC13 chemical intermediate 7

Therefore it is not surprising that iron porphyrins can act as electron transfer mediators in redox processes involving pollutants as well. It could be shown that lindane, a halogenated alicyclic pesticide, can be degraded in systems containing hemoglobin or hemine (Marks et al., 1989). Other reductive transformations of organic pollutants catalysed by iron porphyrins include transformation of nitroaromatic compounds (Schwarzenbach et al., 1990; Wade et al., 1969) and azo compounds (Weber and Wolfe, 1987). Cysteine was chosen as bulk reductant for several reasons: (i) it is a good reductant towards iron porphyrin. and (ii) it also reacts directly with PHAs, although at a much slower rate. Nevertheless, it can be used as reductant representative for -SH groups in natural systems. Cysteine is a proteinogenic amino acid and part of many biological molecules like glutathione, an antioxidant in the liver (Ketterer and Mulder, 1990). Finally, it contains different functional groups (-SH, -NH21 COOH) that can undergo various reactions with reactive intermediates in the degradation of PHAs. Thus, it represents biological molecules as well as NOM1 under sulfate reducing conditions. In order to elucidate the electron transfer mechanism in this model system and to get information about the fate of PHAs after initial one- electron transfer, the major goals of this thesis were

• to provide a systematic evaluation and characterization of the model system • to determine the reaction intermediates and final products of PHAs • to propose degradation mechanisms as a function of different water constituents • to interpret degradation rates of PHAs with respect to structure

1 Natural Organic Matter. 8

In the first part of the thesis (chaper 2) the iron porphyrin/ cysteine system is evaluated. The major goal was to extend the present knowledge about how system parameters influence the rate of electron transfer and which iron porphyrin species are essentially involved in the electron transfer.

To this end, reaction rates of two representative PHAs, namely, CC14 and

C2Cl6, were determined with variation of pH, buffers, cysteine concentration, phosphate concentration, starting concentration of the polyhalogenated compound and ionic strength. Moreover, to get information about the speciation of the iron porphyrin, spectrophotometry and polarography were used. In the second part of this work (chapter 3), the major goal was to identify reaction intermediates and products in order to propose a reaction mechanism. First, the degradation pathway of carbon tetrachloride was stud- 13 14 ied in detail using CCl4 and CC14• Second, the intermediates and products of nine other polyhalogenated methanes and four ethanes were measured. A third and final part of the work (chapter 4) was dedicated to kinetics, especially to activation parameters and to competition experiments between several PHAs. Kinetic. data of other electron transfer systems were compared with the results obtained in this study. Structure- reactivity considerations were made for a large number of PHAs. All experimental details are summarized in Appendix A and B. 9

2. Evaluation of the Iron Porphyrin/ Cysteine System

2.1 Introduction

In this chapter a detailed description of the model system is given. First, the characteristics of the artificial iron porphyrin are presented. Secondly, the influence of important system parameters on the degradation rate of CC14 and C2Cl6 are shown. The influence of system parameters like pH, cysteine concentration and phosphate concentration are interpreted in the light of data obtained from spectrophotochemical and polarographic measurements.

2.2 Meso-tetrakis-(N-methyl-pyridyl)iron porphin

Meso-tetrakis-(N-methyl-pyridyl)-iron porphyin (FeP) is an artificial iron porphyrin. It is synthesized by incorporating dissolved iron (II) in the porphyrin ring (see Appendix A). The four N-methyl-pyridyl groups in the methine-positions protect the ring from being attacked by nucleophiles and make the porphyrin water-soluble as they are positively charged (Figure 2.1). This is important for studies of electron transfer reactions in homogenous aqueous solution.

The aromatic 18-electron system shows two 1t __,. 1t * transitions that can be seen in UV-VIS spectra. At pH 7 a rather weak absorption with Am-.=

1 1 1 598 nm (E 9600 M· cm· ) and an intense absorption at 420 nm (E ~·100'000 M·

1 cm' ), called Q-band and Soret-band respectively, characterize the UV-VIS spectrum of the iron (III) porphyrin (Figure 2.2) (Schoder, 1975). 10

c±l H,C--f-1, \~\-~·

H H

H

~·::>··· ..... ! j@ H3C

Figure 2.1: Meso-tetrakis(N-methyl-pyridyl)iron porphyin, an artificial water- soluble iron porphyrin.

<1.l ~ 0.8 ~

200 300 400 500 600

Wavelength (run)

Figure 2.2: UV-VIS spectrum of iron (III) porphyrin with water as axial ligands at pH 7. 11

The standard redox potential of meso-Tetrakis-(N-methyl- pyridyl)iron porphin is 0.171 V at pH = 0. The redox potential is pH- dependent:

62 105 E0' =0.171+0.062·lo {[Fe(III)P]}+o.062·lo { 1. ' ·[W] } 2 h g [Fe(II)P] g 1+1.62 ·10 5 · [W] .1

at 25°C and IM NaN03 (Schoder, 1975). Thus, at pH 7 the redox potential is 0.065 V. The redox potential is also dependent on the ligands coordinated to the iron porphyrin in its axial positions. Ligands as cysteine or phosphate increase the redox potential to higher values compared to water or hydroxide respectively (see chapters 2.5.1 and 2.5.2). The pK.s of axially coordinated water molecules are 5.21 for iron(III) porphyrin and >10 for iron(II) porphyrin respectively (Schoder, 1975). Thus, at pH 7, reduced iron porphyrin has two water molecules in its axial positions. After oxidation, one water-ligand is deprotonated. The effect of pH and cysteine concentration on the speciation of the iron porphyrin is shown by UV-VIS spectra in the sections where the influence of these parameters on the degradation rate of CCl4 and C2Cl6 is presented (chapters 2.4 and 2.5). Additional UV-VIS experiments are given in Appendix B (effect of oxygen, superoxide radical anion; azide, cyanide and PHAs as ligands). 12

2.3 Kinetics of PHA reduction: determination of rate constants

In the model system containing cysteine as bulk reductant, iron porphyrin (FeP) as a mediator and a PHA as oxidant, the degradation rate of the oxidant can be expressed by:

d[PHAJ . ----=k ·[PHA]' ·[FeP]Y[cysteme]" 2.2 dt

where x, y, and z are reaction orders of PHA, FeP and cysteine respectively. As the direct reduction of PHA by cysteine is very slow compared to the reduction by FeP and as samples were equilibrated at least 12 hours before spiking the substrate (ensuring that all iron porphyrin was in the reduced form), the only influence of cysteine on the reaction rate is by ligation of FeP and thus varying the redox potential. This influence is presented in chapter 2.5. As all standard experiments were performed at constant cysteine concentration that were much higher than the concentration of FeP or substrate, [cysteine]z is integrated in the rate constant k':

- dPHA = k' {PHA]' ·[FeP]Y 2.3 dt

Actually, the dependence of the observed rate constant on iron porphyrin concentration for the reduction of CC14 was linear, thus giving y = 1 (Perlinger, 1994). The order of the reduction reaction for the PHA was determined by the half-life method (Laidler, 1987) and gave a linear dependence as well with a slope x = 0.97 ± 0.15 for CC14 (Perlinger, 1994). Thus, equation 2.3 can be rewritten: 13

dPHA , -~ = k ·[PHAJ·[FeP] kobs ·[PHA] 2.4

where in equation 2.4 the pseudo-first order rate constant, k0 b•' is given by:

2.5

and kFeP is the second-order rate constant of the PHA. Thus, as is illustrated in Figure 2.3 for hexachloroethane the observed rate constant can be determined

) by linear regression from plots with In ([PHA]/[PHAJ0 versus time.

-0.4

;::o -0.8 1 u ;;"" -1.2 -.... ~- -1.6 µ;:i u - 2 ~ .s -2.4 -2.8

-3.2 0 50 100 150 200

time (minutes)

Figure 2.3: Kinetics of the reduction of 1µM hexachloroethane (HCA) in the presence of 30 µM FeP at pH 6. The slope of linear regression analysis is -k00,. 14

2.4 Effect of pH on the reduction rates of CC1 4 and C2Cl 6

The effect of pH on the reaction rates of CC14 and C2Cl6 was studied using different weakly coordinating buffers: pH 6: MES; pH 7: HEPES; pH 8: HEPPS, pH 8.7: TAPS and pH 9.3: CHES (for abbreviations and exact experimental conditions see Appendix A). Figure 2.4 shows that the relative pH-dependence of kFeP was the same for CCl4 and C2Cl6 • In the range from pH 6 to 8.7 rates increased linearly by a factor of about 5 and then reached a plateau (pH 9.3). As the same dependence is seen for a typical representative of C1- and C2- polyhalogenated alkanes

(CCl4 and C2Cl6 respectively), one can exclude different substrate-specific interactions between FeP and the two substrates. If there are any interactions at all, they are the same for CC14 and C2Cl6• The reason for the observed effect has therefore to be found in the solution composition providing different speciation of the iron porphyrin, cysteine and buffer at different pH-values.

6

5 coII 4 ::r::_e,. ~ 3 ..:.:... '- 2 t ~ ..:.:... t

0 6 6.5 7 7.5 8 8.5 9 9.5

pH

Figure 2.4: Second- order rate constants for CC1 4 (D) and C2ClJ •) normalized to pH 6. Error bars indicate one standard deviation from three replicates (see also Appendix A). 15

2.4.1 Effect of pH on iron porphyrln with respect to redox potential and spin state

The redox potential of FeP with two water molecules in its axial positions is pH-dependent (equation 2.1). Reduction at various pH-values requires a proton. At high pH-values protonation of the reduced iron porphyrin becomes difficult. Iron (II) porphyrin with a hydroxide ligand is a better reductant than iron (II) porphyrin ligated by water molecules (where the oxygen atoms of the water molecules are formally positively charged and therefore withdraw charge from the iron center). Thus, the higher the pH, the lower the reduction potential. This pH effect on the redox potential could explain the reduction rate dependence for water-ligated iron porphyrin. As we have a much stronger ligand (cysteine) than water in the solution, however, the pH effect discussed above loses its importance, because cysteine displaces water in the axial positions (see chapter 2.5). However, the effect of pH on the redox potential of cysteine-ligated iron porphyrin is not known. Another important aspect concerning the reduction rate is the spin- state of the iron-center. It is known that electron transfer reactions are faster for low-spin Fe(Il)/Fe(III) than for high-spin Fe(Il)/Fe(III): the oxidation of low-spin cytochrome c is 100 times faster than the oxidation of high-spin hemoglobin, both containing heme as the active center (Kadish and Davis, 1973). pH does not influence the spin-state of the iron center in our system. Iron porphyrins coordinated by six ligands are always in the low-spin state, in both oxidation states +II or +III, with the exception of ligation by very weak ligands such as tetrahydrofuran (Lever and Gray, 1983). Iron porphyrins coordianted by five ligands are only low-spin when strong ligands (strongly electron-donating) are present (Lever and Gray, 1983). 16 As cysteine coordinating the iron center by its thiol-group is a strong ligand (see chapter 2.5) and as meso-tetrakis(N-methyl-pyridyl)iron porphin is always coordinated by two axial ligands (six ligands on the whole) (Schoder, 1975), the iron center of the porphyrin is assumed to be in the low-spin state over the whole pH range. Thus, in this system pH does not influence the spin state of the iron porphyrin in the pH range considered.

2.4.2 Effect of pH on cysteine

Cysteine, a proteinogenic amino acid, contains a thiol functional group with a pK. of 8.33 (Smith and Martell, 1989). In the model system studied, cysteine plays two roles: it serves as a reductant for the iron porphyrin and it is also an axial ligand at the iron center. The reduction of the iron porphyrin by cysteine can be split up into several steps (Figure 2.5): at pH > pK. (thiol group of cysteine), the reduction process of iron porphyrin can be assumed to be faster, because a deprotonated cysteine can reduce the iron center directly after its coordination. With mainly protonated cysteine present (pH< pK.), the coordinated cysteine has to be deprotonated (step 2) before the electron transfer. The higher the pH, the faster the reduction of the iron porphyrin by cysteine. In the model system studied, cysteine was present in high concentrations compared to iron porphyrin (SmM versus 30µM). As all samples were equilibrated for at least 12 hours before starting the reaction, all iron porphyrin was in the reduced form at the start of the reaction. This is confirmed by UV-VIS spectra. Moreover, the ratio of iron porphyrin to substrate was 30:1, so rereduction of the mediator did not play an important role. 17

+ + H3NrCOO HsNrCOO + H-r 9H2 +Ill Ill ste_E 1 ste1:: 2 i Ill ---~e------1e--- coordination.. ---1e--- -H+ .. OH by OH OH I + II III H 3yoo HS l + H;iNrCOO + + 9H2 step3 +i II H;iNfOO + ---FelL-- ---ye--- s I +H20,+W OH2 OH + v IV

Figure 2.5: Reduction of iron (III) porphyrin by cysteine includes coordination (step 1), deprotonation of cysteine (step 2) and release of a cysteine radical (step 3). The electron transfer occurs immediately after deprotonation.

Considering these facts, increasing pH accelerates the reduction of the iron porpyhrin. As the standard experimental conditions provide all iron porphyrin in the reduced form at the start of the reaction and as they exclude the rate of rereduction from kobs' the influence of pH on cysteine as reductant can not be the cause for the observed pH-dependence. On the other hand, pH has an influence on the spedation of cysteine as an axial ligand. Cysteine acts as ligand only at the reduced iron porphyrin, since at the oxidized form it acts as a reductant and is released as a cysteine radical (see Figure 2.5). Five different cysteine complexes with iron (II) porphyrin are possible depending on pH and cysteine concentration (see also chapter 2.5): 18

pH ,....cys ,....cys + Cys-s· ---Fe--··111 ---F.t1---1 o/'5 I -Cys-s· I + OH2 - 5 ~~'if +Cys-SH ---Fe--··I II ---pelL--r M: I - Cys-SH I S>Hz /s, H + Cys VII IX cysteine concentration

Figure 2.6: Speciation of iron porphyrin with cysteine as axial ligand(s) as a function of pH and cysteine concentration.

At very low cysteine concentrations ([cysteine] ~ {FeP]) species V with two water molecules as axial ligands is present (Figure 2.6). The pK, of water at the reduced iron center is > 10 (Schoder, 1975). Water has a pK. of 15.8

(Atkins, 1996). Thus, coordination at the iron (II) center lowers the pK, by., $; 6 units. Analogously, for cysteine with a pK. of 8.33 a cysteine coordinated to the iron center is thought to have a pK. of > 2.5 (involving species VI and VII). The pK. of a second coordinated cysteine would be higher, because one deprotonated cysteine vis-a-vis increases the electron density at the iron center. This enhanced electron density stabilizes the formally positive sulfur of the second cysteine (species X). The pK. of the second cysteine (acid-base pair X and VIII) is therefore estimated to be around 7. 19 At high cysteine:FeP ratios (as given under standard experimental conditions) two cysteine ligands are very probable (see also chapter 2.5). Hence, at high cysteine concentrations and different pH-values species VIII, IX and X are to be expected. With the estimated pK.s of> 4 for one and around 7 for a second cysteine at the iron center the observed increase in rate with increasing pH can be explained: at pH 6 and 7 species X is dominant, whereas at higher pH-values (pH > 8) species VIII is present. Species VIII has a higher electron density at the iron center and should thus accelerate the electron transfer. UV-VIS spectra show that at pH> 8 different species coexist (Figure 2.7 and Figure 2.8). As can be seen in Figure 2.7 the addition of cysteine to an anaerobic iron (III) porphyrin solution at pH 6 leads to a shift of the Soret band to higher wavelengths (420 nm to 443 nm) keeping the form of the band. This is also found for pH 7. At pH 8, 8.7 and 9.3, however, the shift is accompanied by a clear broadening of the band (Figure 2.8). The shift of the Soret band occurs due to ligation of the iron center by cysteine. Ligand field theory states that a ligand change in the order Cl --> S --> 0 -> N --> C increases the energy gap between t28 and e/ for an octahedral geometry of six ligands (Huheey, 1980). An increase in energy gap goes hand in hand with a decrease in wavelength of maximal absorption, t..m.x· A change from 0 to S (in our system represented by a change from H20 to cysteine as axial ligand) should give a shift of A.max to higher wavelengths. On the other hand, the broadening of the band at higher pH-values is due to formation of other 3 species • It implies that there are more species cooccurring than at pH 6 and 7. This corresponds with a pK. of:;;: 7 for a second cysteine ligand.

2 In an octahedral ligand field, the five d-orbitals of the metal center are distributed as follows; d,,_,2 and da in e8 and dxr d:i:~, and dyi: in t211:

3 Slight broadenings of absorptions are also known to be a result of the so called Jahn-Teller effect If ener- gies of two d-orbitals are first on the same level and then a new axial ligand .induces a slight energy change, 20

0.8

~ 0.4 -e0 "' < 0.0

200 300 400 500 600 Wavelength (nm)

Figure 2.7: Shift of Soret band at pH 6 after addition of cysteine to an anaerobic iron (III) porphyrin solution. Dark line: Fe(III)P, bright line: Fe(II) P.

~ 0.8 -e0 <"' 0.4 0.0

200 300 400 500 600 Wavelength (nm)

Figure 2.8: Shift of Soret band at pH 8.7 after addition of cysteine to an anaerobic iron (III) porphyrin solution. Dark line: Fe(III)P, bright line: Fe(II)P.

the UV-VIS spectrum shows two maxima with only slight differences in wavelength. If the two maxima are very near, a broadening of the band results. 21 The observed increase in rate with increasing pH can therefore be plausibly explained by different speciation of the iron porphyrin at different pH-values. As can be seen in Figure 2.4 such different species do not, however, influence the rate of electron transfer in a drastic way (factor of 5).

2.4.3 Buffers

In order to prevent the iron porphyrin from forming complexes with the buffers which would have an influence on its redox properties, weakly coordinating buffers were used. All buffers contained a sulfonyl functional group as the active acid-base center, but no other acidic functional groups. The buffers used are shown in Table 2.1.

Table 2.1: Goods buffers used in the pH experiments

pH Name of buffer Structure pK. (for full name see Appendix A: A.1)

6 MES o..rso,H 6.15

,.-. 7 HEPES Ho__..,--1\._,N -.._S03H 7.55

8 HEP PS ,-.. 8.00 H~ 1'\__)'l""""-..,SO;H

8.7 TAPS HOr-f.~--SO,H 8.40 OH

9.3 CHES HN\_SOH 9.55 03 22 The deprotonated sulfonate group is not a very good ligand for the iron (II) center: sulfonate, in its deprotonated form is a hard base, whereas the iron (m center is soft (Kaim and Schwederski, 1991). In the presence of cysteine, sulfonate should therefore not form stable iron (II) complexes. As a matter of fact, the sulfonate group can be compared with phosphate: similar ligand formation capability is expected (see chapter 2.5.2). The buffers used at different pH-values should therefore not influence the rate of electron transfer by any way.

2.5 Effect of cysteine and phosphate concentration on the

reduction rates of CC1 4 and C2Cl5

Different axial ligands at the iron porphyrin do influence the electron transfer rate by two effects: first, they affect the redox potential by their elec- tron-withdrawing or electron-donating properties. Secondly, they may hinder or promote the approach of the substrate to the iron center by their steric properties. In order to understand more about these effects, reduction rates of

CC14 and C2Cl6 were measured as a function of cysteine concentration in the range from 3 10·5 M to 2 10"2 M and as a function of phosphate concentration in the range from 5 104 M to 5 10·2 M with all other conditions remaining constant (pH 7, concentration of iron porphyrin 30 µM, concentration of substrate= 1 µM, see also Appendix A).

2.5.1 Effect of cysteine concentration

Figure 2.9 shows the dependence of the second-order rate constants of

CC14 and C2Cl6 on cysteine concentration. Again, as for different pH values, for both substrates the same dependence is found. This supports the view that both molecules react in the same way as has been found in chapter 2.4. 23

At very low cysteine concentrations, an increase in rate by a about a factor of 3 was observed. From 6 10-• M to 5 10·3 M (at 5 10·3 M standard experiments were made) a decrease in rate occured, and an increase again at higher cysteine concentrations. This rather complex dependence on cysteine concentration can be explained by the formation of different species at different cysteine concentrations.

Q.? 6 ...... ~ .....ill 5 rfJ >.. u 4 f ~ s + ~ 3 p.. !j QJ ~ "' j ...... : : t p.. t QJ 0 ~"' 0 5 10 15 20 cysteine concentration (mM)

Figure 2.9: Second-order rate constants for CCl 4 (D) and C2Cl 6 (•) normalized to SmM cysteine. Error bars indicate one standard deviation from three repli- cates.

At pH 7, at which standard experiments were performed, cysteine is mainly protonated (pK. = 8.33). At low cysteine concentrations, only one cysteine coordinates to the iron center and is immediately deprotonated after its coordination (pK. > 4, see Figure 2.6, species VI). With polarographic measurements it could be shown that at least up to a ratio of cysteine:FeP of 8:1 (at higher ratios half-wave potentials were unstable) only one cysteine coordinates to FeP as the slope in the ~E versus log[cysteine] diagram (Figure 2.10) approximates a value of --0.06, which corresponds to a single coordination (Clark et al., 1940). 24

-0.02

-0.2S >0\ U') ~ ·0.49 0 r.:t:I ' ·0.73 <1

·Cl96

·U!O ·4.00 log [cysteine J (M)

Figure 2.10: LIE (= with cystetne -E112 without cystetn,)/0.059V as a function of log [cysteine]

Clark et al. (Clark et al., 1940) provide the following relation between AE and the amount of base added for metalloporphyrins:

dE 0.06-r 2.6 dlog(Sb -S)

' where dE Eh- E0 Sb total of base added S concentration of base at E; (here: S 0) r number of ligands (whole number).

Thus, based on the experimental slope of - -0.06 it can be concluded that only one cysteine is bound to the iron (II) center at low cysteine concentrations. This coordination of one cysteine to the iron (II) porphyrin has an association constant of 1085 (estimated with the program FITEQL, see 4 Appendix B) • Although the redox potential of the iron porphyrin increases with increasing cysteine concentration in the range considered (ratio cysteine:FeP = 1:1 to 8:1), an increase in rate was observed (factor 3, see Figure

4 For association constants of various ligands with free aqueous iron (II) and (III) and with iron porphyrin see Table B.4 in Appendix B. 25

2.9). The maximum was reached at [cysteine] = 6 104 M. At higher concentrations the rate decreased. In this range, a second cysteine most probably coordinates to the iron porphyrin and remains protonated (see chapter 2.4, Figure 2.6, species X). As the two axial cysteine molecules are rather bulky, they hinder the electron transfer to substrates. Therefore, the rate decreases up to a cysteine concentration of 5 10'' M. At higher cysteine concentrations, however, a new species is formed, as with increasing cysteine concentration also the amount of deprotonated cysteine increases providing a strong base that can actually deprotonate the second cysteine ligand. So, at higher cysteine concentrations species VIII (Figure 2.6) is dominant. This spe- cies accelerates the electron transfer due to the increase of the electron density at the iron (II) center rationalizing the increase in rate at [cysteine] > 5 mM.

2.5.2 Effect of phosphate concentration

Standard experiments were performed at pH 7 buffered with phosphate (pK01 = 2.1 , pK02 = 7.21, pK,3 11.9, (Sigg and Stumm, 1996)). As

H 2P04-, the dominant species of phosphate at pH 7, and HPOt could be potential ligands for iron porphyrin, the rate dependence of the phosphate concentration was measured for CC14 and C2Cl6• As can be seen in Figure 2.11, increasing phosphate concentration leads to a decrease in degradation rates of CC14 and C2Cl6 • In the range from 5 104 M to 5 10-2 M phosphate, rates decreased by a factor of 2. Thus, phosphate acts apparently as an axial ligand. It diminishes the reactivity of iron (II) porphyrin towards polyhalogenated alkanes. 26

-.!!:! ] 3 p... ffJ 0 ...c: 2.5 p... ~ 2 0 U") 1.5 -I'. ~(;. 1 ...____ ~ 0.5 "-----'-----'----'-----'------"----' ~"' 0 10 20 30 40 50 phosphate concentration (mM)

Figure 2.11: Second-order rate constants for the degradation of CCl 4 (D) and

C2Cl6 (•) normalized to 50mM phosphate. Error bars indicate one standard deviation from three replicates.

Polarographic measurements show that phosphate increases the re- dox potential of the iron porphyrin (Figure 2.12). In the range from from 5 10·4 M to 5 10"2 M phosphate the redox potential increases by 90m V providing a slope of 0.1 V /log phosphate which corresponds to two phosphate molecules at the iron porphyrin (Clark et al., 1940). With FITEQL, an association constant for two phosphate molecules at the iron (II) porphyrin was esti- mated: log p2 = 5.42 ± 0.06 (see Appendix B). Thus, compared to cysteine, phosphate is a rather weak ligand. 27

-0.60 > c::J It) °'0 ·UO 0 '- i:i:i

·2.10 '----~---~---~--~---~ ·3.50 ·3.00 ·2.50 ·2.00 ·150 ·100

log [phosphate] (M)

Figu.re 2.12: L1E (= with cyste1ne -E112 without cys1e1n)/0.059V as a function of log [phosphate]

The observed influence on the reaction rate can be explained by steric hindrance. As phosphate ligands are rather bulky compared to water the access to the iron center for substrates becomes difficult. Summarizing these facts, cysteine is a good ligand for iron (II) porpyhrin. At low cysteine concentrations, one cysteine molecule is axially bound to the iron porphyrin increasing the reactivity. With higher cysteine concentrations a second (protonated) cysteine is bound to the iron center diminishing the reactivity. At concentrations higher than 5mM, the second cysteine is deprotonated due to the presence of increasing free deprotonated cysteine resulting again in an increase in reactivity. Phosphate (RPO/' at pH 7) is a rather weak ligand towards the iron (II) porpyhrin. It decreases the reactivity of the iron center as an electron donor because of steric hindrance. Compared to cysteine, its electron-donating properties are weaker. 28

2.6 Ionic strength

The ionic strength of a solution depends on the concentration and the charge of the ions present:

I=_!" c · z 2 2.7 2~ l l

where ionic strength cl = concentration of ion j zJ =charge of ion j

For ionic reactions between ions A and B, the rate depends on the ionic strength as follows (Laidler, 1987):

2.8

For reactions between non-charged species, equation 2.8 predicts no dependence on ionic strength (zA or zB or both being zero). In our model system, the electron transfer takes place between a charged species (actually the iron (II) porphyrin has an overall charge of +VI) and a neutral molecule (for example C2Cl6). Figure 2.13 shows the influence of ionic strength on the degradation rate of hexachloroethane. The ionic strength was varied with Na2S04• Sulfate is a poor ligand to the iron (II) center (Kaim and Schwederski, 1991). As its charge is -2 there is a large effect on the ionic strength of the solution (see equation 2.7). 29

50 u "' -N u,_, 40 ~ ,.!{"" 30 + (/} • f '::E 20

!l 10 ~';;'.

0 0 20 40 60 80 100 120 140 Ionic Strength (mM)

Figure 2.13: Second-order rate constants for C2Cl6 as a function of ionic strength. Error bars indicate one standard deviation from three replicates.

As is evident, the ionic strength did not have any significant influence on the electron transfer rate from iron (II) porphyrin to hexachloroethane, even at high sulfate concentrations. An explanation can be given by looking at the electron tranfer on a microscopic level. Taube defines an outer-sphere electron transfer between a primary bonding system and a second bonding system, whereas an inner- sphere electron transfer takes place within one primary bonding system (Eberson, 1982). Since UV-VIS spectra did not show any shift of the Soret band while adding CC14 or C2Cl6 to an anaerobic iron (II) porpyhrin solution (Appendix B) and since there was no competition effect when adding

CHBr2Cl, CHBrC12 and CC14 simultaneously to the reaction solution (see chapter 4.2), one can assume that polyhalogenated alkanes do not build up bonds to the iron (II) center. Thus, in an outer-sphere electron transfer, the electron has to overcome a distance between donor and acceptor. Such a distance can be up to 7 A (Eberson, 1987). Compared to a single C-C bond with a length of 1.5 A, the 30 electron transferred can actually "travel" quite a long distance. On its way to the acceptor the electron has to get in contact with the solution (water). From pulse radiolysis experiments (Buhler, 1983), however, it is known, that free aqueous electrons react at near diffusions-controlled rates with water to give hydroxy-radicals. Therefore, the electron seems to reach the acceptor without any contact with water. This could be explained by a quantum mechanical tunnel effect. Estimation of the wavelength of the electron transferred between C2Cl6 and iron (II) porphyrin in a potential gradient of 0.26V (C2Cl6: 5 E0 = 0.33V, iron (II) porpyhrin: E0 = 0.071V ) is difficult and beyond the scope of this dissertation. In a simple system of a plate condenser with the same potential gradient an estimation of the wavelength gives 24 A6, which is much higher than 7A. Butler and Hayes (Butler & Hayes, 1998) did also not find any significant effect on the rate of C2Cl6 reduction by FeS, when varying the ionic strength from 0.06 to O.lM, which is in agreement with our findings. A tunnel effect is a possible explanation for the observation that electrons are transferred from the iron (II) center to the polyhalogenated alkane without any noticeable reaction with water in an outer-sphere electron transfer reaction.

5 This is the E,-value for iron porphyrin complexed by water. As cysteine ligands increase the redox potential (see Figure 2.10), the corresponding E,-value would be even higher for the cysteine-ligated iron porphyrin.

6 ,_ = h 2.9 ~2me·e·U where h= Planck constant mass of electron charge of electron potential difference Inserting the corresponding values, we get: 31

2. 7 Rereduction of the iron porphyrin

The rereduction rate of the mediator may influence the overall reaction rate, especially in the case of quickly degrading substrates and in competition experiments (see chapter 4) where the concentration of the substrates are higher than the concentration of the mediator. Thus, the rereduction rate of the iron porpyhrin by cysteine was determined. The concentration of FeP was 3 10-6 M and constant in all experiments.

Dibromodichloromethane (CBr2Cl2), a fast reacting substrate (see chapter 4.5), was used. Its initial concentration was varied over a range between 1 and 50 µM. The apparent second-order rate constants as a function of the initial concentration of CBr2Cl2 are shown in Figure 2.14:

600 """',_.rJ'J' ~ 500 '-" .8 400 l ;;:;" i [FePJ =no·• M " 300 • ...ltj:a 1:l 200 ....(I) ro 100 0.. ~ 0 ·' • • • 0 110·5 210- 5 -310·5 410- 5 s 10·5 610- 5 initial concentration of CBr C1 (M) 2 2

Figure 2.14: Apparent second-order rate constants as a function of the initial 6 concentration of CBr2Cl 2 with [FeP] = 3 10- M for all experiments.

Apparent second-order rate constants for the degradation of CBr2Cl2 were much higher for [substratelo < [FeP] than for [substrateL > [FeP]. For

[substrate]0 > 30µM the apparent second-order rate constants did not vary 32 anymore indicating that the regeneration of the iron (TI) porphyrin was rate- limiting. Initial concentrations were calculated by taking the intercept of a

linear regression in a {logc/c0 }versus time plot, because concentrations in the samples taken first were smaller than the exact inititial concentrations due to fast degradation (see Appendix A). The regeneration rate of iron (II) porphyrin was calculated by fitting the experimental data with the program AQUASIM (Reichert, 1994) (list of variables, processes and compartments for the fitting of the degradation rates are given in Appendix A). A simple model of two processes was used:

Fe(III)P Fe(II)P k1 Fe(Il)P + CBr2Cl2 Fe(III)P + ... 2.10

1 The best fit (see Figure 2.15) gave values of k1 = 0.116 min.· and k2 1 1 1 1 33126 M- min: = 552 M- s· • The value of k2 is in good agreement with the measured second-order rate constant of CBr2Cl2 determined in experiments 1 1 with [substrate] 0 < [FeP] (see Table 4.9: 610 ± 220 M· s- ). 33

6.00E-05 I ~ I .!:I 5.00E-05 ~------~ I --- B 4.00E-05 1. • ~~---

0 3.00E-05 .~ ... _ .__ ~----• ...____~xp. 5 i:::: ----... • ...... O Ir- • e 2.00E-05 1] .._._._____ Exp. 4

~ LOOE-05 ~·- § l_~ Exp.3 u O.OOE+OO =--·.,.. Ex~---~--··---"------" 0 Exp. l 20 40 60 80 time (min.)

Figure 2.15: Concentration of CBr2Cl 2 versus time for five different initial concentrations: fitting of degradation curves with AQUASIM (Reichert, 1994). For initial concentrations of experiments 1 to 5 see caption of Figure 2.16.

For fast reacting substrates and [substrate]0 > [FeP] it is therefore important to consider the regeneration rate of the iron (II) porphyrin. In Figure 2.16 a simulation of the concentration of iron (II) porphyrin is shown for the experiments shown in Figure 2.15 (see numbers 1-5 for the experiments). As can be seen, for [substratelo >> [FeP] the concentration of iron (II) porphyrin decreases rapidly to less than 5 10·7 M, that is, about only 1/6 of the total iron porphyrin was actually reduced and active in the degradation process. 34 ,..__ .__.,:::E 3.00E-06 _ Exp. l p... ,..__ 2.SOE-06 H .__.,H ~ Exp.2 Q) 2.00E-06 ~ ...... l.SOE-06 0 Exp.3 ~ l.OOE-06 ...... 0 C<:J 5.00E-07 O.OOE+OO ~Q) u § 0 20 40 60 80 u time (min.)

Figure 2.16: Concentration of iron (II) porphyrin during degradation

: experiments of CBr2Cl 2 depending on the initial concentration of CBr2Cl 2 the concentration of Fe(II)P can be quite small compared to the total

6 concentration of FeP. Initial concentrations of CBr 2Cl 2 were: exp. 1: 1 10- M, exp. 2: 3 10-6 M, exp. 3: 1 10-5 M, exp. 4: 3.5 10-s M and exp.5: 5 10·5 M. The concentration of Fe(II)P versus time was simulated with AQUASIM (Reichert, 1994).

2.8 Blank reactions

Two kinds of blank reactions were performed. Firstly the catalytic activity of the sole porphyrin ring (without iron in the center) was checked. Secondly experiments were performed in the absence of mediators at different pH-values and at different cysteine concentrations. 35

2.8.1 Porphyrin ring as electron transfer mediator

It is known that methylviologen (N,N'-dimethyl-bipyridyl) can undergo a one electron reduction reaction (Schumacher and Holliger, 1996). This reduction is accompanied by a colour change from colourless to blue:

2.11

As the porphyrin ring used in this study contains four N-methyl- pyridyl groups (see Figure 2.1), one of these N-methyl-pyridyl groups could be reduced in the same way as methylviologen and act as an electron transfer mediator. This was checked with kinetic experiments using cysteine as a bulk 6 reductant, C2Cl6 as the substrate and 30, 60,100 or 200 µM of the ligand (see Appendix A). Table 2.2 lists the observed reaction rates.

Table 2.2: Observed and second-order rate constants for the degradation of

C2Cl 6 in the presence of porphyrin ligand

concentration of kobs (s-1) kugand (M-1 s-1) ligand

30µM 2.18 10-5 ± 2 10"6 0.727±0.07 60µM 5.33 10"5 ± 4 10"6 0.888± 0.07 lOOµM 1.42 10"4 ± 1 10-s 1.42 ± 0.1 200µM 2.03 10"" ± 1.3 10"5 1.016 ±0.06

Compared to blank reactions with only cysteine present, the ligand shows a slight catalytic activity. For example, at SmM cysteine at pH 7, C2Cl6 is degraded with kobs 3.68 10·7 ± 4.1 10-s s·1. Thus, the presence of 30 µM porphyrin ligand speeds up the reaction by a factor of 60.

6 5, 10, 15, 20 -Tetrakis (1-methyl-4-pyridyl)-21H,23 H-porphine, tetra-p-tosylate salt 36 Compared to reactions catalysed by iron porphyrin, however, the reactions catalysed by the ligand are much slower. At SmM cysteine and 30 3 4 1 µM iron porphyrin at pH 7, C2Cl6 is degraded with kobs = 1.44 10" ± 1.2 10 s· • Thus, the iron in the center of the porphyrin accelerates the reaction by a factor of 70 compared to the ligand alone. Hence, both the iron center and the porphyrin ring are essential for an

2 efficient electron transfer. Studies with free dissolved aqueous Fe +aq did not show inhanced degradation rates (compared to blank reactions). Note 2 however, that Fe + adsorbed on iron(hydr)oxides showed catalytic activity in redox processes (Klausen, 1995; Pecher et al., 1997; Waxweiler, 1996).

2.8.2 Reactions without Iron porphyrin

The reactions of cysteine with CC14 and C2Cl61 respectively, were studied at pH 7 and 9 and at both pH-values for different cysteine concentrations (1, 2, 5 and lOmM, see Appendix A).

Figure 2.17 shows degradation rates for C2Cl6 as a function of cysteine concentration at pH 7 (cysteine mainly protonated) and at pH 9 (cysteine mainly deprotonated). A slight increase with increasing cysteine concentration is observed from 1 to 5 mM cysteine and a leveling off at higher cysteine concentrations. At pH 9 reaction rates are - 10 times higher than at pH 7. A similar behavior is found for CC14 (see Figure 2.18). Compared to the iron porphyrin catalysed reactions, the blank reactions are much slower (by 5 to 7 orders of magnitudes). 37

2.410" 6 --''° 210" 6 u pH9 u"' i.610·6 + 6 ! ~ 1.2 10" ,.,::-- 7 'r.ll B 1(T • ! pH7 .Q"' 4 w-7 (J) ,.:,;;o m 0

0 2 4 6 B 10 12 cysteine concentration (mM)

Figure 2.17: Observed degradation rate constants for C1Cl 6 in systems without iron porphyrin: influence of pH and cysteine concentration.

2.510" 6

g· 2 1lf. -

>-< 1.s 10- 6 .,2 ,.,::-- pH9 'r.ll 1 110"61 T

.Q f "' 1 f ,.:,;;o s 10· ~ I ' pH7 I (J) Oi !I> ~ I ~ 0 2 4 6 B 10 12 cysteine concentration (mM)

Figure 2.18: Observed degradation rate constants for CCl 4 in systems without iron porphyrin: influence of pH and cysteine concentration. 38

2.9 Summary and conclusions

The experimental results presented in this chapter demonstrate that iron porphyrin is an efficient electron transfer mediator in the reduction of polyhalogenated alkanes. Reduction rates were found to be first-order in PHA and iron porphyrin concentration. Thus, the electron transfer from the iron porphyrin to the PHA is the rate-limiting step. Rates increased with increasing pH due to pH dependent speciation of the iron porphyrin: one of the two axially bound cysteine molecules can be deprotonated, the pK. is estimated to be around 7. The axial complexation of the iron porphyrin has an influence on the electron transfer rate. If present at high concentrations compared to FeP, cysteine, a strong ligand, increases the electron transfer rate, whereas phosphate, a rather weak and bulky ligand, decreases the electron transfer rate. On the whole, the influence of axial complexation on the electron trans- fer rate is rather small: factors of maximally 5 are found. Outer- sphere electron transfer processes would suggest a dependence on the ionic strength of the reaction solution, because there is no chemical bond which acts as a bridge for the electron transfer (and thus the electron should come in contact with the reaction medium). However, no such dependence was found. A possible description is electron transfer by tunneling, although a quantitative estimation of the wavelength of the transferred electron is difficult. Considering these results, the proposition of an outer- sphere electron transfer from the iron (II) center of the porpyhrin ring to the polyhalogenated alkane is not contradicted. For the application of reductive dehalogenation processes in remediation of waste disposal sites, results of product analyses and kinetic observation have to provide additional clues. If products are more toxic than the starting material the degradation process is not disirable. Harmless and water-soluble products, on the other hand, can be easily transported and further treatment of such products is not necessary. 39

3. Reaction Intermediates and Reaction Products

3.1 Introduction

An examination of the literature on abiotic transformation reactions of PHAs shows that very often no reaction products were isolated, either because their analysis proved to be too difficult, or because the main focus of the investigation was to learn more about the reactivity of polyhalogenated alkanes (Criddle et al., 1986; Ekstrom, 1988; Schanke and Wackett, 1992; Zoro et al., 1974). Nevertheless, some authors found degradation products with often quite good mass balances. In Table 3.1 a representative selection of the data reported in the literature is given: for several substrates the environmental conditions or the active electron transfer mediator are shown as well as the final product(s). Typical products in the degradation of polyhalogenated methanes are methanes with one halogen less than the parent compound. This halogen is substituted by one hydrogen (hydrogenolysis). The process can proceed until the substrate is completely dehalogenated (CCl4 ~ CHCl3 ~ CH2CI2 ~ CH3Cl

~ CH4 (Lewis et al., 1995)). Often, this chain stops before the methane is completely dehalogenated: the reduction potentials vary with the number of halogens present. Less halogenated compounds are more difficult to reduce (Eberson, 1987). Polyhalogenated ethanes (with at least one hydrogen) can undergo dehydrodehalogenation under basic conditions or in the presence of suitable 40 Table 3.1: Products of polyhalogenated alkanes in different degradation processes

Substrate Product(s) System Reference

(Vogel et al., 1987) Cl3C-CH3 CH,-COOH H20

Cl3C-CH3 c1,c,,,cH, H20 (Vogel and Mc Carty, 1987) Cl,HC-CH,Cl Cl,C=CH, H,O (Walraevens et al., 1974) CHCl, CH, Cr(Il)SO, (Castro and Kray, 1963)

CHC13 Cl-alkylated B12 B22 Co (III) (Wood et al., 1968)

CCJ.. cs, -t co, (80-85%) HS·, minerals: (Kriegmann-King biotite and Reinhard, 1992) CHC13 (5-15%) co venniculite

CCl, CHCl3 (50%) FeS, (Kriegmann-King co, (10-20%) and Reinhard, 1994) HCOOH (5%) CHCI, CH,Cl, Fe(II)Deutero- (Wade and Castro, porpyhrin IX 1973) c,c1. Cl,C=CCI, Fe(II)Deutero- (Wade and Castro, porpyhrin IX 1973)

CCI, CHC13 Fe(Il)Deutero- (Wade and Castro, porpyhrin IX 1973) CH,Br, CH, (3%) Fe(Il)Deutero- (Wade and Castro, H,C=CH, (13%) porpyhrin IX 1973) Rest: unknown

CBrCl, CHC13 Fe(Il)Deutero- (Wade and Castro, porpyhrin IX 1973)

CFCl3 co cytochrome P450cAM (Li and W ackett, 1993) 41 nucleophiles (Barbash and Reinhard, 1989; Roberts et al., 1993). Thus, halogenated ethenes are formed which are either stable or undergo further reductive dehalogenation, particularly catalyzed by microorganisms (Glod et al., 1996; Glod et al., 1997; Holliger et al., 1993). Another fundamental degradation process involves hydrolysis Geffers et al., 1989; Roberts, 1991). The hydrolysis of 1,1,1-trichloroethane results in acetate (Vogel et al., 1987). Products like acetic acid are desired in the degradation of organic pollutants, as they are harmless and water-soluble under environmental conditions. Under sulfate reducing conditions, transformations at the surface of minerals can be important. For example, in a system containing biotite CCl4 is converted to C02 (Kriegmann-King and Reinhard, 1992). Complete mineralization as in this case, however, is usually not achieved. Side- products .such as CHC13 or CO as well as HCOOH are found (Kriegmann-King and Reinhard, 1994). In this chapter the isolation and characterization of the product is shown for CC14 and nine other polyhalogenated methanes in the iron porpyhrin system. A proposal for the major reaction mechanisms is provided, also based on trapped reaction intermediates. Reaction mechanisms are also discussed for four polyhalogenated ethanes.

3.2 Degradation of CCl 4: product analysis

Criddle et al. have summarized all known abiotic and biotic transformations of CC14 with products and intermediates that have been reported in the literature (Criddle and McCarty, 1991). Potential products are

C2Cl6, CHCly CH2Cl21 C021 CO or HCOOH (Figure 3.1). Figure 3.1 shows by which pathways these products may be formed. 42

CC14

Cl C R1 R3 3 F<. K~l R·~R3 R2 R4 x2 112 R4 °CC13 C2Cl6 cell bound - e - - cry j+o~•-,+H' :CC12 CHCl H20 3 2Hc1y tH,O [cC1p2J H'+2e 2HCI t co er HCOOH ! CH2Cl2 0 Cljl_CI HP t 2HCI C02

Figure 3.1: Postulated abiotic and biotic transformations of CC1 4 (Criddle and McCarty, 1991)

As the degradation experiments in the iron porphyrin/ cysteine system were performed under anaerobic conditions the pathway including 0 2 to give phosgene can be excluded. Moreover, the degradation of CCl4 did not show any halogenated products such as CHCl3, CH2Cl2 or C2Cl6• Hence, the corresponding pathways in Figure 3.1 can also be excluded in the iron porphyrin/ cysteine system. In none of the investigated systems summarized 43

by Criddle and McCarty, however, compounds exhibiting -SH or -NH2 groups were present (see Figure 3.1). A recent study by Chiu et al. corroborates our findings: in a system containing vitamine B12 and cysteine as bulk reductant

CC14 was transformed to only 20% of CHC13, 3% of CO and the rest was not identified, being a non-volatile and not-halogenated product (Chiu and Reinhard, 1996).

3.2.1 Product distribution between different phases

14 C-labeled CC14 was used in order to localize the product in different phases. After complete degradation of CC14 the reaction solution was extracted with hexane and with diethyl ether. After both extractions 90 ± 5 % of the activity remained in the aqueous phase. This pointed towards a highly hydrophilic compound.

Table 3.2: List of experiments performed with radiolabeled product(s) of CC1 4

Exp. number and kind of experiment Information aAct.FeP "Act.blank

1) 500 µL sample+ 0.15 mL lN NaOH Overall activity 100 % 100 %

2) 500 µL sample + 0.15 mL lN H 2S04 Non-volatiles 90 % 50% purging with N 2 at pH 0 3) 500 µL sample + 0.15 mL lN NaOH Non-volatiles 95 % 50 % purging with N 2 at pH 14 +co/-

4) Activity (Exp. 3) - activity (Exp. 2) C02 5% 0% 5) Activity (Exp. 1) - activity (Exp. 3) Volatiles 5% 50 %

exceptC02 "Activity found in the solution after performing the experiment 44

Table 3.2 shows three experiments carried out to differentiate between volatile and non-volatile products and to check whether the product was 7 carbonate • The results of the degradation in the iron porphyrin/ cysteine system are shown in Figure 3.2.

volatiles CO 2 5% 5%

non- volatiles 90%

14 Figure 3.2: Iron porpyhrin mediated reaction of CC1 4 : after degradation of

CC1 4 90% of the acitivity was found in the aqueous phase. This phase contained 90% of non-volatiles, 5% of volatiles and 5% of C02•

7 In experiment 1, the reaction solution was made alkaline and the activity was measured without further treatment. Thus, the overall activity was obtained. In experiment 2, the acidified solution was purged with nitrogen in order to obtain the activity of the non-volatile products. Experiment 3 consisted of making the solution alkaline and purging it with nitrogen. The reaction solution contained the non-volatiles plus carbonate. (Carbonate is purged as CO, under acidic conditions, but remains in the solution under basic conditions). The difference between experiment 3 (non-volatiles + CO,) and experiment 2 (non-volatiles) 14 gives the amount of C-labeled C02 in the system. The percentage of the volatile.s (except C02) is given by the difference of experiment I (overall acitvity) and experiment 3 (non-volatiles+ CO,). 45

In the iron porphyrin/ cysteine system, 90% of the activity was found as a non-volatile, hydrophilic product, the rest being 5% volatiles and 5% C02•

In contrast, the degradation of CC14 in the blank system (cysteine only) gave

-50% of volatiles (determined as CHC13, analysed with GC-ECD) and -50% of non-volatiles (see Figure 3.3).

COi 0%

non- volatiles volatiles 50% 50%

Figure 3.3: Blank system: the product distribution is -50% volatiles (CHClJ and -50% non-volatiles.

The distribution experiments show that the degradation product in the iron porphyrin/ cysteine system was non-volatile and water-soluble, whereas in the blank system, 50% of CC14 was reduced to CHC13•

3.2.2 Ion chromatographic analysis

Figure 3.4 shows a typical ion chromatogram of the reaction solution 14 after complete degradation of C-CC14• (For conditions see Appendix A.) Each peak was collected separately and the activity was measured. The whole 46

activity was found to be in one peak eluting directly before SO/. Moreover,

four equivalents of chloride ions were found for one CC14• Apparently, complete dehalogenation took place. The resulting product had most probably two negative charges at pH 12, as under the experimental conditions ions are separated by charges: species with one negative charge are eluted first, followed by species with two negative charges etc.

ci- P043_ 10 µS

7£1 I

'~'I SO/ em ,~, ;~1 •~a s.11~1 I i.S& I I I I ' .II I I \_ 0 ~ '\. Mrutes 0 2 4 6 8 10 12

Figure 3.4: Ion chromatogram of the reaction solution after complete degradation of CCI,. The filled area contains the whole 14C-activity.

The nature of the product was further examined by using 13C-labeled 14 CC14 instead of C-labeled CC14• 47

3.2.3 Product analysis by NMR and MS

Figure 3.5 shows a 2-dimensional NMR spectrum of the isolated 13C- labeled product. The signal of the 13C-labeled C-atom is dominant and occurs at 166 ppm (NMR spectrum taken at 40°C; 164 ppm at 25 °C). Only one signal clearly dominates the spectrum, which agrees with there being only one final product. Compounds containing C-atoms at -164 ppm are urea (163.5 ppm), H- CO-NH-CH3 (165.5 ppm) or H-CO-Nlfz (167.6 ppm) (Pretsch et al., 1986). Thus, a connection of -CO-N= with C being the C of CC14 was postulated to be part of the product. jl( 166 I I -1 16B 170

1'!2

i76

.. 170

180

' ,, 182

ppm

Figure 3.5: 2-dimensional NMR of the 13C-labeled isolated product in DP. The NMR was taken at 40 °C. 48

As can be seen, the 13C-labeled C couples with proton singlets at 8.4 and 8.0 ppm (at 40 °C) (8.1 and 7.9 ppm respectively at 25°C). Protons with such high shifts are found in N-formylic functional groups, for example in methyl-formamide (Pretsch et al., 1986):

H~ 7.9ppm I ('"""H 'll'N-cH 0 3 8.1 ppm

As the rest of the 1H-spectrum is similar to the 1H-spectrum of cysteine or cystine (the oxidation product of cysteine) and as cysteine is the only molecule in the reaction solution with an amine functional group, a proposition for the structure of the product was N-formykysteine:

N-formykysteine was synthesized by formylating cystine and reductively splitting the S-S bond (see Appendix A, Syntheses). NMR-spectra of the resulting N-formykysteine corresponded to the NMR-spectra of the product. Also standard addition of N-formylcysteine to the reaction solution corroborated these findings: the product peak increased after addtion of N- formykysteine (Figure 3.6). 49

a 10 µS

7.61 l

·~'l

il,01 Sf;l! 6r20QJPl I 8.S6 ,,, I I 0 ~ I I l l ' lJ\ ~ M'1U!>< 0 2 4 6 8 10 12

b

10 r~n I µS

"''I 0112 u:o 9 I 0 sre 6119 'f'/f'l ,. I i~~t ~ \_ \ Miflufe'i 0 2 4 6 8 10 12

Figure 3.6: Figure a shows an IC-chromatogram of the reaction solution. The product peak is marked black. Figure b shows the IC-chromatogram after addition of N-formylcysteine. The product peak increased.

Moreover, data from mass spectrometry suggest N-formykysteine as product: the same sample as used for the NMR (13C-labeled and dissolved in 50

DP) was injected with a syringe-pump into a mass spectrometer. The ionization mode was negative electrospray ionization. Masses 298, 299 and 300 (distribution 57%, 74% and 100% respectively) were dominant. They represent the molecular weight of N,N'-diformyl-cystine, twice 13C-labeled at the formyl functional group and deuterated at the carboxylic functional groups (m/z = 300). If one deuterium is exchanged by a proton (the eluent was fizO), m/z = 299 occurs, if both deuteriums are exchanged it is m/z = 298. Apparently, the N-formylcysteine oxidizes completely to N,N' -diformyl- cystine after five days.

The C-atom of CC14 is in the oxidation state +IV. In the product, the C of the N-formyl group is in the oxidation state +II. Thus, the reaction is a reduction and complete dehalogenation of CCI,.

3.3 Degradation of CCl 4 : reaction Intermediates

Iron porphyrin is a one electron mediator (couple Fe(II)P /Fe(III)P).

The first step in the degradation process of CCI4 to N-formylcysteine can therefore be assumed to be a one-electron transfer, where CC14 accepts an electron. As radical anions of aliphatic polyhalogenated compounds have very short lives (Wentworth et al., 1969), a dissociative electron transfer is proposed (equation 3.1):

Fe(II)P Fe(III)P \,, )., + er 3.1

3.3.1 Trapping of CCl3·radlcals

The first intermediate in the degradation of CCI4 is proposed to be a

CCl3 radical as the formation of CCl3 radicals has been demonstrated when 51

hydrated electrons react with CC14 (Swallow, 1978). With dr-isopropyl alcohol, a good D· atom donor, CC13 radicals can be trapped as deuterochloroform (equation 3.2):

if.3D if.3D D'?-OH + D>-OH D DD D 3.2

Different amounts of d 7-isopropyl alcohol were added to the reaction solution (see Appendix A). Table 3.3 shows that in the iron porpyhrin

mediated system1 the amount of CDCl3 found increased with increasing d 7-

isopropyl alcohol present, whereas no CHC13 was detected. In contrast, in the

blank reaction, no CDC13 was found. Only CHC13 was present.

Table 3.3: Portion of deuterated chloroform in the iron porphyrin/cysteine

system as a function of the amount of d7-isopropyl alcohol present in solution

amount of d7-isopropyl alcohol % of deuterated chloroform in the (in % of total vol) degradation of CC14

1 12.62±0.91 0.5 7.97±0.26 0.1 2.4±0.31

This shows clearly that CC~ radicals are formed in the iron porphyrin/ cysteine system. As the final product is in an oxidation state which is +II, in contrast to +IV at the beginning, a second electron has to be transferred during the degradation pathway. A second electron transfer to 52

CC13 radical leads to a CCl3 anion that can either be protonated or decay to dichlorocarbene.

3.3.2 Trapping of carbenes

CHC13 would be a major product if the second electron is transferred to the CCI3 radical. No CHC13 was actually detected in the iron porphyrin/ cysteine system. Nevertheless, it was checked whether :CClz carbenes were present during the degradation process of CCI., Dichlorocarbenes are known to react with secondary amines to form N-fomylic compounds (Frankel et al., 1959). Refluxing secondary amines with chloroform and base results in the corresponding formamide compounds. Figure 3.7 shows a mechanistic interpretation of the formation of N-formylic compounds.

HCC13 + Off - -CC13 + - :CC12 + R2NH + :CC12 -

0 Rz~ + 2HC1 H

Figure 3.7: Proposed mechanism of the formation of N-formylic compounds from dichlorocarbene (Frankel et al., 1959).

It is known that olefins are good carbene traps (Wentrup, 1979) (Fujita et al., 1981). As carbenes do have a non-bonding pair of electrons as well as an electron gap, they are able to form cyclopropanes by adding to olefins: 53 - Dichloro- Olefin carbene Cyclopropane 3.3

Such reactions are also used to differentiate between carbenes in a singlet state (non-bonding electrons in one orbital) or in a triplet state (the two non-bonding electrons are in two different orbitals), as singlets add to the double bond in a concerted reaction forming stereoselective products, whereas triplets do not (Hoffmann, 1976). In the case studied here, a water-soluble olefin was used: 3-pentenok acid. As Fujita et al. got much better yields with 1,1-disubstituted olefins (Fujita et al., 1981) another olefinic acid was used as well: 3-methyl-3-butenoic acid (Figure 3.8) (see Appendix A, Syntheses): r=\_COOH I ~OOH ~OOH Cis-I trans-3-pentenoic acid 3-methyl-3-butenoic acid

Figure 3.8: Olefins used for trapping carbenes: cis-/trans-3-pentenoic acid and 3-methyl-3-butenoic acid.

In the degradation of CCl4 mediated by iron porphyrin, no free carbene was trapped as cyclopropane. In accordance with the absence of any

CHC13 as product, this suggests that the second electron transfer is not an electron transfer to the CCis-radical: the second reduction must occur later in the reaction pathway (Figure 3.9). 54

:CC12 + er

Figure 3.9: In the iron porpyhrin system the second electron is not transferred to the CCl3 radical.

In contrast, in the blank reaction (cysteine only), free carbenes could be trapped with the olefinic acid. This indicates that the first step in the blank reaction could be a nucleophilic attack on a chlorine of CCl4 resulting in a

CC13 anion that decays to free dichlorocarbene and chloride. Nucleophilic attacks at chlorine atoms are known (Li et al., 1987; Zefirov and Makhon'kov,

1982). For example, in reactions of triphenylphosphine with C2Cl6 or CCl4, P- C! bonds resulting from a P-attack on chlorine are found (Appel and Scholer,

1977; Slagle et al., 1981). Thus, a CC13-anion is formed. This anion can either be protonated (52 ± 5% CHC13 are found) or decay to a dichlorocarbene and a chloride anion (carbenes were trapped as cyclopropane). Considering these facts, the difference in reaction mechanisms be- tween the iron porpyhrin/cysteine system and the blank (only cysteine) is clearly shown by reaction intermediates.

3.4 Degradation of CCl4 : proposed reaction mechanism in the iron porphyrin/ cysteine system

With the information gathered in the analysis of product and reaction intermediates, a reaction mechanism is proposed for the degradation of CC14 mediated by iron porpyhrin (Figure 3.10). 55

Fe{II)P Fe{III)P RSSR RS· ,...... , \ )., ' I jcc14! \ ) .. jcc13I + !RS..cCl3: or RS· .... ____ .. '* ....

,...... , r .. - ...... , ! RS-CCl2 : i RS-CCl2- i o I L,.,.,.., ...... ,;1 i...... , ...... ~

r-- ...... ____ , SH : 19 I • RS--< H~":J'~COOH : H I I. .. - ...... ~ H ,.- ...... ____ .. , .• H ., i s+c1 l i ZyNH i ! coo~ : ' .L .... ,.,.,.,.,...... ,.,J

Figure 3.10: Proposed reaction mechanism of the degradation of CCl 4 in the iron porpyhrin mediated system: compounds framed by solid lines have been identified, compounds framed by broken lines are proposed.

As can be seen in Figure 3.10 the first step is a one electron transfer turning CC14 into CC13 radical and chloride ion. CCI;. radicals were trapped with D-atom abstraction from d 7-isopropylic alcohol. As a second electron transfer at this point of the degradation pathway can be excluded (neither

CHCl3 nor :CCl2 were found), a radical reaction with either a cysteine radical or with cystine (short RSSR in Figure 3.10) is proposed. Cystine is present from the beginning of the reaction because it is formed when cysteine reduces the iron porphyrin. The solubility of cystine is 500 µM at maximum (Alfassi,

1988). In the radical reaction proposed, the S-S bond is broken, and RS-CC13 is formed. In RS-CC1 3 the C is formally oxidized again (+III --> +IV). RS-CC1 3 is proposed to be reductively dehalogenated to form a RS-CCl2 radical and a 56

chloride ion or to be transformed to an RS-CC12 anion by an X-philic attack of RS· at chlorine (Li et al., 1987; Zefirov and Makhon'kov, 1982).

Another possibility to form an RS-CC12 radical is that the thiol group of the anion of cysteine attacks the CC13 radical in an SN2-reaction (equation

3.4). Such a reaction would lead directly to the RS-CC12 radical. It is known that Ar-s· (Ar = Aryl) and CF3 radical react to Ar-S-CF3·, a radical anion (Wakselman and Tordeux, 1984). Other nudeophilic attacks at radicals are known (Beadle et al., 1984; Rico et al., 1983): attack of RS" at 0 2N-furan-CH2 or of thiophenoxide at CF2CL In the reaction of an aliphatic thiol (where no delocalization of the negative charge is possible) with CCl3 radical (C-Cl bonds are much weaker than C-F bonds (March, 1992)) the cleavage of a chloride most probably occurs if the thiol group of cysteine attacks the CC13 radical.

RS-CCI2 + Cl" 3.4

In any case, the formation of a RS-CCl2 radical as a reaction intermediate seems a reasonable hypothesis. Radicals substituted by sulfur at the radical center are quite stable because the lone pairs of sulfur stabilize the electron deficiency in the radical center (Cilento, 1960; Pasto et al., 1987).

Further reduction and protonation of the RS-CCl:z radical leads to RS-CHCl2•

H atom abstraction leads also to RS-CHC12• This intermediate can hydrolyze resulting in S-formyl-cysteine. From a thermodynamical point of view, S- formyl-cysteine is less stable than N-formykysteine (March, 1992). Therefore an intramolecular attack of the amine functional group at the formyl group may take place. The amino group could also attack the -CHCl2 of RS-CHCl2 resulting in a cyclic intermediate which would hydrolize to yield N- formykysteine. Although at pH 7 the amine group is protonated (ammonium: pK. = 10.29 (Smith and Martell, 1989)), there is a small number 57 of molecules that are deprotonated. Intramolecular reactions are much faster compared to similar intermolecular reactions because the reactive centers are closer to each other (March, 1992).

3.5 Degradation of CCl 4 : proposed reaction mechanism in the reaction with cysteine (blank system)

In the blank reaction CC14 is converted to 52 ± 5 % CHC13, 53±10 % N- formylcysteine and 5 ± 2 % HCOOH. Here, the reaction mechanism can be formulated as shown in Figure 3.11.

idecay 52±5%

j:cc14 + c1- H20 CO -20-

SH Clyl ~ RS-CHCI2 HCOOH HAN _,(COOH I see Fig. 3.11 5±2% H ~COOH found 53± 10%

Figure 3.11: Proposed reaction mechanism of the degradation of CC14 in the blank system. Compounds framed are isolated products or trapped reaction intermediates. 58

The first step in the degradation of CCl4 is an SN2-attack of cysteine at

the chlorine atom of CCl4• The resulting CC13 anion can be protonated (52 ±

5% of CHCl3 are found) or decay to :CC12 (trapped as cyclopropane) and a chlo- ride ion. 53 ± 10% of N-formylcysteine and traces of formic acid (5 ± 2%) are found. These products are obviously formed by reaction of dichlorocarbene with cysteine and water respectively. The thiol group of cysteine is a good nu- cleophile and attacks the center of the the carbene. Protonation of the result-

ing anion gives RS-CHC12 which is transformed to N-formylcysteine as dis- cussed in the mechanistic description of the iron porphyrin mediated process. On the other hand water can react with dichlorocarbene to form formic acid or carbon monoxide respectively. The presence of dichlorocarbene was proved by trapping it with cis-/trans- 3-pentenoic acid resulting in the corre- sponding cyclopropane.

3.6 Degradation of CBr4, CBr2Cl 2, CBrCl3: comparison with CCl4

Table 3.4 shows the distribution of products and the occurrence of re-

action intermediates in the degradation of CC14, CBr4, CBr2Cl2 and CBrC13 (for conditions, see Appendix A). The distribution is given for the degradation in the iron porphyrin/cysteine system (FeP system) as well as in the blank reac- tion (without iron porpyhrin, only cysteine). The quantification of N-formykysteine by IC measurements showed

that this was the main product of CBr4, CBr2Cl2 and CBrCl3 just like in the

degradation of CCl.- For CBr4, CBr2Cl2 and CBrCl3 , however, this was not 13 proved by applying C-labeled compounds as in the case of CC14•

Halide ions gave quite good mass balances (see Table 3.4): for CBr4,

CBr2Cl2 and CBrCl3 complete dehalogenation was found: 4 Br· I CBr,y 2 Cl", 2 Br·

I CBr2Cl2 and 1 Br·, 3 CtI CBrCl3• This is in accordance with the main product being N-formylcysteine.

As in the degradation of CC14 in the iron porphyrin/ cysteine system no free carbenes could be trapped and only a few of the less halogenated side-

products like CBrC13 -; CHC13 (1±1%), CBr4 -; CHBr3 (3±1%) and CBr2Cl2 ->

CHBrCl2 (5±2%) were found. Thus, the presence of CX3 anions (X Br or Cl) can be excluded. Table 3.4: Product distribution and reaction intermediates of CCl4, CBr4, CBrCl3 and CBr2Cl2 in the iron porphyrin/ cysteine system and in the blank system compound products intermediates halides and other anions halogenated N-formylcysteine methanes iron porphyrin/cysteine system

CC4 420 ± 60 % c1- <1 %CHC]s 125 ±20 % no free carbene CC ls-radical trapped CBq 460 ±70 % Br- 3±1 % CHBr3 95±10% no free carbene CBrClJ 320 ± 40 % CI-; 120 ± 20 % Br- <1 % CHCl3 115± 15 % no free carbene CBr2Cl2 210 ± 30 % CI-; 230 ± 30 % Br- 5±2 %CHBrC!i 110±15 % no free carbene CJ1 "" blank system (cysteine only)

CC4 210 ± 30 % ci-; 5 ± 2 % 52 ± 5 % CHc13a 53±10 % dichlorocarbene trapped HCOO- CBq 150 ± 20 % Br-; 3 ± 2 % 91 ± 10 % CHBr3 12±10% dibromocarbene HCOO- trapped CBrC!J 80 ± 10 % CI-; 115 ± 20 % Br- 80 ±5%CHCl3 5±4% dichloro- and bromochlorocar- bene trapped CBr2Cl2 50 ± 20 % Cl-; 120 ± 20 % Br- 61±22 % 14±10 % dkhlorocarbene CHBrCl2 trapped 1 ± 1 % CHBriCI a Results from measurements carried out at T = 55°C. 60

In the radical trap experiments with d7-isopropyl alcohol, an interesting observation was made: the background reaction of cysteine with

CBr4, CBr2Cl2 and CBrC13 respectively to give CHBr3, CHBrCl2 and CHClg as side-products was faster than the radical reaction of CBr3 radical, CBrC12

radical and CClg radical with d 7-isopropyl alcohol. Therefore, in the GC-MS measurements, where CHX3 and CDX3 coelute, the masses were dominated by

CHX3 - in contrast to the experiment with CC14, where the blank reaction which can be considered as the background reaction in the iron porphyrin/ cysteine system is very slow compared to the trap reaction. Landolt and Bornstein give a second-order rate constant of 79 M·1s·1 for the reaction of isopropyl alcohol with CCis-radical (equation 3.6). Deuterated isopropyl alcohol will react even slower in the same reaction (Landolt and Bornstein).

3.5

On the other hand, the main products in the blank reactions are methanes where one halogen has been replaced by a proton (see Table 3.4). 91

± 10% of CBr4 are found as CHBr3, 61 ± 22% of CBr2Cl2 as CHBrCl2 (side-

product: 1 ± 1% as CHBr2Cl) and 80 ± 5% of CBrClg as CHC13• N-formylcysteine is also found as a product for all three methanes. It is, however, only a side- product (-10%). In accordance with these findings free carbenes were trapped in the blank reactions. The amounts of free carbenes trapped varied with

structure: in the case of CBr4 2 ± 1% :CBr2 were trapped as cyclopropane.

CBrC13 gave 20 ± 5% of :CCl2 and traces of :CBrCl and finally CBr2Cl2 showed

48 ± 12 % :CC12 that was trapped. Two processes determine the amount of free carbene in the reaction solution: the protonation reaction of the C~ anion

and the decay of CX3 anion. 61

Hine and Ehrenson have determined the ratios of these reactions for 8 different haloforms (Hine et al., 1957; Hine and Ehrenson, 1958) •

CHXYZ :CXY + z- + H+ 3.6

In Table 3.5 the ratios of 1

As can be seen in Table 3.5 ·cBr3 anion has the smallest ki/k1 ratio, meaning that protonation is much favoured over decay. Replacing a bromine by chlorine giving ·cBr2Cl gives a higher ki/k.1 ratio. Not every replacement of bromine by chlorine, however, increases the 1

' Hine et al. measured the deprotonation rate of CHXYZ, k1, and expressed the ratio of k.1 (protonation rate of ·cxYZ) and k2 (rate of decay) by a formula with various parameters using CHCJ3 as reference:

= Carbene stabilization factors for halogens at the carbene =Constant for the halogen abstracted in the decay nZ Number of halogens in the anion that are the same as the abstracted one c,d =Proportional constants, dimension- less, function of temperature For values of these parameters (determined by fitting) see next footnote. 62

Table 3.5: Rate of decay/ protonation rate of CXYz- (Hine and Ehrenson, 1958)

anion

-cBr3 2.3 E-6

-cBr2Cl 2.6 E-5

-cc13 7.3 E-5

-cBrCl2 2.9 E-4

-cBr2F 8.2 E-2

-cc12F 8.5 E-2

These ratios are reflected in the amounts of free carbenes trapped and in the amounts of CHX3 found (X = Br or Cl): CBr4 delivers 91 ± 10 % of CHBr3

; and only 2 ± 1% of :CBr2 CBrC13 gives 80 ± 5 % of CHC13 and 20 ± 5 :CC12

(traces of :CBrCI); CBr2Cl2 ends up in 61 ± 22% of CHBrC12 and 1 ± 1 % of

CHBr2Cl with 48 ± 12 % of :CC12 in the reaction solution.

It is somewhat surprising that in the degradation of CBrCl3 both :CC12 and :CBrCl are trapped. C-Br bonds are weaker than C-Cl bonds (March, 1992) 9 (Hanzlik, 1981) and should therefore be cleaved preferentially •

9 Hine and Ehrenson give a formula for the probability with which a halogen Xis cleaved in an anion -ex, relative to chlorine. This formula is limited to anions containing at least one chlorine. kx n log-fi-=Nx -Mx +log-- k2 n -3 where Mx = Carbene stabilization factor for halogen at the carbene Nx = Constant for the halogen abstracted in the decay n = Number of halogens in the anion that are the same as the abstracted one

For example -cBr2CI has k2(Br)/k2(Cl) = 346, whereas -cBrC1 2 has k2(Br) /k2(Cl) = 87. The values of the parameters were achieved by fitting 9 data points. Hine and Ehrenson got values at 0 and

50"C: MF=3.727 (0°C), 3.195 (50°C) M8,=-1.461 (0°C), -1.131 (50°C) M1 = 2.626 (0°C), -2.039 (50°C) N8 , = 0.785 (0°C), 0.936 (50°C)

N1 = 0.118 (0°C), 0.486 (50°C) (c-d) = 0.299 (0°C), 0.196 (50°C) 63

As shown in Table 3.4 the concentrations of halide anions found in the blank reactions balance quite well the distribution of products found, except for relatively high concentrations of chloride in the case of CBrCl3 (80 ±

10 %) and CBr2Cl2 (50 ± 20%). Although in the proposed SN2 attack of the thiol group of cysteine at CX4 an S-X bond is formed, the mass balances for the halide ions are good. Apparently the R-S-X compounds lose the halogen again, most probably by a further attack of cysteine at the sulfur resulting in cystine and a halide ion (equation 3.7):

~~ RS- + R-S-X - R-S-S-R + x- 3.7 , Summarizing these results, the following proposition is made: CBr4

CBr2Cl2 and CBrC13 follow the same reaction pathway as CC14 for both the reaction catalysed by iron porphyrin (see Figure 3.10) as well as the blank reaction (see Figure 3.11). In the porphyrin mediated reaction, a radical is formed in the first electron transfer. The radical reacts either with cystine or by a nucleophilic attack with Rs- to give the final product N-formylcysteine. This product is water-soluble and not toxic. In the blank reaction, the main product of CX4 (X = Br, Cl) is CHX3, the protonated anion which is generated by an SN2 attack of the thiol group of cysteine at a halogen of CX4•

3.7 Degradation of polyhalogenated methanes containing one

hydrogen: CHBr3, CHBr2CI, CHBrCl 2

Degradation products and reaction intermediates of three haloforms were determined: CHBr3, CHBr2Cl and CHBrC12• CHC13 degraded too slowly to be a suitable probe molecule for this investigation (-1 % degradation during three months). 64

In the iron porpyhrin mediated system, the main product of CHBr3,

CHBr2Cl and CHBrC12 was N-formylcysteine (see Table 3.6). The halide ions

; ; found complete the mass balance: 3 Br-/CHBr3 2 c1-, 1 Br"/ CHBrC12 1 c1-, 2 Br·

I CHBr2Cl. Thus, complete dehalogenation took place. For haloforms the following reaction mechanism is proposed (Figure 3.12): the first step is a dissociative electron transfer from the iron (II)

porpyhrin to the haloform. The CHX2 radical reacts with cystine (R-S-S-R) to 10 give RS-CHX2 • This reaction has already been proposed for CX3 radicals. As

in the case of haloforms, RSCHX2 is only formed by this reaction. The reaction

of RS- with CX3 in the degradation of CX4 to give RS-CX 2 radical can be ex-

cluded (see discussion of Figure 3.10). RS-CHX2 is hydrolyzed and intramolecularly reorganized to N-formylcysteine.

Fe(II)P Fe(III)P RSSR RS· r---- ...... , \ J. ' ' CHX1 _\"""--")~. [ ~ ~~~~ ]+ er :RS-CHX2 : I L ...... "'., • .. .I

r------, SH ' ' RS---'10 : H H)(N,(COOH ' I -- L ...... I. H -

Figure 3.12: Proposed reaction mechanism for the degradation of

CHX3 in the iron porphyrin/cysteine system. Compounds framed by solid lines were detected, compounds framed by broken lines are suggested reaction intermediates.

10 4 1 1 Such reactions are known: CH3 + CH3SSCH3 --; [CH3SS(CH3) 2]' k =6 10 M' s' (Alfassi, 1988) Table 3.6: Product distribution and reaction intermediates in the degradation of CHBry CHBr2Cl and CHBrCl2 compound products intermediates halides and other anions halogenated N-formylcysteine methanes iron porphyrin/cysteine system

CHBr3 330 ±50 % Br- 130 ±20 % no free carbene CHBrCl2 160 ± 30 % Cl-; 115 ± 20 % Br- 130 ± 20 % no free carbene CHBr2Cl 100 ± 20 % Cl-; 215 ± 30 % Br- 85±20 % no free carbene blank system (cysteine only)

220 ± 30 % Br; 14 ± 10 % 1±1% CH2Br2 53± 10 % no free carbene HCOO- CHBrC!i' CHBr2Cl'

'Only 5% degradation. 'Degradation too slow to be of any use in product analysis 66

Thus, in the case of halofonns, no overall reduction takes place: the

C-atom in CHX3 is in the oxidation state +II, and in N-formykysteine the C of the N-formyl group is in the same oxidation state. The primary reduction is

compensated by the oxidative radical reaction (CHX2 + R-S-S-R -t RS-CHX2 + RS).

The blank reactions of haloforms are very slow. Only CHBr3 showed ~5% turnover during 3 months under the conditions used (see Appendix A).

The product distribution shows 1 ± 1% of CH2Br21 220 ± 30% of bromide, 14 ± 10% of HCOOH and 13 ± 5% of N-formylcysteine. No :CHBr was trapped, but it can be assumed that with only 5% turnover the concentration of carbenes was too small to be detected even if they were formed. The fact that 2 Br· are

found per CHBr3 implies that CH3Br is a potential product. Methylbromide, however, was not determined. The mass balance in the blank system of

CHBr3 is therefore not complete.

3.8 Degradation of fluorinated methanes: CFBr3, CF2 Br 2, CFCl3

Table 3.7 shows the product distribution and the reaction intermediates in the degradation of the fluorinated polyhalogenated methanes investigated. The results are given for both the iron porphyrin

mediated reaction and the blank system, except for CFC131 where only the iron porphyrin mediated reaction is given because the degradation in the blank system was too slow to be of any use in product analysis.

As can be seen from Table 3.7, CFBr3, CF2Br2 and CFCl3 were also all completely dehalogenated, but the main product is unknown - it was not, as

in the case of CX4 and CHX3 (X = Br, Cl) N-formykysteine. N-formykysteine

was only a side-product (-30% (average in the CFBr3, CF2Br2 and CFC13- system)). Moreover, for all three polyhalogenated methanes, free carbenes were trapped in the iron porphyrin/ cysteine system - which is also in contrast

to all other PHMs investigated. For example, :CFBr and :CBr2 (!) are found in 67

the degradation of CFBr3• In the degradation of CF2Br2 the carbene :CF2 was 11 trapped • And in the case of CFC13, :CFCl was trapped as cyclopropane.

It is rather surprising that in the case of CFBr3, both carbenes, :CFBr and :CBr2 occur. One would expect :CFBr, as C-Br bonds are much weaker than C-F bonds (March, 1992) - still, :CBr2 is also trapped. Fluorine is both a strong cr-acceptor and a relatively strong 7t-donor (stronger compared to chlo- rine or bromine (March, 1992)). In Table 3.5 it is shown that the CFBr2 anion decays 3000 times more easily to :CFBr than CBr2Cl anion decays to :CBrCI. This is due to destabilization of the anion by fluorine. Generally, fluorinated polyhalogenated anions tend to decay more easily than CX3 anions with X = Cl, Br. From these findings it seems that in the iron porphyrin/ cysteine sys- tem, carbanions are formed that decay to carbenes. However, only little of the less halogenated volatile products are found: in the degradation of CFBr3, 12 CHFBr2 was found , but it was only 5% of the amount found in the blank reaction - thus CHFBr2 is only a side-product in the degradation of CFBr3• In the case of CF2Br2 no CHF2Br and in the case of CFC13 no CHFC12 was found - but this can be due to the fact that no standards of CHFCI2 and CHF2Br (like for CHFBr2) were available to compare retention times and determine sensitivity. Moreover, the compounds are very volatile (for example the dimensionless Henry's law constant for CHFC12 was experimentally fpund to be 11 (Mackay et al., 1993)) which implies that most molecules will be in the gas phase. Headspace analysis, however, was not performed in these experiments.

11 By the way, CF,Br, is used as :CF2-supplier in organic syntheses (March, 1992). "Analysis with GC-MS: mlz= 111, 113 (100%, 98%) for the CHFBr fragment and 190, 192, 194 (51%, 100%, 49%) for CHFBr2 Table 3.7: Product distribution and reaction intermediates of CFBr3, Cf2Br2 and CFC13 in the iron porphyrin/ cysteine system and in the blank system (cysteine only) compound products intermediates halides halogenated N- methanes formykysteine iron porphyrin/cysteine system

CFBr3 110 ±30 % P-; 330 ±50 % Br CHFBr2 present 30± 15 % fluorobromo- and (-5% of blank) dibromocarbene trapped CFCl3 100 ± 40 % F-; 360 ± 60 % CI- 40± 15 % fluorochloro- carbene

trapped 0-. 00 CF2Br2 200 ± 70 % F-; 280 ± 50 % Br- 20±15% difluorocarbene trapped blank system (cysteine only)

CFBr3 50 ± 30% F-; 250 ± 50 % Br- CHFBr2 present1' 46± 15% no free carbene CFCh• CF2Bri 105 ± 15 % F-; 260 ± 40 % Br 55±25% no free carbene

' Degradation too slow •not quantified because no standard available 69

The fact that few CHX3 products were found is in accordance with the

findings of complete dehalogenation: 1 P-, 3 Br· I CFBr3; 2 P-, 2 Br· I CF2Br2 and

lP-, 3 Cl" I CFC13• But on the other hand, the precursors of the detected carbenes must be the corresponding anions - and anions can be either protonated or decay (shown for example in Figure 3.11). Moreover, protonation is always faster than decay (see Table 3.5). As the protonated

products (CHX3) are only present in small amounts compared to the blank

system (in the case of CFBr3), alternative concurrent reaction pathways have to be proposed in order to explain the product distribution observed (Table 3.7). The first step is a dissociative one-electron transfer as in the case of the other PHMs (discussed in chapters 3.4, 3.6, and 3.7). A radical is formed. Fluorinated methyl radicals are much more reactive than comparable chlorinated and/or brominated methyl radicals (Jiang et al., 1989; Pasto et al., 1987). An example is shown in Figure 3.13.

cc13 + )-ott HCC13 + )-oH

Figure 3.13: H-atom abstraction of CC13 and CF3 radical from isopropyl alcohol (Landolt and Bornstein)

As fluorinated radicals also tend to have higher electron affinities than chlorinated and/or brominated radicals (Buhler, 1983), three possible re- actions of the primarily produced fluorinated radicals are conceivable (see

Figure 3.14): A first reaction is a radical reaction with cystine to give R-S-CFX2

followed by N-formylcysteine (side-product) or R-S-CH2-0H (proposed). A second reaction could be a second electron transfer to the radical as small 70

13 amounts of CHX3 were found and carbene intermediates were trapped • Li and Wackett proposed a :CFCl carbene (from CFC13) to react with reduced cytochrome P4socAM giving an iron bound carbon-monoxide by an iron-carbene intermediate (Li and Wackett, 1993). Moreover, visible spectral studies of reactions of polyhalogenated methanes with P 450-heme have suggested the generation of iron-carbene complexes (Mansuy, 1980; Nastainczyk et al., 1978; Wolf, 1977). In the system studied here, however, such a reaction is rather not likely, as a free carbene will not replace a cysteine in the axial position at the iron center (see Figure 2.6, species X). In the reactions studied by Li et al. and Mansuy et al. respectively, only one cysteine was in the axial position at the iron-center (Li and Wackett, 1993; Mansuy, 1980). Thus, vis-a-vis, there was free space to form an iron-carbene complex.

13 This would be in clear contrast to the reaction mechanism proposed for CX4 and CHX3 (X =Br, Cl). 71

r ...... -" ...... , l RSH +CHO : f""R5-C~~;;·1 + I HX I !..----·-·---...... - ..... -...... ___ .. t H20 [~~~~~~~] + 0

...... _., ""( Fe(II)P l RS-CHX2 : &.------·-"" ~ Fe(IIl)P RSSR RS· r------.... ., ~ ,.~ ...... , \ )., ; RS-CX3 : ~ ~ RS-Cx2 ! ·----·-··· r;i ...... LJ e·i W r·~c8~-1 ...... , ...... "' l-2ttxl H20 ,...... i ., l RS_f ; ..I __ ...... H ..: ' .' ...... ' 1

Figure 3.14: Possible reaction mechanisms in the degradation of fluorinated methanes. Species framed with solid lines were detected, species framed with broken lines are proposed.

A third reaction has to play a role in the degradation of fluorinated polyhalogenated methanes, because the two reactions proposed do not give a full mass balance. As the corresponding radicals are much more reactive than 72 chlorinated and/or brominated methyl radicals (Figure 3.13), a radical reaction at the porphyrin ring is conceivable. Double bonds are known to react with radicals (March, 1992) (see Figure 3.15).

H Cl Cl3C ·.-Cl k= 7.8·103 M-1s-1 CC13 + ) < H~ 4iiirc1 Cl Cl - Cl E.=33.7 kJ /mol

2 1 1 CC13 + k=2.56'10 M- s- 0 a,cD E.=14.2 kJ/mol Figure 3.15: Reactions of radicals with olefins (Landolt and Bornstein)

Thus, a radical reaction at the P-pyrrolic positions of the porphyrin ring is proposed with further reductive dehalogenation. The electron transfer from the iron center to a "trapped" radical by the ring is an intramolecular electron transfer reaction (Eberson, 1987) and should therefore be favoured. Figure 3.16 shows a proposed reaction scheme for the reaction of the

CFBr2 radical with the porphyrin ring in the P-pyrrolic position (1). An electron is delivered from the iron center in an intramolecular reaction and protonation follows (2). (The electron can also come from another iron porpyhrin or from an anion of cysteine.) Deprotonation results in renewal of the conjugation between the attached CFBr2 radical and the iron center (3). A second electron transfer from the center to the peripheral polyhalogenated carbon is proposed with a concerted loss of bromide (4). Reduction and protonation follow (5). Strong nucleophiles present in large amounts compared to others, for example cysteine, can then provoke an SN2-reaction with bromide as leaving-group (6). Elimination of fluoride is expected next

(7), as for example CF3-0H is unstable at room temperature and known to 73 eliminate HF (Kloter and Seppelt, 1979). Hydrolyzation of the C=SR bond yields a formyl group in the ~-pyrrolic position (8),(9).

(2)

(3) i -H'"

Fe(Ill)P Fe(ll)P ,S CFBr2

)>..-Br- / s'"%HN""" H_ (4) - N- ~ -

U Cys~(; -F ~Cys S1-H 1 I OH-- 1 I - -(7) ifN (8) N (9)

Figure 3.16: Reaction of CFBr2 radical with the porphyrin ring: Possible pathway in the degradation of CFBr3 •

Such a mechanism would explain the fact that complete dehalogenation took place and that the C-atom of CFBr3 was not found in N- formykysteine. Here, a formyl group at the ring is proposed. These three mechanisms for further reactions of fluorinated radicals are shown in Figure 3.14. AB already stated, the information from the results are too sparse to give 74

14 a completely clear mechanistic picture. Further studies with C-labeled CFBr3 need to be made and/ or CO has to be measured because CO could be formed in the reaction of :CX2 and ~O. The product distribution in the blank experiments of fluorinated methanes (Table 3.7) shows N-formylcysteine as a product (46 ± 15% in the

CFBr3 system and 55 ± 25% in the CF2Br2 system). In the case of CFBr3, CHFBr2 was detected with GC-MS. Unfortunately, no standard was available and therefore no quantification could be made. Only relative amounts were determined. In the blank reaction, 20 times more CHFBr2 was formed than in the iron porphyrin/ cysteine system. From the halides found in the blank of

CFBr3, however, it is possible that the amount of CHFBr2 would complete the reaction mass balance: 46 ± 15% is N-formylcysteine (complete. dehalogenation delivers -50% of fluoride and -150% of bromide) and -50%

(proposed) of CHFBr2 (one bromide lost delivers -50% of bromide) would give about 50% of fluoride and 200% of bromide per CFBr3• This corresponds quite well with the amounts of halide ions actually found: 50 ± 30% of fluoride and 250 ± 50% of bromide.

In the CF2Br2 blank system, 1 F and 2 Br· were found per CF2Br2• As -50% of complete dehalogenation took place during the formation of N-for- mylcysteine (55 ± 25%) which delivers IF and 1 Br·, the other -50% of the products should be products that have lost two bromide ions, but no fluoride, for example CH2F2• Because of its high volatility, difluoromethane could, however, not be determined.

In both blank systems (CFBr3 and CF 2Br2) no carbenes were trapped. This is rather amazing as fluorinated anions tend to decay more than compa- rable chlorinated/brominated anions to form the corresponding carbenes (Table 3.5). And in the blank systems SN2-reactions are proposed (Figure 3.17) as in the case of the other polyhalogenated methanes discussed. 75

a) Cys-s· + CFBr3 Cys-S-Br + CHFBr2

Cys-s· + CHFBr2 Cys-S-Br + CH2FBr -----.. Cys-S-CHFBr Br· +

Cys-s- + CH2FBr Cys-S-CHrF + Br-

b) Cys-s· + CF2Br2 Cys-S-Br + CHF2Br

Cys-s· + CHF2Br Cys-S-Br + CH2F2

----.... Cys-S-CHF + Br· 2

Figure 3.17: SN2-reactions in the blank systems of a) CFBr3 and b) CF2Br 2•

In summary, polyhalogenated methanes exhibiting fluorine behave differently than only chlorinated and/ or brominated compounds under the same conditions. Fluorine shows special effects compared to the other halogens. For example, there is an a-fluorine effect: Fluorine strengthens bonds of other halogens (no effect for C-H bonds) at the same C-atom. A systematic increase of bond strength is found with increasing fluorine substitution at a carbon (see Table 3.8) (Slayden, 1983).

Fluorine destabilizes anions with its it-donor properties: pK.(HCF3) =

28; pK.(HCCla) = 15.5; pK.(HCBr3) = 13.7 (Klabunde and Burton, 1971). Fluorine destabilizes radical centers (radical centers are generally denoted electrophilic (Wentrup, 1979)) compared to chlorine and bromine due to its strong cr-acceptor characteristics. This gives a stability order of CF3 < CF2Cl <

CC12F < CFBr2 < CC!Br2• Moreover, CF3 radical is pyramidal, whereas CH,. or

CCl3 are planar (Pasto et al., 1987). Fluorine is also known to make weak H- 76 bridges to acids, whereas the other halogens do not14 (March, 1992) (Howard et al., 1996).

Table 3.8: Bond dissociation energies in kf mo/" 1 (Slayden, 1983)

X=F X=Cl

C-F C-H C-Cl C-H C-F C-CI

CH4 439 439 CCl4 288

CH3X 451 411 349 422 CCl3F 432 314

CH2X2 500 431 338 411 CCl2F2 472 324

CHX3 538 449 346 392 CCIF3 508 362 ex. 545 288 CF4 545

Thus, there are many differences in behaviour between fluorine and the rest of the halogens. Therefore, it is also possible that some mechanism very different from the ones proposed for CX4 and CHX3 (X=Cl, Br) takes place when fluorinated polyhalogenated methanes are degraded. Mechanisms, however, involving an inner-sphere electron transfer and the formation of an iron-fluorine bond, are rather unlikely: UV-VIS spectra did not show any

, shift when adding CFBr3 CF2Br2 or CFC13 to a reduced iron porphyrin solution containing cysteine as axial ligands. An outer-sphere electron transfer mechanism has to be considered, as in the other cases, but the follow- up reactions obviously differ from the ones proposed for CX, and CHX3 (X=Br, Cl).

14 Normally, a C(sp3)-F ·· H-0 bond is -half the strength of a C-0 ··· H-0 bond. 77

3.9 Degradation of polyhalogenated ethanes

In the previous chapters, the focus of product analysis was on polyhalogenated methanes - they are the simplest of all polyhalogented alkanes. In this chapter, polyhalogenated ethanes are discussed regarding their reactivities, their products and reaction mechanisms. Perlinger studied the degradation reaction of eight polyhalogenated ethanes, among them three fluorinated ones (Perlinger, 1994). Table 3.9 shows the degradation rates in the iron porpyhrin system and in the blank system. Reaction products, if found, are given in percent of the initial concentrations.

Table 3.9: Polyhalogenated ethanes: degradation rates and reaction products in the iron porphyrin system and in the blank system, data from (Perlinger, 1994).

compound volatile productsb1aru; productsFeP

1 CCkCCl3 4.8±0.410 a6days 90±19 % C2Cl4

1 0 CF2Cl-CCl3 1.3 ± 0.310 2days not found

CHC12-CCl3 9.5 ± 2.810° 4.8 ± 0.43 E-6 120±20% 100±29%

C2HCl3 C2Cl4

CF3-CCl3 8.6 ± 3.310° • 3 days not found a CH2Cl-CCl3 1.2 ± 0.510° 6days 120±35 %

l,l-C2H 2Cl2 2 CF2Cl-CFC~ 3.8 ± 0.6 10· a 12days not found 2 a CH,-CCl3 3.2±0.2410" 11 days not found 2 7 CHCl2-CHCl2 1.0 ± 0.25 rn- 1.6 ± 0.2 10' 60±9% C2HCl3 100±30%

C2HCl3 0 No detectable degradation within... 78

As all the experiments in Table 3.9 were performed under standard conditions (25°C, see Appendix A), reaction rates of the blank reactions were often too slow to be measured and product analysis of the blanks could therefore not be undertaken. Product analysis in the iron porpyhrin samples consisted of either extracting the solution with hexane and analysing it by GC- ECD or GC-MS or by taking headspace samples (also GC-ECD or GC-MS analysis). Thus, only volatile, non-polar products were detected. Another problem was the lack of reference substances for the analysis of certain volatile compounds that could be generated. No determination of sensitivity could be performed for possible products (Perlinger, 1994) .

.Four of the eight ethanes (CH3-CC13, CF2Cl-CFCl2, CF2Cl-CC13 and CF3- CC13) studied by Perlinger (Perlinger, 1994) were chosen to analyse for non- volatile products with ion chromatography. These four compounds did not yield any volatile products in the iron porpyhrin system (see Table 3.9). Moreover, fluorinated compounds are interesting because they showed different behaviour compared to simply brominated and/or chlorinated compounds (see chapter 3.8). In order to accelerate the degradation rates, experiments with CH3-CC13 and CF2Cl-CFCl2 were performed at 60°C (FeP system as well as blank system, see Appendix A). Generally, polyhalogenated ethanes can undergo a variety of degradation reactions: halothane (HCC1Br-CF3) was found to be transformed to trifluoroacetic acid when cytochrome P450 and oxygen were present (Nastainczyk et al., 1978). Here, an oxidation process took place. Li and

Wackett found in the degradation of CF3-CC13 by cytochrome P450 under reducing conditions CF3-CHC12 and F2C=CCl2 to be the products (Li and

Wackett, 1993). In this case CF3-CHC12 is a hydrohalogenation product, whereas F2C=CC12 is a dihalo-elimination product. Moreover, in the iron porphyrin system studied here, different processes seem to take place: whereas hexachloroethane, pentachloroethane and 1,1,1,2-tetrachloroethane 79 give their dihalo-elimination products in the iron porphyrin system, 1,1,2,2- tetrachloroethane gives the dehydrohalogenation product like in its blank reaction. This is in contrast to findings in a hematin mediated system: Schanke and Wackett found in the degradation of 1,1,2,2-tetrachloroethane

51 % of cis-C2H 2Cl2, 23% of trans-C2H 2Clv 7% of l,l,2-C2H3Cl3 and 3% of C2HC13 (Schanke and Wackett, 1992), which are dominated by dihalo-elimination products (cis- and trans- C2H2Cl2). The results of the analysis of the reaction solution by ion chromatography of the compounds that did not yield any volatile products after complete degradation are given in Table 3.10.

Table 3.10: Anions found in the degradation of CH3-CCly CF2Cl-CFCl2, CF2Cl-

CCly CF3-CCl3 compound 'productsFeP prod uctsblank

CH3-CC13 10±5% acetate, unknown 300±50% c1· product peak at 6.2 min.b 100±20% acetate

CF2Cl-CC13 110±10% F' 110±20%Cl-, 25±5% F

CF3-CC13 100±20% p-d, unknown product 100±15% c1- peak at 4.5 min.b 20±5% F

CF2Cl-CFCl2 100±15% F' no degradation in 3 months

' Chloride was not quantified in the FeP system in these experiments. b Peak in the IC-chromatogram rising proportionally to decay of parent compound. When degradation finished no further rise.

' Fluoride did not occur simultaneously with the decrease of CF2Cl- CC13, only later.

d Occured during disappearance of CF3-CC13. 'Occured during disappearance of CF2Cl-CFCl2. 80

Combining the results of Table 3.9 and Table 3.10, possible degradation mechanisms for polyhalogenated ethanes can be proposed: hexachloroethane is completely converted to perchloroethylene in the iron porpyhrin system 15 (Table 3.9) • The first step in this degradation is most probably a one-electron transfer from iron (II) porphyrin to hexachloroethane as in the degradation of polyhalogenated methanes (see chapter 3.4, 3.5, 3.6 and 3.7). In this dissociative electron transfer, a C2Cl5 radical is formed. Such radicals are quite stable as they can form bridges with the P-halogen:

1 Cl,, • / ci_..->--1- -c1 Cl

A second reduction of the radical would lead to a C2Cl5 anion. This anion can either be protonated or eliminate a chloride ion. As elimination is a unimolecular process and chloride is a good leaving group (HCl has a pl<,, of -7, (March, 1992)), elimination is expected to be much faster than protonation. Actually, deuterium exchange rates were measured for some pentahalo- ethanes (Hine et al., 1961). In an aqueous solution, the second-order rate con- stants for the deuterium exchange of CF3-COCii, CF3CDBrCl and CF3CDBr2 were 3.53 ± 0.12 10" M·1s·1, 1.11 ± 0.03 10·2 M·1s·1 and 2.16 ± 0.04 10·3 M·1s·1 respec- tively. In all cases, protonation of the carbanion was faster than elimination of fluoride. This can be rationalized: the -CF3 group is known to have very good electron acceptor properties without any electron donor characteristics (Klabunde and Burton, 1971), and fluoride is a rather bad leaving group

"An interesting mechanistic study of the degradation ofC2Cl6 is given by M. Elsner (Elsner, 1998). 81 \ considering the high bond dissociation energy of a carbon-fluorine bond (March, 1992). In the case considered here, however, elimination of chloride by forming perchloroethylene is favoured over protonation. Hine et al. state that" the rates of formation of ~-halo carbanions cannot be studied in the case of most saturated organic halides because of the incursion of a much faster concerted elimination reaction or because the competing olefin formation makes it impossible to capture and hence prove the intermediacy of the carbanion" (Hine et al., 1961). Thus, in the case of a pentachloroethylanion, elimination can be assumed to be faster than protonation, which corroborates the findings of 90 ± 10 % of perchloroethylene as the final product of hexachloroethane, but no pentachloroethane nor trichloroethene (which would be the product of the reduction of pentachloroethane, see Table 3.9).

The reaction of C2Cl5 radical with cystine (analogously to the CC14 degradation) can be excluded as trichloroacetic acid would be the final product of this reaction pathway. Another possible reaction of the C2Cl5 radical is a chloro-radical abstraction by cystine to give RS-Cl and perchloroethylene.

The proposition of a second-electron transfer to the radical C2Cl5 with elimination of chloride stands in contrast to the mechanism proposed for

CC14' where the reduction of the CC13 radical can be excluded (no CHC13, no free :CC12 carbenes trapped). As the C2Cl5 radical is stabilized by its ~-halogen, however, the probability of a second electron transfer to the radical increases. Moreover, chlorine substituents increase the electron affinity of the molecule, and in this case five chlorines are present on only two carbon atoms. A second electron transfer to the radical corresponds to an overall two-electron transfer, where the two electrons are transferred successively (Figure 3.18). 82

CCkCOOH -cy not found c1>=p Cl Cl

90± 19 % not found

Figure 3.18: Proposed reaction pathway for the reduction of hexachloroethane in the iron porpyhrin mediated system. Two electrons are transferred successively.

For pentachloroethane and 1,1,1,2-tetrachloroethane, the same mechanism is proposed as for hexachloroethane. The radicals formed in the first step are dichloroalkylradicals with ~-chlorines and are thus quite stable. Further reduction and elimination is proposed. In the case of 1,1,2,2- tetrachloroethane, however, another mechanism has to be active: the main product is trichloroethylene (60 ± 9%) as in the blank reaction (100 ± 30%).

The radical formed is a CHC12-CHC1 radical- its electron-affinity is smaller than those of the higher chlorinated radicals. The CHC12-CHCl radical is also less stable, because the radical center has a chlorine, a hydrogen and an alkyl as substituents (Jiang et al., 1989). Therefore a second electron transfer is rather difficult. Whether an H atom abstraction leads to the formation of trichloroethylene can not be concluded from the results given. In the blank 83 reaction, however, an apparent base catalysed dehydrodehalogenation converts 1,1,2,2- tetrachloroethane to trichloroethylene (100 ± 30 %). Although some information is given in Table 3.10 for the ionic products of CH3-CC13 and the fluorinated ethanes, no final mass-balance can be made and a mechanistic interpretation is rather difficult. Nevertheless, some possible pathways should be discussed. In the iron porpyhrin system, 1,1,l-trichloroethane is most probably reduced to the CH3CC12 radical and a chloride ion in the first step. The

CH3CC12 radical is not very stable (no ~-halogen) and is therefore proposed to react with cystine in a radical reaction like the CHX2 radical in the case of

CHX3 (see chapter 3.7). The resulting product would be N-acetylcysteine analogously to the degradation product found for the haloforms. Actually, the IC-chromatograms showed a peak at 6.2 minutes rising with the degradation of CH3CC13 - but the compound was not analysed further. In the case of the fluorinated ethanes one fluoride per parent compound is found to be lost during the degradation process. For CF3-CC13 and CF2Cl-CFCl2 the rising of the fluoride peak is simultaneous to the

, degradation of the substrate. For CF2Cl-CC13 however, degradation of CF2Cl- CC13 was already completed before fluoride was observed. After three months, one fluoride per CF2Cl-CC13 was found, whereas the degradation of CF2Cl-CC13 was complete within 12 hours. Reaction products could not be identified. In the case of CF3-CC13, a peak at 4.5 min. appeared with decreasing CF3-CC13• This is probably an analoguous product to the supposed water-soluble ionic

CH3-CCls-product (N-acetylcysteine): N-trifluoroacteylcysteine. In the case of the blank reaction, the only identified and quantified product was acetate (100 ± 20%) for CH3-CC13. The amount of chloride observed (300 ± 50 %) supports these findings. Here, most probably nucleophilic substitution of chlorine by hydroxide took place. Haag and Mill had similar findings: in aqueous solution nonreductive elimination 84

converted CH3-CC13 into Cfi;i-COOH (80%) and CH2=CC12 (20%) (Haag and Mill, 1988). It is obvious that these results are insufficient for a dear mechanistic interpretation. Further work has to be done. The water-soluble ionic products

13 of CH3-CCL, and CF3-CC13 should be collected and characterized. C-labeled compounds would give the necessary information in NMR-analysis of the isolated products. Standard compounds for a diverse series of chlorinated and/or fluorinated ethenes would be helpful to analyse such potential final products.

3.1 O Summary and conclusions

In the iron porphyrin/ cysteine system all ten PHMs were completely dehalogenated. Except for the fluorinated PHMs the final product was N- formykysteine, a water-soluble and non-toxic product. In the reaction with cysteine, PHMs were not completely dehalogenated resulting in a mixture of partially dehalogenated products and products resulting from carbene intermediates (HCOOH, N-formykysteine). These data support the proposition of a dissociative one-electron transfer in the case of the iron porphyrin/ cysteine system. On the other hand, in reactions of PHMs involving organic reductants exibiting mercapto groups, (here cysteine) an alternative initial reaction step is an X-philic two-electron transfer. When considering the toxicity of PHMs such reactions should be taken into account, because in organisms, reduction of PHMs could lead to the formylation of NH2 or SH groups. Such groups are present in biologically important molecules (e. g. glutathione). From a practical engineering point of view, the results of this chapter offer an interesting perspective for a fast complete dehalogenation of PHMs, 85 by using a very reactive one-electron donor (i. e., a reactive iron species) in the presence of organic compunds exhibiting reduced sulfur and nitrogen groups. Using such an approach, the formation of less halogenated volatile compounds that are orders of magnitudes less reactive than the parent compound (see Table 1.1), can be prevented. Thus, further treatment of such products is not necessary. 86

4. Kinetics and Reaction Mechanisms. Structure- Reactivity Considerations

4.1 Introduction

In this chapter the proposed outer-sphere one electron transfer reaction in the iron porphyrin/ cysteine system is discussed in more detail. In particular, kinetic aspects such as competition between PHAs and activation energies are evaluated. The degradation mechanism is compared to that of nitroaromatic compounds in the same system. Moreover, the iron porphyrin/ cysteine system is compared with other model systems, one of them being a system with a typical outer-sphere electron transfer mediator

(Co(II)W120 4/-) and one with a reduced mercaptohydroquinone, a mediator exhibiting an -SH group. Finally, relative reactivities of polyhalogenated methanes and ethanes are discussed with respect to their structure and chemical properties.

4.2 Competition

Generally, an electron transfer reaction can be split up in a sequence of several reaction steps: the approach of the electron acceptor to the electron donor (precursor complex formation), the actual electron transfer and the dissociation of the successor complex (Eberson, 1987). One of these reaction steps is rate-determining. It determines to a large extent the overall reaction rate kobs' 87

In the reduction of nitrobenzenes by hydroquinones in the presence of hydrogen sulfide where the reactive species is proposed to be a mercaptoquinone (Perlinger et al., 1996), the correlation of log k,.1 versus Ehv(X-NB) (the one electron reduction potential of the nitrobenzene X-NB; NB = nitrobenzene, X = substituent at the benzene ring) yielded a slope of 1.0 (equation 4.1) (Schwarzenbach et al., 1990).

k EI' (X - NB) log k = log ob' J= a' h + b 4.1 re1 ( RT kob,(4-Cl-NB) 2.303, - F

Moreover, in this system no competition between different nitrobenzenes was observed (Schwarzenbach et al., 1990). This suggests that the actual electron transfer is the rate-limiting step in this electron transfer reaction. The reduction of nitrobenzenes by the FeP, however, yielded a slope of only 0.6 in the correlation of log k,.1 versus Ehv(X-NB) (Schwarzenbach et al., 1990). Furthermore, it was postulated that precursor complex formation was important in determinig the overall reaction rate. This is supported by the fact that competition effects between different nitrobenzenes were found, and it was postulated that relative reactivities reflected relative affinities of the nitrobenzenes towards the iron porphyrin as ortho-substituted nitrobenzenes were found to react faster than meta- and para-substituted NBs. 88

2 1Juglooe / H,S "---a14-Ace •••

~rphyrin/ Cystcine b 2

2-CH3 -1 •

-2

-10 -9 -8 -7 -6 Ehl' (X-NB) I 0.059V

Figure 4.1: Effect of substituents on relative rates of reduction of 10 different monosubstituted NBs. (a) Juglone/ H2S system where mercaptojuglone is postulated to be the reactive species and (b) iron porphyrin/ cysteine system (Schwarzenbach et al., 1997; Schwarzenbach and Gschwend, 1990).

The determination of Ehv for polyhalogenated alkanes is rather difficult (the electron transfer is dissociative implying irreversibility), and measured standard reduction potentials vary by several volts for the same polyhalogenated compound (Curtis, 1992; Eberson, 1987). Thus, correlations of 1 log kre1 versus Eh ' (PHA) are useless for the interpretation of the electron 89 transfer mechanism. In contrast, competition experiments are good tools to gain more information about the rate-determining step, especially about the affinity of PHAs towards the iron center. Thus, competition experiments using binary and ternary mixtures of polyhalogenated alkanes were carried out. A series of competition experiments between PHAs for fast reacting substrates at 25°C and for slowly reacting substrates at 55°C is discussed in the following, including also experiments between different substrate classes such as nitrobenzenes versus polyhalogenated alkanes.

Table 4.1: Competition experiments between CBr 2Cl 2 and CBrCl 3: apparent second-order rate constants

Compound: Concentration of Apparent second- Standard initial iron porphyrin order rate Deviation concentration constant kMed (M-1s-1) (M-1s-1)

CBr2C12: lOµM 3µM 86.34 7.60

CBrC13: lOµM 3µM 60.00 4.6

CBr2Cl2: lOµM 3µM 21.20 4.43 CBrC13: lOµM 19.70 2.27

For the fast reacting substrates, no competition effect was observed in binary mixtures of CBr2Cl2 and CBrC13 (see Table 4.1). Both substrates were reduced more slowly than in experiments containing the single compound only, because the iron porphyrin delivered electrons to both substrates and had to be rereduced. Thus, while CBr2Cl2 reacted with an apparent second- order rate constant of 86 ± 8 M-1s-1 in a single-compound experiment, it reacted 1 1 with 21 ± 4 M- s- when lOµM CBrC13 are added. On the other hand, CBrCl3 90 degraded with an apparent second-order rate constant of 60 ± 5 M·1s·1 in single- compound experiments, wheras the rate was only 20 ± 3 M·1s·1 in the presence of lOµM CBr2Cl2' Neither of the substrates out-competed the other: both substrates degraded with a rate lower than in single-compound experiments due to the regeneration of the iron porphyrin. As shown in chapter 2.7, this is a rather slow process compared to the actual electron transfer from Fe(II)P to

CBr2Cl2 and CBrC13 respectively.

Table 4.2: Competition experiments with CCl4, CHBr2Cl and CHBrCl2

Compound: Apparent second- Standard Ratio initial order rate Deviation rate single compound: concentration constant kMed (Mt s·I) rate mixture 1 (M' s· )

Single compound

16.51 0.88

4.70 0.19

CHBrC12: lOµM 3.12 0.23

Binary mixtures

CCl4: 10µM 11.97 1.48 1.38 CHBr2Cl: 10µM 4.68 0.17 1.00

CCI,: lOµM 9.85 0.38 1.68

CHBrCl2:10µM 2.05 0.06 1.52

Ternary mixture

1.97 CC14: 10µM 8.40 0.16 CHBr Cl: lOµM 3.25 0.10 1.45 2 2.38 CHBrC1 2: 10µM 1.31 0.07 91

Competition experiments between slowly reacting substrates were performed in order to exclude the influence of regeneration of Fe(Il)P on the

overall rate. CC14, CHBr2Cl and CHBrCl2 were used. Apparent second-order rate constants are given in Table 4.2 for each single compound, for binary mixtures and for the ternary mixture. The experiments were carried out at 55°C. It was assumed that temperature affects the regeneration rate of the iron (Il)porpyhrin in the same way as the actual electron transfer. As can be seen in Table 4.2, in mixtures of two substrates, both sub- strates reacted with a lower rate than in single-compound experiments. In the ternary mixture the rates were even lower. The rates of substrates in mixtures were lower because the overall amount of substrate was increased in mixtures compared to single compound experiments while having the same concentration of iron porphyrin in all experiments. The ratios of the rates obtained in single-compound experiments to the rates of binary mixtures and ternary mixtures respectively were more or less the same for all three compounds used here. Thus, no substrate was favoured over the others. There was no competition effect. Also in competition experiments between polyhalogenated alkanes and nitrobenzenes no significant competition effect was found (see Table 4.3). Summarizing these facts, no significant competition between polyhalogenated alkanes could be observed, neither between fast reacting nor between slowly reacting substrates. Thus, there seems to be no specific interaction between the PHA and the iron (II) porpyhrin during electron transfer. In contrast, Castro and coworkers describe the reduction mechanism of allyl chloride, 1,1,l-trichloroethane and DDT by Fe(Il) deuterioporphyrin with a Fe-Cl bond arising (Castro, 1964). Such a bond can be formed if there are no strong ligands in the reaction solution. In the system studied by Castro the iron porpyhrin was reduced with Fe0 powder. Thus, no other potential ligand than the solvent itself (1:1 isopropyl alcohol/ acetic acid) was in the 92 solution and an Fe-Cl bond could be formed during the electron transfer. In the system used in this study, however, cysteine, a strong ligand, and phosphate are present in excess compared to the iron porpyhrin. Therefore polyhalogenated alkanes do not have access to the iron center.

Table 4.3: Competition experiments between two compound classes: polyha- logenated alkanes (here: CCl4) versus nitrobenzenes (here: 4-chloronitroben- zene)

Compound: Concentration of Apparent second- Standard initial iron porphyrin order rate Deviation 1 concentration constant kMed (M'' s· ) (M-1 s-1)

CC14: lOµM 3µM 0.80 0.05

4-Cl-NB: lOµM 3µM 6.62 0.52

CCl4: lOµM 3µM 0.77 0.05 4-Cl-NB: lOµM 4.68 0.18

This is clearly supported when comparing pH effects on PHAs with pH effects on nitroaromatic compounds. Figure 4.2 shows the pH-dependence of the rate for the reduction of 4-chloro-nitrobenzene in the iron porphyrin/ cysteine system (Schwarzenbach et al,, 1990). Compared to polyhalogenated alkanes, relative rates show a stronger pH-dependence, but the same general trend: they increase from pH 6 to 8 (data shown in Figure 2.8) and level off at higher pH-values (data not shown, Ruth Stierli, personal communication). As the spedation of the iron porphyrin is given for a given pH, the stronger pH-dependence found for nitroaromatic compounds compared to 93 polyhalogenated alkanes could be due to different electron transfer mechanisms. 4-chloro-nitrobenzene seems to approach more closely the electron transferring entity of the iron porphyrin than the reacting chlorine atom of CCI, and C2Cl6 respectively. Such a difference in proximity can be explained by the different natures of the compound classes considered: nitroaromatic compounds have a benzene ring which can form an EDA complex (EDA electron donor acceptor) with the aromatic porphyrin ring (Foster, 1969). Moreover, the nitro functional group has two negatively charged oxygen atoms that could possibly interact with the N-methyl-pyridyl groups that are also positively charged.

8,..-,..-~,--~.-~-.--~,..-~,..-~-,----,

7

'°II 6 ::i:: 5 S< 4 ~~ 3 + ...... ~~ 2 1 . • • i

6 6.5 7 7.5 8 8.5 9 9.5 pH

Figure 4.2: Second-order rate constants for 4-chloro-nitrobenzene (•), CCl4 (D) and C2Cl6 ( •) normalized to pH 6.

On the other hand, the hydrophobic surface of CC14 and C2Cl6 hinders their interaction with the polar "sites" of the iron porphyrin. There is no interaction between CCI, or C2Cl6 and the iron center (UV-VIS spectra do not show a shift of the Soret-band when adding CC14 or C2Cl6 to an anaerobic iron 94

(II) porpyhrin solution). In this case, electron transfer occurs through an outer-sphere mechanism, meaning, electrons are transferred via solution and no bond (between iron and chlorine) is formed during electron transfer.

4.3 Activation parameters

The knowledge of the activation parameters of a reaction may give important clues about the mechanism of a reaction. The free enthalpy of activation, .ilG", consists of the enthalpy of activation, AH", and the entropy of activation, .ilS" .

.ilG" = AH" - T AS" 4.2

AG" is the energy needed to reach the transition state. The differences in enthalpy between the initial ground state and the transition state, AH", are due to attraction- and repulsion-potentials between the atoms involved in the reaction, changes in solvation or differences in Coulomb energy. On the other hand AS" includes differences in order between ground states and transition states. This includes possible rotation and translation of atom groups and solvation. If the transition state is more compact than the initial state, a negative AS" is found. Different types of reaction mechanisms have characterstic activation parameters. As shown in Table 4.4, SN2-reactions of sulfide with CH2Brv 1 CH2BrCl and CH2Cl2 have an enthalpy of activation of - 90 kJ mo1- (Roberts, 1991). The corresponding entropies are negative as the transition state is more compact than the ground state: 95

HH H'~ ~ HS- + ,7-Br HS--- -- ..1 • - - --Br Br Br Ground State Transition State 4.3

In the case of CH2Cl2 AS' is more negative than for CH2Br2 and CH2BrC1. This means that the distance between HS- and CH2Cl2 is smaller in the transition state than between HS- and CH2Br2. Considering the higher bond dissociation energy of C-Cl compared to C-Br (March, 1992) this is plausible.

Table 4.4: Activation parameters of different types of reactions

Reaction AH' (kJ moi-1) AS'(Jmol-'K') Reference or E. (kJ mot')

AH'= 89 -28 Roberts, 1991 HS- + CH 2 Br 2 ~ HS-CH2-Br + Br- -25 Roberts, 1991 HS- + CH2BrCl ~ HS-CH2-Cl + Br- AH'= 89.2 1991 Hs- + CH2Cl2 ~ HS-CH2-Cl + c1- AH'= 88.9 -56 Roberts,

... E. = 122 ±32 Kriegmann- HS-+ CC14 ~ King, 1992 Roberts, 1991 Ho- + CH2Cl2 ~ HO-CH2-Cl + ci- AH'= 107 -32 -78 Roberts, 1991 Hp + CH 2Cl2~ HO-CH2-Cl + HCl AH'= 119

2 3 3 2 Waisman, Fe \q + Fe \q ~ Fe \q + Fe \q E. = 41. 3 - 104.5 1977

~ - E. = 38 Atkins, 1996 Ho- + C02 HC03 96

SN2-reactions at carbon involving other nucleophiles (hydroxide or 1 water) show enthalpies of activation of 100 to 120 kJ mo1· • Generally, SN2- reactions have enthalpies of activation between 80 and 120 kJ moi-1. 1n contrast, electron transfer reactions have, in general, much lower

2 enthalpies of activation. For example, the electron transfer from Fe + aq to Fe3\q has an energy of activation, E., of only 41.3 kJ mo\·1 (E. = LiH' + RT; RT -2.5 kJ mo1·1 at 25°C). Considering the fact that in an SN2 reaction two electrons are transferred to a cr* orbital whereas in an electron transfer reaction only one electron is transferrred (to a cr *or ad orbital as in the given example Fe2\q --7 Fe3\q ), the difference in the energy of activation is reasonable: the transfer of one electron needs less energy than the transfer of two electrons. In addition, SN2 reactions often involve quite polar transition states, which needs more overall reorganization energy because of the increased solvent reorganization energy. Addition reactions like the formation of hydrogen carbonate by hydroxide and carbon dioxide also have low activation energies. Here, two electrons are transferred to a 11:* orbital. As n* orbitals have lower energies than cr* orbitals, the activation energies are lower for addition reactions than for SN2 reactions. In order to get more information about the reaction mechanism of the degradation of CC14 and C2Cl6 in the iron porphyrin/ cysteine system studied, activation parameters were determined. Degradation experiments were carried out at 25, 35, 45, 55 and 65°C (with three replicates at each temperature). For the experiments with CCI, 30 µM FeP was used, whereas in the experiments with C2Cl6 only 10 µM FeP was used. All experiments were also carried out with only cysteine. For the exact experimental conditions and the calculation of the activation parameters see Appendix A. 97

0.01 r---.,----,----,---..,,.---,

~ 0.001 .:2 ] 0.0001

ill..._ -- --- (b) ..0"' co 10·7~--~--~--~--~--~ 0.0029 0.003 0.0031 0.0032 0.0033 0.0034 1 1/T (K )

Figure 4.3: Degradation rates of CC14 as a function of the inverse temperature: (a) with 30µM iron porphyrin, (b) without iron porphyrin. Error bars indicate one standard deviation from three replicates.

0.01 .------,----,.---.,----..,.----,

-"' ] 0.001 (a) ~..o ] 0.0001 m ..-... cs..._ -"'c: 10·5 ...... (b) ~':E,"' -.!__ 10·6 - -~ - OJ ..0"' co 10-7 - 0.0029 0.003 0.0031 0.0032 0.0033 0.0034 1/T (K1)

Figure 4.4: Degradation rates of C2Cl6 as a Junction of the inverse temperature: (a) with 10µM iron porphyrin, (b) without iron porphyrin. Error bars indicate one standard deviation from three replicates. 98

Plots of log kobs versus 1/T (Figure 4.3 and Figure 4.4) yield the activation parameters. As can be seen in Table 4.5, the enthalpies of activation for the blank reactions (reactions without iron porpyhrin) were higher than

80 k}mol·1. The reactions of cysteine with CCl4 or C2Cl6 respectively are proposed to be SN2 reactions including a thiol-attack on chlorine. Such a reaction mechanism is supported by the enthalpies of activation found, provided that X-philic SN2 reactions have the same enthalpies of activation as SN2 reactions at carbon.

Table 4.5: Activation parameters in the iron porphyrin/ cysteine system and in the blank system

ca. with 30 µM iron porphyrin 69.29 ± 7.9 -89± 25 without iron porphyrin 83.57±5.96 -95± 18 C2Cl6 with 10 µM iron porphyrin 60.65±1.12 -116±4 without iron porphyrin 97.43 ± 10.53 -48±33

The AS' values for the blank reactions were lower than the ones

found for the reactions of sulfide with CH2Br2 or CH,BrCl. In the case of CCI., a dipole is formed in the transition state including solvation of the Cl-anion

being formed. This leads to a negative AS' (-95 ± 18 J mol"1K'). For C2Cl6 a AS' value of -48 ± 33 J mo1' 1K·1 is found. In this case the standard deviation (three replicates, 95% confidence interval) is too high to give any conclusive information. 99

In the iron porphyrin/ cysteine system, the enthalpies of activation

1 were found to be significantly lower than 80 kJ mo1· • The reaction mechanism is therefore probably different from SN2. One electron transfer reactions have enthalpies of activation of about 40 kJ mo1·1 (Table 4.4). As the degradation of CCl4 and C2Cl6 have enthalpies of activation of 69 ± 8 and 61 ± 1 kJ mo1·1 respectively, there must be an additional need for energy beside the electron transfer: as radical anions of polyhalogenated alkanes are very short- lived and decay to a radical and an anion, the energy needed for bond breaking is part of the activation energy. It is known that polyhalogenated aliphatic compounds react with thermal electrons in a dissociative way (meaning electron transfer and bond breaking are concerted), whereas polyhalogenated aromatic compounds form stable radical anions (Wentworth et al., 1969). Thus, enthalpies of activation between 60 and 70 kJ mo1·1 (higher than .1.H' for typical electron transfer reactions and lower than .1.H' for typical SN2-reactions) do not contradict the mechanistic picture of a dissociative one electron transfer in the system considered. Finally, the entropies of activation are quite negative in the reactions mediated by iron porpyhrin. This is due to formation of a dipole in the transition state which increases the solvation of the radical and the anion 1 being formed. For C2Cl6 the entropy of activation is negative (-116 ± 4 J mo1· 1 K ). Such low entropies of activation are typical for one-electron transfer reactions (Waisman et al., 1977).

Activation energies were determined for the degradation of C2Cl6 in a system containing a hydroquinone as electron transfer mediator and sulfide as the bulk reductant (Perlinger, 1994) (for an exact description of the system see next chapter). Compared to the energies of activation found for the FeP system, E. values were slightly lower in the hydroquinone system. For an experiment containing 20µM of mercaptojuglone (an addition product of sulfide to the hydroquinone juglone (Perlinger et al., 1996) and lmM of

1 1 sulfide at pH 7, the E. was 52 ± 5 kJ mo1· , about 10 kJ mo1· lower than in the 100 iron porphyrin/ cysteine system. Such small differences in enthalpies of activation, however, cannot give any concluding information about differences in reaction mechanisms. In both systems, the mercaptojuglone and the iron porphyrin system, one electron transfer reactions take place (Ea < 80 kJ/mol), but the energies of activation are higher than the ones for typical one electron transfer reactions because bond breaking occures during electron transfer. Thus, in both systems an outer-sphere one-electron transfer is the main first reaction step in the degradation of PHAs.

4.4 Comparison of degradation mechanisms of polyhalogenated methanes in different model systems

Perlinger et al. studied the degradation of polyhalogenated alkanes in a system containing hydrogen sulfide (bulk reductant) and juglone (electron transfer mediator) (Perlinger, 1994; Perlinger et al., 1996). Juglone stands for 5- hydroxy-1,4-naphthoquinone and was used as a model compound to mimic humic acid or natural organic matter in general. The hypothesized reductant in the juglone/hydrogen sulfide system, however, was not the reduced juglone, but an addition product of hydrogen sulfide to juglone. There were several lines of evidence that the completely reduced 5-hydroxy-2-mercapto- 1,4-naphthoquinone (shortly referred to as mercaptojuglone) was the reactive species in the electron transfer reaction (Perlinger, 1994). 101

+8°4.4 ~ + H2S ~ ll)-.)l.SH ~SH 0 OH mercaptojuglone mercaptojuglone oxidized form completely reduced form

In a system containing 200 µM juglone, 1 mM Na2S , 50 mM HEPES (pH = 7) and a starting concentration of 1 µM for the polyhalogenated compound, degradation rates of a series of polyhalogenated methanes were determined (see Table 4.6 (Hofstetter, 1995; Perlinger, 1994)).

Table 4.6: Second-order reaction rates (k1.g) and first-order reaction rates

(kMed""' k 0 b, - kblank and kblank) for the juglone/hydrogen sulfide system (Hofstetter, 1995; Perlinger, 1994)

1 1 kJus (M.1 s·t) ~ed (s. ) kblank (s. )

6 CC14 0.95 ±0.03 3.810- 0

CBr4 too fast too fast too fast 3 5 CBr2Cl2 1600±1400 6.210" 4.510" 3 6 CBrCl3 800±10 3.210" 7.2 10" 4 6 CFBr3 160 ± 50 6.3 10" 2.610.

Under these conditions, 2% of the juglone was estimated to be converted to mercaptojuglone, which was quantified with UV-VIS spectroscopy (the extinction coefficient of juglone was used because the extinction coefficient of mercaptojuglone was unknown) (Perlinger, 1994). In order to know more about the reaction pathway in this system, reaction intermediates and products were analysed. As carbenes are key 102

16 intermediates for the differentiation between mechanisms via CX3 anions 17 and mechanisms without CX3 anions , the focus was to analyse carbenes. The same procedure was used as in the iron porphyrin system (see Appendix A). The conditions were slightly changed: 50 mM HEPES-buffer, pH= 8, 100 µM juglone, 5 mM N a 2S and 100 µM CX4 (instead of lµM CX4 as in the kinetic experiments, see also Appendix A). Table 4.7 shows the percentage of free carbenes trapped as cyclopropane in the juglone mediated system. As can be seen, in the juglone mediated system free carbenes were trapped like in the blank system. In the juglone system, less cyclopropane was found than in the blank system. For all polyhalogenated methanes, analysed carbenes were trapped except for CFBr3, where neither in the juglone system nor in the blank system any cyclopropane was found.

Table 4.7: Percentage of free carbenes trapped as cyclopropane in the jug/one mediated system compound juglone system blank (hydrogen sulfide)

CC14 4±2 % :CCl2 20±5 % :CCl2

CBr4 0.3 ± 0.5 % :CBr2 0.5 ± 0.5 % :CBr2

CBr2Cl, 1 ± 0.5 % :CCl2 2± 1 % :CC12

CBrCl3 1 ± 0.5 % :CCI, 3± 1 % :CC12

CFBr3 0% 0%

In the blank system an SN2-attack of hydrogen sulfide on a halogen of ex. is proposed analogously to the blank system containing cysteine:

"They result from a 2-electron transfer, either simultaneously or successively. " Only one electron is transferred to ex., further electrons are not transferred to the primarily formed ex, radical (Pross, 1985). 103

....--.... ~ HS- X-CX3 - HS-X + -CX3 - :CX2 + X- 4.5 t H+

Such a mechanism is supported by the fact that HCX3 was observed (see Table 4.8). In the mercaptojuglone mediated system, carbene was also trapped as cyclopropane. Table 4.6 shows that the juglone mediated reaction is

100 (CBr2Cl2), 200 (CFBr3) and 400 (CBrC13) times faster than the blank reaction. Th1:1s, only a little fraction of the carbenes found in the juglone system comes from the reaction of hydrogen sulfide with the PHA: another nucleophile, much stronger than hydrogen sulfide, must play an important role in generating CX3 anions and thus carbenes. The only strong nucleophile is mercaptojuglone itself. Therefore the hypothesis is that mercaptojuglone acts not only as a one-electron transfer mediator, but also as a nucleophile towards the polyhalogenated methanes (Figure 4.5). Table 4.8: Products and reaction intermediates in the juglone system

compound products intermediates halides halogenated methanes juglone system: mercaptojuglone/ hydrogen sulfide

40 ± 10 % CHC13 4 ± 2 % dichlorocarbene trapped

330±50 % Br 12 ±5 % CHBr3 0.3 ± 0.5 % dibromocarbene trapped 3±2 % CH2Br2

220±30 % Br- 10 ± 5 % CHBrCl:z 1 ± 0.5 % dichlorocarbene trapped 1 ± 1 % CHBr2Cl

130±20 % Br 28 ± 10 % CHC13 1 ± 0.5 % dichlorocarbene trapped blank system (hydrogen sulfide only)

50±5%CHC13 20 ± 5 % dichlorocarbene trapped

150±20 % Br- 92 ± 10 % CHBr3 0.5 ± 0.5 % dibromocarbene trapped 1±1 % CH2Br2

150 ± 50 % Br- 24 ± 6 % CHBrC12 2 ± 1 % dichlorocarbene trapped 1.5 ± 1 % CHBr2Cl

130 ±20 % Br- 81 ± 15 % CHCI3 3 ± 1 % dichlorocarbene trapped a Reaction at T 40°C. 105

M~~ ~s- X-CX3 + OH

ET C¢t, + C)4 C¢t,. + OH OH

Figure 4.5: Mercaptojuglone in its proposed double function as nucleophile (SN2-reaction) and as electron transfer mediator (ET-reaction).

Arylsulfides are known to be better nucleophiles than hydrogen sulfide (Edwards and Pearson, 1962; Swain and Scott, 1953). Edwards and Pearson give relative nucleophilicities of a series of nucleophiles for substitution at a tetrahedral carbon atom (Edwards and Pearson, 1962):

As Swain and Scott found relative nucleophilicities (relative to water) of cyanide and hydrogen sulfide to be the same (Swain and Scott, 1953), one can conclude from the combination of these findings that arylsulfides are much better nucleophiles than hydrogen sulfide. This is counter-intuitive as for example thiophenol has a pK. of 6.4 (Edwards and Pearson, 1962) and hydrogen sulfide has a pK. of 7 (March, 1992). Apparently, in this case polarizability is more important than basicity (Ritchie, 1978). 106

Table 4.8 gives an overview of the products found in the degradation of CC14, CBr4, CBr2Cl2 and CBrC13• Unfortunately, chloride and the buffer used in this experiment (HEPES) coeluted in the IC-measurements - so quantification of chloride was not possible. In Table 4.8, the halogenated products, bromide and the cyclopropanes are quantified. The amounts of cyclopropane noted are those in experiments with an olefinic carbene trap - all other amounts (given in % of the starting concentration of CX4) are found in the corresponding experiments without a trap. In accordance with the finding of carbenes in both systems (the juglone mediated system and the blank system), hydrogenolysis products are found. This is in clear contrast to the iron porphyrin system, where only in the blank system CHX3 was found, but not in the iron porpyhrin system (see Table 3.4). Thus, the hypothesis of an effective SN2-component in the juglone system is supported - and mercaptojuglone has to be this SN2-component, as no other strong nucleophiles are present. Comparing the kinetics in the juglone system with the rates found for the degradation in the iron porphyrin system (Hofstetter, 1995; Perlinger, 1994) shows that the differences in reactivity in the juglone system are much larger than in the iron porpyhrin system (see Figure 4.6). 107

Mediator: Mediator: iron porphyrin 4 mer cap tojuglone

3

cur,c1, __ 2

CilrCI3--

CBr F 3 --1 cc1 cc1 -- 3 3 cc1F,cc13

CHBr3 /- __CHBr 3 cF3cci, .-/"- 0 cct 4 --

CH,CICC!3 -l _ CHBrC!2 CCIF CCl 2 3 -- cp3cci, CFCl3 -- CH,c1cc1, -2

--CFCl3 -3

Figure 4.6: Second-order rate constants relative to CCl 4 in the iron porphyrin/ cysteine system (left hand side) and in the juglone system (right hand side) for a series of PHAs.

This implies that mercaptojuglone is more sensitive than iron porphyrin towards the structure of the polyhalogenated alkane. A greater sensitivity can be explained by a closer approach of the mediator and the substrate. In a plot of log kJuglone versus logk(Co(II)W120 4/"), where 7 Co(II)W120 40 .is a typical outer-sphere electron transfer mediator (Eberson and Ekstrom, 1988) (see Figure 4.8), the slope is -1.6, whereas the analogous plot of 108

logkFeP versus logk(Co(Il)W1z04t} has a slope of 1.0 (see Figure 4.7). A slope of

1.0 suggests that both FeP as well as Co(II)W120 4t show a similar behaviour towards the polyhalogenated methanes, in this case the behaviour being an outer-sphere one-electron transfer. In contrast, mercaptojuglone and

Co(II)W1z0 4/" apparently behave differently with respect to polyhalogenated methanes.

3.5

'""';.µ, 3 ::E 2.5 c"' 2 "".Q"" ] 1.5 _,,..._. ;.µ, ::E '·§ 0.5 ~ CHB~ ,:,1,- • mercaptoj uglone 0 bO..s -0.5 .5 -4 .3 -2 -1 0

uglvne /). Figure 4.7: Log k1 (•)and log kFeP (•)versus log k(CoW1z0 4 The slope of the regression line in the FeP system is 1, in the juglone system -1.6. 109

Figure 4.8: Structure of Co(II)W1p 4/- (Keggin structure). The Co(II) ion is situated in the center of the tetrahedron and the tungsten atoms are situated in the octahedron centers. Each corner of an octahedron represents an oxygen atom (Eberson and Ekstrom, 1988).

The product distribution in the juglone system implies that some percentage of the mercaptojuglone does react in an SN2-reaction (Table 4.8). The slope of -1.6 does not contradict the hypothesis that mercaptojuglone does not only react as electron transfer mediator, but also as a nucleophile. Another aspect is the distinction of inner- and outer-sphere electron transfer mechanisms: one could suggest that mercaptojuglone reacts in an inner- sphere one-electron transfer mechanism to explain the slope of -1.6. Such an explanation, however, is rather doubtful: for an inner-sphere electron transfer the semiquinone of mercaptojuglone is the species to attack the polyhalogenated methane. The completely reduced mercaptojuglone cannot act in an inner-sphere one-electron transfer. It always delivers two electrons at a time when an S-X bond is formed18 (see Figure 4.9).

18 If only one electron is transferred a very unstable [R-S-Cll' radical anion would be formed, which is energetically not favoured. 110

The semiquinone of mercaptojuglone, however, is not a dominant species in the solution composition: semiquinones were not detected by EPR (Perlinger, 1994).

~ S-X + -cx3 ~ r OH + CX4 co s- coOH OHO- completely ~l

~ + ET ~+ CX4 s· inner- ~S-X OH sphere OH mercaptojuglone semiquinone

Figure 4.9: Completely reduced mercaptojuglone can react as a nucleophile or electron transfer mediator. As an electron transfer mediator, the electron transfer is outer-sphere, because only the semiquinone can act as an inner- sphere electron transfer mediator.

In conclusion, iron porpyhrin is supposed to react in an outer-sphere 7 electron transfer mechanism like Co(II)W 120 40 •• Mercaptojuglone, however, most probably acts as an outer-sphere one-electron transfer mediator and as a nudeophile. This conclusion is both supported by the greater sensitivity of mercaptojuglone towards polyhalogenated alkanes (see Figure 4.6) and by the products and reaction intermediates found in this system. The latter analyses indicate that the SN2-reactivity of the mercaptojuglone does not exceed 10% of the overall reactivity. 111

4.5 Structure- Reactivity Considerations

Table 4.9 shows the reactivities of a series of C1- and C2-polyhalogen- ated alkanes in the iron porphyrin/ cysteine system (compounds in one group are arranged with increasing reactivity). In this system a range of reac- tivity of 4 orders of magnitudes is observed. As a dissociative outer-sphere electron transfer is proposed to be the rate-determinig step, the rate should be a function of both the electron affinity of the PHA and the bond dissociation energy of the bond broken in the transition state. Only very few data on elec- tron affinity are available for PHAs (Chen and Wentworth, 1989). On the other hand, data of bond-dissociation energies are available, but they vary by up to 30 kJ/mol for the same compound (Daubert and Dauner, 1994; Denisov, 1995; Franklin and Huybrechts, 1969; Luke et al., 1987; TRC, ). Thus, it is rather difficult to generate quantitative structure reactivity relationships (Peijnenburg et al., 1991). However, a qualitative interpretation of the kinetic data can be made.

4.5.1 Polyhalogenated methanes

The polyhalogenated methanes can be grouped into two categories: those containing bromide as the leaving group and those containig chloride as the leaving group. The bond dissociation energies of R-X follow the order: C-F > C-Cl > C-Br. Thus, the rate of cleaving a C-Br bond should exceed that of the cleavage of a C-Cl bond. In fact, the most reactive PHMs are the bromi- nated ones: CBr4, CBr2Cl2, CBrC13 and CBr3F. As summarized by Slayden et al. (Slayden, 1983), increasing the num- ber of fluorines on a carbon atom in a PHA increases the C-F and C-Cl bond dissociation energies at this carbon (a-fluorine-effect), but does not have a systematic effect on the C-H bond dissociation energy (see Table 3.8). As can be seen in Table 4.9 CFClo is reduced much more slowly compared to CC14 (see

also CFBr3 compared to CBr4 which was reduced faster than could be measured). Thus substituting a bromine or a chlorine atom by a fluorine atom slows down the rate markedly. 112

Table 4.9: Second-order rate constants for the reaction of halogenated methanes and ethanes with iron porphyrin (kFeP) and without iron por- phyrin (kblank) at 20"C. For kFeP the ± 95 % C.I. are given (Hofstetter, 1995; Perlinger, 1994) compound name formula

chloroform CHC13 n.m.r. c n.m.r. c

1 6 fluorotrichloromethane CFC13 (1.0 ± 0.1) x 10· 1.1x10" bromodichloromethane CHBrCl2 (5.8 ± 0.9) x 10° d 9 days tetrachloromethane co. (5.9 ± 0.9) x 10° d 6 days

6 tribromomethane CHBr3 (8.9 ± 2.1) x 10° 1x10" chlorodibromomethane CHBr2Cl (1.2 ± 0.3) x 101 d 3 days

6 fluorotribromomethane CFBr3 (6.9 ± 2.8) x 101 5.9 x 10"

2 bromotrichloromethane CBrCl3 (2.5 ± 0.6) x 10 5.3x10-s

2 4 dibromodichloromethane CBr2Cl2 (6.1±2.2) x 10 2.6 x 10" tetrabromomethane CBr4 - b - b 1,1,2,2-tetrachloroethane CHC1 -CHCl (1.0 ± 0.25) x 10-2 (1.6±0.2) x 10"7 2 2

2 1, 1,1-trichloroethane CCkCH3 (3.2 ± 0.24) x 10· d 11 days

2 l,1,2-trichlorotrifl uoroethane CCl2F-CClF2 (3.8 ± 0.6) x rn- d 12 days

1,1,1,2-tetrachloroethane CC13-CC1H2 (1.2 ± 0.5) x 10° d 6 days

1, 1, 1-trichlorotrifluoroethane CC13-CF3 (8.6 ± 3.3) x 10° d3 days

6 pentachloroethane CHCl2-CCl3 (9 .5 ± 2.8) x 10° (4.8±0.43) x 10- 1 l, 1-difluorotetrachloroethane CF2Cl-CCl3 (1.3 ± 0.3) x 10 d 2 days hexachloroethane CC1 3-CC13 (4.8 ± 0.4) x 101 d 6 days a At pH 7 .0, 50 mM phosphate buffer, 5 mM cysteine, 25°C; b too fast to be measured; c n.m.r. = no measurable rate. There was no change in concentration measured within 8 weeks.d No degradation within ... 113

The effect of slowing down the rate is even more pronounced by replacing a bromine (or chlorine) atom by a hydrogen (compare CBr4 with

CHBr3 or CC14 with CHC13 that reacted too slowly to be measured). The electron affinity of PHMs containing hydrogen is lower than that of comparable fully halogenated ones. Moreover, hydrogen is unable to stabilize 19 the radical center resulting from the first electron transfer , whereas for example chlorine or bromine can stabilize the radical intermediate through lt- bonding with their free electron pairs.

4.5.2 Polyhalogenated ethanes

In all cases investigated, the leaving group is a chloride ion. Most of the polyhalogenated ethanes contain a CC13 group. In the series CCl3-CCl3 -t

CHCl2-CC13 --+ CH2Cl-CC13 --+ CH3-CCI:i the rate drops by a factor of 5 between both the first and the second and the second and the third compound.

Between CH2Cl-CC13 and CH3-CC13 there is factor of - 40. The first three compounds are able to form bridged radicals due to electron pairs on the halogen atoms. The CH3-group can not easily form such bridged structures.

In the case of CF2Cl-CFC1 21 a fluorine atom is placed at the radical center, destabilizing it Giang et al., 1989; Pasto et al., 1987). Nevertheless, this radical can be stabilized by forming bridged structures (see Figure 3.18). The destabilizing effect of the fluorine atom at the radical center is compensated by the bridge stabilization. If a chlorine at the radical center is replaced by hydrogen, the rate is even lower than for the corresponding replacement by fluorine. This is in agreement with the series of radical stabilities given for the methyl radicals in footnote 19.

19 Jiang et al. (Jiang et al., 1989) observed the following order of relative stabilities of fluorinated an:! chlorinated radicals through pyrolysis experiments: F,C < H,C < F2CH < FCH2 < CICH, < Cl,CH < c1,c The order of radical stabilities shows that chlorine atoms have the largest stabilizing effect on a carbon- centered radical, much greater than fluorine or hydrogen atoms. 114

5. Conclusions

The results of this work are important from both a basic scientific as well as from a practical point of view. As has been demonstrated with nitroaromatic compounds, iron porphyrin cannot only be used as a model reductant for biological molecules exhibiting an iron center but it may also give insight into the first step in the reductive transformation of PHAs with other environmentally important reductants, namely Fe (II) adsorbed on iron oxide surfaces. Cysteine was used as the bulk reductant in the model system representing reduced sulfur species. The results support the conclusions drawn from earlier studies, namely that the one-electron transfer from the iron porphyrin to the PHA is the rate-limiting step. Moreover, activation pa!iameters and the lack of competition effects give hints that the reaction in this model system is an outer-sphere dissociative one-electron transfer reaction. Structure-reactivity relationships were made on a qualitative basis through use of thermochemical data. For a quantitative description of rate constants, however, the presently available data for molecular descriptors (such as redox potentials, electron affinities, bond dissociation energies) are insufficient. The results of this work give also insight into reactions of radicals and carbenes formed during reductive dehalogenation of PHAs in the presence of organic compounds exhibiting groups that may react with these intermediates other than by H-abstraction. It was found that PHAs were converted to less halogenated compounds or even completely dehalogenated compounds depending on the solution conditions: in the presence of iron porphyrin and cysteine, PHMs were transformed to an N-formyl group, except for PHMs comtaining fluorine. The initially formed radicals in this process are not directly reduced in a second one-electron transfer since neither free carbenes nor less halogenated products have been found. They possibly further react, 115 however, in a radical reaction w~th cystine or a cysteine radical. On the other hand, PHMs are converted to less halogenated products in the presence of cysteine. Here, an SN2 reaction of the thiol group at the halogen is proposed. Reactions with other electron transfer mediators can also be important in the degradation of PHMs, both in organisms as well as in the environment. Mercapto groups attached to aromatic compounds as for example found in natural organic matter are very good nucleophiles towards PHMs in X-philic reactions. Such reactions lead, however, to degradation products that are less halogenated and thus much less reactive than the parent compounds. On the other hand, aromatic compounds exhibiting mercapto groups do not only react in an X-philic reaction towards PHMs, they are also very effective one-electron donors which leads to reactive radical intermediates and thus to harmless products provided suitable partners for radical reactions are present. Comparing the relative reactivities of PHAs in the iron porphyrin/ cysteine system with the ones of another model system (mercaptojuglone/ hydrogen sulfide system) it was found that (i) a much larger range of reactiv- ity is observed for a given set of model compounds in the mercaptojuglone system as compared to the iron porphyrin system and (ii) that free carbenes were found in the mercaptojuglone system in contrast to the iron porphyrin system. Such findings could be explained by a closer interaction between mercaptojuglone and the PHA than between the iron porphyrin and the PHA and/ or another reaction mechanism. The mechanism for the mercaptojuglone system is proposed to consist of two parallel reactions, namely an outer-sphere one-electron transfer and an X-philic reaction. The former is supported by the facts that the same order of reactivities for the PHAs was found like in the iron porphyrin/ cysteine system and the activation energies are significantly lower than the ones for typical SN2 reactions, whereas the latter is supported by the finding of free carbenes as intermediates. In the iron porphyrin system, however, a mere dissociative outer-sphere electron transfer is proposed. 116

These mechanistic insights have practical implications: first, in the presence of reactive one-electron donors and components with reduced sulfur and nitro groups PHMs are reduced to halogenated radical intermediates which are, for example, turned into formyl groups. Providing similar conditions in a contaminated aquifer can lead to an effective removal of such compounds. Secondly, reductive transformation of PHAs can be coupled with the oxidation of other pollutants in a system containing iron reducing bacteria (Figure 5.1), which is interesting for remediation purposes. Third, the results gained in this work, especially the relative reactivities of PHMs in different model systems, provide knowledge for the evaluation of other systems.

PHA C02

iron reducing bacteria

HCOOH CHzO

Figure 5.1: The reduction of PHAs in an anoxic aquifer containing reduced iron adsorbed to mineral surfaces and iron reducing bacteria can be coupled with the oxidation of organic pollutants.

In future studies, the focus should be set on product analyses of fluorinated compounds because they show different behavior compared to non-fluorinated PHAs with respect to reduction rates, reaction intermediates and final products. Moreover, a closer look should be taken at polyhalogenated ethanes. As far as that only little is known about reaction intermediates and products as a function of different reductants and other water constituents. Another interesting aspect is the reductive transformation of PHMs in the presence of other organic pollutants like NACs as NACs are known to end up in anilines exhibiting amino groups that could react with the primarily formed PHA radicals or with carbenes to give formylated anilines. 117

6. References

Alfassi, Z. B., 1988, Chemical Kinetics of Small Organic Radicals, CRC-Press Boca Raton. Ando, W., and Moro-Oka, Y., 1988, The Role of Oxygen in Chemistry and Biochemistry: Berlin, Springer-Verlag. ApWel, R., and SchO!er, H., 1977, Notiz Uber ein bequemes Verfahren zur Darstellung von Dichlorphosphoranen: Chemische Berichte, v. 110, p. 2382-2384. Atkins, P. W., 1996, Physikalische Chemie, VCH. Barbash, J. E., and Reinhard, M., 1989, Abiotic Dehalogenation of 1,2-Dichloroethane and 1,2-Dibromoethane in Aqueous Solution Containing Hydrogen Sulfide: Environmental Science and Technology, v. 23, p. 1349-1357. Beadle, C. D., Bowman, R., W., and Prousek, J., 1984, Reaction of nucleophiles with electron acceptors by SN2 or single electron transfer (S. E.T.) mechanisms: Thiolates and 2-halomethyl-5-nitrofurans: Tetrahedron Letters, v. 25, p. 4979- 4982. Bilhler, R. E., 1983, Radiation chemistry of the carbon-halogen bond, in Patai, S. a. R., Zui, ed., The chemistry of the carbon-halogen bond, Part 2. The chemistry of functional groups, p. 795-864. Castro, C. E., 1964, The Rapid Oxidation of Iron(Il) Porphyrins by Alkyl Halides. A Possible Mode of Intoxication of Organisms by Alkyl Halides: Journal of the American Chemical Society, v. 86, p. 2310. Castro, C. E., Wade, R. S., and Belser, N. 0., 1985, Biodehalogenation: Reactions of Cytochrome P-450 with Polyhalomethanes: Biochemistry, v. 24, p. 204 - 210. Castro, D. E., and Kray, W. C. J., 1963, Journal of the American Chemical Society, v. 85, p. 2768-2773. Chen, E. C. M., and Wentworth, W. E., 1989, Mo!. Cryst. Liq. Cryst., v. 171, p. 271. Chiu, P.-C., and Reinhard, M., 1996, Transformation of Carbon Tetrachloride by Reduced Vitamin B12 in Aqueous Cysteine Solution: Environmental Science and Technology, v. 30, p. 1882-1889. Christensen, T. H., Kjeldsen, P., Albrechtsen, H.-J., Heron, G., Nielsen, P. H., Bjerg, P. L., and Holm, P. E., 1994, Attenuation of landfill leachate pollutants in aquifers: Critical Reviews in Environmental Science and Technology, v. 24, p. 119-202. 118

Cilento, G., 1960, The Expansion of the Sulfur Outer Shell: Chemical Reviews, v. 60, p. 147-167. Clark, M. W., Taylor, J. F., Davies, T. H., and Vestling, C. S., 1940, Metlloporphyrins, Coordination with Nitrogenous Bases. Theoretical Relations.: Journal of Biological Chemistry, v. 135, p. 543-568. Criddle, C. S., and McCarty, P. L., 1991, Electrolytic Model System for Reductive Dehalogenation in Aqueous Environments: Environmental Science and Technology, v. 25, p. 973-978. Criddle, C. S., McCarty, P. L., Elliott, C. M., and Barker, J. F., 1986, Reduction of Hexachlorethane to Tetrachlorethylene in Groundwater: Journal of Contaminant Hydrology, v. 1, p. 133 - 142. Curtis, G. P., 1992, Reductive dehalogenation of hexachloroethane and carbon tetrachloride by aquifer sand and humic acid: Unpub. Ph. D. Dissertation thesis, Stanford University. Curtis, G. P., and Reinhard, M., 1994, Reductive Dehalogenation of Hexachloroethane, Carbon Tetrachloride, and Bromoforrn by Anthrahydroquinone Disulfonate and Humic Acid: Environmental Science and Technology, v. 28, p. 2393-2401. Daubert, T. E., and Dauner, R. P., 1994, Physical and Thermodynamic Properties of Pure Chemicals, Design Institute for Phys. Prop. Data. Deriisov, E. T., 1995, Evaluation of Dissociation Energies of C-X (X= Cl, Br, I) Bonds from Kinetic Data for Radical Abstraction Reactions: Chemical Kinetics and Catalysis, v. 69, p. 436-440. Dumnivant, F. M., Schwarzenbach, R. P., and Macalady, D. L., 1992, Reduction of Substituted Nitrobenzenes in Aqueous Solutions Containing Natural Organic Matter: Environmental Science and Technology, v. 26, p. 2133 - 2141. Eberson, L., 1982, Electron- Transfer Reactions in Organic Chemistry: Advances in Physical Organic Chemistry, v. 18, p. 79 - 185. Eberson, L., 1987, Electron Transfer Reactions in Organic Chemistry.: Springer Verlag Berlin. Eberson, L., and Ekstrom, M., 1988, Electron Transfer Reactions in Organic chemistry: The Reaction between carbon Tetrabromide and the Heteropoly Blues 7 Co(II)W12 0 40 - and Co(II)W12 0 40 8- KINETIC AND PRODUCT STUDY: Acta Chimica Scandinavica, v. B 42, p. 101-112. Edwards, J. 0., and Pearson, R., G., 1962, The Factors Determining Nucleophilic Reactivities: Journal of the American Chemical Society, v. 84, p. 16-24. 119

Ekstrom, M., 1988, Electron Transfer in Reductions of Polhalogenated Alkanes: Unpub. Dissertation thesis, Lund University, Chemical Center. Elsner, M., 1998, Reduktive Dehalogenierung von polyhalogenierten Ethanen in Aquifer- Saulen: Unpub. diploma work thesis, ETH Zurich. Fisher, D. A., and Midgley, P. M., 1993, The production and release to the atmosphere of CFCs 113, 114 and 115.: Atmospheric Environment, v. 27 A, p. 271 -276. Foster, R. E., 1969, Organic Charge Transfer Complexes: New York, Academic Press. Frankel, M. B., Feuer, H., and Bank, J., 1959, The Preparation of N-disubstitutd Formamides: Tetrahedron Letters, v. 7, p. 5-7. Franklin, J. A., and Huybrechts, G. H., 1969, Bond Energies in Chloroethanes and Chloroethyl Radicals: International Journal of Chemical Kinetics, v. l, p. 3-9. Fujita, T., Watanabe, S., Suga, K., and Sughara, K., 1981, The Reaction of 3-Alkenoic Acids with Dichlorocarbene: Synthesis, v. 12, p. 1004-1005. Gantzer, C. J., and Wackett, L. P., 1991, Reductive Dechlorination Catalyzed by Bacterial Transition-Metal Coenzymes: Environmental Science and Technology, v. 25, p. 715 - 722. Glod, G., Angst, W., and Schwarzenbach, R., 1996, Corrinoid-Mediated Reduction of Tetrachlorethene, Trichloroethene and Trichlorofluoroethene in Homogenous Aqueous Solution: Reaction Kinetics and Reaction Mechanisms: Environmental Science and Technology. Glod, G., Brodmann, U., Angst, W., Holliger, C., and Schwarzenbach, R., 1997, Cobalamin-Mediated Reduction of cis- and lrans-Dichloroethene, 1,1- Dichloroethene, and Vinyl Chloride in Homogenous Aqueous Solution: Reaction Kinetics and Mechanistic Considerations: Environmental Science and Technology, v. 31, p. 3154-3160. Haag, W. R., and Mill, T., 1988, Effect of subsurface sediments on hydrolysis of haloalkanes and epoxide: Environmental Science and Technology, v. 22, p. 658-663. Hanzlik, R. P., 1981, Reactivity and Toxicity among Halogenated Methanes and related Compounds: Biochemical Pharmacology, v. 30, p. 3027-3030. Heijman, C. G., Schwarzenbach, R. P., Holliger, C., Glaus, M.A., and Zeyer, J., 1993, Abiotic Reduction of 4-Chloronitrobenzene to 4-Chloroaniline in a Dissimilatory Iron-Reducing Enrichment Culture: Applied and Environmental Microbiology, p. 4350-4353. Henschler, D., 1994, Toxikologie chlororganischer Verbindungen, VCH, p. 99. 120

Hine, J., Burske, N. W., Hine, M., and Langford, P. B., 1957, The Relative Rates of Formation of Carbanions by Haloforms: Journal of the American Chemical Society, v. 79, p. 1406-1412. Hine, J., and Ehrenson, S. J., 1958, The Effect of Structure on the Relative Stability of Dihalomethylenes: Journal of the American Chemical Society, v. 80, p. 824-830. Hine, J., Wiesboeck, R., and Ghirardelli, R. G., 1961, The Kinetics of the Base-catalyzed Deuterium Exchange of 2.2-Dihalo-l,l,l-trifluoroethanes: Journal of the American Chemical Society, v. 83, p. 1219-1222. Hoffmann, R., W., 1976, Aufkliirung von Reaktionsmechanismen: Stuttgart, Giitz, Adam. Hofstetter, T., 1995, Kinetics of the reduction of polyhalogenated methanes in model systems: Unpub. Diploma work thesis, ETH Zuerich. Holliger, C., Schraa, G., Starns, A. J. M., and Zehnder, A. J. B., 1993, A highly purified enrichment culture couples the reductive dechlorination of tetrachloroethylene to growth: Appl. Environ. Microbiol., v. 59, p. 2991 - 2997. Howard, J. A. K, Hoy, V. J., 0 Hagan, D., and Smith, G. T., 1996, How good is fluorine as a hydrogen bond acceptor?: Tetrahedron, v. 52, p. 12613-12622. Huheey, J.E., 1980, Inorganic Chemistry. Jafvert, C. T., and Wolfe, L. N., 1987, Degradation of Selected Halogenated Ethanes in Anoxic Sediment-Water Systems: Environmental Toxicology and Chemistry, v. 6, p. 827 - 837. Jeffers, P. M., Ward, L. M., Woytowitch, L. M., and Wolfe, L. N., 1989, Homogenous Hydrolysis Rate Constants for Selected Chlorinated Methanes, Ethanes, Ethenes, and Propanes: Environmental Science and Technology, v. 23, p. 965· 969. Jiang, X.-K, Li, X.-Y., and Wang, K-Y., 1989, Reversal of the Nature of Substituent Effect by Changing the Number of the a:-Substituent. Relative Ease of Formation of the Three a:- Fluoromethyl Radicals: Journal of Organic Chemistry, v. 54, p. 5648-5650. Kadish, K. M., and Davis, D. G., 1973, Electrochemical Studies of Metalloporphyrins: Annals New York Academy of Scienses, v. 206, p. 495-502. Kaim, W., and Schwederski, B., 1991, Bioanorganische Chemie: Stuttgart. Ketterer, B., and Mulder, J. G., 1990, Glutathione Conjugation, in Mulder, J. G., ed., Conjugation Reactions in Drug Metabolism: London, New York, Philadelphia, Taylor and Francis, p. 307 -364. Klabunde, K J., and Burton, D. J., 1971, Acidities of Polyfluorinated Hydrocarbons. Il. Hexafluoropropanes, Trifluoroethanes, and Haloforms. Intermediate 121

Carbanion Stability and Geometry: Journal of the American Chemical Society, v. 94, p. 5985-5990. Klausen, J., 1995, Abiotic Redox Transformations of Aromatic Nitro and Amino Compounds in Suspensions of Soil Minerals, ETH Zurich. Klecka, G. M., and Gonsior, S. J., 1984, Reductive Dechlorination of chlorinated Methanes and Ethanes by Reduced Iron (II) Porphyrins: Chemosphere, v. 13, p. 391 - 402. Kl0ter, G., and Seppelt, K., 1979, Trifluoromethanol (CF30H) and Trifluoromethylamine (CF3NH2): Journal of the American Chemical Society, v. 101, p. 347-349. Kriegmann-King, M., and Reinhard, M., 1992, Transformation of Carbon Tetrachloride in the Presence of Sulfite, Biotite, and Vermiculite: Environmental Science and Technology, v. 26, p. 2189-2206. Kriegmann-King, M. R., and Reinhard, M., 1994, Transformation of Carbon Tetrachloride by Pyrite in Aqueous Solution: Environmental Science and Technology, v. 28, p. 692-700. Kuhn, E. P., and Suflita, J. M., 1989, Dehalogenation of Pesticides by Anaerobic Microorganisms in Soils and Groundwater- A Review, Reactions and Movements of Organic chemicals in Soils: Madison, p. 111-180. Laidler, K. J., 1987, Chemical Kinetics. Landolt, and Bornstein, 1987, Radical reactions : New Series, 13. Lever, A. B. P., and Gray, H. B., 1983, Iron Porphyrins, Part!: Massachusetts. Lewis, T. A., Morra, M. J., Habdas, J., Czuchajowski, L., and Brown, P. D., 1995, Reductive Dechlorination of Carbon Tetrachloride Mediated by Cationic Water-Soluble Metalloporphyrins: Journal of Environmental Quality, v. 24, p. 56-61. Li, S., and Wackett, L. P., 1993, Recuctive Dehalogenation by Cytochrome P450 CAM: Substrate Binding and Catalysis: Biochemistry, v. 32, p. 9355 - 9361. Li, X.-y., Jiang, X.-k., Pan, H.-q., Hu, J.-s., and Fu, W.-m., 1987, Nucleophilic substitution of perhalofluoroalkanes initiated by halophilic attacks: Pure and Applied Chemistry, v. 59, p. 1015-1020. Luke, B. T., Loew, G. H., and McLean, A. D., 1987, Theoretical Investigations of the Anaerobic Reduction of Halogenated Alkanes by Cytochrome P450. 1. Structures, Inversion Barriers, and Heats of Formation of Halomethyl Radicals: Journal of the American Chmical Society, v. 109, p. 1307-1317. 122

Mackay, D., Shiu, W. Y., and Ma, K. C., 1993, Illustrated Handbook of Physical- chemical Properties and Environmental fate for Organic Chemicals, Lewis Publishers. Mansuy, D., 1980, New lron-Porphyrin Complexes with Metal-Carbon Bond - Biological Implications: Pure and Applied Chemistry, v. 52, p. 681 - 690. March, J., 1992, Advanced Organic Chemistry. Marks, T. S., Allpress, J. D., and Maule, A., 1989, Dehalogenation of Lindane by a Variety of Porphyrins and Corrins: Applied and Environmental Microbiology, p. 1258 -1261. McCarty, P. L., Reinhard, M., and Rittmann, B. E., 1981, Trace organics in groundwater: Environmental Science and Technology, v. 15, p. 40- 51. Moore, G. R., and Pettigrew, G. W., 1990, Cytochromes c: Berlin, Springer-Verlag. Nastainczyk, N., Ullrich, V., and Sies, H., 1978, Effect of oxygen concentration on the reaction of halothane with cytochrome P450 in liver microsomes and isolated perfused rat liver.: Biochem. Pharmocol., v. 27, p. 387-392. Nowell, L. H., and Hoigne, J., 1992, Photolysis of aqueous chlorine at sunlight and ultraviolet wavelengths -1. Degradation rates.: Water Res., v. 26, p. 593-598. Pasto, D. J., Krasnansky, R., and Zercher, C., 1987, Stabilization Energies and Structures of Substituted Methyl Radicals: Journal of Organic Chemistry, v. 52, p. 3062-3072. Pecher, K., Waxweiler, T., Haderlein, S. B., and Schwarzenbach, R. P., 1997, in preparation: . Peijrlenburg, W. J. G. M., 't Hart, M. J., Hollander, H. A., Meent, D., Verboom, H. h., and Wolfe, L. N., 1991, QSARs for predicting biotic and abiotic reductive transformation rate constants of halogenated hydrocarbons in anoxic sediment systems: The Science of the Total Environment, v. 109/110, p. 283-300. Penkett, S. A., 1982, Non-methane organics in the remote troposphere, in Goldberg, E. D., ed., Atmospheric Chemistry: Berlin, Springer, p. 329-355. Perlihger, J., 1994, Reduction of Polyhalogenated Alkanes by Electron Transfer Mediators in Aqueous Solution: Unpub. Dissertation thesis, ETH Zurich, Diss. Nr. 10892. Perlihger, J., Angst, W., and Schwarzenbach, R. P., 1996, Kinetics of the Reduction of Hexachloroethane by Juglone in Solutions Containing ·Hydrogen Sulfide: Environmental Science and Technology, v. 30, p. 3408-3417. Pretsch, Clerc, Seib!, and Simon, 1986, Tabellen zur Strukturaufkliirung organischer Verbindungen mit spektroskopischen Verbindungen. 123

Pross, A., 1985, The Single Electron Shift as a Fundamental Process in Organic Chemistry: The Relationship between Polar and Electron-Transfer Pathways: Acc. Chem. Res., v. 18, p. 212-219. Randall, C., and Baselt, R., 1993, Disposition of toxic drugs and chemicals in man: New York. Reichert, P., 1994, Sehr. EA WAG, Diibendorf, Switzerland. Rico, I., D., C., and Wakselman, C., 1983, Reactivity of the Perhaloalkanes CF2X2 (X=Cl, Br) with nucleophiles. 6. Coexistence of Carbene and Radical Processes initiated by Single-Electron Transfer.: Journal of Organic Chemistry, v. 48, p. 1979-1982. Rippen, 1988, Handbuch Umweltchemikalien.

Ritchie, C. D. 1 1978, Problems involved in understanding orders of Nucleophilic Reactivity: Pure & Applied Chemistry, v. 50, p. 1281-1290. Roberts, A. L., Jeffers, P. M., Wolfe, N. L., and Gschwend, P. M., 1993, Structure- Reactivity Relationships in Dehydrohalogenation Reactions of Polychlorinated and Polybrominated Alkanes: Critical Reviews in Environmental Science and Technology, v. 23 (1), p. 1-39. Roberts, L.A., 1991, Dehalogenation Reactions of Polyhalogenated Alkanes in Aquatic Environments: Dissertation thesis, Massachusetts Institute of Technology. Roberts, L. A., and Gschwend, P. M., 1994, Interaction of abiotic and microbial processes in hexachloroethane reduction in groundwater: Journal of Contaminant Hydrology, v. 16, p. 157-174. Rtigge, K., 1997, Fate of xenobiotic compounds downgradient of the Grindsted landfill, Denmark: an anaerobic field injection experiment. Riigge, K., Bjerg, P. L., Mosbrek, H., and Christensen, T. H., 1995, Natural attenuation of xenobiotic compounds: anaerobic field injection experiment, in Hinchee, R. E., Wilson, J. T., and Downey, D. C., eds., Intrinsic Bioremediation, 1: Columbus, Ohio, Batelle Press, p. 127-124. Sthanke, C. A., and Wackett, L. P., 1992, Environmental Reductive Elimination Reactions of Polychlorinated Ethanes Mimicked by Transition-Metal Coenzymes: Environmental Science and Technology, v. 26, p. 830 - 833. Schoder, A., 1975, Metallaustausch von Porphyrinkomplexen Kinetik, Mechanismus, Ligandkatalyse: Unpub. Dissertation thesis, ETHZ. Schumacher, W., and Holliger, C., 1996, The proton/electron ratio of the menaquinone- dependent electron transport from dihydrogen to tetrachloroethene in "Dehalobacter restrictus": Journal of Bacteriology, v. 178, p. 2328-2333. 124

Schwarzenbach, R. P., Angst, W., Holliger, C., Hug, S. J., and Klausen, J., 1997, Reductive transformations of anthropogenic chemicals in natural and technical systems: Chimia. Dec, v. 51, p. 908-914. Schwarzenbach, R. P., and Gschwend, P. M., 1990, Chemical Transformations of Organic Pollutants in the Aquatic Environment, in Stumm, W., ed., Aquatic Chemical Kinetics Reaction Rates of Processes in Natural Waters, p. 199-233. Schwarzenbach, R. P., Gschwend, P. M., and Imboden, D. M., 1993, Environmental Organic Chemistry: New York. Schwarzenbach, R. P., Stierli, R., Lanz, K., and Zeyer, J., 1990, Quinone and Iron Porphyrin Mediated Reduction of Nitroaromatic Compounds in Homogeneous Aqueous Solution: Environmental Science and Technology, v. 24, p. 1566 - 1574. Sigg, L., and Stumm, W., 1996, Aquatische Chemie. Eine Einfuhrung in die Chemie wassriger Li:isungen und natiirlicher Gewasser: Zurich, Stuttgart. Slagle, J. D., Huang, T. T.-S., and Franzus, B., 1981, Mechanism of the Triphenylphosphine-Tetrachloromethane-Alcohol Reaction: Pericyclic or Clustered Ion Pairs?: Journal of Organic Chemistry, v. 46, p. 3526-3530. Slayden, S. W., 1983, Thermochemistry of halogenated organic compounds, in Patai, S. a. R., Zui, ed., The chemistry of halides, pseudo-halides and azides, Supplement D2. The chemistry of functional groups, p. 361-402. Smith, R. M., and Martell, A. E., 1989, Critical Stability Constants: New York. Stumm, W., 1992, Chemistry of the Solid-Water Interface: New York, Wiley Intersdence. Swain, C. G., and Scott, CB., 1953, Journal of the American Chemical Society, v. 75, p. 141-147. Swallow, A. J., 1978, Reactions of free radicals produced from organic compounds in aqueous solution by means of radiation, in Jennings, K R. a. P., D., ed., Progress in reaction kinetics, 9: Pergamon Press, Oxford, New York, p. 195- 366. TRC, Thermodynamic Tables, The Texas A + M University System College Station TX 77843-3111. Tsani-Bazaca, S. A., Mcintyre, J., Lesler, J., and Perry, R., 1982, ambient concentrations and correlations of hydrocarbons and halocarbons in the vicinity of an aeroport.: Chemosphere, v. 11, p. 11-23. Vogel, T. M., Criddle, C. S., and McCarty, P. L., 1987, Transformations of halogenated aliphatic compounds: Environmental Science and Technology, v. 21, p. 722 - 735. 125

Vogel, T. M., and Mc Carty, P., L, 1987, Rate of Abiotic Formation of 1,1- Dichlorethylene from 1,1,1-Trichlorethane in Groundwater: Journal of Contaminant Hydrology, v. 1, p. 299-308. von Diiszeln, J., and Thiemann, W., 1985, Volatile chlorinated hydrocarbons in a coastal urban atmosphere: Sci. Total Environ., v. 41, p. 187 -194. W:ade, R. S., and Castro, C. E., 1973, Oxidation of Iron (II) Porphyrines by Alkyl Halides: Journal of the American Chemical Society, v. 95, p. 226 - 230. Wade, R. S., Havlin, j. R., and Castro, C. E., 1969, The oxidation of iron (Il) porphyrins by organic molecules.: Journal of the American Chemical Society, v. 91, p. 7530- 7534. Waisman, E., Worry, G., and Marcus, R. A., 1977, A Study of the Entropic and Electrolyte Effects in Electron Transfer Reactions: Journal of Electroanalytical Chemistry, v. 82, p. 9-28. Wakselman, C., and Tordeux, M., 1984, SRNl Substitutions of Halogenoperfluoroalkanes (CF3Br or CF2Cl2) under Pressure: J. Chem. Soc. Chem. Commun., p. 793-794. Walraevens, R., Trouillet, P., and Devos, A., 1974, Int. J. Chem. Kinetics, v. 7, p. 777- 786. Waxweiler, T., 1996, Reduktive Dechlorierung von Hexachlorethan in Eisen(hydr)oxidsuspensionen mit gelostem Fe2+: Unpub. Diploma work thesis, Universitat Bayreuth. Weber, E. J., and Wolfe, N. L., 1987, Environmental Toxicological Chemistry, v. 6, p. 911-919. Wentrup, C., 1979, Reaktive Zwischenstufen, Radikale, Carbene, Nitrene, gespannte Ringe. Wentworth, W. E., George, R., and Keith, H., 1969, Dissociative Thermal Electron Attachment to Some Aliphatic Chloro, Bromo, Iodo Compounds: The Journal of Chemical Physics, v. 51, p. 1791-1801. Wolf, C.R., 1977, Microsomes and Drug Oxidations, in Press, P., ed.: New York, p. 240- 246. Wood, J.M., Kennedy, F. S., and Wolfe, R. S., 1968, Biochemistry, v. 7, p. 1707-1713. Zefirov, N. S., and Makhon'kov, D. I., 1982, X-Philic Reactions: Chem. Rev., v. 82, p. 615-624. Zoro, J. A., Hunter, J.M., and Eglinton, G., 1974, Degradation of p, p'- DDT in Reducing Environments: Nature, v. 247, p. 235 237. A-1

Appendix A

A.1 Chemicals

The following chemicals were purchased from Fluka AG (Buchs, Switzerland) and used without further purification: Iron (II) chloride tetra- hydrate, sodium acetate, sodium perchlorate, acetylacetone (>99,5%), amber- lite IRA-401, ethanol (puriss), n-propanol, di-potassium hydrogen phosphate, tetrachloromethane (>98%, IR spectroscopy grade), tetrabromomethane (purum, > 98%), chloroform (IR spectroscopy grade; stabilized with 0.5% etha- nol), bromoform (>99%), dibromomethane (>99%), fluorotribromomethane (>99%), dibromodifluoromethane (>98%), fluorotrichloromethane (>99,5%), bromodichloromethane (>98%), dibromochloromethane (purum, > 97%), cys- tine, formic acid (>98%), zinc dust, ethyl acetate, 3-pentenoic acid, methanol, magnesium, magnesium sulfate, dietylether, diisopropylether, ~- methallyl chloride, tetrahydrofuran, benzyltriethylammoniumchloride, cysteine, ethyleneglycol-monoethylether, N-nitrosotoluol-4-sulfomethylamid, 2-mor- pholinoethanesulfonic acid monohydrate (MES), 4-(2-hydroxyethyl) piperazine-1-propanesulfonic acid (HEPES), 4-(2-hydroxyethyl)-piperazine-l- propanesulfonic acid (HEPPS), N-[tris(hydroxymethyl)methyl]-3-amino- propanesulfonic acid (TAPS), 2-(cydohexylamino)ethanesulfonic acid (CHES) Hexane and pentane were from Burdick & Jackson. Dibromodi- chloromethane (95%), bromotrichloromethane (99%) and 2-iodobenzotri- fluoride were from Aldrich Chemie, Switzerland. 13C-CC4 was purchased from Chemical Isotopes Laboratories CIL (>99%). 14C-CC4 was from NEN Products Du Pont, Boston, with a specific activity of 0.17 GBq/mmol which is 4.6 mCi/ mmol (radiochemical purity: 99%, determined by gas liquid chroma- tography: Porapak QS Durapak Porasil C column at 180°C). Insta-Gel, a uni- A·2 versal liquid scintillation cocktail for aqueous and nonaqueous samples, was purchased from Packard (reorder Nr. 6013009). d7·Isopropy1 alcohol ( 99,5%) was purchased from Armar AG (DOttingen, Switzerland). Deuterochloroform (CDCh) was from Dr. Glaser AG (Basel, Switzerland). The following chemi- cals were synthesized (see 4.2 Syntheses): meso-tetrakis(N-methyl-pyridyl) iron porphyin, N-formykysteine, 3-methyl-3-butenoic acid, diazomethane, 2- (2,2-dichloro-1-methykyclopropyl) ethanoic acid, 2-(2,2-dibromo-1-methyky- clopropyl) ethanoic acid, 2-(2-bromo-2-chloro-1-methylcyclopropyl) ethanoic acid.

A.2 Syntheses

The following abbreviations are used: m. p. = melting point, UV -VIS

Ultraviolet-visible spectrometry, £ extinction coefficient, IR = Infrared spectrometry, NMR= Nuclear magnetic resonance spectrometry, ppm = parts per million, TMS tetramethylsilane, splitting pattern: s = singlet, d doublet, t =triplet, q = quartet1 m =multiplet, J =coupling constant [Hz].

A.2.1 Meso-tetrakls(N-methyl-pyridyl)iron porphin

1431.9 mg (1.05 mmol) 5, 10, 15, 20- tetrakis (1-methyl-4-pyridyl)-21H- 23H- porphin tetra-p-tosylate salt and 417,6 mg (2.1 mmol) iron(II)chloride tetrahydrate were dissolved in 150 mL 0.3 M acetate buffer (see chapter 4.3 Stock solutions and buffers) and transferred into a Schlenk tube. The solution was evacuated and flushed with argon for several times. Then the solution was stored at 75°C for three days. After opening the flask, oxidation of iron (II) to iron (III) occurred. 6 mL of saturated sodium perchlorate was used to pre- cipitate the iron porphyrin as its perchlorate salt. The suspension was warmed, filtered and washed three times with lQ-3 M perchloric acid. A 0.1 M acetylacetone solution in acetate buffer was used to check the washing solu- tion on iron (III). Washing with lQ-3 M perchloric acid was continued until A-3 no iron (III) was observed (visible by a change to red color). The solid was suspended with 75 g amberlite IRA- 401, chloride form, in 300 mL of water. The suspension was passed through a column (60 cm * 4 cm i.d.) filled with amberlite IRA-401 and then completely concentrated. After redissolving the resulting brown solid in 6 mL 10-3 M hydrochloric acid/ 40% ethanol, it was shortly warmed up to 90°C. 15 mL of warm (90°C) n-propanol were added. The solution was cooled to room temperature and then stored at 4°C over night. After filtration and washing with n-propanol, the solid was dried at 70°C. 620 mg (0.7 mmol, 66%) meso-tetrakis(N-methyl-pyridyl) iron porphyin were obtained. lH-NMR (400 MHz, D20, 320K, in ppm downfield of TMS): 5.9 (bs, 8H, N-methyl-H), 10.9 (s, 8H, 3-pyridinium-H), 14.9 (bs, 8H, 2-pyridinium-

H), 67 (bs, 8H, ~-pyrrole-H). UV-VIS: E (598 nm, pH =7): 9600 M·lcm-1.

A.2.2 N·formylcysteine

3g (12.5 mmol) cystine in 15g (326 mmol) formic acid were boiled in a 100 mL flask equipped with a reflux condenser for four hours. The solution was cooled to 4°C. Recrystallization of the white solid was done from water. 1.2g (0.04 mmol, 32%) of N,N'- diformylcystine were obtained. 60 mg (0.2 mmol) N,N'- diformylcystine were suspended in 10 mL of 5% formic acid. 73 mg (1.12 mmol) zinc powder was added. The solution was stirred vigorously at room temperature for 24 hours. The excess zinc was fil- tered off and the solution extracted with ethyl acetate. The organic phase was concentrated and 41 mg (68 %) of N-formylcysteine as a white solid were ob- tained. lH-NMR (400 MHz, MeOD 298K, in ppm downfield of TMS): 2 (m, lH,

SH), 3.98 (m, 2H, CH2), 4.72 (t, JCH-CH2: 4.8 Hz, lH, CH), 8.08 (bs, lH, N-H), 8.14 (s, lH, formyl-H). 13C-NMR (400 MHz, MeOD 298K, in ppm downfield of TMS): 26.83 (CH2), 54.37 (CH), 164.52 (HCO), 172.56 (COOH). A-4

A.2.3 3-methyl-3-butenoic acid

A 1000 mL, two-necked, round-bottom flask was equipped with a re- flux condenser and a 100 mL dropping funnel. 18.26 g (0.75 mol) magnesium were put in the flask and 200 mL diethylether were added. 45.45 g (0.5 mol) ~­ methally l chloride were put into the dropping funnel and 10 mL were dropped into the flask. The start of the reaction was observed by an appearing cloudiness of the solution. 200 mL diethylether and 200 mL tetrahydrofuran were added in 100 mL portions and the rest of the ~- methallyl chloride was dropped to the reaction solution. The solution was boiled for 45 minutes. Af- ter cooling down to room temperature, the reaction mixture was carefully put in a 5 L beaker filled with 1 kg of powdered solid carbon dioxide. A white pre- cipitation occured, with thawing of the mixture it disappeared. After acidify- ing with sulphuric acid (10%), the two phases were separated. The water phase was twice extracted with diethylether. After evaporating the ether and destillation 23.25 g (46%) of 3-methyl-3-butenoic acid were obtained (p = 0.15 torr, b.p. 45°C). IR: 2980 (C-H), 1710 (C=O), 1650 (C=C). lH-NMR (400 MHz, CDCb,

298K, in ppm downfield of TMS): 1.84 (s, 3H, CH3), 3.08 (s, 2H, CH2)1 4.89 (lH, =CH), 4.96 (lH, =CH). 13C-NMR (400 MHz, CDCb, 298K, in ppm downfield of

TMS): 23.08 (CH3), 43.7 (CH2) 1 116.02 (~CH2), 138.6 (=CR2), 177.6 (COOH).

A.2.4 Diazomethane

For this synthesis a diazald-kit apparatus (Aldrich) was used. 36g (642 mmol) potassium hydroxide was dissolved in 60 mL of water (10.7 M), 90 m 1 ethyleneglycol-monoethylether were added. From a dropping funnel 23g (107 mmol) N-nitroso-toluol-4-sulfomethylamid dissolved in 100 mL diethylether was slowly added to the solution when the steam of the heated solution reached 30°C. The tube of the Aldrich apparatus was put in a beaker contain- A-5 ing sufficient diethylether to cover the tube's inlet. The beaker was cooled with ice. A yellow ethereal solution of -0.4 M diazomethane was obtained.

A.2.5 2·(2,2·dichloro·1·methyl cyclopropyl)·ethanoic acid

A three-necked 250 mL flask was equipped with a reflux condenser and a 50 mL dropping funnel. 1.02 g (10.2 mmol) 3-methyl-3-butenoic acid, 15g of a 30% sodium hydroxide solution in water and 0.06g (0.26 mmol) benzyl- triethylammoniumchloride were added. After heating the mixture to 60°C 6.02g (50.4 mmol) chloroform was added during 30 minutes. The reaction mixture was refluxed during 16 hours. The water phase was acidified with 10% sulphuric acid and extracted with diisopropylether. After evaporation the oily residue was bulb-to-bulb destilled. The product solidified and was recrys- tallized from hexane. Yield: 580 mg (31.1%) of 2-(2,2-dichloro-1-methyl cyclo- propyl)-ethanoic acid with an m.p. of 107°C. IR: 2980 (C-H), 1710 (C=O). lH-NMR (400 MHz, CDCls, 298K, in ppm downfield of TMS): 1.42 (q, 2H, CH2 (C-3'), 1.47 (s, 3H, CH3), 2.72 (q, 2H, CJ::h- COOH). 13C-NMR (400 MHz, CDCb, 298K, in ppm downfield of TMS): 21.48 (CH3), 28.27 (C, C-1'), 32.97 (CH2, C-3'), 41.39 (Q-I2-COOH), 66.92 (CC}i), 176.8 (COOH).

A.2.6 2·(2,2-dlbromo-1-methyl cyclopropyl)·ethanoic acid

By following the procedure of 2-(2,2-dichloro-1-methyl cycolpropyl) ethanoic adic and using 12.65 g (50 mmol) bromoform instead of chloroform 0.61g (22.5%) 2-(2,2-dibromo-1-methyl cyclopropyl)-ethanoic acid with an m. p. of91°C was obtained. lH-NMR (400 MHz, CDCl3, 298K, in ppm downfield of TMS): 1.51 (s, 3H, CH3), 1.61 (q, 2H, CH2 (C-3'), 2.78 (q, 2H, Cfu-COOH). 13C- NMR (400 MHz, CDCl3, 298K, in ppm downfield of TMS): 23.03 (CH3), 36.74(C, C-1'), 36.76 (CBr2), 34.19 (CH2, C-3'), 42.79 (C.H2-COOH), 176.3 (COOH). A-6

A.2.7 2-(2-bromo-2-chloro-1-methyl cyclopropyl)-ethanoic acid

By following the procedure of 2-(2,2-dichloro-1-methyl cycolpropyl) ethanoic adic and using 10.4 g (49.9 mmol) dibromochloromethane instead of chloroform O.Sg (21.8%) of 2-(2-bromo-2-chloro-1-methyl cyclopropyl)- ethanoic acid as colourless crystalls (m. p. 91°C) were obtained. Both di- astereomers were obtained. lH-NMR (400 MHz, CDC~, 298K, in ppm down- field of 'IMS): 1.5 (m, SH, CH3 and CH2 (C-3')), 2.75 (m, 2H, C!:U-COOH). 13(:. NMR (400 MHz, CDCh, 298K, in ppm downfield of 'IMS): 20.3/ 23.5 (CH3), 27.26 (C, C-1'), 52.55 (CBrCl), 33.2/ 33.3 (CH2, C-3'), 40.3/ 43.4 (5:82-COOH), 176.80, 176.89 (COOH).

A.3 Stock solutions and buffers

Stock solutions of iron porphyrin were prepared in 50 mM phosphate buffer or in 2 mM phosphate buffer, pH = 7, depending on the buffer concen- tration of the experiment. The concentration of the mediator was about 10 mM. The exact concentration was determined by UV-VIS at 598 nm (£ = 9600 M-1 cm-1). Stock solutions of halogenated methanes were prepared in oxygen- free methanol in the glove-box. Concentrations were 0.1 M. Further dilution for the preparation of spike solutions (1 mM) was done in the glove-box. Buffers were prepared by using HCl or NaOH to get the desired pH after dissolving the corresponding salt in nanopure water. Buffers as well as glassware were autoclaved at 121°C for 30 minutes. The solutions were purged with nitrogen gas (99.995%) for 1 hour/100 mL and immediately trans- ferred into the glove-box. A-7

A.4 Experimental procedures: Kinetic experiments

A.4.1 Standard procedure for kinetic experiments

The samples were prepared by autoclaving 58 ml serum flasks, buffers, stoppers and the glassware needed in preparation of the samples at 121°C for 30 minutes. Viton® stoppers (Maagtechnik, Diibendorf, Switzerland) were used to prevent loss from the serum flasks. Viton® stoppers have been shown to be very tight (Perlinger, 1994). Buffers were stirred and purged with argon gas while they cooled after autoclaving. Each 100 ml of buffer was purged for 1 hour. The buffer was transferred to a glovebox. The preparation of the samples was made in the glovebox: First, 50 ml of buffer was added to the serum flask, then, aliquots of 0.1 M cysteine solution in buffer and an iron porphyrin stock solution (-0.01 M) were added. The rest of volume up to exactly 57 ml was filled with buffer, so 1 ml of headspace was in each sample. The flasks were closed with a Viton® stopper and crimped with a crimp cap. Only then, the samples were taken out of the glovebox and immersed into a water bath with temperature control. The water bath was covered in order to prevent any photo reactions. All samples were equilibrated at least 24 hours. To start the kinetic experiment, 57 µl of a methanolic solution con- taining the substrate and the internal standard were added to the reaction medium. Samples were taken at appropriate time intervals by withdrawing lOOµl of reaction solution into a syringe after injection of an equal volume of nitrogen gas. The sample was placed in a 3 ml glass vial with a plastic screw- top containing 1 ml of hexane or pentane. Then, the solution was shaken for 30 seconds on a vortex mixer in order to extract the sample. The efficiency of extraction is about 80 to 100% depending on the substrate (Perlinger, 1994). The organic phase was transferred to an autosampler vial containing sodium sulfate in order to dry the organic solvent from traces of water. Eight to twelve samples were taken over the course of three half-lives. As the volume A-8 of the headspace increased with every sample taken, the aqueous concentra- tions were corrected for the amount of solute that volatilzed. The corrected aqueous concentration was:

[S] = [S] + [Slmeasured ·KH ·Vo (gas) A-1 corr. measut'OO Vn(liquid)

where = corrected aqueous concentration of the substrate S measured aqueous concentration of S =Henry's law constant = mL + n · 0.1 mL = 57 mL- n · 0.1 mL

KH-values for polyhalogenated alkanes and alkenes used in this thesis are given in Table A.l. A-9

Table A.1: Henry's law constants of polyhalogenated alkanes at 25"C

3 1 compound KH (Pa m moI- ) reference

CC14 2989 (experimental) (Mackay et al., 1993b)

CHC13 427 (experimental) (Mackay et al., 1993b)

CHBr3 46.61 (experimental) (Mackay et al., 1993b)

CH2Br2 86.13 (experimental) (Mackay et al., 1993b)

CBrCl3 42 (estimated) (Hine and Mookerjee, 1975)

CBr2Cl2 11.3 (estimated) (Hine and Mookerjee, 1975)

CHBrC12 162 (experimental) (Mackay et al., 1993b)

CHBr2Cl 86.13 (experimental) (Mackay et al., 1993b)

CFBr3 19.2 (estimated) (Hine and Mookerjee, 1975)

CF2Br2 451 (estimated) (Hine and Mookerjee, 1975)

CFC13 10243 (experimental) (Mackay et al., 1993b)

C2Cl6 846 (experimental) (Mackay et al., 1993b)

C2HC15 252.9 (calculated) (Mackay et al., 1993b)

CC13-CH2Cl 241 (calculated) (Mackay et al., 1993b)

CHC12-CHC12 25.7 (experimental) (Mackay et al., 1993b) CC13-C~ 1763 (experimental) (Mackay et al., 1993b)

CF2Cl-CCl3 1426 (estimated) (Hine and Mookerjee, 1975)

CF3-CC13 8995 (estimated) (Hine and Mookerjee, 1975)

CF2Cl-CFCl2 32323 (experimental) (Mackay et al., 1993b)

Cl2C=CC12 1794 (experimental) (Gossett, 1987) Cl.zC=CHCl 971 (experimental) (Gossett, 1987)

Cl2C=CH2 2646 (experimental) (Gossett, 1987) A-10

A.4.2 Determination of rate constants

All kinetic experiments, if not stated otherwise, were done in tripli-

cates. Plots of ln[S]/[S]0 versus time gave the observed rate constant, kobs' as the reactions were first order in substrate concentration and iron porphyrin con- centration. The data of all three experiments were combined in one ln[S]/[Slo versus time plot. From linear regression analysis kobs (negative slope) and the standard deviation was obtained. The same was done for the blank reactions. For reactions that were fast1 the concentration at time zero was determined by plotting ln[S] versus time: The intercept of the linear regression line is equal to ln[S](t=O). This concentration helps to check the actual initial concentration of the substrate, as often the first sample taken in fast (see footnote 1) reac- tions does not represent the initial concentration because the substrate has al- ready partly degraded. Such an additional value was taken for the ln[SJ/[SJ0 versus time plot. The exact verifying of the initial concentration is important, especially in reactions where rates as a function of the initial concentrations were measured (see for example chapter 2.7). Second-order rate constants were calculated as the difference between kobs and kblank (first order reaction rate constant of blank reaction) devided by the concentration of the mediator (equation A-2).

A-2

A.4.3 pH variation

The conditions in the experiments where the effect of pH on the deg- radation rate of CC14 and C2Cl6 was investigated are given in Table A.2.

1 "Fast" meaning that more than 5% degradation of the substrate occured during the first minute after starting the reaction. A-11

Table A.2: Conditions in pH experiment function compound value mediator M iron porphyrin [M] =30µM, [M] = 0 µM (for blank) bulk reductant R cysteine [R] =5mM substrate S CC14 or C2Cl6 [S] 0 =1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B MES (pH 6), HEPES (pH 7), [BJ =50mM HEPPS (pH 8), TAPS (pH 8.7), CHES (pH 9.3)° additionals temperature T T = 25°C

'Full names are given in chapter A. l.

A.4.4 Variation of cysteine concentration

The conditions for the experiments where the rate was measured as a function of the cysteine concentration are listed in Table A.3.

Table A.3: Conditions in the experiments in which cysteine concentration was varied function compound value mediator M iron porphyrin [M] =30µM, [M] = 0 µM (for blank) bulk reductant R cysteine [R] = 0.03, 0.1, 0.3, 0.6, 1.0, 2.0, 5.0, 10.0, 20.0 mM substrate S CC14 or C2Cl6 [S] 0 =1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ = 50 mM, pH = 7 additionals temperature T T= 25°C A-12

A.4.5 Variation of phosphate concentration

The conditions for the experiments where the rate was measured as a function of the phosphate concentration are listed in Table A.4.

Table A.4: Conditions in the experiments in which phosphate concentration was varied function compound value mediator M iron porphyrin [M] 30 µM, [M] = 0 µM (for blank) bulk reductant R cysteine [R] = 5mM substrate S CC14 or C2Cl6 [S] 0 1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ = 0.5, 2.0, 5.0, 20.0, 50mM,pH=7 additionals temperature T T= 25°C

A.4.6 Variation of ionic strength

The conditions for the experiments where the rate was measured as a function of the ionic strength are listed in Table A.5.

Table A.5: Conditions in the experiments in which the ionic strength was varied function compound value mediator M iron porphyrin [M] = lOµM, [M] 0 µM (for blank) bulk reductant R cysteine [R] =SmM substrate S C2Cl6 [S] 0 =1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ 10 mM, pH = 7 additionals A sodium sulfate Na2S04 [A] = 0, 10, 20, 30, 40 mM temperature T T = 25°C A-13

A.4.7 Variation of ligand concentration

The experimental conditions for the determination of the mediator efficiency of the porphyrin ring are listed in Table A.6.

Table A.6: Conditions in experiments in which ligand concentration was varied function compound value mediator M porphyrin ring" [M] 30, 60, 100, 200 µM, [M] = 0 µM (for blank) bulk reductant R cysteine [R] =5 mM substrate S C2Cl6 [S] 0 =1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ = 0.5, 2.0, 5.0, 20.0, 50 mM,pH=7 additionals temperature T T= 25°C

"5, 10, 15, 20-Tetrakis (1-methyl-4-pyridyl)-21H, 23H-porphine tetra-p- tosylate salt A-14

A.4.8 Reactions without porphyrin

The conditions for the blank reactions where cysteine concentration and pH were varied are given in Table A.7.

Table A.7: Conditions for experiments without iron porphyrin

function compound value

mediator M bulk reductant R cysteine [R] 1, 2, 5, 10 mM [R] = 0 mM (for blank)

substrate S CCl, or C2Cl6 [S] 0 1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate, pH 7 [BJ 50mM,pH=7 TAPS,pH9 [B] = 20 mM, pH 9 additionals A temperature T T 25°C A-15

A.4.9 Variation of initial concentration of CBr2Cl2

The experimental conditions for the evaluation of the regeneration rate of the iron porphyrin by cysteine during the degradation process are given in Table A.8. The initial concentrations of CBr2Cl2 were varied over a range of 1 to 56 µM.

Table A.8: Conditions in the experiments in which the regeneration rate of iron porphyrin was evaluated function compound value mediator M iron porphyrin [M]=3µM [M] = 0 µM (for blanks) bulk reductant R cysteine [R]=5mM substrate S CBr2Cl2 [SL= 1-56µM (15 experiments with 15 different concentrations) internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ = 50 mM, pH = 7 additionals A temperature T T= 25°C

The data of the 15 experiments were fitted with AQUASIM (Reichert, 1994). The variables,processes and compartments that were used in the fitting process are listed in Table A.9. A-16

Table A.9: Definition of variables, processes and compartments for the fitting of the degradation rates as a function of the initial concentration of CBr2C1 2

Variables name of variable function type calcnum number of experiment program variable CBr2Cl2 concentration of substrate dyn. volume variable CBr2Cl2_1 - list of measured concentra- real list variable CBr2Cl2_15 tions versus time CBr2Cl2_exp coordination of calcnum variable list variable with CBr2Cl2_1 - _15 CBr2Cl2_ini gives initial concentrations real list variable of CBr2Cl2 for each experiment FeP(II) concentration of FeP(II) dyn. volume variable FeP(III) concentration of FeP(III) dyn. volume variable FeP_tot total concentration of FeP constant varible kl first-order rate constant for constant variable reduction of FeP k2 second-order rate constant constant variable for ET to CBr2Cl2 time time program variable

Processes kl FeP(III)--> FeP(II) dynamic process k2 FeP(II)+CBr2Cl2--> dynamic process FeP(III)+ ...

Compartments batch active variables: mixed reactor CBr2Cl2, FeP(II), FeP(III) active processes: kl, k2 initial concentrations: CBr2Cl2 (bulk vol.): CBr2Cl2_ini FeP(II) (bulk vol.): FeP _tot x A-17

A.4.10 Competition experiments: CBr2c12 versus CBrCl3

The competition experiments between CBr2Cl2 and CBrCI:, were per- formed under the conditions given in Table A.10.

Table A.10: Conditions in competition experiments between CBr2Cl 2 and

CBrCl3 function compound value mediator M iron porphyrin [M)=3µM [M] =0 µM (for blanks) bulk reductant R cysteine [R) =5mM

substrate S CBr2Cl2 or CBrC13 [S] 0 = 10 µM in single sub- or both together strate experiments and for each compound in mixtures internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ 50 mM, pH 7 additionals A temperature T T 25°C A-18

A.4.11 Competition experiments: CCl4, CHBrCl2, CHBr,CI

The apparent second-order rate constants were measured for CC14,

CHBrC12 and CHBr2Cl in binary and ternary mixtures and also for each single compound under the conditions given in Table A.11.

Table A.11: Conditions in competition experiments between CC!4, CHBrCl2 and CHBr2Cl function compound value mediator M iron porphyrin [M] 3µM [M] = 0 µM (for blanks) bulk reductant R cysteine [RJ =5mM substrate S CC14, CHBr2Cl and [S] 0 = 10 µM in single sub- CHBrCl2, and binary strate experiments and and ternary mixtures for each compound in mixtures internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ 50 mM, pH 7 additionals A temperature T T =55°C A-19

A.4.12 Competition experiments: CCl4 versus 4-chloro-nitrobenzene

The conditions for the competition experiments between carbon tetra- chloride and 4-chloro-nitrobenzene are listed in Table A.12.

Table A.12: Conditions in competition experiments between carbon tetrachloride and 4- chloro-nitrobenzene function compound value mediator M iron porphyrin [M]=3µM [M] = 0 µM (for blanks) bulk reductant R cysteine [R] =5mM substrate S cc1. or 4-chloro-nitro- [S] 0 = 10 µM in single sub- benzene strate experiments and or both together for each compound in mixtures internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ = 50 mM, pH = 7 additionals A temperature T T = 25°C

For the analysis of 4-chloro-nitrobenzene HPLC was used (see Analytical Procedures). A-20

A.4.13 Determination of activation parameters

The determination of the activation parameters, AH' and AS', was done by measuring degradation rates at 25°C, 35°C, 45°C, 55°C and 65°C (with and without iron porphyrin, each temperature in three replicates). In the case of CC14 30 µM of iron porphyrin was used in the catalyzed reactions, in the case of C2Cl6 only 10 µM. The conditions are presented in Table A.13.

Table A.13: Conditions in the experiments where activation parameters were determined function compound value mediator M iron porphyrin [MJ = 30 µM (for cq) ) [M] = 10 µM (for C2Cl6 [M] 0 µM (for blanks) bulk reductant R cysteine [R] =5mM substrate S CC14 or C2Cl6 [S] 0 =1 µM internal standard IS 2-iodobenzotrifluoride [IS]= 1 µM buffer B phosphate [BJ 50 mM, pH = 7 additionals A temperature T T = 25, 35, 45, 55, 65°C

The activation parameters, AHfi and AS•, were determined by plotting

1 1 ln(k/T), where k is equal to k0 b,-kblank (s- ) and T is temperature, versus T- • Equation A-3 gives the connection between activation parameters and degra- dation rate (Hoffmann, 1976).

AH' A-3 R·T where k = kobs-kblank (s-1) T =temperature (K) k Boltzmann constant, 1.381 10-23 JK1 h = Planck constant, 6.626 10·34 Js (Laidler, 1987) R gas constant, 8.314 Jmot'K1 A-21

A.5 Experimental procedures: Analysis of products and reaction intermediates

14 A.5.1 Distribution experiments with C-CCl4

The 14C-CC4 experiments were conducted in a 58-mL serum flask sealed with a Viton® stopper (Maagtechnik, Dubendorf, Switzerland) and an aluminum crimp cap. The reaction mixture was prepared in an anaerobic glove-box: 50 mM phosphate buffer, pH = 7 (see Stock solutions and buffers, chapter 4.3), 5 mM cysteine and 30 µM iron porphyrin were used. 57 µL of an anaerobic methanolic 1 mM 14C-CC4 solution containing 1 mM 2-iodo-ben- zotrifluoride as internal standard was introduced by a gas-tight Hamilton sy- ringe through the Viton stopper. The flask was incubated at 25°C in the dark for 48 hours. After injection of an equal volume of nitrogen gas (purity 99.995%) 100 µL of the solution were withdrawn, extracted with 1 mL of pen- tane and analysed with GC-ECD. Several procedures were used to analyse ra- dioactivity-distribution: 1) 1-mL samples were extracted with either lml of pentane, hexane or ethyl acetate and 500µL of the organic phase were given to 4.5 mL of a univer- sal scintillation cocktail for aqueous and non-aqueous samples. 2) 500µL of the reaction mixture plus 150 µL of a 1 M sodium hydrox- ide solution were given directly to 4.5 mL of the scintillation cocktail and the activity was measured with a scintillation counter (see Analytical Procedures). This was experiment 1 in chapter 3.2.1. Blank samples (just water spiked with the methanolic CC14 solution) served as reference in the activity measure- ments. 3) 500 µL samples were either acidified with 150 µL of lM sulphuric acid (experiment 2 in chapter 3.2.1) or made alkaline with 150 µL of 1 M so- dium hydroxide (experiment 3 in chapter 3.2.1) and flushed with nitrogen gas A-22 for one hour. Activity measurements were performed by mixing the solution with 4.5 mL of the scintillation cocktail. 4) 100-µL samples were injected in an ion chromatographic system (see Analytical prodedures, chapter 4.8). Fractions of each peak in the ionic chromatogram were collected and mixed with 4.5 mL of the scintillation cock- tail in order to localize the product.

A.5.2 "C-CCl4 experiments

The same conditions as in the 14(:-CC4 experiment were used (see chapter A.5.1). 57µL of a 1 mM anaerobic methanolic 13C-CC4 solution con- taining lmM 2-iodobenzotrifluoride were added to start the reaction. 100-µL samples were analysed with an ion chromatographic system. 40 fractions of the peak showing activity were collected in order to accumulate the product. The united samples were lyophylized by freezing them in liquid nitrogen and evaporating the water under high vacuum (O.l torr). Then the sample was redissolved in D20. 13C- and 1H-NMR-spectra were taken (see Analytical Pro- cedures).

A.5.3 a· abstraction experiments

D· abstraction experiments were conducted in 12-mL serum flasks sealed with Viton® stopper and aluminum crimp caps. The reaction mixture contained 2 mM phosphate buffer, pH = 7, (see Stock solutions and buffers, chapter 4.3), 5 mM cysteine and 30 µM iron porpyhrin. Blank samples were prepared without iron porphyrin. Additionally, either 1% (V /V), 0.5% or 0.1 % of d1-isopropyl alcohol were added to the samples. The experiments were started by adding 60 µL of an anaerobic methanolic solution containing 1 mM CX.Y("-•l (X, Y =Cl, Br) and 0.1 mM 2-iodobenzotrifluoride. The concen- tration was 5 µM for CX.Y<"-•l (X, Y = Cl, Br) and 0.5 µM for the internal stan- dard. The conditions are summarized in Table A.14. A-23

Table A.14: Conditions in the D abstraction experiments function compound value mediator M iron porphyrin [M] =30µM, [M] = 0 µM (for blank) bulk reductant R cysteine [R] =5mM , , substrate S CC14 CBr41 CBr2Cl2 [S] 0 =5 µM CBrCl3 internal standard IS 2-iodobenzotrifluoride [IS]= 0.5 µM buffer B phosphate [BJ =2mM,pH= 7 additionals A d1-isopropyl alcohol [A]= 0.1, 0.5, 1% temperature T T = 25°C

The flasks were incubated at 25°C in the dark, exept for the blank sys- tem of CCl4 which was incubated at 60°C in order to accelerate the degradation reaction. 100-µL samples were extracted with 1 mL of pentane and analysed with GC-MS in the SIR-modus. The experiments were done in duplicates and each was measured three times.

A.5.4 Carbene trap experiments

In the iron porphyrin system carbene trap experiments were per- formed in two series with two different carbene traps at slightly different con- ditions: Series 1 was conducted in 58-mL serum flasks under the following conditions: 2 mM phosphate buffer, pH = 7 (see Stock Solutions and Buffers), 2.5 mM cysteine, 250 µM iron porphyrin (in the blank system no iron por- phyrin), 200 µM of CX.Y<4-al (X, Y = F, Cl, Br) and CHX.Y

Table A.15: Conditions for series 1 of carbene trap experiments function compound value mediator M iron porphyrin [M] = 250µM, [M] =0 µM (for blank) bulk reductant R cysteine [R] =2.5mM substrate S CC14, CBr4, CBr2Cl2, [S] 0 = 200 µM CBrC13, CHBr3, CHBr2Cl, CHBrC12, CFBr3, CF2Br2, CFC13 internal standard IS 2-iodobenzotrifluoride [IS] 10 µM buffer B phosphate [BJ = 2 mM, pH = 7 additionals A 3-pentenoic acid [A] =4mM temperature T T= 25°C A-25

Table A.16: Conditions for series 2 of the carbene trap experiments function compound value mediator M iron porphyrin [M] 500 µM, [M] = 0 µM (for blank) bulk reductant R cysteine [R] =5mM , substrate S CC141 CBr41 CBr2Cl2 [S] 0 =100 µM CBrC13 , CHBr:v CHBr2Cl, CHBrC121 CFBr:v CF2Br2, CFC13 internal standard IS 2-iodobenzotrifluoride [IS]= 10 µM buffer B phosphate [B] = 2 mM, pH = 7 additionals A 3-methyl-3-butenoic [A] =4mM acid temperature T T =40°C

In the juglone system the carbene trap experiments were done under the conditions listed in Table A.17.

Table A.17: Conditions of the carbene trap experiments in the juglone system function compound value mediator M juglone [M] = 100 µM, [M] = 0 µM (for blank) bulk reductant R Na2S [R] =5mM , , substrate S CCl4 CBr4 CBr2Cl21 [S] 0 =100 µM CBrCl3 internal standard IS 2-iodobenzotrifluoride [IS] =10 µM buffer B HEP ES [BJ = 50 mM, pH = 8 additionals A 3-methyl-3-butenoic [A] =4mM acid temperature T T = 25°C (exept for CC14: T = 40°C) A·26

A.5.5 Quantification of products: Methods of analyses

Product studies were done with series 2 of the carbene trap experi- ments in the iron porphyrin case and with the carbene trap experiments in the juglone case. Halogenated compounds were identified and quantified by taking 100- µL samples and extracting them by 1, 10 and 20 mL of pentane depending on the concentrations of the products and the linear range of the response factor of each compound analysed. Analysis was done with GC-ECD. Halogenide ions, formiate and N-formylcysteine were analysed by withdrawing 0.8-mL samples, filling them into autosampler vials and injecting 10 µL into an ion chromatographic system.

A.5.6 Product studies of polyhalogenated ethanes

The conditions for the products studies of the four polyhalogenated ethanes CF2Cl-CCly CF3-CC1,, CF2Cl-CFCl2 and CH3-CC1, are listed in Table A.18. Analysis of halogenated and ionic products was done as described in chapter A.5.5.

Table A.18: Conditions in the experiments with polyhalogenated ethanes function compound value mediator M iron porphyrin [MJ =250µM, [MJ = 0 µM (for blank) bulk reductant R cysteine [R] =SmM , substrate S CF2Cl-CCl3 CF3-CC1,, [SJ 0 = 100 µM CF2Cl-CFCl2 and CI-4- CCl, internal standard IS 2-iodobenzotrifluoride [ISJ =10 µM buffer B phosphate [BJ= 2 mM, pH= 7 additionals A temperature T T = 25°C (exept for CF2Cl- CFCl2 and CH3-CC13 : T = 6QOC) A-27

A.6 Analytical prodedures

A.6.1 GC-ECD

The pentane extracts were analysed by GC-ECD on a Carlo Erba HRGC 5160, equipped with an autosampler AS 200 and a Cryo 520 C02 cooling sys- tem. l.5µL were injected. The injection was splitless. A 20m DB-624 column (i.d. 320µm, 0.25µm film thickness) was used. The injector and detector tem- peratures were 200 and 300°C respectively. Hydrogen gas at a flow of 2 mL/min. and a mixture of argon (90%) and methane (10%) were used as car- rier and makeup gas respectively. The detector was an electron capture detec- tor (Carlo Erba ECD 400 with a Ni-63 source). T-programme: 40°C (4 min.); S°C/ min. to 130°C (0 min.); 35°C/ min. to 200 °C (5 min.).

A.6.2 GC-MS

The pentane extracts of the D· abstraction experiments were analysed by GC-MS on a Fisons Instruments GC S165 equipped with a Fisons Instru- ments MD SOO mass spectrometer. 2 µL were injected using an on-column in-

jection. The column was a 30m FS 6S DB-624 (i. d. 250 µm, 2 µm film thick- ness). Single ion recording modus (SIR-modus) was used for quantification, i.e. the ratio of m/z llS, 120, 122 to 119, 121, 123 for CHCIJ to CDCl3 and also the fragment masses S3, S5, S7 to S4, S6, SS (CHClz to CDCl2) in their character- istic isotopic distribution. Helium at a flow of 2 mL/min. was used as carrier gas. T-programme: 40°C (4 min.); S°C/ min. to 130°C (0 min.); 35°C/ min. to 200 °C (5 min.). The methylesters of the dihalocyclopropanes were analysed by GC-MS on a Fisons Instruments GC Sl65 equipped with a Fisons Instruments MD SOO mass spectrometer. 1 µL was injected on column. A glass capillary column (25 A-28 m * 250 µm i.d., coated with PS-089, 0.23 µm) was used. Helium at a flow of 4.4 mL/min. was used as carrier gas. Masses from m/z 30 to 300 were analysed (Full Scan Modus). The solvent delay was 3 min. and the ionisation method was EI+. T-programme: 35°C (2 min.); 10°C/min. to 200°C (2 min.).

A.6.3 IC

Ionic species were analysed on a DX-300 chromatographic system (Dionex) equipped with a standard AGP pump, a PED in the conductivity de- tection mode and a AI-450 software 3.31. The anion separator column (200 * 4 mm) IonPac AS11 and its corresponding guard column (50 • 4 mm) AG11 were used. The eluent (1 mL/min.) conductivity was chemically suppressed by a self-regenerating suppressor (ASRS, current setting 3) that uses the col- umn effluent to supply the proton generated by water electrolysis. An anion trap column (ATC-1, 9 * 24 mm) was installed in front of the injection valve to minimize interferences from anionic impurities in the eluent during gra- dient elution. The gradient programme was: 0-1.2 min. 5 mM NaOH, 1.2-9 min. linear increase to 25 mM NaOH, 9-9.5 min. back to 5 mM NaOH, 9.5-13.8 min. 5 mM NaOH. Samples were injected directly after withdrawing them from the serum flasks.

A.6.4 HPLC

The aqueous samples were analysed on an HPLC without further treatment. The eluent was a mixture of methanol and water in the ratio 3:2. The method was isocratic, the flow 1 mL/min (pump: Jasco PU-980), the col- umn an RP-8. The detection was done by a UV-VIS detector Gasco UV-870) at A= 250 run and the peaks were integrated with the program Chrom-Card®. A·29

A .6.5 Scintillation counter

Radioactivity measurements were carried out with a liquid scintilla- tion counter of Kontron. Samples were measured in plastic vials (20-ACB from Chromacol LTD with plastic screwtops).

A.6.6 NMR

lH- and 13C-NMR spectra were recorded on an ASX-400 NMR spec- trometer from Bruker at 400 MHz. B-1

Appendix B

B.1 UV-VIS experiments

B .1.1 Cysteine-llgated iron porphyrin at different pH

In order to evaluate the pH-dependence of the UV-VIS spectra of the cysteine-ligated iron porphyrin, the following procedure was used: At all pH- values the buffer had a concentration of 10 mM. At pH 6 MES, at pH 7 HEPES, at pH 8 HEPPS, at pH 8.7 TAPS and at pH 9.3 CHES was used (for full names see chapter 4.1). To 2 mL of the buffer 12.7 µL of a 9.4 10"' M iron porphyrin stock solution were added (concentration: 6 µM). 1 mL was filled in a quartz glass cuvette (Hellma® Prazisions-Kiivetten, Kontron, Nr. 117. 104-QS) which was closed with a silicone septum. The cuvette was purged with nitrogen through the septum for five minutes. Then, 20 µL of fresh, anaerobic 0.1 M cysteine solution (prepared in the glove box) was injected through the septum to the anaerobic iron (III) porphyrin solution (resulting concentration of cysteine: 2 mM). Then, a UV-VIS spectrum was taken in the range of 250 to 650 nm. At all pH-values a shift of the Soret-band from 420 to 443 nm was found. At pH 8, 8.7 and 9.3, however, the shift was accompanied with a dear broadening of the band. Additional experiments2 with UV-VIS spectrophotometry were done in order to get information about the behavior of the iron porphyrin in the presence of various solution constituents. The experimental conditions, the results and discussion are given in the following chapters:

' They are not mentioned in chapter 2. B-2

B.1.2 Effect of oxygen: Formation of superoxide radical

When opening the septum for a short moment after a standard ex- periment as described above (B.1.1) in order to study the effect of oxygen on cysteine-ligated iron (II) porphyrin, a decrease in the intensity of the Soret band was observed. The decrease of the absorption went on until the band had completely vanished and the solution lost its brown-reddish color.

Possible explanations for these findings are: Either oxygen was reduced to 0 2 , the superoxide radical anion, and this removed the iron from the center - or o; destroyed the porphyrin ring. Oxygen itself cannot be the cause of the iron porphyrin destruction as iron (IIT) porphyrin is stable under aerobic conditions. Only the combination of cysteine, oxygen and iron porphyrin lead to the apparent destruction of the latter. To check whether superoxide was responsible for the disappearance of

the absorption, experiments with K02 were performed: To an anaerobic 6 µM iron porphyrin solution with 2 mM cysteine two drops of an anaerobic satu-

rated K02-solution in acetonitrile were added through the septum. (K02 is not well soluble in acetonitrile; a saturated solution was made by adding as much

K02 as needed for precipitation of the solid. This solution was prepared in the glove-box.) The same effect was found as for the experiment with oxygen and cysteine: The Soret-band slowly decreased and finally disappeared. This effect can be explained as follows: Cysteine reduces oxygen to the superoxide radical anion. Actually this process was observed when preparing a fresh aqueous cysteine solution: After several hours, cysteine is oxidized to cystine under aerobic conditions whereas it is stable under anaerobic condi-

tions. Thus, 0 2•· or a follow-up product of 0 2" in reaction with water (for ex- 3 ample H 20 2 or hydroxy radical) was probably responsible for the destruction of the iron porphyrin.

3 See for example the iron-promoted Haber - Weiss reaction: o,·. +Fe" ---> Fe2+ + 0 2 2 o,-· + 2 H+ --->H20 2 + 0 2 Fe'• + H,o, ---> Fe>+ + OH" + OR Further superoxide radical anion induced reactions are found in (Aust and White, 1985). 8-3

B.1.3 Azide and cyanide as ligands

An aqueous anaerobic solution of 50 µM NaN3 was added to an an- aerobic iron (Ill) porphyrin solution (pH 7). A shift of the Soret band from 420 to 430 nm occurred. When adding oxygen to the solution, the absorption maximum shifted back to 420 nm. Apparently, N 3- was not able to reduce 0 2 to the superoxide radical anion because no disappearance of the Soret band was observed. Addition of KCN to an anaerobic iron porphyrin solution (pH 9.3) gave a clear shift from 420 to 470 nm in two steps: The first shift was from 420 to 430 nm. Then, the maximum at 430 nm decreased and a new one increased at 470 nm. After addition of oxygen the spectrum remained the same. Neither disappearance of the Soret band nor shift backwards was found. Apparently,

CN- is a very good ligand and cannot be replaced by oxygen (like N 3-) nor by superoxide if it produces superoxide by reduction of 0 2 at all. To test whether the cyanide-ligated iron porphyrin was stable against 0 2•· the following experiment was made: To an anaerobic iron porphyrin solution, pH 9.3, cysteine was added which resulted in a shift in the absorption band from 420 to 443 nm. Afterwards KCN was added, and a further shift resulted to 470 nm. Then, the cuvette was opened in order to check the effect of oxygen and superoxide radical anion respectively: The spectrum remained unchanged. Apparently, the 0 2- radical anion, which was produced under these conditions as cysteine was present, cannot attack the iron center because cyanide is too strong a ligand to be replaced by 0 2-. The condition for a "successful" destruction of the complex seems to be that the superoxide radical anion or one of its follow-up products reaches the iron center. A test was made to know whether free Fe3\q was in the solution after the disappearance of the Soret-band in the corresponding experiments. Acety- 3 lacetone is known to form strong complexes with Fe • which are deep red (Smith and Martell, 1989). Actually, no change in color was observed after B-4 adding acetylacetone to the cuvette. Thus, iron was not removed from the porphyrin ring or the concentrations were too low to see any effect.

B.1.4 Effect of cysteine concentration

The influence of cysteine concentration on the UV-VIS spectrum of iron porphyrin was also checked (see chapter 2.5.1). For this test the conditions were chosen as follows: 6 µM of iron porphyrin in each case, 10 mM phosphate-buffer at pH 7 and 10 mM CHES- buffer at pH 9.3. Cysteine concentration was varied: 6, 20, 60, 120, 200, 400, 1000, 2000 and 4000 µM. 6 and 20 µM were achieved by adding 6 and 20 µL of 0.001 M cysteine solution to 1 mL of the iron porphyrin solution, 60, 120, 200 and 400 µM by adding 6, 12, 20 and 40 µL of a O.Dl M cysteine solution and finally 1000, 2000 and 4000 µM by adding 10, 20 and 40 µL of 0.1 M cysteine solution. Thus, the volume of the solution in the cuvette was increased only slightly (at maximum 4%) which could be ignored in the interpretation of the UV-VIS spectra. The results of this series are: At all cysteine concentrations at pH 7 the shift of the Sort band from 420 to 443 nm was observed. Furthermore, at all cysteine concentrations at pH 9.3 the same shift was found, but the Soret-band had broadened as shown in Figure 2.8. No step-by-step shift of one and two cys- teine ligands were found like in the case of cyanide.

B.1.5 Effect of substrates

Whether precursor complex formation was part of the degradation mechanism of polyhalogenated methanes in the iron porphyrin system studied was checked by taking UV-VIS spectra of cysteine-ligated iron

porphyrin solutions with and without polyhalogenated Cc and C2-com- pounds. To an anaerobic solution of 50 mM phosphate or HEPES buffer, pH 7 and either 5 µMiron porphyrin (to check the absorption at 420 nm) or 50 µM iron porphyrin (to check the absorption at 598 nm) and 2 mM of cysteine 10 µL of an 0.001 M anaerobic methanolic solution (concentration: 10 µM) of ei- B-5 ther hexachloroethane, 1,1,2,2-tetrachloroethane, 1, 1,2-trichlorotrifluoro- ethane, fluorotribromomethane, fluorotrichloromethane or difluorodibro- momethane were added. In all experiments, no shift of the Soret-band was observed. Polyhalogenated alkanes are rather bad ligands as they are quite lipophilic and cannot remove cysteine as axial ligands. Even in solvents like acetonitrile, N,N'-dimethylformamide, ethanol or methanol, where iron por- phyrin was dissolved and purged with nitrogen, no shift of the Soret band was observed when adding 50 or 500 µM of hexachloroethane respectively. The effect of methanol on the UV-VIS spectrum of cysteine-ligated iron porphyrin was also investigated, as all spike solutions (of the polyhalogenated alkanes) were methanolic. Spectra were taken with 5 µM iron porphyrin, 5 mM cysteine, where methanol concentration was varied from 0.025 to 0.050 M in 0.005 M steps. In this range no change in the spectrum was observed.

B.1.6 UV-VIS spectrum of porphyrln ligand

The UV-VIS spectrum of the porphyrin ligand alone (without iron) in a concentration of 7 µM showed a strong absorption at 420 nm and two weak absorption bands at 520 and 580 nm. The extinction coefficients were not determined. After purging the cuvette with nitrogen, cysteine was added and the UV-VIS spectrum showed no change in the absorption maxima. Also after opening the cuvette in order to add oxygen no change was found. Thus, the combination of cysteine, oxygen and porphyrin ligand did not cause a disappearance of the strong absorption at 420 nm like in the case when iron was in the middle of the center. Apparently, whatever species was responsible for the disappearance of the Soret band, it only attacks the porphyrin when iron is in the center of the porphyrin. B-6

B.2 Polarographic measurements

The experimental conditions used in the polarographic measure- ments are given in Table B.1.

Table B.1: Conditions in the polarographic measurements

Working electrode Hanging mercury drop electrode Reference electrode Ag/ AgCl (220 m V versus SHE, stan- dard hydrogen electrode)

Counter electrode Platinum wire Mode Difference polarography 2mV s·1 -dU (U.1s t h e potenha . 1, t 1s . time . ) dt

Purging 5 minutes with N 2 before each meas- urement Drop size 2

First, known redox systems were evaluated by measuring the half- wave potentials. For example, the current of a cystine solution was measured in the potential range from 0.7 V to -0.9 V. A clear current-wave was found at -610 mV. This is in good agreement with the literature value, -600 m V4 (Clark, 1960). On the other hand the oxidation process was checked by taking po- larograms of cysteine in the range from -0.9 V to 0.7 V. At 170 mV, a clear cur- rent-wave showed the oxidation of cysteine, which corresponds to a value of 390 m V versus SHE. The half-wave potential of a 1 mM iron porphyrin solution, pH 7, 50 mM phosphate buffer was measured. A broad, flat current-wave at -160 m V (60 mV versus SHE) was found, which corresponds well with the findings of

4 In order to get the potential versus the standard hydrogen electrode (SHE), one has to add 220 m V: -600 mV + 220 mV = -380 mV (Clark, 1960). B·7

Schoder: + 65 mV versus SHE (Schoder, 1975). The form of the current-wave implies that the redox process at the mercury electrode was not very fast, be- cause the peak was broad. A broad peak means that the speed of the potential change is rather fast compared to the redox process. For reversible, kinetically not hindered redox reactions the half-wave potentials show the redox potentials of the solution, under the given condi- tions (pH, concentration, temperature) (Lund and Baizer, 1991). As the redox- potentials of cysteine and iron porphyrin could be reproduced with ± 10 m V compared to literature values, this method was used to determine the de- pendence of the redox potentials of iron porphyrin as a function of cysteine and phosphate concentration. The experimental conditions for the evaluation of the influence of cysteine concentration on the redox properties of an iron porphyrin solution were: 1 mM iron porphyrin, 50 mM phosphate, pH 7, T = 25°C, and cysteine concentration: 0, 0.3, 0.5, 0.6, 0.8, 1.0, 1.2, 1.6, 2, 3, 4, 5, 8 mM. Addition of ali- quots of cysteine stock solutions to the iron porphyrin solution were chosen such that the increase in volume did not exceed 5%. After each addition of the cysteine aliquot the solution was stirred for 3 minutes and then purged for 5 minutes without further stirring. The influence of the phosphate concentration on the redox potential of the iron porphyrin was checked by titrating a phosphate solution (pH 7) to a 1 mM iron porphyrin solution. Here before each polarographic measure- ment pH was measured as the solution of iron porphyrin without buffer was slightly acidic (pH -5.2). Phosphate concentrations were: 0.3, 1, 2, 5, 10, 20 and 50 mM. The procedure was the same as in the case of cysteine titration. The resulting data were fitted with the program FITEQL (Herbelein and Westall,

1994). Association constants for one axial cysteine ligand (K1) and for two axial phosphate ligands (~2) were determined by fitting log t.iE versus log[ligand]. The fitting parameters for the fitting of the cysteine association con- stant with FITEQL are given in Table B.2. B-8

Table B.2: Table of components and species for the fitting of K1 with FITEQL

Components IDlt component x logX T TYPE 1 Cys-s· 10·' -4 0 I 2 Fell 10-' -6 5 10·• I 3 FeIII 10·' -6 5 10-4 I 6 HPOt 3.12 10-' -1.5 5 10"' I 4 E 1 0 0 II 5 H• 10-7 -7 0 III Species Matrix A IDlt name logK Cys-S- Fell FeIII HP0,2" E H+ 1 Cys-SH 8.2 1 0 0 0 0 1 2 Cys.S- 0 1 0 0 0 0 0 3 Fe2Cyss· 7 1 1 0 0 0 0 4 Fe3Cyss· 5 1 0 1 0 0 0 5 Fe2 0 0 1 0 0 0 0 6 Fe3 0 0 0 1 0 0 0 7 H,PO; 7.21 0 0 0 1 0 1 8 HP0,2' 0 0 0 0 1 0 0 9 Fe2(HPO,'"), 5.435 0 1 0 2 0 0 10 Fe3(HP0.'°), 3.524 0 0 1 2 0 0 11 E 1.904 0 1 -1 0 0 0 Sped es Matrix B ID# name logK Cys-s· Fell FeIII HPO/" E H• 1 Cys-SH 8.2 1 0 0 0 0 0 2 Cys-s· 0 1 0 0 0 0 -1 3 Fe2CysS" 7 1 1 0 0 0 0 4 Fe3Cyss· 5 1 0 1 0 0 0 5 Fe2 0 0 1 0 0 0 0 6 Fe3 0 0 0 1 0 0 0 7 H,PO; 7.21 0 0 0 1 0 0 8 HPO,'" 0 0 0 0 1 0 -1 9 Fe2(HP0,2'), 5.435 0 1 0 2 0 0 10 Fe3(HPO,'"), 3.524 0 0 1 2 0 0 11 E 1.904 0 0 0 0 1 0 Input Data# Cys-s· logE 1 3 10"' -1.356 10·1 2 5 lff' -1.864 10"1 3 6 10-4 -2.542 10·1 4 8 10-4 -3.390 10·1 5 1 10-3 -4.237 10"1 6 1.2 10·3 -4.237 10" 7 1.6 lff3 -5.932 10·1 8 2 10-3 -6.441 10"1 9 3 10'3 -7.627 10- 1 10 410" -7.627 10-1 11 5 10" -8.474 lff' 12 110·2 -9.322 10·1 8-9

Some comments have to be made concerning Table B.2 and the pa- rameters given therein. In the list of components Cys-s· stands for cysteine, Fell for reduced iron porphyrin species, FeIII for oxidized ones, HPOt for phosphate, H+ for proton and E was arbitrarily defined as [Fe(II)P] . This E - [Fe(III)P] value should not be mixed up with E that is often used for the redox potential, the electromotoric force. Hence, the connection between E =

[Fe(II)P] and the measured half-wave potentials is given by the NERNST [Fe(III)P] equation (equation B-1).

E = E _ RT ln [Fe(II)P] B-1 h 0 nF [Fe(III)P] where Bi, = redox potential under given conditions Eo = standard redox potential of the pair Fe(III)P /Fe(II)P R =gas constant: 8.314 Jmol4 K 1 T = temperature (K) n = number of electrons transferred, here n = 1 F = Faraday constant: 9.6485 104 C moJ-1 [Fe(Il)P] = concentration of reduced iron porphyrin [Fe(ill)PJ = concentration of oxidized iron porphyrin

If the difference between Eh (with cysteine) and Eh (without cysteine) is expressed by equation B-2, then with the replacement of [Fe(II)P] by E, setting [Fe(UI)P] n = 1 and turning to a logarithmic expression equation B-3 is obtained.

Eh(with cysteine)-Eh(without cysteine) = 6E =RT ln [Fe(ill)P] B-2 nF [Fe(II)P]

- ~ = lo E = lo [Fe(II)P] B-3 59mV g g [Fe(IIl)P]

The species Cys-SH and Cys-s· stand for the protonated and deproto-

2 nated form of cysteine and analogously fiiP04 and HP04 • for the protonated and deprotonated form of phosphate. The cysteine ligated iron (II) and (III) B-10 porphyrin is expressed by Fe2Cys-s· and Fe3Cys-s· respectively. Fe2 and Fe3 are iron (II) and (III) porphyrin complexes ligated by water and/ or hydroxide. The parameters for the fitting of the results of the phosphate titration with FlTEQL are given in Table B.3.

Table B.3: Table of components and species for the fitting of /32 with FITEQL

Components ID# component x logX T TYPE 1 HPO.'" 10-• -4 0 I 2 Fell 10·6 -6 5 rn-• I 3 FclII 10·• -6 5 io-• I 4 E 1 0 0 n 5 H+ 10·' -7 0 m Species Matrix A ID# name logK HPO,'" Fell Fem E ff 1 H2PO; 7.21 1 0 0 0 1 2 HPO,'" 0 1 0 0 0 0 3 Fe2(HPO.'°), 4 2 1 0 0 0 4 Fe3(HPO,'"), 3.5 2 0 1 0 0 5 Fe2 0 0 1 0 0 0 6 Fe3 0 0 0 1 0 0 7 E 0 0 1 -1 0 0 Species Matrix A ID# name logK HP0,2" Fell FclII E H+ 1 H,PO, 7.21 1 0 0 0 0 2 HPO," 0 1 0 0 0 -1 3 Fe2(HPO/"), 4 2 1 0 0 0 4 Fe3(HPO.'"), 3.5 2 0 1 0 0 5 Fe2 0 0 1 0 0 0 6 Fe3 0 0 0 1 0 0 7 E 0 0 0 0 1 0 Input Data # logH+ HP0,2" logE 1 -5.87 6 10"' -5.1 lo-1 2 -6.32 110" -7.6 10'1 3 -6.64 2 10-3 -1.695 4 -6.93 5 10-3 -1.949 5 -6.93 110" -1.864 6 -6.96 2 10-' -1.695 7 -6.97 510" -1.864

With the conditions given in Table B.3 p2 for Fe(II)(HP0/")2 and for

Fe(III)(HP0/')2 was fitted and with the conditions in Table B.2 K 1 for Fe(II)(Cys-s·) and for Fe(III)(Cys-S") was fitted. Data from Barron et al. (Barron, B·11

1937), where cyanide was titrated to a phosphate-buffered hemine solution, pH 12, were fitted with FITEQL additionally. In Table B.4 the fitted association constants and association constants from literature of Fe2+-, Fe3+-, Fe(II)porphyrin- and Fe(III)porphyrin-complexes are given for an overview. 8·12

Table B.4: Association constants of free aqueous iron and of iron porphyrin with various ligands

Free aqueous iron ligand constant Fe2+aq constant Fe3+aq reference

(Smith and SCN· log K 0.81 log K 2.24 1 1 Martell, 1989)

PO/ log K1 7.03 log K1 11.14 "

H:,C.COO· log K1 1.82 log K1 2.63 HCN 35.4 43.6 log P6 log P.

Off logK, 4.5 log K1 11.81 log P2 7.4 log P2 22.3 9.6 34.4 log p4 log P4

pKa = 6.02 HN-,. log K1 5.88 log K1 4.7 ~~N pl

log K1 0.7 log K1

N pKa =5,28 0.9 0 log P2 log p2 pKa =8.33 11.7 log P2 log p3 SH H30coo·

Iron porphyrin ligand constant Fe(Il)P constant Fe(III)P reference

HPOt log P2 -5.5 log P2 -3.5 this work

HCN log K1 5.63 log K1 5.4 (Goff and Morgan, 8-13

log Pz 9.55 log Pz 10.25 1976) (Barron, 5.52 log K, log K1 1937) pKa = 6.02 HN, log Pz 4.32 log P2 4.67 (Goff and ~l..;N pKa = 1.7 Morgan, 1976) H 3 coo· pKa =9.1

log K1 1.88 log K1 (Brown et () al., 1989) N pKa = 5.28 5.6 3.1 log P2 log Pz (Goff and 7.8 10.2 Morgan, log P, log P2 1976) (Kadish and Bottomley, 1980) pKa =8.33 log K 8.55 logK 5.5 this work SH 1 1 H3N~coo-

P(Bun) 3 log K1 4.66 logK1 (Ellis et al., 1980) ff P(0Et)3 log K1 3.24 log K1 logK 5.44 log K ()NC 1 1

HN, log Pz 5.1 log Pz 6.05 (Goff and l..;N Morgan, 1976) log P2 7.8 log P2 10.5 (Kadish and (( Bottomley, N 1980) 8.3 11.4 ff log P2 log P2

6N ff OUCI log P2 5.5 log Pz 3.4 "" N I log 9.7 log 16.3 "N / Pz P2

6N Curriculum Vitae

Feb. 12,1969 Born in St. Gallen (Switzerland) 1976-1984 Primary and secondary school, Arbon 1984 - 1988 Kantonsschule Romanshorn 1988 Matura type B (Latin)

1988 -1992 Studies in chemistry (Abt. IV), ETH Zurich Summer 1991 Practical work with BP (British Petrol) in Lisbon May 1992 Diploma work in the group of Prof. S. Benner Subject: Synthesis of nucleoside analoga and selective base introduction Nov. 1992 Diploma in chemistry

May 1993 Practical work as chemistry teacher at Kantonsschule Frauenfeld Nov. 1993 "Hoheres Lehramt" in chemistry ETH Zurich

1994 - 1998 Doctoral studies in environmental organic chemistry, EA WAG 1994-1998 Teaching assistant, EAWAG and ETH Zurich