STUDIES ON THE PRODUCTION OF
HYDROGEN PEROXIDE
A Thesis Submitted by
PHILIP ANTHONY CLAPP
In Partial Fulfilment of the Requirements
for the Award of the Degree of
Doctor of Philosophy
of the
University of London.
Department of Chemistry,
Imperial College,
London SW7 2AY. September 1990 The creatures outside looked from pig to man and from man to pig, and from pig to man again, but already it was impossible to say which was which.
George Orwell
'Animal Farm' To my parents, for a debt I can never repay. A bstract.
The production of hydrogen peroxide, H2O2, by the amine-catalysed reduction of a number of substrates by hydrogen sulphide and subsequent re-oxidation by dioxygen is reported. Substrates include alkyltetrahydroanthraquinones, dialkylnaphthoquinones, azocompounds and flavins. The highest yields (>90%) were obtained with the quinone systems, and good turnovers of substrate and catalyst were shown by a series of runs on one reaction solution. Spectroscopic studies have been made in an attempt to elucidate the mechanism of these reactions, and kinetic data are given for the quinone systems. A method for the spectrophotometric determination of H2O2 (Ixl0'5-3xl0-3 M) with titanium(IV) sulphate solution after extraction with ethyl acetate has been developed. This procedure can be applied in the presence of a variety of species which interfere with the conventional colorimetric method. A more sensitive method of spectrophotometric determination of H2O2 in aqueous solution has been reported, using leuco Patent Blue Violet and peroxidase. The method
allows analysis of H2O2 in the range 10~7-10-5 M, and the analysis has been applied to the ethyl acetate extraction method described previously. In an adaptation of the direct LPBV method, glucose has been analysed. A quantitative study has been made of the formation of H2O2 when alkaline solutions of tannic acid, gallic acid and a number of other molecules containing the catechol (1,2- dihydroxybenzene) moiety are exposed to air or dioxygen. For gallic acid, from the pH dependence of the initial rates of reaction, it is shown that over the pH range ca. 7.0-
8.5, the predominant reactive species is the dianion. The photochemical production of H2O2, using 254nm radiation, has been studied for a variety of air-saturated solutions of phenolic compounds at pH 6.0 or below, where the thermal reaction is negligible, and apparent quantum yields are reported. Appreciable formation of H2O2 is observed with 1,2-, 1,3- and 1,4-benzenediols. Possible
mechanisms for the thermal and photochemical reactions are discussed, and it is suggested that benzenediol groups may be involved in the natural photochemical production of H2O2 in surface waters.
2 ACKNOWLEDGEMENTS
I owe a great debt of thanks to Professor Dennis Evans, for giving me the opportunity
to study for a PhD. in the first place, and for his tireless enthusiasm, guidance and
friendship over the past three years.
I would like to thank Peroxid-Chemie GmbH and Interox Oa ^mccAs Ud. fix 'H&k
bursary and funding for this project, and in particular Bill Logan and Don Williams for
their expertise and interest.
Lab. 642 has a certain 'ambience' that is due in no small part to those people who have
worked, and continue to work, there. I am grateful to Jon 'Pilsbury Dough-Boy' Parr
and David 'Splendid' Jacubovic for making my first year there so enjoyable, and to
Caroline 'Deed Poll' Denekamp and Chih 'Hands of Lead' Wong and Nobuhiko
"Anglophile’ Iki for maintaining the friendly and, er, cooperative atmosphere. May
your yields be bountiful and your beakers never empty.
The undergraduate reports of Grace Choi, Ngoc Du, Kate Moore, Adrian Pagan and
Katherine Thompson have been of invaluable use and their hard work deserves much credit.
The technical support of Colin Robinson and Roger Lincoln is gratefully acknowledged.
Peter Beardwood is thanked for his help with the epr spectra, Eric Coker with the x-ray powder diffraction spectrum, Dr. J. Derek Woollins with the Raman spectrum and Dr.
Bill Young for his expertise with software. Chih Wong gets a second plaudit for his assistance with the computer program for the kinetics studies.
Thanks are also due to the dedicated exponents of the white heat of technology on the
5th and 6th floors, without whose acquaintance I would be much the poorer. Dr. Andy
Dengel deserves a special mention for six years of stoic friendship, but also high on the roll of dishonour are Bernie A, Andy B,Vic C, Fitz., Allen F, Mick G, Pete G, Mark
G, Sarah G, Vahe H, Steve H, Dave H, Ben H, Jen, Robin K, Paul K, Rob K, Chris
3 L, Dr. Magnox, Greg McD, Brian McG, Andy M, Jed O, Ivan P and Girish R.
The Young proteges of 1984, messrs. Anderson, Cracknell, Kalia, Shipman and
Shukla are due credit, as are Martin D, Steve E, Suki K and Andy P.
Dr. G Brent Young is gratefully acknowledged for his great patience and friendship and
for forming, with Dr. Don Craig, the artistic base of that renowned Celtic beat trio, The
Atholl Highlanders. How does it go again?
The ragged band of Holbein 1984 are remembered, Neil B, Gord D, Andy R, Jules V
and Simon W, but especially Andy Keelin and Scott Gordon, Fulham 1986-87.
Thanks to Roly Smith for his help, and especially to Trisha Ingles, Nigel Stokes, Kate
Rudlin, Honey Melville Brown and re-apps Rachel C, Andy H, Myles J, Andy J, Alex
Y for making Selkirk Heaven.
Imperial College Rugby Club has detained me more times than I care to remember, both
on the field and off. This is entirely due to the members of the club; a more affable and
friendly bunch would be hard to find. Particular thanks are given to Mike Anderson,
Pete Drew, Dave and Andy Fleming, Pete Galley, Brian Greensmith, Simon Hall, Rob
Hargrove, Chris Hughes, J F Lucas, Jason Jenkins, Fergus MacDonald, Simon
Rowell, Si Smith, Rich Walters and Andy Watson. Thanks also to Dave Boyce, Dr.
Bob Coutts, Tony Morland and Dave Wilson of RCS.
’You can't always get what you want, but if you try sometimes
you just might find that you get what you need.'
4 CONTENTS page.
ABSTRACT. 2
ACKNOWLEDGEMENTS. 3
CONTENTS. 5
LIST OF FIGURES. 12
LIST OF TABLES. 16
ABBREVIATIONS. 18
0.1 INTRODUCTION - Hydrogen Peroxide.
0.1.1 Physical and Chemical Properties. 20
0.1.2 History. 22
0.1.3 Manufacture. 23
0.1.4 Uses. 26
0. 1.5 Occurrence in the Environment. 27
1. PRODUCTION OF HYDROGEN PEROXIDE FROM
HYDROGEN SULPHIDE AND DIOXYGEN.
1.0 INTRODUCTION - Hydrogen Sulphide.
1.0. 1 Sources. 34
1.0. 2 Sulphides as Reducing Agents. 35
1.0. 3 Thermodynamics of Proposed System. 36
1.0. 4 Proposed Reaction Process. 37
1.0. 5 Reduction-Oxidation Potentials of Substrates. 37
1.0. 6 Properties. 39
1.0. 7 Literature Survey. 40
1.1 RESULTS AND DISCUSSION.
1.1.1 Production of H2O2 by Quinone Compounds. 42
1.1.2 Production of H2O2 by 2-Alkyltetrahydroanthraquinones. 44
5 page.
1.1.3 Production of More Concentrated Solutions of H2O2. 45
1.1.4 Measurement of Losses of H2O2 by Reaction with Sulphur. 47
1.1.5 Effect of Amine Catalyst on the Reaction. 47
1.1.6 Catalysis of the Reduction by a Secondary Amine. 50
1.1.7 Continuous Run on One Reaction Solution. 50
1.1.8 UV/Visible Study of the Reduction of ETQ. 51
1.1.9 EPR Study of the Reduction of ETQ. 54
1.1.10 1H NMR Study of the Reduction of BTQ. 54
1.1.11 EPR Study of the Reduction of QS. 57
1.1.12 Infra-Red Study of the Interaction between H2S and Tributylamine. 61
1.1.13 X-Ray Powder Diffraction and Raman Study of Sulphur. 63
1.1.14 Production of H2O2 by 2,3-Dialkyl-1,4-naphthoquinones. 65
1.1.15 Kinetics Studies of Reductions and Oxidations of Quinones. 66
1.1.16 Proposed Mechanism for Amine-Catalysed Reduction of a 75
Quinone by H2S.
1.1.17 Proposed Mechanism for Oxidation of a Quinone by Dioxygen. 77
1.1.18 Production of H2O2 by Azocompounds. 79
1.1.19 Estimate of Yield Based on Hydrogen Sulphide. 80
1.1.20 Production of H2O2 by Flavins. 80
1.2 PHYSICAL MEASUREMENTS
1.2.1 Standard Reaction Apparatus. 82
1.2.2 Standard Reaction Procedure. 83
1.2.3 Standard Reaction Apparatus for Kinetics Studies. 84
1.2.4 Standard Reaction Procedure for Kinetics Studies. 84
1.2.5 UV/Visible Study of the Reduction of ETQ. 87
1.2.6 EPR Study of the Reduction of ETQ. 87
6 page.
1.2.7 EPR Study of the Reduction of ETQ Catalysed by 88
l,8-Bis(dimethyl)aminonaphthalene.
1.2.8 EPR Study of Reduction of QS. 88
1.2.9 1H NMR Study of the Reduction of BTQ. 88
1.2.10 Infra-red Study of Interaction between H2S and Tributylamine. 89
1.3 EXPERIMENTAL PROCEDURES.
1.3.1 Iodometric Analysis. 90
1.3.2 Argon Flushing. 90
1.3.3 Reduction of Sulphonated Anthraquinones. 90
1.3.4 Removal of Sulphur with Carbon Disulphide. 90
1.3.5 Carbon Dioxide Flushing. 91
1.3.6 Continuous Run on One Reaction Solution. 91
1.3.7 Purification of Sulphur for X-Ray Powder Diffraction and Raman. 92
1.3.8 Re-Oxidation of Hydrazobenzene. 92
1.3.9 Reduction of Azobenzene at Atmospheric Pressure of H2S. 92
1.3.10 Reduction of Riboflavin-2’,3',4',5'-tetraacetate. 93
1.4 MATERIALS AND SYNTHESES.
1.4.1 Solvents. 94
1.4.2 Amines. 94
1.4.3 Anthraquinones. 94
1.4.4 Tetrahydroanthraquinones. 94
1.4.5 1,4-Naphthoquinones. 94
1.4.6 Azobenzenes. 96
1.4.7 Flavins. 98
1.5 GENERAL CONCLUSIONS. 99
7 page.
II. SPECTROPHOTOMETRIC DETERMINATION OF
HYDROGEN PEROXIDE.
2. SPECTROSCOPIC DETERMINATION OF HYDROGEN
PEROXIDE WITH TITANIUM (IV) SULPHATE.
2.0 INTRODUCTION. 102
2.1 RESULTS AND DISCUSSION. 103
2.1.1 Spectrophotometric Determination of H2O2 after Extraction 103
with Ethyl Acetate.
2.1.2 Effect of Interfering Species on Extraction. 107
2.1.3 Extraction Method with Isobutanol. 107
2.2 EXPERIMENTAL PROCEDURES.
2.2.1 Analysis Procedure. 109
2.2.2 Analysis of Effect on Extraction of Interfering Species. 109
2.2.3 Use of Catalase to Decompose H2O2 in Sample. 110
2.3 MATERIALS AND SYNTHESES.
2.3.1 Titanium (IV) Sulphate Solution. 110
2.3.2 Hydrogen Peroxide. 110
2.3.3 Ethyl Acetate. 110
3. 'SIX-FOLD' ANALYSIS OF HYDROGEN PEROXIDE.
3.0 INTRODUCTION. Ill
3.1 RESULTS AND DISCUSSION. 112
3.2 EXPERIMENTAL PROCEDURE. 112
3.3 MATERIALS AND SYNTHESES.
3.3.1 Hydrogen Peroxide. 113
3.3.2 Starch. 113
3.3.3 Other Chemicals. 113
8 page.
4. PEROXIDASE-CATALYSED ANALYSES OF HYDROGEN
PEROXIDE.
4.0 INTRODUCTION. 114
4.1 RESULTS AND DISCUSSION.
4.1.1 Direct Analysis of H2O2 by Peroxidase-Catalysed Reactions. 118
4.1.2 Analysis of H2O2 by Peroxidase-Catalysed Reactions after 119
Extraction with Ethyl Acetate.
4.2 EXPERIMENTAL PROCEDURES.
4.2.1 Direct Analysis of H2O2. 120
4.2.2 Extraction Analysis of H2O2. 120
4.3 MATERIALS AND SYNTHESES.
4.3.1 Substrates. 121
4.3.2 Hydrogen Peroxide. 121
4.3.3 Ethyl Acetate. 121
4.3.4 Peroxidase. 121
4.3.5 Buffers. 122
4.3.6 Solutions. 122
5. SPECTROPHOTOMETRIC DETERMINATION OF
HYDROGEN PEROXIDE AND GLUCOSE WITH
LEUCO PATENT BLUE VIOLET.
5.0 INTRODUCTION. 123
5.1 RESULTS AND DISCUSSION.
5.1.1 Direct Spectrophotometric Determination of H2O2 with LPB V. 125
5.1.2 Determination of H2O2 after Ethyl Acetate Extraction. 128
5.1.3 Determination of Glucose. 130
5.1.4 Use of Other Leuco Dyes. 130
9 page.
5.2 EXPERIMENTAL PROCEDURES.
5.2.1 Direct analysis of H2O2 with LPBV. 132
5.2.2 Extraction Analysis of H2O2 with LPBV. 132
5.2.3 Direct Analysis of Glucose with LPBV. 132
5.3 MATERIALS AND SYNTHESES.
5.3.1 Leuco Patent Blue Violet. 133
5.3.2 Hydrogen Peroxide. 134
5.3.3 Ethyl Acetate. 134
5.3.4 Enzymes. 134
5.3.5 Buffers. 134
5.3.6 Solutions. 135
5.3.7 Glucose. 136
GENERAL CONCLUSIONS FOR ANALYSES. 137
III THERMAL AND PHOTOCHEMICAL PRODUCTION
OF HYDROGEN PEROXIDE FROM DIOXYGEN AND
TANNIC ACID, GALLIC ACID AND RELATED
COMPOUNDS IN AQUEOUS SOLUTION.
6.0 INTRODUCTION. 140
6.1 RESULTS AND DISCUSSION. 142
THERMAL PRODUCTION OF HYDROGEN PEROXIDE.
6.1.1 Tannic Acid. 142
6.1.2 Gallic Acid. 142
6.1.3 Pyrogallol (1,2,3-Trihydroxbenznene) and 146
1,2,4-Trihydroxbenzene.
1 0 page.
6.1.4 Dihydroxybenzenes. 146
i. Catechols.
ii. Resorcinols.
PHOTOCHEMICAL PRODUCTION OF HYDROGEN PEROXIDE.
6.1.5 Catechols. 151
6.1.6 Resorcinols. 155
6.2 PHYSICAL MEASUREMENTS.
6.2.1 Thermal Reactions. 156
6.2.2 Photochemical Reactions. 156
6.2.3 Calculation of Quantum Yields. 157
6.3 MATERIALS AND SYNTHESES.
6.3.1 Phenolic Compounds. 158
6.3.2 Buffers. 158
6.3.3 Enzymes. 158
6.4 GENERAL CONCLUSIONS. 159
REFERENCES. 160
APPENDIX - Computer Programme for Calculation 168
of Kinetic Parameters. LIST OF FIGURES
Figure. Page.
0.1 Structure of hydrogen peroxide in the gas phase. 20
0.2 Structure of hydrogen peroxide in the solid phase. 20
0.3 Production of hydrogen peroxide by autoxidation of 24
alkylanthraquinol.
0.4 Shell process for production of hydrogen peroxide. 25
0.5 Applications of hydrogen peroxide as percentage of total 26
capacity for 1985; USA.
0.6 Applications of hydrogen peroxide as percentage of total 26
capacity for 1985; Europe.
0.7 Diel changes in hydrogen peroxide concentration compared to 33
photosynthetically active radiation, May 10 to 14, Jacks Lake,
Ontario, Canada.
1.1 Dialkyl naphthoquinones described in Patent Bel. 652,956. 41
1.2 2-Alkyltetrahydroanthraquinones. 44
1.3 UV/visible absorbance spectrum of the reduction of 2-ethyltetra 52
hydroanthraquinone (0.1M) by hydrogen sulphide. 1mm path
length cell.
1.4 UV/visible absorbance spectrum of the reduction of 2-ethyltetra 53
hydroanthraquinone (0.01M) by hydrogen sulphide. 10mm path
length cell.
1.5 !H nmr spectrum showing aromatic resonances of 2-lbutyltetra 55
hydroanthraquinone (inset).
1.6 nmr spectrum showing aromatic resonances of 2-tbutyltetra 56
hydroanthraquinol (inset)- product of reduction of BTQ with
hydrogen sulphide.
1 2 Figure. Page.
1.7 EPR spectrum of sodium anthraquinone-2-sulphonate reduced 58
by hydrogen sulphide.
1.8 EPR spectra of the reduction of sodium anthraquinone-2- 59
sulphonate by hydrogen sulphide, recorded at timed intervals
after the introduction of H2S.
1.9 EPR spectrum of sodium anthraquinone-2-sulphonate reduced 60
by sodium dithionite.
1.10 Infra-red spectrum of hydrogen sulphide in tetrachloroethylene. 62
1.11 Infra-red spectrum of hydrogen sulphide in tetrachloroethylene 62
plus 1% tri-n-butylamine.
1.12 X-ray powder diffraction spectrum of sulphur produced by 64
reduction of 2-ethyltetrahydroanthraquinone by hydrogen
sulphide.
1.13 Raman spectrum of sulphur produced by reduction of 2-ethyl 64
tetrahydroanthraquinone by hydrogen sulphide.
1.14 Kinetics Data for reduction of varying concentrations of 69
2-amyltetrahydroanthraquinone by hydrogen sulphide.
1.15 Kinetics data for reduction of 2-amyltetrahydroanthraquinone 71
by hydrogen sulphide, varying concentrations of amine catalyst.
1.16 Kinetics data for reduction of 2,3-dialkyl-1,4-naphthoquinones 72
by hydrogen sulphide.
1.17 Kinetics data for oxidation of varying concentrations of 2,3- 74
dialkyl-1,4-naphthoquinones by oxygen.
1.18 Proposed mechanism for the amine-catalysed reduction of a 76
quinone by hydrogen sulphide.
1 3 Figure. Page.
1.19 Proposed mechanism for the oxidation of a hydroquinone by 78
oxygen.
1.20 Riboflavin-2',3',4',5'-tetraacetate. 81
1.21 Standard reaction apparatus. 82
1.22 Standard reaction apparatus for kinetics studies. 84
2.1 Ethyl acetate extraction analysis with 5ml of Ti(IV) reagent. 104
2.2 Ethyl acetate extraction analysis with 3ml of Ti(IV) reagent. 105
2.3 Ethyl acetate extraction analysis with 1ml of Ti(IV) reagent. 106
2.4 Isobutanol extraction analysis with lml of Ti(IV) reagent. 108
4.1 3-D structure of horse-radish peroxidase. 115
4.2 Structure of porphyrin ring in peroxidase. 116
4.3 Reaction scheme for peroxidase-catalysed oxidation of 117
substrate DH2.
5.1 Crystal Violet. 123
5.2 Leuco Patent Blue Violet, C.I. 42045. 124
5.3 Direct analysis with LPBV reagent - high concentration range. 126
5.4 Direct analysis with LPBV reagent - low concentration range. 127
5.5 Ethyl acetate extraction analysis with LPBV reagent. 129
5.6 Direct analysis of glucose with LPBV-glucose reagent. 131
6.1 Tannic acid. 141
6.2 Concentrations of hydrogen peroxide produced at various 143
times for 2x10-4 mol dm-3 aqueous solutions of tannic acid
saturated with air at a) pH 8.58, b) pH 7.99, c) pH 7.51,
d) 7.00.
1 4 Figure. Page.
6.3 Concentrations of hydrogen peroxide produced at various 144
times for IxlO-3 mol dm-3 aqueous solutions of gallic acid
saturated with air at a) pH 8.52, b) pH 8.07, c) pH 7.48,
d) 7.00.
6.4 Galloflavin. 145
6.5 Concentrations of hydrogen peroxide produced at various times 147
for air-saturated aqueous solutions of a) lxlO-3 mol dm-3
pyrogallol at pH 8.00, b) 1x10-3 mol dm-3 gallic acid at pH 8.07,
c) lxlO-3 mol dm*3 1,2,4-trihydroxybenzene at pH 6.00,
d) 2xl0-3 mol dm-3 catechol at pH 8.50.
6.6 Concentrations of hydrogen peroxide produced at various times 148
for air-saturated 2x10-3 mol dm-3 aqueous solutions of a) catechol
at pH 8.50, b) 3-0-methyl gallic acid at pH 8.58, c) 3,4-dihydroxy
benzoic acid at pH 8.50.
6.7 Possible mechanism for formation of H2O2 from oxygen and a 149
catechol.
6.8 Possible mechanism for photochemical formation of H2O2 from 154
oxygen and a catechol.
6.9 Apparatus for photochemical reactions. 157
1 5 LIST OF TABLES
T able. Pag
1.1 The effect of substituents on redox potentials of anthraquinones. 38
1.2 The effect of substituents on redox potentials of naphthoquinones. 38
1.3 The effect of substituents on redox potentials of azocompounds. 38
1.4 Solubility data for hydrogen sulphide in organic solvents. 39
1.5 Summary of reactions of quinones by standard procedure. 43
1.6 Summary of reactions of alkyltetrahydroanthraquinones by 45
standard procedure.
1.7 Summary of a series of reduction/oxidation runs on a solution 46
of ATQ.
1.8 Values of pK for a range of bases, and yields of H2O2 obtained 49
when used in standard reaction procedure.
1.9 Yields of H2O2 and analyses of losses for a continuous mn of 51
reduction/oxidation reactions on a solution of ATQ.
1.10 1H nmr chemical shifts for the aromatic protons of BTQ and its 57
reduction product.
1.11 Summaiy of reactions of 2,3-disubstituted-1,4-naphthoquinones by 66
standard procedure.
1.12 Comparison of kinetic data for reduction of varying concentrations 70
of ATQ.
1.13 Comparison of kinetic data for reduction of ATQ catalysed by varying 70
concentrations of tri-n-octylamine.
1.14 Comparison of kinetic data for reduction of naphthoquinones. 73
1.15 Comparison of kinetic data for oxidation of naphthoquinones. 73 Table. Page.
1.16 Summary of reactions of azobenzenes by standard procedure. 79
2.1 Limits of interfering species upon extraction method. 107
4.1 Summary of results for analysis of hydrogen peroxide by various 119
peroxidase-catalysed colorimetric methods.
4.2 Summary of results for analysis of hydrogen peroxide after 120
extraction with ethyl acetate by various peroxidase-catalysed
colorimetric methods.
5.1 Summary of results obtained with commercial samples of ethyl 130
acetate.
6.1 Initial rates of production of hydrogen peroxide as a function of pH 145
for lxlO-3 mol dm-3 aqueous solutions of gallic acid saturated with
air.
6.2 Variation of quantum efficiency with pH for an air-saturated 4x10-3 151
mol dm-3 aqueous solution of gallic acid. 30 minutes irradiation at
254nm.
6.3 Variation of quantum efficiency with pH for an air-saturated 4x 10-3 152
mol dm-3 aqueous solution of pyrogallol. 30 minutes irradiation at
254nm.
6.4 Efficiencies for the production of hydrogen peroxide for air-saturated 153
4x10-3 mol dm-3 aqueous solutions of aromatic compounds at
pH 6.00±0.01. 30 minutes irradiation at 254nm.
1 7 ABBREVIATIONS.
AAP - 4-Aminoantipyrine.
ABTS - 2,2'-Azinodi(3-ethylbenzthiazoline-6-sulphonic acid).
ATQ - 2-Amyltetrahydroanthraquinone.
BTQ - 2-tButyltetrahydroanthraquinone.
BzlNAH • N-Benzyl-1,4-nicotinamide.
DBC - Diisobutylcarbinol.
DCPS - 2,4-Dichlorophenolsulphonate.
DL-DOPA - 3-(3,4-dihydroxyphenyl)-DL-alanine (DL-Dopamine).
DMF - N,N-Dimethylformamide.
DMAB - 3-(Dimethylamino)benzoic acid.
DMH - 2,6-Dimethyl-4-heptanol.
DMN - 2,3-Dimethyl-1,4-naphthoquinone.
EMN - 2-Ethyl-3-methyl- 1,4-naphthoquinone.
EPPS - 4-(2-Hydroxy)-1-piperazinepropanesulphonic acid.
EQ - 2-Ethylanthraquinone.
ETQ - 2-Ethyltetrahydroanthraquinone.
FMN-Na - Flavin mononucleotide, sodium salt.
GH4 - Gallic acid.
HEPES - 4-(2-Hydroxyethyl)-l-piperazine-ethanesulphonic acid.
HRP - Horse radish peroxidase.
LPBV - Leuco Patent Blue Violet.
MN - 2-Methyl-1,4-naphthoquinone.
MOPS - 4-Morpholinepropanesulphonic acid.
NMP - 1-Methylpyrolidinone (N-Methylpyrollidone).
PHBS - p-Hydroxybenzoic acid.
PIPES - l,4-Piperazinebis(ethanesulphonic acid).
PTFE - Polytetrafluoroethylene. Q - Quantum yield.
QDS - Disodium anthraquinone-2,6-disulphonate.
QS - Sodium anthraquinone-2-sulphonate.
TBP - Tri-n-butyl phosphate.
TOPS - N-Ethyl-n-sulphopropyl-m-toluidine. p-Xyl - p-Xylene. 0.1 INTRODUCTION.
Hydrogen Peroxide.
0.1.1 Physical and Chemical Properties.
As might be expected, hydrogen peroxide has many physical properties that are similar to water, although it isd e n s e r 1.44 at 25°C). Pure H2O2 is a colourless liquid, boiling at 150.2°C and freezing at -0.43 °C. The H2O2 molecule has an unusual skew structure, which it adopts due to bonding pair-lone pair repulsions at each O atom. The
0 -0 bond distance in H2O2, as determined by X-ray measurements, corresponds to the value expected for a single 0 -0 bond, 1.47±0.02A1. In the gas phase, the dihedral angle 0 is 111.5°, Fig.0.1, decreasing to 90.2° in the solid state as a result of H- bonding, Fig.0.2.
Fig. 0.1 Gas phase Fig. 0.2 Solid phase
In aqueous solution, H2O2 is a weak acid, with pKa = 11.66 at 25°C2.
A excellent general text (published in 1955) by Schumb, Satterfield and Wentworth1
details the early history of hydrogen peroxide and includes methods of production, uses
and chemical and physical properties.
2 0 Hydrogen peroxide has a rich and well-studied chemistry which can be broadly divided into four categories of reaction; i. oxidation/reduction reactions; ii. formation of inorganic and organic peroxy compounds; iii. formation of perhydrates; iv. the self-decomposition (disproportionation) of H2O2.
The oxidation state (OS) of oxygen in H202is -1, which is intermediate between the
OS of oxygen in dioxygen (0S=0) and water (OS=-2). This enables H2O2 to act as either an oxidising or reducing agent.
The oxidation/reduction chemistry of H2O2 in aqueous solution can be summarised by the following potentials;
H20 2 + 2H+ + 2e- —------** 2H20 E=1.77V ....0.1
HO2" + H2O + 2e- 30H- E=0.88V ....0.2
0 2 + 2H+ + 2e- —------^ H20 2 E=0.70V. ....0.3
From the above it is clear that in either acidic or basic solution, H2O2 acts as a strong oxidising agent. Among many examples of this oxidising nature are the rapid oxidation
of ferrous Fe2+ ions to ferric Fe3+ ions, and the oxidation of organic sulphides to
sulphoxides;
2Fe2+ + H20 2 + 2H+ ------► 2Fe3+ + 2H20 ....0.4
R2S + H2O2 ------^ R2SO + H2O2 ------^ R2SOz + 2H20...... 0.5
However, in the presence of a stronger oxidising species, H2O2 will act as a reducing
agent. For example, ceric Ce4+ ions will be reduced to cerous Ce3+ ions, and silver
oxide will be reduced to metallic silver;
2Ce4+ + H20 2 ------► 2Ce3+ + 2H+ + 0 2 ....0.6
21 Ag20 + H2O2 2Ag + O2 + H20. ....0.7
It can also be seen that the above reactions broadly divide into one- and two-electron processes (eg. ferrous to ferric and organic sulphide to sulphoxide respectively).
Examples of organic and inorganic peroxy compounds include peracetic acid and uranium peroxide respectively. The latter is important in the purification of uranium from uranium ore (see later);
CH3COOH + H20 2 ------► CH3COOOH + H20 ....0.8
U022+ + H20 2 + nH20 ------► UO4.11H2O + 2H+. ....0.9
H2O2 forms both perhydrates and peroxo compounds with alkali metal salts. The most well-known of these are sodium perborate (Na2[B2(02)2(0H)4].6H20) and sodium carbonate hydroperoxidate (2Na2C03.3H2C>2), both of which are major components of heavy duty washing powders.
In the absence of any catalysing substance, H2O2 is relatively stable; spontaneous decomposition to O2 and H2O appears to occur very slowly, even by the so-called base- catalysed reaction;
H2O2+HO2------► H20 + OH- + 02 ....0.10
Evans and Upton3 quote a tentative rate constant for this uncatalysed reaction of
3.7xl0-8dm3 mol-1 s-1 at 35°C.
In the presence of a wide variety of substances ranging from traces of transition metal
ions to simple dust, the rate of decomposition is greatly accelerated.
0.1.2 History.
The discovery of hydrogen peroxide was reported by the French chemist Louis-Jacques
Thenard in July 18184. In fact, Thenard stated that he had produced a new class of
compounds which he termed "oxygenated acids", formed by the reaction of a range of
mineral acids with barium peroxide. Unknown at that time to Thenard, the identity of
2 2 the acid in each case was hydrogen peroxide.
Having later recognised that what he had in fact formed was oxygenated water,
Thenard produced the most systematic study of H2O2 in the nineteenth century. The preparative route followed by Thenard is summarised in the equations below;
B a0 2 + 2H+ ------► Ba2+ + H202. ....0.11
Thenard was able to produce a 33% by weight solution of hydrogen peroxide by this method5, and by subsequent vacuum distillation, was able to concentrate the solution of
H2O2 to ca. 98%6, an achievement not equalled for over seventy years7.
Thenard worked further on the stability of hydrogen peroxide in acid, alkali and neutral
solutions8, and the role that certain substances played in its decomposition9.
Schonbein attempted to explain many of Thenard’s observations in terms of negatively
or positively polarised oxygen, the so-called "ozone-antozone" theory10.11. This
followed Schonbeiris discovery of ozone itself, and his proof of it as an allotrope of
oxygen12. The theory of polarised oxygen was widely held until the 1870's.
Only at the beginning of the twentieth century was systematic research pursued into the
formation of addition products by H20213’U, its oxidative/reductive action 1536 and its
catalysed decomposition17.18. 19. The work of Wolffenstein7 and Schone20 appeared to
demonstrate, by their respective preparations of 99.1% solutions, that it was possible to
produce stable hydrogen peroxide in high concentrations.
0.1.3 Manufacture.
The industrial preparation of H2O2 remained essentially unchanged for almost one
hundred years, being essentially an adaptation of the method employed by Thenard.
Different solvent acids were used for different purity grades of H2O221’ with
hydrofluoric acid gradually gaining favour alongside hydrochloric and phosphoric
acids22. A wide variety of stabilisers were employed in an attempt to slow the
2 3 decomposition of the peroxide, ranging from sodium chloride, ethanol, sulphuric acid
and acetic acid23 to acetanilide and quinine salts24.
A number of electrochemical processes25 came into industrial use in the early 1900's
after the original discovery of the production of H2O2 by the electrolysis of H2SO4 by
Meidlinger in 185326. This was subsequently shown by Berthelot to be due to
hydrolysis of peroxydisulphuric acid at the anode27. These processes were particularly
widespread in Germany and Austria during WWII, and were later adopted worldwide
to such an extent that by 1955, they accounted for 80% of the world's production of
H20 2i .
The autoxidation process that has almost entirely superceded these electrochemical
methods in the industrial production of H2O2, primarily due to the rising costs of
electricity, was patented by IG Farbenindustrie in 193428. The autoxidation occurs, as
shown in fig.0.3, upon reaction of an afkylanthraquinol with molecular oxygen.
Fig.0.3. Production of hydrogen peroxide by autoxidation of alkylanthraquinol.
2 4 The alkylanthraquinol itself is formed by catalytic hydrogenation of the alkylanthraquinone. An alkylanthraquinone is used instead of the unsubstituted form due to solubility considerations. 2-tbutylanthraquinone is widely used, as are other similar derivatives, and modifications are still being made to the substrate to improve the industrial process29. The original German process used Raney nickel as a catalyst, whereas more recent processes use a palladium metal catalyst instead2^
The requirements for the solvent in the process are very exacting. It must; i. not be liable to oxidation; ii. be of low volatility, hence little is lost in removal of unreacted gases; iii. display high solubility for both the oxidised and reduced form of the substrate; iv. form no explosive mixtures with peroxides or oxygen. v. be practically immiscible with water.
The original German patent suggested a 1:1 mixture of benzene and various secondary alcohols, the latter being collectively termed "Paralk". However, due to modem safety considerations, solvents in use today contain mixtures of alkylbenzenes, and longer chain alcohols or esters. The H2O2 formed is extracted into water and concentrated by vacuum distillation.
A process patented by Shell Ltd. that has gained limited use since the 1950's involves the oxidation of isopropanol to acetone and hydrogen peroxide30. This is performed mostly in the liquid phase at 15-20 atm. of oxygen and 100°C, fig.0.4.
Me2C=0 + H20
Fig.0.4. Shell process for production of hydrogen peroxide. Despite the ca. 80% selectivity of the process for the production of hydrogen peroxide, secondary reactions occur to yield peroxides of aldehydes and acids. The H2O2 formed is extracted into water and purified by ion-exchange.
0.1.4 Uses.
The industrial importance of H2O2 is due to its wide range of applications. In 1987 the
equivalent of 883,000 tonnes of H2O2 were produced worldwide (excluding E.
Europe)30. The relative amount used in each application varies from one country to the
next. Figs.0.5 and 0.6 show the % of total capacity (1985) for each use in the
USA and Europe30 respectively. The 'other* category includes a wide range of minor
applications, including those in the food, electronics and cosmetics industries, as well
as the purification of uranium ore in the nuclear power industry (see later). Cellulose
bleaching includes paper and wood, while the figures for chemical synthesis apply to
the manufacture of perborates and percarbonates in Europe, and contain a 6%
contribution from the production of peroxo compounds in the case of the USA.
Figs. 0.5,0.6; Applications of hydrogen peroxide as a percentage of total capacity for
1985; USA, Europe respectively.
2 6 It can be seen from Figs.0.5 and 0.6 that the predominant use of H2O2 is as a bleach, its bleaching activity apparently seeming to depend upon its oxidising ability. Although there are probably several different ways in which hydrogen peroxide acts as a bleaching agent, the predominant pathway is presumed to be by oxidation of chromophores, changing the relative energies of molecular orbitals between which the electronic transition is occurring and moving the wavelength of light which is absorbed out of the visible region1.
As early as the 1880's bleaching of hair, wool, cotton, straw and jute were practised on a small-scale. Advances in cotton bleaching, particularly the advent of the Du Pont
"continuous" process in 1939 made the procedure much more efficient, and by 1948 most large-scale cotton bleaching was carried out in this way. Wool bleaching with
H2O2, which by comparison with the cotton process gave much superior results, was widespread and predominant by 1880263i.
All bleaching solutions were alkaline due to added sodium hydroxide or ammonia.
Bleaching action is far greater at high pH, presumably due to the formation of the HO2- ion, but decomposition is accelerated as was discussed earlier. A compromise pH of ca.
10 was usually used, the exact pH depending on the nature of bleaching required.
H2O2 as a bleach for wood and paper pulp has only attained widespread use in the last thirty years. As environmental considerations have become increasingly important, and the benign breakdown products of H2O2 - water and oxygen - have been seen as increasingly attractive in comparison to the environmentally polluting byproducts of
established chlorine-based bleaches, so consumer acceptance of decreased brightness
and whiteness in bleached products has grown.
The use of H2O2 in chemical synthesis is a comparatively recent development, dating
from the 1940,s13>32,33) although its value in chemical analysis has been recognised
for over one hundred years34. The major use of H2O2 in chemical synthesis in
organic chemistry ȣ> in the metal catalysed oxidation of alkenes35.
2 7 The separation and purification of metal ions by H2O2 in metallurgy is another recent application. The most important use is in the extraction of uranium from its ore via ferric ions which oxidise insoluble U4+ to U6+;
2Fe2+ + H2S 0 5 + 2H+ ------► 2Fe3+ + H2S 0 4 + H20 ....0.12
U02 + 2Fe3+ ------U022+ + 2Fe2+. ....0.13
Uranium may then be purified;
U 0 2 + H20 2 + x H20 ------► U 0 4.xH20 + 2H+...... 0.14
Hydrogen peroxide also found use between 1930 and 1955 as a possible mono propellant in military applications, most notably in the launching mechanism of the
German V-l flying bomb used in the later stages of WWII. However, the explosive nature of H2O2 has to a great extent curtailed its use in this direction as anything other
than a minor fuel to give boost at the launching of rockets or jets.
Initially widespread use of H2O2 in the field of medicine has more recently concentrated
on its role as a disinfectant and antiseptic, particularly in the disinfection of surgical
instruments.
0.1.5 Occurrence in the Environment.
A great deal of research has been pursued into both the occurrence and activity of H2O2
in the environment. Recently, this work has taken on a new significance as the possible
role of H2O2 in the the production of "acid rain" has come under scrutiny.
Almost all biological organisms obtain energy by oxidation of organic substances,
yielding CO2 and water. Participation of molecular oxygen in the oxidation facilitates
the possible formation of H2O2. Indeed, hydrogen peroxide has been detected in a
number of enzymatic and biological systems. Such detection is only possible in the
absence of heavy metals and other peroxide-decomposing species such as the enzymes
peroxidase and catalase. It is for this reason that H2O2 is not found in aerobic cells
28 (which contain catalase) but has been detected in anaerobic bacteria such as pneumococci and streptococci upon exposure to atmospheric oxygen. It is thought that the destruction of these cells upon contact with air may proceed via formation of H2O2, which is poison to their metabolisms.
Some biological processes function only in the presence of H2O2. The white rot fungi
Phanerochaete chrysosporium requires H2O2 in order to degrade lignin in wood36, apparently via a single mechanism of oxidation involving H2O2 and ligninase, a haem enzyme.
A number of enzymes are known which catalyse the oxidation of a variety of substrates by dioxygen, leading to the production of H2O2. Examples of these include glucose oxidase and oxalate oxidase in equations 0.15 and 0.16 respectively;
Glucose + 2H2O + O2 ------► Gluconic acid + H2O2 ....0.15
Oxalate + O2 ------2CO^+ H2O2. ....0.16
It has been shown that H2O2 production is also widespread among blue-green algae isolated from both marine and freshwater sources. Of the three types of algae characterised, one, an example of which is Nostoc mucsomm, is found to produce
H2O2 in darkness, the rate of production increasing upon illumination and falling back to the original value once illumination is ceased37.
The Bombardier beetle is an amusing example of H2O2 production in biological systems. When provoked, the beetle ejects a jet of steam and obnoxious benzoquinone from a pair of glands at the top of the abdomen. This emission is accompanied by an audible detonation as H2O2 (25%, 8.5M) and hydroquinone are forced into a vestibule in the beetle's body. This contains both catalase and peroxidase; the latter facilitates the oxidation of hydroquinone to 1,4-benzoquinone while the former decomposes the
H2O2 to produce steam38.
Remarkably high concentrations of H2O2 - up to 300ppb -are found in polar ice and
2 9 snow39. Such is the disparity between concentrations of H2O2 in summer and winter snow layers (high and low concentrations respectively) that they can offer a useful method for the dating of polar ice by counting annual layers. Similarly polar precipitation archives can be used to trace back H2O2 levels in the atmosphere for high latitudes39. It is remarkable that supposedly unstable and highly reactive hydrogen peroxide can withstand not only the fimification process, but can also still be detectable in ice several thousand years old. This has been explained partly by the suppressed chemical activity of H2O2 in the very low temperatures of the polar snow, and also by the ease with which H2O2 is built into the ice matrix. However, this suppressed chemical activity does not totally preclude reaction40.
The major gaseous sources of H2O2 in the troposphere are believed to involve reactions of HO and hydrated HO2 radicals. Reaction of hydroxyl radicals with carbon monoxide
(eqn. 0.17) leads to the formation of HO2 via the recombination of H atoms with O2
(eqn. 0.18)41.
OH + CO ------H + C02 ....0.17
H + 02 + M ------H 0 2 + M ....0.18 where M is some other molecule which can remove the energy of the exothermic process.
Ki HO2 + HO2 ------^ H2O2 + O2 ....0.19
HO2 + H2O «». H 0 2.H20 ....0.20
k 3 H2O.HO2 + H 0 2 ------► H2O2 + O2 + H2O. ....0.21
The rate constant ratio K3/K1 is sufficiently high for equation 0.21 to play a significant
role in H2O2 production even though only a few% of total HO2 concentration is present as H2O.HO2. Increased H2O2 levels during periods of high humidity4 M2
3 0 result from equations 0.20 and 0.21.
Major sources of aqueous (and hence in-cloud) formation of H2O2 are thought to result from scavenging of OH and HO2 radicals, and subsequent HO2 radical disproportionation, according to the pH-dependent reactions41*43;
Ki h o 2 + h o 2 ------► H20 2 + O2 Ki = 7.5x105 mol-1 dm3 s-1 ...0.22 and
H 0 2 = 5 = ^ H+ + Oz pKa = 4.9 ....0.23
h 2o 02 + 02 ------H 0 2 + Oz + HO- K = 1x102 moi-l dm3 s-i ....0.24 and
H02 + H+ ^ —'Js H2o2 ...0.25
H 0 2 + O2 ------H O j + 0 2 K = 8.5x102 mol-1 dm3 s- l ....0.26
The rate constant for the reaction represented by eqn. 0.24 is much lower than those for the other reactions due to the repulsion effect between the two negatively charged reactants. Equation 0.26 predominates at pH 4.9, equation 0.22 is important at lower pH, and under alkali conditions equation 0.24 prevails. However, calculations suggest that gas phase peroxide formation is favoured over production in aqueous solution41.
Thus aqueous phase concentrations are increased by incorporation of gas phase H2O2 into droplets42.
These are viewed as the dominant processes in the production of H2O2 in the
atmosphere. However, many other processes are thought to contribute, ranging from photolysis of formaldehyde44 to the reaction of ozone with terpenes45.
Eqns. 0.22,0.23 and 0.26 represent very important reactions. Although theoretically, the reduction of O2 to H2O2 is a two-electron process, a one electron reduction is
sufficient to yield HO2, which then reacts with a similar species to produce H2O2.
31 Production and occurrence of hydrogen peroxide in natural waters has been the subject of much research over the last twenty five years. Many pathways to the production of
H2O2 have been postulated46*47, but the processes now accepted are in the main photochemical, with some augmentation from peroxide-containing natural rainwater.
The extent of this rainwater contribution varies enormously according to site and conditions. A graphic display of the photochemical nature of H2O2 production is provided by the work of Cooper and Lean48. Fig.0.7 overleaf shows the diel changes in H2O2 concentration compared to incidence of photosynthetically active radiation over a five day period at Jacks River, Ontario in 1988.
32 Tkm of Day (hour*)
Fig. 0.7 Diel changes in hydrogen peroxide concentration/nM (top) compared to photosynthetically active radiation/jxE m-2s-i (bottom), May 10 to May 14 1988, Jacks
Lake, Ontario, Canada48.
33 I. PRODUCTION OF HYDROGEN PEROXIDE FROM
HYDROGEN SULPHIDE AND DIOXYGEN.
1.0 INTRODUCTION.
Hydrogen Sulphide.
1.0.1 Sources.
Hydrogen sulphide is found in abundant quantities in natural gas deposits. Particularly
large, industrially important reserves occur in the natural gas fields of Canada, the USA
and France49. H2S concentrations as high as 42% have been reported in gas from
Wyoming, USA, where reserves are estimated to be ca. 5.9xl010kg. The sulphur
content of petroleum deposits in the USA varies from ca. 0.04-5%, consisting entirely
of H2S and organo-sulphur compounds. The action of steam on inorganic sulphides at
high temperatures is thought to be responsible not only for the production of H2S from
volcanic gases, but also in geothermal 'wells'. In contrast, bacteriological
decomposition of mineral sulphates is thought to be key in the formation of H2S in
natural springs and wells. It is also formed as a primary bacteriological decomposition product of protein matter, particularly of animal origin50-51.
There is currently a great deal of interest in the bacteria, found in hydrothermal vents on
the deep ocean floor, which assimilate hydrogen sulphide by chemosynthesis (ie. by oxidation ofH 2S to elemental sulphur)52. These bacteria, which derive all the energy they require for metabolism and growth from this simple chemical reaction, are being put forward not only as a possible means for the control of 'acid rain' and other forms of sulphur pollution, but also as a source of synthetic fuel.
H2S is a byproduct of a number of industrial processes. Very substantial quantities are released in coking plants and in the production of gases from coal51. Several steps in the refining of petroleum products require the removal and recovery of H2S and other sulphur compounds. Processes for the conversion of thiols, sulphides, thiophenes and
34 benzothiophenes, all of which form the main sulphur-containing constituents of crude oil, are based on their hydrodesulphurisation to form H2S, which is then absorbed from the gas stream and subsequently regenerated.
Other sources of H2S include the Kraft method of wood-pulping in the production of paper, where ca. 0.5-20kg of H2S can be produced per 1000kg of dry pulp5*, and in the production of viscose rayon (ca. 6-9 tonnes per 100 tonnes of rayon).
1.0.2 Sulphides as Reducing Agents.
Ammonium hydrogen sulphide has been used for many years as a reducing agent in synthetic organic chemistry. More recently sodium sulphide has gained use in various reactions where selective reduction is required. For example, in the reduction of 2,4- dinitro phenol, the reagent selectively attacks the nitro grouportho to the hydroxy group to yield 2-amino-4-nitrophenol53.
The general term for the reduction of nitro groups to amino groups by negative divalent sulphur is the Zinin reduction54, named after the first recorded use of the method, by
Zinin in 184255, in the preparation of aniline from nitrobenzene. With the advent of catalytic reduction processes, the method has fallen into disuse, but can generally be employed in the mild reduction of more sensitive compounds which are not compatible with acid media or which would be reduced further than desired by the established catalytic procedures.
NaHS in alcoholic solution has been used as the favoured reagent for the reduction of compounds of the general formula x,y'-dinitrodiphenyl, to give the product of formula x-amino-y'-nitrodiphenyl. For example, 4,4'-dinitrodiphenyl is reduced in 79% yield to 4-amino-4'-nitrodiphenyl56.
The stoichiometry of the reduction in general is illustrated by Zinin's original reaction;
4C6H5N 0 2 + 6S2- + 7H20 ------4C6H5NH2 + 3S20 32- + 60H- ....1.1
The main reductant species appears to be the disulphide, S22-, and it is worthy of note
35 that thiosulphate is produced by the reduction. This is in contrast to the reductions by
H2S herein, where the reductant is oxidised to elemental sulphur.
A few kinds of compound other than nitroarenes have been reduced under the conditions of the Zinin reduction. These include reduction of certain azocompounds to the hydrazo species57 (in some cases cleavage of the N-N bond is observed5k59).
1.0.3 Thermodynamics of Proposed System.
The proposed method by which hydrogen peroxide will be produced involves the use of hydrogen sulphide as a reducing agent, analogous to the catalytic hydrogenation stage of the alkylanthraquinol process for the industrial preparation of H2O2 which currently prevails. Aside from quinones, proposed substrates include azocompounds and flavins. It is to be noted that all three proposed substrates contain conjugated systems.
The most obvious consideration for the proposed method is the viability of H2S as a reducing agent. Thermodynamic data60-61 for the existing and the proposed process, equations 1.2 and 1.3 respectively, are compared below;
H2(g) + 0 2(g) ----- H2 0 2 (i) .... 1.2
AH°298 = -187.9 kJmoH
AG°298 = -120.4 kJ mol-i
H2S(g) + 0 2(g) ------H2Q2(i) + S(s) ..... 1.3
AH°298 =-167.8 kJmoM
A(j °298 = - 88.6 kJ mol-1
Consideration of the above data leads us to the conclusion that although the reduction with hydrogen is thermodynamically more favoured, reduction with hydrogen sulphide is certainly feasible in free energy terms.
36 1.0.4 Proposed Reaction Process.
In an method analogous to the catalytic reduction of alkylanthraquinone in the industrial manufacture of H2O2 described in the general introduction (p24-), it is proposed to use
H2S as a reducing agent. However, since H2S and H2O2 react together under a variety of conditions1, a two stage reaction process is required;
H2S + R ------► RH2 + S .....1.4
RH2 + 0 2 ------► R + H2O2 .....1.5
Thus after reduction of substrate R by H2S, the unreacted gas will be evacuated (and any precipitated sulphur will be removed by filtration if possible) before introduction of oxygen.
Initially, the most likely candidates for substrate R appear to be quinones (cf. industrial manufacture), azocompounds and flavins.
1.0.5 Redox Potentials of Substrates.
For the proposed method to operate successfully, not only must reduction of the starting substrate be feasible, but also once the reduced compound is formed, it must be capable of being oxidised back to the original material. The most convenient way to assess this possibility is to study reduction-oxidation potentials for the proposed substrates. Redox potentials for the quinones and azocompounds are widely available62-
65. In each case, it can be seen that the presence of electron-donating groups on the molecule serve to lower the redox potential, ie. to stabilise the oxidised form of the substrate. Conversely, electron-withdrawing substituents lead to an increase in redox potential and stabilise the reduced form.
These effects are illustrated in tables 1.1, 1.2 and 1.3 below, for anthraquinones62, naphthoquinones63 and azocompounds164,65 respectively.
37 Table 1.1 The Effect of Substituents on the Redox Potentials of Anthraquinones.
9,10-Anthraquinone 0.155
2-Methyl-9,10 anthraquinone 0.150 l-Chloro-9,10-anthraquinone 0.174 l-Hydroxy-9,10-anthraquinone 0.132 a Potentials recorded in ethanol-water at 25°C.
Table 1.2 Effect of Substituents on the Redox Potentials of Naphthoquinones.
Naphthoquinone e 025/ v
1,4-Naphthoquinone 0.484
2-Methyl-1,4-naphthoquinone 0.408
2,3-Dimethyl-1,4-naphthoquinone 0.340
^Potentials recorded in dUnol-water at 25°C.
Table 1.3 Effect of Substituents on the Redox Potentials of Azocompounds.
Azocompound E 025/V T/°C
Azobenzene-4,4'-disulphonate64 0.424 25
3,3'-Diamino-4,4'-dimethylazobenzene65 0.367 25
3,3'-Diamino-4,4'-dimethylazobenzene64 0.374 18
3,3'-Diaminoazobenzene65 0.414 18
The data in tables 1.1, 1.2 and 1.3 indicate that should a suitable substrate for the reversible reaction be found, there might be scope for 'fine-tuning' of its propensity for reduction or re-oxidation by slightly altering the substituents on the molecule.
38 1.0.6 Properties.
Hydrogen sulphide has, at latm., a freezing point of -85.5°C and a boiling point of
-60.3°C66. In connection with the present work, the most important of its physical
properties is its solubility in the organic solvents used. Although this data is not
available for the solvents actually employed, literature results67 are given for related
solvents in table 1.4 below.
It can be seen that H2S has appreciable solubility in a variety of solvents (a mole
fraction of 0.0358 in benzene equates to a solution of ca. 0.4 mol dm*3, while a value
of 0.018 in ethanol represents a solution of ca. 0.3 mol dm-3), thus considerable
concentrations of H2S will be available for reaction in solution. Equally importantly
from both a practical and an industrial viewpoint, due to the large volume of gas in
solution, effective evacuation of the unreacted H2S will be necessary.
Table 1.4 Solubility data for H2S in organic solvents.
Solvent Mole Fraction H 2S in liquid3 TIK
Benzene 0.0358 298.15
Ethylbenzene 0.0420 298.15
Propylbenzene 0.0520 298.15
Toluene 0.0660 293.15
o-Xylene 0.0700 293.15
Ethyl Acetate 0.0866 298.15
Hexane 0.0372 298.15
Ethanol 0.0180 293.15
1-Butanol 0.0315 298.15
1-Methyl-2-pyrrolidinone 0.1800 298.15
aMole fraction recorded at a partial pressure of H2S of 101.325kPa.
pKai for H2S at 25°C is 6.8368, while the most recent studies state that
39 pKa2 =c*n-o 69,70.
1.0.7 Literature survey.
There exist a number of industrial patents that describe production of hydrogen peroxide from hydrogen sulphide. All depend upon the reduction of a substrate by
H2S and subsequent oxidation of the reduced form of the substrate with O2 to yield
H2O2. It is-fo Joa rvtedrflcrf 'ftfl- perfect iw/ouuj ^ o f U iJoes l ngt use . HZS i. US 4,592,905 ; Plummer P A, Beazley P M. (Marathon Oil Co.).
This patent outlines the amine-catalysed reduction of 2-ethyl- and 2-tbutylanthraquinone by H2S. This is effected in a wide variety of organic solvents, and it is concluded that solvents of high polarity such as 1-methylpyrolidinone produced the highest yields of
H2O2 upon re-oxidation with O2. ii. US 3,923,966 ; Vaughan L G. (Du Pont). J97S
Vaughan outlines a similar process to that of Plummer and Beazley, except that a wider variety of both substrates and amine catalysts are considered. Aside from alkyl anthraquinones, alkyl tetrahydroanthraquinones are also discussed. Of the amines put forward as catalyst, tertiary species are favoured due to their enhanced stability to attack by O2 and H2O2 relative to secondary and primary amines. Preferred organic solvents are mixtures of diisobutylcarbinol, alkylbenzene derivatives and methyl or dimethyl naphthalene. Over a range of temperatures (10-200°C), pressures (l-10atm.) and reduction periods, yields of between 50 and 70% based on substrate are claimed. iii. US 3,311,453 Lusby G L. (Canadian Industries). \% 1 '
Water-soluble sulphonated anthraquinones are reduced by H2S in borate-buffered
(pH 9) aqueous solution. No additional catalyst is necessary. Aeration of the reduced substrate is effected in the presence of calcium hydroxide. Thus the H2O2 is regenerated afterwards from the calcium peroxide formed in the initial oxidation. This overcomes problems of separation of H2O2 from the reaction solution. Yields of the
40 order of 30% are claimed, according to weight of calcium peroxide formed, iv. Bel. 652,956 ; Laporte Chemicals. I%3
A process is reported for the catalytic hydrogenation of a variety of 2,3-dialkyl substituted naphthoquinones, yielding H2C>2upon oxidation with O2. The dialkyl naphthoquinones considered are;
Rx = Alkyl Cj - C6.
R2 = Alkyl Cx - C6.
o Fig. 1.1 Dialkyl naphthoquinones described in Patent Bel. 652,956.
Mixed alkyl substituted species are preferred on solubility grounds. It is claimed that degradation of the naphthoquinones by O2 is less marked, and solubility much better, than with other quinones. Yields of H2O2 of up to 100% according to substrate are claimed. v. Pol. 106,154 ; Bogdal S, Kuczynski H, Hirszowski J and Sciazko A.
The patent describes (in Polish) the reduction of azobenzene by H2S in the presence of a strong organic base at 60-65°C. High yields of hydrazobenzene are claimed.
41 1.1. RESULTS AND DISCUSSION.
1.1.1 Production of H 2O 2 by Quinone Compounds.
Table 1.5 shows the results for initial attempts to produce H2O2 by reduction with H2S and subsequent reoxidation with O2 of a number of different quinone compounds by the standard reaction procedure (p.83). Reduction period was two hours.
Of those reactions where H2O2 was produced, only reduction and subsequent oxidation of 2-ethylanthraquinone (EQ) in 1-methylpyrolidinone (NMP) produced a satisfactory yield. This method, following the patent work of Plummer and Beazley, had the disadvantage that NMP is miscible with water. Hence upon addition of water, centrifugation of the substrate and precipitated sulphur were required, and also extraction of the H2O2 is difficult to envisage in an industrial context. Ideally the H2O2 should be formed in a organic layer immiscible with water and be subsequently extracted into an aqueous layer.
Other points raised by the results are that an increase in reaction temperature from 25°C to 60°C had little effect in increasing the yield of H2O2, and that even employing all reasonable methods to avoid contact between the H2O2 and dissolved sulphur in the organic layer, yields for EQ were particularly disappointing.
As would be expected from the redox potential discussion put forward in the introduction, 2-methyl-1,4-naphthoquinone (MN) exhibited a greatly increased ease of reduction (indicated by a rapid darkening of colour on introduction of the H2S), but oxidation did not occur.
42 Table 1.5 Summary of Reactions of Quinones by Standard Procedure (p.83).
S u b s .a mmol Solventb T/°C Cat.c Exp. Notesd Yield/% e
a. EQ 5.2 p-Xyl/DMH 25 6.3 — 0.62 b. EQ 5.2 p-Xyl/DMH 60 6.3 Ar 0.88 c. MN 2.6 p-Xyl/DMH 25 6.3 Ar ....f
d. EQ 5.2 NMP 60 — Ar — e.QDS 3.0 H2Og 25 — Ar 1.4
f.QDS 1.5 H2Og 25 — Ar 2.1
g.E Q 0.52 NMP 60 — Ar,CS2 41.8
h. EQ 1.25 DMF 25 — Ar,CCl4 11.0
i. EQ 1.25 DMF 60 — Ar,CCl4 13.3 j- QS 0.5 H2Og 25 — Ar,C02,CCl4 10.1
k. QS 2.5 H2Og 25 — Ar,C02,CCl4 18.2
1. EQ 2.5 p-Xyl/DMH 25 4.2 A r,C02 1.1 m. EQ 2.5 p-Xyl/DMH 60 4.2 A r,C02 8.9 n. EQ 2.5 TBP 60 4.2 A r,C02 14.7 o. EQ 2.5 p-Xyl/DMH 60 4.2 A r,C02 12.5
^Substrate - see materials and syntheses section. Abbreviations; DMF - N,N-dimethyl formamide; DMH - 2,6-dimethylheptanol; EQ - 2-ethyl anthraquinone; NMP - 1- methylpyrolidinone; p-xyl; p-xyleneQDS - sodium anthraquinone-2,6-disulphonate; QS - sodium anthraquinone-2-sulphonate; TBP - tii-n-butyl phosphate. bSolvent - 25ml of reaction solvent. cmmol of tri-n-butylamine added (4.2mmol catalyst = 1ml). dSee experimental procedures section. eYield based on substrate. fReduced substrate does not reoxidise. gBorate buffer, pH 9.0.
43 1.1.2 Production of H 2O 2 by 2-Alkyltetrahydroanthraquinones.
In view of the poor yields obtained with the anthraquinone compounds detailed in table
1.5, and after consultation with representatives of Interox Ltd., 2-alkyltetrahydroanthra quinones were obtained and used as substrates in the standard reaction procedure
(p.83). Those compounds obtained are shown in fig 1.2;
Fig. 1.2 2-Alkyltetrahydroanthraquinone.
The patent work of Vaughan indicated that these substrates, which can be effectively regarded as 2,3-dialky 1-1,4-naphthoquinones, are readily reduced by H2S in the presence of an amine, and hence could be of great use in the desired reaction.
Table 1.6 overleaf summarises the results of initial experiments with the tetrahydro anthraquinones in the standard reaction procedure (p.83). It can be seen that the initial results seemed very promising. 2-tButyltetrahydroanthraquinone (BTQ) was only partially soluble in the reaction solvent, dissolving upon reduction and coming out of solution upon oxidation. The fact that the unreduced form came back out of solution upon reoxidation was undesirable from an industrial viewpoint, where it would then be mixed with sulphur precipitated in later runs on the same reaction solution. It was also apparent that extending the reduction period from 2 to 48 hours had little appreciable effect on the yield of H2O2.
For 2-ethyltetrahydroanthraquinone (ETQ), the situation was reversed in that the reduced form was partially insoluble in the reaction solvent. Unlike BTQ, this appeared to have an adverse effect on the yield of H2O2. A 50% reduction in the concentration of
44 substrate produced a much higher yield.
Both the reduced and oxidised forms of 2-amyltetrahydroanthraquinone (ATQ) were soluble in the reaction solvent in the concentration studied, producing a high yield of h 2o 2.
For solutions of all three substrates, the reduction produced a colour change from bright yellow through dark amber-brown to a very pale yellow.
Table 1.6 Summary of Reactions of Alkyltetrahydroanthraquinones by Standard
Procedurea.
Substrate mmol YieId/% Experimental Notes a. BTQ 4.8 73.3 Oxidised form partially insoluble. b. BTQ 4.8 75.9 Reduction period 48 hours. c. ETQ 5.2 42.1 Reduced form partially insoluble d. ETQ 2.6 80.5 e. ATQ 4.8 85.8 aReaction solvent 25ml 60:40 Shellsol/diisobutycarbinol, 0.091ml (0.3mmol) tri-n- butylamine. Reduction period two hours at room temperature. Yields based on substrate.
1.1.3 Production of More Concentrated Solutions of H20 2.
In the previous experiments, sulphur had not precipitated out of the reaction solution upon reduction of the substrate. It was presumed that this was because the solvent had not yet become saturated with sulphur, and should a series of reduction/ oxidation runs be performed on one quinone/solvent solution by the standard reaction procedure
(p.83), this sulphur would begin to come out of solution. This was of interest for a number of reasons. Firstly, once the solution was saturated with sulphur, its precipitation would be theoretically stoichiometric with the reduction of the substrate.
Thus filtration and weighing of the precipitated sulphur prior to oxidation made it possible to gauge the extent of reduction. Comparison with the yield of H20 2 enabled
45 an estimate of the loss of peroxide by reaction with the dissolved sulphur to be made.
Similarly, by varying the amount of water added prior to oxidation, more concentrated solutions of H2O2 could be produced, and their reaction with the dissolved sulphur gauged. Table 1.7 shows the results for a series of reduction/oxidation runs on ATQ.
Table 1.7 Summary of a Series of Reduction/Oxidation Runs on a Solution of ATQa.
Run Y ield A T Q b Y ie ld s c H 20 /m Id [H20 2]/M
1 27.9 — 25 0.053
2 63.5 — 50 0.061
3 69.9 — 10 0.333
4 53.6 79.3 1 0 0.256
5 55.8 86.5 1 0 0.266
6 69.2 87.7 15 0.220 a1.35g, (4.77mmol) ATQ, reaction solvent 25ml 60:40 Shellsol/ diisobutylcarbinol,
0.091ml (0.3mmol) tri-n-butylamine. Reduction period two hours at room temperature.bYield based on ATQ. cYield based on precipitated sulphur. dVolume of water added prior to reoxidation with O2.
A number of points were apparent from the above results. Sulphur was precipitated out during run 3. The fact that the first disappointing yield has no effect upon later yields would seem to indicate that full reoxidation occurs. In this and later work a wide variance was found in the yields of H2O2 for runs under exactly the same conditions.
Since the iodometric titration has an accuracy in the region of ca. 1%, this was attributed both to adventitious losses of H2O2 by catalysed decomposition and practical losses on transfer and separation of reaction solutions.
Any possible reaction between the dissolved sulphur and the hydrogen peroxide solution does not prevent the production of quite strong solutions of H2O2.
46 1.1.4 Measurement of Loss of H 2O 2 by Reaction with Sulphur.
Analysis of the discrepancies between the yields calculated according to the sulphur and
those according to the substrate indicated that some peroxide was being lost, probably
due to reaction with sulphur;
3H20 2 + S — H2S04 + 2H20 .....1.6
In order to gauge the extent of this reaction, the reduction of ATO was repeated firstly
by the standard reaction procedure, whereby water was added to the reduced solution
prior to oxidation, and secondly by amending this process and adding the water after
oxidation. Thus in the latter case the H20 2 would be formed in the organic sulphur-
containing layer and not immediately extracted into the aqueous layer. It was of interest
to see if this longer period in contact with the potentially highly reactive dissolved
sulphur would have any great effect upon the yield of H20 2. Results are shown below.
Addition of water prior to oxidation ; Yield = 86.5%
Addition of water after oxidation ; Yield = 76.4%
Thus there appeared to be an appreciable effect on hydrogen peroxide yield when the
H20 2 was allowed to remain in contact with the dissolved sulphur.
Gravimetric analysis of the peroxide-containing aqueous solution as outlined by
Vogel71 was another method by which the loss of H20 2 by reaction with sulphur could
be measured. Comparison with the ideal yield of H20 2 (100%) is made below.
Addition of water prior to oxidation; Loss = 4.5%
Addition of water after oxidation; Loss = 4.8%
Although this seemed to indicate that, in terms of reaction between sulphur and hydrogen peroxide, there was little difference between the two methods, a loss of ca.
5% of H20 2 on each mn was not inappreciable.
1.1.5 Effect of Amine Catalyst on the Reaction.
Following the patent work of Vaughan tri-n-butylamine had been the sole catalyst
47 employed in the reductions, but it is indicated in the patent that a variety of sterically hindered bases catalyse the reaction satisfactorily. Tributylamine is relatively involatile; this was considered an important property if an appreciable amount of catalyst was not to be lost upon evacuation of H2S in a series of runs. However, its toxic and corrosive nature, coupled with its unpleasant odour, led to the study of other longer chain amines, such as trioctylamine, (bpt. = 365-367°C compared to 216°C for tributylamine), which would be less unpleasant to deal with in an industrial context.
Aside from these essentially practical considerations, it was of interest to study the effect on yield of H2O2 and rate of reduction (qualitatively) of employing catalysts with a wide range of pK values. Table 1.8 overleaf shows the yields of H2O2 and experimental observations for a number of catalysts when employed in the standard reaction procedure.
The results indicate that the optimum pK range for the effective catalysis of the reaction is ca. 7.8-12.3, with a dramatic loss of efficiency occurring between pK 7.8-7.4. This region is close to pKal for H2S,which is 6.8368. In view of the different solvation and other effects to be expected in the solvent mixture used compared to those found in aqueous solution, this would appear to have little significance. The upper value of the optimum pK range was taken as that for l,8-bis(dimethyl) naphthalene ("proton sponge") where the reaction proceeds very quickly to the normal highly coloured intermediate stage in the reduction before more slowly reaching completion. Although a satisfactory yield was obtained for 1,1,3,3-tetramethylguanidine, the reaction route in this case was unusual in that a red-brown oil separated out from the reaction mixture upon reduction, and an unusual colour change was observed. The fact that the oil redissolved upon re-oxidation led us to believe that it might be the amine-HS salt coming out of solution, but that enough of the catalyst remained in solution to facilitate the reaction.
Steric effects do not seem to be important. 1-Methylimidazole and 4-lbutylpyridine, neither of which are sterically hindered, fail to catalyse the reduction effectively. The same is true for the sterically hindered bases 2,4,6-collidine and 2,6 lutidine. Yet
48 1,2,2,6,6,-pentamethyl piperidine, with strong steric hindrance, is an excellent catalyst for the reduction, indicating that pK is the most important factor in determining the effectiveness of the catalyst.
Table 1.8 Values of pK for a range of Bases and Yields of H2O2 obtained when used in the Standard Reaction Procedure*.
Catalyst pKb Yield/% Experimental Notes
1,1,3,3-Tetramethyl 13.6072 76.1 0*usual Reaction Route . guanidine
1,8-Bis(dimethyl) 12.3473 62.5 Reaction "held” halfway aminonaphthalene
1,2,2,6,6-Penta 11.2574 79.6 Very rapid reduction methylpiperidine
Tri-n-octylaminec • 98.4 11.0473 Tri-n-butylaminec 80.5
4-Dimethylamino 9.7073 90.8 pyridine
N,N'-Dimethyl 9.2575 88.6 ethanolamine
2-Methylimidazole 7.8673 60.9
Triethanolamine 7.7673 67.6 Proceeds with turbidity
2,4,6-Collidine 7.4373 5.6 Very slow reduction
1 -Methylimidazole 7.0673 1.3 Very slow reduction
2,6-Lutidine 6.7275 — No reaction
4-tButylpyridine 5.9973 — No reaction a0.38mmol catalyst, 0.625g (2.6mmol) ETQ, reaction solvent 25ml 60:40 p-xylene/ diisobutylcarbinol, room temperature. Yields according to substrate.bpK at 25°C in water. cSince the pKa value for neither trioctylamine nor tributylamine was available, the quoted figure is for tributylamine at 18°C.
49 1.1.6 Catalysis of the Reduction by a Secondary Amine.
A wide variety of tertiary amines had been shown to successfully catalyse the reduction
stage of the reaction. The tertiary compounds were originally chosen to avoid attack by
H2O2. Although secondary amines are more labile to H2O2 attack than the tertiary analogues, pK values for the two are broadly similar; for example diethylamine pK = 10.98, triethylamine pK = 10.7575. Use of di-n-octylamine (0.115ml,
O.38mmol) to catalyse the reduction of ATQ (0.729g, 2.6mmol) by the standard reaction procedure produced H2O2 in high yield (88.9%) upon reoxidation with O2.
Thus the reaction can be catalysed equally well by tertiary or secondary amines, with little apparent increase in loss of H2O2 by reaction with the secondary amine as compared to the tertiary species.
1.1.7 Continuous Run on One Reaction Solution.
A continuous cycle of reductions and oxidations allowed a number of important measurements to be made. In the main these involved possible losses of materials to the extraction process and by side-reactions. The reaction on a laboratory scale would offer little indication of side-reactions that might be important if the process were operated at pilot-plant scale. However, loss of catalyst by extraction into the aqueous layer and loss of H2O2 by oxidation of sulphur could be gauged by extraction of combined aqueous layers into a volatile organic solvent and gravimetric analysis respectively. Table 1.3 overleaf shows both the yields of H2O2 and analyses of losses for a continuous cycle of reduction/oxidations on ETQ by the standard reaction procedure. Precipitated sulphur was filtered off in each case prior to oxidation of the reduced substrate.
From the results shown in table 1.9 overleaf, it can be seen that loss of catalyst by extraction into the aqueous layer was negligible. Measurement of the loss of solvent into the aqueous layer gave similar results.
50 Table 1.9 Yields of H2O2 and Analysis of Losses for a Continuous Run of
Reduction/Oxidation Reactions on ATQ by the Standard Procedures
R un Yield/ % Analysis of Losses
1 87.7 8xl0-6mol SO4- = 2.4xl0-5mol H20 2 lost =0.92%
2 86.7 3.1xl0-5mol SO4- = 9.4xl0-5mol H20 2 lost=3.62%
3 77.3 3.4x10-5mol SO4- = 1.0xl0*4mol H20 2 lost = 3.92%
4 75.6
5 84.4
6 76.7
7 84.4
8 82.9 Average 0.6% Tri-n-octylamine lost each run
9 77.3
10 71.3
11 69.6 Average 0.4% Tri-n-octylamine lost each run
12 70.6
13 67.3
a0.625g (2.6mmol) ETQ. Reaction solvent 25ml 60:40 p-xylene/ diisobutylcarbinol,
0.167ml (0.38mmol) tri-n-octylamine, reduction period two hours at room temperature.
Yields based on substrate.
1.1.8 UV/Visible Study of the Reduction of ETQ.
It was observed in all the reductions of the 2-alkyltetrahydro anthraquinone systems that considerable darkening in colour occurred upon introduction of H2S. Over the period of one hour, this lightened to eventually yield a pale yellow solution. This colour change seemed to afford some scope for study of the reaction by ultra-violet/visible
absorption spectrophotometry. In order to allow a suitable timescale for study of the change, the amount of amine catalyst was reduced, and due to the practical limitations of the cell, reactions were carried out at atmospheric pressure of H2S .
Figs. 1.3 and 1.4 show the spectra obtained for both reductions undertaken. It can be
51 0.0 into the cell, (t = 0, unbroken line). 1mm path length cell. length path 1mm line). unbroken 0, = (t cell, the into Fig. 1.3 UV/visible absorbance spectra of reduction of 2-ethyltetrahydroanthraquinone 2-ethyltetrahydroanthraquinone of reduction of spectra absorbance UV/visible 1.3 Fig. (0.1M) by hydrogen sulphide. Figures refer to minutes after initial introduction of H2S of introduction initial after minutes to refer Figures sulphide. hydrogen by (0.1M) 52 e m
ABS 0 .5 Fig. 1.4 UV/vis absorbance spectra of reduction of 2-ethyltetrahydroanthraquinone of reduction of spectra absorbance UV/vis 1.4 Fig. H2S to the cell (t = 0, unbroken line). 10mm path length cell. length path 10mm line). unbroken 0, = (t cell the to H2S (0.01M) by hydrogen sulphide. Figures refer to minutes after initial introduction of introduction initial after minutes to refer Figures sulphide. hydrogen by (0.01M) nm 53
ABS seen that (as had been expected from the observed changes in colour), the reaction appears to occur in two stages. In both cases, during the first ten minutes after introduction of the H2S, the shoulder on the absorption curve shifts markedly to longer wavelength as the dark intermediate forms, gradually disappearing as the reaction proceeds to completion and the solution lightens in colour.The other notable feature of the two spectra is their marked similarity.
The intermediate is probably either a dimolecular quinhydrone species, or some form of radical.
1.1.9 EPR Study of the Reduction of ETQ
It was noted in the study of the effect of pK of base catalyst on the rate of reduction of the 2-alkyltetrahydranthraquinones that catalysis by l,8-bis(dimethyl)aminonaphthalene led to the reaction remaining at the dark brown intermediate stage for over two hours. It was hoped that if the highly-coloured intermediate was a radical species, epr study of this reaction under these conditions would afford the best opportunity of confirming its presence.
No radical signal was detected .
Attempts to confirm by epr the presence of a radical intermediate in the reduction of
ETQ catalysed by trioctylamine were similarly unsuccessful. It was possible in both cases that a radical species was being formed, but that it was unstable and rapidly converted to the reaction product, and thus was present in a very low equilibrium concentration.
1.1.10 1H NMR Study of the Reduction of BTQ.
Figs. 1.5 and 1.6 show the 1H nmr spectra of the unreduced and fully reduced reaction solutions, recorded at 270MHz in 50:50 C6D6/CD3OD. Attempts were made to follow the reduction by nmr, particularly to observe any broadening of the aromatic resonances caused by the presence of a radical species as the intermediate stage state was
approached. Unfortunately the rapid rate of reduction of the small amount of BTQ in
54 o (1) *Bu
(3)
o
PPM i------r —i------1------1------1------r —T------, 8.1 8.0 7.9 7.8 7.7
Fig. 1.5 nmr spectrum showing aromatic resonances of 2-tbutyltetrahydroanthraquinone (inset).
55 Fig. 1.6 nmr spectrum showing aromatic resonances of 2-tbutyltetrahydroanthra
quinol (inset) - product of reduction of 2-tbutyltetrahydroanthraquinone with
hydrogen sulphide.
56 the nmr tube made this impossible.
In both figures, it can be seen that the resonances due to methyl protons on the lbutyl group and the methylene protons on the saturated ring of the tetrahydroanthraquinone are masked by the aliphatic protons of the tributylamine catalyst. For this reason, only aromatic protons are assigned in table 1.10 overleaf.
It can be seen that the effect of aromatisation of the central ring upon reduction of the
BTQ was to shift protons Hi and H4 downfield, whereas H3 was unexpectedly shifted upfield. The fact that fig. 1.6 shows no sign of the original aromatic BTQ resonances indicates that the reduction went to completion.
Table 1.10 !H Chemical Shifts for the Aromatic Protons of BTQ and its Reduction
Product.
5/ppm H i H 3 H4
BTQ 8.08 7.70 7.99
Reduced BTQa 8.35 7.51 8.31 aResonance at 5 7.27ppm is due to C6D5H. H2S is seen as a broad resonance at
4.85ppm.
1.1.11 EPR Study of the Reduction of QS.
Initial experiments on the reduction of sodium anthraquinone-2-sulphonate by H2S in water indicated that an epr signal could be found at g = 2.0046, fig. 1.7. If this was due to a semiquinone radical intermediate, it was hoped that the epr signal of the semiquinone radical would be seen to grow and recede with time as the anthraquinone was reduced gradually via the semiquinone to the dihydroanthraquinone. Fig. 1.8 shows a series of epr spectra recorded at 10 minute intervals after initial introduction of
H2S. It can be seen that the signal does not grow or lessen appreciably - within the noise limits - over the period of study. Spectra recorded for the unreduced anthraquinone and the empty epr tube discounted either as possible sources for the
57 Fig. 1.7 EPR spectrum of sodium anthraquinone-2-sulphonate
reduced by hydrogen sulphide.
58 0 mins.
Fig. 1.8 EPR spectra of the reduction of sodium anthraquinone-2-sulphonate by hydrogen sulphide, recorded at timed intervals after the introduction of H2S.
59 Fig. 1.9 EPR spectrum of sodium anthraquinone-2-sulphonate
reduced by sodium dithionite.
60 signal. Other possibilities included the presence of sulphur atoms prior to polymerization. In order to ascertain that the signal was that of the semiquinone the substrate was reduced with sodium dithionite and the epr spectrum recorded, fig. 1.9.
The signal obtained on reduction of the QS by dithionite was similarly found to occur at g = 2.0046, and thus could be confirmed as that of the semiquinone species.
1.1.12 Infra-red Study of the Interaction between H 2S and
Tributylamine.
The nature of the interaction between H2S and the amine catalyst was not known. In aqueous solution;
Tri-n-butylamine pK = 11.0473,
Hydrogen Sulphide pKi = 6.8368.
ie.,
i. H2S ----- H+ + us- .....1.7
[H+][HS- ] Ki 10-6.83 1 8 [H2S]
ii. Bu3NH+ ^ ------Bu3N + H+ .... 1.9
Ki : 10-H-O4 ....1.10 [Bu3NH+]
By comparison of values of Kj above, it can be seen that in aqueous solution, with an excess of H2S, the Bu3N will be almost completely protonated. However, due to a number of influences, particularly solvation effects, it was uncertain as to whether proton transfer would occur in the organic solvents employed.
The symmetric and asymmetric stretches of H2S in CCI4 occur at 2595 and 2610 cm-1 respectively76. Fig. 1.10 shows the IR spectmm obtained with tetrachloroethylene saturated with H2S. The symmetric stretch of H2S is seen at 2593.2 cm-1. The asymmetric stretch is probably too weak to be seen. Should proton transfer occur when
61 C l.830 102.235
XT XT
99.121 97.518
96.411 92.740
93.702 87.963
90.992 83.185
0B.2B3 78.408
85.573 73.630
82.864 68.853
80.154 64.075
77.445 59.299
74.735 54.520 3231 2000 3231 !900
-1 cm cm-1
Fig. 1.10 IR spectrum of H2S in Fig. 1.11 IR spectrum of H2S in
tetrachloroethylene. tetrachloroethylene plus 1%
tributylamine.
62 a base is added, it would be expected that the position of the peak would alter, and that appreciable broadening of the observed band would be seen due to the effects of hydrogen-bonding. From fig. 1.11, the IR spectrum of H2S in tetrachloroethylene plus
1% v/v tributylamine, it can be seen that this is not the case. The observed symmetric occurs at 2593.0 cm-i, a negligible shift from its original position. If it is assumed that similar interactions would occur in the solvent mixtures employed in the reductions, it can be concluded that there is negligible proton transfer from H2S to the base.
1.1.13 X-Ray Powder Diffraction and Raman Study of Sulphur.
The exact nature of the sulphur that precipitated out from the reaction solution upon reduction with H2S was primarily of interest for two reasons. Firstly, the practical
considerations of an industrial process demanded that it should be easily removed from the reaction mixture preferably by some form of filtration. It was thought that should the sulphur come out of solution in a form not easily filterable, this would be made much more difficult. From a theoretical viewpoint, the form of sulphur produced is relevant to the postulated reaction mechanism.
It was found that the sulphur precipitated out of the reaction solution in a form that was
easily filtered. The x-ray powder diffraction spectmm for a sample of sulphur (obtained
in this case from a series of reduction/oxidation runs on ETQ) was recorded, fig. 1.12.
Comparison of d values obtained for the powder sample and those quoted in the
literature77 prove conclusively that the micro-crystalline sulphur present is in the
orthorhombic (a) form of Ss. Where a slight discrepancy between d values occurs, this
is usually due to averaging out of two quoted values to produce one d value in the
obtained spectmm.
However, the spectrum did not confirm that the sample was composed of 100% of the
orthorhombic form, since some could be in the form of amorphous sulphur which
would not be detected by the powder diffraction technique.
A Raman spectmm for the same sulphur sample was recorded, fig. 1.13. Comparison
of this spectrum compared with that reported by Steudel78 proved conclusively that the
63 2.00 CuKai+2
< 10.380 x : 2 theia y ; 2381, Linear 60,880>
Fig. 1.12 X-ray powder diffraction spectrum of sulphur produced by reduction of
2-ethyltetrahydroanthraquinone by hydrogen sulphide.
Fig. 1.13 Raman spectrum of sulphur produced by reduction of 2-ethyltetrahydro anthraquinone by hydrogen sulphide.
64 sulphur was present entirely as orthorhombic Ss within the accuracy of the technique.
Thus any reaction mechanism postulated for the reduction had to explain the formation of Sg from H2S.
1.1.14 Production of H 2O 2 by 2,3-Dialkyl-l,4-naphthoquinones.
The use of 2,3-dialkyl-1,4-naphthoquinones as substrates in the standard reaction procedure was attempted with three main points borne in mind. Firstly, the tetrahydroanthraquinones used in earlier work could be regarded as analogous to these compounds. Secondly it was hoped that side reactions such as the epoxidation found in oxidation of the reduced form of the tetrahydro anthraquinones would be curtailed by the new substrate. Finally, initial attempts to use the monosubstituted 2-methyl-1,4- naphthoquinone as a substrate failed due to the stability of the reduced form to oxidation.
Reduction and subsequent oxidation of the dialkyl compounds would have been a further indication of the delicate balance of redox potential required for the process, and the two alkyl substituents would offer the opportunity of "fine-tuning" the relative rates of the two stages. Patent Bel. 652956 outlines use of a variety of 2,3-disubstituted -1,4- naphthoquinones as substrates for the conventional process using hydrogenation.
Slower degradation of the substrate by the oxidation process, and greater solubility
(particularly of those containing mixed alkyl substituents eg. 2-methyl-3-ethyl-1,4- naphthoquinone) are cited as advantages of these substrates in the process.
Table 1.11 shows the results obtained for the 2,3-disubstituted naphthoquinones in the standard reaction procedure.
65 Table 1.11 Summary of Reactions of 2,3-Disubstituted- 1,4-Naphthoquinones by
Standard Procedure (p83) a.
Substrate mmol Yield H20 2/%
DMN 1.3 96.3
DMN 2.6 97.4
EMN 2.6 94.3
Vit. Ki 2.2 94.8 aReaction solvent 25ml 50:50 p-xylene/O-methylcyclohexylacetate (sextate), 0.166ml
(0.38mmol) trioctylamine. Reduction period two hours at room temperature. Yields based on substrate.
Very high yields of H2O2 were found for all the 2,3-disubstituted -1,4- naphthoquinones. Due to its symmetry, DMN exhibited low solubility in the reaction solvent, coming out of solution upon reduction. This appeared to have no effect on the yield of H2O2 compared to a reaction solution of half concentration, where no reduced product precipitated. No problems of solubility were exhibited by the mixed alkyl
substituted EMN. There was little evidence of H2O2 attack on the carbon-carbon double bond in the phytyl substituent of Vitamin Ki, 2-methyl-3- phytylnaphthoquinone.
The solutions underwent a similar colour change to that witnessed for the reactions of the tetrahydroanthraquinones, namely noticeable darkening of the bright yellow
solution on the introduction of H2S, with subsequent lightening to a very pale yellow
colour. Although the period allowed for reduction was maintained at two hours, qualitative observations of this colour change indicated that both the reduction and
oxidation stages are markedly faster in the case of the dialkylnaphthoquinones
compared to the tetrahydroanthraquinones.
1.1.15 Kinetic Studies of Reductions and Oxidations of Quinones.
Until this point, all observations of relative rates of reduction and oxidation for the
66 tetrahydroanthraquinones and naphthoquinones had been purely qualitative. Despite the practical difficulties of monitoring H2S uptake, a series of kinetic mns were undertaken on a variety of reaction mixtures.
In the set-up used, the amount of H2S was in large excess to the amount of quinone.
For this reason, the reaction between H2S and the quinone could be studied as an example of pseudo-first order kinetics; ie., if the reaction were second order, then
-d[A] = k[A][H2S] ....1.11
dt
Where [A] is the concentration of quinone, and k the rate constant for the reaction.
However, since we can suppose [H2S] = constant, then
-d[A] = k'[A] ....1.12
dt
Where k' = kpH^S] = the observed rate constant for the reaction.
We can also state
[A]t =c(pt-poo), ....1.13
and [A]0 = c(p0-poo), ....1.14 where [A]0 and [A]t are the concentrations of the quinone A at times 0 and t respectively, pQ>pt and poo are the recorded pressures at times 0, t and at the perceived end of the reaction respectively, and c is a constant primarily related to the volume of the apparatus.
From first order reaction kinetics;
[A]t = [A]0e-k’t ....1.15 ie., In [A]t - In [A]0 = -k't , ....1.16 or In {c(pr poo)} - In {c(p0-poo)} = -k't , ....1.17 therefore In c + In (pt-poo) - In c(p0-poo) = -k't. ....1.18
Therefore a plot of In(pt-poo) against t should give a straight line of gradient -k'.
In the graphs plotted for both the reduction and oxidation studies, pt-poo was plotted
67 against time t, giving exponential curves. It should be noted that although the concentration of H2S is taken as effectively constant in the calculations above, the reaction itself is followed by measurement of changes in the pressure of H2S. This was due to the practical difficulties encountered in attempting to devise a method whereby the concentration of quinone could be measured with any accuracy, and to the availability of a digital manometer that allowed accurate measurement of comparatively small variations in pressure.
A computer program was devised (see Appendix 1) that allowed calculation of 'best-fit' values of both k' and c for the data.
This fits the equation;
(PrPoo) = c(l-e-k’t) ....1.19
The theoretical curve for the optimum values of k' and c produced by the program is that shown in each of the graphs, with the experimentally observed data plotted. It should be noted that the value of poo in each case is the value for the pressure at completion of the reaction, generated by the program from the experimental data it has received. Thus, particularly in the case of the oxidation data where the reaction does not appear to reach completion, it is open to error.
The solubility of H2S in the solvent systems can be estimated from the data obtained from the blank runs on the solvent/catalyst mixtures alone. Neglecting the effects of the vapour pressure of the solvents, at NTP H2S dissolves in 60:40 p-xylene/diisobutyl carbinol to yield a ca. 0.6 mol dm-3 solution, and in 50:50 p-xylene/sextate to yield a solution that is ca. 0.5 mol dm-3 in H2S. In each case, the solvent mixture also contains
0.38 mol dm-3 trioctylamine. These approximate values are broadly in line with those literature values quoted in the introduction for similar organic solvents.
In fig. 1.14, the relative rates of reaction for the reduction 0.25M and 0.5M solutions of ATQ respectively are shown, all other conditions remaining similar. It can be seen from fig. 1.14 that, as would be expected, the 0.5M solution took up essentially twice the volume (ie. there was twice the drop in pressure) of H2S as the 0.25M solution.
Comparison of the rate constants k' in table 1.12 indicates that there was little
68 (pt-B»)/mbar. Fig. 1.14 Kinetics Data for Reduction of varying concentrations of ATQ. of concentrations varying of Reduction for Data Kinetics 1.14 Fig. Time (secs)/ ATQ-0.25M o ATQ-0.5M •
O) CO appreciable difference in the rates of reaction for different concentrations of quinone.
Table 1.12 Comparison of kinetic data for reduction of varying concentrations of
ATQ.
Quinone Concn./ M O ptim um k' O ptim um c
ATQ 0.5 0.0013 0.1545
ATQ 0.25 0.0018 0.0830
10ml of reaction solution. Reaction solvent 60:40 p-xylene/diisobutylcarbinol, plus
0.167ml (0.38mmol, 38mM) tri-n-octylamine.
Fig. 1.15 shows the effect of varying the amount of catalytic amine in the reduction of
ATQ. It can be seen that the amount of amine had little effect on the final extent of
reaction. Comparison of values of k' in table 1.13 shows that to a first approximation,
the rate of reaction is first order with respect to the amine (ie., k' doubled as
concentration of amine doubled).
Table 1.13 Comparison of kinetic data for reduction of ATQ catalysed by varying
concentrations of tri-n-octylamine.
Q uinone Concn. Cat./mM Optimum k’ Optimum c
ATQ 19 0.00107 0.0783
ATQ 38 0.00180 0.0830
ATQ 76 0.00333 0.0843
10ml of reaction solution. 0.25M in ATQ. Reaction solvent 60:40 p-xylene/diisobutyl
carbinol.
Consideration of the data shown in fig. 1.16 and table 1.14 indicates that although
there is little difference between the rates of reduction of 2,3-dimethyl- and 2-ethyl-3-
methyl-1,4-naphthoquinone (and there is no difference in the extents to which they
were reduced by H2S), the higher value of k' for the dimethyl species indicates that it
was the more readily reduced. This is as would have been expected from the electron
7 0
ATQ-0.038MCat • ATQ-0.017MCat □ ATQ-0.076MCat o Time / (secs) Time Fig. 1.15 Kinetics for ReductionData ofATQ, varying concentrations of amine catalyst.
•jBqrn / (®d -*d) CM r*-
o o EMN • DMN Time / (secs) Time Fig. 1.16 Kinetics Data for Reduction of2,3-Dialkyl-1,4-naphthoquinones.
•jequi / («°d - ld) density considerations put forward in the introduction.
Table 1.14 Comparison of kinetic data for reduction of naphthoquinones.
Quinone Optimum k' Optimum c
EMN 0.0036 0.0941
DMN 0.0045 0.0912
10ml of reaction solution. 0.25M in substrate. Reaction solvent 50:50 p-xylene/sextate, plus 0.167ml (0.38mmol, 38mM) tri-n-octylamine.
Fig. 1.17 and table 1.15 show that for the oxidation of the 2,3-dialky 1-1,4-naphtho quinones, the reaction of the reduced ethyl,methyl- species proceeded slightly faster than the dimethyl analogue (k' = 0.00038,0.00034 respectively). This is the reverse of the case for the reduction stage of the reaction, and would again have been anticipated with regard to electron density considerations.
Table 1.15 Comparison of kinetic data for oxidation of naphthoquinones.
Quinone Concn./M O ptim um k' O ptim um c
EMN 0.5 0.00036 0.0373
EMN 0.25 0.00038 0.0284
DMN 0.25 0.00034 0.0253
10ml of reaction solution. Reaction solvent 50:50 p-xylene/sextate, plus 0.167ml
(0.38mmol, 38mM) tri-n-octylamine.
73 (Pt - Poo) / mbar- i. .7 ieisDt o xdto fvrigcnetain f23Dakl 1,4-naphthoquinone. 2,3-Dialkyl- of concentrations varying of Oxidation for Data Kinetics 1.17 Fig. Time (secs) / EMN-0.25M o EMN-0.5M • DMN-0.25M □
1.1.16 Proposed Mechanism for Amine-Catalysed Reduction of a
Quinone by Hydrogen Sulphide.
The proposed mechanism for the amine-catalysed reduction of a quinone (which is represented as a 2-alkyltetrahydroanthraquinone, but which could equally well be a 2,3- dialkyl-1,4-naphthoquinone) is shown in fig. 1.18. This mechanism is proposed in the light of experimental and physical measurements detailed in this work. The initial reaction is termolecular, between the quinone, hydrogen sulphide, and the base.
Although infra-red studies indicate that there is no appreciable interaction between the amine and H2S, the possibility of a weak interaction is not disproved, and in view of the relatively high concentrations of amine used, such a termolecular collision cannot be precluded. Following attack by the H2S on one of the quinone C=0 groups, an unstable intermediate is formed which rapidly decomposes to yield the semiquinone radical.
The semiquinone can then react by two possible routes. Firstly, further reaction with
H2S and subsequent protonation by the protonated base yields the hydroquinone.
Secondly, the semiquinone can pick up a proton, and disproportionate with a similar species to give the starting quinone and the hydroquinone.
Since the semiquinone radical could not be detected by epr spectroscopy, it is to be concluded that one or both of these reactions proceeds at a much faster rate than the reaction which produces the semiquinone itself, leading to a low equilibrium concentration of the radical.
The formation of sulphur in the orthorhombic form , ie. as Ss, is not easily explained.
A chain mechanism is proposed, where the HS- produced by the cleavage of the C-SH bond in the second stage of the proposed reduction mechanism recombines with a similar species to yield H2S2. This then reacts with the starting quinone in much the same mode as H2S, subsequent cleavage of the C-SSH bond yielding the HS2- radical.
Combination with similar or other hydrosulphide radicals will eventually produce
H2S8. The route by which this could eliminate two hydrogen atoms to yield Ss is not known.
7 5 Fig. 1.18 Proposed mechanism for amine-catalysed reduction of a quinone by H2S.
+ BH+
disproportionation
etc ? 2H S ^ = ^ H 2S2 h s 2 — — - h 2s 8 — - s 8
76 1.1.17 Proposed Mechanism for Oxidation of Hydroquinone by
Dioxygen.
The proposed mechanism for the oxidation of a hydroquinone (although the reduced form of a 2-alkyltetrahydroanthraquinone is shown, the mechanism also applies to 2,3- dialkyl-l,4-naphthohydroquinones) is detailed in fig. 1.19. It is to be noted that the amine catalyst plays no part in the proposed mechanism, since hydroquinones produced by catalytic hydrogenation, where no such amine is present, undergo similar oxidation with dioxygen.
Attack upon the hydroquinone by dioxygen yields the protonated semiquinone radical and HO2. As in the proposed reduction mechanism, the semiquinone can then react by two routes. Firstly, disproportionation yields the hydroquinone and the original quinone. Secondly, further reaction with dioxygen produces the quinone and HO2. It is in the initial reaction of the reduced form with dioxygen, as well as in the second of the two possible reactions of the protonated semiquinone, that HO2 is formed. This is a well-known precursor to H2O2 by the reaction shown at the end of fig. 1.19.
7 7 Fig. 1.19 Proposed mechanism for oxidation of hydroquinone by 0 2.
+ h o 2
2 H 0 2 ------^ H 20 2 + 0 2
7 8 1.1.18 Production of H 2O 2 by Azocompounds.
Tests for the production of hydrogen peroxide by oxidation of a commercial sample of
hydrazobenzene indicated that a yield of the order of 60% could be expected for the oxidation stage. The results obtained using the azobenzenes as substrates in the
standard reaction procedure (p83) are shown in table 1.16.
Table 1.16 Summary of Reactions of Azobenzenes by the Standard Procedure.
Substrate Solvent3 T°C Cat.b Yield /% c
Azobenzene p-Xyl/DBC 60 1.0 44.5
Azobenzene p-Xyl/DBC 60 0.5 59.2
Azobenzene p-Xyl 60 0.25 36.2
4,4'-Dimethyl p-Xyl/DBC 60 0.5 60.3 azobenzene
4,4'-Diethyl p-Xyl/DBC 60 0.5 49.3 azobenzene a25ml of reaction solvent. bml of tri-n-butylamine added. cYield based on substrate.
Reduction period two hours. See below for details of oxidation.
Although the yields of H2O2 for the above substrates are comparable, the relative rates of oxidation were markedly different. Azobenzene reduced rapidly, undergoing a colour change from deep orange to pale yellow on reduction, which was characteristic of all the azocompounds used as substrates. However, it was only possible to effect re oxidation with oxygen at ca. 3atm. pressure for 1 hour. The electron-pushing nature of the methyl groups in stabilising the oxidised form was such that oxidation of hydrazotoluene could be achieved simply by bubbling oxygen through the solution for two hours, while 4,4'-diethylhydrazobenzene was readily oxidised by this method over the period of one hour. However, this increase in ease of re-oxidation was balanced by an increase in reduction time from 30 minutes for azobenzene to over an hour for 4,4- diethylazobenzene.
79 Comparison of the experimental yields of H2O2 with the ca. 60% yield found for the oxidation of hydrazobenzene in initial experiments indicates the reduction stage went to completion.
Decreasing the amount of amine in the reduction of azobenzene from an approximately equimolar quantity had no effect on the yield of H2O2, indicating the catalytic nature of the base in the reduction.
The 60:40 p-xylene/diisobutylcarbinol solvent mixture was primarily used for the quinone systems to ensure solubility of both the oxidised and reduced compounds.
Attempts to dispense with the alcohol component in the reduction and re-oxidation of azobenzene inexplicably led to a lower yield of H2O2 and a markedly slower rate of reduction.
1.1.19 Estimate of Yield Based on Hydrogen Sulphide.
The yields of H2O2 quoted in the reduction and subsequent re-oxidation until this point had all been based on the substrate.An industrial process based upon the reduction/ oxidation cycle of any substrate would primarily require information on the yield in relation to the amount of hydrogen sulphide.
Reduction of azobenzene at 25 °C by a measured initial volume of H2S at atmospheric pressure allowed a yield of 46.7% according to the H2S to be calculated. The reduction and subsequent re-oxidation was effected in 2:1 p-xylene/tributylamine in order to ensure rapid and complete reduction of the substrate.
1.1.20 Production of H2O2 by Flavins.
Initial attempts to reduce a 0.01M solution of the sodium salt of flavin mononucleotide
(FMN-Na) in aqueous solution at pH 6.8 (0.1M l,4-piperazinebis(ethanesulphonic acid) (PIPES) buffer) with sodium borohydride and sodium dithionite respectively in anaerobic conditions failed. At best only partial reduction was achieved to give a highly coloured intermediate.
80 The work of Tamao et al79 had shown that the riboflavin-2',3',4',5'-tetraacetate, fig. 1.20, would produce H2O2 in a cycle when reduced with N-benzyl-1,4- dihydronicotin amide (BzlNAH). Reduction of this flavin with H2S catalysed by tributylamine, and subsequent re-oxidation with O2, produced a yield of H2O2 of 28.8%, based on the flavin.
Although an encouraging result ,the problems of the expensive nature of the compound and the low solubility of the substituted flavin at room temperature led us to the conclusion that the system held few possibilities for an industrial process.
81 1.2. PHYSICAL MEASUREMENTS.
E.P.R.spectra were recorded on a Varian E-12 spectrometer.
1H n.m.r. spectra were recorded on a Jeol GSX270 (270MHz) spectrometer.
I.R. spectra were recorded on a Perkin-Elmer 1720 Fourier Transform spectrometer.
Pressure measurements during kinetics studies were made using a Honeywell DPM140 series digital manometer.
Raman spectra were recorded on a Spex Ramalog V instrument.
U.V./Visible spectra were recorded on a Perkin-Elmer Lambda 2 Series PECSS spectrophotometer.
X-Ray powder diffraction spectra were recorded on a Siemens D500 Kristalloflex diffractometer.
1.2.1 Standard Reaction Apparatus.
The apparatus used for all the H2S reductions performed under pressure is shown in fig. 1.21 below; Vacuum Pump with Cold Trap
H2S Cylinder
reaction vessel.----
—Water Bath
Stirred Reaction Solution
Fig. 1.21 Standard Reaction Apparatus.
8 2 1.2.2 Standard Reaction Procedure.
The following general reaction procedure was used for the reductions of the quinone, azobenzene and flavin systems under pressure of hydrogen sulphide, and subsequent re oxidation with oxygen, except where slight deviations are noted in the text.
The substrate was dissolved in 25ml of reaction solvent, with a catalytic amount of amine. The reaction vessel was placed in a water bath maintained at the desired temperature (or ambient room temperature), evacuated and the solution thoroughly degassed. Hydrogen sulphide, up to a pressure of 2-3atm, was introduced and the reaction mixture stirred for two hours. During this time, any change in colour of the reaction solution was noted. The H2S was evacuated from the vessel and attached tubes, as was that dissolved in solution. The apparatus was opened up to the air, and
10ml of water added. Oxygen was bubbled through the solution for a further hour with stirring. The aqueous/organic solvent mixture was poured into a separating funnel and the lower aqueous layer taken. An aliquot of appropriate volume was taken and analysed iodometrically as detailed below.
83 1.2.3 Standard Reaction Apparatus for Kinetics Studies.
The standard reaction apparatus used for kinetics studies is shown below in fig. 1.22
Manometer -ive Port
Fig. 1.22 Standard reaction apparatus for kinetics studies.
This apparatus was used in both the reduction by H2S and the oxidation by O2 stages of the reaction.
1.2.4 Standard Reaction Procedure for Kinetic Studies.
The volume of the apparatus above was calculated as follows; i. The gas bulb was filled with water and its volume calculated from the weight. ii. Simple geometry was used to calculate the volume of the cone in the neck of the bulb and the small length of pipe leading to Tap A.
Total volume of bulb + cone + pipe = Vj = 585.8ml. iii. With the bulb kept at latm. pressure of air (Tap A. closed), the vacuum line from
Taps F to D was evacuated (Tap D closed, manometer connected at negative differential port to Tap E, Tap E open, reaction vessel containing only magnetic follower).
8 4 Pressure Pi recorded after Tap F closed.
Pi = 0.993 iv. Bulb opened at Tap A. Pressure P2 recorded.
P2 = 0.201
Let volume of apparatus excluding Vi = x.
From P1.V1 = P2.V2,
0.993x = 0.201(585.8 +x),
therefore x = 148.7ml,
and total volume of system = 585.8ml + 148.7ml = 734.5ml. v. Subtract volume of solvent/catalyst to give volume of gas.
Volume of solvent/catalyst = 10.17ml
so volume of gas = 724.3ml.
For the measurement of H2S pressure, the use of mercury manometers is excluded due to the attack of H2S on the mercury. Accordingly, a Honeywell DPM140 digital manometer was used. The suppliers informed us that they had previously successfully tested the instrument with H2S. The instrument relies on a pressure sensor utilising piezo resistors based on a 2.5mm2 silicon chip. The gas comes into contact only with silicon, polyester and epoxy adhesive. The range of the instrument was 0-1 bar, and it has a claimed accuracy of better than 1%, but a discrimination of 0.1% at 1 bar.
Problems were caused initially by the pipes that connected both the reaction vessel and the manometer to the vacuum line. Rubber pressure tubing and, to a lesser extent,
PVC tubing both allowed diffusion of H2S out of the system, causing a steady drop in pressure even when reaction or dissolution of the gas had ceased. Since a partial vacuum was produced by uptake of H2S, any leak in the system would have manifested itself by an increase in the pressure reading. Glass tubing was one possible solution to this problem. Yet for practical reasons, the connections to the manometer required flexibility. To remedy this, 4mm glass tubing, jointed at intervals by short lengths of butyl rubber tubing, was used. Even this improved set-up led to diffusion of
8 5 H2S out of the system, indicated by a drop in pressure of ca. 0.003atm. every thirty minutes.
The standard kinetic studies of the reduction stage were effected as follows. The manometer was attached at the positive differential port to Tap E, the negative port to
Tap G. 10ml of the reaction solution plus a catalytic amount of amine was added to the reaction vessel, which was immersed in a bath maintained at 25±0.5°C and rapidly stirred by means of a magnetic follower. The system was thoroughly evacuated and the reaction solution degassed. Then with only Taps C and F closed, the H2S cylinder was opened and H2S introduced until the manometer reading was ca. 1.2atm.. Tap A was then closed and the H2S evacuated thoroughly from the system by opening Tap F.
To start the run , Taps F and D were closed, and Tap C opened. At time T =0, Tap A was opened and the pressure noted. The pressure was then recorded at intervals initially of 20 seconds, increasing as the rate of pressure drop slowed.
Oxidation runs were performed subsequent to the above reduction runs. After complete reduction, Tap F was opened and the system completely evacuated of H2S. Then Taps
F and C were closed and Tap H opened. Once the reading on the manometer was ca. latm., Taps H and A were closed, and Tap F opened once more. After evacuation of air from the vacuum line, Tap F was closed. To start the run, Tap C was opened. At time T = 0, Tap A was opened and the pressure measurements recorded as before.
Blank runs were carried out for both the reduction and oxidation stages of each system by monitoring the uptake of H2S for 10ml of the appropriate solvent/catalyst mixture as in the standard reaction runs. These readings were subtracted from the experimental readings to give the data for the substrate alone.
8 6 1.2.5 UV/Visible Study of the Reduction of ETQ. i. A 1mm path length quartz cell, connected to a PTFE Rotaflo tap by a graded glass joint, was used. 0.187g (0.78mmol) of 2-ethyltetrahydroanthraquinone was dissolved in 7.5ml of 60:40 p-xylene/diisobutylcarbinol. 0.012ml (O.Olg, 0.028mmol) of tri-n- octylamine was added and the solution stirred thoroughly. An aliquot of the solution was added to the quartz cell and the uv/visible spectrum recorded from A^350-550nm, using a solvent reference cell. The cell was evacuated, flushed with H2S (Time T=0) and the reaction monitored by recording the UV/Visible spectrum at intervals over the time period indicated. ii. A 10mm path length quartz cell fitted with a Rotaflo tap was used. 0.019g
(0.078mmol) of 2-ethyltetrahydroanthraquinone dissolved in 7.5ml of reaction solvent plus 0.012ml (0.028mmol) of tri-n-octylamine was treated as above, and the reaction similarly monitored.
1.2.6 EPR Study of the Reduction of ETQ.
0.0187g (0.078mmol) of ETQ was dissolved in 7.5ml of 60/40 p-xylene/dbc and
0.012ml trioctylamine added. The solution was placed in an epr tube fitted with a rubber septum cap and degassed with argon. Hydrogen sulphide was then bubbled through the solution for 2-3 minutes. The tube was immediately placed in the epr spectrometer and over a period of 60mins the epr spectrum was monitored continuously.
In order to locate the semiquinone radical that might be present, the work of Bolton et also Was followed. 1.20g (5mmol) of ETQ was dissolved in 40ml of ethanol and 0.76g of zinc dust (90%) added. To the hot solution, 0.83g of NaOH in 10ml of water was added dropwise.
A small amount of the solution was filtered and transferred anaerobically to an epr tube and the epr spectrum recorded.
8 7 1.2.7 EPR Study of the Reduction of ETQ Catalysed by 1,8-
Bis(dimethyl) aminonaphthalene.
An argon-degassed solution of 0.188g (0.78mmol) of ETQ and 0.024g (O.Ollmmol) of l,8-bis(dimethyl)aminonaphthalene was reduced in an epr tube by bubbling H2S for
30mins, and the epr spectrum recorded.
1.2.8 EPR Study of Reduction of QS.
0. 066g (0.2mmol) sodium anthraquinone-2-sulphonate was dissolved in 10ml of water and 0.0lg (0.03mmol) of sodium tetraborate added. The reaction tube was thoroughly flushed with argon, and hydrogen sulphide was then bubbled through the solution for five minutes until the pale yellow solution changed to a deep green colour. The solution was then transferred anaerobically to a sealed epr tube previously flushed with argon and the epr spectrum recorded every ten minutes over a period of two hours.
For comparison of the signal, 0.0174g (O.lmmol) of sodium dithionite was added under an argon atmosphere to 0.066g of the anthraquinone in 10ml of solution as above, the solution transferred anaerobically to an epr tube, and the epr spectrum recorded.
1.2.9 1H NMR Study of the Reduction of BTQ.
Attempts were made to find a suitable solvent system for the reduction such that the reaction proceeded as normal and deuteriated forms of the solvent components were readily available.
1. 0.625g (2.6mmol) ETQ was reduced in 25ml of CHCI3 with 0.092ml (0.07g,
O.38mmol) tributylamine by the standard reaction procedure. The reduced form proved insoluble in chloroform. ii. i was repeated in 25ml of N,N-dimethylformamide. The reaction failed. iii. With BTQ (0.699g, 2.6mmol) as the substrate in the reduction in CHCI3, reduction proceeded as usual. However, due to a poor yield of H2O2 upon re-oxidation (17.7%),
88 another solvent system was sought. iv. Analogous to the p-xylene/dbc system used in our normal system, (0.699g,
2.6mmol) of BTQ was dissolved in 25ml of 60:40 benzene/methanol with 0.092ml tributylamine and reduced by H2S, the reaction proceeding as usual. Due to the high yield of H2O2 produced (75.0%), this was the system studied by nmr, using the deuteriated forms of the solvents, CeD6 and CD3OD respectively.
Three spectra were recorded. These were; i. Substrate reference; 0.028g (O.lmmol) BTQ dissolved in 0.75ml CDCI3 and 0.1%
TMS. ii. Argon saturated reaction solution; 0.028g (O.lmmol) BTQ dissolved in 0.75ml
50:50 C6D6/CD3OD, 0.0037ml tributylamine and 0.1% TMS. iii. The above solution saturated in situ with H2S, and the nmr spectrum recorded over the time period of an hour.
1.2.10 IR Study of the Interaction between H 2S and Tributylamine.
An infra-red solution cell with sodium chloride windows was used. H2S was bubbled through a test-tube fitted with a rubber septum and containing 10ml of tetrachloroethylene. After 5 minutes, some of the saturated solution was transferred anaerobically to the solution cell, and the infra-red absorption spectrum (4000-250cm*1) was recorded. The spectrum of tetrachloro ethylene alone was then recorded, and this was subtracted from the initial spectrum to give the absorption spectrum of H2S.
The process was repeated with another 10ml of tetrachloroethylene, except 0.1ml of tributylamine was added prior to saturation with H2S. The spectrum of this saturated solution was then recorded, and the separately-run spectrum of tetrachloroethylene containing 1% v/v tributylamine subtracted to yield the spectrum of H2S in the presence of tributylamine.
89 1.3 EXPERIMENTAL PROCEDURES.
1.3.1 Iodometric Analysis.
The hydrogen peroxide produced by all the reactions involving hydrogen sulphide reduction was analysed iodometrically following the method described by Vogel81.
An aliquot of the hydrogen peroxide-containing aqueous layer was added to a solution of 10ml of 10% potassium iodide in 100ml of 1M sulphuric acid. To this was added 3 drops of 3% ammonium molybdate solution, and a small piece of solid carbon dioxide to prevent aerial oxidation of the iodide. The liberated I2 was titrated against freshly- prepared 0.1M sodium thiosulphate, a few drops of a freshly prepared 1% aqueous solution of soluble starch being added near the end-point as an indicator.
1.3.2 Argon Flushing.
This process was used after the reduction of 2-ethylanthraquinone and the sulphonated anthraquinones in an attempt to remove any residual H2S after evacuation. After evacuation of the hydrogen sulphide, prior to oxidation, a pressure of argon of 1-2 atm was applied to the reaction system for several minutes, the system then evacuated once more and argon introduced again. After several flushings with argon, the oxidation and extraction was carried out as usual.
1.3.3 Reduction of Sulphonated Anthraquinones.
The anthraquinone (4mmol) was stirred in 25ml of water, to which was added 0.125g
(2mmol) of boric acid, and the pH adjusted to 9.0 with saturated sodium hydroxide solution. After evacuation and reduction/oxidation as in the standard reaction, 10ml aliquots of the reaction mixture were centrifuged and analysed for H2O2 as above.
1.3.4 Removal of Sulphur with Carbon Disulphide.
Sulphur was removed from the reduced reaction mixture of both 2-ethylanthraquinone and the sulphonated anthraquinones prior to oxidation in order to eliminate the
90 possibility of reaction between both colloidal and dissolved sulphur and the hydrogen peroxide formed upon re-oxidation of the substrate. The reduced reaction mixture was shaken with 25ml of carbon disulphide, the two layers separated and the reaction mixture oxidised in the normal way. In later experiments, the less hazardous carbon tetrachloride was substituted for carbon disulphide.
1.3.5 Carbon Dioxide Flushing.
This was an attempt to acidify the reaction solution and convert any HS- present to the more easily removed H2S and was used with both 2-ethylanthraquinone and the sulphonated anthraquinones. After addition of water to the reduced reaction mixture, but prior to re-oxidation, CO2 was bubbled through the reaction mixture until H2S could no longer be detected by lead acetate paper.
1.3.6 Continuous Run on One Reaction solution.
The continuous run process was an adaptation of the standard reaction procedure described previously. 0.625g (2.6mmol) ETQ was reduced by H2S in the presence of
0.07g (0.091ml, O.38mmol) tributylamine. After evacuation of H2S, the reaction solution was transferred anaerobically by means of a filter-stick to a second argon- flushed flask to which 10ml of water was added and the substrate re-oxidised with oxygen. Thus at no stage did the H2O2 formed come into contact with the colloidal sulphur that began to precipitate out of the sulphur-saturated solution during and after the third reduction. The aqueous layer was separated and H2O2 was analysed for iodometrically as detailed earlier.
Reaction of H2O2 with sulphur was measured using a method for the gravimetric analysis of sulphate as barium sulphate, outlined by Vogel71. 5ml of the extracted aqueous layer was added to 25ml of water, and 0.25ml of concentrated hydrochloric acid added. The solution was heated to boiling, and 10ml of 0.5% barium chloride solution added dropwise from a burette. After standing overnight, the suspension was filtered through a previously oven-dried and weighed sintered glass crucible. The
91 precipitate of barium sulphate was heated to 150°C for several hours to constant
weight. From the weight of barium sulphate, and the equation;
3H20 2 + S -----► H2S04 + 2H20 ....1.20
the amount of hydrogen peroxide lost by reaction with dissolved sulphur was
calculated. Loss of catalyst by extraction into the aqueous layer was analysed by a two
stage process. Since extraction of the aqueous layer into dichloromethane would
remove not only any catalyst, but also any dissolved solvent, the combined aqueous
extracts from three runs were shaken with 0.5ml of 1M H2S 04 and extracted into 10ml
of dichloromethane. At this stage the amine would be present in aqueous solution as the
sulphate salt, and would not be extracted. The dichloromethane layer was poured into a
preweighed tube and allowed to evaporate. A pellet of sodium hydroxide was added to
the combined aqueous extracts, and these were again extracted into 10ml of dichloro
methane, which was treated as before. Upon addition of the sodium hydroxide, the
amine salt is converted back to the amine and thus can be extracted into the
dichloromethane.
1.3.7 Purification of Sulphur for X-ray Powder Diffraction and Raman.
The sulphur filtered from later runs on the same quinone/solvent system above was
washed with a little cold methanol to remove any traces of reaction solvents and
allowed to air dry.
1.3.8 Re-Oxidation of Hydrazobenzene.
After addition of 25ml of water to extract H20 2 as it was formed, the reduction product of azobenzene required a pressure of ca. 3atm. of oxygen for several hours in order that re-oxidation occurred.
1.3.9 Reduction of Azobenzene at Atmospheric Pressure of H2S.
This was performed in order to calculate the yield based on the H2S rather than the substrate. 1.82g (0.01 mol) of azobenzene was dissolved with stirring in 30ml of the
92 2:1 p-xylene/tri-n-butylamine. Evacuation was effected so as to minimise the loss of
xylene (the vacuum valve was quickly opened and closed every couple of minutes,
allowing the reaction vessel to equilibrate with the vacuum). The stirrer was then
turned off, and H2S slowly let into the system. Any pressure over atmospheric was
quickly let off by means of a tube connected to a Dreschel bottle containing concentrated
NaOH solution. The reaction was then stirred at 25°C for two hours. The system was
then opened up to the air, and 25ml of water added. No argon nor carbon dioxide
flushing was deemed necessary since all of the H2S should have reacted - a vacuum
existed within the vessel when it was opened to the air. The reduced substrate was then
re-oxidised as oudined above, the aqueous layer being analysed for H2O2.
1.3.10 Reduction of RibofIavin-2',3',4',5'-tetraacetate.
Due to its limited solubility, the flavin (0.3 lg, 0.625mmol) was reduced at 606C in
25ml of 60:40 p-xylene/diisobutylcarbinol in the presence of tributylamine (0.5ml,
2.1mmol). After a reaction period of two hours, the system was cooled, then treated as
usual with evacuation and addition of 10ml of water. Flushing with carbon dioxide was performed before bubbling of oxygen through the solution, which was continued for
thirty minutes.
93 1.4 MATERIALS AND SYNTHESES.
1.4.1 Solvents.
Diisobutylcarbinol (DBC, 2,6-dimethylheptanol), Sextate (O-methylcyclohexylacetate)
and Shellsol (a mixture of alkylbenzenes) were obtained from Interox Ltd. The
substitute solvents 2,6-dimethylheptanol (DMH, technical grade) and p-xylene were purchased from Aldrich, as were N-methylpyrollidinone (NMP) and dimethyl formamide (DMF). Tri-n-butyl phosphate (TBP) was obtained from BDH.
1.4.2 Amines.
All amine catalysts were purchased from Aldrich and used without further purification.
1.4.3 Anthraquinones.
Disodium anthraquinone-2,6-disulphonate (QDS), 2-ethyl anthraquinone (EQ) and sodium anthraquinone-2-sulphonate (QS) were purchased from Aldrich and used without further purification.
1.4.4 Tetrahydroanthraquinones.
2-Amyl, 2-lbutyl and 2-ethyltetrahydroanthraquinone (ATQ, BTQ and ETQ respectively) were obtained from Interox Ltd. 2-tButyl tetrahydroanthraquinone was characterised by !H N.m.r. (1H n.m.r. (CDCI3) 270MHz; d 8.08 (1H, d, Hi), 7.99
(1H, d, H4), 7.70 (1H, dd, H3), 2.58 (4H, m, H5,5\8,8'), 1.74 (4H, m, H6.6\7,7’)}
1.38 (9H, s, (CH3)3).
1.4.5 1,4-Naphthoquinones.
2,3-Dimethyl-1,4-naphthoquinone (DMN), 2-methyl-1,4-naphthoquinone (MN) and 2- methyl-3-phytyl- 1,4-naphthoquinone (Vitamin K4) were obtained from Aldrich and used without further purification.
9 4 2-Ethyl-3-methyl- 1,4-naphthoquinone (EMN) was initially prepared as outlined by
Fieser and Change, a solution of 13.76g (0.06mol) of 2-methyl- 1,4-napthoquinone
and 9.6ml (0.075mol) of ethyl acetoacetate in 150ml of propionic acid was heated to 85-
95°C with stirring and 180g (4.37mol equiv.) of Pb3C>4 added in ca.l5g portions.
Upon addition of the first load of oxide, evolution of gas indicated commencement of
reaction. Each time evolution of gas slowed, a further load of oxide was added; addition
of all of the oxide took around three hours. The mixture was heated with stirring for a
further three hours, then cooled a little and ether added with continuous stirring. The
directions given by Fieser et al at this point are not clear. It is stated that the "semi
solid" mixture was extracted several times with ether. In this preparation, it was found
that the mixture solidified completely at temperatures below ca.60°C, well above the
boiling point of diethyl ether (34-36°C), and that addition of ether to the hot mixture
cooled it to such an extent that solidification rapidly ensued, hindering effective
extraction. Therefore a series of extractions (addition of ether to the cooled solution,
leaving the former to boil - adding more ether as necessary - decanting off the ether
extract and heating the reaction mixture until it melted once more) were carried out,
leaving a final volume of extract of ca.500ml. The ether solution was washed with
3x300ml of saturated sodium bicarbonate solution (until neutral), dried with calcium
chloride, filtered and evaporated. Cooling and manipulation of the resultant brown oil
yielded crystals which were recrystallised from methanol in an ice-bath. The crystals
formed were recrystallised from methanol twice more (once with decolourising
charcoal), yielding yellow-brown needles.
Yield= 3.18g (21.2%) Melting Point= 72-3°C (Lit. Value82=73°C)
Analysis; Calculated for Ci 3H 120 2 %C 77.98 %H 6.04
Found %C 78.27 %H 6.00
In view of the poor yield and unsatisfactory nature of the preparation above, the method
outlined by Thomson8^ was followed. 12.0g (0.07mol) of 2-methyl naphthoquinone was partially dissolved in 120ml of glacial acetic acid and 12ml of acetaldehyde. The
95 suspension was cooled in an ice-bath and dry hydrogen chloride passed for 25 minutes.
After leaving overnight, the dark red solution was poured onto 250ml of ice, and the
purple precipitate washed with a little cold
ethanol. The solid was then recrystallised from the same solvent (charcoal) and isolated
as yellow needles.
Yield= 6.36g (38.9%), Melting Point= 146-7°C (Lit. V alu ed 145-6°C)
Analysis; Calculated for C 13H11O2CI %C 66.53 %H 4.72
Found %C 66.64 %H 4.63
2.0g of 2-chloroethyl-3-methylnaphthoquinone was dissolved in 80ml of glacial acetic
acid with anhydrous sodium acetate (2.0g), and hydrogenated in the presence of 5%
palladium on barium sulphate (0.3g). After absorption of 2 molar equivalents of
hydrogen, the suspension was filtered and the filtrate oxidised for 15 minutes at 60°C
with a solution of lg of chromium trioxide in 10ml of water. The product was isolated
as a yellow solid from the green solution by dilution and cooling in an ice-bath. It was
then recrystallised from methanol in bright yellow needles.
Yield= 0.98g (48.7%), Melting Point= 72-3°C
Analysis; Calculated for Ci3H i20 2 %C 77.98 %H 6.04
Found %C 77.50 %H 5.99
1.4.6 Azobenzenes.
4,4'-Dimethylazobenzene was prepared by reduction of p-nitrotoluene with stannous
chloride, according to Cook84.
Melting Point= 142-4°C (Lit. Value84= 145°C)
In another method, 4-methylnitrobenzene was reduced with zinc and sodium hydroxide
according to Carlin and Wich85.
Melting Point= 141.5-3.5°C
Another preparation of this compound was carried out by reduction of p-nitrotoluene according to Chung et al86.
96 Melting Point= 141-3°C
Analysis; Calculated for Q 4H 14N2 %C 79.97 %H6.71 %N 13.32
Found %C 79.46 %H 6.60 %N 13.34
4,4-Diethylazobenzene was prepared by reduction of 4-ethylnitro benzene as outlined by Carlin and Wich85.
Melting Point = 54-55°C (Lit Value87 = 63°C)
Analysis; Calculated for C 16 H 18N2 %C 80.63 %H7.61 %N 11.75
Found %C 80.59 %H7.63 %N 11.82
In each of the above preparations lower yields than those quoted in the literature were obtained. This was in part due to the desire to synthesise a substrate simply for use in the standard reaction procedure, and in the case of the work of Chung et al86 due to practical difficulties in maintaining a vacuum tight seal on a glass vessel that expands appreciably when heated to 150°C for 10 hours. Attempts to produce 4,4'-diethyl azobenzene by this method were unsuccessful.
4,4'-Dimethoxyazobenzene. Attempted preparation by the method of Carlin and Wich8^ gave results that led us to believe that 4,4-dimethoxyazoxybenzene was the product.
Yield= 7.63g (56.0%), Melting Point= 118°C (Lit. Value for 4,4-dimethoxy azobenzene88 = 163-165°C, 4,4'-dimethoxyazoxybenzene89 = 118.5°C).
Analysis; Calculated for C 14H 14N2O2 %C 69.40 %H5.82 %N 11.56
Calculated for Ci4N 14N203 %C 65.11 %H 5.46 %N 10.85
Found %C 65.27 %H 5.42 %N 10.86
Attempts to repeat the preparation according to Cook84 seemed to produce a mixture of the desired product and the azoxy compound. Attempts to separate or reduce the mixture proved unsuccessful.
97 1.4.7 Flavins.
Flavin-2',3',4',5'-tetraactetate was prepared according to Kyoguku and Yu90. 3.0g
(8.0mmol) of riboflavin was stirred for 50 hours at room temperature with 35ml
(34.2g, 0.43mol), 5ml (5.4g, 0.053mol) of acetic anhydride and 50mg (0.4mmol) of
4-dimethylaminopyridine as a catalyst. Over this period of time, the solution became clear, and its colour changed to deep orange with green fluorescence. The pyridine was removed by successive addition of ethanol and evaporation. The brown-yellow solid residue was recrystallised from ethanol to give a yellow crystalline solid.
Yield = 3.81g, (70.2%). Melting point = 245°C (Lit. Valued = 245°C).
98 1.5 GENERAL CONCLUSIONS.
The production of hydrogen peroxide from hydrogen sulphide and dioxygen is feasible from a thermodynamic viewpoint. However, the probability of reaction between H2S and H 2O2 requires a two stage reaction, where a substrate is first reduced by H 2S, which is then evacuated, and then oxidised with oxygen to form H 2O2.
Possible candidates for this substrate are quinones, azocompounds and flavins.
Attempts to produce H2O2 by means of the reduction of anthraquinones by H 2S and their subsequent oxidation with oxygen were generally unsuccessful, since the substrates were not initially reduced by H 2S.
Reduction was achieved in the case of 2-ethylanthraquinone and sodium anthraquinone-
2-sulphonate in highly polar solvents 1(-methylpyrolidinone and water respectively) where an amine catalyst was not required. Unfortunately, the miscibility of these solvents with water would make extraction of H2O2 difficult in an industrial context.
EPR studies of the reduction of sodium anthraquinone-2-sulphonate in aqueous solution by H2S indicate the presence of a radical intermediate which can be identified as the semiquinone species by comparison with the spectrum obtained upon reduction of the substrate with dithionite.
Reduction and subsequent oxidation of 2-alkyltetrahydroanthraquinones produces
H2O2 in high yield (yields based on substrate in the range ca. 75-95%). The solvent used in these systems, a mixture of alkylbenzenes and a long chain aliphatic alcohol, serves to solubilise both the starting quinone and the hydroquinone reduction product.
All of these reductions require an amine catalyst. The most important consideration for the suitability of an amine as a catalyst in these systems appears to be the pK value. An optimum pK range of ca. 7.8-12.3, with a rapid drop in efficiency between pK 7.8-
7.4. This is explained by the proposed mechanism which requires that the base must be able to initially gain a proton, and give it up at a later stage in the reduction. Steric factors do not appear important for the catalyst.
Sulphur was precipitated during a continuous reduction/oxidation cycle on the same 2- alkyltetrahydroanthraquinone/solvent solution in a form that was easily filtered. This
99 was an important consideration in assessing the viability of the process in an industrial
context, where filtration would be the easiest way of removing sulphur from the
reaction mixture. X-ray powder diffraction and Raman spectroscopy analyses of this
sulphur indicated that it was produced in the orthorhombic, a, form of S8- Losses of
amine and reaction solvent caused by the extraction of H2O2 into water in the
continuous runs was negligible.
UV/visible absorbance studies of the reduction of 2-ethyltetrahydroanthraquinone
indicate that the reaction occurs in two stages. NMR studies of the same reduction
show that it goes to completion, and that, as expected, the hydroquinone form is the
reduction product.
IR studies of the interaction between H2S and the amine catalyst suggest that negligible
amounts of HS- are present in solution in the solvents systems used in the reductions of
both the alkyltetrahydranthraquinones and the naphthoquinones.
The 2-alkyltetrahydroanthraquinones, however, allow little scope for adjusting the
realtive rates of reduction and oxidation by altering the substituents on the central
quinone ring (the alkyl substituent on the molecule is, as in the case of the alkyl
anthraquinones, largely for solubility purposes). For this reason, the reduction and
subsequent oxidation of the 2,3-dialkylnaphtho quinones was studied. Early attempts to
use 2-methyl-1,4-naphthoquinone in the process indicated that it was easily reduced,
but that re-oxidation could not be effected even under pressures of oxygen for extended
periods. From the study of redox potentials for the quinones in general, it can be seen
that the presence of electron-pushing groups on the molecule serves to lower the redox
potential and stabilise the oxidised (or unreduced) form. It was found that the presence
of another electron-donating alkyl group on the alkynaphthoquinone was sufficient to
allow re-oxidation of the reduced substrate to occur. This process was effected in a
solvent mixture of alkylbenzenes and O-methylcyclohexylacetate.
The amine-catalysed reduction of 2,3-dialkyl-1,4-naphthoquinones by hydrogen
sulphide, and their subsequent oxidation with oxygen , produced hydrogen peroxide in very high yields (yields based on the substrate in the range ca. 90-100%). On the
100 grounds of both solubility and favourable rates of reaction for the two stages, 2-ethyl-3- methyl-1,4-naphthoquinone is the preferred substrate. Patent literature for the use of these substrates in the catalytic hydrogenation process indicates that negligible degradation of the quinone occurs during the process.
The major disadvantages of the process from an industrial viewpoint are the need to filter large quantities of sulphur from the reaction solution, and the much higher solubility of hydrogen sulphide in organic solvents compared to hydrogen. Thus a great deal more effort will be expended removing unreacted gas. In balance, however, hydrogen sulphide is an extremely cheap commodity, with many companies actually paying for its disposal.
The use of azocompounds as substrates produced moderate yields ( yields based on substrate in the range ca. 50-60%, yield based on H 2S for azobenzene 47%).
Reductions were effected at 60°C. As with the tetrahydroanthraquinones, reduction requires an amine catalyst, and the preferred solvent is p-xylene. Unlike the tetrahydroanthraquinone systems, however, no alcohol component seems to be required for the reduced form to remain in solution. As expected from study of the redox potentials, ease of oxidation increases from azobenzene through azotoluene to
4,4'-diethylazobenzene; only in the case of the alkylated compounds can oxidation of the hydrazo form be effected without recourse to pressures of oxygen. Even so, re oxidation requires longer periods than that encountered in both the2 - alkyltetrahydroanthraquinone and 2,3-dialkyl-l,4-naphtho quinone substrate systems.
The only success met with in the use of flavins as substrates was with riboflavin-
2',3,,4',5'-tetraacetate (yield based on substrate 28.8%). An amine catalyst was once more required. Problems with solubility of the substrate (reaction solvent a mixture of alkylbenzenes and a long chain aliphatic alcohol) meant that this reaction had to be effected at 60°C. The low yield of H 2O2 obtained and the high cost of the substrate, indicates that this sytem has little potential in an industrial context.
101 II. SPECTROPHOTOMETRIC DETERMINATION OF
HYDROGEN PEROXIDE.
2. SPECTROPHOTOMETRIC DETERMINATION OF HYDROGEN
PEROXIDE WITH TITANIUM (IV) SULPHATE.
2.0 INTRODUCTION
Although iodometric analysis is in itself an excellent technique, its use is limited when
dealing with highly coloured reaction solutions, or where other species are present that
are equally capable of oxidising iodide to iodine. A technique was required for analysis
of highly coloured aqueous solutions of2 HO2, in particular those produced in the
experiments detailed in section HI of this thesis. Since standard colorimetric methods
were obviously inapplicable, an ethyl acetate extraction method, based on the
colorimetric determination of H2O2 with titanium(IV) sulphate, was developed.
The reaction of hydrogen peroxide with titanium (TV) in acidic solution, to give a
yellow peroxytitanium species, provides a well-known method for the
spectrophotometric determination of hydrogen peroxide in aqueous solution91. A
variety of titanium (IV)compounds have been employed in this reaction, including
titanium sulphate91, oxalate92, chloride93 and 8-hydroxyquinolinate94. However, the method is subject to a number of interferences93 from species such as phosphate, which react with titanium (IV), or those which either liberate hydrogen peroxide or react with it on acidification (eg. inorganic peroxy compounds93 and chromate96)- In addition,
substances which absorb near 410nm will interfere strongly.
It was hoped that extraction of aqueous hydrogen peroxide into ethyl acetate and back- extraction into acidic titanium (TV) sulphate solution could be used to overcome the problems of interference. The yellow species formed, possibly [(H2 0 )5 Ti(0 2 )]2+, absorbs at = 407nm (e = 770 mol-1 dm 3 cm*1).
102 It has been known for many years 97 that H2O2 partitions itself appreciably between water and a variety of polar solvents. For ethyl acetate, Walton and Lewis98 reported a value for the |H202]EtOAc/[H202]H20 partition coefficient of 0.250. This was obtained using strong (ca. 3 mol dm-3) solutions of 2HO2. For more dilute solutions (initially lxlO -3 mol dm-3), we have found a smaller value of 0.121.
Although only a fraction of the hydrogen peroxide is extracted into the ethyl acetate, the back-extraction is efficient owing to the high stability constant for the peroxytitanium species". Hence, by using a small volume of titanium (IV) solution, it is possible to attain a higher sensitivity than with the conventional method.
2.1 RESULTS AND DISCUSSION.
2.1.1 Spectrophotometric Determination of H 2O 2 after Extraction with Ethyl Acetate.
Figs. 2.1, 2.2 and 2.3 show plots of absorbance against original concentration of hydrogen peroxide for the method employing 5, 3 and 1ml of titanium (IV) solution respectively. Good straight lines are obtained, indicating that the partition coefficient of
0.121 previously stated applies over the concentration range 0-3xl0-3mol dm-3 H2O2.
When 1ml of the titanium(IV) solution is used, a lxl0-4mol dm-3solution of hydrogen peroxide gives an absorbance of 0.114, compared to 0.069 with the standard method, using a volume ratio of H2O2 solution:Ti(IV) reagent of 9:196.
It is assumed, in applying the extraction technique, that the distribution coefficient of hydrogen peroxide between the aqueous solution and ethyl acetate remains constant.
Any change in readings over a room temperature range of ca. 20-25°C appeared negligible.
103 Absorbance Fig. 2.1 Ethyl Acetate Extraction Analysis with 5ml of Ti(IV) Reagent. Ti(IV) of 5ml with Analysis Extraction Acetate Ethyl 2.1 Fig.
104 Absorbance Fig. 2.2 Ethyl Acetate Extraction Analysis with 3ml o f Ti(IV) Reagent. Ti(IV) f o 3ml with Analysis Extraction Acetate Ethyl 2.2 Fig. [H 2 O 2 / ‘ 0 ]/1 3 dm ld o m 3
105 Absorbance Fig. 2.3 Ethyl Acetate Extraction Analysis with 1ml of Ti(IV) Reagent. Ti(IV) of 1ml with Analysis Extraction Acetate Ethyl 2.3 Fig. [H 2 0 2] / / 2] 0 10 ' 4mol dm 4mol ' 3
106 2.1.2 Effect of Interfering Species on Extraction.
Analyses were repeated in the presence of a variety of salts, some of which contain ions
reported to interfere with the direct analysis (HPO 42 - C1O 42-’ Ni2+, Al3, F-)93.Table
2.1 shows the salts studied, and the molar concentrations at which the resultant error
reached ca. 2 %.
Table 2.1 Limits of Interfering Species upon Extraction Method.
NaCl Na2S 0 4 Phosphate Buffer* HEPES* NiCl 2 A12(S 04)3 KF Na2C r04
0.05 0.05 0.02 0.3 0.1 0.05 0.1 0.01
*Phosphate buffer - dihydrogen potassium phosphate and dipotassium hydrogen
phosphate pH 6 .8; HEPES- [4-(2-hydroxyethyl) -1-piperazineethane sulphonic acid]
pH 7.5.
Higher concentrations could often be tolerated if a recalibration was carried out in the
presence of the salt. The present method seems more effective in reducing the
interference by chromate than that reported by Rynasiewicz96 where only a 91%
recovery of hydrogen peroxide was obtained with 0.01M chromate. Aside from the
superior results obtained in this respect by the extraction process, it was felt that the use
of exchange resins with hydrogen peroxide, as described by Rynasiewicz, was potentially fraught with difficulties.
2.1.3 Extraction Method with Isobutanol.
A number of other solvents were tested in the extraction method in an attempt to increase sensitivity. In most cases, the separation of the aqueous and organic layers in the initial extraction was unsatisfactory. Of those tested, the most successful was
2-methyl -1-propanol (isobutanol). The results for the use of this solvent in the extraction analysis method with 5ml of Ti(IV) solution are shown in fig. 2.4.
Although this is a more sensitive process (cf. A=0.47 for lxlOPM 2H0 2 with
107 Absorbance Fig. 2.4 Isobutanol Extraction Analysis with 1ml of Ti(IV) Reagent. Ti(IV) of 1ml with Analysis Extraction Isobutanol 2.4 Fig. [H2O2] / lt r t n d dm d n t r lt / [H2O2] 3 "
108 isobutanol compared to A=0.265 with ethyl acetate), and the obtained calibration line is excellent, even in this case the separation between the aqueous and organic layers in the initial extraction is unsatisfactory. Attempts to use phase-separating filters failed, since the dissolved alcohol lowered the surface tension of the aqueous layer to such an extent that both components passed through the silicon-treated paper.
2.2 EXPERIMENTAL PROCEDURES.
U.V./Visible spectra were recorded on a Perkin-Elmer 551 and a Perkin-Elmer Lambda
2 Series PECSS spectrophotometer.
2.2.1 Analysis Procedure.
10ml of the sample to be analysed were vigorously shaken with 25ml of ethyl acetate in a stoppered tube for ca. 30 seconds. The two layers were allowed to separate, and 20ml of the upper (ethyl acetate) layer transferred to a separating funnel fitted with a Rotaflo tap. Either 5ml,3ml or, for dilute solutions (less than SxlO^mol dm-3), 1ml of the titanium(IV) solution was added. The mixture was shaken for ca. 15 seconds, the two layers allowed to separate, and the lower layer run directly into a 1cm pathlength glass cell (semi-micro type if 1ml of extractant was used). The absorbance at 407nm was measured against a blank of the titanium(IV) solution.
2.2.2 Analysis of Effect on Extraction of Interfering Species.
In each case, the analysis of 10ml aliquots of a standard solution of H2O2 was carried out in the presence of a variety of salts, some of which contained ions reported to interfere with the direct analysis (HPO 42-* C 1O 42-. Ni2+, Al3+, F-)93.
The concentration of salt in each case was increased and the sample analysed for H2O2 until the resultant error in absorbance at 407nm was ±2% relative to the expected value.
109 2.2.3 Use of Catalase to Decompose H 2O 2 in Sample.
This was used to give blank readings at 407nm by decomposing the H 2O2 prior to
analysis. Ca. 250 units of catalase were added to 10ml of the sample to be analysed,
and the standard extraction method of analysis carried out after 10 minutes.
2.3 MATERIALS AND SYNTHESES.
2.3.1 Titanium (IV) Sulphate Solution.
This was prepared by a modification of the procedure of Egerton100. lOg of AnalaR potassium titanium oxalate was heated in 55ml of concentrated sulphuric acid for at
least an hour until all effervescence had ceased. The solution was then cooled and
diluted to 500ml with water.
2.3.2 Hydrogen Peroxide.
Hydrogen peroxide for analysis was prepared by dilution of a stock solution of AristaR
H2O2 (BDH) that was standardised by iodometric titration.
2.3.3 Ethyl Acetate.
Ethyl acetate was of AnalaR purity, although similar results were obtained using GPR
grade solvent.
110 3. 'SIX-FOLD' ANALYSIS OF HYDROGEN PEROXIDE.
3.0 INTRODUCTION.
A great deal of time was spent attempting to develop a method for the routine colorimetric analysis of H2O2 based upon the iodometric method. This was done in the hope of both increasing the sensitivity of the method, and also to push Xmax to longer wavelength, out of range of expected interferences. Initial experiments indicated that in the pH range 4.5-6.8 (succinate buffer) molar extinction coefficients in the range 25-
32.000 mol-1 dm3 cm *1 at ^max = ca. 590nm. could be achieved. However, difficulty was found in attaining reproducibity of results on a day-to-day basis. This appeared to be primarily due to the instability of the soluble starch solutions over extended periods.
Attempts to produce stable starch solutions (ie. yielding reproducible results, on analysis of H2O2, over periods longer than 24 hours) by the addition of mercuric iodide as preservative, or by preparation of starch-glycerol solution, were unsuccessful for this application. Nor was replacement of the soluble starch with amylose successful, the amylose-15 - complex101 dropping out of solution upon formation.
An attempt was made to adapt the successful aspects of the iodometric analysis method to a process which made use of the six-fold increase in sensitivity produced by the following equations;
I- + 3Br2 + 3H20 ------► I0 3- + 6 HBr .....3.1
I 0 3- + 51- + 6 H+ ------3I2 + 3H20 ...... 3.2
The H2O2 sample to be analysed was back-extracted from ethyl acetate into buffered molybdate (which catalyses the oxidation of I-) and potassium iodide solution. Thus the iodide was oxidised to iodine. The organic layer was then shaken with sodium sulphite to reduce the iodine back to iodide. The iodide was oxidised to iodate in aqueous solution with acetate-buffered bromine water (eqn. 3.1). After removal of excess bromine and any residual ethyl acetate, potassium iodide (eqn. 3.2) and starch solution were added and the solution colorimetrically analysed at Xmax = 608nm.
Thus the I- in equation 3.2, which results from addition of potassium iodide to the solution, amplifies what is effectively half a mole of iodine in eqn 3.1 to produce 3 moles in the final analysis; essentially a six-fold amplification.
3.1 RESULTS AND DISCUSSION.
Despite initially encouraging results, it was found that the blank reading was extremely high (A = ca. 1.2). This was thought t o be due to the presence of iodine in the bromine used. It would seem that extremely pure bromine is necessary.
3.2 EXPERIMENTAL PROCEDURE.
U.V./Visible spectra were recorded on a Perkin-Elmer Lambda 2 Series PECSS spectrophotometer.
10ml of the H2O2 solution to be analysed was shaken with 30ml of ethyl acetate for ca.
30 seconds. 25ml of the upper organic layer was shaken with 5ml of buffered molybdate solution and lOOpl of 50% KI solution. This mixture was periodically shaken over 30 minutes. The upper organic layer was shaken with 3ml of 5X10-4M aqueous sodium sulphite solution for ca.1 minute, and to the lower aqueous layer was added a few drops of bromine water (in 0.1M acetate buffer adjusted to pH 4.5).
Excess bromine was removed by the addition of a few drops of saturated phenol solution, and the solution gently heated to remove any residual ethyl acetate. After cooling, the solution was left for one minute, then lOOpl of 50% KI solution and a few drops of starch added. The resultant blue colour was analysed colorimetrically at
Xmax=608nm .
112 3.3 MATERIALS AND SYNTHESES.
3.3.1 Hydrogen Peroxide.
Samples of hydrogen peroxide for analysis were prepared by appropriate dilution of a stock solution which was standardised periodically by iodometric analysis or titanium
(IV) solution.
3.3.2 Starch.
Soluble starch was AnalaR from BDH. (+)-Amylose was from Aldrich.
3.3.3 Other Chemicals.
All other chemicals used were of AnalaR grade.
113 4.PEROXIDASE-CATALYSED ANALYSES OF HYDROGEN
PEROXIDE.
4.0 INTRODUCTION.
There exist a wide variety of sensitive colorimetric analyses for hydrogen peroxide, all
of which involve the oxidation of a substrate or substrates by H2O2, catalysed by the
enzyme peroxidase. These oxidations yield in each case a highly coloured species that
can be determined spectrophotometrically.
Peroxidases, which have been found in many plant, animal and bacterial organisms102,
catalyse the oxidation of a range of electron donor molecules (DH2) by H 2O2, as
summarised in equation 4.1;
DH2 + H2O2 ------► 2H20 + D .....4.1
Although there are a variety of forms of peroxidase, that most often used in enzymatic
analysis is horse-radish peroxidase (HRP). This itself is believed to consist of several
"isoenzymic" forms that differ in structure, composition and physico-chemical properties (some studies estimate that there exist ca. 40 of these isoenzymic forms of
HRP)!°3. 104.
Horse radish peroxidase is highly specific for the hydrogen acceptor, which should be hydrogen peroxide, methyl or ethyl peroxide (the relative rates are quoted as 100, 0.17,
0.04104 ).
However, HRP is much less specific for the hydrogen donor molecule, with a wide range of donors including 4-hydroxydiphenyl-T,2-phenylenediamine (relative rate =
100), 1,3-phenylenediamine (rate = 3), hydroquinone (rate = 6 ), pyrogallol (rate =
0.6), 2-methoxyphenol (guaiacol, rate = 0.5), ascorbate (rate = 0.04) and NADH (rate
= 0.006).
The structure of HRP has been the subject of a great deal of study. It is thought that it consists of around 300 amino acid residues, arranged around an essentially spherical porphyrin (or protoheme) group, and has attached a number of carbohydrate residues, presumably to facilitate water solubility !°5. It has a molecular weight of ca. 44,000.
114 Fig. 4.1 below shows the proposed 3-dimensional structure of peroxidase105. The cylinders indicate a-helices, while the sheets represent so-called (3-pleated sheets. The numbered forks show the points of attachment of the carbohydrate molecules.
Fig. 4.1 3-D structure of Horse-radish peroxidase.
The fifth ligand around the Fe(HI) (the first four being the nitrogens on the porphyrin ring) is thought to be a histidine imidazole, providing a 'covalent link' between the heme and the protein. However the nature of the sixth ligand is uncertain. Water is often proposed in this position106, despite spectral data that suggests otherwise107.
It is known that cyanide, azide and fluoride all coordinate in this position, accounting for their inhibitory effect upon the action of HRP104.
% nmr studies show that the porphyrin has the structure shown in fig. 4.2 overleaf;
115 c h 2 II
CH=CH2
c h 3
COOH COOH Fig. 4.2 Structure of porphyrin ring in peroxidase.
Two Ca2+ions per HRP molecule have been detected, although these are thought to be primarily involved in structural stability rather then having any chemical function 108
The first application of stopped-flow methods to study the actions of enzymes allowed
Chance1Q9 to propose the reaction scheme shown below for the oxidation of the hydrogen donor DH2 catalysed by the enzyme E;
ki E + H2U2 ------Ei + H20 .... 4.2 (2e- transfer)
k 2 E i + DH2 e 2 + d h 2+ .... 4.3
k3 e 2 + d h 2 E + DH2 + .... 4.4
Ei has two oxidising 'equivalents', while E2 has only one. This latter can oxidise another DH2 donor molecule (as shown in eqn. 4.4) or can cause the first donor molecule to become fully oxidised.
Arguments exist for the direct oxidation of the heme group in Ei, with the simultaneous
116 formation of an iron-oxygen double bond110,111,112. Magnetic circular dichroism,
Mossbauer and resonance Raman scattering all indicate the presence of Fe (IV), with spin S=l, in both Ej and E 2. It is thought that the second oxidising equivalent in Ej is due to the occurrence of a cation-radical form of the radical ring, although the exact location of the radical is unknown.
The reaction scheme shown in fig. 4.3 below is consistent with the above and also with 11a experimental data . X represents the porphyrin ring.
-DH2,+
Fig.4.3 Reaction scheme for peroxidase-catalysed oxidation of substrate DH2.
The principal use of peroxidase is in analytical chemistry, where it has been found to be an important tool in the assay of a number of systems which produce H2O2 in the presence of O2 and a specific enzyme such as glucose oxidase. In this way, peroxidase has been shown to be important in the analysis of glucose, cholesterol, creatine, phospholipids, triglycerides, oxalate and uric acid, and is a very important component of modem diagnostic chemistry114.
However, HRP is not stable under a variety of conditions. Its optimum pH range is 6.0-
6.5, and its action is inhibited by alkaline pH!04 Similarly, it has a short lifetime in solution; at 25°C it is stable for 24 hours, at 60°C it is stable for up to 10 minutes.
Vigorous shaking of solutions of HRP is to be avoided since this leads to denaturisation of the enzyme.
4.1 RESULTS AND DISCUSSION.
4.1.1 Direct Analysis of H 2O 2 by Peroxidase-Catalysed Reactions.
Table 4.1 overleaf summarises the results obtained and observations made when attempts were made to analyse H 2O2 by a number of previously reported peroxidase- catalysed reactions. Although all the reactions considered are highly sensitive in the analysis of H2O2, none offer a combination of the advantages required of the desired substrate reaction; ie. high sensitivity, good stability and long wavelength. Most are excellent for fixed-time methods but would be difficult to adapt to the extraction process. Those reagent solutions involving MBTH display the problem of a blank that steadily increases after addition of the enzyme. Not only are similar disadvantages found with the ABTS reagent, but the coloured species is bleached rapidly in sunlight.
Several analyses also show marked deviation from Beer-Lambert behaviour.
118 Table 4.1 Summary of results for analysis of hydrogen peroxide by various peroxidase- catalysed colorimetric methods.
Substrates3 eeff/moI-idm3cm-i Xmax/nm
AAP + DCPS115 28,000 515 Non-Beer's Law. Low?i. Colour stable.
TOPS + AAPH6 18,000 550 Moderate Xmax. Low e. Colour stable.
ABTSH7 30,000 743 Non-Beer's Law. Sample and blank bleach in light.
AAP + PHBS ns 6,400 505 Low 8. Low7.max.
MBTH + TOPS 119 37,600 589 Blank high but fairly stable. Reagent unstable
MBTH + DMAB 120 30,000 590 Readings unstable due to change in blank.Unstable reagent. aAbbreviations; AAP - 4-aminoantipyrine; ABTS - 2,2'-azinobis(3-ethylbenz thiazoline-
6 -sulphonic acid), diammonium salt; DCPS - 2,4-dichlorophenolsulphonate; DMAB - 3-
(dimethylamino)benzoic acid; M BIH - 3-methyl-2-benzothiazolinone hydrazone;
PHBS - /?-hydroxybenzenesulphonate; TOPS - N-ethyl-N-sulphopropyl-ra-toluidine.
4.1.2 Analysis of H2O 2 by Peroxidase-Catalysed Reactions after
Extraction into Ethyl Acetate.
It was hoped that use of the ethyl acetate extraction into 1ml of combined substrate/ peroxidase reagent would lead to a similar increase in sensitivity as seen in the Ti(IV) analysis method, and would also avoid interference by coloured species.
Unfortunately, the several of the substrate reactions appeared to be partially extracted into the ethyl acetate layer. Comparison of the extinction coefficients for the direct and extraction methods where extraction was successful is given in table 4.2.
119 Table 4.2 Summary of results for analysis of hydrogen peroxide after extraction with ethyl acetate by various peroxidase-catalysed colorimetric methods.
Substrate AAP + DCPS ABTS AAP+PHBS MBIH + DMAB
^direct 28,000 30,000 6,400 30,000
^extract 40,000 34,000 6,500 35,000
Of those analyses studied, only the reaction between AAP and 2,4-dichlorophenol sulphonate exhibits an appreciable increase in sensitivity. Unfortunately this analysis has been previously shown not to obey the Beer-Lambert law in the direct analysis of
H 20 2.
4.2 EXPERIMENTAL PROCEDURES.
U.V./Visible spectra were recorded on a Perkin-Elmer Lambda 2 Series PECSS spectrophotometer.
4.2.1 Direct Analysis of H20 2.
10(il aliquots of H20 2 were added to 2ml of water in a 1cm path length quartz uv cell. lm l of the reagent solution was added to this cell and also to another matched cell containing only 2ml of water. The cells were stoppered and gently inverted for ca. 30 seconds. The sample cell was placed in the uv/vis spectrophotometer, with the cell not containing the H20 2 as a reference, and the absorption spectrum monitored over 30 minutes. The maximum value of A and ^max were recorded.
4.2.2 Extraction Analysis of H20 2.
10ml of H20 2 solution were shaken with 25ml of ethyl acetate in a stoppered tube for
120 ca. 30 seconds and the two layers allowed to separate. 20ml of the upper layer was
taken and added to 1ml of the reagent solution in a separating funnel fitted with a
Rotaflo tap. The funnel was gently inverted for 5 minutes and the two layers allowed to
separate. The lower aqueous layer was run directly into a 1cm path length glass semi
micro cell (internal width 4mm) and the cell placed in the uv/vis spectrophotometer. The
absorption spectrum was measured against a reference of the reagent solution.
4.3 MATERIALS AND SYNTHESES.
4.3.1 Substrates. I
2,2'-Azinodi(3-ethylbenzthiazoline-6-sulphonic acid) (ABTS), 4-aminoantipyrine
(AAP), 3-(dimethylamino)benzoic acid (DMAB), N-ethyl-n-sulphopropyl-m-toluidine
(TOPS), p-hydroxybenzene sulphonate(FHBS) and 3-methyl-2-benzothiazolinone
hydrazone (MBTH) were purchased from Sigma and used without further purification.
A 2% stock solution of 2,4-dichlorophenolsulphonate (DCPS) was prepared as
described by Barham and Trinder115.
4.3.2 Hydrogen peroxide.
Samples of H2O2 for analysis were prepared by dilution of a stock solution prepared
from AristaR hydrogen peroxide.
4.3.3 Ethyl Acetate.
This was HiPerSolv grade from BDH.
4.3.4 Peroxidase.
Peroxidase (from horse-radish) was obtained from Sigma.
121 4.3.5 Buffers.
All buffering agents were of AnalaR grade. Adjustment to the required pH for MOPS buffer was effected by addition of a saturated solution of AnalaR sodium hydroxide.
4.3.6 Solutions.
A number of different combinations of substrates in solution were employed in both the direct and ethyl acetate extraction analysis of H 2O2. The reagent solutions are described below;
i. 0.2mM in MBTH, lOmM in DMAB, 0.1M phosphate buffer
pH 6.5, 5U/ml peroxidase.
ii. 0.06% ABTS in 0.25M Na2HP04 adjusted to pH 4.4 with citric
acid. 5U/ml peroxidase. iii. 0.5mM in AAP, 20mM in HBS, 0.1M phosphate buffer pH 6 .8, lOU/ml peroxidase. iv. a. 1.2mM in AAP, 0.1M phosphate buffer pH 6.9.
b. 0.5% in DCPS, 0.1M phosphate buffer pH 7.0, 20U/ml peroxidase.
v. a. 2mM in TOPS, 0.1 MOPS buffer pH 7.0, lOU/ml peroxidase,
b. 2mM in AAP, O.l MOPS buffer pH 7.0.
vi. 0.2mM in MBTH, 2mM in TOPS, 0.1M phosphate buffer pH 7.0,
lOU/ml peroxidase.
In iv. and v. the final reagent solution was prepared by adding equal volumes of solutions a. and b. together. 5. SPECTROPHOTOMETRIC DETERMINATION OF HYDROGEN
PEROXIDE AND GLUCOSE WITH LEUCO PATENT BLUE VIOLET.
5.0 INTRODUCTION.
Since none of the established peroxidase-catalysed methods for the determination of hydrogen peroxide could be adapted to the ethyl acetate extraction technique, it was decided to consider, and possibly synthesise a new substrate for this purpose.
The widespread use of triphenylmethane dyes in the foodstuff and textile industries has been curtailed by their determination as cancer suspect agents. The dyestuffs have characteristically intense bands (e > 50,000 mol -1 dm3 cm-1) at long wavelength (Amax
> ca. 600nm). Thus they meet two of the requirements of the desired substrate reaction.
The reduced leuco base form of one particular species, Crystal Violet, fig 5.1, was used by Mottola et al in the spectrophotometric determination of H2O2 in exactly the kind of peroxidase-catalysed reaction desired121.
Fig. 5.1 Crystal Violet.
However, although a molar extinction coefficient of 75,000 mol-1 dm3 cm -1 was recorded at ^ a x = 596nm for the coloured species, a number of problems were
123 encountered with the substrate, particularly non-zero intercept on the calibration curve
for the determination of H2O2.
In view of the fact that the substrate reaction is ideally to be adapted to the ethyl acetate
extraction method, the presence of water-solubilising substituents on the substrate molecule, particularly sulphonate groups, would be particularly desirable. These, as electron-withdrawing groups, would also serve to stabilise the reduced form of the
substrate. Since many of the triphenylmethane dyestuffs have enjoyed commercial use,
there is little evidence that the oxidised forms will require similar stabilisation.
The dyestuff Patent Blue Violet is discussed in the literature as two similar dyes, C.I.
42045 and 42051. Structurally, they differ with regard to the substituents on the
aromatic rings, C.I. 42045 (sometimes called Sulphan Blue), fig. 5.2, is 2,4- disulphonated;
Fig. 5.2 Patent Blue Violet C.I. 42045
while C.I. 42051 is 2,4-disulphonate-5-hydroxy substituted (Acid Blue 3). There also exists the 2,5-disulphonated isomer, so-called Isosulphan Blue. Although commercial samples of Patent Blue Violet contain solely the disulphonated forms (ie. no Acid Blue
3), variance in appearance and erratic results of the leuco form in the analysis of H2O2 indicate that some might contain mixtures of the 2,4- and 2,5-disulphonated
124 compounds. For example, two samples of Patent Blue Violet obtained from Aldrich
differed markedly in colour, one was violet, similar in colour to the Sigma sample, the
other was olive green. The leuco form produced from the latter sample gave
absorbances ca. 50% lower than those found with the reduced Sigma compound.
Dye content of commercial samples of Patent Blue Violet similarly varies between 50
and 75%, the predominant contaminant being sodium sulphate. For our purposes,
wherever Patent Blue Violet is referred to in this work, reference is being made to the
compound C.I. 42045, the 2,4-disulphonated species.
5.1 RESULTS AND DISCUSSION.
5.1.1 Direct Spectrophotometric Determination of H 2O 2 with LPBV.
Figs. 5.3 and 5.4 show plots of absorbance against the final concentration of hydrogen
peroxide for two different concentration ranges. Beer’s law is closely obeyed, and the
effective value of e^ax is I.OO5 XIO5 mol*1 dm3 cm-i, which is the highest yet recorded.
The absorbance values also show very good stability with time (±<1% over 3 hours,
±<3% over 24 hours) and the long wavelength for maximum absorption (639nm) tends
to reduce the possibility of interference from other substances present in the samples.
Experiment showed that the reagent is fully operative in the pH range 6 3 -1 3for both
phosphate and PIPES buffer systems.
A number of problems were encountered in reaching the optimum conditions for the
analysis. Initially non-zero intercepts were obtained for the calibration lines. This has
been noted elsewhere by Mottola et al in work on Leuco Crystal Violet^i, who
suggested addition of a small amount of H2O2 to the reagent solution. Although this had a beneficial effect, the problem was not entirely eradicated until the work of Capaldi and Taylori22 was noted. These workers found that the intercept varied according to the concentration of peroxidase in their studies of the MBTH reagent. By decreasing the
125 Absorbance Fig. 5.3 Direct Analysis with LPBV Reagent - High Concentration Range. Concentration High - Reagent LPBV with Analysis Direct 5.3 Fig. [H 2 0 2] / lff^nol dm lff^nol 2]/ 0 -3
126 Absorbance Fig. 5.4 Direct Analysis with LPBV Reagent - Low Concentration Range. Concentration Low - Reagent LPBV with Analysis Direct 5.4 Fig. [H 2 O 2 ] / 10 mol dm 3 dm mol 10 / ]
127 amount of peroxidase in the LPB V solution to 4U/ml, the problem of non-zero intercepts was overcome.
High concentrations of LPB V initially used in the reagent solution similarly produced erratic results at low concentrations of 2HO2.
5.1.2 Determination of H 2O2 after Ethyl Acetate Extraction.
Fig. 5.5 is a plot of absorbance against the concentration of hydrogen peroxide in the
10ml samples taken. Appreciably higher values of absorbance are obtained than with the direct method. More importantly, the extraction can be used in the presence of substances which interfere with the direct determination. These include "poisons" for peroxidase, materials which absorb at 639nm and buffers and other electrolytes at high concentrations. For example, the extraction procedure gave results accurate to within
±2 % for the analysis of hydrogen peroxide in the presence of one of the following; hydroxylamine (0.01M, pH 2), potassium fluoride (0.5M), methylene blue (lxlCMM) and PIPES buffer (0.5M, pH 7). All of these interfere fully or partially with the direct analysis method.
Use of higher concentrations of LPB V in the reagent solution in initial analyses led to the low optical densities with concentrations of 2HO2 < 2xl0-6mol dm-3.
Appreciable concentrations of H2O2 were found in commercial samples of ethyl acetate, analysed with the LPBV reagent, table 5.1. Tests with purified ethyl acetate to which known amounts of H2O2 had been added indicated that the range of absorbances recorded corresponded to concentrations of hydrogen peroxide (or hydroperoxides) in the range ca. 0.5-lxl0- 7M. No absorbance was observed in the absence of peroxidase. Absorbance Fig. 5.5 Ethyl Acetate Extraction Analysis with LPBV Reagent. LPBV with Analysis Extraction Acetate Ethyl 5.5 Fig. [H 2 O 2 ] /1 0 1mol dm 3 1moldm 0 /1 ] Table 5.1 Summary of Results obtained with Commercial Samples of Ethyl Acetate.
Supplier Grade Absorbance at 639nma
BDHGPR 0.0476
BDH AnalaR 0.1075
BDH AristaR 0.1117
Aldrich HPLC 0.0627
BDH 'Purified' AnalaRb 0.01555 aAbsorbances recorded after shaking 25ml of ethyl acetate with 2ml of LPBV/ peroxidase direct analysis reagent, against reagent blank. bPurified as detailed in
Materials and Syntheses section.
5.1.3 Glucose Determination. p-D-glucose reacts with dissolved atmospheric oxygen in the presence of glucose oxidase to give gluconic acid and hydrogen peroxide. Mutarotase catalyses the conversion of a-D-glucose to p-D-glucose. Fig 5.6 shows a plot of absorbance against the final concentration of glucose when the LPBY-glucose reagent was used.
The effective value of e^x is I.OI4XIO5 mol-1 dm3 cm-1, which is in good agreement with that obtained in the direct hydrogen peroxide analysis. Problems of a non-zero intercept of the calibration line in this case could only be overcome by addition of a very small amount of glucose to the reagent solution. Even with this addition, blank readings were only ca. 0.05.
5.1.4 Use of Other Leuco Dyes.
In addition to Patent Blue Violet, a number of other anionic triphenylmethane dyestuffs are commercially available, and also Methyl Green, a dicationic complex. Of these dyestuffs, Light Green SF Yellowish, Fast Green FCF, Erioglaucine, Alphazurine A
130 Absorbance Fig. 5.6 Direct Analysis of Glucose with LPBV-Glucose Reagent. LPBV-Glucose with Glucose of Analysis Direct 5.6 Fig. Glucose / 10 6mol dm 6mol 10 / Glucose "3 and Methyl Green were reduced in a similar manner to LPB V, but in contrast the leuco forms of these compounds were all were oxidised quite rapidly by atmospheric oxygen.
5.2 EXPERIMENTAL PROCEDURES.
U.V./Visible spectra were recorded on a Perkin-Elmer Lambda 2 Series PECSS spectrophotometer.
5.2.1 Direct Analysis of H 2O 2 with LPBV.
2ml of the sample to be analysed and 1ml of the LPBY reagent were added together in a lcm path length cell. After 15 minutes the absorbance at 639nm was measured against a similar reference cell containing 2ml of water and 1ml of the LPBV reagent.
5.2.2 Extraction Analysis of H 2O 2 with LPBV.
10ml of the sample to be analysed was vigorously shaken with 25ml of purified ethyl acetate in a stoppered tube for ca. 30 seconds. After separation of the layers, 20ml of the upper layer was transferred into a separating funnel fitted with a Rotaflo tap. 1ml of the LPBV extraction reagent was added and the mixture gently inverted for 5 minutes, or 10 minutes for concentrations >5xl0*6 M. The two layers were allowed to separate, and the lower layer run into a lcm path length glass semi-micro cell. The absorbance at
639nm was measured relative to a blank obtained by following the same procedure, but with 10ml of purified water in place of the hydrogen peroxide solution.
5.2.3 Direct Analysis of Glucose with LPBV.
The procedure in this case was as for the direct analysis of hydrogen peroxide, but 1ml of the LPBV glucose reagent was added to 2ml of the sample, and the absorbance at
639nm measured after 30 minutes.
132 5.3 MATERIALS AND SYNTHESES.
5.3.1 Leuco Patent Blue Violet.
Patent Blue Violet (Sigma P-6396, 3.0g) and anhydrous sodium carbonate (1.7g) were stirred in water (30ml) under nitrogen or argon. Sodium hydrosulphite (85%, 1.4g) was added, and stirring continued for 30 minutes. Sodium chloride (9.0g) was added, and the solution cooled to 4°C. The solid which precipitated out was centrifuged off in air, and washed with an almost saturated solution of sodium chloride (4x30ml). The solid residue was stirred with water (30ml) and the solution centrifuged. The clear supernatant was stirred at 0-4°C, and sodium chloride (9.0g) added. After 10 minutes the mixture was centrifuged and the residue washed with almost saturated sodium chloride solution (4x30ml). After drying in vacuo in the dark, the solid was extracted into ethanol (3x25ml), the solvent removed and the product dried in vacuo in the dark to yield a pale blue solid; Yield 1.2g.
The compound is quite stable if kept in a dark bottle. Analysis indicated the presence of small amounts of sodium chloride, water and ethanol but otherwise the compound was pure by iH N.m.r. (1H N.m.r. (D 20) 270 MHz; 8 0.92 (12H, m, CH3), 3.09 (8H, m,
CH2), 6.50 (1H, CH), 6.68 (4H, d, H3.5), 6.95 (4H, d, H2.6), 7.29 (1H, d, H*'),
7.67 (1H, dd, H5'), 8.33 (1H, d, H3').
The large 8 value for H3' indicates that the sulphonate groups are in the 2' and 4' positions as in fig. 5.7
Fig. 5.7 Leuco Patent Blue Violet. 5.3.2 Hydrogen Peroxide.
Due to the possible rapid decomposition of the very dilute solutions of 2HO2 analysed by this method, the samples for analysis were freshly prepared in each case by dilution of 10(il aliquots of a stock solution of hydrogen peroxide, which was standardized by iodometric titration or with titanium(IV) sulphate solution.
5.3.3 Ethyl Acetate.
Problems were encountered due to the greater sensitivity of the method for the analysis of H2O2 compared to the Ti(IV) extraction method. Thus when 25ml samples of ethyl acetate from BDH and Aldrich (G.P.R., AnalaR, Aristar and HPLC grades) were inverted for 5 minutes with 2ml of the LPB V/peroxidase reagent, the aqueous layers gave absorbances in the range of 0.05-0.12 at 639nm.
Purification of the ethyl acetate was performed by washing 500ml six times with an equal volume of water, then stirring the washed solvent for three hours with ca. 50ml of the direct hydrogen peroxide analysis reagent detailed above. After separation from the aqueous layer the ethyl acetate was stirred with ca. 50g of anhydrous sodium sulphate, filtered, and slowly passed through a column of oven-dried 3A molecular sieves. Analysis of this purified solvent in the manner outlined above gave absorbances of <0.005.
5.3.4 Enzymes.
Peroxidase (from horse-radish), glucose oxidase and mutarotase were obtained from
Sigma.
5.3.5 Buffers.
Disodium hydrogen orthophosphate and sodium dihydrogen orthophosphate were of
AnalaR grade, and sodium hydroxide of Aristar grade. Piperazine-N,N'-bis-2-ethane sulphonic acid (PIPES) was obtained from BDH or Aldrich. With certain samples of
134 PIPES erratic results were obtained and it proved necessary to pass the prepared buffer solution down a column containing Dowex chelating resin (Sigma) before use. The distilled water used was also passed down a similar column.
5.3.6 Solutions.
A 0.05M phosphate buffer (pH 7.0) was prepared by dissolving 1.37g of
Na2HP0 4 .12H2O and 0.38g of NaH2P0 4 .2H2 0 in 200ml of water and diluting to
250ml. In cases where interference was likely to be experienced with the phosphate buffer (eg. the presence of certain metal ions), a PIPES buffer was prepared by dissolving 3.78g of PIPES in 200ml of water, adjusting the pH to 7.0 with saturated sodium hydroxide and diluting to 250ml. The buffer solution was then passed down a column of Dowex chelating resin (typically 400mmxl 1mm) obtained from Sigma, rejecting the first 50ml of eluant. The reagent for direct analysis of hydrogen peroxide was prepared by dissolving 2.9mg of LPBV (final concentration 5xlO* 5 M) and 400 units of peroxidase in 100ml of the phosphate or PIPES buffer, using gentle stirring or inversion.
For the hydrogen peroxide extraction reagent, l.Omg of LPBV and 400 units of peroxidase were dissolved in 100ml of buffer.
When stored in a dark bottle at room temperature, both reagents were stable for at least
24 hours.
For the glucose analysis reagent, 2.9mg of LPBV, 400 units of peroxidase, 3000 units of glucose oxidase and 3000 units of mutorotase were dissolved in 100ml of phosphate buffer by gentle inversion, 1ml of 2xlO-4M glucose solution added and the solution left for 30 minutes. In the absence of added glucose, non-zero intercepts were encountered.
The reagent was stored in a dark bottle.
135 5.3.7 Glucose. a-D-Glucose was an A.C.S. reagent obtained from Aldrich.
136 GENERAL CONCLUSIONS FOR ANALYSES.
The spectrophotometric determination of hydrogen peroxide after extraction with ethyl acetate, using titanium (IV) sulphate solution, is a potentially very useful adaptaton of the standard method. Its main advantage is that it allows the use of the colorimetric technique where direct analysis would be interfered with by species present in the solution to be analysed. This is illustrated to good effect by its use in the work on production of H2O2 by tannic acid, gallic acid and other related compounds in aqueous solution, found in section HI of this thesis. The highly-coloured nature of the reaction solutions would otherwise preclude the use of the Ti(IV) colorimetric method.
Its utility in excluding chemical interferences experienced in the standard analysis method has also been demonstrated, particularly in the case of chromate interference.
The previous method for removing this species prior to analysis produced inferior results to the extraction technique, and involved the passing of the samples to be analysed down an exchange column. This is viewed by us as particularly difficult for a compound as easily decomposed as H2O2.
Although there is a marked increase in sensitivity for the ethyl acetate extraction method compared to direct analysis with Ti(IV) solution ( eeff for the extraction technique was calculated as 1140 mol-1 dm3 cm-1, compared to the quoted value of 8=770 mol-1 dm3 cm-1 for the yellow peroxytitanium species formed), the sensitivity of the technique is still limited by this low value of £ for the Ti(IV) reagent, and the value of ^max» 407nm, is prone to interference from compounds which manage to get through the two extractions into the final Ti(IV) solution. Attempts to increase the sensitivity by use of a solvent with a higher value of partition coefficient [H2O2]soiv/[H 2Q 2] H20 were unsuccessful due to poor separation of the solvent/aqueous layers in the initial extraction.
In an attempt to find a more sensitive analysis system that could be adapted to the
137 extraction method, and where the value of A^ax was appreciably higher, a number of
variations on the standard starch-iodide determination method were explored. However,
it was found that the formation of the blue starch-15 - species was prone to a range of
interferences, including temperature, the type of starch used and dissolution of ethyl
acetate into the aqueous layer. However, one variation of the standard reaction,
employing a six-fold amplification factor in a series of extractions, shows particular
promise, although our studies were ironically curtailed by amplification of the blank
reading due to iodine present in the bromine used in one stage of the analysis. Use of
AristaR or suitably purified bromine offers the prospect of further work on this method.
Attempts to adapt established peroxidase-catalysed reactions for the
colorimetric determination of H2O2 to the extraction technique were unsuccessful. The
criteria we had set for the suitability of the substrate reaction, aside from the obvious
requirement that neither substrates nor coloured species be extracted into the ethyl
acetate, were a high value of Emax for the coloured product (ideally £>50,000 mol-1
dm 3 cm-1), a high value of A^ax (ideally >600nm), and reasonable stability of the
coloured product once formed. Although most of the methods in the literature are
excellent for direct fixed-time colorimetric determination of H2O2, none of the reactions
matched these requirements.
Development of leuco Patent Blue Violet (LPBV) as a reagent for the colorimetric
determination of H2O2 required a great deal of purification of not only the substrate, but
also the distilled water and buffers employed in the reagent solution. The peroxidase-
catalysed direct analysis of H 2O2 with LPB V produces a coloured product of eeff =
I.OO5XIO5 mol-1 dm 3 cm-1 at A^ax = 639nm. The maximum absorbance of the coloured
species also exhibits excellent stability; to within 1% over 3 hours, and 3% over 24 hours.
138 Adaptation of the LPB V analysis to the ethyl acetate extraction method requires not only slight modification of the reagent solution (a reduced concentration of substrate), but also surprisingly purification of the ethyl acetate to remove residual concentrations of peroxide in the range 0.5-lxl0 -7 mol dm-3. The fact that the oxidation of the LPBV reagent by these impurities proceeded very rapidly led us to believe that hydrogen peroxide, rather than some other peroxy species, was present (see introduction to peroxidase-catalysed analysis section for relative activity of peroxidase to H2O2 and other peroxy species).
Aside from the appreciably higher absorbances obtained with the extraction technique, analysis in the presence of a number of species that interfere with the direct method is possible. These include both 'poisons' for peroxidase, such as hydroxylamine, as well as materials which absorb at 639nm (eg. methylene blue) and buffers and other electrolytes in high concentration (eg. 0.5M potassium fluoride, 0.5M PIPES, pH 7).
Further adaptation of the direct LPBV method to the analysis of glucose required the preparation of a complex mixture of enzymes to perform a chain of reactions in solution. For this reason, the method is perhaps more prone to interferences than the relatively straightforward analysis of H 2O2.
The stability of the leuco form to aerial oxidation in comparison to the reduced forms of other triphenylmethane dyestuffs cannot be explained simply in terms of substituents or steric effects. It might be expected that the presence of electron withdrawing sulphonate groups on the molecule would serve to stabilise the reduced, leuco, form in the case of
Patent Blue Violet. However, this cannot be the only reason for the relative stability of
LPBV, since Alphazurine A, where one of the ethyl groups on each nitrogen of LPBV is replaced by a benzyl group, undergoes rapid oxidation when reduced with dithionite and exposed to air. In both molecules, the sulphonate groups are in the 2,4- positions, so some other factor must be involved.
139 III. THERMAL AND PHOTOCHEMICAL PRODUCTION OF
HYDROGEN PEROXIDE FROM DIOXYGEN AND TANNIC
ACID, GALLIC ACID AND RELATED COMPOUNDS IN
AQUEOUS SOLUTION.
6.0 INTRODUCTION.
As discussed in the general introduction, hydrogen peroxide occurs naturally in the environment. It is a common component of cloudwater, being formed by photochemical processes involving atmospheric trace gases41-123, and probably plays an important role in the oxidation of SO2 to H2SO4 in the production of 'acid rain'124.
It is also present in natural surface waters, with concentrations typically in the range
1-I00xl0-8 mol dm*3 46>125. Here, the H 2O2 is thought to arise mainly from the absorption of sunlight by organic materials such as humic substances. Exposure of natural waters to sunlight or UV in the laboratory normally causes a very considerable increase in the concentration of H202125-126. It has been suggested 125 that either the light-absorbing materials generate free electrons, or that the excited species transfers an electron directly to O2 to yield the O2- ion, which is a common precursor of2 HO2 by the well known reaction;
H 0 2 + 0 2------H 0 2- + 0 2 .....6.1
Very recently, it has been shown that H 2O2 occurs in some ground-waters126, where photochemical production is excluded.
In I860128, Schonbein reported that H 2O2 was formed when weakly alkaline solutions of tannin (tannic acid), pyrogallol, gallic acid (3,4,5-trihydroxybenzoic acid) and haemotoxylin were shaken in air. Tannins of various types occur in almost all plant and vegetable material. The tannic acid from chemical suppliers is a gallotannin (normally
140 Chinese tannin from oak galls), and is a mixture of gallate esters of glucose, containing mono-, di- and tri-galloyl groups, with ca. 9-10 gallic acid residues per glucose molecule129, fig. 6 .1;
Fig. 6.1 Tannic acid.
Tannic acid and the other compounds studied by Schonbein all contain the catechol (1,2- dihydroxybenzene) moiety. In 1955 Brackman and Havinga 130 showed that H2O2 in ca. 40% yield was obtained when O 2 was bubbled through an alkaline solution of catechol, provided that morpholine was present to trap1 ,2-benzoquinone as soon as it was formed, and prevent its further reaction with H2O2.
In the work detailed herein, both the thermal and the photochemical production of H2O2 from air or oxygen and aqueous solutions of tannic acid, gallic acid and a variety of related compounds has been studied quantitatively. One of the aims of the work was to elucidate, using relatively simple molecules, the structural features required for the production of H2O2.
141 THERMAL PRODUCTION OF HYDROGEN PEROXIDE.
6.1 RESULTS AND DISCUSSION.
6.1.1 Tannic Acid.
The hydrogen peroxide produced when air was bubbled through aqueous solutions of tannic acid at various pH values is shown in fig. 6.2. The molecular weight of the tannic acid was taken as 1701, corresponding to one glucose molecule and ten gallate residues, so the maximum yield of H 2O2 is ca.0.33 moles per gallate residue. When pure O2 was used instead of air (pH 8.5), the maximum yield was ca. 0.40 moles. The rate of production of H2O2 increased very markedly as the pH was increased. In view of the complex nature of "tannic acid", further measurements were made using gallic acid.
6.1.2 Gallic Acid.
The results obtained in air are shown in fig. 6.3. The maximum yield of H 2O2 in the pH range 7.5-8.5 is ca. 1 mole H 2O2 per mole of gallic acid. When O2 was used in place of air, the rate of production of 2HO2 was considerably increased (at pH 8.5 the initial rate was greater by a factor of ca. 3.3), and the maximum yield of H 2O2 reached a value of ca. 1.4 moles. It has been known for many years that oxidation of gallic acid in alkaline solution by air or O2 gives the highly coloured compoundgalloflavin131.i32? fig. 6.4 overleaf.
The mechanism involved is unknown, but conversion of two molecules of gallic acid into one molecule of galloflavin involves the loss of2H "C6 O2". If all the H atoms were involved in the formation of H2O2, then the maximum possible yield of H 2O2 would be
1.50 moles per mole of gallic acid. The greater yields obtained with O 2 than air are presumably due to faster production of H2O2 which allows less time for its subsequent removal by reaction or catalysed decomposition.
142 7
Fig. 6.2 Concentrations of H2O2 produced at various times for 2x10‘4 mol dm"3 aqueous solutions of tannic acid saturated with air at (a) pH 8.58, (b) pH 7.99, (c) pH
7.51, (d) pH 7.00.
143 12
Fig. 6.3 Concentrations of H2O2 produced at various times for lxlO'3 mol dm"3 aqueous solutions of gallic acid saturated with air at (a) pH 8.52, (b) pH 8.07, (c) pH
7.48, (d) pH 7.00.
144 o
The pK values reported for gallic acid (GH4) at 20°C are 4.26, 8.70 and 11.45133. For
the solutions studied here, the carboxylic acid group is effectively completely ionized,
and it is possible to calculate the concentrations of the species GH22- and GH3-, where
one and two respectively of the hydroxylic protons are removed. Table 6.1. below
shows the approximate initial rates of production of2 HO2, and the corresponding concentrations of GH22- and GH3-.
Table 6.1 Initial rates of production of H2O2 as a function of pH for lxlO-3 mol dm-3 aqueous solutions of gallic acid saturated with air.
pH 7.00 7.48 8.07 8.52
Initial Rate/l(Hmoldm-3m in -i 0.81 2.41 9.36 27.5
Relative Initial Rate 1 2.9 11.5 33.9
Relative [GH22-1 1 3.2 9.7 20.3
R elative [G H 3-] 1 8.8 114 673
It is clear that, over the pH range 7-8.5, the predominant reactive species is GH22-.
This contrasts with the situation found for tetramethylhydroquinone, which reacts
1 45 quantitatively with O 2 in alkaline EkO/EtOH solution to give tetramethylquinone and
H2O2. Here, the reactive species is apparently the dianion, where both hydroxylic
protons have been removed134.
6.1.3 Pyrogallol (1,2,3-Trihydroxybenzene) and 1,2,4-
Trihydroxybenzene.
H2O2 is produced even more rapidly with pyrogallol than with gallic acid, and higher
yields are obtained, fig. 6.5. At pH 8.0 with 1,2,4-trihydroxybenzene, the reaction was
too fast to be studied conveniently. The results obtained at the much lower pH of 6.00
are shown in fig. 6.5. In spite of the very high reactivity of 1,2,4-trihydroxy benzene
towards O2 (0.37mol) is less than that produced by gallic acid and pyrogallol. Since
1,2,4-trihydroxybenzene itself does not react with H 202, presumably attack by one or
more of the reaction products is involved.
6.1.4 Dihydroxybenzenes.
i. Catechols.
Fig. 6.6 shows the production of H2O2 by catechol, 3-O-methyl gallic acid and 3,4-
dihydroxybenzoic acid. The observed formation of H 2O2 by catechol alone, in
contradiction of an earlier report130, may be partly associated with the low
concentration of substrate used in this work. The yields of H2O2 are much lower than
those found with gallic acid or pyrogallol. This means that the observed initial rates of reaction will be considerably lower than the true values. Nevertheless, it is clear that for
the carboxylic acids, the actual reactivities towards oxygen are in the order gallic acid >
3-O-methylgallic acid > 3,4-dihydroxybenzoic acid, and also that 1,2,4-trihydroxy
benzene and pyrogallol are considerably more reactive than catechol. These differences cannot be accounted for by variations in the value of pK for ionization of the first OH
146 time/min
Fig. 6.5 Concentrations of H2O2 produced at various times for air-saturated aqueous solutions of (a) lxlO' 3 mol dm"3 pyrogallol at pH 8.00, (b) lxlO"3 mol dm"3 gallic acid at pH 8.07, (c) lxlO' 3 mol dm"3 1,2,4-trihydroxybenzene at pH 6.00, (d) 2xl0 "3 mol dm ' 3 catechol at pH 8.50.
147 3.00 (6 )
a Fig. 6.6 Concentrations of H2O2 produced at various times for air-saturated 2x10 mol dm aqueous solutions of (a) catechol at pH 8.50, (b) 3-O-methyl gallic acid at pH 8.58, (c) 3,4-dihydroxybenzoic acid at pH 8.50.
148 proton. (The actual values at 25°C are gallic acid 8.64±0.09135, 2,4-dihydroxybenzoic
acid 8.79±0.07136, pyrogallol 8.96±0.02137, 1,2,4-trihydroxybenzene 9.02137 and
catechol 9.30±0.01138).
With "tiron" (4,5-dihydroxy- 1,3-benzenedisulphonate), which contains two electron-
withdrawing sulphonate groups, negligible amounts of H2O2 were produced (2xlO- 3M,
air, pH 8.5), and also, in contrast to the other catechols studied, the solution remained
almost colourless.
ii. Resorcinols.
Under similar conditions, 3,5-dihydroxybenzoic acid, resorcinol and phloroglucinol
also gave negligible amounts of H2O2, whilst a phenol (syringic acid, 4-hydroxy-3,5-
dimethoxybenzoic acid) produced only 0.013 moles H 2O2 after 300 minutes.
For formation of hydrogen peroxide from oxygen and a catechol, the possible
mechanistic pathway shown in fig. 6.7. can be written;
1 further products highly coloured
02‘ + H+ H02
02‘ + H02 ---- ► H02" + 02
Fig. 6.7 Possible mechanism for the formation of H2O2 from oxygen and a catechol.
149 This mechanism can only account for the formation of 1.0 moles of2O H2 for each
mole of the catechol. The higher yields with gallic acid and pyrogallol suggest that
H2O2 is also produced in the further steps leading to the final product.
The differences in reactivity for substituted catechols can be qualitatively interpreted in
terms of the electron density on the O - of the anion. The COO- group is weakly
electron-withdrawing as shown by the Hammett parameters (am=0.02,ap=0.1 1)139
which accounts for the greater reactivity of pyrogallol compared to gallic acid.
Conversely OH and OCH 3 groups increase the electron density when in a para (and
also ortho) position (OH, am=0.13, ap=-0.38, OCH3, cym=0.10, The lack of reactivity of resorcinols towards O2 is presumably because they cannot give a stabilised semiquinone radical-ion. As mentioned earlier, the main reactive species in hydroquinones seems to be the dianion. This apparent difference from the catechols may be partly due to the difficulty, for electrostatic reasons, of losing two protons from ortho OH groups (at 25°C,2 is pK 13.3±0.3138 for catechol, 11.50140 for hydroquinone). Finally, the thermal reaction of O2 with catechol moieties in humic substances could be involved in the reported occurrence of H2O2 in groundwaters127. 150 PHOTOCHEMICAL PRODUCTION OF HYDROGEN PEROXIDE. 6.1.5 Catechols. Aqueous solutions of the catechols (4xlO*3M, 0.05M succinate buffer) were saturated in air, and irradiated at a wavelength of 254nm. This was done at pH 6.0 or below, where the thermal reaction was normally negligible. The quantum yields Q given (see physical measurements) ( molecules H2O2 form ed) ------x 100% (quanta incident on the solution) are effective ones only, being those observed after 30 minutes irradiation, with the quanta incident upon the solution lying in the range 2.5-3.3xl0-7 mol s-i. For gallic acid and pyrogallol respectively, the pH dependence of the quantum yields are tabulated in tables 6.2. and 6.3. The decrease in Q as the pH was lowered may arise from protonation of the excited species, leading to lower reactivity, or possibly to changes in reactions subsequent to the primary step. Table 6.2 Variation of Quantum Efficiency with pH for an Air-Saturated 4x10-3M Aqueous Solution of Gallic Acid. 30 Minutes Irradiation at 254nm. pH Efficie 6.02 2.40 4.98 1.72 4.80 1.48 4.61 1.32 4.42 1.03 4.04 0.67 3.91 0.63 3.85 0.54 3.80 0.47 3.70 0.43 3.60 0.34 151 Table 6.3. Variation of Quantum Efficiency with pH for an Air-Saturated 4xlO*3M Aqueous Solution of Pyrogallol. 30 Minutes Irradiation at254nm. pH Efficie 6.00 2.18 5.60 1.96 5.20 1.85 5.00 1.60 4.80 1.54 4.40 1.33 4.00 1.13 3.60 1.02 The effective quantum yields for a wide variety of catechols at pH 6.00±0.01 are given in Table 6.4. In most cases small but definite amounts of H 2O2 were produced. In spite of the comparatively low quantum yields, appreciable concentrations of H2O2 were formed, for example with gallic acid after 30 minutes irradiation, [H2O2] was 5.6xl0-4mol dm-3. Absorption of 254nm radiation will normally result in ak -k* transition to the lowest excited singlet state of the molecule, which corresponds to the level in benzene shifted to lower energies by the substituents141. This singlet state can react directly or it can undergo rapid radiationless conversion to the lowest triplet.It has been shown that phenols are more acidic in the lowest excited singlet state than in the ground state (or the lowest triplet). For phenol itself, the values of pK are 10.00, ca. 4.0 and ca. 8.5 for the ground state, the lowest excited singlet state and the lowest excited triplet state respectively142. Flash photolysis of aqueous phenol solutions is known to produce hydrated electron s 143>144j although continuous photolysis using low intensity radiation apparently does not145. 152 Table 6.4. Efficiencies for the Production of Hydrogen Peroxide for Air-saturated 4xlO-3M Aqueous Solutions of Aromatic Compounds at pH = 6.00±0.01. 30 Minutes Irradiation at 254nm. Compound Efficiency / % 3,5-dihydroxybenzoic acid 3.29 Gallic acid 2.40a Pyrogallol 2.20 2,5-dihydroxybenzoic acid 1.22 3-O-methyl gallic acid 0.89 2 ,3-dihydroxynaphthalene- 6 -sulphonic acid 0.80 Tannic acid 0.72b DL-DOPA 0.71 3,4-dihydroxy benzoic acid 0.66 Resorcinol 0.57 2,3-dihydroxybenzoic acid 0.55 Phloroglucinol 0.50 3-hydroxy-5-methoxybenzoic acid 0.48 1,2,4-trihydroxybenzene 0.46c 2,3,4-trihydroxybenzoic acid 0.42 3-hydroxy tyramine hydrochloride 0.40 5-bromo-3,4-dihydroxybenzoic acid 0.31 2,4-dihydroxybenzoic acid 0.30 Tiron 0.30 3,4-dihydroxyhydrocinnamic acid 0.25 Syringic acid 0.24 Methyl-2,3,4-trihydroxybenzoate 0.22 Catechol 0.21 Chlorogenic acid 0.11 153 Benzoic acid < 0.10% 2 ,6 -dihydroxybenzoic acid " 3,4-dihydroxy cinnamic acid " 4-hydroxybenzoic acid " 3-hydroxy-2-naphthoic acid " Salicylic acid " a 4 x 10"3M solution with oxygen 3.40%. b 4xlO'3M solution with oxygen 0.83%.c at pH 4.00 due to appreciable thermal reaction at pH 6.00. One possible mechanism for the photochemical formation of2 OH2 from catechols is given in fig. 6 .8; (o) Fig. 6 .8(a) Possible mechanism for the photochemical formation of2 OH2 from oxygen and a catechol. Alternatively, the electron can be transferred directly from the excited state to the O2. The remaining steps are as for the thermal reaction, fig. 6.7, except that in these more acidic solutions disproportionation (b) of the semiquinones may occur. 1 5 4 With gallic acid, the solution formed on photolysis had an almost identical visible spectrum to that of the product obtained in the thermal reaction (after subsequent adjustment to the same pH). 4-hydroxybenzoic acid, a substituted phenol, gave negligible yields of H 2O2 (table 6.4) . The greater reactivity of catechols in their excited states, as compared to phenols, may be due to the increased stability of the semiquinone radicals formed by the loss of an electron. This would allow some of the electrons to escape from the solvent cage before recombination occurs. 6.1.6 Resorcinols. It was shown earlier that, as expected, resorcinols in alkaline solution did not react thermally with O 2 to give H2O2. However, in the photochemical reactions, the highest quantum yield for the formation of H2O2 was found with 3,5-dihydroxy benzoic acid (Table 6.4) and smaller but significant yields are obtained with other resorcinols. There is thus a complete contrast between the ground state and excited state reactivity of the resorcinol moiety. This may be associated with the much greater interaction which is known to occur between meta substituents in the excited states of substituted benzenes, compared to the ground states14^. In the thermal production of H2O2, trihydroxy compounds such as gallic acid and resorcinol are clearly best regarded as OH substituted catechols. In the photochemical production of H2O2, this is not necessarily so, and they could equally well be considered as OH substituted resorcinols. The negligible quantum yield for 2,6-dihydroxybenzoic acid may arise from H-bonding of both OH groups with the carboxylate ion, which can result in rapid radiationless conversion of the excited state to the ground level. 2,5-dihydroxybenzoic acid was the only hydroquinone studied, and an appreciable yield of H 2O2 was obtained, (Table 6 .4 ) . The wavelength of the radiation used in this work (254nm) is shorter than that corresponding to the solar energy "cut-off' at sea-level (ca.295nm)147. However, most 155 of the compounds studied here have absorption bands which extend beyond 295nm (eg 1-3 1 8=50 mol" dm cm for gallic acid at 316nm, 3,4-dihydroxybenzoic acid at 314nm 3,5-dihydroxy benzoic acid at 353nm and tannic acid at 436nm, while 8=10 mol" dm cm"1 for resorcinol at 295nm and catechol at 299nm). 6.2 PHYSICAL MEASUREMENTS. 6.2.1 Thermal Reactions. Air or O2 were passed first through a pre-saturator, and then through the phenolic solution contained in a Dreschel bottle fitted with a sintered glass disc and magnetic follower, immersed in a water bath at 20±0.5°C. The Dreschel bottle was coated on the outside with Kodak Linagraph Stabilising Lacqueur to absorb ambient UV radiation. Periodically, 10ml samples were removed and analysed for H 2O2 by means of the ethyl acetate extraction method described elsewhere. Blank readings, obtained by treating an aliquot with ca. 250 units of catalase and leaving for 10 minutes before analysing, were normally < 1% of the observed maximum absorbances. 6.2.2 Photochemical Reactions. The apparatus used for the photochemical reactions is shown to scale in fig. 6.9 overleaf. 20ml of the phenolic solution in 10a cm path length cylindrical silica cell was exposed along the axis of the cell to 254nm radiation from a low-pressure mercury lamp whilst air or O2 was bubbled through using two lengths of 1mm diameter PTFE tubing. The temperature was maintained at 25±2°C. 1 5 6 atr or oxygen low pressure mercury lamp emitting 254nm radiation Fig. 6.9 Apparatus for photochemical reactions. 6.2.3 Calculation of Quantum Yields. Quantum yields Q for the formation of H2O2 were obtained using the ferrioxalate actinometer148’149. The solution was replaced by 20ml of 0.006M K3Fe(ox)3, and N2 bubbled through during irradiation for a known time. [H202].t2.Q’ Then Q = ...... 6.2 [Fe2+].t! Molecules of H2O2 formed where Q = ...... 6.3 Number of quanta absorbed [H2O2] and [Fe2+] are the concentrations of these substances formed in times tj and t2 respectively, and Q' is the quantum yield for the formation of Fe from K3Fe(ox)3 (1.25 at 254nm ). For the measurements, a slight error will be introduced by the presence of the Hg emission lines in the region 313-436nm, which will be absorbed by the actinometer but only slightly by the phenolic substances. From the known emission of low-pressure mercury lamp149, this error should not exceed 5%. Calibration of the 157 lamp in this way was carried out daily before and after each series of photochemical runs. 6.3 MATERIALS AND SYNTHESES. 6.3.1. Phenolic Compounds. Tannic acid and gallic acid (supplied by BDH Chemicals Ltd) were used directly. The tannic acid lost 6.9% by weight of H 2O on heating to 110°C. 3,4-dihydroxybenzoic acid, 3-hydroxy-2,4-dimethoxybenzoic acid (syringic acid), 2,3-dihydroxynaphthalene- 6-sulphonic acid, 3-hydroxytyramine hydrochloride, 3-hydroxy-2-naphthoic acid, 2,3,4-trihydroxybenzoic acid, 2,3-dihydroxybenzoic acid, 4-hydroxybenzoic acid , benzoic acid, 2-hydroxybenzoic acid (salicylic acid), chlorogenic acid and caffeic acid were recrystallised from water. 3-Methoxy-4,5-dihydroxybenzoic acid (3-O-methyl gallic acid) was prepared as outlined by Scheline150 (MPt 218-220°C, Lit150 220°C) 6.3.2 Buffers. The following buffers were used at a concentration of 0.05M over the stated pH ranges; 4-(2-hydroxyethyl)-l-piperazinepropane sulphonic acid, EPPS , pH=ca.8.5- 8.0; 4-morpholinepropane sulphonic acid, MOPS, pH=ca.7.5-7.0; succinate, pH=ca.6.0-3.6. 6.3.3 Enzymes. Catalase (from bovine liver) was obtained from Sigma. 158 6.4 GENERAL CONCLUSIONS. A quantitative study has been made of the formation of hydrogen peroxide when alkaline solutions of tannic acid, gallic acid and a number of other molecules containing the catechol (1,2-dihydroxybenzene) moiety are exposed to air or oxygen. For gallic acid, from analysis of the pH dependence of the initial rates of reaction with air, it can be concluded that over the pH range ca. 7.0-8.5, the predominant reactive species is the dianion, where only one of the hydroxylic protons has been removed. This is in clear contrast to the situation found for tetramethylhydroquinone, which reacts quantitatively with O2 in alkaline water-ethanol to give tetramethylquinone and H2O2. Here, the reactive species is apparently the dianion, where both hydroxylic protons have been removed. Thermal reactions of both tannic and gallic acid with oxygen produce higher yields than those with air, the yield for gallic acid approaching the maximum theoretical value of a 150% yield of H 2O2 based on the gallic acid. The proposed mechanism for the formation of2 HO2 from oxygen and a catechol involves the reduction of O2 to O2'. Subsequent protonation of 02" yields the HO2 precursor, which can combine with O2- to yield HO 2*, and hence H 2O2. Thus what is ostensibly the two-electron reduction of O2 to H2O2 is effected by an initial one- electron reduction step. The photochemical production of hydrogen peroxide, using 254nm radiation, has been studied for air-saturated solutions of a variety of phenolic compounds at pH 6.0 below, where the thermal reaction outlined above is negligible. Apparent quantum yields have been reported from measurements made using a fenioxalate actinometer to calibrate the output of the lamp used. Unexpectedly the highest yield of H 2O2 was obtained with a resorcinol derivative, 3,5-dihydroxybenzoic acid. It was found that appreciable formation of H2O2 is observed with 1,2-, 1,3- and 1,4-benzene diols. These residues are found quite commonly in plant materials, including humic substances. This suggests that the photochemical production of 2HO2 in the environment could well 159 involve these species, possibly in conjunction with other groups which would shift the absorption to longer wavelengths. Although the wavelength of radiation used is shorter than that corresponding to the solar energy cut-off (ca. 295nm), most of the compounds studied in the photochemical experiments have absorption bands which extend beyond 295nm. The proposed mechanism for the photochemical formation of2O H2 from oxygen and a catechol proceeds via loss of the more acidic hydroxylic proton of the catechol in the excited state. 160 References. 1. W.C. Schumb, C.N. Satterfield and R.L. Wentworth, 'Hydrogen Peroxide', Reinhold Publishing Corp., New York, 1955. 2. A.M. Held, D.J. Halko and J.K. Hurst, J. Am. Chem. Soc., 1978, 100, 5732. 3. D. F. Evans and M. W. Upton, J. Chem. Soc. Dalton Trans., 1985, 2525. 4. L J. Thenard, Ann. Chim. et Phys., 1818, 8, 306. 5. L.J. Thenard, Ann. Chim. et Phys., 1818, 9, 441. 6. L.J. Thenard, Ann. Chim. et Phys., 1819, 10, 335. 7. R. Wollfenstein, J. Chem. Soc. Ind., 1895, 14, 505. 8. J.W. Mellor, 'Comprehensive Treatise on Inorganic and Theoretical Chemistry', Longmans, London, I, 1946. 9. LJ. Thenard and P. Thenard, Ann. Chim. et Phys., 1856, 47, 173. 10. C.F. Schonbein, J. Sci., 1859, 27, 19. 11. C. Weltzien, Ann.. 1866. 138. 129. 12. E. Farber, 'Oxygen and Oxidation, Theories and Techniques in the Nineteenth Century and the first part of the Twentieth', Washington Acad. Sci., 1967, 4. 13. R. Willstatter, J. Chem. Soc. Abs., 1903, 537. 14. S. Tanatar, J. Chem. Soc. Abs., 1902, (if). 26. 15. T. Fairley, J. Chem. Soc., 1877, 31, 1. 16. W. Manchot, Ann., 1903, 325, 105. 17. D. McIntosh, J. Chem. Soc. Abs., 1902, (iiL 310. 18. G. Bredig and M. Fortner, Ber., 1904, 37, 798. 19. G. Bredig and W. Reinders, J. Chem. Soc. Abs., 1901, (ii), 442. 20. E. Schone, Ann., 1881, 195, 228. 21. H. Wilson, J. Soc. Dyers. Colourists, 1922, 38, 139. 22. G.E. Davis, Chem. News, 1879, 39, 221. 23. C.T. Kingzett, J. Soc. Chem. Ind., 1890, 9, 3. 1 61 24. P. Poetschke, J. Ind. Eng. Chem., 1922, 181. 25. E.S. Shanley and M.E. Bretschger, J. Electrochem. Soc., 1952, 99(11), 311C. 26. H. Meidinger, J. Chem. Soc., 1854, 7, 251. 27. M. Berthelot, Compt. rend., 1878, 86, 71. 28. C.A. Cramp ton, G. Farber, R. Jones and J. Pleaver, 'The Modern Inorganic Chemical Industry', Ed. R. Thompson, Special Publication No.31, Chemicals Society, London, 1977. 29. M. Eggsdorfer, J. Henkelmann and W Grosch, Patent DE 3,732,015. 30. W. Buchner, R.Sddiebs, G. Winter and K.H. Biichel, 'Industrial Inorganic Chemistry’, V.C.H., W einheimf X .6,., 1989. 31. P. Ebell, J. Soc. Chem. Ind., 1882,1, 65. 32. A.W. Johnson, Science Progress, 1951, 39, 96. 33. H.Y. Taylor, The Industrial Chemist, 1953, 91. 34. A. Classen and E Bauer, Ber., 1883,16, 1062. 35. J. March, 'Advanced Organic Chemistry; Reactions, Mechanisms and Structure', 3rd Edn., Wiley-Interscience, New York, 1985, 733. 36. P.J. Harvey, H.E. Schoemaker and J.M. Palmer, Plant Cell Environ., 1987, 12, 709. 37. S.E. Stevens Jr., C.O. Patterson and J. Myers, J. Phycol., 1973, 9, 427. 38. D.J. Aneshansley, T. Eisner, J.M. Widom and B. Widom, Science, 1969, 165. 61; D.J. Aneshansley, J.H. Jones, D. Aslop, J. Meinwald and T. Eisner, Experimenta, 1983, 39(4), 366. 39. A. Sigg and A. Neftel, Annals. Glaciology, 1988, 10, 157. 40. P. Laj, S.M. Drummey, M.J. Spencer, J.M. Palais and H. Sigurdsson, Nature, 1990, M6, 45. 41. W.J. McElroy, Atmos. Environ. 1986, 20, 427 and references therein. 42. A.M. Thompson, M.A. Owens and R.W. Stewart, Geophys. Res. Lett., 1989, 16(1), 53. 162 43. Farhataziz and A.B. Ross, NSRD-National Bureau of Standards-59, U.S. Dept, of Commerce, Washington, D.C., 1977. 44. J.J. Bufalini, B.W. Gray Jr. and K.L. Brubaker, Environ. Sci. Tech., 1972, 6(9), 816. 45. K.H. Becker, K.J. Brockmann and J. Bechara, Nature, 1990, 346. 256. 46. C. van Baalen and J.E. Marler, Nature, 1966, 211. 9511. 47. A.J. Swallow, Nature, 1969, 222. 369. 48. W.J. Cooper and D.R.S. Lean, Environ. Sci. Tech., 1989, 23(11), 1425. 49. H. Fischer and D. Mayer, 'Ullman's Encykl. Tech. Chem.', 4th Edn., 1982, 21, 167. 50. National Research Council, Committee on Medical and Biologic Effects of Environmetal Pollutants, Hydrogen Sulphide, Baltimore, Md., 1979, pp 1/183, 1/9, 47/8, 82/3. 51. P. Macaluso, 'Kirk-Othmer Encycl. Chem. Technol.', 2nd Edn., 1969, 19, 375. 52. H.W. Jannasch and M.J. Mottl, Science, 1985, 229, 717. 53. W.W Hartman and H.L. Silloway, Org. Synth., 1955, Coll. Vol. 3, 83, 1955. 54. For a general review of the Zinin reaction see; H.K. Porter, Org. React., 1973, 20, 453. 55. N. Zinin, J. Prakt. Chem., 1842, [1] 27, 149. 56. J.P. Idoux, J. Chem. Soc., 1970, C, 435. 57. A, Laurent and C. Gehrhardt, Ann., 1850, 75, 74. 58. H. Heidelberger and W.A. Jacobs, J. Am. Chem. Soc., 1919, 41, 1459. 59. H. Heidelberger and W.A. Jacobs, J. Am. Chem. Soc., 1919, 41, 1467. 60. M.C. Ball and A.H. Norbury, 'Physical Data for Inorganic Chemists’, Longman, London, 1974, and references therein. 61. D.D. Wagman, National Bureau of Standards-Technical Note 270 Parts 1 (1965) and 2 (1966). 1 63 62. J.B. Conant and L.F. Fieser, J. Am. Chem. Soc., 1924, 46, 1858. 63. L.F. Fieser and M. Fieser, J. Am. Chem. Soc., 1935, 57, 491. 64. J.B. Conant and M.F. Pratt, J. Am. Chem. Soc., 1926, 48, 3178; J. Am. Chem. Soc., 1926, 48 3220. 65. E. Biilmann and J.H. Blom, J. Chem. Soc., 1924, 125, 1719. 66. "Matheson Gas Data Book', W. Braker and A.L. Mossman, Matheson Gas Products Ltd., 5th Edn., 1971, 319. 67. IUPAC Solubility Data Series, Vol. 32, 'Hydrogen Sulphide, Deuterium Sulphide and Hydrogen Selenide', Pergamon, Oxford, 1988, and references therein. 68. M.B. Goldhaber and I.R. Kaplan, Marine Chem., 1975, 3, 83. 69. S. Licht, F. Farouzan and K. Longo, Anal. Chem., 1990, 62, 1356. 70. S. Licht, J. Electrochem. Soc., 1988, 135, 2971. 71. A.I. Vogel, 'A Textbook of Quantitative Inorganic Analysis', 3rd Edn., Longmans, London, 1961, 462. 72. S.J. Angyal and W.K. Warburton, J. Chem. Soc., 1951, 2492. 73. D.D. Perrin, Dissociation Constants of Organic Acids and Bases in Aqueous Solution', Butterworth's Scientific Publishing, London, 1965, and references therein. 74. L.E. Overman, D. Lesuisse and M. Hashimoto, J. Am. Chem. Soc., 1983, 105. 5373. 75. 'Critical Stability Constants', Vol. 2, Eds. R.M. Smith and A.E. Martell, Plenum Press, New York, 1975, and references therein. 76. Gmelin Handbook, S Supplement Vol. 4A, Springer-Verlag Chemie GmbH, Berlin, 1983, 79. 77. Powder Diffraction File, 8-247, American Society of Testing and Materials, Philadelphia, 1967. 78. R. Steudel, Top. Curr. Chem., 1982, 102, 149. 1 64 79. K. Tamao, T. Hayashi and Y. Ito, J. Chem. Soc., Chem. Comm., 1988, 795. 80. J.R. Bolton, A. Carrington and J. dos Santos-Veiga, Mol. Phys., 1962, 5, 465. 81. A.I. Vogel, 'A Textbook of Quantitative Inorganic Analysis', 3rd Edn., Longmans, London, 1961, 357. 82. L.F. Fieser and F.C. Chang, J. Am. Chem. Soc., 1942, 64, 2043. 83. R.H. Thomson, J. Chem Soc., 1953, 1196. 84. A.H. Cook, J. Chem. Soc., 1938, 876. 85. A.C. Carlin and G.S. Wich, J. Am. Chem. Soc., 1958, 80, 423. 86. T.F. Chung, Y.M. Yu and C.H. Cheng, J. Org. Chem., 1984, 49, 1215. 87. G. Schultz, Ber., 1884, 17, 463. 88. J. Gasparic, I. Gemzova and D. Snobi, Coll. Cz. Comm., 1966, 31, 1712. 89. W. Davies and R.A.R. Down, J. Chem. Soc., 1929, 586. 90. Y. Kyogoku and B.S. Yu, Bull. Chem. Soc. Jpn., 1969, 42, 1387. 91. 'Colorimetric Determination of Non-Metals', D.F. Boltz and J. A. Howell (Eds.), Vol. 8, 2nd Edn., Wiley-Interscience, New York, 1978, 301. 92. R.M. Sellers, Analyst, 1980, 105. 93. G. Janicek and J. Pokorny, Chem. Listy, 1955, 49, 1315. 94. I.R. Cohen and T.C. Purcell, Anal. Chem., 1967, 39, 131. 95. J.S. Reichert, S.A. McNeight and H.W Rudel, Ind. Eng. Chem., Anal. Ed., 1939, n , 194. 96. J. Rynasiewicz, Anal. Chem, 1954, 26, 355. 97. W.C. Schumb, C.N. Satterfield and R.L. Wentworth, 'Hydrogen Peroxide', Reinhold Publishing Corp., New York, 1955, 294. 98. J.H. Walton and H.A. Lewis, J. Am. Chem. Soc., 1916, 38, 633. 99. V.P. Vasil'ev and P.N. Vorob'ev, Izv. Vyssh. Uchebn. Zaved. Khim. Khim. Technol., 1968, 11, 971. 165 100. A.C. Egerton, AJ. Everett, G.J. Minkoff, S. Rudrakanchana and K.C. Salooja, Anal. Chim. Acta, 1954, 10, 422. 101. TJ. Marks, J. Am. Chem. Soc., 1978,100, 3215. 102. T. Yonetani and H. Yamanoto, 'Oxidases and Related Systems', T.E. King (Ed.), Vol. 1, University Park Press, Maryland, 279. 103. H.B. Dunford and J.S. Stillman, Coord. Chem. Rev., 1976, 19, 187. 104. Biochemica Information, J. Keesey (Ed.), 1st Edn., Boehringer Mannheim Biochemicals Ltd., 1989, 56. 105. K.G. Welinder, Eur. J. Biochem., 1985, 151. 497. 106. M. Morrison and G.K. Schonbaum, Ann. Rev. Biochem., 1976, 435. 861. 107. C. Vuk-Pavlovic and Y. Siderer, Biochem. Biophys. Res. Comm., 1977, 79. 885. 108. T. Nozawa and N. Koboyashi, Biochim. Biophys. Acta, 1976, 427. 652. 109. B.C. Chance, Science, 1949, 109, 204. 110. R. Makino and Y. Ishimura, J. Biol. Chem., 1986. 261. 8376. 111. G.K. Schonbaum and S. Lo, J. Biol. Chem., 1972, 247. 3353. 112. G. Rakshit and T.G. Spiro, Biochem. Biophys. Res. Commun., 1976, 71. 803. 113. R. Lichtenstein, 'Chemical Physics of Redox Metalloenzyme Catalysis', Wiley-Interscience, New York, 1989, 136. 114. Biochemicals, Organic Compounds for Research and Diagnostic Reagents, Sigma Chemical Co. Ltd., Poole, Dorset, 1990, 7rr3. 115. D. Barham and P.Trinder, Analyst, 1972, 97, 142. 116. T. Katsumi, Y. Murao, K. Akiura and Y Ohkura, Anal. Chim. Acta., 1982, 136. 121. 117. A. Szutowicz, R.D. Kobes and P. Orsulak, Anal. Biochem., 1984, 138. 86. 118. P. Trinder, Ann. Clin. Biochem., 1969, 6, 24. 119. M. Shiga, M. Saito and K. Kina, Anal. Chim. Acta., 1983, 153. 191. 166 120. T.T. Ngo and H.M. Lenhoff, Anal. Biochem., 1980, 105, 389. 121. H.A. Mottola, B.E. Simpson and G. Gorin, Anal. Chem., 1970, 42, 470. 122. D.T. Capaldi and K.E. Taylor, Anal. Biochem., 1983,129, 329. 123. TJ. Kelly, P.H. Daum and S.E. Schwartz, J. Geophys. Res., 1985, 90. 7861. 124. A.S. Chandler, T.W. Chourlaton, G.J. Dollard, A.E.J. Eggleton, M.J. Gay, T.A. Hill, B.M.R. Jones, B.J. Tyler, B J. Bandy and S.A. Penkett, Nature, 1988, 336, 562. 125. W.J. Cooper, R.G. Zika, R.G. Petasme and J.M.C. Plane, Environ. Sci. Technol., 1988, 22, 1156, and references therein. 126. R.G. Petasme and R.G. Zika, Nature, 1987, 325. 1987. 127. T. Holm, G.K. George and M.C. Barcelona, Anal. Chem., 1987, 59, 582. 128. C.F. Schonbein, J. Pr. Chem., 1860, 81, 257. 129. 'Rodd's Practical Chemistry', E. Coffey (Ed.), Vol. 3, Elsevier, Amsterdam, 1976, 227. 130. W. Brackman andE. Havinga, Reel. Trav. Chim., 1955, 74, 110. 131. R. Bohn and C. Graebe, Ber., 1887, 20, 2328. 132. J. Grimshaw and R.D. Howarth, J. Chem. Soc., 1956, 418. 133. 'Critical Stability Constants', Vol. 3, Eds. R.M. Smith and A.E. Martell, Plenum Press, New York, 1975, and references therein. 134. T.H. James and A. Weissberger, J. Am. Chem. Soc., 1938, 60, 98. 135. N.P. Slabbert, Tetrahedron, 1977, 33, 821. 136. L.H.J. Lajunen, A. Kostama and M. Karvo, Acta. Chem. Scand., Sect. A, 1979, 33, 681. 137. B. Carlini, P. Silvestri and M. Vicedomini, Ann. Chim. (Rome), 1979, 69, 533. 138. R.J. Motekaitis and A.E. Martell, Inorg. Chem., 1984, 23, 18. 1 6 7 139. J. March, 'Advanced Organic Chemistry; Reactions, Mechanisms and Structure', 3rd Edn., Wiley-Interscience, New York, 1985, 244. 140. P. Silvestri, F. Talarico and M. Vicedomini, Ann. Chim. (Rome), 197, 69, 19. 141. 'The Theory of Electronic Spectra of Organic Molecules', J.N. Murrell (Ed.), Methuen, London, 1963, 127. 142. 'Fluorescence', G.G. Guilbault (Ed), Arnold, London, 1967, 47, and references therein. 143. L.I. Grossweiner, G.W. Swenson and E.F. Zwicker, Science, 1963, 141. 805. 144. J-C. Mialocq, J. Sulton and P. Goujon, J. Chem. Phys., 1980, 72, 12. 145. J. Jortner, M. Ottolenghi and G. Stein, J. Am. Chem. Soc., 1963, 85, 2712. 146. 'Techniques of Organic Chemistry', A. Weissberger (Ed.), Vol. XIV, Interscience, New York, 1969, 184. 147. 'Light, Chemical Change and Life', J.D. Coyle, R.R. Hill and D.R. Roberts (Eds.), Open University Press, Milton Keynes, 1983, 338. 148. G.C. Hatchard and C.A. Parkes, Proc. R. Soc. London, Ser. A, 1956, 235, 518. 149. 'Photochemistry', J.G. Calvert and J.N. Pitts, Wiley, New York, 1966, 64. 150. R.R. Scheline, Acta. Chem. Scand., 1966, 20, 1182. 168 A ppendix 1 Shown below is the computer program (BBC Basic) used to obtain optimum values of k' and c for the kinetic studies of the reductions and oxidations of various quinones (see 1.1.15, p.66). By way of illustration of its use, at the end of the program is data for the reduction of 2-ethyl-3-methyl- 1,4-naphthoquinone, shown in fig. 1.16. 10 REMPHIL1 20 OPTION BASE 1 30 NO. OF DATA POINTS...... 40 PO=19 50 DIM Y(PO), T(PO), YCAL(PO), SUM( 100) 60 REM CONSTANTS FOR OPTIMIZATION 70 REM STARTING VALUES AND INCREMENTS 80 C=.023: CBIT=.0001: B=.0001: BBIT=.00001 90 INTC=C: INTB=B: SMALLSUM=1000: SMALL=1000 100 REM READ IN DATA POINTS...... 110 FOR A=1 TO PO 120 READ T(A) 130 NEXT A 140 FOR A=1 TO PO 150 READ Y(A) 160 NEXT A 170 REM 2ND OPTIMIZATION LOOP 180 FOR BLOOP= 1 TO 100 190 FOR A=1 TO 100 200 SUM(A)=0 210 NEXT A 215 C=INTC 220 REM 1ST OPTIMIZATION LOOP 230 FOR CLOOP=l TO 100 240 FOR A=1 TO PO 250 YCAL(A)= C*(1-EXP(-B*T(A))) 260 DIFF=(YCAL(A)-Y(A))A2 270 SUM(CLOOP) 280 NEXT A 290 REM OPTIMUM C...... 300 IF SUM(CLOOP) SMALLSUM=SUM(CLOOP):OPTIC=C 310 C=C+CBIT 320 NEXT CLOOP 330 REM OPTIMUM B ...... 340 IF SMALLSUM 350 B=B+BBIT 360 PRINT BLOOP 370 LPRINT "OPTIMUM B= ";OPTIB, "RANGE= ";INTB;"-";B 380 LPRINT "OPTIMUM C= ";OPTIC, "RANGE= ";INTC;"-";C 390 LPRINT "LEAST SQUARES= ";SMALL 400 LPRINT 410 LPRINT "T", "YCAL", "YOBS" 420 FOR A=1 TO PO 430 Y CAL( A)=OPTIC* (1 -EXP(-OPTIB *T (A))) 440 LPRINT T(A), YCAL(A), Y(A 450 NEXT A 460 DATA 150, 300, 600, 900, 1200, 1500, 2100, 2400, 3000, 3600. 4200. 4800. 5400, 600, 7200,10000, 13800 470 DATA 0.001, 0.002, 0.005, 0.008, 0.010, 0.012, 0.014, 0.016, 0.017, 0.020, 0.022, 0.023, 0.024, 0.025, 0.025, 0.026, 0.027, 0.028, 0.028 490 END The progam works by performing a least squares fit on the data for the equation A=C(l-e-Bt) Where Y, B and C are referred to as (pt - poo), k' and c in the text. Values in the region of those expected are taken form the graphs according to the half-life relationship for a first order reaction; ti/2 = In 2 k' Taking this value of k' and fitting i t , and the appropriate experimentally-obtained value of (p-t “Poo) at ti/2, into the equation; (Pt -Poo) = c(l-e-k't) gives an approximate value for c. These approximate values are put into the program along with the size of increments in each (line 80 of the program) which the computer will make until it has arrived at the best fit for the experimental data. At this point, the computer prints out both the experimental values of Y (YOBS') and those calculated according to the best-fit parameters for the equation (YCAL).