摘要:

Journal of the Chemical Society, Faraday Transactions I SUBJECT INDEX-VOLUME 68, 1972 PAGE I, 1 Adsorption (including absorption, chemisorption, physical adsorption, clathration, ion Absorption of Hydrogen by Palladium-Platinum Alloys under High Pressures of Hydrogen Adsorption and Incorporation on Copper. (Saleh) . Adsorption of Krypton on Alkali Halide Crystals. Part 1 : Theory. (Takaishi) : Adsorption of Krypton on Alkali Halide Crystals. Part 2: On NaCI, KC1 and RbCi. Adsorption of n-Alkane Vapour; on Graphon. <Clintj : Adsorption of n-Alkanols at Alkane/Water Interfaces. (Aveyard & Briscoe) : Adsorption of Neopentane on Tungsten and Palladium Films. (Ross, Roberts & Kembali) Adsomtion of Xenon and Hydrogen Atoms and their Interaction on Silver Films. (Knaup exchange, etc.) Gas.(Baranowski, Lewis, Majchrzak & Widniewski) . . . . . (Takaishi & Mohri) . . . - - . - - &Stiddard) . . . Adsomtion of Xenon on Films of Goid. Aluminium and Silver. (Bruce' & Sheridan) : Chemisorption of Hydrogen on Evaporated Copper Films. (Alexander & Pritchard) . Effkct of Silica Surface Dehydroxylation on Adsorption of Aromatic Hydrocarbons from Electronegativity, Work Function, and Heat of Adsorption of' Hy&ogen * on Metals: Field Emission Study of ;he Formation and Desorption of Oxide Layers on Tungsten Infra-red Studies of Carboxylic Acid VapoGs Adsorbed on Evaporated Germakum Films'. Infra-red Studies of Oxygen Adsorbed on Evapoiated Germanium Film's. (Howe,'Liddy Infra-red Studies af Reactions on Oxide Surface's. Part 1 : 'Boron Trifluoride on 'Silica: Infra-red Studies of Rutile Surfaces. Part 3 : Adsorption of Water and Dehidroxylation Infra-red Study of the Adsorption of Ammonia on'MgO: Pait 1 : The Dehydrated Surface: Infra-red Study of the Adsorption of Ammonia on Mgb.Part 2:'The Hydrated Surface: Infra-red Study of the Reactions of Skon,'Titanium and Tin Tetrachlorides 'with Rutile: Infra-red Study of the Surface Properties of Rutiie. Adsorpiion of Ethanol, n-Butanol and Infra-red Study of the Surface Properties of Rutile: Deuterium Exdhange, Carbon Dioxide and But-1-ene Adsorption. (Jackson & Pafitt) Interaction of Hydrogen Chloride with Evaporated Metal Films. .Part 1 : Chiorinakon of Iron, Nickel, Palladium, Silver and Lead. (Dadiza & Saleh) Interaction of Hydrogen Chloride with Evaporated Metal Films.Part 2: Adiorption and Incorporation on Films of Pd, Ag and Pb, Deposited on Iron Substrates. (Dadiza & Saleh) . . . . . . . . - . Interaction of Oxygen with Polycrystalline Tungsten. Part 3 : Electron Skmulated Desorp- tion. (King, Madey & Yates, Jr.) . . . . . Ion Exchange Involving Several Groups of Homogeneous Sites. (Barre; & Kiinowski) . Molecular Sieve Properties of the Carbonaceous Products prepared by Chemical Dehydro- halogenation of Polyvinylidene Chloride. (Bartcn, Beswick & Harrison) . Order-Disorder Transitions at the Liquid/Solid Interface. Volumetric Behaviour of Primary Aliphatic Alcohols near the Graphon Surface. (Findenegg) . . Oxidation of a Pyrophoric Iron. Part 1 : Kinetics of Chemisorption of Oxygen on the finely divided Iron-Carbon Substrate.(Galwey & Gray) Oxygen Species Adsorbed on Oxides. Part 1 : Formation and Reactivity of (O')s on'MgO: (Tench, Lawson & Kibblewhite) . Oxygen Species Adsorbed on Oxides. Part'2: Foknation of (O&'on MgO. '(Tendh) : Solid-Solution Interface Equilibria for Aromatic Molecules Adsorbed from Non-aromatic Media, Aromatic Hydrocarbons. (Wright & Powell) . . . Sorption of Carboxylate Ions by Strongly Basic Anion Exchangers. (Gregory & S e k e n s j Solution in n-Alkanes. (Eltekov, Khopina & Kiselev) . (Trasatti) . . . . Surfaces. (Goymour & King) (Howe & Metcalfe) . & Metcalfe) (Morrow & Devi) . . . of Rutile. (Jones & Hockey) . . (Tench & Giles) .(Tench) . . . (Parfitt, Ramsbotham & Rochester) . . n-Hexanol. (Jackson & Parfitt) . . 1 824 1520 801 1921 2239 478 914 21 39 997 202 889 229 280 393 1595 403 907 193 197 17 1443 896 269 1513 1347 73 1647 1799 1935 1169 1181 1908 10452 SUBJECT INDEX-VOLUME 68, 1972 Sorption of Ethylene, Propylene and Cyclopropane in 5A Zeolite. Analysis of Equilibrium Sorption of Light Paraffins in Type-A Zeolites. Analysis and'Interpretation of 'Equilibrium Spectral and Energetic Aspects of Water Adsorpiion by Li-, Na-, K- and Ci-X Glites: (Kiselev, Lygin & Staradubceva) . Thermodynamics of Adsorption of a Series of Related Organic Moieculei on GrapGte and a Carbon Black. Part 1 : Heat of Adsorption. (Dollimore, Heal & Martin) . . Isotherms. (Derrah, Loughlin & Ruthven) .Isotherms. (Ruthven & Loughlin) . . . . I, 2 Biophysical Chemistry I, 3 Catalysis (including heterogeneous and homogeneous catalysis and surface reactivity) Acidic Sites on Silica-Alumina Catalysts with Reduced Aluminium Content. Infra-red and Thermogravimetric Studies. (Ballivet, Barthomeuf & Pichat) . . Catalytic Activities of Polynaphthoquinone, Containing various Metal Halides. Dehydro: genation of Formic Acid, Cyclohexane, Ammonia and Alcohols and Isomerization of Butene. (Iwasawa, Soma, Onishi & Tamaru) Catalytic Activity and Electron Configuration of the EDA Compiexes of Phfhalocyanink with Alkali Metals. Hydrogen Adsorption and the H2-D2 Exchange Reaction. (Naito, Ichikawa & Tamaru) . . . Catalytic Activity and Selectivity of NaOH-Doped y-Aluminas.Dehydrakon and Dehydro- genation of 3-Pentanol. (Chuang & Dalla Lana) . . . Chemisorption and Decomposition of Tetramethylsilane over'Tungsten and Irbn Sukaces: (Roberts &Ross) . . . . Effect of Adsorbed Hydrocarbon on the Catalytic Exchange of n-Butane and Deuterikn on Tungsten Films. (Arroyo & Kemball). . . . Effect of Pressure Changes on the Oxidation Rate of Aluminium in 'the Tkmpeiature Range 323-673 K. (Hunt & Ritchie) . . Heterogeneous Decomposition of Hydrazine. Part 1 : On a Suppoked Rhodium Catalp;. (Davis & Sayer) . . . Homogeneous Catalysis by Pt(G)-Sn(iI) Chloride Complex.' Part 2 : Mechanism of the Isomerization of n-Butanes. (Hirabayashi, Saito & Yasumori) . Hydrogenolysis of Ethane on Evaporated Copper-Nickel Alloy Films.(Plunkett & Hydrogenolysis of Saturated Hydrocarbons on Evaporated Plaiinum' Fi&. (Dowie: Infra-red Spectra and Catalitic Activity of Supportkd Molybdenum Hexacarbonyl. (Howe; Low Temperature Parahydrogen' Enrichmen; on Non-s;oichiometric Rutile. (Richardson; Oxidation of Aluminium by Nitrous Oxide in the Tempeiature Range 3231683 K. criunt & Role of the Catalyst Suppok in the Oxidation of Methane over PaliadiA. (Cullis, Neveil & Trimm) . . Surface and Bulk Reduction of' Bismuth/Molybdenum' Oxides Studied 'by Eiectron Spin Resonance Spectroscopy. (Burlamacchi, Martini & Ferroni) . . . . . Clarke) . . . Whan & Kemball) . . Davidson & Whan) Rudham, Tullett & Wagstaff) . . . Ritchie) . I, 4 Colloid Science (including birefringence, electrophoresis, light-scattering, sedimentation, thixotropy, soluble and insoluble monolayers) Absorption and Emission Studies of Solubilization in Micelles.Part 1 : Pyrene in Long- chain Cationic Micelles. (Dorrance & Hunter) . Adhesion at the Alkane/Water and Ester/Water Interfaces. (Aveyaid, Bri'scoe & Chapmanj Aerosol Formation from Sulphur Dioxide in the Presence of Ozone and Olefinic Hydro- carbons. (Cox & Penkett) Colloid and Surface Chemistry of Polymer &ti&. Part 1 :'Adso&ion and *Wetting Be: haviour of n-Alkanols. (Ottewill & Vincent) Contact Angle Studies of Some Low Energy Polymer Su;.faces: (Murphy, Robkrts & Rossj Dynamic Contact Angles. Part 7: Impact Spreading of Water Drops in Air and Aqueous Solutions of Surface Active Agents in Vapour on Smooth Paraffin Wax Surfaces.(Elliott & Ford) Fluorocarbon SurfactantlVjater Mesomorphic Phases. 'Part i : A Study of the Ammonium Perfluoro-octanoate+ Water System by Optical Microscopy and Low Angle X-ray Diffraction. (Tiddy) . Fluorocarbon Surfactant/Water Mesomorphic Phases. 'Part 3 : 2H'and iH Puised N.M.R. Study of Water in the Lamellar Phase of the Ammonium Perfluoro-octanoate. (Tiddy) FJuorocarbon SurfactantlWater Mesomorphic Phases. Part 4: Study of the Distribution of Alkyl Chain Mobilities in the Lamellar Phase of the Sodium 12,12,13,13,14,14,14- PAOB 1947 696 1793 832 1712 1617 1451 773 221 1029 1413 1884 978 600 21 50 2266 2203 1423 1406 1586 1312 10 1735 1533 1190 1814 608 653SUBJECT INDEX-VOLUME 68, 1972 3 PAQP Heptafluoromyristate+D20 System by 19F and 1H Spin Echo Nuclear Magnetic Resonance.(Tiddy) . . Light Scattering Studies on n-DodecyltrirnethylaIknonkm Bromide and n-Dddecylpyridi- niumIodide. (Jones & Piercy) . . . . . . . Measurement of Coagulation Forces by Ultracentrifugation. (Meiville & W i k ) N.M.R. Relaxation Times of the Lamellar Phases of the System Sodium Caprylate/Decanol/ Water. (Tiddy) . . . Propagation of Surface Tension Changes over a Surfak withLimiied Aiea. iLucassen & Barnes) . . . . . . . Salt Effects in Foam Films: (Inhami : Solubilization of p-Xylene in Sodium Dodecyl Sulphate 'Micelies. (Fox, Robb'& Smith) : Stability of Aqueous Films on Hydrophobic Methylated Silica. (Blake & Kitchener) . I, 5 Combustion and Flames (including explosions, shock waves; see also I,9) Interpretation of Collisional Ionization Rates in Flames.(Priest) . . . Kinetics of Thermal Ionization of Alkali Metals in Flames. (Hayhurst & Teiford) . Comparison of the Transport Properties of Normal and Expanded Forms of a Cation- exchange Membrane. Part 2 : Self-diffusion and Electrical Properties of Membranes in the Sodium Form in Concentrated Sodium Chloride. (Ferguson, Gardner & Paterson) Comparison of the Transport Properties of Normal and Expanded Forms of a Cation- exchange Membrane. Part 3 : Application of Irreversible Thermodynamics and Nernst- Planck Theories to Membranes in Concentrated NaCl Solutions. (Gardner & Paterson) Double Membrane Diaphragm Technique for Absolute Measurements of Diffusion CO- Effect of Salts on the Diffusion of Dissolved Non-electrolytes.(Tham & Gubbins) . Liquid State Thermal Conductivity of n-Paraffin Hydrocarbons. (Kandiyoti, Mclaughlin Lithium+ Methvlamine Solutions. Part 2: Viscosiiies. iYamamoto. NakamGa & I, 6 Diffusion (including thermal diffusion, viscosity, thermal conductivity; see also II,8) efficients. (Elrick, Smiles & Wooding) . * . . . & Pittman) . . . . Shimoji) . . . . . . . . . . Physical Properties of Heavy-Oxygen Water. Absolute Viscosity of H2180 'between 15 and 35°C. (Kudish. Wolf & Steckel) . . . . . . . . . Pressure and Temperature Dependence' of the Self-diffusion of Benzene. (McCool, Coolings & Woolf) . Pressure and Temperature Dependence of' the Self-diffusion of ' Carbon Titrachioride: (McCool & Woolf) .Self-diffusion and Conduction in Hydrogen-Bonded Solids. Part 1 : Labilk Proton Diffusioi in Pivalic Acid. (Hood, Lockhart & Sherwood) . Self-diffusion and Conduction in Hydrogen Bonded Solids. Part 2:. Se1f:diffus;on ~k C~HS~~COOH and C6H5C003H in Single Crystals of Benzoic Acid. (McGhie & Sherwood) . . . . . . . Thermal Conductiviiies of Gaseous ' Mixtures Containing Hydrocarbons. Part 1 : n- Butane, iso-Pentane and neo-Pentane, and Their Binary Mixtures With Argon. (Parkinson & Gray) . . . Thermal Conductivities of Gaseous Mixtures Containing Hidrocarbons. Pak 2: Cyc10: propane, Propene, But-1-ene and trans-But-2-ene, and Their Binary Mixtures with Argon. (Parkinson, Mukhopadhyay & Gray) .Transition State Theory of Zeolitic Diffusion. Diffusion of CH4 and CFq in' 5A Zeolite: (Ruthven & Derrah) . . . . . . Viscosity of the Isotopes of Hydrogen and their Intermolecular Force Potentials. (Kestin; Ro & Wakeham) . * . . . . . . I, 7 Electrochemistry (including electrolytes, activity coefficients, electrical conductivity, electrode processes) Anodic Displacement of Adsorbed H in Electrochemisorption of Organic Molecules at Conductance Study on the Association Behaviour of Quaiexmar; Arkonium Salts in Hydrogen-bonding Solvents. (Geoffredi & Triolo) Dilute Aqueous Solutions of Unsymmetrical Quaternary Akmonium Iodides. Part 2: Conductance Measurements. (Lowe & Rendall) . Effect of Irradiation on Electrode Processes. Part 1 : The Hydrogen Eiectrode in Dilutk Solutions.( Airey) . Effect of Pressure on Electrical Conductivit'ies of Fused Alkali Metal Halide; and' SilveE Halides. (Cleaver, Smedley & Spencer) Electrical Conductance in Molten KCl, KI, TI1 and KILTlI.' (Bu'ckle & Tsaoussoglou) Electrical Conductivity of Tetracyanoquinodimethane Crystals. (Hurditch, Vincent & Wright) . Electro-catalytic Rediction' of Eihylenk on Gold and other Sibstrafes. iByrne & Kuhn) : Platinum. (Conway, MacDougall & Kozlowska) . . . 670 1839 450 3 69 2129 2230 445 1435 661 237 2021 2030 591 1339 860 135 2041 1489 1971 736 533 1065 1077 2332 2316 1566 2324 2191 1299 1720 1024 465 18984 SUBJECT INDEX-VOLUME 68, 1972 Electrocatalytic Reduction of Ethylene on Platinum and Ruthenium.(Byrne & Kuhn) . Electrocrystallization of Ruthenium and Electrocatalysis of Hydrogen Evolution. (Fleisch- E.M.F. Measurements with Solid Eleckolyte Galvanic Cells on the LimeCZirconia System. Hydrogen Ion Mobility in Aqueous Electrolyte Solutions. Comparison of Polarographic and Diaphragm Cell Methods. (Roberts & Northey) . Interaction of Inert Gases with Ionic Melts. Solubility of He, Ar,'N2, 0 2 and CH4'in the (Na,K)NO3 Eutectic Solvent. (Paniccia & Zambonin) Ionic Conductances in Aqueous Solution and an Empirical Conductance Equation. (Davies) Ionic Mobility in Mi&ocapillaries. A Test 'for Anomaious Water Strucfures. * (Anderson &Quinn) . Mechanistic Conclus'ions from 'the Curvaiure of Soivent Isotope Effects. * (Albery & Davies) Molar Conduct'ivity of Sodium Fluoride in' Aqueous Solution at 25'C.- Applicatibns 0.f Pitts' Conductivity Equation.(Duer, Robinson & Bates) . Oxygen Electrode Reaction. Part 2: Behaviour at Ruthenium Black Eiectrodes. (Burke Plateau Potentials of' the ;+/I Palladim Hydride Electrode at Temperatures betwken 25 Ring-Disc Electrodes. Part 16: A Comparison of kalytidal and Numerical Solutions: Some Expenmental Factors which Goiern the Potential of the'Palladium Hydride Ele'ctrode Standard Potentials of Silver/Silver Halide Electrodes and Ion Solvafion in Gly &rol +'Wate; Studies of Free Radical Reactions in Molten'Salts by GI&-Discharge Electrolysis. Part 1 ': The Oxygen Electrode. Part 3: Inhibition of the Oxygen Evoluiion Reaction. (Burke; Use of a Perchlorate-responsive Membrane Cell for the Determination of Activity Co- mann & Grenness) .. Determination of the Phase Relationships. (Pizzini & Morlotti) . . . . & O'Meara) . . . . and 195°C. (Dobson, Dagless & Thirsk) . (Albery.&Drury) - Mixtures at 25°C. (Khoo) . . at 25 to 195°C. (Dobson, Dagless & Thirsk) Molten Nitrates. (Hamilton & Ingram) McCarthy & O'Meara) . . . efficients of Aqueous Perchloric Acid. (Mussini, Galli & Dubini-Paglia) . I, 8 Kinetics of Reactions (including photochemistry, reactions of gases, solids, liquids, two phase-systems, such as gaslsolid; see also I,2) Addition of Formaldehyde to Slowly Reacting Hydrogen + Oxygen Mixtures. (Baldwin, Fuller, Longthorn & Walker) Analysis of Kinetic Data for a First-order Reaction' with' UnGown 'Initial and Finai Readings by the Method of Non-linear Least Squares.(Moore) Arrhenius Parameters of the Reaction CH3 - +C2H6+CH4+C2Hs. (Pacey & Purnell) .* Chlorine Abstraction Reactions of Fluorine. Part 2: The Kinetic Determination of the Bond Dissociation Energy D",CC12F--CI). (Foon & Tait) . Chlorine Abstraction Reactions of Fluorine. Part 3 : Thermochemical 'Data ' for Chloro: fluoroalkanes. (Foon & Tait) Competitive Study of the Reactions Br+Rk+Brz+R (R 2 CF;, C2F5). (Ferguson & Whittle) . . . . . . Decomposition of &oia in a Microwave Disdharge. (Barker)' Electron Attachment and Negative Ion-Molecule Reactions in Nitrous Oxide. ' (Parkes) : Electron Spin Resonance Studies of Elementary Processes in Radiation- and Photo- chemistry.Part 8 : Thermal and Photochemical Annealing of y-irradiated Glycine. (Ayscough & Mach) . . . . . . Electron Spin Resonance Studies of' Elementary Processes in Radiation- and Photo: chemistry. Part 9 : Carboxylic Esters at Cryogenic Temperatures. (Ayscough & Oversby) . . . . . Electron Spin Resonance Studies of Elementary Processes in Radia'tion- 'and Photo: chemistry. Part 10: Carboxylic Acid Anhydrides at Cryogenic Temperatures. (Ayscough & Oversby) . Electron Spin Resonance Studies of Elementaiy Prdcesses' in Radiation- and Photb-chemistr;. Part 11 : Radical Reactions of a-Aminobutyric Acid between 26 and 4-40 K. (Ayscough & Olsen) . . Electron Spin Resonance Studies'of Photo-Oxidation b; Metal Ions' in Rigid Media at Low Temperatures.(Greatorex & Kemp) . Electron Spin Resonance Studies of Photo-oxidation b; Metal Ions in Rigid Media'at low Temperatures. Part 4: Survey of Photo-oxidation by the Uranyl Ion. (Greatorex, Hill, Kemp & Stone) . Energy Partition in the Photolysis of I% and of His. (Older~haw,'Port~r & Sniith) : PAGE 355 2305 1601 1528 2083 1117 744 167 71 6 839 764 456 749 554 785 1086 1322 1362 1890 1462 1 04 1121 64 1 31 5 2103 1139 1153 1164 1635 121 2059 2218SUBJECT INDEX-VOLUME 6 8 , 1972 5 PAGE Excited Products in the Photodissociation of Methyl Thiocyanate and Isothiocyanate in the Vacuum Ultra-violet. (D'Amario, Di Stefano, Lenzi & Mele) . Flash Photolysis and Electron Spin Resonance Study of p-Chloranil. (Wong, Fa&; Green &Wan) .* . . . . . Flash Photolysis of H2S. iLangford & Oldershaw) . . . . . Flash Photolysis of Naphthalenedicarboxyiic Anhydrides. (Lohmann) . * . Formation of Nitrous Acid in a Gas-phase Stirred Flow Reactor. (Graham & Tyler) . Free Radical Addition to Olefins. Part 8: Addition of n-Heptafluoropropyl Radicals to Free Radical Addition to Olefins. Part 9: Addition of Methyl'Radicals to Fluoro-ethylenes. Kinetic Equations for Consecutive Reversible Reactions with Special Reference to Protein Kinetic Hydrogen Isotope Effect in the Fl'uorination of H2, CH; and CHC13. &on; Kinetic Investigation of ;he Lanthanideim-Nitrate ' Comilexation Reaction. (Silber; Fluoroethylenes. (Tedder, Walton & Winton) . . . . . (Tedder, Walton & Winton) . . . . . . . .Denaturations. (Hijazi & Laidler) . . I Reid & Tait) . .~ Scheinin, &kinson & Grecsek) . . . Kinetics and Thermodvnamics of the Comdex-formation of Niikelfli) and CobaltCZI) Ions with h o 6 a and Pyridine-2-azodimethylaniline in Water; at Pressures from Kinetics and Thermodynamics of the Formation of the Donor-Acceptor Complex of Tet- racyanoethylene with Hexamethylbenzene in Solution, by a Microwave Temperature- jump method. (Caldin, Crooks, O'Donnell, Smith & Toner) . . . . . Kinetics of Hydrogen Abstraction by Difluoroamino Radicals from n-Propyl Formate and the n-Propoxycarbonyl Radical Decomposition. (Cadman, White & Trotman- Dickenson) . Kinetics of Reaction of Metal Alkyls with Alkenks. Part 71 n-Butyl Lithium and N: N.: N': "-Tetramethyl Ethylene Diamine with Butadiene.(Hay, McCabe & Robb) . Kinetics of Reaction of Metal Alkyls with Alkenes. Part 8: Oligomerization of Ethylene by n-Butyl-lithium-NNN'N'-Tetramethylethylenediamine. (Hay, McCabe & Robb) . Kinetics of Ternary Complex Formation between Magnesium Species and 5-Nitrosalicylic Acid. (Cayley & Hague) . . . . . . Kinetics of Ternary Complex Formation Beiween'Nickel(1I) Species and Pyridine-2-azo-p- dimethylaniline. (Cobb & Hague) . . . . . . Kinetics of the Oxidation of Formic Acid by Aquomanganese(II1) Ions in Aqueous Per- chlorate Media. (Wells 8c Whatley) . . Kinetics of the Reaction Between HBr and CF3Br. ' Determination of the Bond Dissociation Energy D(CF3-Br). (Ferguson & Whittle) . . . Kinetics of the Reaction Between HBr and CzFsBr.Determination of the Bond Dissocia- tion Energy D(CzF5-Br). (Ferguson & Whittle) . . . Kinetics of the Reaction between Methyl Iodide and Silver Nitraie in Aqueous Solution. Kinetics of the 'Readions of Hydrogen Ato'ms with Ethylene and Vinyl' Fluohde. (Teng & Jones) . . Kinetics of the Reactions of Oxygen Atoms and 'Nitrogen Atoms'with Sulphur Trioxide. (Jacob & Winkler) . . . Lithium Hydroxide and the Dissociation Kinetics'of Water Vapok. (Cithro '& Mackie). Negative Ion Reactions in Nitrous Oxide+Carbon Dioxide Mixtures. (Parkes) . . Oxidation of Ferrous Ions by Hydroxyl Radicals. (Jayson, Parsons & Swallow) Oxygen Negative Ion Reactions with Carbon Dioxide and Carbon Monoxide. Cparkes) Photochemistry and Radiation Chemistry of Anthraquinone-2-sodium-sulphonate in Aqueous Solution.Part 1 : Photochemical Kinetics in Aerobic Solution. (Clark & Stonehill) . . . Photochemistry and Radiaiion Chemistry of 9,1O-Anthraquinone-2-sodium Sulphonate in Aqueous Solution. Part 2: Photochemical Products. (Clark & Stonehill) . . Photochemistry of Anhydrides. Part 1 : Photolysis of Perfluoroacetic Anhydride Vapour: A New Source of CF3 Radicals. (Chamberlain & Whittle) . Photochemistry of Anhydrides. Part 2: Photolysis of Perffuoropropionic ' Anhidride': A New Source of C2F5 Radicals. (Chamberlain & Whittle) . . . Photochemistry of Aromatic Compounds. Part 5 : Measurement of Quantum Yields of Photochemistry of 1,3-Dichloro-tetr~uoroacetone. Part 1 : The Singlet State.(I.iacke't Photochemistry of 1,3-Dichloro-tetrafluoroacetone. Part 2: 'The Triple; State. O-Iackett Photochemistry of 113-Didhloro~tetrafluoroacetone. Part 31 Quenching of 'the Excited Photolysis of Dibromodifluoromethane at 265 nm: (Waltonj : ; : : 1 1 bar to 2 kbar. (Caldin, Grant & Hasinoff) . . . . . . . . . . . . Triplet State Formation by Flash Photolysis. wander Donckt & Leitaer) & Phillips) . . , . . . . & Phillips) . . . . States. (Hackett & Phillips) . 940 221 1 1550 814 683 160 1866 1235 1131 1 200 2247 849 506 1 1227 2259 932 434 295 306 1824 1267 2077 150 2121 2053 627 577 1676 88 96 112 323 329 335 15596 SUBJECT INDEX-VOLUME 6 8 , 1972 Prediction of Transition State Configuration in Concerted Reactions from the Energy Requirements of the Separate Processes. (Critchlow) Primary Photochemical Processes in Aromatic Molecules.Part 15 : The Photochemistry of Aromatic Carbonyl Compounds in Aqueous Solution. (Ledger & Porter) Protonation of Aromatic Carboxylic Acids in the First Excited Singlet State. (Watkins) . Quenching and Isomerization in the Photolysis of trans-But-2-ene and cis-Pent-2-ene at 185 nm and 203 nm. (Borrell & Cervenka) Quenching Effects Upon the Life-Time of Protonated 2:Naphthami'de Sl 'in Aqueous Solu: tion. (Hussain &Wyatt) . Radiation Chemistry and Photochemistry of Oxychlorine Ions. ' Part' 1 : Radiolysis of Aqueous Solutions of Hypochlorite and Chlorite Ions. (Buxton & Subhani) Radiation Chemistry and Photochemistry of Oxychlorine Ions. Part 2 : Photodecomposi- tion of Aqueous Solutions of Hypochlorite Ions.(Buxton & Subhani) . Radiation Chemistry and Photochemistry of Oxychlorine Ions. Part 3 : Photodecomposi- tion of Aqueous Solutions of Chlorite Ions. (Buxton & Subhani) . Rate of Combination of Trimethylsilyl Radicals in the Gas Phase. (Cadman, Tilsley & Trotman-Dickenson) . Reaction between Ammonia and Nitrogen Dioxide. (Bedford & Thomas) ' Reaction of C2FS Radicals with HCI. Determination of the Bond Dissociaiion Energ; D(C2Fs--H). (Bassett & Whittle) Reaction of Hot Hydrogen Atoms with COS. (Oldershaw & Porter) Reaction of Oxygen Atoms with Alkynes. Part 1: The Butyne-2 Reaction.' (Aiery & Heath) Reactions of CF3 Radicals with Benzotrifluoiide and the C-H Bond Strength in CsHsCF; and C6H6.(BCrces, Mfirta & Szikigyi) Reactions of Cyanogen Radicals. Part 2: Reactions'with '(CN); and' 0 2 . '(Bullock & Cooper) Reactions of banogen Radicals. Part 3: kheniu; Parameters fo; Reactions with Alkanes. (Bullock & Cooper) . . Reactions of Group 3 Metal Alkyls in the' Gas Phase'. Pait 8: Homogeneous Therrnai Unimolecular Elimination of Ethylene from Triethylaluminium. (Cocks & Egger) . Reactions of Group 3 Metal Alkyls in the Gas Phase. Part 9: Addition of Propene to Trimethylaluminium. (Egger) . . . Reactions of Perfluoroethyl Radicals with Propaie and Neopentane, iWhyiock, Clark; &Gray) . Reactions of the 0- Negative Ion with Hydrogen Hnd the Lower Hidrocarbon's. (Parkesj Reactions of Trifluoromethyl Radicals with Organic Halides.Part 7 : Chloro- and Fluoro- chloro-ethanes. (Quick & Whittle) . . . Reduction of Methylene Blue by Dihydro-compounds, Cat'alyzed by Aliphatic Amines; and the Reaction of Methylene Blue with Aliphatic Amines. (Iwasawa, Soma, Onishi & Tamaru) Role of Metals in Enzymat'ic Reaction's. Part 5: Kinetics of'Ternary Complex Formation Between Magnesium and Manganese(I1) Species and 8-Hydroxyquinoline. (Hague, Martin & Zetter) Self-heating and Spontaneous Ighition' in the Oxidation of Gaseoui Hydrazine. (Gray & O'Neill) Specificity in the Intkractibn be'tween' some Chiial Co(II1) Complex Ions and DNA in Aqueous Solution. (Ascoli, Branca, Mancini & Pipisa) Thermal Decomposition of Ethyl, Isopropyl and t-Butyl Fluorides in the' Gas Phase: (Dastoor & Emovon) .Thermal Isomerization of Cyclobutenes. Part 19: 3,3,4,4-Te~rafluorocy~lobutene. '(Frey; Hopkins &Vinall) . . . Thermal Unimolecular Decomposition of i-Eth&yclohexa-i ,4-diene, 1;2-Dimethyicyclo: hexa-l&diene and Bicyclo[4,3,0]nona-l(6),3-diene. (Cocks, Frey & Hopkins) Threshold Energy in the Abstraction Reaction between Hydrogen Atoms and Cyclohexane.. (Fink & Nicholas) . . . . . I, 9 Polymers and Polymerization (including physical properties of polymers and their solutions ; see also I,8) Analysis of the Reversible Aggregation of Macromolecules in a Centrifugal Field. (Skerrett) Effects of Salts of Metals on Vinyl Polymerization. Part 8: Polymerization of Vinyl Chloride in the presence of Ferric Chloride. (Bengough & Chawdry) .Fluorescence Depolarization Measurements of Polymer Segmental Mobility. (North d Soutar) . Kinetics of Anionic Polymekzation of o-Methylstyiene in 2-Methy1tetrah;drofu;an and Tei- Kinetics of Anionic Polymerizations of Styrene and Its m- and p-Derivatives: Hammett's Relations. (Hirohara, Nakayama & Ise) . , . . . , , . . rahydrofuran. (Hirohara, Nakayama, Kawabata & Ise) . . PAQE 1774 539 28 345 130 947 958 970 1849 21 63 492 709 512 867 21 75 21 85 423 1017 689 61 3 878 1697 37 564 1213 2098 1874 1287 1706 1328 1807 1101 51 $8SUBJECT INDEX-VOLUME 6 8 , 1972 Microwave Absorption in a Helical Polypeptide Molecule. (Davies, Maurel & Price) . Photo-initiation of Polymerization by Manganese(II1) Chelates. (Bamford & Ferrar) , Photolytic Decomposition of Polymethacrylates.Part 3 : Poly(ethy1 methacrylate). Reaction Kinetics of Polymer Substituents.' Neighbok-ing-substituent Effects in Pairing Reaction Kinetics of Polymer Substi;uents: Neighbouring-iubstituent Effects in Single: Surface-Photopolymerization from Hexachlorobuiadiene. (Kunz & Wright) Telomerization of Methvl Methacrvlate with Bromotrichloromethane. (Barson, Luxton (MacCallum & Schoff) . . . Reactions. (Boucher) . . . substituent Reactions. (Boucher) . . . &Robb) . radical Formation. (Bamford & Hughes) . Tetrakis(tripheny1 phosphke)nickel(O) $ Maieic khydiide : Ligand Exdhange and' Free: I, 10 Radiolysis (including nuclear transformation in solids, neutron capture, etc.) Dissociation Equilibrium of Nitroform in Mixtures of Polar Solvents Studied by Pulse Dissociation Equilibrium of Nitroform in Polar' Solvents Studied by 'Pulse' Radiolysis: Effects of Gamma Radiation on'vitamin Bi2 Systems.. (Blackburn, Cox & Phillips) : Energy Partition in the y-Radiolysis of Gaseous H2S+Nz, H2S+Ar and H2SfXe Mix- Interpretation of the Effects of Ionic Scavengers a; High L.E.T. in Irradiated Cyclohexane; Nanosecond Pulse Radiolysis of Acetone. Kinetic and Thermodynamic Properties of Pulse Radiolysis of 9,lO-Anthraquinones. Part 1 : Radicals. * (Hulme, Land & Phillips) Pulse Radiolysis of 9,lO-Anthraquinones. Part 2: Triplet Excited States. (Hulme, Land Pulse Radiolysis of Hexamethylphosphoric Triamide. iNau;a & Van Huis) . Radiation Chemistry of Butan-1-01. Radiolysis of Carboxylic Compounds. Part 1 : Comparison of Potassium Acetate and Acetic Acid Radiolysis. (LukBE, Teplf & Vacek) .y-Radiolysis of Diethyl Succinate. Reactions of the H02 Radical in Aqueous Solution with Bromine and Reiated Compounds. (Sutton & Downes) Unstable Intermediates. Part i07: Radiaiion Damage Products in Choline Chloride. (Symons) , . Radiolysis. (Chaudhri & Asmus) . (Chaudhri & Asmus) . tures. (Ahmad, Huyton & Woodward) Based on Ambipolar Diffusion. (Burns & Reed) . . . Some Aromatic Radical Cations. (Rodgers) & Phillips) . . (Ackerman, Basson & Van Der Liqde) (Mills & Nosworthy Peto) I, 11 Solid-state Chemistry Crystal Growth in Manganese-Doped Magnesium Oxide Powders. . Dislocations in Orthorhombic Minerals. Part 1 : Naturally-occurring Barytes.(Williams, Electrical Conduction in Ammonium Perchl'orate.' (Owen, Tnama's & Williams) * Influence of Framework Charge Density on Ion-exchange Properties of Zeolites. (Barre; Lattice Defects in Plastic Organic Crystals. ' Part'7: Isotope 'Mass 'Effecis in Self-diffusion Magnetic Studies of Zeolites. Part 1 : The Magneiic Properties of' COY and' CoA. Self- and Impurity Diffusion in Anthracene Single Crystals. '(Burns & Sherwood) : : I, 12 Thermodynamic and Equilibrium Properties (including multiphase systems ; see also I,7) Activity Coefficients for the System HCl-MnC12-H20 at 25°C. (Downes) . Apparent Molar Volumes of Multiply Charged Electrolytes. (Indelli & Zamboni) . Complex Formation in Molten Salts. Part 2: Association Constants of Silver Chloro- complexes in Molten Eutectic NaN03+Ba(N03)2*.(Gaur & Bansal) . Component Interactions in Aqueous Acetone. (Fox) . Cryoscopic Comparison of Nitric and Deuteronitric Acids as Solvents. (Kureki & Wyatt) . . . Dipole Moments of Election Donor-Acceptor Complexes ok Iodine with Oxygen 'Bases: (Sobczyk & Danel) Drop Calorimetric Determination of Enthaipy Content'of the Systems AgIt-MI (M = K; Rb, Cs, NH4, (CH3)4N). (Johnson & Dudley) . Effect of the Solution Vapour Pressure on the Temperature Dependence of the Dissociation Constant of Acetic Acid in Water. (Lown & Thirsk) . (Guilliatt & Brett) Tennakoon, Thomas & Jacobs) & Klinowski) . in Cyclohexane. (Chadwick & Sherwood) (Egerton, Hagan, Stone & Vickerman) .Enthalpy of Solvation of Carbonium Ions. (Gold) . . . P A W 1041 1243 499 228 1 2295 140 1666 1474 1010 385 1687 1857 67 1278 1992 2003 647 1258 1377 1626 1498 216 429 1987 2356 1956 47 723 1036 1964 1831 1368 1294 676 1544 201 5 1982 16118 SUBJECT INDEX-VOLUME 68, 1972 Excess Enthalpies and Weak Interactions in Liquid Mixtures of Methylene Chloride with Benzene, Toluene and Xylenes. (Nigam & Mahl) . Excited Singlet and Triplet pK Values of Xanthone in Aqueous' Solukon. . (Ireland & Wyatt) . Flame-Photometric Determination of *the Standaid Enihalpiks of 'Formation 'of Ai(0H); and A10. (Jensen & Jones) . . . Heats of Dilution of Sub-micellar Aqueous Surfadtant Solutions. iBirch & Hall) . Ion Association of Rubidium Chloride in Aqueous Solutions at 25°C.(Dunsmore, Jalota & Paterson) . Ionic Association of Some Tetraheptylammonium and Tetraoctylphosphonium Salts in Ethanol Studied by Vapour Pressure Osmometry. (PaligoriC & Gal) . Kinetic Isotope Effects. Part 7 : Ionization of (d)-Phenylmethylacetophenone. iEarls; Jones & Rumney) . Mass Spectrometric Determination of the Thermodynamic Properties of the Vapour Species from Alumina. (Farber, Srivastava & Uy) . Measurements of Heats of Combustion by Flame Calorimetry. Part 8: Methane, Ethane; Microcalorimetric Studies : Enthalpies of Formation of Amino-acid and Peptide Complexes of CopperOI) and Nickel(1I). (Tipping & Skinner) Microcalorimetric Studies. Thermal Decomposition and Iodination of 'Metai Carbonyls: (Connor, Skinner & Virmani) .Phase Diagram and Thermodynamic Data kor t'he HydrogenlVanadium System: (Griffiths, Pryde & Righini-Brand) . Physico-chemical Studies in Non-aqueous Solvents. Part 61 Cryoscopi'c Studies 0; 1 : i Electrolytes in Formamide. (Paul, Singla & Gill) Polarographic Determination of Hydrogen Ion Activities in Sirongl y Acidic Media: A Nek Acidity Function. (Janata & Jansen) . . . Solid/Liquid Phase Equilibria and Solid Compound Formation in' Mixtures of Dimethyl: . Solubility of Cadmium Cyanide and the Formation Constants of the Cadmium-Cyanide Solubility of p-Nitroaniline and the Basic Strength of Aqueous Alcohols (and oiher Oxygen- Standard Enthalpy of Formation of Crystalline Gold(III) Okide. * (Ashcroft '& Schwarz: Studies in Ion Solvat'ion in Non-aqueous Sdlvents and their Aquedus Mixture's.Part 14. Free Energies of the Alkali-metal Chlorides from Water to 10-99 % (w/w) Methanol+ Water Mixtures at 25°C. (Feakins & Voice) . Thermodynamic Studies of Hydrochloric Acid in Propan-2-01 from Electromotive ' Force Measurements between 5 and 45°C. (Roy, Vernon & Bothwell) Thermodynamic Study of Disorder in Zinc Fluoride Tetrahydrate. (Cook, Daiies & Staveley) . Thermodynamics of Autoionizatibn of Methanol f'Wate; Mixiures. (Parsons & Rochesterj Use of p-Nitroaniline to Determine Protonation Equilibrium Constants of Oxygen-contain- ing Molecules in Aqueous Solution. (Wells) Vapour-Liquid Equilibrium and Excess Free Energies for Benzene+ Dioxan and Carbon Tetrachloride+ Dioxan Systems.(Deshpande & Oswal) . . . . Propane, n-Butane and 2-Methylpropane. (Pittam & Pilcher) . . . sulphoxide with c c 4 , CHC13, and CC13CHC12. (Goates, Ott, Reeder & Lamb) System in Dimethylformamide at 25°C. (Jagtiani, Siddiqi & Tyrrell) Containing Compounds. (Sierra, Teixid6 & Wyatt) mann) . . . . . . PAGE 1508 1053 259 2350 1583 1093 925 249 2224 1764 1754 2344 1894 1656 2171 2090 290 1360 1390 2047 1384 523 993 1059AUTHOR INDEX-VOLUME 68, 1972 PAGE Ackerman, L. 4;. J., Basson, R. A. and Van Der Linde, H. J. . 1258 Ahmad, M., Huyton, D. W. and Wood: Airey, P. L. . 1299 Albery, W. J. and Davies, M.. H. : . 167 - and Drury, J. S. . . 456 Alexander, C. S. and Pritchaid, J. 1 . 202 Anderson, John L., and Quinn, John A. . 744 Ascoli, F., Branca, M., Mancini, C.and Pispisa, B. . 1213 Ashcroft, S. J. and Schwarzmann, E. . 1360 Asmus, K.-D. See Chaudhri, Shamiiz A. and Atkinson, G. See Silber, H. B., Scheinin, N., Atkinson, G. and Grecsek, J. J. Avery, H. E. and Heath, S. J. . . 512 Aveyard, R. and Briscoe, B. J. . . 478 -and Briscoe, B. J. and Chapman, J. . 10 Ayscough, P. B. and Mach, K. . . 1139 - and Olsen, K. J. . 1635 - and Oversby, J. P. . : 1153, 1164 ward, T. W. . . 1857 Baldwin, R. R., Fuller, A. R., Longthorn, D. and Walker, R. W. . 1362 Ballivet, D., Barthomeuf, D. and Pichat, P. 17 12 Bamford, C. H. and Ferrar, A. N. . . 1243 - and Hughes, E. 0. . . . 1474 Bansal, N. P. See Gaur, H. C and Baranowski, B., Lewis, F. A., Majchrzak, S. and Wisniewski, R. . 824 Barker, R. . . 315 Barnes, G.T. See Lucassen, J. and Barrer, R. M. and Klinowski, J. . 73, 1956 Barson, C. A., Luxton, A. R. and Robb, J. C. . . 1666 Barthomeuf, D. ske Bailivet, * D., jartho- meuf, D. and Pichat, P. . Barton, S. S., Beswick, P. G. and Harrison, B. H. . . . 1647 Bassett, J. E. and'Whitbe, E. . . 492 Basson, R. A. See Ackerman, L. G. J., Basson, R. A . and Van Der Linde, H. J. Bates, Roger G. See h e r , Wayne C., Robinson, R. A . and Bedford, G. and Thomas, J. H. . 2163 Bengough, W. I. and Chawdry, N. M. . 1807 Berces, T., Marta, F. and Szilagyi, I. 867 Beswick, P. G. See Barton, S. S., Beswick; P. G . and Harrison, B. H . Birch, B. J. and Hall, D. G. . . 2350 Blackburn, R. Cox, D. L. and Phillips, G.O. . . 1687 Blake, T. D. and Kitchener, J. A. : . 1435 Borrell, Peter and Cervenka, A.. 345 Bothwell, A. L. M. See Roy, k. N., Boucher, E. A. . . 2281, 2295 Branca, M. See Ascoli, F., Branca, M., Vernon, W. and Mancini, C. and Pispisa, B. PAGE Brett, N. H. See Guilliatt, I. F. and Briscoe, B. J. See Aveyard, R. and - S e e Aveyard, R., Briscoe, B. J. and Bruce, Linda A. and Sheridan, Margaret Buckle, E. R. and'Tsaoussoglou, P.'E. Bullock, G. E. and Cooper, R. Burke, L. D., McCarthy, F. and O'Meara, Burlamacchi, L., Martini, G. and Ferroni, E. . 1586 Burns, G. and Sherwood, J. N. : . 1036 Burns, W. G. and Reed, C. R. V. 67 Buxton, G. V. and Subhani, M. S. '947, 958, 970 Byrne, M. and Kuhn, A. T. . . 355, 1898 Chapman, J. H. . 997 1024 21j5, 2185 T. 0. . . . . 1086 - and O'Meara, T. 0. . * 839 Cadman, P., Tilsley, G. M. and Trotman- - White, A. J.and Trotman-Dickenson, A. F. 506 Caldin, E. F., Crooks, J. E.; O'Donnell, D., Smith, D. and Toner, S. . 849 - Grant, M. W. and Hasinoff, B. B. . 2247 Cathro, W. S. and Mackie, J. C. . . 150 Cayley, C. R. and Hague, D. N. . 2259 Cervenka, A. See Borrell, Peter and Chadwick, A. V. and Sherwood, J. N. . 47 Chamberlain, G. A. and Whittle, E. 88, 96 Chapman, J. See Aveyard, R., Briscoe, Chaudhri, Shamin A. and Asmus, K.-D. 385, 1011 Chawdry, N: M. See Beigough, W.'I. and Chuang, T. T. and Dalla Lana, I. G. 773 Clark, K. P. and Stonehill, H. I. . 577, 1676 Clarke, J. D. See Whytock, D . A., Clark, J . D. and Gray, P. Clarke, J. K. A. See Plunkett, T. J . and Cleaver, B., Smedley, S. I. and Spencer, P. N. . . . . 1720 Clint, J. €3. . . 2239 Cobb, M. A. and k g u e , D.N. . . 932 Cocks, Alan T. and Egger, Kurt W. . 423 Cocks, A. T., Frey, H. M. and Hopkins, R. G. . . 1287 Callings, A. F. 'See ;Mccok, h;I. A., Collings, A. F. and Woou, L. A. Connor, J. A., Skinner, H. A. and Virmani, Y. 1754 Conway, B: E., . MacDougall, B. ani Kozlowskn, H. A. . . 1566 Cook, R. O., Davies, A. and Skveley, L. A. K. . 1384 Cooper, R. See Bullock,' G. E.' and * Cortes Arroyo, Antonio and Kemball, C. . 1029 COY, D. L. See Blackburn, R., Cox, D. L. Dickenson, A. F. . . 1849 B. J. and and Phillips, G. 0. Cox, R. A. and Penkett, S. A. . . 1735 Critchlow, J. E. . . 1774 910 AUTHOR INDEX-VOLUME 68, 1972 PAQE Crooks, J. E. See Caldin, E. F., Crooks, J . E., O’DonneII, D., Smith, D. and Toner, S. Cullis, C. F., Nevell, T. G. and Trimm, D.L.. . 1406 Dadiza, Y. Ad. and Saleh, J. M. 269, 1513 Dagless, M. N. See Dobson, i. V., Dugless, M . N. and Thirsk, H. R. Dalla Lana, I. G. See Chuang, T. T. and D’Amario, P., Di Stefano, G., Lenzi, M. and Mele, A. . 940 Danel, J. See Sobczyk, i. and Dastoor, P. N. and Emovon, E. U. . 2098 Davidson, D. E. See Howe, R. F.,Davidson, Davies, A. See Cook, R. O., Davies, A . Davies, Cecil W. . 1117 Davies, M. H. See Albeiy, W: J. and Davies, Mansel, Maurel, P. and Price, A. H. . 1041 Davis, K. hi. C. and Sayer, C. F. : . 1884 Derrah, R. I., Loughlin, K. F. and Ruthven, D.M. . . . . 1947 - See Ruthven, D. M. and Deshpande, D. D. and Oswal, S. L. . 1059 Devi, A. See Morrow, B. A . and Di Stefano, G. See D’Amario, P., Di Stefano, G., Lenzi, M . and Mele, A . Dobson, J. V., Dagless, M.N. and Thirsk, H. R. . . . 749, 764 Dollirnore, D., Heal, G. R. and Martin, D. R. . 832 Dorrance, R: C. and Hunter, T. F. * 1312 Dowie, R. S., Whan, D. A. and Kemball, C. 2150 Downes, C. J. . . . 1964 Downes, M. T. See Sutton, €3. C. and Drury, J. S. See Albery, W. J. and Dubini-Paglia, E. See Mussini, T., Galli, R. and Dudley, J. See Johnson, K. E., Sime, S . J. and Duer, Wayne C., Robinson, R. A. and Bates, Roger G. . . 716 Dunsmore, H. S., Jalota, S. K. andhater- son, R. f . . . 1583 Earls, D. W., Jones, J. R. and Rumney, T. G. . . . . . 925 Egerton, T. A., Hagan, A., Stone, F. S. and Vickerman, J. C. . . 723 Egger, Kurt W. . . . . 1017 - See Cocks, Alan T. and Elliott, T. A. and Ford, D. M. . . 1814 Elrick, D. E., Smiles, D. E. and Wooding, R.A.. . 591 Eltekov, Yu. A.,’ Khopina, ‘V. V. and Kiselev, A. V. . . . 889 Emovon, E. U. See Dastoor, P. N. and Fabes, L, See Wong, S. K., Fabes, L., Green, W. J . and Wan, J . K . S . Farber, Milton, Strivastava, R. D. and Uy, O.M. . . . . 249 Feakins, D. and Voice, P. J. . . 1390 Ferguson, H., Gardner, C. R. and Pater- son, R. 202 1 Ferguson, K. C. aid Whittle, E. 295, 306, 641 D. E. and Whan, D. A. and Staveley, L. A . K . PAQE Ferrar, A. N. See Bamford, C . H. and Ferroni, E. See Burlamacchi, L., Martini, Findenegg, G. H. . . . 1799 Fink, R. D. and Nichokk, J. E. . . 1706 Fleischmann, M. and Grenness, M.. . 2305 Foon, Ruby, Reid, G. P. and Tait, K. B. . 1 13 1 Ford, D. M. See Elliott, T. A.’ and Fox, K. K., Robb, I. D. and Smith, R. . 445 Fox, Malcolm F. . 1294 Frey, H.M., Hopkins, ‘R. G: and Vinall; I. c. . 1874 -See Cocks, A.‘ T., Frey, ‘H. i. and Hopkins, R. G. Fuller, A. R. See Baldwin, R. R., Fuller, A . R., Longthorn, D. and Walker, R. W . G. and - and Tait, K. B. . . 104,1121 Gal, I. J. See Paligoric, 1. and Galli, R. See Mussini, T., Galli, R . and Dubini-Paglia, E. Galwey, A. K. and Gray, P. . Gardner, C. R. and Paterson, R. -See Ferguson, H., Gardner, C. R. and Gaur, H. C. and Bansal, N. P. Giles, J. See Tench, A. J. and Gill, Dip Singh. See Paul, Rum Chand, Singla, Jai Parkash and Goates, J. R., Ott, J. B., Reeder, J. and Lamb, J. D. . Goffredi, M. and Triolo, R. . Gold, V. . . . Goymour, Clarence G. and King, David A. Graham, R. F. and Tyler, B. J. Grant, M. W. See Caldin, E. F., *Grani, M . W. and Hasinoff, B.B. Gray, P. and O’Neill, E. P. . - See Gufwey, A . K. and - See Parkinson, C. and - See Parkinson, C., Mukhopadhay, P. and - See Whytock, D. A., Clarke, J . D. and Greatorex, D., Hill, R. J., Kemp, T. J. and Stone, T. J. . - and Kemp, T. J. . Grecsek, J. J. See Silber, H. B., Scheinin, N., Atkinson, G. and Green, W. J. See Wong, S. K., Fabes, L., Green, W . J. and Wan, J. K . S. Gregory, J. and Semmens, M. J. . Grenness, M. See Fleischmann, M. and Griffiths, R., Pryde, J. A. and Righini- Gubbins, K. E. See Tham, M.’K. and ’ Guilliatt, I. F. and Brett, N. H. . Puterson, R. . Brand,A. . , 1935 2030 1368 2171 2324 161 1 280 683 564 2059 121 1045 2344 429 Hackett, Peter A. and Phillips, David 323, 329, 335 Hagan, A. See Egerton, T. A., Hagan, A., Stone, F. S. and Vickerman, J.C. Hague, D. N., Martin, S. R. and Zetter, M.S. . . . . . 37 - See CayIey, G. R. and - See Cobb, M. A . and Hall, D. G. See Birch, B. J. and Hamilton, L. W. and Ingram, M. D. . 785AUTHOR INDEX-VOLUME 6 8 , 1972 11 PAQE Harrison, B. H. See Barton, S. S., Beswick, Hasinoff, B. B. See Caldin, E. F., Grant, Hay, J. N., McCabe, J. F. and Robb, Hayhurst, A. N. and Telford,N. R. Heal, G. R. See Dollimore, D., Heal, G. R. Heath, S. J. See Avery, H. E. and Hijazi, N. H. and Laidler, K. J. . . 1235 Hill, R. J. See Greatorex, D., Hill, R. J., Hirabayashi, Kazuo, Saito, Shuji and Hirohara, H., Nakayama, M.’ and Ise, N. - Nakayama, M., Kawabata, R. and Ise, Hockey, J. A. See’Jonei, P. and Hood, G. M., Lockhart, N. C. and Sher- Hopkins, R. 6. See Cocks, A.T., Frey, -See Frey, H. M., Hopkins, R. G. and Howe, R. F., Davidson, D. E. and Whan, P. G. and M. W. and J. C. . 1, 1227 23 7 and Martin, D. R. Kemp, T. J. and Stone, T. J. Yasumori, Iwao . . 978 58 N. . 51 ’ wood, J. N. . . 736 H. M. and Vinall, I. C. D. A. . . 2266 - Liddy, J. P. and Metcalfe, A. . . 1595 - and Metcalfe, A. . 393 Hughes, E. 0. See Bamford, C. H. and Hulme, B. E., Land, E. J. and Phillips, G. 0. . . 1992, 2003 Hunt, G. L. and Ritchie; I. G. . 1413, 1423 Hunter, T. F. See Dorrance, R. C. and Huq, Rokeya 1824 Hurditch, R. J., ‘Vincent, Vera M. and Wright, J. D. . . 465 Hussain, S. K. and Wyatt, P.’A. H: . 130 Huyton, D. W. See Ahmad, M., Huyton, Ichikawa, M. See Naito, S., Ichikawa, M . Indelli, A. and Zamboni, R. . . 1831 Ingram, B. T. . . 2230 Ingram, M.D. See Hamilton,’L. k and Ireland, J. F. and Wyatt, P. A. H. 1053 Ise, N. See Hirochara, H., Nakayama, M, and - See Hirohara, H., Nakayama, A!. , Kawubata, R. and Iwasawa, Y., Soma, M., Onishi, T. and Tamaru, K. . . 1617, 1697 Jackson, P. and Parfitt, G. D. . 896, 1443 Jacob, A. and Winkler, C. A. . 2077 Jacobs, P. W. M. See Williams, ‘J. O., Tennakoon, D. T. B., Thomas, J. M. and Jagtiani, I. A., Siddiqi, I. W. and Tyrrell, H. J. V. . . . . 2090 Jalota, S. K. See Dunsmore, h. S., Jalota, S. K. and Paterson, R. Janata, Jiri and Jansen, Geertje . . 1656 Jansen, Geertje. See Janata, Jiri and Jayson, G. G., Parsons, B. J. and Swallow, A. J. . . 2053 Jensen, D. E. and Jonei, G. A. : . 259 D. W. and Woodward, T . W. and Tamaru, K. Johnson, K. E., Sime, S. J.and Dudley, J. Jones, G. A. See Jensen, D. E. and Jones, J. R. See Earls, D. W., Jones, J. R. and Rumney, T. G. Jones, M. N. and Piercy, J. . Jones, P. and Hockey, J. A. . Jones, W. E. See Teng, L. and Kandiyoti, R., McLaughlin, E. and Pittman, J. F. T. . Kawabata, R. See Hirohara, H., Naka- yama, M., Kuwabata, R. and h e , N. Kemball, C. See Cartes Arroyo, Antonio and -See Dowie, R . S., Whan, D. A. and - See Ross, J. R . H., Roberts, M . W. and Kemp, T. J. See Greatorex, D. and -See Greatorex, D., Hill, R. J., Kemp, T. J. and Stone, T. J. Kestin, J., Ro, S. T. and Wakeham, W. A. Khoo, K. H. Khopina, V. V. ’See Eltekdv, Yu. A.; Khopina, V. V. and Kiselev, A. V. Kibblewhite, J. F. J. See Tench, A. J., Lawson, T. and King, D. A., Madey, T. E. and Yates, Jr., J.T. King, David ‘A. See Goymour,’CIarence G: and Kiselev, A. V., Lygim, V. I. and Staro- dubceva, R. V. . - See Eltekov, Yu. A., Khopina, V. V. and Kitchener, J. A. See Blake, T. D. and Klinowski, J. See Barrer, R. M. and Knapp, A. G. and Stiddard, M. H. B. . Kozlowska, H. A. See Conway, B. E., Mac Dougall, B. and Kudish, A. I., Wolf, D. and Steckel, F. . Kuhn, A. T. See Byrne, M. and Kunz, C. 0. and Wright, A. N. Kureishi, A. W. and Wyatt, P. A. G. : Laidler, K. J. See Hijazi, N. H . and Lamb, J. D. See Goates, J. R., Ott, J. B., Reeder, J. and Land, E. J. See Hulme, B. E., Land, E. J. and Phillips, G. 0. Langford, R. B. and Bldershaw, G. A. . Lawson, T. See Tench, A. J., Lawson, T. and Kibblewhite, J. F. J . Ledger, M. B. and Porter, G. Leitaer, D. See Vander Donckt, E.and Lenzi, M. See D’Amario, P., Di Stefano, G., Lenzi, M . and Mele, A. Lewis, F. A. See Baranowski, B., Lewis, F. A., Majchrzak, S. and Wisniewski, R. Liddy, J. P. See Howe, R. F., Liddy, J. P. and Metcalfe, A. Lockhart, N. C. See Hood, G. M., Lock- hart, N . C. and Sherwood, J. N. Lohmann, J. . Longthorn, D. See Baldwin, R. R., Fuller; A. R., Longthorn, D. and Walker, R. W. Loughlin, K. F. See Derrah, R. I., Loughlin, K. F. and Ruthven, D. M . - See Ruthven, D. M. and . PAQE 201 5 1839 907 860 2316 5 54 1347 1793 2139 2031 140 676 1550 539 81412 AUTHOR INDEX-VOLUME 6 8 , 1972 Lowe, B. M. and Rendall, H. M. . . Lown, D. A. and Thirsk, H. R. . . Lucassen, J. and Barnes, G. T. , . Lukac, S., Teply, J. and Vacek, K. . Luxton, A. R. See Bdrson, C . A., Luxton, A.R. and Robb, J. C. Lygin, V. I. See Kiselev, A. V., Lygin, V. I. and Starodubceva, R . V. McCabe, J . F. See Hay, J. N., McCabe, J. F. and Robb, J. C. NlacCallum, J. R. and Schoff, C. K. . McCarthy, F. See Burke, L. D., McCartliy, I;. and O’Meara, T. 0. II’IcCool, M. A., Collings, A. F. and Woolf, L. A. - and Woolf, L. A. . MacDougall, B. See Conway, b. E.; MacDougall, B. and Kozlowska, H. A. McGhie, A. R. and Sherwood, J. N. . McLaughlin, E. See Kandiyoti, R., Mchughlin, E. and Pittman, J. F, T. Mach, K. See Ayscough, P. B. and Mackie, J. C. See Cathro, W. S. and Madey, T. E. See King, D. A., Madey, T. E. and Yates, Jr., J. T. Mahl, B. S. See Nigam, R. K. and Majchrzak, S. See Baranowski, B., Lewis, F. A,, Majchrzak, S. and Wisniewski, R . Mancini, C .See Ascoli, F., Branca, M., Mancini, C. and Pispisa, B. Marta, F. See Berces, T., Marta, F. and Szilagyi, I. Martin, D. R. See Dollimore, D., Heal, G. R. and Martin, S. R. See Hague, D. N., Martin, S. R. and Zetter, M. S. Martini, G. See Burlamacchi, L., Martini, and Ferroni, E. Maurel, P. See Davies, Munsel, Maurel, P. and Price, A. H. Mele, A. See D’Arnario. P.. DiStefano. G.. , , ” I . Lenzi, M. and Melville. J. B. and Willis. E. Metalfe, A. See Howe, R. F. and - See Howe, R. F., Liddy, J. P. and Mills, K. J. and Nosworthy Peto, J. M. Mohri, M. See Takaishi, T. and Moore, P. . Morlotti, R. See Pizzini,. S. and Morrow, B. A. and Devi, A. . Mukhopadhyay, P. See Parkinson, C., Mukhopddhyay, P. and Gray, P. Murphy, W. J., Roberts, M. W. and Ross, J. R. H. Mussini, T., Gal$ R.and Dubini-Paglia; E. . . . . Naito, S., Ichikawa, M. and Tamaru, K. . Nakamura, Y. See Yamamoto, M., Naka- mura, Y. and Shimoji, M. Nakayama, M. See Hirohara, H., Naka- yama, M. and Ise, N. - See Hirohara, H., Nakayama, M., Kawabdta, R. and Ise, N. Niuta, H. and Van Huis, C. . * PAGB 2191 1982 21 29 1377 499 1489 1971 533 450 1626 1890 403 1190 1322 1451 647 PAOL Nevell, T. G. See Cullis, C. F., Nevell, Nicholas, J . E. See Fink, R. D. and Nigam, R. K. and Mahl, B. S. . . 1508 Northey, H. L. See Roberts, N. K. and Nosworthy Peto, J. M. See Mills, K. J. and O’DonnelI, D. See Caldin, E. F., Crooks, J. E., O’Donnell, D., Smith, D. and Toner, S. Oldershaw, G. A. and Porter, D. A. . 709 - Porter, D. A. and Smith, A. . . 2218 - See Langford, R. B. and Olsen, K.J. See Ayscough, P. B. and O’Meara, T. 0. See Burke, L. D. and - See Burke, L . D., McCarthy, F. and O’Neill, E. P. See Gray, P. and Onishi, T. See Iwasawa, Y., Soma, M., Onishi, T. and Tamaru, K. Oswal, S. L. See Deshpunde, D. D. and Ott, J. B. See Goates, J. R., Ott, J. B., Reeder, J . and Lamb, J. D. Oversby, J. P. See Ayscough, P. B. and Owen, G. P., Thomas, J. M. and Williams, T. G. and Trimm, D. L. North, A. M. and Soutar, I. . . 1101 Ottenwill, R. H. and Vincent, B. . . 1533 J. 0. . . . 2356 Pacey, P. D. and Purnell, J. H. . . 1462 Paligoric, I. and Gal, I. J. . 1093 Paniccia, F. and Zambonin, P. G. . 2083 Padtt, G. D., Ramsbotharn, J. and Rochester, C. H. . . 17 - See Jackson, P. and Parkes, David A. . . 613, 627, 2103, 2121 Parkinson, C. and Gray, P.. . 1065 - Mukhopadhyay, P. and Gray, P. . 1077 Parsons, B. J. See Jayson, G. G., Parsons, Parsons, G. H. and Rochester, C. H. . 523 Paterson, R. See Dunsrnore, H. S., Jalota, - See Ferguson, H., Gardner, C.R. and - See Gardner, C . R . and Paul, Ram Chand, Singla, Jai Parkash and Penkett, S. A. See Cox, R. A: and . Phillips, David. See Hackett, Peter A . Phillips, G. 0. See Blackburn, R., Cox, - See Hulme, B. E., Land, E. J. and Pichat, P. See Ballivet, D., Barthomeuf, D. Piercy, J. See Jones, M. N. and Pilcher, G. See Pittam, D. A. and Pispisa, B. See Ascoli, F., Branca, M., Pittam, D. A. and Pilcher, G. . 2224 Pittman, J. F. T. see Kandiyoii, R., Pizzini, S. and Morlotti, R. . . 1601 Plunkett, T. J. and Clarke, J. K. A. . 600 Porter, D. A. See Oldershaw, G.A . and - See Oldershaw, G. A., Porter, D. A. Porter, G. See Ledger, M. B. and Powell, A. V. See Wright, E. H. M. and B. J. and Swallow, A. J. S. K. and Gill, Dip Sin@ . . 1894 and D. L. and and Mancini, C. and McLaughlin, E. and and Smith, A.AUTHOR INDEX-VOLUME 6 8 , 1972 13 Price, A. H. See Davies, Mansel, hfaurel, Priest, T. W. . . Pritchard, J. See Alexaider, C. S. and Pryde, J. A. See Grifiths, R., Pryde, J. A. and Righini-Brand, A. Purnell, J. H. See Pacey, P. D. and Quick, L. M. and Whittle, E. Quinn, John A. See Anderson, John i. and Ramsbotham, J. See Parftt, G. D., Ramsbotham, J. and Rochester, C. H. Reed, C. R. V. See Burns, W. G. and Reeder, J. See Goates, 9. R., Ott, J. B., Reeder, J. andhmb, J. D. Reid, G. P. See Roon, Ruby, Reid, G. P.and Tait, K. B. Rendall, H. M. See Lowe, B. M. and Richardson, P. C., Rudham, R., Tullett, A. D. and Wagstaff, K. P. Righini-Brand, A. See GrfltG, R.; Pryde, J. A. and Ritchie, I. M. See Hunt, G. L. and Ro, S. T. See Kestin, J., Ro, S. T. and Wukeham, W. A. Robb, I. D. See Fox, K. K., Robb, I. D. and Smith, R. Robb, J. C. See Barson, C. A., Luxton, A. R. and - See Hay, J. N,, McCabe, J. F. and Roberts, M. W. and Ross, J. R. H. . - S e e Murphy, W. J., Roberts, M. W. and Ross, J. R. H. - See Ross, J. R. H., Roberts, M. W. and Kemball, C. Roberts, N. I(. and Northey, H. L. Robioson, R. A. See Duer, Wuyne C.; Robinson, R. A. and Bates, Roger G. Rochester, C. H. See Parfitt, G. D., Ramsbotham, J. and - See Parsons, G. H. and Rodgers, M. A. J. Ross, J. R. H., Roberts, M.W. and Kemball, C. . . . . - See Murphy, W. J., Roberts, M. W. and - See Roberts, M. W. and Roy, R. N., Vernon, W. and Bothwell, A.L. M. . Rudham, R. See Richards& P. C.; Rudham, R., Tullett, A. D. and Wagstaff, K. P. Rumney, T. G. See Earls, D. W., Jones, J. R. and Ruthven, D. M. and Derrah, R. I. . - and Loughlin, K. F. . . . . - See Derrah, R. I., Loughlin, K. F. and Saito, Shuji. See Hirabaymhi, Kazuo, Saito, Shuji and Yasumori, Iwao Saleh, J. M. - See Dadiza, Y.* M. a$d Sayer, C. F. See Davis, K. M. C. and Scheinin, N. See Silber, H. B., Scheinin, N,, Atkinson, G. and Grescek, J. J. Schoff, C. I(. See MacCallum, J. R. and Schwarzmann, E. See Ashcrof, S. J. and Semmens, M. J. See Gregory, J. and P. and ' PAOE 66 1 878 2203 221 1528 1278 914 2047 2332 696 1520 Sheridan, Margaret H.See Bruce, Linda A. Sherwood, J. N. See Burns, G. and - See Chadwick, A. V. and - See Hood, G. M., Lockhart, N. C. and - See McGhie, A. R. and Shimoji, M. See Yamamoto, M., Nukamura, Y. and Siddiqi, I. W. See Jagtiani, I. A., Siddiqi, I. W. and Tyrrell, H. J. V. Sierra, Jose, Teixido M., Emilio and Wyatt, P. A. H. . . . Silber, H. B., Scheinin, N., Atkinson, G. and Grecsek, J. J. . . Sime, S. J. See Johnson, K. E., Sime, S. J. and Dudley, J, Singla, Jai Parkash. See Paul, Ram Chand, Singla, Jai Parkash and Gill Dip Singh. Skerrett, R. J. . . skinner, H. A. see Connor, J.'A., sicinner; H. A. and Virmani, Y. - See Tipping, E. W. and Smedley, S. I. See Cleaver, B., Smedley, S. I. and Spencer, P. N. Smiles, D. E. See Elrick, D. E., Smiles, D.E. and Wooding, R. A. Smith, A. See Oldershaw, G. A., Porter, D. A. and Smith, D. See Caldin, E. F., Crooks, J. E., O'Donnell, D., Smith, D. and Toner, S. Smith, R. See Fox, K. K., Robb, I. D. and Sobczyk, L. and Danel, J. . . Soma, M. See Iwusawa, Y., Soma, M., Onishi, T. and Tamaru, K. Soutar, I. See North, A . M. and Spencer, P. N. See Cleaver, B., Smedley, S. I. and Srivastava, R. D. See Farber, Milton, Srivastava, R. D. and Uy, 0. M. Starodubceva, R. V. See Kiselev, A . V., Lygin, V. I. and Staveley, L. A. K. See Cook, R. O., Davies, A. and Steckel, F. See Kudish, A. I., WOK, D. and Stiddard, M. H. B. See Knapp, A. G. and Stone, F. S. See Egerton, T. A., Hagan, A., Stone, F. S. and Vickerman, J. C. Stone, T. J. See Greatorex, D., Hill, R. J., Kemp, T. J.and Stonehill, H. I. See Clark, K . P. and Subhani, M. S. See Buxton, G. V. and Sutton, H. C. and Domes, M. T. . Swallow, A. J. See Jayson, G. G., Parsons, Symons, M. C. R. . Szilagyi, I. See Berces, T., Marta, F. and Tait, K. B. See Foon, Ruby, and - See Foon, Ruby, Reid, G. P. and Takaishi,T. . . . . -andMohri,M. . . . . Tamaru, I(. See Iwasawu, Y., Soma, M., - See Naito, S., Ichikawa, M. and Tedder, J. M., Walton, J. C. and Winton, and B. J. and Onishi, T. and K.D.R. . . . 160, PAGE 290 1200 1328 1544 1498 216 80 1 1921 186614 AUTHOR INDEX-VOLUME 68, 1972 PADE Teixido, M., Emilio. See Sierra, Jose, Teixido, M., Emilio and Wyatt, P. A. H . Telford, N. R. See Hayhurst, A. N. and Tench, A. J. . . . 197, 1181 - and Giles, J. . . . . . 193 -- Lawson, T. and Kibblewhite, J. F.J. . 11 69 Teng, L. and Jones, W. E. . 1267 Tennakoon, D. T. B. See WiZliams,'J. O., Tennakoon, D. T. B., Thornas, J. M. and Jacobs, P. W. M . Teply, J. See Lukac, S., Teply, J. and Vacek, K. Tham, M. K. and Gubbins, K. E. . . 1339 Thirsk, H. R. See Dobson, J. V., Dagless, M. N. and -- See Lown, D. A. and Thomas, J. H. See Bedford, G. and Thomas, J . M. See Owen, G. P., Thomas, J. M. and Williams, J. 0. -- See Williams, J. O., Tennakoon, D. T. B., Thomas, J. M. and Jacobs, P. W . M. Tiddy, G. J. T. . 369, 608, 653, 670 Tilsley, G. M. See Cadman, P., Tilsley, G. M. and Trotman-Dickenson, A. F. Tipping, E. W. and Skinner, H. A. . 1764 Toner, S. See Caldin, E. F., Crooks, J. E., O'Donnell, D., Smith, D. and Trasatti, S. . 229 Trimm, D. L. Sek Culiis, C: F., Neveli, T. G.and Triolo, R. See Gofredi, M. and Trotman-Dickenson A. F. See Cadman, P., Tilsley, G. M. and - See Cadman, P., White, A. J. and Tsaoussoglou, P. E. See Buckle, E. R. and Tullett, A. D. See Richardson, P. C., Rudham, R., Tullett, A. D. and Wagstaf, K. P. Tyler, B. J. See Graham, R. F. and Tyrrell, H. J. V. See Jagtiani, I. A., Siddiqi, I. W . and Uy, 0. M. See Farber, Milton, Strivastava, R. D. and Vacek, K. SeeLukac, S., Teply, J. and Vander Donckt, E. and Leitaer, D.. . 112 Van Der Linde, H. J. See Ackerman, L. G. J., Basson, R. A. and Van Huis, C. See Nauta, H. and Vernon, W. See Roy, R. N., Vernon, W. and Bothwell, A. L. M. Vickerman, J. C. See Egerton, T. A., Hagan, A., Stone, F. S. and Vinall, I. C. See Frey, H. M., Hopkins, R. G.and Vincent, B. See Ottewill, R. H. and Vincent, Vera M. See Hurditch, R . J,, Vincent, Vera M. and Wright, J. D. Virmani, Y. See Connor, J. A., Skinner, H. A. and 'C70ice, P. J. See Feakins, D. and Wagstaff, K. P. See Richardson, P. C., Rudham, R., Tullett, A. D. and Wakeham, W. A. See Kestin, J., Ro, S. T. and Walker, R. W. See Baldwin, R . R., Fuller, A. R., Longthorn, D. and Walton, J. C. . -See Tedder, J. M., Walton, J. C. anh Winton, K. D. R. Wan, J. K. S. See Wong, S. K., Fabes, L., Green, W. J. and Watkins, A. R. , . . . Wells, C. F. . . - and Whatley, D. : Whan, D. A. See Dowie, R: A., 'Whan', D. A. and Kemball, C. - See Howe, R. F., Davidson, D. E. and Whatley, D. See Wells, C. F, and White, A. J. See Cadman, P., White, A. J. and Trotman-Diekenson, A. F.Whittle, E. See Bassett, J. E. and - See Chamberlain, G. A. and - See Ferguson, K. C. and - See Quick, L. M. and Whytock, D. A., Clarke, J. D. and Gray, P. Williams, J: O., 'Tennakoon; D. T. B.; Thomas, J. M. and Jacobs, P. W. M. - See Owen, G. P., Thomas, J. M. and Willis, E. See Melville, J. B. and Winkler, C. A. See Jacob, A. and Winton, K. D. R. See Tedder, J. M., Walton, J. C. and Wisniewski, R. See Baranowski, B., Lewis, F. A., Majchrzak, S. and Wolf, D. See Kudish. A. I.. Wolf. D. and ~ - I Steckel, F. Wong, S. K., Fabes, L., Green, W. J. and Wan, J. K. S. . . . Wooding,-R. A. See Elrick, D. E., Smiles, Woodward, T. W. See Ahmad, M., Woolf, L. A. See McCool, M. A. and - See McCool, M. A., Collings, A. F. and Wright, A. N. See Kunz, C. 0. and Wright, E. H.M. and Powell, A. V. . Wright, J. D. See Hurditch, R. J., Vincent, Vera, M. and Wyatt, P. A. H. See Hussain, S. K. and - See Ireland, J. F. and - See Kureishi, A. W. and - See Sierra, Jose, Teixido M. Emilio Yamamoto, M., Nakamura, Y. and Yasumori, Iwao See Hiiabayashi, Kazuo; Yates, Jr., J. T. See King, D. A., Madey, Zamboni, R. See Indelli, A, and Zambonin, P. G. See Paniecia, F. and Zetter, M. S. See Hague, D. N., Martin, D. E. and Wooding, R. A. Huyton, D. W. and and Shimoji, M. . Saito, Shuji and T. E. and S. R. and PAGE 1559 28 993 434 689 1987 221 1 1908 135INDEX TO REVIEWS-VOLUME 68, 1972 Acidity Functions. C. H. Rochester . . . . . . . Actions Chimiques et Biologiques des Radiations. Vol. 14.' Edited by M. Haissinsky . Adsorption of Organic Compounds on Electrodes.B. B. Damaskin, 0. A. Petrii and V. V. Batrakov . . . Advances in Chemical Physics. Voi. 17. ' Ediied b;I. Prigogink and's. A.'Rice . . Advances in Chemical Physics. Vol. 18. Edited by I. Prigogine and S. A. Rice . Advances in Physical Organic Chemistry. Vol. 8. Edited by V. Gold . . . . Applied Spectroscopy Reviews. Vol. 3. Edited by E. G. Brame Aquatic Chemistry. An Introduction Emphasizing Chemical Equilibria in Nathal Waters: Werner Stumm and James J. Morgan . . . Atomic Collision Phenomena in Solids. Edited by D. W. Paimer, M. W. Thompson an; P. D. Townsend . . . . . . . . . . . Band Theory of Metals. The Elements. S. L.. Altman . . . . . . . Classical Dynamics of Particles and Systems (2nd ed.). J. B. Marion Compilation of Reported 19F NMR Chemical Shifts. C. H. Dungan and J. R. Van Wazer .* Creation and Detection of the Excited State. Vol. 1. Edited by Angel0 A. Lamola . . Crystals and X-rays 13. H. S. Lipson . . . . . . . Diffusion Processes. Vol. 1 and 2: Proceedings or The'Thomas Graham Memorial Sym- posium. Edited by J. N. Sherwood, A. V. Chadwick, W. M. Muir and F. L. Swinton . Electrochemical Reactions in Non-aqueous Systems. . Chiles K: Mark and Karen K. Barnes Elektronenmikroskopische Methodik. G. Schimmel . . . . Formulations of Classical and Quantum Dynamical Theory. G. Rosen . . . : Handbook of Atomic Elements. R. A. Williams Handbook of Metal Ligand Heats and Related Thermdynamic Quantiiies. Jakes'J. Chkten: Introduction to the Kinetics of Chemical Chain Reactions. . F. G: R. Grirnbiett Introduction to Polymer Science (No. 9. The Wykeham Science Series). L. R. G. Treloa; Introductory Group Theory for Chemists. G. Davidson. . . . . . . Ion-Molecule Reactions. E. W. McDaniel, V. Cermaj, A. Dalgarno, E. E. Ferguson and L. Friedman . . . * . Magnetism and Metallurgy. Vol. i and 2. Edited by A.'Berkowitz and E. Kneiler . . Mathematical Methods for Physicists. G. Arfken . . . Modem Aspects of Electrochemistry, No. 6. Edited by J.'O'M.'Boc&is and B. E. Conway . Modern Methods of Surface Analysis. (Proceedings of the Symposium on Modern Methods of Surface Analysis, Bell Telephone Labs. Murray Hill, N.J., U.S.A.). Edited by P. R. Mark and J. D. Levine . , . Molecular Interactions and Electronic Spectra. Noboru Metaga and Tanekazu Kubota . Physical Acoustics. Vol. VII. Edited by W. P. Mason and R. N. Thurston . . . Polyelectrolytes. Fumio Oosawa . . . Preparation and Properties of Solid State Materials'. Vk. 1. Aspects of Crystal Giowth. Principles of Activation Analysis. P. Kruger . Proceedings of the 3rd Symposium on Coordination Chemishy. Debrecan, Hungary, 1970 1 Progress in Polymer Science. Vol. 2. Edited by A. D. Jenkins . . . . . Quantum Mechanics. Vol. 1. Fundamentals. K. Gottfried. . . . . Reverse Osmosis. S. Sourirajan . . . . . . . . * . . Selection of Oxidants in Synthesis. L. J. Chinn . . Semiconductors and Semimetals. Vol. 5 . Infrared Detectors. Edited by 'R. K: WiH&dson and A. C.Beer . . . . . . . . . Spin Temperature and Nuclear Magnetic Resonance in Solids. M. Goldman : . . Techniques of Chemistry. Vol. 1. Edited by A. Weissberger and B. W. Rossiter . . Techniques of Chemistry. Vol. 2. Edited by A. Weissberger and B. W. Rossiter . The Chemistry of the Cyano Group. Edited by Zvi Rappoport The Chemistry of the Nitro and Nitroso Groups. Part 2. Edited by HenriFeue; : The Photographic Image. Formation and Structure. Edited by S. Kikuchi The Royal Institution Library of Science, Physical Sciences. Vol. 1-10. Edited by SiE William Lawrence Bragg and Prof. George Porter . . sen and Reed M. Izatt . . . . . Edited by R. A. Lefever . . . . . . . The Chemistry of Inorganic Ring Systems. I. Haiduc . . . . . 15 PAGE 182 381 1152 185 189 191 186 798 3 84 187 191 190 799 382 1151 797 384 183 186 185 381 185 991 188 186 182 1151 798 192 381 201 3 1376 991 382 576 188 187 798 192 184 799 1375 797 576 576 189 38316 INDEX TO REVIEWS-VOLUME 6 8 , 1972 PAQE Thennochemistry of Transition Metal Complexes. S. J. Ashcroft and C. T. Mortimer . . 182 Three Approaches to Electron Correlation in Atoms. Oktay Sinanoglu and Keith A. Brueckner 183 Topics in Organic Mass Spectrometry. Edited by A. L. Burlingame. (Vol. 8 of Advances in Analytical Chemistry and Instrumentation. Edited by C. N. Reilley and F. W. UnimolecuIar Reactions. P. i.. Robinson' and K. A.'Holb;ook : . 2013 McLaffert y) . 190 Valence Theory. J. N. Murrell, S. F. A. Kettle and J. M. Tedder . . 190

ISSN:0300-9599    

DOI:10.1039/F197268BA001

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Adhesion at the Alkane/Water and Ester/Water Interfaces BY R. AVEYARD, B. J. BRISCOE," AND J. CHAPMAN Chemistry Department, The University, Hull. Received 26th July, 197 1 The surface tension of 1-decanol and of several methyl esters of n-alkanoic acids, and the inter- facial tension of these materials against water, have been determined in the temperature range 20-32°C. The interfacial tensions of n-nonane and n-undecane against water in the range 20-30°C are also reported. A simple model is proposed to account for the variation with chain length of the thermo- dynamic parameters of adhesion of the liquids with water. The results are consistent with n-alkanes assuming an orientation parallel to the water surface and the esters being essentially normal to the surface. A previous study of the adhesion between liquid n-alkanes and water has been extended to include two further alkanes (nonane and undecane), a series of methyl esters of n-alkanoic acids, and n-decanol.Measurements of the surface tension of the esters and decanol, and of the interfacial tensions of all of the organic liquids against water, have been made in the temperature range 20-32°C so that the temperature variation of the work of adhesion could be determined. Previous suggestions concerning the orientations of n-alkanes at alkanelwater interfaces differ. Two extreme possibilities are that the alkanes are extended and normal to the surface, or that only configurations parallel to the surface exist. In this paper we consider these two extremes, neither of which is likely to be entirely true, to find which is more nearly the case.A crude model is used, which is formulated in terms of the surface and interfacial tensions of the various components. The results obtained for the esters are particularly useful since for these systems it can be supposed that the molecules will not assume an orientation parallel to the surface. EXPERIMENTAL Surface and interfacial tensions were determined using the drop-volume technique as described e1sewhere.l All results were reproducible to at least rt0.2 %. The water thermo- stat had a temperature constancy to better than L-0.02". All necessary densities were deter- mined to within k2x g using a 25 cm3 density bottle. In the calculation of surface tension, allowance was made for the displacement of air by the drop of liquid.The water was twice distilled from a Pyrex apparatus and had a surface tension of 72.50 f 0.10 mN rn-' at 20°C. The alkanes were obtained from Newtsn-Maine (>99 %). Further purification was effected by percolation through alumina prior to use. All samples had a final purity, as estimated by g.l.c., of >99.5 %. The decanol was apuriss. grade Koch-Light sample of purity 98.9 %. The methyl esters of the C6, C8, C12 and C14 acids were Flukapuriss. grade with g.1.c. purities of, respectively 99.6,99.8,99.4, and 100 %. The methyl ester of decanoic acid was a Fluka purum grade of 98.5 % purity. Decanol and the esters were passed through alumina twice before use. For the alkanes and the esters, pre-equilibration with water did not significantly affect the value of the interfacial tensions and the values reported relate to systems which have not been mutually saturated.Decanol and water were, however, allowed to reach saturation equilibrium before the determination of the interfacial tensions were made. * present address : Cavendish Laboratory, The University, Cambridge. 10R . AVEYARD, B . J . BRISCOE A N D J . CHAPMAN 11 RESULTS AND DISCUSSION The surface tensions of the alkanes were not determined; values reported in ref. (2) have been used. The interfacial tensions yap for the alkane + water systems are given in table 1. Values for hexane/water and tetradecanelwater, taken from ref. (I), are included for convenience. Values for the octane/water and dodecane/watar systems agree well with those reported in ref.(1). TABLE 1 .-INTERFACIAL TENSIONS (mN m-') OF ll-ALKANES AGAINST WATER 6 7 8 *,4\N 20.0 (50.80) (51.23) 51.69(51.68) 25.0 (50.41) - 51.2q51.22) 27.5 (50.16) - - 30.0 (49.92) - 50.83(50.78) 32.5 (49.74) - - 35.0 (49.49) - 50.37(50.31) - - 22.5 - 9 10 11 12 14 51.96 52.26(52.30) 52.51 52.86(52.78) (53.32) 51.76 - 52.39 52.74 - 51.54 - 52.1 1 52.50(52.46) (52.92) 51.30 - 51.92 52.28(52.21) (52.69) 51.11 - 51.72 52.07(51.99) (52.46) - - 51.82(51.74) (52.27) - - - 51.61(51.50) (52.04) Values given in brackets are from ref. (1). The surface tensions of the esters and of decanol are given in table 2. The values for methyl dodecanoate are interpolated from the results of V0ge1.~ The present values for methyl hexanoate, and methyl decanoate are in good agreement with those of Vogel.Our values for methyl octanoate are, however, approximately 0.6 mN m-l lower than his at a given temperature. The general trend of the resuIts suggests that Vogel's value is in error. TABLE 2.-sURFACE TENSION (mN m-') OF ESTERS AND DECANOL TIT hexanoate octanoate decanoate dodecanoate tetradecanoate 1 -decanol Me Me Me Me Me 20.0 26.3 7( 26.36) 27.3 q27.99) 28.47(28.46) (29.58) 30.16 28.60 22.0 26.16 27.19 I (29.40) 29.97 - 24.0 25.94 26.97 28.09 (29.23) 29.75 28.24 26.0 25.75 26.75 I (29.05) 29.56 - 28.0 25.56 26.57 27.69 (28.87) 29.38 27.91 30.0 25.35 26.40 - (28.69) 29.22 - 32.0 25.14 26.21 27.32 (28.51) 29.06 27.59 Values in brackets estimated from values of VogeL3 The interfacial tensions of the esters and of decanol against water are given in table 3. Both the surface and interfacial tensions (yp and yap respectively) are, within the reproducibility of the results, linear functions of the temperature in the range studied.Constants of the equation where Tis the temperature in K, have been obtained using the method of least squares, and are listed in table 4. The constants A and -B are respectively the enthalpy and entropy of extension of the surface or interface. We are concerned with the work of adhesion WF of the organic liquid (fl phase) and water (a phase), where = A+BT, (1) wy = y=+yB-y@. (2)12 ADHESION AT WATER SURFACES In the present work, y" and yp refer to the pure rather than the mutually-saturated m.aterials. As mentioned earlier, yap is not significantly dependent upon whether or not the phases have attained saturation equilibrium.The mutual solubilities of water and the organic liquids are likely to be very small for all the systems studied. TABLE 3.-INTERFACIAL TENSIONS (mN m-') OF ESTERS AND DECANOL AGAINST WATER TIT Me hexanoate Me octanoate Me decanoate Me dodecanoate Me tetradecanoate 1-decanol 20.0 17.83 20.62 22.53 23.92 25.14 8.97 - 23.98 25.22 - 22.0 17.87 20.66 24.0 17.90 20.70 22.66 24.05 25.30 9.25 26.0 17.94 20.73 - 24.12 25.39 - 28.0 17.99 20.79 22.80 24.19 25.45 9.50 30.0 18.01 20.83 - 24.25 25.52 - 32.0 18.04 20.88 22.90 24.32 25.58 9.78 Enthalpies Ah:! and entropies AsIp corresponding to Wip can be obtained from the temperature variation of Wip in the usual way. All the terms refer to the isothermal separation of unit area of the liquid/liquid interface to yield unit areas of the two TABLE 4.-cONSTANTS OF EQN (1) FOR ESTERS AND DECANOL surface tension interfacial tension A - B A B Me hexanoate 56.15 0.101 61 12.60 0.017 86 Me octanoate 56.1 1 0.098 04 14.28 0.021 61 Me decanoate 56.69 0.096 25 13.37 0.031 25 Me dodecanoate* 55.71 0.089 13 14.08 0.033 57 Me t etradecan oat e 57.20 0.092 32 14.31 0.036 96 1 -decanol 53.21 0.084 00 - 10.67 0.067 00 * For the surface tension, the values of Vogel between 13 and 87°C have been used.liquid surfaces. temperature, it can be expressed as Since, for all the systems studied, Wip is a linear function of the w.,s = C-DT. (3) The constants C and D are, respectively, Ahlp and A,$. Values of C and D for the temperature range 20-32°C are given in table 5.In the calculation of W f , values for the surface tension of water were taken from ref. (4). TABLE s.-cONSTANTS OF EQN (3) FOR w:p IN THE RANGE 20-32°C" chain lengtht 6 8 9 10 11 12 14 alkanelwater C D esterlwater C D 91.21 0.1735 162.53 0.2771 90.86 0.1644 160.60 0.2766 91.59 0.1636 - - -- - 162.02 0.2843 92.46 0.1624 - - 92.43 0.1611 160.40 0.2796 92.84 0.1598 162.03 0.2875 Wzp is given in mJ m-2 ; -f for the esters, the chain length refers to the parent acid ; * for the decanoliwater interface, C = 182.59 and D = 0.3078.R . AVEYARD, B . J . BRISCOE A N D J . CHAPMAN 13 For the orientation of the organic molecules at the liquid/liquid interface, it appears likely that singly adsorbed n-alkane molecules on a water surface will assume mainly those configurations which are parallel to the s~rface.~ However, when the alkanes are present at the interface in a condensed state, as at the liquid/liquid inter- face, it may be more favourable for the chains to align at a steep angle to the interface.On energy considerations, Ohki and Fukuda have concluded that a vertical orienta- tion, where the chains can interact mutually along their long axes, is more probable than an orientation parallel to the interface. On the other hand, Pomerantz, Clinton and Zisman believe that the alkanes (at least up to octane) are probably parallel to the alkane/water interface, and state that there is much recent evidence to support this. For long chain polar compounds it is likely that a near vertical orientation will be assumed in order to pack as many polar groups as possible into contact with water at the interface.We consider first the work of adhesion in alkanefwater systems. We suppose the alkanes are parallel to the alkane/water interface and that the alkane side of the interface can be considered as a mixture of -CH2- and -CH3 groups. If the work of adhesion Wl of the methyl group differs from that, W2, of the methylene group, Wip should vary with chain length in a way to be determined. The mol fractions, x1 and x2, of -CH, and -CH2- groups in an alkane con- taining N carbon atoms, are given by If the molar surface areas of -CH3 and -CH2- groups are, respectively, c1 and a2, and if the surface is monomolecular with respect to alkane, then Q, the area of the interface is = 2 / N ; ~2 = 1-2fN.(4) 0 = iz9,0,+n~a2, ( 5 ) where ni and 12; are the numbers of moles of -CH, and -CH2- groups respectively in the surface. Thus, WIp can be expressed as Combination of (4), (5) and (6) yields w5y2(r, - 1) 3. N ) = N w. + 2(r, w, - W,), (7) where r1 is defined as a,/o,. parallel to the interface, we obtain Using similar arguments for the esters, assumed to be W;yN + r l f r3 - 2) = N W, + (7.1 W1+ r3 w3 - 2W,), (8) where W3 is the work of adhesion of the -COOMe group with water and r3 = a,/a2 ; o3 is the molar surface area of the -COOMe group. N in (8) is the number of carbon atoms in the parent acid. Implicit in the above arguments is the assumption that the orientation of the chains is the same at the liquid/vapour interface and the liquid/liquid interface.This may not be true but it seems reasonable to suppose that the major contribution to WT will arise from removing the liquid/liquid interactions, and that the reorient- ation of the chains, if it occurs, gives rise to a relatively small change in free energy. However, the validity of the assumption can only be judged by the measure of agree- ment obtained between values of Wl and W2 (and the corresponding enthalpy changes) calculated using the theory, and values obtained independently. A plot of the left-hand side of (7) or (8) against N should be linear, with the slope of the line equal to W,, if the assumed model approximates to the real situation.14 ADHESION AT WATER SURFACES Further, for the alkanes, Wl can be calculated from the intercept of the plot according to (7).Values of Wl and W2 can then be compared with independently-obtained values. Adam and Elliott,8 from a study of contact angles of water on various solid hydrocarbons, concluded that the work of adhesion of the methylene group is about 54mJm-2 and of the methyl group about 40rnJnr2. These values were obtained using the mean of the advancing and receding contact angles and are not therefore precise, but they may be used as a guide. Plots according to (7) and (8) for the alkanes and esters respectively are shown in fig. 1. Values of crl and c2 have been taken as 0.1 1 and 0.05 nm2 respectively,' and cr3 has been assigned a value of 0.20nm2. Both plots are apparently linear. The value of W, obtained from the plot for the alkanes is 52 mJ m-2 in good agreement with the findings of Adam and Elliott.From the same plot W1 is found to be 30 mJ ~tl.-~, which is lower than the value of 40 mJ m-2 found from contact angles. Some of this difference may arise from different values of c1 at the liquid and at the solid surfaces. 1 1 N E m i 6 8 I0 12 14 N FIG. 1.-Plots according to eqn (7) and (8) for 20°C. 0 and left-hand ordinate refer to alkanes and and right-hand ordinate to esters. Aveyard and Haydon reported that the plot of WiB against 1/N for the alkanes is linear. Inspection of (7) shows that strictly, according to the model proposed, the plot should only be linear if r1 = 1. However, in the range N = 6-14 such a plot is almost linear for rl = 2.2 (i.e., for ol = 0.1 1 and cr2 = 0.05 nm2 molecule-'). Assuming linearity, the value obtained for W2 (at 1/N = 0) is very close to the value found using (7).The value obtained for Wl, however, is about 30 % lower than that obtained using (7). W, calculated from a plot according to (8) for the esters is 73 mJ m-2, which is unacceptably high. Different choice of magnitudes for rl and r3 does not significantly alter this value. It is thus reassuring that the use of the model for " flat " molecules on a system where the molecules are likely to be essentially vertical, does not fortuit- ously lead to a reasonable value for W2. Ahip for the alkane + water systems (table 5) is almost independent of N indicatingR . AVEYARD, B . J . BRISCOE A N D J . CHAPMAN 15 that, if the alkane molecules are parallel to the surface, the heats of adhesion of the -CH3 and -CH2- groups are of similar magnitude.The enthalpies of adhesion of surfaces comprised of solely methyl or methylene groups can be obtained using an equation analogous to (7), for Ah?. Values so obtained are for the methyl group 88 mJ m-2 and for the methylene group 95 mJ m-2. Confirmation of the reason ableness of these quantities has been sought by comparison with standard equilibrium heats of desorption (-d,H*) of n-alkanes from the surface of water. Although the process of parting a liquid/liquid interface is not equivalent to desorption from a dilute alkane monolayer, the magnitude of the enthalpy changes for the two processes should be similar, if the orientation of the alkanes at the interface is the same in both cases.From the unit area values already quoted, the molar enthalpy of adhesion of the -CH3 group is 5.8 kJ mol-I and of the -CH2- is 2.7 kJ mol-I. The values of AHiB for the whole molecules are plotted in fig. 2 together with -A,H* obtained by Jones and Ottewill at 7.5OC.’ The two sets of data are in good agreement. For example, for octane, AHiB = 28.8 kJ mo1-I and -A,H* = 30.6 kJ mol-I. In addition, the -CH2- increment in -AaHe is about 2.8 kJ mol-l, close to the cor- responding increment of 2.7 kJ mol-l in AHiP. In view of this agreement, and since the adsorbed alkanes are likely to be flat on the water surface, the adhesion data are consistent with a parallel orientation at the alkanelwater interface. 6 a 10 12 14 N at 7.5% and 0 to AHZP in the temperature range 20-32°C.FIG. 2.-Enthalpies of desorption and adhesion for systems of alkanefwater. refers to -Aa,Hf3 Ah? for the esters is also independent of chain length (see table 5 ) and is 161.64 1 mJ m-2. The high value is indicative of the interaction of the -COOMe group with water. The constancy of Ahip indicates that the molecules are not parallel to the surface. A flat orientation would give rise to a marked chain length dependence of Ahib since the enthalpies of adhesion of the -CH3 and -CH2-- groups are much smaller than that for the -COOMe group. AhIP for n-decanol is 182.6 mJ m-2. If the surface area of decanol is supposed to be the same as the close-packed area in insoluble films of long-chain n-alkanols on water, i.e., 0.216 nrn2,l0 the molar enthalpy of adhesion is 23.8 kJ mol-’.As might be expected the value is not very different from the heat of desorption (-31 kJ mol-I) of alkanols from the alkanelwater interface to dilute solution in the alkane.16 ADHESION AT WATER SURFACES To summarize, using a simple theory, values of the work and enthalpies of ad- hesion are consistent with a predominantly flat orientation of alkanes, and an essen- tially vertical orientation of long-chain methyl esters, at the interface between these liquids and water. Probably the main objection to the proposed theory is the neglect of the effects which could arise as a result of the organic molecules assuming different configurations on parting the liquid/liquid interface. B. J. B. and J. C. are indebted to the Science Research Council for provision of maintenance grants. R.Aveyard and D. A. Haydon, Trans. Faraday SOC., 1965, 61, 2255. Selected Properties of Hydrocarbons and related compounds, API Project 44 (Thermodynamics Research Center, Texas, 1966). A. I. Vogel, J. Chem. SOC., 1948, 638. Int. Crit. Tables (McGraw-Hill, 1928), 4, 447. J. H. de Boer, The DynamicaZ Character of Adsorption (Oxford, 1953), p. 151. S. Ohki and N. Fukuda, J. CoZZoid Interface Sci., 1968, 27, 208. ' P. Pomerantz, W. C. Clinton and W. A. Zisman, J . CoZZoid Interface Sci., 1967, 24, 16. * N. K. Adam and G. E. P. Elliott, J. Chem. SOC., 1962,2206. D. C. Jones and R. H. Ottewill, J. Chem. Soc., 1955,4076. lo N. K. Adam, The Physics and Chenzistry of Srtrfaces (Oxford, 1938), p. 50. Adhesion at the Alkane/Water and Ester/Water Interfaces BY R.AVEYARD, B. J. BRISCOE," AND J. CHAPMAN Chemistry Department, The University, Hull. Received 26th July, 197 1 The surface tension of 1-decanol and of several methyl esters of n-alkanoic acids, and the inter- facial tension of these materials against water, have been determined in the temperature range 20-32°C. The interfacial tensions of n-nonane and n-undecane against water in the range 20-30°C are also reported. A simple model is proposed to account for the variation with chain length of the thermo- dynamic parameters of adhesion of the liquids with water. The results are consistent with n-alkanes assuming an orientation parallel to the water surface and the esters being essentially normal to the surface.A previous study of the adhesion between liquid n-alkanes and water has been extended to include two further alkanes (nonane and undecane), a series of methyl esters of n-alkanoic acids, and n-decanol. Measurements of the surface tension of the esters and decanol, and of the interfacial tensions of all of the organic liquids against water, have been made in the temperature range 20-32°C so that the temperature variation of the work of adhesion could be determined. Previous suggestions concerning the orientations of n-alkanes at alkanelwater interfaces differ. Two extreme possibilities are that the alkanes are extended and normal to the surface, or that only configurations parallel to the surface exist. In this paper we consider these two extremes, neither of which is likely to be entirely true, to find which is more nearly the case.A crude model is used, which is formulated in terms of the surface and interfacial tensions of the various components. The results obtained for the esters are particularly useful since for these systems it can be supposed that the molecules will not assume an orientation parallel to the surface. EXPERIMENTAL Surface and interfacial tensions were determined using the drop-volume technique as described e1sewhere.l All results were reproducible to at least rt0.2 %. The water thermo- stat had a temperature constancy to better than L-0.02". All necessary densities were deter- mined to within k2x g using a 25 cm3 density bottle. In the calculation of surface tension, allowance was made for the displacement of air by the drop of liquid.The water was twice distilled from a Pyrex apparatus and had a surface tension of 72.50 f 0.10 mN rn-' at 20°C. The alkanes were obtained from Newtsn-Maine (>99 %). Further purification was effected by percolation through alumina prior to use. All samples had a final purity, as estimated by g.l.c., of >99.5 %. The decanol was apuriss. grade Koch-Light sample of purity 98.9 %. The methyl esters of the C6, C8, C12 and C14 acids were Flukapuriss. grade with g.1.c. purities of, respectively 99.6,99.8,99.4, and 100 %. The methyl ester of decanoic acid was a Fluka purum grade of 98.5 % purity. Decanol and the esters were passed through alumina twice before use. For the alkanes and the esters, pre-equilibration with water did not significantly affect the value of the interfacial tensions and the values reported relate to systems which have not been mutually saturated.Decanol and water were, however, allowed to reach saturation equilibrium before the determination of the interfacial tensions were made. * present address : Cavendish Laboratory, The University, Cambridge. 10R . AVEYARD, B . J . BRISCOE A N D J . CHAPMAN 11 RESULTS AND DISCUSSION The surface tensions of the alkanes were not determined; values reported in ref. (2) have been used. The interfacial tensions yap for the alkane + water systems are given in table 1. Values for hexane/water and tetradecanelwater, taken from ref. (I), are included for convenience. Values for the octane/water and dodecane/watar systems agree well with those reported in ref.(1). TABLE 1 .-INTERFACIAL TENSIONS (mN m-') OF ll-ALKANES AGAINST WATER 6 7 8 *,4\N 20.0 (50.80) (51.23) 51.69(51.68) 25.0 (50.41) - 51.2q51.22) 27.5 (50.16) - - 30.0 (49.92) - 50.83(50.78) 32.5 (49.74) - - 35.0 (49.49) - 50.37(50.31) - - 22.5 - 9 10 11 12 14 51.96 52.26(52.30) 52.51 52.86(52.78) (53.32) 51.76 - 52.39 52.74 - 51.54 - 52.1 1 52.50(52.46) (52.92) 51.30 - 51.92 52.28(52.21) (52.69) 51.11 - 51.72 52.07(51.99) (52.46) - - 51.82(51.74) (52.27) - - - 51.61(51.50) (52.04) Values given in brackets are from ref. (1). The surface tensions of the esters and of decanol are given in table 2. The values for methyl dodecanoate are interpolated from the results of V0ge1.~ The present values for methyl hexanoate, and methyl decanoate are in good agreement with those of Vogel.Our values for methyl octanoate are, however, approximately 0.6 mN m-l lower than his at a given temperature. The general trend of the resuIts suggests that Vogel's value is in error. TABLE 2.-sURFACE TENSION (mN m-') OF ESTERS AND DECANOL TIT hexanoate octanoate decanoate dodecanoate tetradecanoate 1 -decanol Me Me Me Me Me 20.0 26.3 7( 26.36) 27.3 q27.99) 28.47(28.46) (29.58) 30.16 28.60 22.0 26.16 27.19 I (29.40) 29.97 - 24.0 25.94 26.97 28.09 (29.23) 29.75 28.24 26.0 25.75 26.75 I (29.05) 29.56 - 28.0 25.56 26.57 27.69 (28.87) 29.38 27.91 30.0 25.35 26.40 - (28.69) 29.22 - 32.0 25.14 26.21 27.32 (28.51) 29.06 27.59 Values in brackets estimated from values of VogeL3 The interfacial tensions of the esters and of decanol against water are given in table 3.Both the surface and interfacial tensions (yp and yap respectively) are, within the reproducibility of the results, linear functions of the temperature in the range studied. Constants of the equation where Tis the temperature in K, have been obtained using the method of least squares, and are listed in table 4. The constants A and -B are respectively the enthalpy and entropy of extension of the surface or interface. We are concerned with the work of adhesion WF of the organic liquid (fl phase) and water (a phase), where = A+BT, (1) wy = y=+yB-y@. (2)12 ADHESION AT WATER SURFACES In the present work, y" and yp refer to the pure rather than the mutually-saturated m.aterials. As mentioned earlier, yap is not significantly dependent upon whether or not the phases have attained saturation equilibrium.The mutual solubilities of water and the organic liquids are likely to be very small for all the systems studied. TABLE 3.-INTERFACIAL TENSIONS (mN m-') OF ESTERS AND DECANOL AGAINST WATER TIT Me hexanoate Me octanoate Me decanoate Me dodecanoate Me tetradecanoate 1-decanol 20.0 17.83 20.62 22.53 23.92 25.14 8.97 - 23.98 25.22 - 22.0 17.87 20.66 24.0 17.90 20.70 22.66 24.05 25.30 9.25 26.0 17.94 20.73 - 24.12 25.39 - 28.0 17.99 20.79 22.80 24.19 25.45 9.50 30.0 18.01 20.83 - 24.25 25.52 - 32.0 18.04 20.88 22.90 24.32 25.58 9.78 Enthalpies Ah:! and entropies AsIp corresponding to Wip can be obtained from the temperature variation of Wip in the usual way. All the terms refer to the isothermal separation of unit area of the liquid/liquid interface to yield unit areas of the two TABLE 4.-cONSTANTS OF EQN (1) FOR ESTERS AND DECANOL surface tension interfacial tension A - B A B Me hexanoate 56.15 0.101 61 12.60 0.017 86 Me octanoate 56.1 1 0.098 04 14.28 0.021 61 Me decanoate 56.69 0.096 25 13.37 0.031 25 Me dodecanoate* 55.71 0.089 13 14.08 0.033 57 Me t etradecan oat e 57.20 0.092 32 14.31 0.036 96 1 -decanol 53.21 0.084 00 - 10.67 0.067 00 * For the surface tension, the values of Vogel between 13 and 87°C have been used.liquid surfaces. temperature, it can be expressed as Since, for all the systems studied, Wip is a linear function of the w.,s = C-DT. (3) The constants C and D are, respectively, Ahlp and A,$. Values of C and D for the temperature range 20-32°C are given in table 5.In the calculation of W f , values for the surface tension of water were taken from ref. (4). TABLE s.-cONSTANTS OF EQN (3) FOR w:p IN THE RANGE 20-32°C" chain lengtht 6 8 9 10 11 12 14 alkanelwater C D esterlwater C D 91.21 0.1735 162.53 0.2771 90.86 0.1644 160.60 0.2766 91.59 0.1636 - - -- - 162.02 0.2843 92.46 0.1624 - - 92.43 0.1611 160.40 0.2796 92.84 0.1598 162.03 0.2875 Wzp is given in mJ m-2 ; -f for the esters, the chain length refers to the parent acid ; * for the decanoliwater interface, C = 182.59 and D = 0.3078.R . AVEYARD, B . J . BRISCOE A N D J . CHAPMAN 13 For the orientation of the organic molecules at the liquid/liquid interface, it appears likely that singly adsorbed n-alkane molecules on a water surface will assume mainly those configurations which are parallel to the s~rface.~ However, when the alkanes are present at the interface in a condensed state, as at the liquid/liquid inter- face, it may be more favourable for the chains to align at a steep angle to the interface.On energy considerations, Ohki and Fukuda have concluded that a vertical orienta- tion, where the chains can interact mutually along their long axes, is more probable than an orientation parallel to the interface. On the other hand, Pomerantz, Clinton and Zisman believe that the alkanes (at least up to octane) are probably parallel to the alkane/water interface, and state that there is much recent evidence to support this. For long chain polar compounds it is likely that a near vertical orientation will be assumed in order to pack as many polar groups as possible into contact with water at the interface.We consider first the work of adhesion in alkanefwater systems. We suppose the alkanes are parallel to the alkane/water interface and that the alkane side of the interface can be considered as a mixture of -CH2- and -CH3 groups. If the work of adhesion Wl of the methyl group differs from that, W2, of the methylene group, Wip should vary with chain length in a way to be determined. The mol fractions, x1 and x2, of -CH, and -CH2- groups in an alkane con- taining N carbon atoms, are given by If the molar surface areas of -CH3 and -CH2- groups are, respectively, c1 and a2, and if the surface is monomolecular with respect to alkane, then Q, the area of the interface is = 2 / N ; ~2 = 1-2fN.(4) 0 = iz9,0,+n~a2, ( 5 ) where ni and 12; are the numbers of moles of -CH, and -CH2- groups respectively in the surface. Thus, WIp can be expressed as Combination of (4), (5) and (6) yields w5y2(r, - 1) 3. N ) = N w. + 2(r, w, - W,), (7) where r1 is defined as a,/o,. parallel to the interface, we obtain Using similar arguments for the esters, assumed to be W;yN + r l f r3 - 2) = N W, + (7.1 W1+ r3 w3 - 2W,), (8) where W3 is the work of adhesion of the -COOMe group with water and r3 = a,/a2 ; o3 is the molar surface area of the -COOMe group. N in (8) is the number of carbon atoms in the parent acid. Implicit in the above arguments is the assumption that the orientation of the chains is the same at the liquid/vapour interface and the liquid/liquid interface.This may not be true but it seems reasonable to suppose that the major contribution to WT will arise from removing the liquid/liquid interactions, and that the reorient- ation of the chains, if it occurs, gives rise to a relatively small change in free energy. However, the validity of the assumption can only be judged by the measure of agree- ment obtained between values of Wl and W2 (and the corresponding enthalpy changes) calculated using the theory, and values obtained independently. A plot of the left-hand side of (7) or (8) against N should be linear, with the slope of the line equal to W,, if the assumed model approximates to the real situation.14 ADHESION AT WATER SURFACES Further, for the alkanes, Wl can be calculated from the intercept of the plot according to (7).Values of Wl and W2 can then be compared with independently-obtained values. Adam and Elliott,8 from a study of contact angles of water on various solid hydrocarbons, concluded that the work of adhesion of the methylene group is about 54mJm-2 and of the methyl group about 40rnJnr2. These values were obtained using the mean of the advancing and receding contact angles and are not therefore precise, but they may be used as a guide. Plots according to (7) and (8) for the alkanes and esters respectively are shown in fig. 1. Values of crl and c2 have been taken as 0.1 1 and 0.05 nm2 respectively,' and cr3 has been assigned a value of 0.20nm2. Both plots are apparently linear. The value of W, obtained from the plot for the alkanes is 52 mJ m-2 in good agreement with the findings of Adam and Elliott.From the same plot W1 is found to be 30 mJ ~tl.-~, which is lower than the value of 40 mJ m-2 found from contact angles. Some of this difference may arise from different values of c1 at the liquid and at the solid surfaces. 1 1 N E m i 6 8 I0 12 14 N FIG. 1.-Plots according to eqn (7) and (8) for 20°C. 0 and left-hand ordinate refer to alkanes and and right-hand ordinate to esters. Aveyard and Haydon reported that the plot of WiB against 1/N for the alkanes is linear. Inspection of (7) shows that strictly, according to the model proposed, the plot should only be linear if r1 = 1. However, in the range N = 6-14 such a plot is almost linear for rl = 2.2 (i.e., for ol = 0.1 1 and cr2 = 0.05 nm2 molecule-'). Assuming linearity, the value obtained for W2 (at 1/N = 0) is very close to the value found using (7).The value obtained for Wl, however, is about 30 % lower than that obtained using (7). W, calculated from a plot according to (8) for the esters is 73 mJ m-2, which is unacceptably high. Different choice of magnitudes for rl and r3 does not significantly alter this value. It is thus reassuring that the use of the model for " flat " molecules on a system where the molecules are likely to be essentially vertical, does not fortuit- ously lead to a reasonable value for W2. Ahip for the alkane + water systems (table 5) is almost independent of N indicatingR . AVEYARD, B . J . BRISCOE A N D J . CHAPMAN 15 that, if the alkane molecules are parallel to the surface, the heats of adhesion of the -CH3 and -CH2- groups are of similar magnitude. The enthalpies of adhesion of surfaces comprised of solely methyl or methylene groups can be obtained using an equation analogous to (7), for Ah?.Values so obtained are for the methyl group 88 mJ m-2 and for the methylene group 95 mJ m-2. Confirmation of the reason ableness of these quantities has been sought by comparison with standard equilibrium heats of desorption (-d,H*) of n-alkanes from the surface of water. Although the process of parting a liquid/liquid interface is not equivalent to desorption from a dilute alkane monolayer, the magnitude of the enthalpy changes for the two processes should be similar, if the orientation of the alkanes at the interface is the same in both cases. From the unit area values already quoted, the molar enthalpy of adhesion of the -CH3 group is 5.8 kJ mol-I and of the -CH2- is 2.7 kJ mol-I.The values of AHiB for the whole molecules are plotted in fig. 2 together with -A,H* obtained by Jones and Ottewill at 7.5OC.’ The two sets of data are in good agreement. For example, for octane, AHiB = 28.8 kJ mo1-I and -A,H* = 30.6 kJ mol-I. In addition, the -CH2- increment in -AaHe is about 2.8 kJ mol-l, close to the cor- responding increment of 2.7 kJ mol-l in AHiP. In view of this agreement, and since the adsorbed alkanes are likely to be flat on the water surface, the adhesion data are consistent with a parallel orientation at the alkanelwater interface. 6 a 10 12 14 N at 7.5% and 0 to AHZP in the temperature range 20-32°C. FIG.2.-Enthalpies of desorption and adhesion for systems of alkanefwater. refers to -Aa,Hf3 Ah? for the esters is also independent of chain length (see table 5 ) and is 161.64 1 mJ m-2. The high value is indicative of the interaction of the -COOMe group with water. The constancy of Ahip indicates that the molecules are not parallel to the surface. A flat orientation would give rise to a marked chain length dependence of Ahib since the enthalpies of adhesion of the -CH3 and -CH2-- groups are much smaller than that for the -COOMe group. AhIP for n-decanol is 182.6 mJ m-2. If the surface area of decanol is supposed to be the same as the close-packed area in insoluble films of long-chain n-alkanols on water, i.e., 0.216 nrn2,l0 the molar enthalpy of adhesion is 23.8 kJ mol-’. As might be expected the value is not very different from the heat of desorption (-31 kJ mol-I) of alkanols from the alkanelwater interface to dilute solution in the alkane.16 ADHESION AT WATER SURFACES To summarize, using a simple theory, values of the work and enthalpies of ad- hesion are consistent with a predominantly flat orientation of alkanes, and an essen- tially vertical orientation of long-chain methyl esters, at the interface between these liquids and water. Probably the main objection to the proposed theory is the neglect of the effects which could arise as a result of the organic molecules assuming different configurations on parting the liquid/liquid interface. B. J. B. and J. C. are indebted to the Science Research Council for provision of maintenance grants. R. Aveyard and D. A. Haydon, Trans. Faraday SOC., 1965, 61, 2255. Selected Properties of Hydrocarbons and related compounds, API Project 44 (Thermodynamics Research Center, Texas, 1966). A. I. Vogel, J. Chem. SOC., 1948, 638. Int. Crit. Tables (McGraw-Hill, 1928), 4, 447. J. H. de Boer, The DynamicaZ Character of Adsorption (Oxford, 1953), p. 151. S. Ohki and N. Fukuda, J. CoZZoid Interface Sci., 1968, 27, 208. ' P. Pomerantz, W. C. Clinton and W. A. Zisman, J . CoZZoid Interface Sci., 1967, 24, 16. * N. K. Adam and G. E. P. Elliott, J. Chem. SOC., 1962,2206. D. C. Jones and R. H. Ottewill, J. Chem. Soc., 1955,4076. lo N. K. Adam, The Physics and Chenzistry of Srtrfaces (Oxford, 1938), p. 50.

ISSN:0300-9599    

DOI:10.1039/F19726800010

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Infra-Red Study of the Reactions of Sihcon, Titanium and Tin Tetrachlorides with Rutile BY G. D. PARFITT,* J. RAMSBOTHAM AND C. H. ROCHESTER Chemistry Dept., The University, Nottingham, NG7 2RD Received 26th May, 1971 Hydroxyl groups on rutile react with silicon tetrachloride to form hydrogen chloride and TiOSi bonds. Nucleophilic attack of the silicon atom by oxide ions in the surface also occurs, and causes the displacement of chlorine atoms which adsorb on titanium ion sites. Unreacted chlorine atoms in the chemisorbed species are hydrolyzed by water to SiOH groups. The hydrogen chloride formed adsorbs on remaining exposed rutile surface. In the presence of adsorbed ammonia or pyridine the modified surface, after desorption of hydrogen chloride, exhibits Brsnsted acidity which is enhanced by the addition of water.Brarnsted acidity was also observed for a rutile surface treated with tin tetrachloride, hydrolyzed, and evacuated at 400°C to remove hydrogen chloride. The similar reactions of silica and rutile surfaces with titanium tetrachloride followed by hydrolysis and evacuation at 400°C conferred Brsnsted acidity on silica but not on rutile. The modified silica surface also contained Lewis acid sites. Studies of the adsorption of silicon tetrachloride and alkyl substituted chlorosilanes on silica have helped in the characterization of hydroxyl groups on the oxide surface. Infra-red measurements of the kinetics of the reactions have given information on the proportions of isolated and geminal hydroxyl groups. Studies using chloromethyl- silanes, titanium tetrachloride, and boron trichloride together with gravimetric measurements of water adsorption and desorption led Armistead et aL2 to identify thermally stable isolated hydroxyl groups and hydrogen-bonded groups held on an array of different crystal planes.At high temperatures, chloromethylsilanes react not only with hydroxyl groups but also with oxygen atoms (in surface siloxane bridges) which act as nucleophiles towards adsorbate molecules. The modification of the surface which is caused by adsorption of silicon tetrachloride followed by hydrolysis has been discussed by Peri and can lead to the formation of a polysiloxane layer which restricts access of adsorbate molecules to the underlying ~urface.~ This paper reports an investigation of the use of silicon tetrachloride and the related compounds titanium tetrachloride and tin(1V) tetrachloride for the characterization of hydroxyl groups on rutile. Ammonia and pyridine adsorption has been used to monitor the changes in acidity which occur when the surface is modified by adsorption of MCl, (M = Si, Ti, or Sn) followed by hydrolysis.EXPERIMENTAL The source and purification of rutile (BTP code numbers DD/CL/9 and 10, CL/D338’), ammonia, oxygen and hydrogen and pyridine and deuterium oxide have been described. A further sample of rutile (CL/D428) was similar in preparation and purity to those above and had a surface area of 24.3 m2 8-l. Titanium tetrachloride was supplied by British Titan Products Co. Ltd. and tin(IV) tetrachloride and silicon tetrachloride by B.D.H.Chemicals Ltd. All were distilled under vacuum before use and cold-pumped to remove dissolved gases. * present address : Tioxide International Ltd., Billingham, Teesside. 1718 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES Spectra were recorded using a Perkin-Elmer 125 infra-red spectrometer with a pro- grammed slit width control giving resolutions of 7 cm-' at 4000 cm-' falling to 3.8 cm-' at 2000 cm-l and 4.5 cm-l at 1500 cm-l ; resolutions of ca. 1 cm-' over these ranges gave no additional jnformation. Spectra were recorded with the rutile discs at ambient beam temperatures (ca. 45°C) using the cell described elsewhere.8 Interference from atmospheric water vapour in the spectrometer beams was minimized by purging with dry air.RESULTS A rutile sample was heated in oxygen (4 h, 400"C, 6.7 kN m-2) evacuated (3 h, 400°C) and then rehydroxylated by equilibrating with excess water at 45°C to produce a starting surface similar to that used in previous adsorption studies.6* '* Evacua- tion produced a surface giving rise to bands at 1615, 3420 and 3660cm-l with a shoulder at 3690 cm-l (fig. la). Contact with excess silicon tetrachloride vapour and subsequent evacuation reduced the band at 3660cm-I to a shoulder, reduced the intensity of the bands at 1615 and 3420cm-1 and the shoulder at 3690cm-l, 3700 3500 3300 3100 1650 1550 wavmumber/cm-' FIG. 1.-The reaction of rutile with SiCL at 45°C. (a) rutile surface, equilibrated with saturated HzO vapour and evacuated (17 h, 50°C) ; (b) in contact with SiC14 (+ h, 45"C, 4.7 kN m-2) and evacuated (1-5 h, 45°C) ; (c) in contact with H20 vapour (13 h, 45"C, 2 kN m-2) ; (d) evacuated (1 h, 45°C).and gave a new band at 3580 cm-I (fig. lb). Equilibration of this surface with water vapour produced a broad band centred on 3400cm-l with shoulders between 3620 and 3700 cm-I and at ca. 3520 cm-l together with bands at 1605 and 1580 cm-I and a shoulder at ca. 1620 cm-1 (fig. lc). Evacuation for periods of up to 1 h at 45°C resulted in the steady growth of a band centred on 1565 cm-l and the initial growth of bands at 1580 and 1605 cm-l followed by the disappearance of the former and the diminution of the latter (fig. Id). The broad band at 3400 cm-l was reduced in intensity and bands appeared at 3730 and 3550 cm-I with a shoulder at 3660 cm-1 (fig.Id). Similar variations in the intensities of bands at 1565, 1580 and 1605 cm-1 with evacuation time were obtained using starting surfaces which were evacuated at 250°C (fig. 2) and 400°C before treatment with silicon tetrachloride at 45°C. ThereG. D. PARFITT, 3 . RAMSBOTHAM AND C. H. ROCHESTER 19 were, however, minor differences in the rates of change of band intensities. The rates of growth of the band at 1565 cm-I and disappearance of that at 1580 cm-I increased with increasing outgassing temperature. The change in intensity of the bands around 1600 cm-I as a function of time of evacuation is shown in fig. 2. Evacuation at > 150°C removed the peak at 1565 cm-1 and it was not reformed when water was subsequently admitted and then pumped off.Further treatment of this surface with silicon tetrachloride followed by water vapour produced only a very weak band at 1565 cm-l together with one much more intense at 1605 cm-l. 1650 1550 1650 1550 wavenumber/cm-l FIG. 2.-Spectra of rutile previously equilibrated with H,O and evacuated (134 h, 250"C), treated with SiC14 (+ h, 45"C, 4.7 kN m-2), evacuated (1% h, 45°C) equilibrated with saturated H20 vapour (2% h, 45°C) and evacuated at 45°C ; the numbers indicate pumping time in min. Comparable information was obtained using surfaces exchanged with deuterium oxide vapour. A rutile disc was heated in oxygen and evacuated at 400°C as described above, cooled to ca. 45"C, equilibrated with saturated deuterium oxide vapour (3 h, 400°C) and allowed to cool in contact with the vapour.Evacuation at 50 or 250°C resulted in the spectra shown in fig. 3Aa and 3Ba respectively. Silicon tetrachloride reduced the intensity of bands at 2540, 2695 and 2750 cm-l, caused the appearance of a new band at 2650 cm-l and removed a band at 2725 cm-l (fig. 3B) or a corre- sponding shoulder at 2715 cm-1 (fig. 3A). Equilibration of the surface initially outgassed at 50°C with saturated deuterium oxide vapour and subsequent evacuation resulted in intense bands at 2540 and 2750 cm-I together with a shoulder at 2650 cm-l20 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES (fig. 3Ac and d). Analogous results from a surface initially evacuated at 250°C showed an additional shoulder at 2695 cm-l while that at 2650 cm-l was not as marked (fig.3Bc and d). 2900 2700 2500 2900 2700 2500 wavenumber/cm- FIG. 3.-The spectra of deuterated rutile surfaces. A: (a) pumped free of excess DzO vapour (6$ h, 50°C) ; (b) in contact with SiC14 (3 h, 45"C, 4.7 kN m-2) and evacuated (14 h, 45°C) ; (c) in contact with saturated D,O vapour (23 h, 45°C) and evacuated (1 min, 45°C) ; ( d ) further evacuation (3 h, 45°C). B : (a) pumped free of excess DzO vapour (16 h, 250°C) ; (b) in contact with SiCI4 (3 h, 45"C, 4.7 kN m-2) and evacuated (14 h, 45°C) ; (c) in contact with saturated DzO vapour (23 h, 45°C) and evacuated (1 min, 45°C) ; (d) further evacuation (% h, 45°C). Experiments were carried out with a reduced rutile surface to test whether com- parison with the results for a surface heated in oxygen gave any evidence for the participation of oxide sites in the adsorption process.Rutile was reduced by heating it in hydrogen (+ h, 400"C, 8 kN m-2) and then evacuating (I h, 400°C) before treating with silicon tetrachloride. Reduction did not change the intensity of the infra-red bands arising from hydroxyl stretching vibrations and treatment with silicon tetrachloride (4 h, 45"C, 6.7 kN m-2) left only a broad band centred at 3580 cm-l (cf. fig. lb). Contact with saturated water vapour (9 h, 45°C) followed by evacuation (<+ h, 45°C) produced a band at 1605 cm-l with a shoulder at 1625 cm-l. Further evacuation (>4 h, 45°C) left only the band at 1605 cm-l and an absorption maximum centred on 3400 cm-l while pumping for 1 h at 400°C led to a surface for which only an absorption band at 3660 cm-l remained In order to study the reaction of silicon tetrachloride with rutile at elevated temperatures, a starting surface was evacuated (13 h, 250°C) and gave rise to infra- red bands at 3660 and 3700 cm-l with a broad shoulder at ca.3400 cm-1 which were removed on treating with excess silicon tetrachloride (1 h, 25"C, 2 kN m-2). Subsequent equilibration with excess water vapour (2 h, 45°C) and evacuation pro- duced an absorption band at 1615 cm-l due to molecular water together with a band at 3730 cm-l and a broad absorption maximum centred on ca. 3400 cm-l. Evacua- tion (2 h, 250°C) left only an intense band at 3740 cm-l together with a weaker oneG. D. PARPITT, J . RAMSBOTHAM AND C. H . ROCHESTER 21 at 3670 crn-l.The interaction of silicon tetrachloride vapour with a rutile surface exchanged with deuterium oxide vapour gave similar results. After exchange with deuterium oxide vapour the rutile was evacuated and gave bands at 2540, 2600, 2695 and 2750 cm-l (fig. 4a). A residual band at 3660 cm-l was due to incomplete exchange of OH group^.^ Treatment with excess silicon tetrachloride, cooling and 2700 2500 2300 1600 1500 wavenumber /cm- FIG. 4.-Reaction of SiC14 with deuterated rutile at 250°C. (a) Rutile pumped free of saturated DzO vapour (JJ h, 50°C) ; (6) in contact with SiC14 (3 h, 250°C. 4.7 kN m-2) ; cooled to 45" and evacu- ated (I+ h) ; (c) in contact with saturated H20 vapour (3 h, 45°C) and evacuated at 250°C (10 h) then 400°C (3 h) ; ( d ) in contact with NH3 (3 h, 45"C, 0.67 kN m--2) and evacuated (5 min).evacuation left small bands at 2540 and 2750 cm-' (fig. 4b) and these decreased in intensity but were not removed after contact with saturated water vapour. Evacua- tion (-$min, 45°C) left bands at 1615, 2760 and 3750cm-l with a broad band at ca. 3450 cm-l and pumping at 400°C left bands at 2760 and 3750 cm-l with a shoulder at ca. 3400 cm-I (fig. 4c). The acidity of this surface was investigated by ammonia and pyridine adsorption. Ammonia vapour rapidly interacted with the species resulting in the band at 276Ocm-l and produced bands at 3360 and 1605cm-l (fig. 4d). The band at 3750 cm-1 due to SiOH groups was hardly affected by ammonia treatment but a band due to the ammonium ion generated on surface Brarnsted sites appeared at 1420 cm-l.Evacuation (lh, 400°C) left a surface similar to that before ammonia adsorption except that the intensity of the 2760 cm-l peak was reduced. Hydrogen chloride vapour did not produce any further change in intensity of the latter band but gave rise to a shoulder at 3400 crn-l. The adsorption of pyridine vapour onto a similar surface to that treated with ammonia (fig. 5a) greatly reduced the intensity of the peak at 3750 cm-1 giving a broad band at ca. 3400 cm-l together with bands at 1450, 1490, 1575, 1610, 1640 and 1540cm-l (fig. 5b and c). The latter two bands arise from the pyridinium ion and confirm the presence of Brnrnsted sites on the modified surface. Evacuation followed by contact with water vapour (fig. 5 4 then further evacuation (fig.5e) increased the intensity of the bands at 1540 and 1640 cm-l while decreasing that of the other bands in this region.22 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES The interaction of tin(1V) tetrachloride with rutile was investigated to determine if it was comparable to that of silicon tetrachloride and produced surfaces which exhibited Brarnsted acidity in the presence of adsorbed ammonia or pyridine. The spectrum of rutile after evacuation (17* h, 400°C) showed a band at 3700 cm-l with weak shoulders at 3660 and ca. 3500 cm-l. Reaction with tin(1V) tetrachlorjde wavenumber/cm- FIG. 5.-Surface acidity of rutile treated with SiC14 at 250°C. (a) Rutile surface with a pretreatment similar to that described in fig. 4a-c followed by evacuation ($ h, 400°C) ; (b) in contact with pyridine (0.67 kN m-2, 45°C) ; (c) in contact with pyridine (3 h, 45"C, 8 kN m-2) and evacuated (2 h, 45°C) ; (d) in contact with H20 vapour (45"C, 1.33 kN m-2) ; (e) evacuated (3 h, 45°C).(% h, 45"C, 2.7 kN m-2) followed by evacuation (3 h, 45°C) removed the band at 3700cm-l and resulted in a peak of slightly lower intensity at 3650cm-l with a more intense shoulder than before now centred on 342Ocm-l. Contact with saturated water vapour (13 h, 45°C) and subsequent pumping (+ h, 45°C) resulted in bands at 3350 and 1610cm-l which resolved into a band at 3360cm-l with a shoulder at 358Ocm-l and a broad absorption between 1550 and 1610cm-l. Evacuation (1 h, 400°C) left a peak at 3660 cm-l with a weak shoulder at 3500 cm-l. Contact of tin(1v) tetrachloride (+ h, 45"C, 1.3 kN m-2) with a rutile surface after equilibration with saturated water vapour and evacuation (5 s, 45°C) shifted a band originally at 1625 cm-l to two centred at 1565 and 1600 cm-l.The reaction of excess tin(1V) tetrachloride with a rutile surface followed by evacuation equilibra- tion with saturated water vapour, oxygen treatment and evacuation gave a surface which exhibited Brarnsted and Lewis acidity and hydrogen bonding interactions with pyridine vapour (fig. 6b). Bands characteristic of Brarnsted acidity at 1540 and 1640 cm-l were enhanced by water vapour which removed the bands characteristic of hydrogen bonding at 1440 and 1607 cm-l (cf. fig. 6b and c). A band at 1450 cm-l was removed by further evacuation to leave bands at 1455,1495,1575 and 1618 cm-l, due to pyridine on Lewis acid sites and bands at 1540 and 1640 cm-l due to the pyridinium ion (fig.64. The interaction of titanium tetrachloride with rutile was studied to test whether the modifications to the rutile surface produced by t i n o tetrachloride and silicon tetrachloride also arise when the metal in the adsorbate is the same as that in theG . D. PARFITT, J . RAMSBOTHAM AND C. H. ROCHESTER 23 50 1; 'Jb - - 'I I 1 I I I ao 50 40 30 u 3 9 0 0 3700 3 5 0 0 33OC I I I I I : *. - I I I I ~ 1650 1550 1450 wavenumber /cm- FIG. 7.-The reaction of TiCI4 with silica. (a) SO2 evacuated (13 h, 400°C) ; (b) in contact with saturated Tic14 (5 h, 250°C) ; (c) in contact with excess TiC14 at 400°C ( >3 h), evacuated (+ h, 45"C), hydrolyzed with saturated water vapour (3; h, 45"C), and evacuated (t h, 45°C) ; this treatment repeated five times followed by evacuation (+ h, 250°C) ; (d) in contact with saturated pyridine vapour (+ h, 45°C) and evacuated (+ h, 45°C) ; (e) equilibrated with excess water vapour and evacuated (& min, 45°C).24 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES adsorbing oxide.After equilibration of a rutile surface with saturated water vapour and evacuation (1 h, 2OO0C), a sharp peak at 3665 cm-1 with a shoulder at ca. 3450 cm-1 remained. Adsorption of excess titanium tetrachloride vapour (4 h, 45"C, 0.5 kN m-2) followed by evacuation (3 h, 45°C) and further equilibration with excess water vapour and pumping (23 h, 45°C) gave bands at 1565, 1610 and 3400 cm-' with a shoulder at 3660 cm-l.In the presence of pyridine vapour, this surface did not give bands at 1540 and 1640 cm-l arising from Bronsted acidity. A silica (Degussa aerosil) disc was evacuated and gave infra-red bands at 1635 and 3750 cm-' (fig. 7a). Reaction with excess titanium tetrachloride did not change the band at 1635 but reduced that at 3750cm-l to a broader band at 3680cm-l (fig. 7b). Subsequent equilibration with saturated water vapour (23 h, 45°C) and evacuation (4 h, 400°C) produced bands at 3680 and 3750 cm-l. Fig. 7c shows the result of five similar consecutive treatments with titanium tetrachloride followed by hydrolysis and evacuation. Contact with pyridine vapour and evacuation left bands characteristic of Lewis acidity at 1445, 1490, 1575 and 1610 cm-l and weaker ones, characteristic of Brarnsted acidity, at 1540 and 1640 cm-l (fig.7d). The addition of water enhanced the Bronsted acidity of the surface (fig. 7e). Excess pyridine vapour was completely removed by evacuation (< 3 h, 45°C) from a silica disc which had not been pretreated with titanium tetrachloride. No infra-red bands characteristic of pyridine adsorbed on Lewis or Brarnsted acid sites were observed. Pyridine is adsorbed on silica by hydrogen-bonding with the surface hydroxyl groups.16 DISCUSSION The bands arising from hydroxyl groups and water at 3700, 3660, 3620, 3520, 3410 and ca. 1610cm-l and their deuterium analogues at 2730, 2700, 2670, 2620, 2540 and ca. 1185 cm-1 have been discussed elsehwere.1° The small shoulder observed in some spectra at 3730 cm-l (2750 cm-I on deuterated surfaces) results from trace amounts of silica in the samples.The reaction of silicon tetrachloride with hydroxyl groups on rutile (fig. l a and b) is consistent with the formation of surface TiOSiC13 species and the elimination of hydrogen chloride. Individual silicon tetrachloride molecules may also react with two hydroxyl groups 1 * although the relative proportions reacting with one or two hydroxyl groups cannot be deduced from the present results. A band observed at 3580 cm-1 after adsorption arises from weak hydrogen bonding of the new surface species with unreacted hydroxyl groups. Molecular water is displaced from the surface but does not cause hydrolysis of Sic1 groups to SiOH as shown by the absence of an absorption band at ca.374Ocm-l. However, the addition of water vapour to the treated surface causes hydrolysis of unreacted Sic1 bonds to SiOH groups for which the characteristic absorption band appeared at a position in the range 3730-3750 cm-l depending upon the precise experimental conditions. Hydrogen chloride produced by the hydrolysis leads to the appearance of infra-red bands at 3400, 1605, 1580, and 1565 cm-l. The 3400, 1605 and 1565 cm-l frequencies are similar to those of 3360, 1600 and 1565 cm-l for the corresponding bands which appear when hydrogen chloride alone is adsorbed on r ~ t i l e . ~ The tentative assign- ment of these bands to molecular water present on the surface either as hydronium ions or coordinately bound to Ti4+ ions is discussed el~ewhere.~ The band at 1580 cm-l which decreases in intensity as water is evacuated from the rutile surface and whose lifetime is a function of the pretreatment evacuation temperature of the starting surface is only observed after adsorption of silicon tetrachloride followed by hydrolysis.It did not appear after analogous treatments with hydrogen chloride, titanium tetra-G. D . PARFITT, J . RAMSBOTHAM AND C. H . ROCHESTER 25 chloride or tin tetrachloride. The decrease in intensity of the 1580 cm-l band on evacuation is accompanied by a corresponding increase in the intensity of the 1565 cm-I band (fig. 2) suggesting that the species responsible for the former is a precurser of that responsible for the latter. The following formulation is consistent with the previous assignment of the 1565 em-I band.g (1580 cm') (1565 cm') The 1580 cm-I band is ascribed to the bending vibration of molecular water associated via hydrogen-bonding with a hydrogen chloride molecule on the surface.Evacuation causes loss of water and also results in the growth (fig. Ic and d ) of the 3730 cm-l absorption band of isolated SiOH groups. The interaction of silicon tetrachloride with a deuterated rutile surface and sub- sequent treatment with deuterium oxide gives results which are in accord with those for the hydroxylated surface although the removal of deuteroxyl groups is apparently more complete than that of hydroxyl groups. This arises because there is a broad absorption maximum at ca. 3660cm-l due to hydroxyl groups which cannot react with silicon tetrachloride and which may be attributed either to bulk hydroxyl groups or to interparticulate hydrogen-bonding.Thus exchange with deuterium oxide (fig. 3a) left a broad absorption band at 3660cm-l which was unaffected by the adsorption of silicon tetrachloride. The removal of the deuteroxyl groups which give absorption bands at 2695 and ca. 2720cm-l is caused in part by chemical reaction and also by hydrogen-bonding (which leads to the appearance of a band at 2650 cm-l) of the TiOSiC13 or (TiO)2SiC12 surface species with unreacted groups. Treatment with deuterium oxide causes deuterolysis of Sic1 groups to SiOD and the formation of adsorbed deuterium chloride which on evacuation leads to an intense band at 2540 cm-l (cf. 3400 cm-1 for an hydroxylated surface). No infra-red bands appear in the range 1550-1620 cm-l.The deuterium analogues of the species responsible for these bands should give corresponding bands in the range 1140- 1180 cm-l but these cannot be reliably studied because of strong background absorp- tion by the bulk rutile. Heat treatment of rutile in hydrogen results in a reduction in the number of oxide ions in the surface but leaves the hydroxyl population ~nchanged.~ Hydrogen chloride therefore interacts with reduced rutile predominantly by reaction with H / \ hydroxyl ions to give the species Ti4+0 C1- (infra-red band at 1565 cm-l) rather H than by reaction with oxide ions which would give hydroxide and chloride ions on adjacent titanium ion sites.g The reaction of reduced rutile, evacuated at 400"C, with silicon tetrachloride followed by hydrolysis contrasts with the corresponding reaction for an oxygen-treated rutile in that neither the 1565 nor the 3730 cm-l band are observed for the former.The difference is inexplicable in terms of reaction of silicon tetrachloride with surface hydroxyl groups because the reduced and oxidized surfaces have similar low hydroxyl populations. This suggests that an oxidized surface which has been evacuated at 400°C before the addition of silicon tetrachloride reacts26 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES predominantly, even at 45"C, by the nucleophilic attack of silicon atoms by surface oxide ions.3 Reaction (2) -0SiCl3 I c1- Ti4+02-Ti4+ + SiC14+Ti4+ Ti4+ (2) is probably occurring.This reaction does not occur on the reduced surface because the appropriate oxide sites are removed by heating rutile in hydrogen. The addition of water vapour to a silicon tetrachloride treated reduced rutile causes rehydroxylation of the treated surface and the reappearance of an absorption band at 3660 cm-l. No band at 3700 cm-l appears because the appropriate hydroxyl groups are not removed by the pretreatment (400°C evacuation lo) and have therefore all reacted with silicon tetrachloride. Rutile, in common with anatase l2 and alumina,ll displays Lewis acidity in the presence of adsorbed ammonia6 or pyridine.' The addition of water does not promote Brarnsted acidity. Brnrnsted acidity is observed for silica-aluminas and zeolites and is enhanced by the adsorption of water.At 250°C silicon tetrachloride undergoes reaction with all the hydroxyl groups on rutile. Extensive reaction with oxide ion sites probably also occurs. Hydrolysis and evacuation lead to an infra-red spectrum which is more characteristic of silica than of rutile. Subsequent adsorption of ammonia gave bands at 1605 and 3360cm-l due to ammonia coordinated to titanium ions and a broad shoulder at 1420 cm-1 due to ammonium ions formed by adsorption of ammonia on Brnrnsted acid sites. Alternatively, the adsorption of pyridine gave bands characteristic of both the pyridinium ion (1 540 and 1640 cm-l) and coordinated complexes (fig. 5c). The marked changes in intensity of the band at 3750 crn-l suggest that the pyridine molecules may also coordinate with the rutile surface while perturbing adjacent Si-OH groups. The adsorption of water on the surface enhances Brnrnsted acidity at the expense of Lewis acidity (cf.fig. 5c and e) suggesting that water interacts with Lewis sites formerly occupied by pyridine mole- cules. Similar Brarnsted activity was exhibited by rutile surfaces treated with tin(1V) tetrachloride and water but not by surfaces treated with titanium tetrachloride and water. These results indicate that M-OH groups on rutile where M is not Ti show Brarnsted acidity towards ammonia and pyridine vapour and behave like surfaces in mixed oxide systems.l Similar reactions of titanium tetrachloride and water on a silica aerosil produced four infra-red bands in the range 1400-1700 cm-I, which are characteristic of pyridine coordinated to the surface, together with two bands at 1540 and 1640 cm-l which arise from Brarnsted acidity and are enhanced by adsorbed water vapour (fig.7). These data indicate that surface hydroxyl groups on heterometal atoms exhibit Brarnsted acidity and also the Lewis acidity of the surface is modified. Electrophoresis studies indicate that the presence of Si-OH groups on rutile lower the isoelectric point towards that of ~i1ica.l~ However, electrophoresis studies by Parfitt and Ramsbotham l4 using rutile coated with varying proportions of silica and alumina indicated that although hydroxyl groups on surfaces containing a high proportion of silica appear to be similar to those on a pure silica surface, those on surfaces of high alumina content are not similar to those on alumina.Infra-red data by Smith l5 for coatings on rutile pigments suggest that bands arising from hydroxyl groups in the coatings " tended to lose their individual identity ". These latter points suggest that hydroxyl groups on metals other than titanium on rutile may not necessarily be characteristic of their parent oxide, with the possible exception of Si-OH. The reaction of t i n 0 tetrachloride and water with rutile differs from the corre- sponding reaction of silicon tetrachloride in that the 1580 cm-' band never appearsG . D . PARFITT, J . RAMSROTHAM AND C . H . ROCHESTER 27 and the 1565 crn-l band is less intense. The addition of pyridine to the rutile surface coated with SnOH groups shows (fig. 6) infra-red bands arising from hydrogen-bonded pyridine (1443 and 1605 crn-l), pyridine coordinated to Lewis acid sites (1455 and 1617 cm-I) and the pyridinium ion (1450, 1540 and 1640cm-I).The bands at 1495 and 1575 cm-l are common to all types of pyridine on the surface.16 Hydrogen- bonded pyridine is displaced by the adsorption of water which enhances the formation of the pyridinium ion. The authors thank British Titan Products Co. Ltd. for a fellowship (J. R.) and for rutile samples. ' M. L. Hair and W. Hertl, J. Phys. Chem., 1969,73, 2372. ' C. G. Arrnistead, A. J. Tyler, F. H. Hambleton, S. A. Mitchell and J. A. Hockey, J . Phys. Chern., 1969,73, 3947. J. A. Hockey, J. Phys. Chem., 1970,74, 2570. J. B. Peri, J. Phys. Chem., 1966, 70, 2937. A. A. Isirikyan, A. V. Kiselev and E.V. Ushakova, KolZoidZhur., 1964,26,45 ; A. A. Isirikyan, A. V. Kiselev and E. V. Ushakova, KolZoidZhur., 1965,27, 690. G. D. Parlitt, J. Ramsbotham and C. H. Rochester, Trans. Faraduy Soc., 1971,67, 841. ' G. D. Parfitt, J. Ramsbotham and C . H. Rochester, Tram. Faruday Soc., 1971,67, 1500. * A. Buckland, J. Ramsbotham, C. H. Rochester and M. S . Scurrell, J. Sci. Imtr., 1971,4,146. G. D. Parfitt, J. Ramsbotham and C. H. Rochester, Tram. Faraday SOC., 1971, 67, 3100. lo P. Jackson and G. D. Parfitt, Trans. Faraday Soc., 1971,67,2469. 11 L. H. Little, Infrared Spectra of Adsorbed Species, (Academic Press, London, 1966). l3 G. D. Parfitt, J. Ramsbotham and C. H. Rochester to be published. l4 G. D. Parfitt and J. Ramsbotham, J.O.C.C.A., 1971,54, 356. Is I. T. Smith, Nature, 1964, 201, 67.l 6 E. P. Parry, J. Catalysis, 1963,2, 371. N. D. Parkyns, Symp. Chemisorption and Catalysis, (Institute of Petroleum, London, 1970). Infra-Red Study of the Reactions of Sihcon, Titanium and Tin Tetrachlorides with Rutile BY G. D. PARFITT,* J. RAMSBOTHAM AND C. H. ROCHESTER Chemistry Dept., The University, Nottingham, NG7 2RD Received 26th May, 1971 Hydroxyl groups on rutile react with silicon tetrachloride to form hydrogen chloride and TiOSi bonds. Nucleophilic attack of the silicon atom by oxide ions in the surface also occurs, and causes the displacement of chlorine atoms which adsorb on titanium ion sites. Unreacted chlorine atoms in the chemisorbed species are hydrolyzed by water to SiOH groups. The hydrogen chloride formed adsorbs on remaining exposed rutile surface.In the presence of adsorbed ammonia or pyridine the modified surface, after desorption of hydrogen chloride, exhibits Brsnsted acidity which is enhanced by the addition of water. Brarnsted acidity was also observed for a rutile surface treated with tin tetrachloride, hydrolyzed, and evacuated at 400°C to remove hydrogen chloride. The similar reactions of silica and rutile surfaces with titanium tetrachloride followed by hydrolysis and evacuation at 400°C conferred Brsnsted acidity on silica but not on rutile. The modified silica surface also contained Lewis acid sites. Studies of the adsorption of silicon tetrachloride and alkyl substituted chlorosilanes on silica have helped in the characterization of hydroxyl groups on the oxide surface.Infra-red measurements of the kinetics of the reactions have given information on the proportions of isolated and geminal hydroxyl groups. Studies using chloromethyl- silanes, titanium tetrachloride, and boron trichloride together with gravimetric measurements of water adsorption and desorption led Armistead et aL2 to identify thermally stable isolated hydroxyl groups and hydrogen-bonded groups held on an array of different crystal planes. At high temperatures, chloromethylsilanes react not only with hydroxyl groups but also with oxygen atoms (in surface siloxane bridges) which act as nucleophiles towards adsorbate molecules. The modification of the surface which is caused by adsorption of silicon tetrachloride followed by hydrolysis has been discussed by Peri and can lead to the formation of a polysiloxane layer which restricts access of adsorbate molecules to the underlying ~urface.~ This paper reports an investigation of the use of silicon tetrachloride and the related compounds titanium tetrachloride and tin(1V) tetrachloride for the characterization of hydroxyl groups on rutile. Ammonia and pyridine adsorption has been used to monitor the changes in acidity which occur when the surface is modified by adsorption of MCl, (M = Si, Ti, or Sn) followed by hydrolysis.EXPERIMENTAL The source and purification of rutile (BTP code numbers DD/CL/9 and 10, CL/D338’), ammonia, oxygen and hydrogen and pyridine and deuterium oxide have been described. A further sample of rutile (CL/D428) was similar in preparation and purity to those above and had a surface area of 24.3 m2 8-l.Titanium tetrachloride was supplied by British Titan Products Co. Ltd. and tin(IV) tetrachloride and silicon tetrachloride by B.D.H. Chemicals Ltd. All were distilled under vacuum before use and cold-pumped to remove dissolved gases. * present address : Tioxide International Ltd., Billingham, Teesside. 1718 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES Spectra were recorded using a Perkin-Elmer 125 infra-red spectrometer with a pro- grammed slit width control giving resolutions of 7 cm-' at 4000 cm-' falling to 3.8 cm-' at 2000 cm-l and 4.5 cm-l at 1500 cm-l ; resolutions of ca. 1 cm-' over these ranges gave no additional jnformation. Spectra were recorded with the rutile discs at ambient beam temperatures (ca.45°C) using the cell described elsewhere.8 Interference from atmospheric water vapour in the spectrometer beams was minimized by purging with dry air. RESULTS A rutile sample was heated in oxygen (4 h, 400"C, 6.7 kN m-2) evacuated (3 h, 400°C) and then rehydroxylated by equilibrating with excess water at 45°C to produce a starting surface similar to that used in previous adsorption studies.6* '* Evacua- tion produced a surface giving rise to bands at 1615, 3420 and 3660cm-l with a shoulder at 3690 cm-l (fig. la). Contact with excess silicon tetrachloride vapour and subsequent evacuation reduced the band at 3660cm-I to a shoulder, reduced the intensity of the bands at 1615 and 3420cm-1 and the shoulder at 3690cm-l, 3700 3500 3300 3100 1650 1550 wavmumber/cm-' FIG.1.-The reaction of rutile with SiCL at 45°C. (a) rutile surface, equilibrated with saturated HzO vapour and evacuated (17 h, 50°C) ; (b) in contact with SiC14 (+ h, 45"C, 4.7 kN m-2) and evacuated (1-5 h, 45°C) ; (c) in contact with H20 vapour (13 h, 45"C, 2 kN m-2) ; (d) evacuated (1 h, 45°C). and gave a new band at 3580 cm-I (fig. lb). Equilibration of this surface with water vapour produced a broad band centred on 3400cm-l with shoulders between 3620 and 3700 cm-I and at ca. 3520 cm-l together with bands at 1605 and 1580 cm-I and a shoulder at ca. 1620 cm-1 (fig. lc). Evacuation for periods of up to 1 h at 45°C resulted in the steady growth of a band centred on 1565 cm-l and the initial growth of bands at 1580 and 1605 cm-l followed by the disappearance of the former and the diminution of the latter (fig.Id). The broad band at 3400 cm-l was reduced in intensity and bands appeared at 3730 and 3550 cm-I with a shoulder at 3660 cm-1 (fig. Id). Similar variations in the intensities of bands at 1565, 1580 and 1605 cm-1 with evacuation time were obtained using starting surfaces which were evacuated at 250°C (fig. 2) and 400°C before treatment with silicon tetrachloride at 45°C. ThereG. D. PARFITT, 3 . RAMSBOTHAM AND C. H. ROCHESTER 19 were, however, minor differences in the rates of change of band intensities. The rates of growth of the band at 1565 cm-I and disappearance of that at 1580 cm-I increased with increasing outgassing temperature. The change in intensity of the bands around 1600 cm-I as a function of time of evacuation is shown in fig.2. Evacuation at > 150°C removed the peak at 1565 cm-1 and it was not reformed when water was subsequently admitted and then pumped off. Further treatment of this surface with silicon tetrachloride followed by water vapour produced only a very weak band at 1565 cm-l together with one much more intense at 1605 cm-l. 1650 1550 1650 1550 wavenumber/cm-l FIG. 2.-Spectra of rutile previously equilibrated with H,O and evacuated (134 h, 250"C), treated with SiC14 (+ h, 45"C, 4.7 kN m-2), evacuated (1% h, 45°C) equilibrated with saturated H20 vapour (2% h, 45°C) and evacuated at 45°C ; the numbers indicate pumping time in min. Comparable information was obtained using surfaces exchanged with deuterium oxide vapour.A rutile disc was heated in oxygen and evacuated at 400°C as described above, cooled to ca. 45"C, equilibrated with saturated deuterium oxide vapour (3 h, 400°C) and allowed to cool in contact with the vapour. Evacuation at 50 or 250°C resulted in the spectra shown in fig. 3Aa and 3Ba respectively. Silicon tetrachloride reduced the intensity of bands at 2540, 2695 and 2750 cm-l, caused the appearance of a new band at 2650 cm-l and removed a band at 2725 cm-l (fig. 3B) or a corre- sponding shoulder at 2715 cm-1 (fig. 3A). Equilibration of the surface initially outgassed at 50°C with saturated deuterium oxide vapour and subsequent evacuation resulted in intense bands at 2540 and 2750 cm-I together with a shoulder at 2650 cm-l20 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES (fig.3Ac and d). Analogous results from a surface initially evacuated at 250°C showed an additional shoulder at 2695 cm-l while that at 2650 cm-l was not as marked (fig. 3Bc and d). 2900 2700 2500 2900 2700 2500 wavenumber/cm- FIG. 3.-The spectra of deuterated rutile surfaces. A: (a) pumped free of excess DzO vapour (6$ h, 50°C) ; (b) in contact with SiC14 (3 h, 45"C, 4.7 kN m-2) and evacuated (14 h, 45°C) ; (c) in contact with saturated D,O vapour (23 h, 45°C) and evacuated (1 min, 45°C) ; ( d ) further evacuation (3 h, 45°C). B : (a) pumped free of excess DzO vapour (16 h, 250°C) ; (b) in contact with SiCI4 (3 h, 45"C, 4.7 kN m-2) and evacuated (14 h, 45°C) ; (c) in contact with saturated DzO vapour (23 h, 45°C) and evacuated (1 min, 45°C) ; (d) further evacuation (% h, 45°C).Experiments were carried out with a reduced rutile surface to test whether com- parison with the results for a surface heated in oxygen gave any evidence for the participation of oxide sites in the adsorption process. Rutile was reduced by heating it in hydrogen (+ h, 400"C, 8 kN m-2) and then evacuating (I h, 400°C) before treating with silicon tetrachloride. Reduction did not change the intensity of the infra-red bands arising from hydroxyl stretching vibrations and treatment with silicon tetrachloride (4 h, 45"C, 6.7 kN m-2) left only a broad band centred at 3580 cm-l (cf. fig. lb). Contact with saturated water vapour (9 h, 45°C) followed by evacuation (<+ h, 45°C) produced a band at 1605 cm-l with a shoulder at 1625 cm-l.Further evacuation (>4 h, 45°C) left only the band at 1605 cm-l and an absorption maximum centred on 3400 cm-l while pumping for 1 h at 400°C led to a surface for which only an absorption band at 3660 cm-l remained In order to study the reaction of silicon tetrachloride with rutile at elevated temperatures, a starting surface was evacuated (13 h, 250°C) and gave rise to infra- red bands at 3660 and 3700 cm-l with a broad shoulder at ca. 3400 cm-1 which were removed on treating with excess silicon tetrachloride (1 h, 25"C, 2 kN m-2). Subsequent equilibration with excess water vapour (2 h, 45°C) and evacuation pro- duced an absorption band at 1615 cm-l due to molecular water together with a band at 3730 cm-l and a broad absorption maximum centred on ca. 3400 cm-l.Evacua- tion (2 h, 250°C) left only an intense band at 3740 cm-l together with a weaker oneG. D. PARPITT, J . RAMSBOTHAM AND C. H . ROCHESTER 21 at 3670 crn-l. The interaction of silicon tetrachloride vapour with a rutile surface exchanged with deuterium oxide vapour gave similar results. After exchange with deuterium oxide vapour the rutile was evacuated and gave bands at 2540, 2600, 2695 and 2750 cm-l (fig. 4a). A residual band at 3660 cm-l was due to incomplete exchange of OH group^.^ Treatment with excess silicon tetrachloride, cooling and 2700 2500 2300 1600 1500 wavenumber /cm- FIG. 4.-Reaction of SiC14 with deuterated rutile at 250°C. (a) Rutile pumped free of saturated DzO vapour (JJ h, 50°C) ; (6) in contact with SiC14 (3 h, 250°C. 4.7 kN m-2) ; cooled to 45" and evacu- ated (I+ h) ; (c) in contact with saturated H20 vapour (3 h, 45°C) and evacuated at 250°C (10 h) then 400°C (3 h) ; ( d ) in contact with NH3 (3 h, 45"C, 0.67 kN m--2) and evacuated (5 min).evacuation left small bands at 2540 and 2750 cm-' (fig. 4b) and these decreased in intensity but were not removed after contact with saturated water vapour. Evacua- tion (-$min, 45°C) left bands at 1615, 2760 and 3750cm-l with a broad band at ca. 3450 cm-l and pumping at 400°C left bands at 2760 and 3750 cm-l with a shoulder at ca. 3400 cm-I (fig. 4c). The acidity of this surface was investigated by ammonia and pyridine adsorption. Ammonia vapour rapidly interacted with the species resulting in the band at 276Ocm-l and produced bands at 3360 and 1605cm-l (fig.4d). The band at 3750 cm-1 due to SiOH groups was hardly affected by ammonia treatment but a band due to the ammonium ion generated on surface Brarnsted sites appeared at 1420 cm-l. Evacuation (lh, 400°C) left a surface similar to that before ammonia adsorption except that the intensity of the 2760 cm-l peak was reduced. Hydrogen chloride vapour did not produce any further change in intensity of the latter band but gave rise to a shoulder at 3400 crn-l. The adsorption of pyridine vapour onto a similar surface to that treated with ammonia (fig. 5a) greatly reduced the intensity of the peak at 3750 cm-1 giving a broad band at ca. 3400 cm-l together with bands at 1450, 1490, 1575, 1610, 1640 and 1540cm-l (fig. 5b and c). The latter two bands arise from the pyridinium ion and confirm the presence of Brnrnsted sites on the modified surface.Evacuation followed by contact with water vapour (fig. 5 4 then further evacuation (fig. 5e) increased the intensity of the bands at 1540 and 1640 cm-l while decreasing that of the other bands in this region.22 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES The interaction of tin(1V) tetrachloride with rutile was investigated to determine if it was comparable to that of silicon tetrachloride and produced surfaces which exhibited Brarnsted acidity in the presence of adsorbed ammonia or pyridine. The spectrum of rutile after evacuation (17* h, 400°C) showed a band at 3700 cm-l with weak shoulders at 3660 and ca. 3500 cm-l. Reaction with tin(1V) tetrachlorjde wavenumber/cm- FIG.5.-Surface acidity of rutile treated with SiC14 at 250°C. (a) Rutile surface with a pretreatment similar to that described in fig. 4a-c followed by evacuation ($ h, 400°C) ; (b) in contact with pyridine (0.67 kN m-2, 45°C) ; (c) in contact with pyridine (3 h, 45"C, 8 kN m-2) and evacuated (2 h, 45°C) ; (d) in contact with H20 vapour (45"C, 1.33 kN m-2) ; (e) evacuated (3 h, 45°C). (% h, 45"C, 2.7 kN m-2) followed by evacuation (3 h, 45°C) removed the band at 3700cm-l and resulted in a peak of slightly lower intensity at 3650cm-l with a more intense shoulder than before now centred on 342Ocm-l. Contact with saturated water vapour (13 h, 45°C) and subsequent pumping (+ h, 45°C) resulted in bands at 3350 and 1610cm-l which resolved into a band at 3360cm-l with a shoulder at 358Ocm-l and a broad absorption between 1550 and 1610cm-l.Evacuation (1 h, 400°C) left a peak at 3660 cm-l with a weak shoulder at 3500 cm-l. Contact of tin(1v) tetrachloride (+ h, 45"C, 1.3 kN m-2) with a rutile surface after equilibration with saturated water vapour and evacuation (5 s, 45°C) shifted a band originally at 1625 cm-l to two centred at 1565 and 1600 cm-l. The reaction of excess tin(1V) tetrachloride with a rutile surface followed by evacuation equilibra- tion with saturated water vapour, oxygen treatment and evacuation gave a surface which exhibited Brarnsted and Lewis acidity and hydrogen bonding interactions with pyridine vapour (fig. 6b). Bands characteristic of Brarnsted acidity at 1540 and 1640 cm-l were enhanced by water vapour which removed the bands characteristic of hydrogen bonding at 1440 and 1607 cm-l (cf.fig. 6b and c). A band at 1450 cm-l was removed by further evacuation to leave bands at 1455,1495,1575 and 1618 cm-l, due to pyridine on Lewis acid sites and bands at 1540 and 1640 cm-l due to the pyridinium ion (fig. 64. The interaction of titanium tetrachloride with rutile was studied to test whether the modifications to the rutile surface produced by t i n o tetrachloride and silicon tetrachloride also arise when the metal in the adsorbate is the same as that in theG . D. PARFITT, J . RAMSBOTHAM AND C. H. ROCHESTER 23 50 1; 'Jb - - 'I I 1 I I I ao 50 40 30 u 3 9 0 0 3700 3 5 0 0 33OC I I I I I : *. - I I I I ~ 1650 1550 1450 wavenumber /cm- FIG. 7.-The reaction of TiCI4 with silica.(a) SO2 evacuated (13 h, 400°C) ; (b) in contact with saturated Tic14 (5 h, 250°C) ; (c) in contact with excess TiC14 at 400°C ( >3 h), evacuated (+ h, 45"C), hydrolyzed with saturated water vapour (3; h, 45"C), and evacuated (t h, 45°C) ; this treatment repeated five times followed by evacuation (+ h, 250°C) ; (d) in contact with saturated pyridine vapour (+ h, 45°C) and evacuated (+ h, 45°C) ; (e) equilibrated with excess water vapour and evacuated (& min, 45°C).24 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES adsorbing oxide. After equilibration of a rutile surface with saturated water vapour and evacuation (1 h, 2OO0C), a sharp peak at 3665 cm-1 with a shoulder at ca. 3450 cm-1 remained. Adsorption of excess titanium tetrachloride vapour (4 h, 45"C, 0.5 kN m-2) followed by evacuation (3 h, 45°C) and further equilibration with excess water vapour and pumping (23 h, 45°C) gave bands at 1565, 1610 and 3400 cm-' with a shoulder at 3660 cm-l.In the presence of pyridine vapour, this surface did not give bands at 1540 and 1640 cm-l arising from Bronsted acidity. A silica (Degussa aerosil) disc was evacuated and gave infra-red bands at 1635 and 3750 cm-' (fig. 7a). Reaction with excess titanium tetrachloride did not change the band at 1635 but reduced that at 3750cm-l to a broader band at 3680cm-l (fig. 7b). Subsequent equilibration with saturated water vapour (23 h, 45°C) and evacuation (4 h, 400°C) produced bands at 3680 and 3750 cm-l. Fig. 7c shows the result of five similar consecutive treatments with titanium tetrachloride followed by hydrolysis and evacuation.Contact with pyridine vapour and evacuation left bands characteristic of Lewis acidity at 1445, 1490, 1575 and 1610 cm-l and weaker ones, characteristic of Brarnsted acidity, at 1540 and 1640 cm-l (fig. 7d). The addition of water enhanced the Bronsted acidity of the surface (fig. 7e). Excess pyridine vapour was completely removed by evacuation (< 3 h, 45°C) from a silica disc which had not been pretreated with titanium tetrachloride. No infra-red bands characteristic of pyridine adsorbed on Lewis or Brarnsted acid sites were observed. Pyridine is adsorbed on silica by hydrogen-bonding with the surface hydroxyl groups.16 DISCUSSION The bands arising from hydroxyl groups and water at 3700, 3660, 3620, 3520, 3410 and ca.1610cm-l and their deuterium analogues at 2730, 2700, 2670, 2620, 2540 and ca. 1185 cm-1 have been discussed elsehwere.1° The small shoulder observed in some spectra at 3730 cm-l (2750 cm-I on deuterated surfaces) results from trace amounts of silica in the samples. The reaction of silicon tetrachloride with hydroxyl groups on rutile (fig. l a and b) is consistent with the formation of surface TiOSiC13 species and the elimination of hydrogen chloride. Individual silicon tetrachloride molecules may also react with two hydroxyl groups 1 * although the relative proportions reacting with one or two hydroxyl groups cannot be deduced from the present results. A band observed at 3580 cm-1 after adsorption arises from weak hydrogen bonding of the new surface species with unreacted hydroxyl groups.Molecular water is displaced from the surface but does not cause hydrolysis of Sic1 groups to SiOH as shown by the absence of an absorption band at ca. 374Ocm-l. However, the addition of water vapour to the treated surface causes hydrolysis of unreacted Sic1 bonds to SiOH groups for which the characteristic absorption band appeared at a position in the range 3730-3750 cm-l depending upon the precise experimental conditions. Hydrogen chloride produced by the hydrolysis leads to the appearance of infra-red bands at 3400, 1605, 1580, and 1565 cm-l. The 3400, 1605 and 1565 cm-l frequencies are similar to those of 3360, 1600 and 1565 cm-l for the corresponding bands which appear when hydrogen chloride alone is adsorbed on r ~ t i l e .~ The tentative assign- ment of these bands to molecular water present on the surface either as hydronium ions or coordinately bound to Ti4+ ions is discussed el~ewhere.~ The band at 1580 cm-l which decreases in intensity as water is evacuated from the rutile surface and whose lifetime is a function of the pretreatment evacuation temperature of the starting surface is only observed after adsorption of silicon tetrachloride followed by hydrolysis. It did not appear after analogous treatments with hydrogen chloride, titanium tetra-G. D . PARFITT, J . RAMSBOTHAM AND C. H . ROCHESTER 25 chloride or tin tetrachloride. The decrease in intensity of the 1580 cm-l band on evacuation is accompanied by a corresponding increase in the intensity of the 1565 cm-I band (fig. 2) suggesting that the species responsible for the former is a precurser of that responsible for the latter.The following formulation is consistent with the previous assignment of the 1565 em-I band.g (1580 cm') (1565 cm') The 1580 cm-I band is ascribed to the bending vibration of molecular water associated via hydrogen-bonding with a hydrogen chloride molecule on the surface. Evacuation causes loss of water and also results in the growth (fig. Ic and d ) of the 3730 cm-l absorption band of isolated SiOH groups. The interaction of silicon tetrachloride with a deuterated rutile surface and sub- sequent treatment with deuterium oxide gives results which are in accord with those for the hydroxylated surface although the removal of deuteroxyl groups is apparently more complete than that of hydroxyl groups.This arises because there is a broad absorption maximum at ca. 3660cm-l due to hydroxyl groups which cannot react with silicon tetrachloride and which may be attributed either to bulk hydroxyl groups or to interparticulate hydrogen-bonding. Thus exchange with deuterium oxide (fig. 3a) left a broad absorption band at 3660cm-l which was unaffected by the adsorption of silicon tetrachloride. The removal of the deuteroxyl groups which give absorption bands at 2695 and ca. 2720cm-l is caused in part by chemical reaction and also by hydrogen-bonding (which leads to the appearance of a band at 2650 cm-l) of the TiOSiC13 or (TiO)2SiC12 surface species with unreacted groups. Treatment with deuterium oxide causes deuterolysis of Sic1 groups to SiOD and the formation of adsorbed deuterium chloride which on evacuation leads to an intense band at 2540 cm-l (cf.3400 cm-1 for an hydroxylated surface). No infra-red bands appear in the range 1550-1620 cm-l. The deuterium analogues of the species responsible for these bands should give corresponding bands in the range 1140- 1180 cm-l but these cannot be reliably studied because of strong background absorp- tion by the bulk rutile. Heat treatment of rutile in hydrogen results in a reduction in the number of oxide ions in the surface but leaves the hydroxyl population ~nchanged.~ Hydrogen chloride therefore interacts with reduced rutile predominantly by reaction with H / \ hydroxyl ions to give the species Ti4+0 C1- (infra-red band at 1565 cm-l) rather H than by reaction with oxide ions which would give hydroxide and chloride ions on adjacent titanium ion sites.g The reaction of reduced rutile, evacuated at 400"C, with silicon tetrachloride followed by hydrolysis contrasts with the corresponding reaction for an oxygen-treated rutile in that neither the 1565 nor the 3730 cm-l band are observed for the former.The difference is inexplicable in terms of reaction of silicon tetrachloride with surface hydroxyl groups because the reduced and oxidized surfaces have similar low hydroxyl populations. This suggests that an oxidized surface which has been evacuated at 400°C before the addition of silicon tetrachloride reacts26 REACTIONS OF RUTILE WITH VARIOUS TETRACHLORIDES predominantly, even at 45"C, by the nucleophilic attack of silicon atoms by surface oxide ions.3 Reaction (2) -0SiCl3 I c1- Ti4+02-Ti4+ + SiC14+Ti4+ Ti4+ (2) is probably occurring.This reaction does not occur on the reduced surface because the appropriate oxide sites are removed by heating rutile in hydrogen. The addition of water vapour to a silicon tetrachloride treated reduced rutile causes rehydroxylation of the treated surface and the reappearance of an absorption band at 3660 cm-l. No band at 3700 cm-l appears because the appropriate hydroxyl groups are not removed by the pretreatment (400°C evacuation lo) and have therefore all reacted with silicon tetrachloride. Rutile, in common with anatase l2 and alumina,ll displays Lewis acidity in the presence of adsorbed ammonia6 or pyridine.' The addition of water does not promote Brarnsted acidity.Brnrnsted acidity is observed for silica-aluminas and zeolites and is enhanced by the adsorption of water. At 250°C silicon tetrachloride undergoes reaction with all the hydroxyl groups on rutile. Extensive reaction with oxide ion sites probably also occurs. Hydrolysis and evacuation lead to an infra-red spectrum which is more characteristic of silica than of rutile. Subsequent adsorption of ammonia gave bands at 1605 and 3360cm-l due to ammonia coordinated to titanium ions and a broad shoulder at 1420 cm-1 due to ammonium ions formed by adsorption of ammonia on Brnrnsted acid sites. Alternatively, the adsorption of pyridine gave bands characteristic of both the pyridinium ion (1 540 and 1640 cm-l) and coordinated complexes (fig.5c). The marked changes in intensity of the band at 3750 crn-l suggest that the pyridine molecules may also coordinate with the rutile surface while perturbing adjacent Si-OH groups. The adsorption of water on the surface enhances Brnrnsted acidity at the expense of Lewis acidity (cf. fig. 5c and e) suggesting that water interacts with Lewis sites formerly occupied by pyridine mole- cules. Similar Brarnsted activity was exhibited by rutile surfaces treated with tin(1V) tetrachloride and water but not by surfaces treated with titanium tetrachloride and water. These results indicate that M-OH groups on rutile where M is not Ti show Brarnsted acidity towards ammonia and pyridine vapour and behave like surfaces in mixed oxide systems.l Similar reactions of titanium tetrachloride and water on a silica aerosil produced four infra-red bands in the range 1400-1700 cm-I, which are characteristic of pyridine coordinated to the surface, together with two bands at 1540 and 1640 cm-l which arise from Brarnsted acidity and are enhanced by adsorbed water vapour (fig.7). These data indicate that surface hydroxyl groups on heterometal atoms exhibit Brarnsted acidity and also the Lewis acidity of the surface is modified. Electrophoresis studies indicate that the presence of Si-OH groups on rutile lower the isoelectric point towards that of ~i1ica.l~ However, electrophoresis studies by Parfitt and Ramsbotham l4 using rutile coated with varying proportions of silica and alumina indicated that although hydroxyl groups on surfaces containing a high proportion of silica appear to be similar to those on a pure silica surface, those on surfaces of high alumina content are not similar to those on alumina. Infra-red data by Smith l5 for coatings on rutile pigments suggest that bands arising from hydroxyl groups in the coatings " tended to lose their individual identity ". These latter points suggest that hydroxyl groups on metals other than titanium on rutile may not necessarily be characteristic of their parent oxide, with the possible exception of Si-OH. The reaction of t i n 0 tetrachloride and water with rutile differs from the corre- sponding reaction of silicon tetrachloride in that the 1580 cm-' band never appearsG . D . PARFITT, J . RAMSROTHAM AND C . H . ROCHESTER 27 and the 1565 crn-l band is less intense. The addition of pyridine to the rutile surface coated with SnOH groups shows (fig. 6) infra-red bands arising from hydrogen-bonded pyridine (1443 and 1605 crn-l), pyridine coordinated to Lewis acid sites (1455 and 1617 cm-I) and the pyridinium ion (1450, 1540 and 1640cm-I). The bands at 1495 and 1575 cm-l are common to all types of pyridine on the surface.16 Hydrogen- bonded pyridine is displaced by the adsorption of water which enhances the formation of the pyridinium ion. The authors thank British Titan Products Co. Ltd. for a fellowship (J. R.) and for rutile samples. ' M. L. Hair and W. Hertl, J. Phys. Chem., 1969,73, 2372. ' C. G. Arrnistead, A. J. Tyler, F. H. Hambleton, S. A. Mitchell and J. A. Hockey, J . Phys. Chern., 1969,73, 3947. J. A. Hockey, J. Phys. Chem., 1970,74, 2570. J. B. Peri, J. Phys. Chem., 1966, 70, 2937. A. A. Isirikyan, A. V. Kiselev and E. V. Ushakova, KolZoidZhur., 1964,26,45 ; A. A. Isirikyan, A. V. Kiselev and E. V. Ushakova, KolZoidZhur., 1965,27, 690. G. D. Parlitt, J. Ramsbotham and C. H. Rochester, Trans. Faraduy Soc., 1971,67, 841. ' G. D. Parfitt, J. Ramsbotham and C . H. Rochester, Tram. Faruday Soc., 1971,67, 1500. * A. Buckland, J. Ramsbotham, C. H. Rochester and M. S . Scurrell, J. Sci. Imtr., 1971,4,146. G. D. Parfitt, J. Ramsbotham and C. H. Rochester, Tram. Faraday SOC., 1971, 67, 3100. lo P. Jackson and G. D. Parfitt, Trans. Faraday Soc., 1971,67,2469. 11 L. H. Little, Infrared Spectra of Adsorbed Species, (Academic Press, London, 1966). l3 G. D. Parfitt, J. Ramsbotham and C. H. Rochester to be published. l4 G. D. Parfitt and J. Ramsbotham, J.O.C.C.A., 1971,54, 356. Is I. T. Smith, Nature, 1964, 201, 67. l 6 E. P. Parry, J. Catalysis, 1963,2, 371. N. D. Parkyns, Symp. Chemisorption and Catalysis, (Institute of Petroleum, London, 1970).

ISSN:0300-9599    

DOI:10.1039/F19726800017

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Protonation of Aromatic Carboxylic Acids in the First Excited Singlet State BY A. R. WATKINS” Department of Chemistry, University of Sheffield, Sheffield Received 1st July, 1971 Previous methods for determining the excited state acid-base characteristics of molecules from kinetic data are applicable only to molecules which undergo dissociation upon excitation. This paper presents a method of obtaining kinetic data from fluorescence measurements for compounds, which, upon excitation, take on a proton. The kinetic data so obtained can be used to calculate the excited state acidity constants of the protonated molecules. Comparison, for two naphthoic acids, of the results so obtained with data from the spectral method due to Forster and with data from the pH at which the fluorescence spectrum undergoes a transformation reveals considerable discrepancies : these are discussed briefly.Weber was the first to discover that, with change in pH, some organic molecules undergo a change in wavelength of fluorescence emission ; this was later interpreted by Forster as being due to a change in acid-base properties of the excited species. Since then a considerable literature dealing with acid-base properties of excited states has accumulated. 3-5 There are essentially three methods of gaining information about the acid-base properties of excited molecules. The first is the direct observation of the pH at which the fluorescence of the excited molecule is replaced by the fluorescence of its conjugate ; under certain conditions this “transformation” pH, referred to later in this paper as pHtr, gives a measure of the pK, or pK, of the excited The second method, proposed by Forster,6 relates the 0-0 transitions of the molecule and its conjugate acid or base to the equilibrium constants of the proton transfer reaction in the ground and excited states; a knowledge of the ground state equilibrium con- stant together with the relevant spectral data allows calculation of the excited state equilibrium constant.The third method, developed by Weller,’ consists of measuring the forward and reverse rate constants for the proton transfer reaction in the excited state, from which the equilibrium constant can be calculated. and generally good agreement is found. We were interested in extending this comparison to weak bases which become more basic on excitation, for this purpose I-naphthoic acid and 2- naphthoic acid B* + Hf + B*H+ were chosen as the probe compounds.The three methods have been compared by several authors,3* EXPERIMENTAL The compounds studied were obtained commercially and were purified by repeated crystal- lization ; in all cases the observed melting point agreed with the literature value. Fluoresc- ence intensities were measured with an Aminco-Bowman spectrofluorometer, variations in * present address : Max-Planck-Institut fur biophysikalische Chemie, D 34 Gottingen-Nikolaus- burg, West Germany. 28A . R. WATKINS 29 lamp intensity being taken into account by the use of a standard solution M quinine sulphate in water or a solution of the compound at a fixed pH) whose fluorescence quantum yield remained constant during the course of the experiment.All experiments were carried out at 25 fO.l”C in aqueous solution ; the pH was varied by addition of known amounts of hydrochloric acid. The concentration of the naphthoic acids used in these experiments was The effect of overlap between the emission spectra of the fluorescing base and its conjugate 2~ 10-5 M. acid was corrected for by use of the equations I’ - k l q’= ~ I - k‘I’ y=- I-kk” 1 - kk’ ’ where 1 and I’ are the measured intensities at the emission wavelengths of the protonated and unprotonated species, k and k’ represent the proportion of overlap from the protonated and unprotonated emitting species at the wavelengths used, and y and q’ represent the correspond- ing true intensities (throughout this paper primed quantities refer to the free base, unprimed quantities to the protonated base).The wavelengths at which the emission intensities were measured are given in table 1. TABLE 1 .-EXCITATION AND EMISSION WAVELENGTHS USED IN THE FLUORESCENCE MEASUREMENTS excitatian emission wavelength wavelength nm compound nm B* B*H+ 1-naphthoic acid 295 405 463 2-naphthoic acid 270 380 454 The relative quantum yields of fluorescence from the excited states of the base and its protonated conjugate are shown in fig. 1 and 2 (q/qo measures the relative quantum yield of fluorescence for the protonated excited state, y’/q& the relative quantum yield for the un- protonated excited state) as a function of the acidity of the solution for both naphthoic acids.--logKl [HCll FIG. 1.-Relative fluorescence quantum yields of 1-naphthoic acid as a function of acidity. The open circles refer to the unprotonated excited species (q’/q&), the closed circles to the protonated excited species (q/qo). In all measurements it was important to ensure that the compound was present in the ground state as the undissociated molecule, and that formation of the anion or the protonated carboxylic acid in the ground state, or of the anion in the excited state, in the pH range used was negligible. For this reason the maximum pH used was, in every case, smaller than the30 pH at which the extent of dissociation into the anion in the ground state is 3 %. Since the protonation of aromatic carboxylic acids generally occurs at quite high acidities (correspond- ing to Hammett acidity function values l2 of about -6), it is clear that significant proton- ation of the carboxylic acids in the ground state in these experiments could not occur.Dissociation of the excited naphthoic acids into the carboxylate anions also lies well outside this ~H-range.~ PROTONATION OF AROMATIC CARBOXYLIC ACIDS -0 F 2 F 0 E F U -0.5 0 a5 1.0 1.5 2D --log,, W l l FIG. 2.-Relative fluorescence quantum yields for 2-naphthoic acid as a function of acidity. The open circles refer to the unprotonated excited species (q’/qA), the closed circles to the protonated excited species (q/qo). Fluorescence lifetimes were measured with an apparatus consisting of a repeating spark light source, the light from which intermittently excited the fluorescent sample. The emitted light was Gollected by a 54AVP fast-response photomultiplier, the output of which was fed to a Hewlett-Packard 185 sampling oscilloscope. In all cases the resulting decay curves fitted an exponential decay law.All lifetime measurements were made at 25f0.1°C, and TABLE 2.-LIFETIMES OF PROTONATED AND UNPROTONATED EXCITED STATES. tlns 7‘lns compound (protonated base) (unprotonated base) 1-naphthoic acid 4.5 f1.5 5.2 ~t0.4 2-naphthoic acid 7.1 &0.7 15.2 h0.2 were corrected, where necessary, for the response time of the instrument. To avoid the possibility of dissociation in the ground state, measurements of the fluorescence lifetimes of the unprotonated carboxylic acids were made in benzene. The relation TbenzenehHzO = qbenzene/rHzO (2) allows the corresponding lifetimes in aqueous solution to be calculated from these measure- ments.The fluorescence lifetimes of the protonated excited molecules were obtained from measurements at high acid concentrations where the proton transfer reaction ceases to have any influence on the observed lifetimeY8 and were converted to the true lifetimes in water by means of an equation similar to (2). The method by which ~ H ~ O (qo in the discussion below) is obtained is described in the following section. Table 2 gives the life-times so obtained, together with an estimate of the error due to the response time of the instrument ; z refers to the protonated excited state, z’ to the unprotonated excited state.A . R. WATKINS 31 THEORY The equations relating quantum yield of fluorescence to pH for the protonation of a weak base in the excited state have been derived by Kokubun in his analysis of the protonation of acridone ; although the quantity (1 + klz')/k2z can be derived, where kl and k2 are the forward and back rate constants for protonation in the excited state, he found no way of obtaining these rate constants separately.The present analysis, on the other hand, allows kl, kz, and nQ, the quenching constant for the quenching of the protonated excited molecules by the added hydrochloric acid, to be calculated. Addition of acid to a solution of a base can produce two effects : the protonation of the base, either in the excited state or in the ground state or both, and quenching of the fluorescences of the emitting species.The systems studied here consist of carboxylic acids which, as discussed above, remain substantially undissociated and unprotonated in the ground state in the range of acidity used here. Proton transfer occurs only in the excited state, according to the following scheme, in which the mol- ecule B is excited by absorption of light, and then undergoes the reactions shown : O+hvJ 0 B B H + + ~ V BH+ BH+ where n; and ne are rate constants for emission of fluorescence, n;l and nL describe radiationless decay of the excited state, n;2 and nQ refer to the quenching action of the added acid, and kl and k2 are the forward and reverse rate constants for the protona- tion reaction. The relative quantum yields of fluorescence from the protonated and unprotonated excited states are given by the following equations : which can be obtained by application of steady state kinetics to the reaction scheme given above.7 From these equations the following expression for the ratio of relative quantum yields can be derived : Thus, when the ratio of the experimentally determined quantum yields is plotted against the reciprocal of the acid concentration, the points are expected to fall on a straight line of slope (1 + k 2 ~ ) / k l ~ ' and intercept nQz/klz'.At low acid concentrations it is found experimentally that a plot of y against y ' is linear, i.e., the sum of the relative quantum yields of fluorescence of B* and B*H+ becomes independent of acid concentration. This may be written, from eqn (4) and (5), as = const.(7)32 PROTONATION OF AROMATIC CARBOXYLIC ACIDS This will only be the case when the term containing [Hell2 in the denominator be comes negligible, and when klrf 4- npT > nQTf + n;2zfk2T (8) in which case the sum of the quantum yields tends to unity. Proceeding from this inequality and eqn (4), the following relation may be derived : and, with the value of (1 + k27)/k17’ previously obtained, the left-hand side of eqn (9) may be plotted against acid concentration to give a straight line of slope YtQz. It is clear that, with the values of nQz, nQz/kl.c’ and (1 +k2z)/klz’ obtained as shown above, together with the measured lifetimes, the rate constants k l , k2 and nQ can be calculated from the experimental measurements. It should be mentioned that ylo and qb, the fluorescence intensities o f the proton- ated and unprotonated carboxylic acids in the absence of any proton transfer reaction, could not be measured directly, since the dissociation of the carboxylic acid in the ground state becomes important at high pH values, and in strongly acid solutions the fluorescence of the protonated carboxylic acid is heavily quenched.Instead the equation which follows from eqn (7) and the discussion above, and which is expected to be valid in the low acidity region where no fluorescence quenching by the added acid occurs, can be applied. DISCUSSION Fig. 3 and 4 show the application of eqn (6), and fig. 5 and 6 the application of eqn (9), to the data for both naphthoic acids. Table 3 lists the quantities derived from these graphs, and table 4 gives the values of the rate constants derived from table 3.-0 20 LO 60 80 too 120 IbO 1 / [ H a eqn (6)). FIG. 3.-Plot of q‘q~/q-& against the reciprocal of the acid concentration for 1-naphthoic acid (seeA . R. WATKINS 33 1/[Ha eqn (6)). FIG. 4.-Hot of q'q,Jq~& against the reciprocal of the acid concentration for 2-naphthoic acid (see [Ha FIG. 5.-Application of eqn (9) to the experimental data for 1-naphthoic acid. 234 PROTONATION OF AROMATIC CARBOXYLIC ACIDS The last column of table 4 gives the calculated diffusion-controlled rate constants, assuming an average distance of reaction of 7 A.1° The discrepancy between the experimental and calculated values of kl for I-naphthoic acid can be explained by the steric requirements of the proton transfer step of eqn (3).For this reason the value [HClI FIG. 6.-Application of eqn (9) to the experimental data for 2-naphthoic acid. of k2 enclosed in brackets for 2-naphthoic acid, where nQz/klz' is too small to be measured, was calculated from the experimentally determined quantities in conjunc- tion with a value of kl of 2 x 1 O 1 O 1. mol-1 s-l. TABLE 3.-PARAMETERS DERIVED FROM THE APPLICATION OF EQN (6) AND (9) TO THE EXPERI- MENTAL MEASUREMENTS 1 +k2r nQt nQr kid klr' - compound 1-naphthoic acid 0.0301 0.104 9.54 2-naphthoic acid 0.0972 0 0.352 In view of the mathematical complexity involved, no attempt has been made to incorporate a correction for the possible presence of non-steady-state conditions at high acid concentration^.^ The small value of the intercept in fig.3, together with possible effects due to the changing medium on the kinetics in the measurement of Y t Q q will mean that, even in the absence of non-steady-state effects, the derived rate constants will be subject to some uncertainty. TABLE 4.-RATES OF PROTONATION (ki), DEPROTONATION OF THE EXCITED NAPHTHOIC ACIDS (k2), TOGETHER WITH QUENCHING CONSTANTS ( n ~ ) FOR THE EXCITED PROTONATED ACIDS ki k2 nQ ki(alc) compound I. mol-1 s-1 S-1 1. mol-1 s-1 1. mol-1 s-1 1-naphthoic acid 1 . 7 ~ 10'' 1 . 7 ~ 10' 2.1 x 109 4.7x 1 O 1 O 2-nap ht ho i c acid - ( 4 . 0 ~ 109) 2.3 x 107 4.7x loio In table 5, values of the ground state and excited state acidity constants of the protonated carboxylic acids are presented, the latter calculated from the kinetic data of table 4.For comparison, excited state acidity constants calculated from spectralA . R. WATKINS 35 data have been included together with the acidity constants of the protonated car- boxylic acids in the ground state." The discrepancies between the acidity constants derived from kinetic and spectral data in table 5 cannot be accounted for by the uncertainties in the measurements of the rate constants. The difference of 2.2 units for 1-naphthoic acid, for example, TABLE 5.-ACIDITY CONSTANTS OF THE PROTONATED NAPHTHOIC ACIDS IN THE EXCiTED AND GROUND STATES (pKZ AND pK,, RESPECTIVELY) compound pK: (kinetic) pKX (spectral) PKa 1-naphthoic acid 2.0 - 0.20 - 7.70 2-naphthoic acid 0.70 - 0.35 - 7.68 a, this work; b, ref.(11). would mean, in the event that the pKif' from spectral data was correct, that the ratio of the rate constants k2/kl differed by a factor of 100 from the true value, which is too large to be accounted for by the experimental uncertainty (maximally+ 50 %) in the rate constants. At least part of the discrepancy must be attributed to the fact that the ground state pK, values,ll from which the values of pK,* (spectral) are derived, are based on Hammett acidity functions whose applicability to measurements of equilibria in concentrated acid solutions has been opened to doubt by a number of a ~ t h 0 r s . l ~ This is best illustrated by a comparison of acidity scales (Ho scales) measured with different types of bases. For example, the value of the acidity function Ho obtained for 80 % H2S04 (the acidity region in which protonation of carboxylic acids occurs) is - 7.34 based on primary aromatic amines, - 9.44 (tertiary aromatic amines), - 1 1.84 (triarylcarbinols) and - 6.35 (benzophenone~).~~ No acidity scale based on aromatic carboxylic acids has yet been reported, but, in view of the variations found above, it is highly likely that such a scale would behave differently to the traditional scale based on primary aromatic amines (used in evaluating the pK, values quoted in table 5).The pH at which change-over of the fluorescence spectrum from one component of the acid base equilibrium to the other (pH,,) occurs is sometimes taken as being equal to pK: ; however, this has only limited validity. It is clear that, depending on the values of t?Q and nh, this pH will be determined not only by the kinetic constants of the proton transfer reaction, but will also be determined by competing quenching reactions occurring between the added acid and the fluorescing species.From the analysis given above, the pH at which the quantum yields of fluorescence of proton- ated and unprotonated emitting species are equal is given by : This corresponds to the pH for fluorescence transformation measured by Urban and Weller,14 and to the pKgH+ values determined by Hopkinson and Wyatt? The values of these quantities determined by these authors for 1-napthoic acid and 2- naphthoic acid, together with the pH,, values obtained in the present work, are listed in table 6. The measured pH,, will only give an accurate indication of pK: when nQ7 e klzf and k27 9 1, (12) i.e., when the added mineral acid is an inefficient quencher of B*H+ and when the dissociation reaction is fast in comparison with the other processes leading to the disappearance of B*H+.36 PROTONATION OF AROMATIC CARBOXYLIC ACIDS The equality of pH,, with pK,* depends on the order of magnitude of K:( = k2/k1). If k , is taken to be diffusion controlled (2: loio 1.mol-1 s-l) and if the lifetimes z' and z are approximately equal, then the range of possible values which pK: can take can be divided into two regions of interest : (i) pK: <2, k2/kl > s, then k2z> 1 ; however, pH,, = pK: only when nQz<k,z'. (ii) pKZ >2, k2/kl < In this case PHtr will be independent of pK:, and will be given (neglecting the second term in eqn (1 1)) by Here, in contrast to case (i), the rate of protonation is too slow to compete with the other processes that deactivate B*, so that, as the pH is increased, the fluorescence of B*H+ disappears before it is expected to.If z is taken as By the same reasoning k2z < 1. pH,, N log,,klz' N 2. (13) TABLE 6a-COMPARISON OF THE pH OF FLUORESCENCE TRANSFORMATION PHtr FROM VARIOUS SOURCES compound PHtr" PHtrb PHtrC PKt" nQ 1 -naphthoic acid 1.5 0 1.48 2.0 2 . 1 ~ 109 2-naphthoic acid 1.5 0 1.04 0.70 2 . 3 ~ lo7 Q, ref. (14); b, ref. (11); c, this work. From table 3, it can be seen that both naphthoic acids belong to category (i) above, and that the conditions (12) will be fulfilled by 2-naphthoic acid but only partially by 1-naphthoic acid ; the measure of agreement between columns 4 and 5 of table 6 is reasonably consistent with this analysis.The disagreement between the PHtr values of Weller and Urban,14 and Hopkinson and Wyatt,ll may be a consequence of the use of different mineral acids to fix the pH of the solutions in which the experiments were carried out. Thanks are due to the Science Research Council for a fellowship, during the tenure of which this work was carried out, and to Prof. P. A. H. Wyatt for helpful discussions. l K. Weber, 2. phys. Chem. B., 1931,15, 18. Th. Forster, Naturwiss., 1949, 36, 186. A. Weller, Progress in Reaction Kinetics, 1961, 1, 189. B. L. Van Duuren, Chem. Rev., 1962,62,325. E. Vander Donclct, Progress in Reaction Kinetics, 1970, 5, 273. Th.Forster, 2. Elektrochem., 1950,54,42. A. Weller, 2. Elektrochem., 1952,56,662. H. Kokubun, 2. Elektrochem., 1958,62,599. l o M. v. Smoluchowski, 2. phys. Chem., 1917,92, 129. A. C. Hopkinson and P. A. H. Wyatt, J. Chem. SUC. B, 1967, 1333. l2 M. A. Paul and F. A. Long, Chem. Rev., 1957,57, 1. l3 A. R. Katritsky, A. J. Waring and K. Yates, Tetrahedron, 1963,19,465. J. T . Edward and I. C. Wang, Can. J . Chem., 1962,40,966. E. M. Arnett and J. N. Anderson, J . Amer. Chem. SOC., 1963,85,1542 ; R. B. Homer and R. B. Moodie, J. Chem. SOC., 1963,4377 ; E. M. Arnett and G. W. Mack, J. Amer. Chem. SOC., 1966,88,1177 ; T. G. Bonner and J. Phillips, J. Chem. SOC. B, 1966,650. * J. B. Birks and I. H. Munro, Prog. Reaction Kinetics, 1967, 4, 239. l4 A. Weller and W. Urban, Angew.Chem., 1954,66,336. Protonation of Aromatic Carboxylic Acids in the First Excited Singlet State BY A. R. WATKINS” Department of Chemistry, University of Sheffield, Sheffield Received 1st July, 1971 Previous methods for determining the excited state acid-base characteristics of molecules from kinetic data are applicable only to molecules which undergo dissociation upon excitation. This paper presents a method of obtaining kinetic data from fluorescence measurements for compounds, which, upon excitation, take on a proton. The kinetic data so obtained can be used to calculate the excited state acidity constants of the protonated molecules. Comparison, for two naphthoic acids, of the results so obtained with data from the spectral method due to Forster and with data from the pH at which the fluorescence spectrum undergoes a transformation reveals considerable discrepancies : these are discussed briefly.Weber was the first to discover that, with change in pH, some organic molecules undergo a change in wavelength of fluorescence emission ; this was later interpreted by Forster as being due to a change in acid-base properties of the excited species. Since then a considerable literature dealing with acid-base properties of excited states has accumulated. 3-5 There are essentially three methods of gaining information about the acid-base properties of excited molecules. The first is the direct observation of the pH at which the fluorescence of the excited molecule is replaced by the fluorescence of its conjugate ; under certain conditions this “transformation” pH, referred to later in this paper as pHtr, gives a measure of the pK, or pK, of the excited The second method, proposed by Forster,6 relates the 0-0 transitions of the molecule and its conjugate acid or base to the equilibrium constants of the proton transfer reaction in the ground and excited states; a knowledge of the ground state equilibrium con- stant together with the relevant spectral data allows calculation of the excited state equilibrium constant.The third method, developed by Weller,’ consists of measuring the forward and reverse rate constants for the proton transfer reaction in the excited state, from which the equilibrium constant can be calculated. and generally good agreement is found. We were interested in extending this comparison to weak bases which become more basic on excitation, for this purpose I-naphthoic acid and 2- naphthoic acid B* + Hf + B*H+ were chosen as the probe compounds. The three methods have been compared by several authors,3* EXPERIMENTAL The compounds studied were obtained commercially and were purified by repeated crystal- lization ; in all cases the observed melting point agreed with the literature value.Fluoresc- ence intensities were measured with an Aminco-Bowman spectrofluorometer, variations in * present address : Max-Planck-Institut fur biophysikalische Chemie, D 34 Gottingen-Nikolaus- burg, West Germany. 28A . R. WATKINS 29 lamp intensity being taken into account by the use of a standard solution M quinine sulphate in water or a solution of the compound at a fixed pH) whose fluorescence quantum yield remained constant during the course of the experiment.All experiments were carried out at 25 fO.l”C in aqueous solution ; the pH was varied by addition of known amounts of hydrochloric acid. The concentration of the naphthoic acids used in these experiments was The effect of overlap between the emission spectra of the fluorescing base and its conjugate 2~ 10-5 M. acid was corrected for by use of the equations I’ - k l q’= ~ I - k‘I’ y=- I-kk” 1 - kk’ ’ where 1 and I’ are the measured intensities at the emission wavelengths of the protonated and unprotonated species, k and k’ represent the proportion of overlap from the protonated and unprotonated emitting species at the wavelengths used, and y and q’ represent the correspond- ing true intensities (throughout this paper primed quantities refer to the free base, unprimed quantities to the protonated base).The wavelengths at which the emission intensities were measured are given in table 1. TABLE 1 .-EXCITATION AND EMISSION WAVELENGTHS USED IN THE FLUORESCENCE MEASUREMENTS excitatian emission wavelength wavelength nm compound nm B* B*H+ 1-naphthoic acid 295 405 463 2-naphthoic acid 270 380 454 The relative quantum yields of fluorescence from the excited states of the base and its protonated conjugate are shown in fig. 1 and 2 (q/qo measures the relative quantum yield of fluorescence for the protonated excited state, y’/q& the relative quantum yield for the un- protonated excited state) as a function of the acidity of the solution for both naphthoic acids.--logKl [HCll FIG. 1.-Relative fluorescence quantum yields of 1-naphthoic acid as a function of acidity. The open circles refer to the unprotonated excited species (q’/q&), the closed circles to the protonated excited species (q/qo). In all measurements it was important to ensure that the compound was present in the ground state as the undissociated molecule, and that formation of the anion or the protonated carboxylic acid in the ground state, or of the anion in the excited state, in the pH range used was negligible. For this reason the maximum pH used was, in every case, smaller than the30 pH at which the extent of dissociation into the anion in the ground state is 3 %. Since the protonation of aromatic carboxylic acids generally occurs at quite high acidities (correspond- ing to Hammett acidity function values l2 of about -6), it is clear that significant proton- ation of the carboxylic acids in the ground state in these experiments could not occur.Dissociation of the excited naphthoic acids into the carboxylate anions also lies well outside this ~H-range.~ PROTONATION OF AROMATIC CARBOXYLIC ACIDS -0 F 2 F 0 E F U -0.5 0 a5 1.0 1.5 2D --log,, W l l FIG. 2.-Relative fluorescence quantum yields for 2-naphthoic acid as a function of acidity. The open circles refer to the unprotonated excited species (q’/qA), the closed circles to the protonated excited species (q/qo). Fluorescence lifetimes were measured with an apparatus consisting of a repeating spark light source, the light from which intermittently excited the fluorescent sample.The emitted light was Gollected by a 54AVP fast-response photomultiplier, the output of which was fed to a Hewlett-Packard 185 sampling oscilloscope. In all cases the resulting decay curves fitted an exponential decay law. All lifetime measurements were made at 25f0.1°C, and TABLE 2.-LIFETIMES OF PROTONATED AND UNPROTONATED EXCITED STATES. tlns 7‘lns compound (protonated base) (unprotonated base) 1-naphthoic acid 4.5 f1.5 5.2 ~t0.4 2-naphthoic acid 7.1 &0.7 15.2 h0.2 were corrected, where necessary, for the response time of the instrument. To avoid the possibility of dissociation in the ground state, measurements of the fluorescence lifetimes of the unprotonated carboxylic acids were made in benzene.The relation TbenzenehHzO = qbenzene/rHzO (2) allows the corresponding lifetimes in aqueous solution to be calculated from these measure- ments. The fluorescence lifetimes of the protonated excited molecules were obtained from measurements at high acid concentrations where the proton transfer reaction ceases to have any influence on the observed lifetimeY8 and were converted to the true lifetimes in water by means of an equation similar to (2). The method by which ~ H ~ O (qo in the discussion below) is obtained is described in the following section. Table 2 gives the life-times so obtained, together with an estimate of the error due to the response time of the instrument ; z refers to the protonated excited state, z’ to the unprotonated excited state.A .R. WATKINS 31 THEORY The equations relating quantum yield of fluorescence to pH for the protonation of a weak base in the excited state have been derived by Kokubun in his analysis of the protonation of acridone ; although the quantity (1 + klz')/k2z can be derived, where kl and k2 are the forward and back rate constants for protonation in the excited state, he found no way of obtaining these rate constants separately. The present analysis, on the other hand, allows kl, kz, and nQ, the quenching constant for the quenching of the protonated excited molecules by the added hydrochloric acid, to be calculated. Addition of acid to a solution of a base can produce two effects : the protonation of the base, either in the excited state or in the ground state or both, and quenching of the fluorescences of the emitting species. The systems studied here consist of carboxylic acids which, as discussed above, remain substantially undissociated and unprotonated in the ground state in the range of acidity used here.Proton transfer occurs only in the excited state, according to the following scheme, in which the mol- ecule B is excited by absorption of light, and then undergoes the reactions shown : O+hvJ 0 B B H + + ~ V BH+ BH+ where n; and ne are rate constants for emission of fluorescence, n;l and nL describe radiationless decay of the excited state, n;2 and nQ refer to the quenching action of the added acid, and kl and k2 are the forward and reverse rate constants for the protona- tion reaction.The relative quantum yields of fluorescence from the protonated and unprotonated excited states are given by the following equations : which can be obtained by application of steady state kinetics to the reaction scheme given above.7 From these equations the following expression for the ratio of relative quantum yields can be derived : Thus, when the ratio of the experimentally determined quantum yields is plotted against the reciprocal of the acid concentration, the points are expected to fall on a straight line of slope (1 + k 2 ~ ) / k l ~ ' and intercept nQz/klz'. At low acid concentrations it is found experimentally that a plot of y against y ' is linear, i.e., the sum of the relative quantum yields of fluorescence of B* and B*H+ becomes independent of acid concentration.This may be written, from eqn (4) and (5), as = const. (7)32 PROTONATION OF AROMATIC CARBOXYLIC ACIDS This will only be the case when the term containing [Hell2 in the denominator be comes negligible, and when klrf 4- npT > nQTf + n;2zfk2T (8) in which case the sum of the quantum yields tends to unity. Proceeding from this inequality and eqn (4), the following relation may be derived : and, with the value of (1 + k27)/k17’ previously obtained, the left-hand side of eqn (9) may be plotted against acid concentration to give a straight line of slope YtQz. It is clear that, with the values of nQz, nQz/kl.c’ and (1 +k2z)/klz’ obtained as shown above, together with the measured lifetimes, the rate constants k l , k2 and nQ can be calculated from the experimental measurements. It should be mentioned that ylo and qb, the fluorescence intensities o f the proton- ated and unprotonated carboxylic acids in the absence of any proton transfer reaction, could not be measured directly, since the dissociation of the carboxylic acid in the ground state becomes important at high pH values, and in strongly acid solutions the fluorescence of the protonated carboxylic acid is heavily quenched.Instead the equation which follows from eqn (7) and the discussion above, and which is expected to be valid in the low acidity region where no fluorescence quenching by the added acid occurs, can be applied. DISCUSSION Fig. 3 and 4 show the application of eqn (6), and fig. 5 and 6 the application of eqn (9), to the data for both naphthoic acids.Table 3 lists the quantities derived from these graphs, and table 4 gives the values of the rate constants derived from table 3. -0 20 LO 60 80 too 120 IbO 1 / [ H a eqn (6)). FIG. 3.-Plot of q‘q~/q-& against the reciprocal of the acid concentration for 1-naphthoic acid (seeA . R. WATKINS 33 1/[Ha eqn (6)). FIG. 4.-Hot of q'q,Jq~& against the reciprocal of the acid concentration for 2-naphthoic acid (see [Ha FIG. 5.-Application of eqn (9) to the experimental data for 1-naphthoic acid. 234 PROTONATION OF AROMATIC CARBOXYLIC ACIDS The last column of table 4 gives the calculated diffusion-controlled rate constants, assuming an average distance of reaction of 7 A.1° The discrepancy between the experimental and calculated values of kl for I-naphthoic acid can be explained by the steric requirements of the proton transfer step of eqn (3).For this reason the value [HClI FIG. 6.-Application of eqn (9) to the experimental data for 2-naphthoic acid. of k2 enclosed in brackets for 2-naphthoic acid, where nQz/klz' is too small to be measured, was calculated from the experimentally determined quantities in conjunc- tion with a value of kl of 2 x 1 O 1 O 1. mol-1 s-l. TABLE 3.-PARAMETERS DERIVED FROM THE APPLICATION OF EQN (6) AND (9) TO THE EXPERI- MENTAL MEASUREMENTS 1 +k2r nQt nQr kid klr' - compound 1-naphthoic acid 0.0301 0.104 9.54 2-naphthoic acid 0.0972 0 0.352 In view of the mathematical complexity involved, no attempt has been made to incorporate a correction for the possible presence of non-steady-state conditions at high acid concentration^.^ The small value of the intercept in fig.3, together with possible effects due to the changing medium on the kinetics in the measurement of Y t Q q will mean that, even in the absence of non-steady-state effects, the derived rate constants will be subject to some uncertainty. TABLE 4.-RATES OF PROTONATION (ki), DEPROTONATION OF THE EXCITED NAPHTHOIC ACIDS (k2), TOGETHER WITH QUENCHING CONSTANTS ( n ~ ) FOR THE EXCITED PROTONATED ACIDS ki k2 nQ ki(alc) compound I. mol-1 s-1 S-1 1. mol-1 s-1 1. mol-1 s-1 1-naphthoic acid 1 . 7 ~ 10'' 1 . 7 ~ 10' 2.1 x 109 4.7x 1 O 1 O 2-nap ht ho i c acid - ( 4 . 0 ~ 109) 2.3 x 107 4.7x loio In table 5, values of the ground state and excited state acidity constants of the protonated carboxylic acids are presented, the latter calculated from the kinetic data of table 4.For comparison, excited state acidity constants calculated from spectralA . R. WATKINS 35 data have been included together with the acidity constants of the protonated car- boxylic acids in the ground state." The discrepancies between the acidity constants derived from kinetic and spectral data in table 5 cannot be accounted for by the uncertainties in the measurements of the rate constants. The difference of 2.2 units for 1-naphthoic acid, for example, TABLE 5.-ACIDITY CONSTANTS OF THE PROTONATED NAPHTHOIC ACIDS IN THE EXCiTED AND GROUND STATES (pKZ AND pK,, RESPECTIVELY) compound pK: (kinetic) pKX (spectral) PKa 1-naphthoic acid 2.0 - 0.20 - 7.70 2-naphthoic acid 0.70 - 0.35 - 7.68 a, this work; b, ref.(11). would mean, in the event that the pKif' from spectral data was correct, that the ratio of the rate constants k2/kl differed by a factor of 100 from the true value, which is too large to be accounted for by the experimental uncertainty (maximally+ 50 %) in the rate constants. At least part of the discrepancy must be attributed to the fact that the ground state pK, values,ll from which the values of pK,* (spectral) are derived, are based on Hammett acidity functions whose applicability to measurements of equilibria in concentrated acid solutions has been opened to doubt by a number of a ~ t h 0 r s . l ~ This is best illustrated by a comparison of acidity scales (Ho scales) measured with different types of bases. For example, the value of the acidity function Ho obtained for 80 % H2S04 (the acidity region in which protonation of carboxylic acids occurs) is - 7.34 based on primary aromatic amines, - 9.44 (tertiary aromatic amines), - 1 1.84 (triarylcarbinols) and - 6.35 (benzophenone~).~~ No acidity scale based on aromatic carboxylic acids has yet been reported, but, in view of the variations found above, it is highly likely that such a scale would behave differently to the traditional scale based on primary aromatic amines (used in evaluating the pK, values quoted in table 5).The pH at which change-over of the fluorescence spectrum from one component of the acid base equilibrium to the other (pH,,) occurs is sometimes taken as being equal to pK: ; however, this has only limited validity.It is clear that, depending on the values of t?Q and nh, this pH will be determined not only by the kinetic constants of the proton transfer reaction, but will also be determined by competing quenching reactions occurring between the added acid and the fluorescing species. From the analysis given above, the pH at which the quantum yields of fluorescence of proton- ated and unprotonated emitting species are equal is given by : This corresponds to the pH for fluorescence transformation measured by Urban and Weller,14 and to the pKgH+ values determined by Hopkinson and Wyatt? The values of these quantities determined by these authors for 1-napthoic acid and 2- naphthoic acid, together with the pH,, values obtained in the present work, are listed in table 6. The measured pH,, will only give an accurate indication of pK: when nQ7 e klzf and k27 9 1, (12) i.e., when the added mineral acid is an inefficient quencher of B*H+ and when the dissociation reaction is fast in comparison with the other processes leading to the disappearance of B*H+.36 PROTONATION OF AROMATIC CARBOXYLIC ACIDS The equality of pH,, with pK,* depends on the order of magnitude of K:( = k2/k1).If k , is taken to be diffusion controlled (2: loio 1. mol-1 s-l) and if the lifetimes z' and z are approximately equal, then the range of possible values which pK: can take can be divided into two regions of interest : (i) pK: <2, k2/kl > s, then k2z> 1 ; however, pH,, = pK: only when nQz<k,z'. (ii) pKZ >2, k2/kl < In this case PHtr will be independent of pK:, and will be given (neglecting the second term in eqn (1 1)) by Here, in contrast to case (i), the rate of protonation is too slow to compete with the other processes that deactivate B*, so that, as the pH is increased, the fluorescence of B*H+ disappears before it is expected to.If z is taken as By the same reasoning k2z < 1. pH,, N log,,klz' N 2. (13) TABLE 6a-COMPARISON OF THE pH OF FLUORESCENCE TRANSFORMATION PHtr FROM VARIOUS SOURCES compound PHtr" PHtrb PHtrC PKt" nQ 1 -naphthoic acid 1.5 0 1.48 2.0 2 . 1 ~ 109 2-naphthoic acid 1.5 0 1.04 0.70 2 . 3 ~ lo7 Q, ref. (14); b, ref. (11); c, this work. From table 3, it can be seen that both naphthoic acids belong to category (i) above, and that the conditions (12) will be fulfilled by 2-naphthoic acid but only partially by 1-naphthoic acid ; the measure of agreement between columns 4 and 5 of table 6 is reasonably consistent with this analysis. The disagreement between the PHtr values of Weller and Urban,14 and Hopkinson and Wyatt,ll may be a consequence of the use of different mineral acids to fix the pH of the solutions in which the experiments were carried out. Thanks are due to the Science Research Council for a fellowship, during the tenure of which this work was carried out, and to Prof. P. A. H. Wyatt for helpful discussions. l K. Weber, 2. phys. Chem. B., 1931,15, 18. Th. Forster, Naturwiss., 1949, 36, 186. A. Weller, Progress in Reaction Kinetics, 1961, 1, 189. B. L. Van Duuren, Chem. Rev., 1962,62,325. E. Vander Donclct, Progress in Reaction Kinetics, 1970, 5, 273. Th. Forster, 2. Elektrochem., 1950,54,42. A. Weller, 2. Elektrochem., 1952,56,662. H. Kokubun, 2. Elektrochem., 1958,62,599. l o M. v. Smoluchowski, 2. phys. Chem., 1917,92, 129. A. C. Hopkinson and P. A. H. Wyatt, J. Chem. SUC. B, 1967, 1333. l2 M. A. Paul and F. A. Long, Chem. Rev., 1957,57, 1. l3 A. R. Katritsky, A. J. Waring and K. Yates, Tetrahedron, 1963,19,465. J. T . Edward and I. C. Wang, Can. J . Chem., 1962,40,966. E. M. Arnett and J. N. Anderson, J . Amer. Chem. SOC., 1963,85,1542 ; R. B. Homer and R. B. Moodie, J. Chem. SOC., 1963,4377 ; E. M. Arnett and G. W. Mack, J. Amer. Chem. SOC., 1966,88,1177 ; T. G. Bonner and J. Phillips, J. Chem. SOC. B, 1966,650. * J. B. Birks and I. H. Munro, Prog. Reaction Kinetics, 1967, 4, 239. l4 A. Weller and W. Urban, Angew. Chem., 1954,66,336.

ISSN:0300-9599    

DOI:10.1039/F19726800028

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Role of Metals in Enzymatic Reactions Part 5.-Kinetics of Ternary Complex Formation Between Magnesium and ManganeseOI) Species and 8-Hydroxyquinoline BY D. N. HAGUE, S. R. MARTIN AND M. S. ZETTER University Chemical Laboratory, Canterbury, Kent Received 23rd June, 1971 The temperaturejump relaxation method has been used to measure the activation parameters for the formation and dissociation of the 1 : 1 complexes between magnesium and manganesem) and 8-hydroxyquinoline (oxine) and of the ternary complexes between oxine and the magnesium and manganese@) complexes of nitrilotriacetate, uramil-NN-diacetate, adenosine-S-diphosphate, adenosine-5’-triphosphate and polytriphosphate. The results strengthen the previous conclusion that the first ligand has comparatively little influence on the kinetics of the reaction of magnesium with a secondligand,whereas manganesem) shows similarities to zinc in the reiactivityof its complexes. The enzymatic behaviour of magnesium and manganese(I1) is discussed in the light of these results.The kinetics of formation and dissociation of simple 1 : 1 complexes of labile metal ions in aqueous solution have received considerable attention recently and the results, many of them obtained with the help of chemical relaxation techniques, have been reviewed.l It is usually found that the rate constant for the formation of any complex of a given metal k, is constant, to within an order of magnitude or so, unless some special feature of the incoming ligand causes a deviatioa2 The mechanistic implications of the invariance in kf are the subject of debate,l* but the formation of complexes of bivalent metal ions in aqueous solution is usually discussed in terms of a two-step mechanism in which the hydrated metal ion and the ligand diffuse together rapidly to form an outer-sphere complex, and then one or more water molecule in the inner coordination sphere of the metal is replaced by the ligand.For most metals the second step is fast but does not approach the diffusion- controlled value of step (1) and, provided that the concentration of the outer-sphere intermediate is small, kf is given by K,,k>, where KS is the formation constant for the outer-sphere complex and k j is the rate constant for water-exchange at the metal ion. In most cases where the incoming ligand contains a second or further binding group, the rate-limiting step is the formation of the first metal-ligand bond.Comparatively little is known about the influence exerted by a ligand which is already bound to a metal ion on the subsequent reaction of the metal with another ligand. An understanding of the effects of bound ligands on ternary complex forma- tion is important because many enzymic and other catalytic processes requiring metal ions are thought to involve such complexes. The rate constants and activation parameters, as measured by the temperature-jump relaxation method, are reported for the formation in aqueous solution of the complexes of 8-hydroxyquinoline (oxine) with hydrated magnesium(II) and manganese(II), and with the magnesium@) and manganese(I1) complexes of nitrilotriacetate (NTA), uramil-NN-diacetate (UDA), adenosine-5’-diphosphate (ADP), adenosine-5’-triphosphate (ATP) and polytri- phosphate (TP).The rate constants at 16°C have already been published for four of the Mg2+ systems (aquo, UDA, ATP and TP)4 and five of the Mn2+ systems 3738 KINETICS OF TERNARY COMPLEX FORMATION (aquo, NTA, UDA, ATP and TP).5 It is particularly important to determine the influence of bound ligands on the reactivity of a do metal ion such as Mg2+, because the latter, by virtue of the lack of directionality imposed on the metal-ligand bonds by partial covalency and the lack of asymmetry in the electrical field around the metal ion caused by a partially-filled d electronic sub-shell, provides a norm against which the results on other metals in which these factors are present may be compared.Mn2+ is also electronically spherosymmetrical since it has a high-spin d5 configura- tion; in addition, it is found that many enzyme systems which require Mg2+ will function in vitro with Mn2+. EXPERIMENTAL Tri-sodium adenosine-5'-diphosphate (Boehringer) was used without further purification. Details of other materials and the methods used for the kinetic and equilibrium measure- ment have been p~blished.~. s The data reported here refer to the relaxation effects in the 1-100 ms (Mg2+) and 10 ps-1 ms (Mn2+) regions associated with metal-oxine complex formation. Temperatures were accurate to &0.2"C. RESULTS The mechanism postulated 4 9 for the formation of the magnesium or manganese complex (MOx+) is k3 2 3 HOx + M2+ - MOxH2+ k23 (3) (4) k43 k34 kiz 11 kzi k4 1 H+ + Ox- + M2+ + MOx+ + H+.k14 This represents a two-path mechanism for the formation of MOx+ from the free metal M2+ in which the path involving the oxine anion Ox- is important at higher pH and the other, involving the oxine molecule HOx, is important at lower pH. Since H+ is buffered, M2+ is present in large excess over oxine and the deprotonation steps can be treated as rapid pre-equilibria, the observed relaxation time z is given by where c, and cH are the equillibrium concentrations of M2+ and H+, respectively, K2 = k43/k34 and K3 = k12/k2,. Similar expressions may be used in the ML+ oxine systems, where L is the firmly bound ligand NTA, UDA, ADP, ATP, or TP. (Charges have been omitted, and activity coefficients neglected. Eqn (1) is slightly different from the expressions used previ~usly.~. This is because in t h s paper we prefer to consider the general situation in which significant concentrations of the species MOxH2+ (and MLOxH) may exist. In most systems, however, the steady- state approximation may be applied and in this case eqn (1) reduces to the more restricted expression used previously since &CH < 1 .) the data were analyzed by plotting z-l against cH at low pH or against (1 +K2cH)-l at high pH (at a fixed M2+ or ML concentration), and by plotting 7-l against cM or cML at a fixed pH.It was sometimes possible to obtain a complete internal check on the mechanism and rate constants from the symmetrical nature of the pH profile of z-l.Although this method could have been used in most systems studied here, the data were actually analyzed with the help of a curve-fitting computer programme. Values of k41, k23K3, k32 and K3 were varied until the pH profile of 7-l calculated from eqn (1) fitted the profile indicated by the experimental points. The criterion used for the best fit was that the sum of the moduli of the percentage T-l = cM/(l +K2cH)(k41 +k3ZK2CH) + l/(l +K3CH){k14+k23K3CH), (1) PreviouslyY4*D. N. HAGUE, S. R. MARTIN AND M. S . ZETTER 39 IOOC rl 8CX L \ r( I 60( 40C I I I I t t 1 6.7 7.0 7.3 7-6 7.9 0.2 PH 7.7 8.0 8.3 8.6 8.9 9.2 9.5 PH FIG. 1.-Typical plots showing the variation of 2' with pH for the reaction of (a) Mg2+ (4 x 4O"C, where K3q&1) ; and (b) MgUDA- (8 x M, M, 40°C, where K~CH is not <l).Full lines on each point indicate experimental scatter.40 KINETICS OF TERNARY COMPLEX FORMATION deviations of the observed reciprocal relaxation times from the calculated values should be a minimum. Such a criterion was considered more reasonable than one based on a least-squares treatment since the latter would have placed too much weight on the higher values of 7-l. However, it was not possible to incorporate it into any statistical method for analyzing the errors associated with a particular rate constant and the uncertainties quoted are estimates. Two typical runs, together TABLE 1 .-RATE AND EQLJILIBRIUM CONSTANTS FOR THE REACTION OF MAGNESIUM SPECIES WITH OXINE (APPROX. 4 x M) AT VARIOUS TEMPERATURES Mg2+ 10.0 20.0 30.01 40.0 MgNTA- 10.0 20.0 30.0 40.0 MgUDA- 10.0 30.0 40.0 MgADP- 10.0 20.0 30.0 40.0 20.0 30.0 40.0 20.0 30.0 MgATP2- 10.0 MgTP3- 10.0 6.5-8.8 6.8-8.6 6.6-8.5 6.7-8.4 7.4-9.6 7.3-9.2 6.9-9.2 7.0-9.1 8.5-10.5 8.0-10.0 7.9-9.7 7.9-10.1 7.5-9.7 7.4-9.6 7.3-9.4 7.9-10.3 7.8-9.9 7.6-9.7 7.6-9.7 8.4-10.5 8.2-10.2 8.1-10.1 20 39 83 180 11 25 40 4.6 1.9 7.8 9.2 14 19 37 58 4.5 9.3 17 27 1.4 3.5 8.7 0.46 1.1 1.8 2.7 0.86 1.7 3.0 7.9 5.3 11 16 0.79 2.2 2.9 6.6 0.94 1.6 3.4 5.8 1.4 3.1 6.5 6.5 9.2 26 89 1 .o 2.5 6.8 10 43 230 430 32 37 120 1 80 15 35 61 140 12 35 4.4 k23K31 (1010 M-1 S-1) 0.14 0.16 0.23 0.39 0.17 0.23 0.34 0.57 110 130 1 40 2.3 2.3 3.6 5.1 2.6 3.4 4.4 7.7 3.6 6.0 9.7 The approximate concentration of the magnesium species was 8 x M except for Mg2+ and MgNTA- at 40T, where it was 4 x M.Ionic strength (KNO,) was 0.10 M for Mg2+, MgNTA- and MgUDA- and 0.30 M for MgADP-, MgATP2- and MgTP3-. Values of K3 (lo8 M-l) for MgUDA- were 8.8 (lO"C), 3.0 (30°C) and 2.O(4O0C). with their respective computed " best-lines " fits, are shown in fig. 1, and the rate and activation data are given in tables 1-4 (the fact that the 16°C rate constants given in tables 1 and 2 sometimes differ slightly from the ones quoted previously 4* is attributable to the different methods used to evaluate them ; the ones given here are considered to be the more reliable). The normal statistical adjustments have been made 4* to the formation rate constants, the activation entropies and the stability constants in order to allow for the different numbers of water molecules available for substitution by Ox- or HOx. k i l t ki2, etc., have been estimated on the assump- tion that the metal ions are hexahydrated and that in ML the ligand L is n-dentate (tables 3 and 4).The dependence of the dissociation constant of HOx on temperature was deter- mined spectrophotometrically. The method, which was similar to the one used previo~sly,~ employed a curve-fitting computer programme. The values of log K2 obtained at various temperatures were : 9.87 (16.0°C), 9.70 (24.0°C), 9.59 (32.0"C) and 9.46 (40.0"C).D . N. HAGUE, S . R . MARTIN AND M. S . ZETTER 41 TABLE 2.-RATE AND EQUILIBRIUM CONSTANTS FOR THE REACTION OF hUNGANESE(IT) SPECIES WITH OXINE (APPROX. 4 x M) AT vmous TEMPERATURES temp. approx. k41 k32/ system ("0 pH range (106 M-1 s-1) (106 M-1 s-1) Mn2+ 16.0 22.8 29.9 35.0 39.5 MnNTA- 7.0 16.0 28 .O 40.0 MnUDA- 7.0 16.0 28.0 40.0 MnADP- 16.0 20.0 26.0 32.5 36.0 MnATP2- 10.0 16.0 28.0 40.0 MnTP3- 10.0 16.0 28.0 40.0 5.3-7.2 5.2-7.2 5.0-7.0 5.0-6.9 5 .O-6.6 6.5-8.7 6.3-8.5 6.5-8.6 6.5-8.5 6.6-8.8 6.5-8.6 6.5-8.5 6.8-8.1 6.8-8 .O 7.1-8.0 6.9-7.9 7.0-7.9 7.5-8.9 7.1-9.0 7.3-9.2 7 .O-8.4 7.9-9.7 7.8-9.8 7.7-9.3 7.5-9.1 6.4-8.3 260 220 420 730 720 20 33 38 10 33 26 38 78 87 98 100 120 8.1 0.96 1.7 3.1 9.1 0.34 0.58 1 .o 2.0 0.31 0.41 0.57 0.90 1 .o 0.83 1.4 1.9 2.7 0.86 1 .o 1.5 12.2 1.3 1.2 1.3 1.9 1.8 0.56 0.58 1.4 2.3 0.24 0.23 0.46 0.72 k141 (s-1) 350 370 680 730 860 54 96 210 450 28 180 200 310 670 1800 2200 2100 3300 20 44 86 430 65 170 450 1400 k2 3K3l (1010 M-1 s-1) 0.30 0.33 0.38 0.31 0.36 5.9 4.9 5.5 9.1 2.6 4.0 4.9 5.3 8.8 15 14 14 16 11 15 27 39 44 80 120 9.9 The approximate concentration of the manganese(II) species was 1 .2 ~ M except for MnNTA- and MnUDA- at 28 and 40°C, where it was 6 x M. Ionic strength (KN03) was 0.10 M for Mn2+, MnNTA- and MnUDA-, 0.15 M for MnADP-, and 0.30 M for MnATP2- and MnTP3-. DISCUSSION The results are discussed under three main headings. The magnesium and manganese systems are considered in turn, and then some implications of the results on the behaviour of the metals in enzymic reactions are discussed. In the previous study with magnesium it was shown that the bound ligand had little effect other than reducing the solid angle of the metal ion over which effective substitution may occur.This implies that the local charge density at the surface of the metal ion rather than the overall charge on the complex is the factor which controls the reaction rate, and that this is not affected much by the presence of negatively-charged groups on the other side of the metal ion. The 25°C rates and activation parameters (table 3) confirm this. The activation enthalpies for the formation reactions (i.e., AHZl and AH&) are approximately the same, with the exception of the MgUDA-+ HOx system which is discussed separately. They are also close to the activation enthalpy for the water-exchange reaction (AH* = 10.2 kcal mol-', ref. (6)). Whereas previously 4* it was suggested that the rate-limiting step for the reaction of the oxine molecule might not be the loss of the first water molecule from the inner coordination sphere of the metal, Johnson and Wilkins7P N TABLE 3.-RATE AND EQUaIBRIUM CONSTANTS AND ACTIVATION PARAMETERS FOR THE REACTIONS OF MAGNESIUM SPECIES WITH OXINE (ESTIMATED ERRORS IN BRACKETS) The rate constants refer to 25°C at ionic strength 0.10 (Mg2+, MgNTA-, MgUDA-) or 0.30 (MgADP-, MgATP2-, MgTP3-), n is the number of coordination positions at Mg2+ assumed to be occupied by the ligand L.Mg2+ (6HzO) 6.0 (0.5)~ lo5 5.78 12.3 (0.6) +9 (2) 1.3 (0.2) x 104 4.11 9.6 (1.6) -7 (5) 20 ( 5 ) 15.0 (2.9) -2 (10) 4.48 MgNTA- (n = 4) 1.5 ( 0 . 2 ) ~ 105 +9 (3) 5.65 12.3 (0.9) 2.5 ( 0 . 3 ) ~ lo4 4.88 12.2 (1.3) f 5 (4) 3.7 (0.5) 13.3 (1.5) -11 (5) 5.08 MgUDA- (n = 4) 5.5 ( 0 . 4 ) ~ 104 5.23 11.2 (1.5) +3 (6) 9.3 (0.7)~ 104 5.45 5.7 (0.8) -14 (3) 150 (40) 13.0 (3.0) - 5 (10) 3.05 MgADP- (n = 2) 2.5 (0.3)~ lo5 5.58 10.3 (0.6) +2 (2) 2.5 ( 0 .3 ) ~ lo4 4.58 11.2 (1.6) 0 (5) 73 (10) 10.4 (2.6) -15 (9) 3.72 MgATP2- ( n = 3) 1.2 ( 0 . 3 ) ~ 105 0 (2) 2.4 (0.3) x lo4 4.68 10.3 (0.5) -2 (5) 48 (8) 12.0 (0.7) -11 (3) 5.38 10.0 (0.5) 3.70 MgTP3- (n = 3) 5.6 ( 0 . 5 ) ~ lo4 5.04 14.7 (1.5) + 14 (6) 4.5 ( 0 . 5 ) ~ 104 4.95 12.3 (1.3) +6 ( 5 ) 21 (6) 17.1 (3.0) + 5 (10) 3.72TABLE 4.-RATE AND EQUILIBRIUM CONSTANTS AND ACTIVATION PARAMETERS FOR THE REACTIONS OF MANGANESE(I1) SPECIES WITH OXINE (ESTIMATED ERRORS IN BRACKETS) The rate constants refer to 25°C at ionic strength 0.10 (MnZ+, MnNTA-, MnUDA-), 0.15 (MnADP-) or 0.30 (MnATP2-, MnTP3-). n is the number of coordination positions at Mn2+ assumed to be occupied by the ligand L.k14/(s- '> AH&/(kcal mol-') ASf4 /(cal mol-' K-l) log Ki (kinetic) Mn2+ (6H20) 3.3 (0.6)~ lo8 8.52 8.8 (4.6) + 10 (16) 4.9 ( 0 . 7 ) ~ 105 5.69 9.1 (1.5) -2 (5) 480 (60) 7.0 (0.5) -22 (9) 5.84 MnNTA- (n = 4) 2.3 (0.3)~ lo7 7.84 6.6 (2.4) 0 (8) 1.7 (0.3)~ lo6 6.71 5.5 (0.7) -9 (2) 180 (30) 10.6 (2.2) - 12 (7) 5.58 MnUDA- (n = 4) 2.2 ( 0 . 3 ) ~ 107 7.82 6.4 (2.5) -1 (8) 1.4 (0.2) x lo6 6.62 4.5 (0.5) -13 (2) 130 (60) 12.4 (2.0) -7 (7) 5.71 MnADP- (n =i. 3) 8.5 ( 0 . 6 ) ~ lo7 8.23 4.4 (2.5) -6 (8) 1.4 ( 0 . 2 ) ~ lo6 6.45 4.0 (1.5) -15 (6) 1400 (700) 12.0 (3.0) -3 (9) 4.98 MnATP2- (n = 4) 2.5 ( 0 . 3 ) ~ lo6 6.88 12.0 (1.2) +I3 (4) 9.3 ( 0 . 8 ) ~ 105 6.45 8.3 (1.3) +1 (5) 90 (40) 16.4 (2.2) + 6 (8) 4.92 MnTP3- (n = 3) 9.1 (0.6) x 105 6.66 9.8 (0.6) + 5 (2) 4.0 ( 0 .5 ) ~ lo5 5.90 6.4 (1.3) - 10 (4) 350 (100) 16.8 (1.3) + 10 ( 5 ) 4.1244 KINETICS OF TERNARY COMPLEX FORMATION found that HOx behaves as a “ typical ” ligand in its reaction with hydrated Ni2+. The data in tables 3 and 4 indicate that the differences between k i l and k& for a given ML are largely entropic in origin, and reaction (3)-(2) will therefore be taken as typical for complex formation at the MgL species. The larger value of AHzl for MgTP3- might be associated with a charge repulsion between TPs- and Ox-, although AH:4 is also higher for the MgTP3- system than for the others. The entropies of activation for the formation reactions and AS,’,cs)) are also similar (except for MgUDA-, see below) but the values for Mg2+ + HOx and MgTP3-+ Ox- are out of line.The values of AH* and AS* for the dissociation process show a similar slight fluctuation across the series but again there seems to be no trend which can be attributed directly to a repulsion between the oxine anion and the negatively- charged bound ligand L. Unfortunately, the experimental accuracy is insufficiently high to attempt a more detailed interpretation of the variation in activation parameters along the series. 0 I The reaction between MgUDA- and HOx is the only magnesium reaction reported here for which the activation parameters differ significantly from the rest. Although it is possible that the low value of AH* and large negative value of AS* are associated with an increased lability of the bound water molecules, as has been suggested for the reaction of CoNTA- with the neutral bidentate ligand pyridine-2-azo-p-dimethyl- aniline (PADA),8 we prefer to invoke an increase in the value of KO, (with correspond- ingly favourable AHos but unfavourable ASos) to explain the enhanced reactivity because the reduced AH& is not reflected in a change in AHzl or AHT4.If 11.5 kcal mol-1 is taken as the typical activation enthalpy for complex formation at magnesium and 0 cal mol-1 K-l as the corresponding statistically-corrected activation entropy for a neutral incoming ligand (these values being estimates based on the data in table 3) then, the difference in AH,, and ASos between the MgUDA- + HOx system and the rest will be, respectively, - 5.8 kcal mol-l and - 14 cal mol-l K-l.These correspond to a value of AGos which is more favourable by approx. 1.5 kcal mol-l than average-equivalent to an equilibrium constant of 10-15 M-l. It is possible,D. N. HAGUE, S . R. MARTIN AND M. S . ZETTER 45 using the value of K3 estimated from the computer fit of the relaxation data (fig. l(b) and table l), to calculate the value of K4 (= k32/k23) ; it has a similar magnitude to the value estimated from the increments in AHos and ASos. The question of the physical significance of these figures, which represent differences in AGos and KO, from normal, should not be pressed too hard in view of the fact that the exact mechanism of metal complex formation is still uncertain.l* However, the geometry of the barbiturate ring of UDA is such that a complex reminiscent of the adenine-thymine base-pair found in double-stranded DNA may be formed between MgUDA- and HOx.In this outer-sphere complex (I), the oxine molecule is conveniently placed for substitution at the metal ion. The values of AH* and the(second-order) rate constant at 25°C for water-exchange at Mn2+ are, respectively, 8.8 (k 1.0) kcal mol-l and approx. 3 x lo6 M-l s-l (ref. (9)) and it is assumed, as in the magnesium case, that reaction (3)-(2) may be taken as typical for complex formation at the MnL species. For this reaction with manganese(I1) there is essentially the same pattern (table 4) as has been found lo with zinc in its reaction with PADA : the activation enthalpy for the substituted metal is a few kcal mol-1 less than the value for the unsubstituted metal, but this difference in AH* is largely compensated by a corresponding difference in AS*.This is also found for the reaction of Ox- and the fact that the activation enthalpies for the two series are similar suggests that the variation in rate constant, etc., for different manganese species originates in the water-exchange process at the metal ion. We would therefore expect the activation enthalpies for the latter process in MnL to reflect the values of AH:, and AHZ2 (AHz2 for MnATP2- and MnTP3-). The general similarity between AH:, and AH,+, for a given MnL suggests that here, as with magnesium, there is usually no direct interaction between the negative groups on the bound ligand and those on the incoming ligand (although we attribute the high value of AH& for MnTP3- to electrostatic repulsion between the TP5- moiety and the Ox- anion, as for MgTP3-).No explanation is suggested for the trend in AH&, but it is opposite in direction to that predicted on the basis of a significant electrostatic repulsion between the bound L and Ox-. The low values of k i l for MnATP2- and MnTP3- noted in ref. (5) do not appear to be associated with the presence of phosphate groups as such, since MnADP- behaves in a similar way to MnNTA- and MnUDA-. There has been discussion concerning the nature of the binding between ATP and metal ions,ll and recent n.m.r. results suggest l2 that the nitrogen at the 7-position in the adenine ring is bound to Mn2+ through a complexed water molecule. It is possible that this outer- sphere interaction is responsible for the high value of AH& and contributes to the high value of AH& observed for the MnATP2- system.No such abnormally high values are found with the corresponding magnesium system (table 3). For any enzyme system, efficient turnover is dependent on obtaining a suitable balance between the time required for the enzyme-substratelproduct complex to form and the time taken for it to dissociate : the lifetime must be sufficiently long to permit the subsequent bond reorganization and possible conformational changes involving the ligands to take place, but it should not be so long as to tie up the active site unnecessarily. In part lY4 a kinetic model was discussed which could explain the frequent mutual antagonism of Ca2+ and Mg2+ and in part 2 it was shown that this model would also rationalize the frequent interchangeability of Mg2+ and Mn2+- neither of which can be explained simply in equilibrium terms.Essentially, if the catalytic pathway includes a conformational change of the enzyme which can only take place when the metal ion is bound, then the dissociation rate constant of the enzyme-metal complex must be less than the turnover number, which is typically46 KINETICS OF TERNARY COMPLEX FORMATION 102-103 s-I. (Although this was formulated for reactions which involve metal- bridged intermediates, it applies equally well for reactions in which the active form of the substrate is a metal complex.) Caution should be exercised in applying con- clusions drawn from model systems to enzyme reactions, but the results presented here confirm this rationalization.They also confirm the error in the widely-held view that ternary complexes form less readily and dissociate more easily than binaries because of the charge neutralization of the metal by the bound ligands or the electro- static repulsion between the two ligands. With magnesium, the activation enthalpy for the formation of the complex is abnormally high only in the most extreme case, where the bound ligand carries a charge of - 5 and the incoming ligand one of - 1. With manganese, the bound ligand exerts a greater influence on the reactivity of the metal ion ; however, it normally reduces AHZ1 rather than increases it. It is interesting that the reactivity of the MnATP2- complex (and, to a lesser extent, the MnTP3- com- plex) appears to be atypical of substituted manganese species in view of the fact that the magnesium/manganese interchangeability is prevalent in enzyme reactions which involve ATP.The only reaction reported here for which the activation parameters are markedly out of line is that between MgUDA- and HOx. This result illustrates one way in which an enzyme can accelerate a step in a reaction sequence. If the oxine molecule were part of the metal-enzyme (represented by MgUDA-) and complex formation between it and the metal ion was an elementary step in the reaction sequence, then the favourable AHos would not be compensated by an unfavourable ASos (since the outer-sphere complex would already exist) and the net metal- substitution rate would be even more enhanced than it is here.We are grateful to the S.R.C. for grants for maintenance (for S. R. M.) and the purchase of the temperature-jump apparatus, and to the University of Kent at Canterbury for a University Studentship (for M. S . Z.). e.g., M. Eigen and R. G. Wilkins, Mechanisms of Inorganic Reactions, R. F. Gould, ed., Adv. Chem. Series, no. 49 (Amer. Chem. SOC. Washington, D.C., 1965), p. 55 ; D. J. Hewkin and R. H. Prince, Coord. Chem. Rev., (Elsevier, Amsterdam), 1970,545. e.g., D. B. Rorabacher, Inorg. Chem., 1966,5,1891. H. P. Bennett0 and E. F. Caldin, J. Chem. SOC. A , 1971,2198. D. N. Hague and M. Eigen, Trans. Furaday Soc., 1966,62,1236. D. N. Hague and M. S. Zetter, Trans.Furaday SOC., 1970,66, 1176. J. Neely and R. Connick, J . Amer. Chem. SOC., 1970,92,3476. ' W. A. Johnson and R. G. Wilkins, Inorg. Chem., 1970,9,1917. * M. A. Cobb and D. N. Hague, Trans. Faraday SOC., 1971, 67, 3069. lo G. R. Cayley and D. N. Hague, Truns. Faruduy Soc., 1971,67,786. l 1 L. Rimai and M. E. Heyde, Biochem. Bipohys. Res. Cumm., 1970, 41, 313. See also ref. l2 T. A. Glassman, C. Cooper, L. W. Harrison and T. J. Swift, Biochemistry, 1971,10,843. M. Grant, H. W. Dodgen and J. P. Hunt, Inorg. Chem., 1971,10,71. quoted therein and in ref. (12). Role of Metals in Enzymatic Reactions Part 5.-Kinetics of Ternary Complex Formation Between Magnesium and ManganeseOI) Species and 8-Hydroxyquinoline BY D. N. HAGUE, S. R. MARTIN AND M. S. ZETTER University Chemical Laboratory, Canterbury, Kent Received 23rd June, 1971 The temperaturejump relaxation method has been used to measure the activation parameters for the formation and dissociation of the 1 : 1 complexes between magnesium and manganesem) and 8-hydroxyquinoline (oxine) and of the ternary complexes between oxine and the magnesium and manganese@) complexes of nitrilotriacetate, uramil-NN-diacetate, adenosine-S-diphosphate, adenosine-5’-triphosphate and polytriphosphate.The results strengthen the previous conclusion that the first ligand has comparatively little influence on the kinetics of the reaction of magnesium with a secondligand,whereas manganesem) shows similarities to zinc in the reiactivityof its complexes. The enzymatic behaviour of magnesium and manganese(I1) is discussed in the light of these results.The kinetics of formation and dissociation of simple 1 : 1 complexes of labile metal ions in aqueous solution have received considerable attention recently and the results, many of them obtained with the help of chemical relaxation techniques, have been reviewed.l It is usually found that the rate constant for the formation of any complex of a given metal k, is constant, to within an order of magnitude or so, unless some special feature of the incoming ligand causes a deviatioa2 The mechanistic implications of the invariance in kf are the subject of debate,l* but the formation of complexes of bivalent metal ions in aqueous solution is usually discussed in terms of a two-step mechanism in which the hydrated metal ion and the ligand diffuse together rapidly to form an outer-sphere complex, and then one or more water molecule in the inner coordination sphere of the metal is replaced by the ligand.For most metals the second step is fast but does not approach the diffusion- controlled value of step (1) and, provided that the concentration of the outer-sphere intermediate is small, kf is given by K,,k>, where KS is the formation constant for the outer-sphere complex and k j is the rate constant for water-exchange at the metal ion. In most cases where the incoming ligand contains a second or further binding group, the rate-limiting step is the formation of the first metal-ligand bond. Comparatively little is known about the influence exerted by a ligand which is already bound to a metal ion on the subsequent reaction of the metal with another ligand.An understanding of the effects of bound ligands on ternary complex forma- tion is important because many enzymic and other catalytic processes requiring metal ions are thought to involve such complexes. The rate constants and activation parameters, as measured by the temperature-jump relaxation method, are reported for the formation in aqueous solution of the complexes of 8-hydroxyquinoline (oxine) with hydrated magnesium(II) and manganese(II), and with the magnesium@) and manganese(I1) complexes of nitrilotriacetate (NTA), uramil-NN-diacetate (UDA), adenosine-5’-diphosphate (ADP), adenosine-5’-triphosphate (ATP) and polytri- phosphate (TP). The rate constants at 16°C have already been published for four of the Mg2+ systems (aquo, UDA, ATP and TP)4 and five of the Mn2+ systems 3738 KINETICS OF TERNARY COMPLEX FORMATION (aquo, NTA, UDA, ATP and TP).5 It is particularly important to determine the influence of bound ligands on the reactivity of a do metal ion such as Mg2+, because the latter, by virtue of the lack of directionality imposed on the metal-ligand bonds by partial covalency and the lack of asymmetry in the electrical field around the metal ion caused by a partially-filled d electronic sub-shell, provides a norm against which the results on other metals in which these factors are present may be compared.Mn2+ is also electronically spherosymmetrical since it has a high-spin d5 configura- tion; in addition, it is found that many enzyme systems which require Mg2+ will function in vitro with Mn2+.EXPERIMENTAL Tri-sodium adenosine-5'-diphosphate (Boehringer) was used without further purification. Details of other materials and the methods used for the kinetic and equilibrium measure- ment have been p~blished.~. s The data reported here refer to the relaxation effects in the 1-100 ms (Mg2+) and 10 ps-1 ms (Mn2+) regions associated with metal-oxine complex formation. Temperatures were accurate to &0.2"C. RESULTS The mechanism postulated 4 9 for the formation of the magnesium or manganese complex (MOx+) is k3 2 3 HOx + M2+ - MOxH2+ k23 (3) (4) k43 k34 kiz 11 kzi k4 1 H+ + Ox- + M2+ + MOx+ + H+. k14 This represents a two-path mechanism for the formation of MOx+ from the free metal M2+ in which the path involving the oxine anion Ox- is important at higher pH and the other, involving the oxine molecule HOx, is important at lower pH.Since H+ is buffered, M2+ is present in large excess over oxine and the deprotonation steps can be treated as rapid pre-equilibria, the observed relaxation time z is given by where c, and cH are the equillibrium concentrations of M2+ and H+, respectively, K2 = k43/k34 and K3 = k12/k2,. Similar expressions may be used in the ML+ oxine systems, where L is the firmly bound ligand NTA, UDA, ADP, ATP, or TP. (Charges have been omitted, and activity coefficients neglected. Eqn (1) is slightly different from the expressions used previ~usly.~. This is because in t h s paper we prefer to consider the general situation in which significant concentrations of the species MOxH2+ (and MLOxH) may exist.In most systems, however, the steady- state approximation may be applied and in this case eqn (1) reduces to the more restricted expression used previously since &CH < 1 .) the data were analyzed by plotting z-l against cH at low pH or against (1 +K2cH)-l at high pH (at a fixed M2+ or ML concentration), and by plotting 7-l against cM or cML at a fixed pH. It was sometimes possible to obtain a complete internal check on the mechanism and rate constants from the symmetrical nature of the pH profile of z-l. Although this method could have been used in most systems studied here, the data were actually analyzed with the help of a curve-fitting computer programme. Values of k41, k23K3, k32 and K3 were varied until the pH profile of 7-l calculated from eqn (1) fitted the profile indicated by the experimental points.The criterion used for the best fit was that the sum of the moduli of the percentage T-l = cM/(l +K2cH)(k41 +k3ZK2CH) + l/(l +K3CH){k14+k23K3CH), (1) PreviouslyY4*D. N. HAGUE, S. R. MARTIN AND M. S . ZETTER 39 IOOC rl 8CX L \ r( I 60( 40C I I I I t t 1 6.7 7.0 7.3 7-6 7.9 0.2 PH 7.7 8.0 8.3 8.6 8.9 9.2 9.5 PH FIG. 1.-Typical plots showing the variation of 2' with pH for the reaction of (a) Mg2+ (4 x 4O"C, where K3q&1) ; and (b) MgUDA- (8 x M, M, 40°C, where K~CH is not <l). Full lines on each point indicate experimental scatter.40 KINETICS OF TERNARY COMPLEX FORMATION deviations of the observed reciprocal relaxation times from the calculated values should be a minimum.Such a criterion was considered more reasonable than one based on a least-squares treatment since the latter would have placed too much weight on the higher values of 7-l. However, it was not possible to incorporate it into any statistical method for analyzing the errors associated with a particular rate constant and the uncertainties quoted are estimates. Two typical runs, together TABLE 1 .-RATE AND EQLJILIBRIUM CONSTANTS FOR THE REACTION OF MAGNESIUM SPECIES WITH OXINE (APPROX. 4 x M) AT VARIOUS TEMPERATURES Mg2+ 10.0 20.0 30.01 40.0 MgNTA- 10.0 20.0 30.0 40.0 MgUDA- 10.0 30.0 40.0 MgADP- 10.0 20.0 30.0 40.0 20.0 30.0 40.0 20.0 30.0 MgATP2- 10.0 MgTP3- 10.0 6.5-8.8 6.8-8.6 6.6-8.5 6.7-8.4 7.4-9.6 7.3-9.2 6.9-9.2 7.0-9.1 8.5-10.5 8.0-10.0 7.9-9.7 7.9-10.1 7.5-9.7 7.4-9.6 7.3-9.4 7.9-10.3 7.8-9.9 7.6-9.7 7.6-9.7 8.4-10.5 8.2-10.2 8.1-10.1 20 39 83 180 11 25 40 4.6 1.9 7.8 9.2 14 19 37 58 4.5 9.3 17 27 1.4 3.5 8.7 0.46 1.1 1.8 2.7 0.86 1.7 3.0 7.9 5.3 11 16 0.79 2.2 2.9 6.6 0.94 1.6 3.4 5.8 1.4 3.1 6.5 6.5 9.2 26 89 1 .o 2.5 6.8 10 43 230 430 32 37 120 1 80 15 35 61 140 12 35 4.4 k23K31 (1010 M-1 S-1) 0.14 0.16 0.23 0.39 0.17 0.23 0.34 0.57 110 130 1 40 2.3 2.3 3.6 5.1 2.6 3.4 4.4 7.7 3.6 6.0 9.7 The approximate concentration of the magnesium species was 8 x M except for Mg2+ and MgNTA- at 40T, where it was 4 x M.Ionic strength (KNO,) was 0.10 M for Mg2+, MgNTA- and MgUDA- and 0.30 M for MgADP-, MgATP2- and MgTP3-. Values of K3 (lo8 M-l) for MgUDA- were 8.8 (lO"C), 3.0 (30°C) and 2.O(4O0C).with their respective computed " best-lines " fits, are shown in fig. 1, and the rate and activation data are given in tables 1-4 (the fact that the 16°C rate constants given in tables 1 and 2 sometimes differ slightly from the ones quoted previously 4* is attributable to the different methods used to evaluate them ; the ones given here are considered to be the more reliable). The normal statistical adjustments have been made 4* to the formation rate constants, the activation entropies and the stability constants in order to allow for the different numbers of water molecules available for substitution by Ox- or HOx. k i l t ki2, etc., have been estimated on the assump- tion that the metal ions are hexahydrated and that in ML the ligand L is n-dentate (tables 3 and 4).The dependence of the dissociation constant of HOx on temperature was deter- mined spectrophotometrically. The method, which was similar to the one used previo~sly,~ employed a curve-fitting computer programme. The values of log K2 obtained at various temperatures were : 9.87 (16.0°C), 9.70 (24.0°C), 9.59 (32.0"C) and 9.46 (40.0"C).D . N. HAGUE, S . R . MARTIN AND M. S . ZETTER 41 TABLE 2.-RATE AND EQUILIBRIUM CONSTANTS FOR THE REACTION OF hUNGANESE(IT) SPECIES WITH OXINE (APPROX. 4 x M) AT vmous TEMPERATURES temp. approx. k41 k32/ system ("0 pH range (106 M-1 s-1) (106 M-1 s-1) Mn2+ 16.0 22.8 29.9 35.0 39.5 MnNTA- 7.0 16.0 28 .O 40.0 MnUDA- 7.0 16.0 28.0 40.0 MnADP- 16.0 20.0 26.0 32.5 36.0 MnATP2- 10.0 16.0 28.0 40.0 MnTP3- 10.0 16.0 28.0 40.0 5.3-7.2 5.2-7.2 5.0-7.0 5.0-6.9 5 .O-6.6 6.5-8.7 6.3-8.5 6.5-8.6 6.5-8.5 6.6-8.8 6.5-8.6 6.5-8.5 6.8-8.1 6.8-8 .O 7.1-8.0 6.9-7.9 7.0-7.9 7.5-8.9 7.1-9.0 7.3-9.2 7 .O-8.4 7.9-9.7 7.8-9.8 7.7-9.3 7.5-9.1 6.4-8.3 260 220 420 730 720 20 33 38 10 33 26 38 78 87 98 100 120 8.1 0.96 1.7 3.1 9.1 0.34 0.58 1 .o 2.0 0.31 0.41 0.57 0.90 1 .o 0.83 1.4 1.9 2.7 0.86 1 .o 1.5 12.2 1.3 1.2 1.3 1.9 1.8 0.56 0.58 1.4 2.3 0.24 0.23 0.46 0.72 k141 (s-1) 350 370 680 730 860 54 96 210 450 28 180 200 310 670 1800 2200 2100 3300 20 44 86 430 65 170 450 1400 k2 3K3l (1010 M-1 s-1) 0.30 0.33 0.38 0.31 0.36 5.9 4.9 5.5 9.1 2.6 4.0 4.9 5.3 8.8 15 14 14 16 11 15 27 39 44 80 120 9.9 The approximate concentration of the manganese(II) species was 1 .2 ~ M except for MnNTA- and MnUDA- at 28 and 40°C, where it was 6 x M.Ionic strength (KN03) was 0.10 M for Mn2+, MnNTA- and MnUDA-, 0.15 M for MnADP-, and 0.30 M for MnATP2- and MnTP3-. DISCUSSION The results are discussed under three main headings. The magnesium and manganese systems are considered in turn, and then some implications of the results on the behaviour of the metals in enzymic reactions are discussed. In the previous study with magnesium it was shown that the bound ligand had little effect other than reducing the solid angle of the metal ion over which effective substitution may occur. This implies that the local charge density at the surface of the metal ion rather than the overall charge on the complex is the factor which controls the reaction rate, and that this is not affected much by the presence of negatively-charged groups on the other side of the metal ion.The 25°C rates and activation parameters (table 3) confirm this. The activation enthalpies for the formation reactions (i.e., AHZl and AH&) are approximately the same, with the exception of the MgUDA-+ HOx system which is discussed separately. They are also close to the activation enthalpy for the water-exchange reaction (AH* = 10.2 kcal mol-', ref. (6)). Whereas previously 4* it was suggested that the rate-limiting step for the reaction of the oxine molecule might not be the loss of the first water molecule from the inner coordination sphere of the metal, Johnson and Wilkins7P N TABLE 3.-RATE AND EQUaIBRIUM CONSTANTS AND ACTIVATION PARAMETERS FOR THE REACTIONS OF MAGNESIUM SPECIES WITH OXINE (ESTIMATED ERRORS IN BRACKETS) The rate constants refer to 25°C at ionic strength 0.10 (Mg2+, MgNTA-, MgUDA-) or 0.30 (MgADP-, MgATP2-, MgTP3-), n is the number of coordination positions at Mg2+ assumed to be occupied by the ligand L.Mg2+ (6HzO) 6.0 (0.5)~ lo5 5.78 12.3 (0.6) +9 (2) 1.3 (0.2) x 104 4.11 9.6 (1.6) -7 (5) 20 ( 5 ) 15.0 (2.9) -2 (10) 4.48 MgNTA- (n = 4) 1.5 ( 0 . 2 ) ~ 105 +9 (3) 5.65 12.3 (0.9) 2.5 ( 0 . 3 ) ~ lo4 4.88 12.2 (1.3) f 5 (4) 3.7 (0.5) 13.3 (1.5) -11 (5) 5.08 MgUDA- (n = 4) 5.5 ( 0 . 4 ) ~ 104 5.23 11.2 (1.5) +3 (6) 9.3 (0.7)~ 104 5.45 5.7 (0.8) -14 (3) 150 (40) 13.0 (3.0) - 5 (10) 3.05 MgADP- (n = 2) 2.5 (0.3)~ lo5 5.58 10.3 (0.6) +2 (2) 2.5 ( 0 . 3 ) ~ lo4 4.58 11.2 (1.6) 0 (5) 73 (10) 10.4 (2.6) -15 (9) 3.72 MgATP2- ( n = 3) 1.2 ( 0 .3 ) ~ 105 0 (2) 2.4 (0.3) x lo4 4.68 10.3 (0.5) -2 (5) 48 (8) 12.0 (0.7) -11 (3) 5.38 10.0 (0.5) 3.70 MgTP3- (n = 3) 5.6 ( 0 . 5 ) ~ lo4 5.04 14.7 (1.5) + 14 (6) 4.5 ( 0 . 5 ) ~ 104 4.95 12.3 (1.3) +6 ( 5 ) 21 (6) 17.1 (3.0) + 5 (10) 3.72TABLE 4.-RATE AND EQUILIBRIUM CONSTANTS AND ACTIVATION PARAMETERS FOR THE REACTIONS OF MANGANESE(I1) SPECIES WITH OXINE (ESTIMATED ERRORS IN BRACKETS) The rate constants refer to 25°C at ionic strength 0.10 (MnZ+, MnNTA-, MnUDA-), 0.15 (MnADP-) or 0.30 (MnATP2-, MnTP3-). n is the number of coordination positions at Mn2+ assumed to be occupied by the ligand L. k14/(s- '> AH&/(kcal mol-') ASf4 /(cal mol-' K-l) log Ki (kinetic) Mn2+ (6H20) 3.3 (0.6)~ lo8 8.52 8.8 (4.6) + 10 (16) 4.9 ( 0 .7 ) ~ 105 5.69 9.1 (1.5) -2 (5) 480 (60) 7.0 (0.5) -22 (9) 5.84 MnNTA- (n = 4) 2.3 (0.3)~ lo7 7.84 6.6 (2.4) 0 (8) 1.7 (0.3)~ lo6 6.71 5.5 (0.7) -9 (2) 180 (30) 10.6 (2.2) - 12 (7) 5.58 MnUDA- (n = 4) 2.2 ( 0 . 3 ) ~ 107 7.82 6.4 (2.5) -1 (8) 1.4 (0.2) x lo6 6.62 4.5 (0.5) -13 (2) 130 (60) 12.4 (2.0) -7 (7) 5.71 MnADP- (n =i. 3) 8.5 ( 0 . 6 ) ~ lo7 8.23 4.4 (2.5) -6 (8) 1.4 ( 0 . 2 ) ~ lo6 6.45 4.0 (1.5) -15 (6) 1400 (700) 12.0 (3.0) -3 (9) 4.98 MnATP2- (n = 4) 2.5 ( 0 . 3 ) ~ lo6 6.88 12.0 (1.2) +I3 (4) 9.3 ( 0 . 8 ) ~ 105 6.45 8.3 (1.3) +1 (5) 90 (40) 16.4 (2.2) + 6 (8) 4.92 MnTP3- (n = 3) 9.1 (0.6) x 105 6.66 9.8 (0.6) + 5 (2) 4.0 ( 0 . 5 ) ~ lo5 5.90 6.4 (1.3) - 10 (4) 350 (100) 16.8 (1.3) + 10 ( 5 ) 4.1244 KINETICS OF TERNARY COMPLEX FORMATION found that HOx behaves as a “ typical ” ligand in its reaction with hydrated Ni2+.The data in tables 3 and 4 indicate that the differences between k i l and k& for a given ML are largely entropic in origin, and reaction (3)-(2) will therefore be taken as typical for complex formation at the MgL species. The larger value of AHzl for MgTP3- might be associated with a charge repulsion between TPs- and Ox-, although AH:4 is also higher for the MgTP3- system than for the others. The entropies of activation for the formation reactions and AS,’,cs)) are also similar (except for MgUDA-, see below) but the values for Mg2+ + HOx and MgTP3-+ Ox- are out of line. The values of AH* and AS* for the dissociation process show a similar slight fluctuation across the series but again there seems to be no trend which can be attributed directly to a repulsion between the oxine anion and the negatively- charged bound ligand L.Unfortunately, the experimental accuracy is insufficiently high to attempt a more detailed interpretation of the variation in activation parameters along the series. 0 I The reaction between MgUDA- and HOx is the only magnesium reaction reported here for which the activation parameters differ significantly from the rest. Although it is possible that the low value of AH* and large negative value of AS* are associated with an increased lability of the bound water molecules, as has been suggested for the reaction of CoNTA- with the neutral bidentate ligand pyridine-2-azo-p-dimethyl- aniline (PADA),8 we prefer to invoke an increase in the value of KO, (with correspond- ingly favourable AHos but unfavourable ASos) to explain the enhanced reactivity because the reduced AH& is not reflected in a change in AHzl or AHT4.If 11.5 kcal mol-1 is taken as the typical activation enthalpy for complex formation at magnesium and 0 cal mol-1 K-l as the corresponding statistically-corrected activation entropy for a neutral incoming ligand (these values being estimates based on the data in table 3) then, the difference in AH,, and ASos between the MgUDA- + HOx system and the rest will be, respectively, - 5.8 kcal mol-l and - 14 cal mol-l K-l. These correspond to a value of AGos which is more favourable by approx. 1.5 kcal mol-l than average-equivalent to an equilibrium constant of 10-15 M-l.It is possible,D. N. HAGUE, S . R. MARTIN AND M. S . ZETTER 45 using the value of K3 estimated from the computer fit of the relaxation data (fig. l(b) and table l), to calculate the value of K4 (= k32/k23) ; it has a similar magnitude to the value estimated from the increments in AHos and ASos. The question of the physical significance of these figures, which represent differences in AGos and KO, from normal, should not be pressed too hard in view of the fact that the exact mechanism of metal complex formation is still uncertain.l* However, the geometry of the barbiturate ring of UDA is such that a complex reminiscent of the adenine-thymine base-pair found in double-stranded DNA may be formed between MgUDA- and HOx. In this outer-sphere complex (I), the oxine molecule is conveniently placed for substitution at the metal ion.The values of AH* and the(second-order) rate constant at 25°C for water-exchange at Mn2+ are, respectively, 8.8 (k 1.0) kcal mol-l and approx. 3 x lo6 M-l s-l (ref. (9)) and it is assumed, as in the magnesium case, that reaction (3)-(2) may be taken as typical for complex formation at the MnL species. For this reaction with manganese(I1) there is essentially the same pattern (table 4) as has been found lo with zinc in its reaction with PADA : the activation enthalpy for the substituted metal is a few kcal mol-1 less than the value for the unsubstituted metal, but this difference in AH* is largely compensated by a corresponding difference in AS*. This is also found for the reaction of Ox- and the fact that the activation enthalpies for the two series are similar suggests that the variation in rate constant, etc., for different manganese species originates in the water-exchange process at the metal ion.We would therefore expect the activation enthalpies for the latter process in MnL to reflect the values of AH:, and AHZ2 (AHz2 for MnATP2- and MnTP3-). The general similarity between AH:, and AH,+, for a given MnL suggests that here, as with magnesium, there is usually no direct interaction between the negative groups on the bound ligand and those on the incoming ligand (although we attribute the high value of AH& for MnTP3- to electrostatic repulsion between the TP5- moiety and the Ox- anion, as for MgTP3-). No explanation is suggested for the trend in AH&, but it is opposite in direction to that predicted on the basis of a significant electrostatic repulsion between the bound L and Ox-. The low values of k i l for MnATP2- and MnTP3- noted in ref.(5) do not appear to be associated with the presence of phosphate groups as such, since MnADP- behaves in a similar way to MnNTA- and MnUDA-. There has been discussion concerning the nature of the binding between ATP and metal ions,ll and recent n.m.r. results suggest l2 that the nitrogen at the 7-position in the adenine ring is bound to Mn2+ through a complexed water molecule. It is possible that this outer- sphere interaction is responsible for the high value of AH& and contributes to the high value of AH& observed for the MnATP2- system. No such abnormally high values are found with the corresponding magnesium system (table 3).For any enzyme system, efficient turnover is dependent on obtaining a suitable balance between the time required for the enzyme-substratelproduct complex to form and the time taken for it to dissociate : the lifetime must be sufficiently long to permit the subsequent bond reorganization and possible conformational changes involving the ligands to take place, but it should not be so long as to tie up the active site unnecessarily. In part lY4 a kinetic model was discussed which could explain the frequent mutual antagonism of Ca2+ and Mg2+ and in part 2 it was shown that this model would also rationalize the frequent interchangeability of Mg2+ and Mn2+- neither of which can be explained simply in equilibrium terms.Essentially, if the catalytic pathway includes a conformational change of the enzyme which can only take place when the metal ion is bound, then the dissociation rate constant of the enzyme-metal complex must be less than the turnover number, which is typically46 KINETICS OF TERNARY COMPLEX FORMATION 102-103 s-I. (Although this was formulated for reactions which involve metal- bridged intermediates, it applies equally well for reactions in which the active form of the substrate is a metal complex.) Caution should be exercised in applying con- clusions drawn from model systems to enzyme reactions, but the results presented here confirm this rationalization. They also confirm the error in the widely-held view that ternary complexes form less readily and dissociate more easily than binaries because of the charge neutralization of the metal by the bound ligands or the electro- static repulsion between the two ligands.With magnesium, the activation enthalpy for the formation of the complex is abnormally high only in the most extreme case, where the bound ligand carries a charge of - 5 and the incoming ligand one of - 1. With manganese, the bound ligand exerts a greater influence on the reactivity of the metal ion ; however, it normally reduces AHZ1 rather than increases it. It is interesting that the reactivity of the MnATP2- complex (and, to a lesser extent, the MnTP3- com- plex) appears to be atypical of substituted manganese species in view of the fact that the magnesium/manganese interchangeability is prevalent in enzyme reactions which involve ATP. The only reaction reported here for which the activation parameters are markedly out of line is that between MgUDA- and HOx. This result illustrates one way in which an enzyme can accelerate a step in a reaction sequence. If the oxine molecule were part of the metal-enzyme (represented by MgUDA-) and complex formation between it and the metal ion was an elementary step in the reaction sequence, then the favourable AHos would not be compensated by an unfavourable ASos (since the outer-sphere complex would already exist) and the net metal- substitution rate would be even more enhanced than it is here. We are grateful to the S.R.C. for grants for maintenance (for S. R. M.) and the purchase of the temperature-jump apparatus, and to the University of Kent at Canterbury for a University Studentship (for M. S . Z.). e.g., M. Eigen and R. G. Wilkins, Mechanisms of Inorganic Reactions, R. F. Gould, ed., Adv. Chem. Series, no. 49 (Amer. Chem. SOC. Washington, D.C., 1965), p. 55 ; D. J. Hewkin and R. H. Prince, Coord. Chem. Rev., (Elsevier, Amsterdam), 1970,545. e.g., D. B. Rorabacher, Inorg. Chem., 1966,5,1891. H. P. Bennett0 and E. F. Caldin, J. Chem. SOC. A , 1971,2198. D. N. Hague and M. Eigen, Trans. Furaday Soc., 1966,62,1236. D. N. Hague and M. S. Zetter, Trans. Furaday SOC., 1970,66, 1176. J. Neely and R. Connick, J . Amer. Chem. SOC., 1970,92,3476. ' W. A. Johnson and R. G. Wilkins, Inorg. Chem., 1970,9,1917. * M. A. Cobb and D. N. Hague, Trans. Faraday SOC., 1971, 67, 3069. lo G. R. Cayley and D. N. Hague, Truns. Faruduy Soc., 1971,67,786. l 1 L. Rimai and M. E. Heyde, Biochem. Bipohys. Res. Cumm., 1970, 41, 313. See also ref. l2 T. A. Glassman, C. Cooper, L. W. Harrison and T. J. Swift, Biochemistry, 1971,10,843. M. Grant, H. W. Dodgen and J. P. Hunt, Inorg. Chem., 1971,10,71. quoted therein and in ref. (12).

ISSN:0300-9599    

DOI:10.1039/F19726800037

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Lattice Defects in Plastic Organic Crystals Part 7.-Isotope Mass Effects in Self-diffusion in Cyclohexane A. V. CHADWICK* AND J. N. SHERWOOD Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow, C. 1 ., Scotland Received 5th July, 1971 A study has been made of the diffusion of the species 12C514C11H12 and 12C62H113H1 in single crystals of cyclohexane at 242 K. The measured diffusion coefficients are in good agreement with previously published values. An estimate of the isotope mass effect yields a value of the mass factor E; = 0.003 f0.004. This virtually-zero value indicates that self-diffusion in cyclohexane proceeds by a complicated mechanism in which many molecules are involved ( >12). This discovery confirms previous speculations on the defect structure of the low entropy of fusion plastic solids.Recent studies of the processes of self-diffusion and plastic deformation in a series of rotator phase (plastic) and non-rotator phase (brittle) organic crystals have shown that for both types of solid high-temperature plastic deformation is likely to occur by a self-diffusion controlled, dislocation-climb, mechanism. The difference in plasticity between the two solid types is associable with the considerable difference in self-diffusion rates. This, in turn, is a consequence of some difference in the structure or nature of the mobile point defect in the solids. Analysis of the self- diffusion data has indicated the possibility of a structural differen~e.~ Recently it has been shown by isotope mass effect measurements that in benzene (a brittle crystal), self-diffusion occurs by the migration of lattice vacancies.In order to test the pre- vious speculations on the nature of point defects in plastic crystals, we have extended these measurements to the plastic crystal cyclohexane. For this substance the mass ratio of the two '' isotopes " is 1.128. This considerable increase with respect to benzene (1.062) should improve the accuracy of the separation of the diffusion coefficients of the two species. EXPERIMENTAL Cyclohexane- l-14C (specific activity 10 mCi mmol-') was purchased from the Radiochemical Centre, Amersham. Cyclohexane-d1 1-3H was synthesized from cyclohexanol-dI2 (Merck, Sharpe and Dohme 99+ %) by chlorination of the hydroxyl followed by hydrolysis of the corresponding Grignard's compound with tritiated water.g The product (0.6 g, specific activity 0.05 mCi mmol--') was dried and distilled.Attempts to prepare the deuterated tracer via chlorination of the deuterated hydrocarbon with N-chloro succinimide l o were unsuccessful. It proved impossible to prepare the Grig- nard's compound of the resulting chloride, although trial syntheses with the normal hydro- carbon were successful. Diffusion experiments were carried out at 242 K in the manner described previ~usly.~ Since the mercury encapsulant, which was essential to the accurate performance of this experiment, freezes at 234K it was not possible to carry out experiments at significantly lower temperatures. Attempts to dispense with the mercury were unsuccessful since the * present address : The University Chemical Laboratories, The University of Kent, Canterbury. The purification and preparation of the crystals is described elsewhere.6* 4748 ISOTOPE EFFECTS I N SELF-DIFFUSION I N CYCLOHEXANE combination of the high specific activity of the tracers, plus the high volatility of cyclohexane led to contamination problems.Because of the nature of the results, experiments at higher temperatures would have yielded no additional information. RESULTS AND DISCUSSION The geometry of source and sample used in the experiment approximate to that of diffusion from a thin source of total activity Q into a semi-infinite solid. For this geometry, the integration of Fick's law yields an expression for the concentration (specific activity) A of the diffusing species at distance x into the crystal after time t of the form, D, is the diffusion coefficient of the tracer.The applicability of this equation to the description of the present experimental data is shown in fig. 1. The average value of A = Q/(nDxt)* exp(-x2/4D,t). (1) x2/104 pm2 FIG. l.-Typical plots of the logarithmic form of eqn (1) for experiments 3 (0) and 4 (0). 0, evaluated graphically from the separate experiments is given in table 1. This value is in good agreement with a value predicted from previously published data (table 1). TABLE RESULTS OF THE ISOTOPE EFFECT EXPERIMENTS experiment mean Dx* number (Da/Db- 1) /m2 s-1 1 - 0.005 f0.005 2 0.010 *0.005 3 0.005 f0.004 4 0.003 f0.004 7.33 x 10-13 * cf. Dx = 9.33 x m2 s-l (Hood and Sherwood) ' D, represents the correlated motion of the tracer in the lattice.This coefficient is related to the random self-diffusion coefficient D by an equation D, = fD. (2) the correlation factor, is a constant which depends on the mechanism by which the diffusion occurs and the geometry of the crystal lattice. The experimental evaluationA. V. CHADWICK AND J . N. SHERWOOD 49 offand the comparison of this value with theoretical values for the different possible mechanisms can lead to the definition of the mechanism of diffusion.ll For the simultaneous diffusion of two isotopic species of mass mu and mb in a lattice of molecules of mass m, the two diffusion coefficients D, and Db are related by the expression, 2-1 n is the number of molecules which change positions in the diffusive jump, i.e., n = 1 €or a normal vacancy process.The " mass factor ", E$ = fAK. AK is a term which allows for the disturbance caused in the surrounding lattice during the jump process. If there is no disturbance, then AK = 1. As the coupling between the migrating molecules and the neighbouring molecules increases, thus causing some motion in the surrounding lattice, AK+O. For a simple mechanism, viz., a single molecule jump which does not affect the surrounding lattice, eqn (3) reduces to l2 andfcan be evaluated from the experimental determination of the ratio of the diffusion coefficients for the isotopic species. Since the rate of distribution in the solid of both isotopes will be governed by eqn (l), we can compound the two forms of the equation to yield l 5 (DtzlDb- 1) = f[(mb/%)*- 11, (4) I .55 1.45 n \ 3 D s - en 4 1.35 1.25 I A t ' ( 3: I t 1 4 (2: log10 A, FIG.2.-Plots of loglo(A,/Ab) against loglo A, (eqn (5)) for all experiments. The dashed lines at the head of the figure represent the expected behaviour for many (n) body processes assuming that f= I,A.K= 1.50 ISOTOPE EFFECTS I N SELF-DIFFUSION I N CYCLOHEXANE Thus, comparison of the relative specific activities of the two isotopic species at different depths into the crystal, after the completion of a diffusion experiment in which both isotopes are diffused simultaneously, yields (D,/Db- 1). Fig. 2 shows several plots of the experimental data in the form of eqn (5). An estimate of the total error in each experiment is given in table 1 which summarizes all the data.Within a small error the lines have zero slope. This result is completely different from the previously reported results for benzene.5 Since the mass ratio in the present case is much more favourable than for the benzene experiment we conclude that this result is meaningful. Reference to eqn (3) shows that for the ratio (D,/Db- 1) to be approximately zero, then the factorsf, or AK, must approach zero, or n must be large. For the f.c.c. lattice, f for the most normal mechanisms (including a multi-molecule exchange process) is greater than 0.48 and it seems unlikely that this is the cause of the anomaly. Thus, we have either a simple mechanistic process (n = 1) in which the surrounding lattice is considerably involved (AK N O), or a multimolecular process (n> I), or what is more probable in the circumstances, a combination of the two, i.e., a co-operative process involving a number of molecules in which the surrounding lattice participates.If we accept thatfAK = 1 then only 12-18 molecules on average need to be considered as the maximum number involved in the jump process to yield an approximately zero value for (Du/Db-l). Thus, even at the worst we should expect only a very localized complete disorder in the solid and hence more probably a limited disorder over only a few neighbouring sites. Such a disordered defect could be highly mobile and hence could yield the more rapid self-diffusion found in these solids. It could also lead to the difference found between radiotracer and n.m.r.studies of translational motion in plastic solids of low entropy of fusion.' We conclude that self-diffusion in cyclohexane proceeds by a co-operative motion involving either a complicated defect or a vacancy around which the lattice is dis- ordered. A. V. C. thanks Imperial Chemical Industries Ltd. for the award of an I.C.I. Fellowship and J. N. S . the Science Research Council for the award of an apparatus grant. Materials were supplied, in part, through the Science Research Council Sponsored Organic Crystal Laboratory at this University. H. M. Hawthorne and J. N. Sherwood, Trans. Farahy Soc., 1970,66, 1783. N. T. Corke and J. N. Sherwood, J . Materials Sci., 1971,6,68. J. N. Sherwood, Bull. SOC. Franc. Min. Cryst., in press. A. V. Chadwick and J.N. Sherwood, Difusion Processes, ed. J. N. Sherwood, A. V. Chad- wick, W. M. Muir and F. L. Swinton (Gordon and Breach, London, 1971), p. 475. R. Fox and J. N. Sherwood, Trans. Faraduy SOC., 1971, 67, 3364. G. M. Hood and J. N. Sherwood, Brit. J. AppI. Phys., 1963,14,215. J. F. Norris and R. S. Mulliken, J. Amer. Chem. Soc., 1920,42,2093. L. Melander, Arkiv Kemi, 1950,2,260. ' G. M. Hood and J. N. Sherwood, Mof. Cryst., 1966, 1,97. lo N. Ph. Buu Hoi and P. Demerseman, J. Org. Chem., 1953,18,649. l 1 A. D. LeClaire, Physical Chemistry-An Advanced Treatise, VoI. 10-The Solid State, ed. l2 A. Schoen, Phys. Rev. Letters, 1958, 1, 138. l3 J. G. Mullen, Phys. Rev., 1961,121, 1649. l4 A. D. LeClaire, Phil. Mag., 1966, 14, 1271. l5 L. W. Barr and A. D. LeClaire, Proc.Brit. Cerum. SOC., 1964,1, 109. l6 P. Bladon, N. C. Lockhart and J. N. Sherwood, MoZ. Phys., 1971,20, 577. H. Eyring, D. Henderson and W. Jost (Academic Press, N.Y., 1970), p. 261. Lattice Defects in Plastic Organic Crystals Part 7.-Isotope Mass Effects in Self-diffusion in Cyclohexane A. V. CHADWICK* AND J. N. SHERWOOD Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow, C. 1 ., Scotland Received 5th July, 1971 A study has been made of the diffusion of the species 12C514C11H12 and 12C62H113H1 in single crystals of cyclohexane at 242 K. The measured diffusion coefficients are in good agreement with previously published values. An estimate of the isotope mass effect yields a value of the mass factor E; = 0.003 f0.004. This virtually-zero value indicates that self-diffusion in cyclohexane proceeds by a complicated mechanism in which many molecules are involved ( >12).This discovery confirms previous speculations on the defect structure of the low entropy of fusion plastic solids. Recent studies of the processes of self-diffusion and plastic deformation in a series of rotator phase (plastic) and non-rotator phase (brittle) organic crystals have shown that for both types of solid high-temperature plastic deformation is likely to occur by a self-diffusion controlled, dislocation-climb, mechanism. The difference in plasticity between the two solid types is associable with the considerable difference in self-diffusion rates. This, in turn, is a consequence of some difference in the structure or nature of the mobile point defect in the solids.Analysis of the self- diffusion data has indicated the possibility of a structural differen~e.~ Recently it has been shown by isotope mass effect measurements that in benzene (a brittle crystal), self-diffusion occurs by the migration of lattice vacancies. In order to test the pre- vious speculations on the nature of point defects in plastic crystals, we have extended these measurements to the plastic crystal cyclohexane. For this substance the mass ratio of the two '' isotopes " is 1.128. This considerable increase with respect to benzene (1.062) should improve the accuracy of the separation of the diffusion coefficients of the two species. EXPERIMENTAL Cyclohexane- l-14C (specific activity 10 mCi mmol-') was purchased from the Radiochemical Centre, Amersham.Cyclohexane-d1 1-3H was synthesized from cyclohexanol-dI2 (Merck, Sharpe and Dohme 99+ %) by chlorination of the hydroxyl followed by hydrolysis of the corresponding Grignard's compound with tritiated water.g The product (0.6 g, specific activity 0.05 mCi mmol--') was dried and distilled. Attempts to prepare the deuterated tracer via chlorination of the deuterated hydrocarbon with N-chloro succinimide l o were unsuccessful. It proved impossible to prepare the Grig- nard's compound of the resulting chloride, although trial syntheses with the normal hydro- carbon were successful. Diffusion experiments were carried out at 242 K in the manner described previ~usly.~ Since the mercury encapsulant, which was essential to the accurate performance of this experiment, freezes at 234K it was not possible to carry out experiments at significantly lower temperatures.Attempts to dispense with the mercury were unsuccessful since the * present address : The University Chemical Laboratories, The University of Kent, Canterbury. The purification and preparation of the crystals is described elsewhere.6* 4748 ISOTOPE EFFECTS I N SELF-DIFFUSION I N CYCLOHEXANE combination of the high specific activity of the tracers, plus the high volatility of cyclohexane led to contamination problems. Because of the nature of the results, experiments at higher temperatures would have yielded no additional information. RESULTS AND DISCUSSION The geometry of source and sample used in the experiment approximate to that of diffusion from a thin source of total activity Q into a semi-infinite solid.For this geometry, the integration of Fick's law yields an expression for the concentration (specific activity) A of the diffusing species at distance x into the crystal after time t of the form, D, is the diffusion coefficient of the tracer. The applicability of this equation to the description of the present experimental data is shown in fig. 1. The average value of A = Q/(nDxt)* exp(-x2/4D,t). (1) x2/104 pm2 FIG. l.-Typical plots of the logarithmic form of eqn (1) for experiments 3 (0) and 4 (0). 0, evaluated graphically from the separate experiments is given in table 1. This value is in good agreement with a value predicted from previously published data (table 1).TABLE RESULTS OF THE ISOTOPE EFFECT EXPERIMENTS experiment mean Dx* number (Da/Db- 1) /m2 s-1 1 - 0.005 f0.005 2 0.010 *0.005 3 0.005 f0.004 4 0.003 f0.004 7.33 x 10-13 * cf. Dx = 9.33 x m2 s-l (Hood and Sherwood) ' D, represents the correlated motion of the tracer in the lattice. This coefficient is related to the random self-diffusion coefficient D by an equation D, = fD. (2) the correlation factor, is a constant which depends on the mechanism by which the diffusion occurs and the geometry of the crystal lattice. The experimental evaluationA. V. CHADWICK AND J . N. SHERWOOD 49 offand the comparison of this value with theoretical values for the different possible mechanisms can lead to the definition of the mechanism of diffusion.ll For the simultaneous diffusion of two isotopic species of mass mu and mb in a lattice of molecules of mass m, the two diffusion coefficients D, and Db are related by the expression, 2-1 n is the number of molecules which change positions in the diffusive jump, i.e., n = 1 €or a normal vacancy process.The " mass factor ", E$ = fAK. AK is a term which allows for the disturbance caused in the surrounding lattice during the jump process. If there is no disturbance, then AK = 1. As the coupling between the migrating molecules and the neighbouring molecules increases, thus causing some motion in the surrounding lattice, AK+O. For a simple mechanism, viz., a single molecule jump which does not affect the surrounding lattice, eqn (3) reduces to l2 andfcan be evaluated from the experimental determination of the ratio of the diffusion coefficients for the isotopic species.Since the rate of distribution in the solid of both isotopes will be governed by eqn (l), we can compound the two forms of the equation to yield l 5 (DtzlDb- 1) = f[(mb/%)*- 11, (4) I .55 1.45 n \ 3 D s - en 4 1.35 1.25 I A t ' ( 3: I t 1 4 (2: log10 A, FIG. 2.-Plots of loglo(A,/Ab) against loglo A, (eqn (5)) for all experiments. The dashed lines at the head of the figure represent the expected behaviour for many (n) body processes assuming that f= I,A.K= 1.50 ISOTOPE EFFECTS I N SELF-DIFFUSION I N CYCLOHEXANE Thus, comparison of the relative specific activities of the two isotopic species at different depths into the crystal, after the completion of a diffusion experiment in which both isotopes are diffused simultaneously, yields (D,/Db- 1). Fig.2 shows several plots of the experimental data in the form of eqn (5). An estimate of the total error in each experiment is given in table 1 which summarizes all the data. Within a small error the lines have zero slope. This result is completely different from the previously reported results for benzene.5 Since the mass ratio in the present case is much more favourable than for the benzene experiment we conclude that this result is meaningful. Reference to eqn (3) shows that for the ratio (D,/Db- 1) to be approximately zero, then the factorsf, or AK, must approach zero, or n must be large. For the f.c.c. lattice, f for the most normal mechanisms (including a multi-molecule exchange process) is greater than 0.48 and it seems unlikely that this is the cause of the anomaly.Thus, we have either a simple mechanistic process (n = 1) in which the surrounding lattice is considerably involved (AK N O), or a multimolecular process (n> I), or what is more probable in the circumstances, a combination of the two, i.e., a co-operative process involving a number of molecules in which the surrounding lattice participates. If we accept thatfAK = 1 then only 12-18 molecules on average need to be considered as the maximum number involved in the jump process to yield an approximately zero value for (Du/Db-l). Thus, even at the worst we should expect only a very localized complete disorder in the solid and hence more probably a limited disorder over only a few neighbouring sites.Such a disordered defect could be highly mobile and hence could yield the more rapid self-diffusion found in these solids. It could also lead to the difference found between radiotracer and n.m.r. studies of translational motion in plastic solids of low entropy of fusion.' We conclude that self-diffusion in cyclohexane proceeds by a co-operative motion involving either a complicated defect or a vacancy around which the lattice is dis- ordered. A. V. C. thanks Imperial Chemical Industries Ltd. for the award of an I.C.I. Fellowship and J. N. S . the Science Research Council for the award of an apparatus grant. Materials were supplied, in part, through the Science Research Council Sponsored Organic Crystal Laboratory at this University. H. M. Hawthorne and J. N. Sherwood, Trans. Farahy Soc., 1970,66, 1783. N. T. Corke and J. N. Sherwood, J . Materials Sci., 1971,6,68. J. N. Sherwood, Bull. SOC. Franc. Min. Cryst., in press. A. V. Chadwick and J. N. Sherwood, Difusion Processes, ed. J. N. Sherwood, A. V. Chad- wick, W. M. Muir and F. L. Swinton (Gordon and Breach, London, 1971), p. 475. R. Fox and J. N. Sherwood, Trans. Faraduy SOC., 1971, 67, 3364. G. M. Hood and J. N. Sherwood, Brit. J. AppI. Phys., 1963,14,215. J. F. Norris and R. S. Mulliken, J. Amer. Chem. Soc., 1920,42,2093. L. Melander, Arkiv Kemi, 1950,2,260. ' G. M. Hood and J. N. Sherwood, Mof. Cryst., 1966, 1,97. lo N. Ph. Buu Hoi and P. Demerseman, J. Org. Chem., 1953,18,649. l 1 A. D. LeClaire, Physical Chemistry-An Advanced Treatise, VoI. 10-The Solid State, ed. l2 A. Schoen, Phys. Rev. Letters, 1958, 1, 138. l3 J. G. Mullen, Phys. Rev., 1961,121, 1649. l4 A. D. LeClaire, Phil. Mag., 1966, 14, 1271. l5 L. W. Barr and A. D. LeClaire, Proc. Brit. Cerum. SOC., 1964,1, 109. l6 P. Bladon, N. C. Lockhart and J. N. Sherwood, MoZ. Phys., 1971,20, 577. H. Eyring, D. Henderson and W. Jost (Academic Press, N.Y., 1970), p. 261.

ISSN:0300-9599    

DOI:10.1039/F19726800047

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Kinetics of Anionic Polymerization of o-Methylstyrene in 2-Methyltetrahydrofuran and Tetrahydrofuran * BY HIDEO HIROHARA, MASATOSHI NAKAYAMA, RYOICHI mWABATA AND NORIO ISE Dept. of Polymer Chemistry, Kyoto University, Kyoto, Japan Received 26th July, 1971 The kinetics of anionic polymerization of o-methylstyrene were investigated in 2-methyltetra- hydrofuran (MTHF) at 25°C with Li+, Na+, K+ and Cs+ as gegenions in the presence and absence of an electric field. Conductance studies showed that the dissociation constants K vary remarkably with the gegenion and that the lithium and sodium salts are dissociated to larger extents than the respective polystyryl salts. The free ion rate constant k$ was -2 X lo4 M-' s- which was smaller than that of polystyrene. From the conductivity data, the field effect was concluded to be due to an increase in k i with increase of the field strength as in other systems.The k i reached a limiting value of w3 x lo4 M-l s-' above 3 kVlcm. The ion pair rate constants kb were 2, 1, 16 and 30 M-' s-l for Li+, Na+, K+ and CS+, respectively. Comparison with polystyrene systems led to the conclusion that the behaviour of poly(o-methylstyryl) salts are due to steric hindrance effects of the methyl group in the ortho position. The field-accelerating effects reflected in the k i term as in other systems are smaller than that of poIystyrene systems. This is discussed in terms of the ortho effect. The kinetics of the caesium salt of poly(o-methylstyrene) were also studied in tetrahydrofuran (THF) at 25°C. The results are compared with those in MTHF.In a previous paper we reported briefly on anionic polymerizations of poly(0- and p-methy1styrene)s in MTHF.l The study revealed that poly(o-methylstyrene) is less reactive and more strongly dissociated than poly(p-methylstyrene). The work was, however, limited to sodium salts in MTHF in the absence of an electric field. We report here an anionic polymerization study of poly(o-methylstyrene) with Li+, Na+, Kf and Cs+ as gegenions in MTHF together with the caesium salt in THF. Furthermore, it is of interest to apply an electric field to the systems of monomers having a substituent group in the vicinity of the active centre as a continuation of our research work on high intensity electric field effect on polymerization.2 EXPERIMENTAL The detailed techniques and purification of reagents have been described in earlier paper^.^-^ More recently we noted that the purity of reagents is one of the most crucial factors in an accurate determination of the rate constant, and that the influence of impurities on the rate constants is remarkable in MTHF.Therefore, extreme care was exercised in purification of the materials. Commercial 0-methylstyrene (0-MS) (Monomer-Polymer Lab.) was purified by a method similar to that for styrene and p-methoxystyrene.6 The polymerizations of the lithium and caesium salts were started by the corresponding polystyryl salts which were obtained from ethyllithium and cumylcaesium ; for the sodium and potassium salts, sodium and potassium a-methylstyrene tetramer dianions, respectively, were used as initiators.Except for ethyllithium, the initiators were prepared in THF and then the solvent was replaced by MTHF. The replacemeni was confirmed by gas chromatography. The lithium and caesium salts were of the oneended type; the others were twoended. * This article is part 19 of Ionic Polymerization under an Electric Field. 5152 ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE Reaction rates were measured at 25°C using a Hitachi EPS-3T spectrophotometer provided with a temperature-controlled cell holder ; the reactions were monitored by follow- ing the disappearance of the o-methylstyrene absorption at 298.2 mp. The initial concentra- tion of o-methylstyrene was usually 25-50 times greater than the active centre concentration, which was also determined spectrophotometrically using the poly(o-methylstyryl) anion absorption band (& : Li, 329 mp ; Na, 338 mp ; K, 340 mp ; Cs, 343 mp).The molar extinction coefficients were assumed to be 1.3 x lo4 M-' cm-1 for all four salts in the light of the extinction coefficient data of living polystyrenes and other polymers in ethereal solvents. The electric conductance was calculated from the applied voltage and the current passing through the polymerizing solution. The conductances of all four salts were practically independent of time after application of the electric field when elaborately purified MTHF was used as the solvent. The electric field was applied in a conversion range between 20 and 75 %; the rate constant was also calculated in the same conversion range.RESULTS The apparent propagation rate constants k, in the presence and absence of an electric field for all four salts in MTHF are presented in the form of the (kp, l/[LE]+) plots in fig. 1 and 2, where [LEI is the total concentration of active centres. The intercept gives the ion-pair rate constant kk and the slope kgK4, where kg is the free ion rate constant and K the ion-pair dissociation constant. Except for the caesium salt the slope for which was zero, a high intensity electric field increased the slope, whereas the intercept remained unaffected as previously observed in various sy~terns.~'~ 0 50 100 150 200 1 / d [LEI FIG. 1 .-Dependence of the apparent propagation rate constant on the poly(o-methylstyry1)lithium and poly(o-methylstyry1)potassium concentration in MTHF at 25°C. The kinetic data and field effects obtained from fig.1 and 2 are collected in table 1. The ki values of both the lithium and sodium salts seem to be practically equal and much smaller than those of the respective salts of polysytrene (25 and 23 M-l s-l for lithium and sodium salts, respecti~ely).~ The reactivity of the potassium ion pair isH. HIROHARA, M. NAKAYAMA, R. KAWABATA A N D N. ISE 53 comparable to that of polystyrene and the kb value of the caesium salt appeared to be larger than that of polystyrene (23 M-’ s - ’ ) . ~ The k%K4 values at 0 kV/cm decreased \ @ 50 0 50 100 150 200 1 /1/[LEI FIG. 2.-Dependence of the apparent propagation rate constant on the poly(o-methylstyry1)sodium and poly(o-methylstyry1)caesium concentration in MTHF at 25°C.@ : from ref. (1). in the order Lif > Naf > K+ > Cs+, and are much smaller than those of polystyryl salts. The field acceleration factor (given in the fifth column) reached a limiting value (1.3 - 1.4) at E = 2- 3 kV/cm and is smaller than that for polystyrene (1.6 N 1.7).5 TABLE KINETIC RESULTS AND FIELD EFFECTS ON THE ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE IN MTHF AT 25°C kgEK% E k;, kgK3 gegenion /kV cm-1 /M-1 s-1 lM-& s-1 k&, K$ Li+ 0 2 1.06 1 0.6 2 1.17 1.1 1.3 2 1.42 1.3 2.3 2 1.56 1.5 1.5 1 0.60 1.2 3.0 1 0.68 1.3 4.5 1 0.68 1.3 1.5 16 0.30 1.2 3.0 16 0.34 1.3 4.5 16 0.34 1.3 cs+ 0 30 0 - 3.0 30 0 Na+ 0 1 0.51 1 K+ 0 16 0.26 154 - 1.4 -1.6- ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE - - 1.0 \ 0 0 cs(l)o ooo The equivalent conductances A for the four salts are shown in fig.3. We note that the A values of each salt fall on a line irrespective of the applied field strength. Thus all the data are shown with blank circles without specifying the field strength. f2A[LE]/Fx lo6 20 f2A[LE]/Fx lo5 FIG. 4.-Fuoss plots of the conductance of living poly(o-methylstyrene) in MTHF at 25°C.H. HIROHARA, M. NAKAYAMA, R . KAWABATA A N D N. ISE 55 The lower concentration part for the four salts gave straight lines of a slope of -3. For potassium and caesium salts, however, the slopes became larger at higher con- centrations. In fig. 4, the conductances for lithium, sodium and potassium salts TABLE 2.-DISSOCIATION CONSTANT AND FREE ION RATE CONSTANT OF LIVING POLY(0-METHYL- STYRENE) IN MTHF AT 25” K x lO*O/M K(o-MeSt) k$ x lO-4/M-1 s-1 St b K(St) o-MeSt St A0 = gegenion /cmz mho-1 o-MeSt Li+ 58 25 8.7 3 2.1 6.8 Na+ 58 5.5 2.4 2 2 .2 7.4 K+ 58 3.6 3.4 1 1.4 7.1 7.3 cs+ 75 0.04 2.3 =O a assumed : see text for details b from ref. (5). - were plotted according to the method of FUOSS.~~ To evaluate K, we used the limit- ing conductance A. of polystyryl salts estimated in the previous work s assuming that the effect of the methyl group on A. is neghgible. The K values are given in TABLE 3.-KINETIC RESULTS OF THE CAESIUM SALTOF POLY(0-METHYLSTYRENE) IN THF AT 25°C k;, kp”& A0 K x lo9 k g x 10-4 monomer /M-1 s-1 /M-3 s-1 /cmz mho-1 /M /M-1 s-1 O-MS 15 f5 1.7 92.5 1.8 4 styrene” 20f5 6.8 92.5 2.7 13 b from Smarc’s data in ref.(1 1). from ref. (7) table 2, together with those of polystyryl salts. The K value, in conjunction with the slopes of the (kp, I/[LE]*) plot, allows us to calculate k;. The results are also given in table 2. Experiments were also performed for caesium salts in THF. The kinetic data are represented in table 3. DISCUSSION Alkali poly(o-methylstyryl) salts with the exception of the caesium salt are more strongly dissociated than corresponding polystyryl salts. The K values of the former were strongly dependent on the gegenion and decreased along the series Li+ > Na+ > K+ > Cs+, but those of the four salts of the latter polymer were not greatly different from each other. The ratio of the K values of the two polymers decreases along the series Li+ > Na+ > K+ > Cs+.For poly(p-methoxystyrene) in THF and poly(p- methylstyrene),12 the order of the ratio was reversed. These polymer salts were less dissociated than polystyryl salts. These facts are inexplicable in terms of the electron-donating nature of the methyl group, but suggest that the methyl group in the vicinity of the active centre prevents the gegenion from re-approaching the nega- tively charged free ion, except in the caesium case. This steric hindrance would depend on the effective ionic radius of the gegenion. It is believed that the lithium ion is strongly solvated and the caesium ion is hardly solvated by MTHF molecules. The solvated lithium ion having a large effective ionic radius would be effectively hindered from association with the free anion. This might account for much larger K value of the lithium poly(o-methylstyrene) than that of the corresponding salt of56 ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE polystyrene.The situation would be similar for the sodium salt. For the caesium salt, on the other hand, the methyl group might prevent the caesium ion from dis- sociating as a consequence of the small effective ionic radius. The interionic distance a in the ion pairs can be calculated from the K value using the “ sphere in continuum ” model which gives the Fuoss eqn l3 K = (3000/4nNa3)exp( - e2/aDkT). The a value for -(o-MS)-, Cs+ ion pairs was found to be 3.1 A, whereas it is 3.7A for -S-, Cs+ ion pairs in MTHF,’ where w(o-MS)- and -S- are poly(o- methylstyryl) anion and polystyryl anion, respectively. The Stokes radius for the caesium ion in MTHF is probably 2.2-2.4 A in the light of the data in THF.14 From these, we have = 1.4 A for the radius of the polystyrylanion in MTHF.If poly(o- methylstyryl) anion has the same radius, the radius of the cation for the -(o-MS)-, Cs+ ion pair in MTHF is estimated to be M 1.7 A, which is practically equal to the crystal radius of the caesium ion. This might suggest that in the formation of the -(o-MS)-, Cs+ ion pair, the MTHF molecules weakly bound to the caesium ion are removed as a result of the high electron density on the carbanion due to the electron- donating nature of the methyl group. On the other hand, in THF, the solvating power of which is stronger than that of MTHF, this situation might not be en- countered.As a matter of fact, the caesium poly(o-methylstyrene) is slightly less dissociated in THF than the caesium polystyrene (see table 3). Thus, the ortho effect is very sensitive to the solvation state. The methyl group in the ortho position also prevents monomer from reaching the active centre by the steric effect, in addition to the electron-donating nature of the methyl group. Thus, the kk and k i values are much smaller than the respective values of poly(m- and p-methy1styiene)s ’* l2 as well as polystyrene.’ The free - (0-MS)- ion is 3-$ as reactive as the free - 5 ion not only in MTHF but also in THF. This behaviour is reminiscent of that found for living poly(2-vinylpyridine) and poly(o-methoxystyrene),l although the intramolecular “ solvation ” of the cation took place in these cases.For the field effect, we conclude that, since A was insensitive to the field strength, i.e., K did not increase with the field strength, augmentation in k: is a primary factor for the observed effect and may be attributed to removal of the solvent molecules weakly bound to the free carbanion by the electric field, as suggested in other sys- tems. 3-6 The observed field-acceleration effects (kiE/k%,) were smaller than those for polystyrene systems, suggesting that solvation of the negatively charged active centre of poly(o-methylstyrene) is partly hindered sterically. The kLE value of 3 x lo4 M-l s-I, which was obtained from the K value and the kg K*, is interpreted as the “de- solvated ’’ free ion rate constant for living poly(o-methylstyrene).Finally, there should be discussion of the slope of the (log A, log[LE]) plots in potassium and caesium systems. Deviation of the slope from -4 toward positive values was reported for sodium poly(2- and 4-vinylpyridine)~ in THF, ’ alkali poly(methylmethacry1ate)s in THF, * and lithium polystyrene in benzene-dimethoxy- ethane mixtures.19 No explanation is offered for this phenomenon except in the polystyryl lithium case. In this case, intermolecular triple ion formation was pro- posed. In the present system also, we can tentatively assume that triple ions exist in addition to free ions and ion pairs. Negatively charged triple ions -(o-MS)-, (+)--? (0-MS)-- are much more stable than positive ones (+I, (o-MS)--, (+) as discussed by Fuoss 2o and Wooster.21 We note here that the Syracuse group was the first to propose formation of triple ions,22 which were, however, of an intra- molecular type.Following the treatment of Wooster 21 for the unilateral triple ionH . HIROHARA, M. NAKAYAMA, R . KAWABATA A N D N . ISE 57 formation, we calculated approximate dissociation constants for the triple ion k and the ion pair K. The results are as follows. For the caesium salt, there can exist about 15 times as many triple ions as free ions at [LEI = 1 x M and K = 8 x 10-13 M. For the potassium salt, about 3 times as many triple ions as free ions exist at [LEI = 1 x M and K = 1.5 x 10-lo M, which, by combining with the slope of the (kp, l/[LE]*) plot, leads to Pi = 2.1 x lo4 M-l s-' . This value is in excellent agreement with the k i values from the other two salts (see table 2).We note that positive slopes were observed in the (kp, 1/[LEIB) plot, though three ionic species (free ions, ion pairs and triple ions) contribute to the propagation process. The positive slope implies kg (triple ion rate constant) < k; in any case. This is quite reasonable since triple ions for the potassium and caesium salts would be of the contact type as suggested by Szwarz et aZ.22 Furthermore, contributions from the triple ions would explain the apparently large kk value for the -S-, K+ and -S-, Cs+ ion pairs. Although we feel compelled to draw attention to contribution of the triple ions, since it fits the results mentioned above, we are unhappy with this inter- pretation. It is rather difficult to understand why the triple ion is formed for living poly(o-methylstyrene) in spite of the fact that no triple ion was formed for living polystyrene in the same solvent.In conclusion, we emphasize that the steric hindrance effect depends on the effective size of the gegenion and that the kinetic study of ortho-substituted living polymer can be useful for obtaining information on solvation of active centres in polymerization systems. M. Nakayama, H. Hirohara, K. Takaya and N. Ise, J. Polymer Sci., A-1,1970,8,3653. N. Ise, Ado. Polymer Sci., 1969, 6, 347, in which previous work is referred to. N. Ise, H. Hirohara, T. Makino and I. Sakurada, J. Phys. Chem., 1968, 72,4543. H. Hirohara, M. Nakayama, K. Takaya and N. Ise, Trans. Fmaday Soc., 1970,66,1165.H. Hirohara, K. Takaya, M. Nakayama and N. Ise, Trans. Farahy SOC., 1970,66, 3163. K. Takaya, H. Hirohara, M. Nakayama and N. Ise, Trans. Faraday SOC., 1971,67, 119. H. Hirohara, K. Takaya, N. Ise, Macromolecules, 1971,4,288. K. Ziegler and H. Dislich, Ber., 1957, 90, 1107. (b) R. M. Fuoss and F. Accascina, ElectroZytic Conductance (John Wiley and Son., Inc., New York, 1959), chap. 17. ' M. Morton, A. Rembaum and J. I. Hall, J . PoZymer Sci. A-1, 1963, 1,461. lo (a) R. M. FUOSS, J . Amer. Chem. Soc., 1935,57,488. l1 T. Shimomura, K. J. Tolle, J. Smid and M. Szwarc, J . Amer. Chem. Soc., 1967, 89, 796. '' H. Hirohara, M. Nakayama and N. Ise, J. C.S. Faraday I, 1972,68,58. l3 (a) R. M. FUOSS, J . Amer. Chem. SOC., 1958, 80, 5059. l4 (a) C. Carvajal, K. J. Tolle, J.Smid and M. Szwarc, 1. Amer. Chem. SOC., 1965, 87, 5548. l5 M. Fisher and M. Szwarc, Macromolecules, 1970, 3,23. l6 J. Geerts, M. van Beylen and G. Smets, J . Polymer Sci., A-1, 1969,7,2859. l7 M. Tardi, D. Rougk and P. Sigwalt, Europ. Polymer J., 1967,3, 85. l 8 J. E. Figueruelo, Makromol. Chem., 1970,131,63. l9 N. Ise, H. Hirohara, T. Makino, K. Takaya and M. Nakayama, J . Phys. Chem., 1970,74,606. 'O ref. (lob), chap. 18. 21 C. B. Wooster, J . Amer. Chem. SOC., 1937,59,377; 1938,60,1609. " (a) D. N. Bhattacharyya, J. Smid and M. Szwarc, J. Amer. Chem. Soc., 1964, 86, 5024. (b) ref. (lob), chap. 16. (b) M. Szwarc, Accounts Chem. Res., 1969, 2, 87. (b) D. N. Bhattacharyya, C. L. Lee, J. Smid and M. Szwarc, J. Phys. Chem., 1965, 69, 612. Kinetics of Anionic Polymerization of o-Methylstyrene in 2-Methyltetrahydrofuran and Tetrahydrofuran * BY HIDEO HIROHARA, MASATOSHI NAKAYAMA, RYOICHI mWABATA AND NORIO ISE Dept.of Polymer Chemistry, Kyoto University, Kyoto, Japan Received 26th July, 1971 The kinetics of anionic polymerization of o-methylstyrene were investigated in 2-methyltetra- hydrofuran (MTHF) at 25°C with Li+, Na+, K+ and Cs+ as gegenions in the presence and absence of an electric field. Conductance studies showed that the dissociation constants K vary remarkably with the gegenion and that the lithium and sodium salts are dissociated to larger extents than the respective polystyryl salts. The free ion rate constant k$ was -2 X lo4 M-' s- which was smaller than that of polystyrene. From the conductivity data, the field effect was concluded to be due to an increase in k i with increase of the field strength as in other systems.The k i reached a limiting value of w3 x lo4 M-l s-' above 3 kVlcm. The ion pair rate constants kb were 2, 1, 16 and 30 M-' s-l for Li+, Na+, K+ and CS+, respectively. Comparison with polystyrene systems led to the conclusion that the behaviour of poly(o-methylstyryl) salts are due to steric hindrance effects of the methyl group in the ortho position. The field-accelerating effects reflected in the k i term as in other systems are smaller than that of poIystyrene systems. This is discussed in terms of the ortho effect. The kinetics of the caesium salt of poly(o-methylstyrene) were also studied in tetrahydrofuran (THF) at 25°C. The results are compared with those in MTHF.In a previous paper we reported briefly on anionic polymerizations of poly(0- and p-methy1styrene)s in MTHF.l The study revealed that poly(o-methylstyrene) is less reactive and more strongly dissociated than poly(p-methylstyrene). The work was, however, limited to sodium salts in MTHF in the absence of an electric field. We report here an anionic polymerization study of poly(o-methylstyrene) with Li+, Na+, Kf and Cs+ as gegenions in MTHF together with the caesium salt in THF. Furthermore, it is of interest to apply an electric field to the systems of monomers having a substituent group in the vicinity of the active centre as a continuation of our research work on high intensity electric field effect on polymerization.2 EXPERIMENTAL The detailed techniques and purification of reagents have been described in earlier paper^.^-^ More recently we noted that the purity of reagents is one of the most crucial factors in an accurate determination of the rate constant, and that the influence of impurities on the rate constants is remarkable in MTHF.Therefore, extreme care was exercised in purification of the materials. Commercial 0-methylstyrene (0-MS) (Monomer-Polymer Lab.) was purified by a method similar to that for styrene and p-methoxystyrene.6 The polymerizations of the lithium and caesium salts were started by the corresponding polystyryl salts which were obtained from ethyllithium and cumylcaesium ; for the sodium and potassium salts, sodium and potassium a-methylstyrene tetramer dianions, respectively, were used as initiators.Except for ethyllithium, the initiators were prepared in THF and then the solvent was replaced by MTHF. The replacemeni was confirmed by gas chromatography. The lithium and caesium salts were of the oneended type; the others were twoended. * This article is part 19 of Ionic Polymerization under an Electric Field. 5152 ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE Reaction rates were measured at 25°C using a Hitachi EPS-3T spectrophotometer provided with a temperature-controlled cell holder ; the reactions were monitored by follow- ing the disappearance of the o-methylstyrene absorption at 298.2 mp. The initial concentra- tion of o-methylstyrene was usually 25-50 times greater than the active centre concentration, which was also determined spectrophotometrically using the poly(o-methylstyryl) anion absorption band (& : Li, 329 mp ; Na, 338 mp ; K, 340 mp ; Cs, 343 mp).The molar extinction coefficients were assumed to be 1.3 x lo4 M-' cm-1 for all four salts in the light of the extinction coefficient data of living polystyrenes and other polymers in ethereal solvents. The electric conductance was calculated from the applied voltage and the current passing through the polymerizing solution. The conductances of all four salts were practically independent of time after application of the electric field when elaborately purified MTHF was used as the solvent. The electric field was applied in a conversion range between 20 and 75 %; the rate constant was also calculated in the same conversion range.RESULTS The apparent propagation rate constants k, in the presence and absence of an electric field for all four salts in MTHF are presented in the form of the (kp, l/[LE]+) plots in fig. 1 and 2, where [LEI is the total concentration of active centres. The intercept gives the ion-pair rate constant kk and the slope kgK4, where kg is the free ion rate constant and K the ion-pair dissociation constant. Except for the caesium salt the slope for which was zero, a high intensity electric field increased the slope, whereas the intercept remained unaffected as previously observed in various sy~terns.~'~ 0 50 100 150 200 1 / d [LEI FIG. 1 .-Dependence of the apparent propagation rate constant on the poly(o-methylstyry1)lithium and poly(o-methylstyry1)potassium concentration in MTHF at 25°C.The kinetic data and field effects obtained from fig. 1 and 2 are collected in table 1. The ki values of both the lithium and sodium salts seem to be practically equal and much smaller than those of the respective salts of polysytrene (25 and 23 M-l s-l for lithium and sodium salts, respecti~ely).~ The reactivity of the potassium ion pair isH. HIROHARA, M. NAKAYAMA, R. KAWABATA A N D N. ISE 53 comparable to that of polystyrene and the kb value of the caesium salt appeared to be larger than that of polystyrene (23 M-’ s - ’ ) . ~ The k%K4 values at 0 kV/cm decreased \ @ 50 0 50 100 150 200 1 /1/[LEI FIG. 2.-Dependence of the apparent propagation rate constant on the poly(o-methylstyry1)sodium and poly(o-methylstyry1)caesium concentration in MTHF at 25°C.@ : from ref. (1). in the order Lif > Naf > K+ > Cs+, and are much smaller than those of polystyryl salts. The field acceleration factor (given in the fifth column) reached a limiting value (1.3 - 1.4) at E = 2- 3 kV/cm and is smaller than that for polystyrene (1.6 N 1.7).5 TABLE KINETIC RESULTS AND FIELD EFFECTS ON THE ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE IN MTHF AT 25°C kgEK% E k;, kgK3 gegenion /kV cm-1 /M-1 s-1 lM-& s-1 k&, K$ Li+ 0 2 1.06 1 0.6 2 1.17 1.1 1.3 2 1.42 1.3 2.3 2 1.56 1.5 1.5 1 0.60 1.2 3.0 1 0.68 1.3 4.5 1 0.68 1.3 1.5 16 0.30 1.2 3.0 16 0.34 1.3 4.5 16 0.34 1.3 cs+ 0 30 0 - 3.0 30 0 Na+ 0 1 0.51 1 K+ 0 16 0.26 154 - 1.4 -1.6- ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE - - 1.0 \ 0 0 cs(l)o ooo The equivalent conductances A for the four salts are shown in fig.3. We note that the A values of each salt fall on a line irrespective of the applied field strength. Thus all the data are shown with blank circles without specifying the field strength. f2A[LE]/Fx lo6 20 f2A[LE]/Fx lo5 FIG. 4.-Fuoss plots of the conductance of living poly(o-methylstyrene) in MTHF at 25°C.H. HIROHARA, M. NAKAYAMA, R . KAWABATA A N D N. ISE 55 The lower concentration part for the four salts gave straight lines of a slope of -3. For potassium and caesium salts, however, the slopes became larger at higher con- centrations. In fig. 4, the conductances for lithium, sodium and potassium salts TABLE 2.-DISSOCIATION CONSTANT AND FREE ION RATE CONSTANT OF LIVING POLY(0-METHYL- STYRENE) IN MTHF AT 25” K x lO*O/M K(o-MeSt) k$ x lO-4/M-1 s-1 St b K(St) o-MeSt St A0 = gegenion /cmz mho-1 o-MeSt Li+ 58 25 8.7 3 2.1 6.8 Na+ 58 5.5 2.4 2 2 .2 7.4 K+ 58 3.6 3.4 1 1.4 7.1 7.3 cs+ 75 0.04 2.3 =O a assumed : see text for details b from ref. (5). - were plotted according to the method of FUOSS.~~ To evaluate K, we used the limit- ing conductance A. of polystyryl salts estimated in the previous work s assuming that the effect of the methyl group on A. is neghgible. The K values are given in TABLE 3.-KINETIC RESULTS OF THE CAESIUM SALTOF POLY(0-METHYLSTYRENE) IN THF AT 25°C k;, kp”& A0 K x lo9 k g x 10-4 monomer /M-1 s-1 /M-3 s-1 /cmz mho-1 /M /M-1 s-1 O-MS 15 f5 1.7 92.5 1.8 4 styrene” 20f5 6.8 92.5 2.7 13 b from Smarc’s data in ref. (1 1).from ref. (7) table 2, together with those of polystyryl salts. The K value, in conjunction with the slopes of the (kp, I/[LE]*) plot, allows us to calculate k;. The results are also given in table 2. Experiments were also performed for caesium salts in THF. The kinetic data are represented in table 3. DISCUSSION Alkali poly(o-methylstyryl) salts with the exception of the caesium salt are more strongly dissociated than corresponding polystyryl salts. The K values of the former were strongly dependent on the gegenion and decreased along the series Li+ > Na+ > K+ > Cs+, but those of the four salts of the latter polymer were not greatly different from each other. The ratio of the K values of the two polymers decreases along the series Li+ > Na+ > K+ > Cs+.For poly(p-methoxystyrene) in THF and poly(p- methylstyrene),12 the order of the ratio was reversed. These polymer salts were less dissociated than polystyryl salts. These facts are inexplicable in terms of the electron-donating nature of the methyl group, but suggest that the methyl group in the vicinity of the active centre prevents the gegenion from re-approaching the nega- tively charged free ion, except in the caesium case. This steric hindrance would depend on the effective ionic radius of the gegenion. It is believed that the lithium ion is strongly solvated and the caesium ion is hardly solvated by MTHF molecules. The solvated lithium ion having a large effective ionic radius would be effectively hindered from association with the free anion. This might account for much larger K value of the lithium poly(o-methylstyrene) than that of the corresponding salt of56 ANIONIC POLYMERIZATION OF 0-METHYLSTYRENE polystyrene.The situation would be similar for the sodium salt. For the caesium salt, on the other hand, the methyl group might prevent the caesium ion from dis- sociating as a consequence of the small effective ionic radius. The interionic distance a in the ion pairs can be calculated from the K value using the “ sphere in continuum ” model which gives the Fuoss eqn l3 K = (3000/4nNa3)exp( - e2/aDkT). The a value for -(o-MS)-, Cs+ ion pairs was found to be 3.1 A, whereas it is 3.7A for -S-, Cs+ ion pairs in MTHF,’ where w(o-MS)- and -S- are poly(o- methylstyryl) anion and polystyryl anion, respectively.The Stokes radius for the caesium ion in MTHF is probably 2.2-2.4 A in the light of the data in THF.14 From these, we have = 1.4 A for the radius of the polystyrylanion in MTHF. If poly(o- methylstyryl) anion has the same radius, the radius of the cation for the -(o-MS)-, Cs+ ion pair in MTHF is estimated to be M 1.7 A, which is practically equal to the crystal radius of the caesium ion. This might suggest that in the formation of the -(o-MS)-, Cs+ ion pair, the MTHF molecules weakly bound to the caesium ion are removed as a result of the high electron density on the carbanion due to the electron- donating nature of the methyl group. On the other hand, in THF, the solvating power of which is stronger than that of MTHF, this situation might not be en- countered.As a matter of fact, the caesium poly(o-methylstyrene) is slightly less dissociated in THF than the caesium polystyrene (see table 3). Thus, the ortho effect is very sensitive to the solvation state. The methyl group in the ortho position also prevents monomer from reaching the active centre by the steric effect, in addition to the electron-donating nature of the methyl group. Thus, the kk and k i values are much smaller than the respective values of poly(m- and p-methy1styiene)s ’* l2 as well as polystyrene.’ The free - (0-MS)- ion is 3-$ as reactive as the free - 5 ion not only in MTHF but also in THF. This behaviour is reminiscent of that found for living poly(2-vinylpyridine) and poly(o-methoxystyrene),l although the intramolecular “ solvation ” of the cation took place in these cases.For the field effect, we conclude that, since A was insensitive to the field strength, i.e., K did not increase with the field strength, augmentation in k: is a primary factor for the observed effect and may be attributed to removal of the solvent molecules weakly bound to the free carbanion by the electric field, as suggested in other sys- tems. 3-6 The observed field-acceleration effects (kiE/k%,) were smaller than those for polystyrene systems, suggesting that solvation of the negatively charged active centre of poly(o-methylstyrene) is partly hindered sterically. The kLE value of 3 x lo4 M-l s-I, which was obtained from the K value and the kg K*, is interpreted as the “de- solvated ’’ free ion rate constant for living poly(o-methylstyrene).Finally, there should be discussion of the slope of the (log A, log[LE]) plots in potassium and caesium systems. Deviation of the slope from -4 toward positive values was reported for sodium poly(2- and 4-vinylpyridine)~ in THF, ’ alkali poly(methylmethacry1ate)s in THF, * and lithium polystyrene in benzene-dimethoxy- ethane mixtures.19 No explanation is offered for this phenomenon except in the polystyryl lithium case. In this case, intermolecular triple ion formation was pro- posed. In the present system also, we can tentatively assume that triple ions exist in addition to free ions and ion pairs. Negatively charged triple ions -(o-MS)-, (+)--? (0-MS)-- are much more stable than positive ones (+I, (o-MS)--, (+) as discussed by Fuoss 2o and Wooster.21 We note here that the Syracuse group was the first to propose formation of triple ions,22 which were, however, of an intra- molecular type.Following the treatment of Wooster 21 for the unilateral triple ionH . HIROHARA, M. NAKAYAMA, R . KAWABATA A N D N . ISE 57 formation, we calculated approximate dissociation constants for the triple ion k and the ion pair K. The results are as follows. For the caesium salt, there can exist about 15 times as many triple ions as free ions at [LEI = 1 x M and K = 8 x 10-13 M. For the potassium salt, about 3 times as many triple ions as free ions exist at [LEI = 1 x M and K = 1.5 x 10-lo M, which, by combining with the slope of the (kp, l/[LE]*) plot, leads to Pi = 2.1 x lo4 M-l s-' . This value is in excellent agreement with the k i values from the other two salts (see table 2).We note that positive slopes were observed in the (kp, 1/[LEIB) plot, though three ionic species (free ions, ion pairs and triple ions) contribute to the propagation process. The positive slope implies kg (triple ion rate constant) < k; in any case. This is quite reasonable since triple ions for the potassium and caesium salts would be of the contact type as suggested by Szwarz et aZ.22 Furthermore, contributions from the triple ions would explain the apparently large kk value for the -S-, K+ and -S-, Cs+ ion pairs. Although we feel compelled to draw attention to contribution of the triple ions, since it fits the results mentioned above, we are unhappy with this inter- pretation. It is rather difficult to understand why the triple ion is formed for living poly(o-methylstyrene) in spite of the fact that no triple ion was formed for living polystyrene in the same solvent.In conclusion, we emphasize that the steric hindrance effect depends on the effective size of the gegenion and that the kinetic study of ortho-substituted living polymer can be useful for obtaining information on solvation of active centres in polymerization systems. M. Nakayama, H. Hirohara, K. Takaya and N. Ise, J. Polymer Sci., A-1,1970,8,3653. N. Ise, Ado. Polymer Sci., 1969, 6, 347, in which previous work is referred to. N. Ise, H. Hirohara, T. Makino and I. Sakurada, J. Phys. Chem., 1968, 72,4543. H. Hirohara, M. Nakayama, K. Takaya and N. Ise, Trans. Fmaday Soc., 1970,66,1165. H. Hirohara, K. Takaya, M. Nakayama and N. Ise, Trans. Farahy SOC., 1970,66, 3163. K. Takaya, H. Hirohara, M. Nakayama and N. Ise, Trans. Faraday SOC., 1971,67, 119. H. Hirohara, K. Takaya, N. Ise, Macromolecules, 1971,4,288. K. Ziegler and H. Dislich, Ber., 1957, 90, 1107. (b) R. M. Fuoss and F. Accascina, ElectroZytic Conductance (John Wiley and Son., Inc., New York, 1959), chap. 17. ' M. Morton, A. Rembaum and J. I. Hall, J . PoZymer Sci. A-1, 1963, 1,461. lo (a) R. M. FUOSS, J . Amer. Chem. Soc., 1935,57,488. l1 T. Shimomura, K. J. Tolle, J. Smid and M. Szwarc, J . Amer. Chem. Soc., 1967, 89, 796. '' H. Hirohara, M. Nakayama and N. Ise, J. C.S. Faraday I, 1972,68,58. l3 (a) R. M. FUOSS, J . Amer. Chem. SOC., 1958, 80, 5059. l4 (a) C. Carvajal, K. J. Tolle, J. Smid and M. Szwarc, 1. Amer. Chem. SOC., 1965, 87, 5548. l5 M. Fisher and M. Szwarc, Macromolecules, 1970, 3,23. l6 J. Geerts, M. van Beylen and G. Smets, J . Polymer Sci., A-1, 1969,7,2859. l7 M. Tardi, D. Rougk and P. Sigwalt, Europ. Polymer J., 1967,3, 85. l 8 J. E. Figueruelo, Makromol. Chem., 1970,131,63. l9 N. Ise, H. Hirohara, T. Makino, K. Takaya and M. Nakayama, J . Phys. Chem., 1970,74,606. 'O ref. (lob), chap. 18. 21 C. B. Wooster, J . Amer. Chem. SOC., 1937,59,377; 1938,60,1609. " (a) D. N. Bhattacharyya, J. Smid and M. Szwarc, J. Amer. Chem. Soc., 1964, 86, 5024. (b) ref. (lob), chap. 16. (b) M. Szwarc, Accounts Chem. Res., 1969, 2, 87. (b) D. N. Bhattacharyya, C. L. Lee, J. Smid and M. Szwarc, J. Phys. Chem., 1965, 69, 612.

ISSN:0300-9599    

DOI:10.1039/F19726800051

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Kinetics of Anionic Polymerizations of Styrene and Its m- and p-Derivatives: Hammett’s Relations * BY HIDEO HIROHARA, MASATOSHI NAKAYAMA AND NORIO ISE Dept. of Polymer Chemistry, Kyoto University, Kyoto, Japan Received 9th August, 1971 The kinetics of anionic polymerizations of styrene, m-methylstyrene, p-isopropylstyrene, and p- methylstyrene were investigated in tetrahydrofuran (THF) at 25°C with Cs+ as gegenion in the presence and absence of an electric field. Anionic polymerizations of m- and p-methylstyrenes in 2-methyl- tetrahydrofuran (MTHF) with Na+ as gegenion were also studied. The ion-pair rate constants kk, the free ion rate constants k; and the dissociation constants K of the styrene derivatives were generally smaller in both solvents than those of styrene. Conductance studies showed that the K values of the caesium salts are practically the same in THF.It was found that the free anions of these living polymers in MTHF are half as reactive as in THF. The field acceleration effect was hardly per- ceptible in THF. Following earlier studies of the field effect, the solvation ability of ethereal solvents to free carbanions was concluded to decrease along the series dimethoxyethane @ME) >THPe MTHF > THF. A value of (12&2)x lo4 M-l s-1 was attributed to the ‘‘ desolvated ” free ion rate constant for living polystyrene. Hammett’s relations of the two systems Cs+/THF and Na+/ MTHF were examined for k,, k;, k;, and K, independently. The reactivity constants p for k, and K were dependent on systems whereas those for k,” were not.In this laboratory, electric field effects on polymerization reactions have been studied.l Special attention has been paid to the solvent and gegenion dependences of the kinetics of anionic polymerization of ~tyrene.~’~ We report here the kinetics and field effects for anionic polymerizations of styrene and its meta- and para- substituted derivatives, having an electron-donating substituent, in THF with Csf as gegenion. This combination of solvent-gegenion was chosen because the field effect has not yet been studied in THF and because caesium ion is hardly coordinated with the solvent. Also, we report the kinetics of anionic polymerizations of rn- and p-methylstyrenes in MTHF with Naf as gegenion in the presence and absence of an electric field.EXPERIMENTAL All the experiments were carried out under high vacuum in sealed apparatus. Com- mercial solvents and monomers [styrene(S), m-methylstyrene(m-MS), p-methylstyrene@-MS) TABLE l.-sPECTRA OF SUBSTITUTED STYRENES AND THEIR ALKALI SALTS AT 25°C max. of monomer solvent monomer (w) S THF 291.4 m-MS THF 296.4 p-iPS THF 294.2 p-MS THF 295.5 m-MS MTHF 296.2 p-MS MTHF 295.5 extinction coefficient 6 X lO-4/cm-1 M-1 max. of of living end living end (w) 342 1.25 355 1.3 350 1.3 350 1.3 352 1.3 337.5 1.3 a assumed value, except the extinction coefficient of styrene is taken from ref. (9). * This article is part 20 of Ionic Polymerization under an Electric Field. 58H. HIROHARA, M. NARAYAMA AND N. ISE 59 and p-isopropylstyrene(p-iPS)] were purified by techniques and methods described in earlier papers.2* 4* Extreme care was exercised in purification of thematerials because of our findings on the effects of impurities in anionic polymerization.6 The general procedures for kinetic and conductance measurements have been described 2* ; the reaction rates were measured at 25°C spectrophotometrically. The spectra and extinction coefficient of the systems are listed in table 1.The polymerization was started by the caesium salt of living polystyrene (E N 10) or sodium a-methylstyrene tetramer dianions. The caesium salt of living poly- styrene was prepared from cumylcaesium, which was obtained in THF by the method of Ziegler et aL7 The conductances of all caesium salts in THF began to increase markedly with time 15-20 s after application of an electric field even when THF was purified very thoroughly.Similar behaviour was reported by Bywater et al. for a sodium salt of polystyrene in THF.8 Thus, the equivalent conductance before the increase was taken for all salts. RESULTS Fig. 1, 2 and 3 show the kinetic data in the presence and absence of an electric field plotted in the form proposed by Szwarc and Schulz,l* where [LEI is the con- centration of living ends. The intercept gives the ion-pair rate constant k; and the slope KiK4 (where kg is the free ion rate constant and K the ion-pair dissociation constant). For the caesium salts of polystyrene and poly(p4sopropylstyrene) in THF, the slope appears to increase very slightly at E = 0.6 kV/cm. Strangely 21 I.! n I .+ 8 rl I.( W I 2 2 X 0.1 C 0 : 1.5 KVIcm 0 : 0.6 om0 -s;cs+ / @ 50 100 I50 200 25( [LEI-* FIG.1.-Dependence of the apparent propagation rate constant on the polystyrylcaesium and poly(m-methylstyry1)caesium concentration in THF at 25°C (ZJ : from ref. (6). enough, the slope of the styrylcaesium decreased again at E = 1.5 kV/cm and is equal to that at zero field. The field effects are hardly perceptible for the caesium salts of poly(m- and p-methy1styrene)s in THF. The slopes for these sodium salts in MTHF increased with field strength, in contrast to the caesium salts in THF, whereas the intercepts of all systems remained unaffected as observed p r e v i o ~ s l y . ~ ~ ~ ~ Essential kinetic data and field effects obtained from fig. 1,2 and 3 are summarized in table 2. The k; values of caesium salts in THF appear to be insensitive to the 11.1260 HAMMETT’S RELATIONS IN ANIONIC POLYMERIZATIONS 1 0 .0 1 X 2 C J 0 50 100 150 200 250 [LEI-* FIG. 2.-Dependence of the apparent propagation rate constant on the poIy(p-methylstyry1)caesium and poly(p-isopropylstyry1)caesium concentration in THF at 25°C. 200 I50 n ?, @ -I I z 100 50 0 0 : 1.5 .o:o r I I I I 50 100 150 200 [LEI-* FIG. 3.-Dependence of the apparent propagation rate constant on the poly(m-methylstyry1)sodium and ply@-methylstyry1)sodium concentration in MTHF at 25°C.H. HIROHARA, M. NAKAYAMA AND N. ISE 61 TABLE 2.-KINETIC DATA AND FIELD EFFECTS ON THE ANIONIC POLYMERIZATION OF m- AND p- SUBSTITUTED STYRENES AT 25°C E k;, kg K3 monomer /kV cm-1 IM-1 s-1 lM-3 s-1 gegenion, Cs': solvent, THF S 0 23 6.8 0.6 23 (8.4) 1.5 23 6.8 m-MS 0 20 2.9 0.6 20 2.9 1.5 20 2.9 p-iPS 0 -0 1.1 0.6 -0 1.3 p-MS 0 17 4.1 0.3 17 4.1 0.6 17 4.1 1 .O (-0) (4.0) gegenion, Na+: m-MS 0 1.5 3.0 4.5 1.5 3.0 4.5 p-MS 0 solvent, MTHF 8 0.71 8 0.82 8 0.97 8 0.97 5 0.25 5 0.28 5 0.36 5 0.36 3 c i3; 2 1 0 2 4 6 8 1 0 1O5[LE]AJ2/F FIG.4.-The Fuoss conductivity plot for caesiwn salts of various living polymers in THF at 25°C. 0, polystyrene ; A, poly@-methylstyrene) ; 0, polyfm-methylstyrene) ; x , poly(p-isopropyl- styrene).62 HAMMETT’S RELATIONS IN ANIONIC POLYMERIZATIONS substituent group, except for p-ips. The field acceleration effects for sodium salts of poly(m- and p-methy1styrene)s in MTHF reached a limiting value at about 3 kV/ cm, as observed in the polystyrene ~ystems.~ The conductance data in both THF and MTHF as a function of log[LE] gave a straight line of a slope -3 irrespective of the applied field strength as observed in other systems.4* 5 9 1 1 * l2 I c c ’ 0 4 8 1 1 O6 [ LE]AJ2 / F FIG.5.-The Fuoss conductivity plot for sodium salts of various living polymers in MTHF at 25°C. The method of Fuoss l 3 was used to determine K. The Fuoss plots for various living polymers are shown in fig. 4 and 5. The conductance of the sodium salt of polystyrene in MTHF, which was previously analyzed by the Kraus and Bray method was replotted by the Fuoss method in fig. 5. To calculate the parameters in the Fuoss relations, the limiting conductance (Ao) values, of appropriate polystyryl salts reported in previous papers 5 * l4 were used with an assumption that the effect of the substituent groups on A.is negligible. This is justified by the fact the conductance of living polymer anions is much smaller than that of alkali cations. The A,, l/A; K and K values are given in table 3 together with k: values in the absence of an electric field. Also the results for the casium salt of poly(p-methoxystyrene) and the sodium salt of polystyrene are given for comparison. DISCUSSION KINETIC BEHAVIOUR The k; value of the living polystyrene in THF is in good agreement with that reported by Schulz et aZ.lO. l 5 As all the selected substituent groups are electron- donating, the electron density of the carbanions of styrene derivatives must be higher than that of styrene, making stronger the interaction in the ion-pairs.Moreover, the substituent groups increase the electron density of the double bond of monomers.H. HIROHARA, M. NAKAYAMA A N D N . ISE 63 Living polymers of substituted styrenes are, thus, expected to be less dissociated and less reactive than living polystyrene. Except for the K values of caesium salts, this is really the case as shown in table 2 and 3, although the k; values in THF could not be determined very accurately because of the large slopes of the (kp, [LEI-%) plots. TABLE 3.-DISSOCJATION CONSTANT AND FREE ION RATE CONSTANT OF VARIOUS LIVING POLYMERS AT 25°C monomer A0 1 /A$K Kx 10-10 k; x 10-4 /cm-2 ohm-1 equiv.-l /M IM-1 s-1 gegenion, Cs+ : solvent, THF S 92.5 4 . 3 ~ 104 27 13 m-MS 92.5 8.8 13 8.0 p-iPSa 92.5 2.8 <41 > 1.7 p-MS 92.5 4.3 27 7.9 p-MOS 92.5 4.3 27 3.9 gegenion, Na+ : solvent, MTHF S 58.0 1.3 x lo6 2.3 7.5 m-MS 58.0 1.8 1.7 5.5 p-MS 58.0 6.0 0.50 3.7 * see text for details It is very interesting to note that the free carbanions are half as reactive in MTHF as in THF.This might imply that the coordination of MTHF molecules to the free carbanions is stronger than that of THF molecules. The reactivity of living poly(p- isopropylstyrene) was abnormally small." This might be due to the steric hindrance due to the isopropyl group as suggested by Natta et al.l6 in the complex polymer- ization systems. The behaviour of K for the caesium salts in THF will be discussed later. FIELD EFFECT The field accleration effect was distinct in MTHF, whereas it was hardly per- ceptible in THF.From the insensitivity of A to field strength, it can be concluded that increase of k: is a primary factor for the observed accleration and is due to desolvation by an electric field of solvent molecules weakly bound to free carbanions, as proposed previously.2 ' 9 11* l 2 Some comments are needed on the absence of field effects in THF systems. From the steep increase of conductance in THF systems mentioned in the experimental section, it is believed that ionic impurities were produced in the electric field so that the field effect was marked. Moreover, it should be noted that the field effect was clearly observed to be saturated at 0.6- 1.0 kV/cm for living poly(p-methoxystyrene) in THF.ll A value of * 12 x lo4 M-' s-l was deduced for the rate constant of " desolvated " free ion of living polystyrene in previous studies.2* 4 9 The k; value (13 x lo4 M-l s-l ) found in the present work * Gas chromatographic analyses of p-isopropylstyrene showed the presenm of a small quantity of an impurity. An attempt to remove this impurity by distillation was not completely successful. At the start of the propagation reactions of poly(p-isopropylstyryl)caesium, a shoulder of absorption appeared at 450 mp in addition to the absorption peak (Amax) at 350 my.The shoulder disappeared rather quickly as the polymerization progressed whereas the peak at 350 mp was stable during the polymerization. Therefore, the rate constant was calculated and an electric field was applied in a conversion range between 50 and 85 % after the shoulder at 450mp had disappeared.In view of this complication, the results obtained for poly(p-isopropylstyry1)caesium should be viewed with caution. When the shoulder at 450 my existed, smaller kp and higher A values were obtained than in its absence.64 for living polystyrene in THF in the absence of an electric field was found to be nearly equal to the rate constant of “ desolvated ” growing ends. Thus, it can be concluded that the solvation of THF molecules to free polystyryl anion in the transition state of the reaction is extremely weak. The field intensity dependence of the free ion rate constant in THF is shown in fig. 6 together with the results obtained in other solvents. From fig. 6, we assert HAMMETT’S RELATIONS I N ANIONIC POLYMERIZATIONS ‘1 0 0 1.5 3D 4.5 E(kV/cm) FIG. 6.-Dependence of the free ion rate constant of polystyryl salt on field intensity in various solvents.0, Li salt ; A, Na salt ; 13, K salts ; x , Cs salt. again that the k; value of (12+2) x lo4 M-l s-l is the rate constant of the desolvated free anion of living polystyrene. Furthermore, from the field effects observed in various systems, the order of coordination ability of various ethereal solvents to free carbanions decreases along the series DME > THF-MTHF > THF. This view seems to be supported by the fact that the reactivity of polystyryl free anions decreased along the series THF > MTHFETHP > DME in the absence of an electric field. HAMMETT’S RELATIONS The relative rate constants and dissociation constants of living polymers of sub- stituted styrene should obey the Hammett relations.This is demonstrated in fig. 7. The results for homopolymerization are plotted. Therefore, it should be noted that the reactivities of not only monomers but also of living active ends are reflected in the results, although it is found that the reactivity of very similar active ends is not significantly affected by structure.17 The data for both caesium and sodium salts are plotted in fig. 7 for comparison. Szwarc et al. have studied l8 addition reactions of a series of analogous monomers to sodium polystyryl in THF to give a reactivity constant p = 5.0 for the apparent propagation rate constant k,. The value of k , is not, however, very significant, since it is a function of kk, k$ and K. Accord- ingly we studied Hammett’s relations for &, and K, separately.H.HIROHARA, M . NAKAYAMA AND N . ISE 65 For the apparent rate constant, p = 1.7 was obtained for caesium salts in THF and p = 4.2 for sodium salts in MTHF. The latter value is in agreement with Szwarc’s data in spite of the difference in solvent. For k; and K, also, the p values were found to depend on the gegenion. It should be noted that the gegenion de- pendence of k, is a consequence of the gegenion dependences for kb and K. For k;, P = 0.5 -2.0 for caesium salts and 3.8 for sodium salts. For K, p = 0 for caesium A I I n 1 4 -0.5 -0.3 -Q2 -0.1 D 0 I (I FIG. 7.-Hammett plots for living poly(substituted styrene) at 25°C. x , sodium salt in MTHF. A, apparent rate constant ([LEI = 1 x 0, caesium salt in THF; M) ; B, ion-pair rate con- stant ; C, dissociation constant ; D, free ion rate constant.salts and 4.2 for sodium salts. These results show that the solvation state and inter- ionic distance of contact ion pairs of caesium salts are not affected by anions having the electron donating substituent group because of the large Pauling ionic radius of Csf ion. This explanation, however, might be questionable. As a matter of fact, poly(0-methylstyry1)caesium was much less dissociated in MTHF than polystyryl- caesium, although in THF the former salt was dissociated to the same extent as the latter.12 On the other hand, Na+ ion has a small Pauling ionic radius, so that the interionic distance of contact ion pairs will be small, and in the formation of solvent- separated ion pair its interionic distance will be large since Na+ ion is coordinated strongly with the ethereal solvents.Consequently, the solvation state and interionic distance of sodium ion pairs vary greatly with the anion. The relative reactivity of the free ions fell on the same line of a slope of 1.7 ir- respective of the solvent and the gegenion. This might be reasonable since k: should be independent of gegenion. The identity of the values implies that the relative66 HAMMETT'S RELATIONS IN ANIONIC POLYMERIZATIONS reactivity of free carbanions in THF is equivalent to that in MTHF although the magnitudes of k; are different in both solvents as shown in table 3. Finally, it is safe to conclude that, in the case of sodium salts, the k, values of various living polymers reflect the dissociation state of these polymers, which is determined by solvation state of the ion pairs.On the contrary, for caesium salts, the most important factor influencing the k, value is the reactivity of the free ions. e.g., N. he, Ad. Polymer Sci., 1969, 6, 347, in which previous work is referred to. N. Ise, H. Hirohara, T. Makino and I. Sakurada, J. Phys. Chem., 1968,72,4543. N. Ise, H. Hirohara, I. Makino, K. Takaya and M. Nakayama, J. Phys. Chem., 1970,74,606. H. Hirohara, M. Nakayama, K. Takaya and N. Ise, Trans. Favaday SOC., 1970,66,1165. H. Hirohara, K. Talcaya, M. Nakayama and N. be, Tram. Faraday Soc., 1970,66,3163. H. Hirohara, K. Takaya and N. Ise, Macromolecules, 1971,4,288. D. J. Worsfold and S . Bywater, J. Chem.SOC., 1960, 5234. D. N. Bhattacharyya, C. L. Lee, J. Smid and M. Szwarc, J. Phys. Chem., 1965,69,612. l o H. Hostalka and G. V. Schulz, 2. phys. Chern., 1965,45,286. K. Takaya, H. Hirohara, M. Nakayama and N. Ise, Trans. Faraby SOC., 1971,67, 119. l2 H. Hirohara, M. Nakayama, R. Kawabata and N. Ise, J.C.S. Fczraday I, 1972, 68, 51. l3 (a) R. M. Fuoss, J. Amer. Chem. Soc., 1935,57,488. ' K. Ziegler and H. Dislich, Ber., 1957,90, 1107. (b) R. M. Fuoss and F. Accascina, Electrolytic Conductance (John Wiley and Son Inc., New York, 1959), chap. 17. l4 J. Shimomura, K. J. Tolle, J. Smid and M. Szwarc, J. Amer. Chem. Soc., 1967,89, 796. l 5 B. J. Schmitt and G. V. Schulz, Makrornol Chem., 1971,142,325. l6 G. Natta, F. Danusso and D. Sianesi, Makromol. Chem., 1959,30,238. l' M.Szwarc, Carbanions, Living Polymers and Electron-Transfer Process (John Wiley and Son Inc., New York, 1968). M. Shima, D. N. Bhattacharyya, J. Smid and M. Szwarc, J. Amer. Chem. SOC., 1963,85,1306. Kinetics of Anionic Polymerizations of Styrene and Its m- and p-Derivatives: Hammett’s Relations * BY HIDEO HIROHARA, MASATOSHI NAKAYAMA AND NORIO ISE Dept. of Polymer Chemistry, Kyoto University, Kyoto, Japan Received 9th August, 1971 The kinetics of anionic polymerizations of styrene, m-methylstyrene, p-isopropylstyrene, and p- methylstyrene were investigated in tetrahydrofuran (THF) at 25°C with Cs+ as gegenion in the presence and absence of an electric field. Anionic polymerizations of m- and p-methylstyrenes in 2-methyl- tetrahydrofuran (MTHF) with Na+ as gegenion were also studied.The ion-pair rate constants kk, the free ion rate constants k; and the dissociation constants K of the styrene derivatives were generally smaller in both solvents than those of styrene. Conductance studies showed that the K values of the caesium salts are practically the same in THF. It was found that the free anions of these living polymers in MTHF are half as reactive as in THF. The field acceleration effect was hardly per- ceptible in THF. Following earlier studies of the field effect, the solvation ability of ethereal solvents to free carbanions was concluded to decrease along the series dimethoxyethane @ME) >THPe MTHF > THF. A value of (12&2)x lo4 M-l s-1 was attributed to the ‘‘ desolvated ” free ion rate constant for living polystyrene.Hammett’s relations of the two systems Cs+/THF and Na+/ MTHF were examined for k,, k;, k;, and K, independently. The reactivity constants p for k, and K were dependent on systems whereas those for k,” were not. In this laboratory, electric field effects on polymerization reactions have been studied.l Special attention has been paid to the solvent and gegenion dependences of the kinetics of anionic polymerization of ~tyrene.~’~ We report here the kinetics and field effects for anionic polymerizations of styrene and its meta- and para- substituted derivatives, having an electron-donating substituent, in THF with Csf as gegenion. This combination of solvent-gegenion was chosen because the field effect has not yet been studied in THF and because caesium ion is hardly coordinated with the solvent.Also, we report the kinetics of anionic polymerizations of rn- and p-methylstyrenes in MTHF with Naf as gegenion in the presence and absence of an electric field. EXPERIMENTAL All the experiments were carried out under high vacuum in sealed apparatus. Com- mercial solvents and monomers [styrene(S), m-methylstyrene(m-MS), p-methylstyrene@-MS) TABLE l.-sPECTRA OF SUBSTITUTED STYRENES AND THEIR ALKALI SALTS AT 25°C max. of monomer solvent monomer (w) S THF 291.4 m-MS THF 296.4 p-iPS THF 294.2 p-MS THF 295.5 m-MS MTHF 296.2 p-MS MTHF 295.5 extinction coefficient 6 X lO-4/cm-1 M-1 max. of of living end living end (w) 342 1.25 355 1.3 350 1.3 350 1.3 352 1.3 337.5 1.3 a assumed value, except the extinction coefficient of styrene is taken from ref.(9). * This article is part 20 of Ionic Polymerization under an Electric Field. 58H. HIROHARA, M. NARAYAMA AND N. ISE 59 and p-isopropylstyrene(p-iPS)] were purified by techniques and methods described in earlier papers.2* 4* Extreme care was exercised in purification of thematerials because of our findings on the effects of impurities in anionic polymerization.6 The general procedures for kinetic and conductance measurements have been described 2* ; the reaction rates were measured at 25°C spectrophotometrically. The spectra and extinction coefficient of the systems are listed in table 1. The polymerization was started by the caesium salt of living polystyrene (E N 10) or sodium a-methylstyrene tetramer dianions. The caesium salt of living poly- styrene was prepared from cumylcaesium, which was obtained in THF by the method of Ziegler et aL7 The conductances of all caesium salts in THF began to increase markedly with time 15-20 s after application of an electric field even when THF was purified very thoroughly.Similar behaviour was reported by Bywater et al. for a sodium salt of polystyrene in THF.8 Thus, the equivalent conductance before the increase was taken for all salts. RESULTS Fig. 1, 2 and 3 show the kinetic data in the presence and absence of an electric field plotted in the form proposed by Szwarc and Schulz,l* where [LEI is the con- centration of living ends. The intercept gives the ion-pair rate constant k; and the slope KiK4 (where kg is the free ion rate constant and K the ion-pair dissociation constant). For the caesium salts of polystyrene and poly(p4sopropylstyrene) in THF, the slope appears to increase very slightly at E = 0.6 kV/cm.Strangely 21 I.! n I .+ 8 rl I.( W I 2 2 X 0.1 C 0 : 1.5 KVIcm 0 : 0.6 om0 -s;cs+ / @ 50 100 I50 200 25( [LEI-* FIG. 1.-Dependence of the apparent propagation rate constant on the polystyrylcaesium and poly(m-methylstyry1)caesium concentration in THF at 25°C (ZJ : from ref. (6). enough, the slope of the styrylcaesium decreased again at E = 1.5 kV/cm and is equal to that at zero field. The field effects are hardly perceptible for the caesium salts of poly(m- and p-methy1styrene)s in THF. The slopes for these sodium salts in MTHF increased with field strength, in contrast to the caesium salts in THF, whereas the intercepts of all systems remained unaffected as observed p r e v i o ~ s l y .~ ~ ~ ~ Essential kinetic data and field effects obtained from fig. 1,2 and 3 are summarized in table 2. The k; values of caesium salts in THF appear to be insensitive to the 11.1260 HAMMETT’S RELATIONS IN ANIONIC POLYMERIZATIONS 1 0 . 0 1 X 2 C J 0 50 100 150 200 250 [LEI-* FIG. 2.-Dependence of the apparent propagation rate constant on the poIy(p-methylstyry1)caesium and poly(p-isopropylstyry1)caesium concentration in THF at 25°C. 200 I50 n ?, @ -I I z 100 50 0 0 : 1.5 .o:o r I I I I 50 100 150 200 [LEI-* FIG. 3.-Dependence of the apparent propagation rate constant on the poly(m-methylstyry1)sodium and ply@-methylstyry1)sodium concentration in MTHF at 25°C.H.HIROHARA, M. NAKAYAMA AND N. ISE 61 TABLE 2.-KINETIC DATA AND FIELD EFFECTS ON THE ANIONIC POLYMERIZATION OF m- AND p- SUBSTITUTED STYRENES AT 25°C E k;, kg K3 monomer /kV cm-1 IM-1 s-1 lM-3 s-1 gegenion, Cs': solvent, THF S 0 23 6.8 0.6 23 (8.4) 1.5 23 6.8 m-MS 0 20 2.9 0.6 20 2.9 1.5 20 2.9 p-iPS 0 -0 1.1 0.6 -0 1.3 p-MS 0 17 4.1 0.3 17 4.1 0.6 17 4.1 1 .O (-0) (4.0) gegenion, Na+: m-MS 0 1.5 3.0 4.5 1.5 3.0 4.5 p-MS 0 solvent, MTHF 8 0.71 8 0.82 8 0.97 8 0.97 5 0.25 5 0.28 5 0.36 5 0.36 3 c i3; 2 1 0 2 4 6 8 1 0 1O5[LE]AJ2/F FIG. 4.-The Fuoss conductivity plot for caesiwn salts of various living polymers in THF at 25°C. 0, polystyrene ; A, poly@-methylstyrene) ; 0, polyfm-methylstyrene) ; x , poly(p-isopropyl- styrene).62 HAMMETT’S RELATIONS IN ANIONIC POLYMERIZATIONS substituent group, except for p-ips.The field acceleration effects for sodium salts of poly(m- and p-methy1styrene)s in MTHF reached a limiting value at about 3 kV/ cm, as observed in the polystyrene ~ystems.~ The conductance data in both THF and MTHF as a function of log[LE] gave a straight line of a slope -3 irrespective of the applied field strength as observed in other systems.4* 5 9 1 1 * l2 I c c ’ 0 4 8 1 1 O6 [ LE]AJ2 / F FIG. 5.-The Fuoss conductivity plot for sodium salts of various living polymers in MTHF at 25°C. The method of Fuoss l 3 was used to determine K. The Fuoss plots for various living polymers are shown in fig. 4 and 5. The conductance of the sodium salt of polystyrene in MTHF, which was previously analyzed by the Kraus and Bray method was replotted by the Fuoss method in fig. 5.To calculate the parameters in the Fuoss relations, the limiting conductance (Ao) values, of appropriate polystyryl salts reported in previous papers 5 * l4 were used with an assumption that the effect of the substituent groups on A. is negligible. This is justified by the fact the conductance of living polymer anions is much smaller than that of alkali cations. The A,, l/A; K and K values are given in table 3 together with k: values in the absence of an electric field. Also the results for the casium salt of poly(p-methoxystyrene) and the sodium salt of polystyrene are given for comparison. DISCUSSION KINETIC BEHAVIOUR The k; value of the living polystyrene in THF is in good agreement with that reported by Schulz et aZ.lO.l 5 As all the selected substituent groups are electron- donating, the electron density of the carbanions of styrene derivatives must be higher than that of styrene, making stronger the interaction in the ion-pairs. Moreover, the substituent groups increase the electron density of the double bond of monomers.H. HIROHARA, M. NAKAYAMA A N D N . ISE 63 Living polymers of substituted styrenes are, thus, expected to be less dissociated and less reactive than living polystyrene. Except for the K values of caesium salts, this is really the case as shown in table 2 and 3, although the k; values in THF could not be determined very accurately because of the large slopes of the (kp, [LEI-%) plots. TABLE 3.-DISSOCJATION CONSTANT AND FREE ION RATE CONSTANT OF VARIOUS LIVING POLYMERS AT 25°C monomer A0 1 /A$K Kx 10-10 k; x 10-4 /cm-2 ohm-1 equiv.-l /M IM-1 s-1 gegenion, Cs+ : solvent, THF S 92.5 4 .3 ~ 104 27 13 m-MS 92.5 8.8 13 8.0 p-iPSa 92.5 2.8 <41 > 1.7 p-MS 92.5 4.3 27 7.9 p-MOS 92.5 4.3 27 3.9 gegenion, Na+ : solvent, MTHF S 58.0 1.3 x lo6 2.3 7.5 m-MS 58.0 1.8 1.7 5.5 p-MS 58.0 6.0 0.50 3.7 * see text for details It is very interesting to note that the free carbanions are half as reactive in MTHF as in THF. This might imply that the coordination of MTHF molecules to the free carbanions is stronger than that of THF molecules. The reactivity of living poly(p- isopropylstyrene) was abnormally small." This might be due to the steric hindrance due to the isopropyl group as suggested by Natta et al.l6 in the complex polymer- ization systems.The behaviour of K for the caesium salts in THF will be discussed later. FIELD EFFECT The field accleration effect was distinct in MTHF, whereas it was hardly per- ceptible in THF. From the insensitivity of A to field strength, it can be concluded that increase of k: is a primary factor for the observed accleration and is due to desolvation by an electric field of solvent molecules weakly bound to free carbanions, as proposed previously.2 ' 9 11* l 2 Some comments are needed on the absence of field effects in THF systems. From the steep increase of conductance in THF systems mentioned in the experimental section, it is believed that ionic impurities were produced in the electric field so that the field effect was marked.Moreover, it should be noted that the field effect was clearly observed to be saturated at 0.6- 1.0 kV/cm for living poly(p-methoxystyrene) in THF.ll A value of * 12 x lo4 M-' s-l was deduced for the rate constant of " desolvated " free ion of living polystyrene in previous studies.2* 4 9 The k; value (13 x lo4 M-l s-l ) found in the present work * Gas chromatographic analyses of p-isopropylstyrene showed the presenm of a small quantity of an impurity. An attempt to remove this impurity by distillation was not completely successful. At the start of the propagation reactions of poly(p-isopropylstyryl)caesium, a shoulder of absorption appeared at 450 mp in addition to the absorption peak (Amax) at 350 my. The shoulder disappeared rather quickly as the polymerization progressed whereas the peak at 350 mp was stable during the polymerization.Therefore, the rate constant was calculated and an electric field was applied in a conversion range between 50 and 85 % after the shoulder at 450mp had disappeared. In view of this complication, the results obtained for poly(p-isopropylstyry1)caesium should be viewed with caution. When the shoulder at 450 my existed, smaller kp and higher A values were obtained than in its absence.64 for living polystyrene in THF in the absence of an electric field was found to be nearly equal to the rate constant of “ desolvated ” growing ends. Thus, it can be concluded that the solvation of THF molecules to free polystyryl anion in the transition state of the reaction is extremely weak.The field intensity dependence of the free ion rate constant in THF is shown in fig. 6 together with the results obtained in other solvents. From fig. 6, we assert HAMMETT’S RELATIONS I N ANIONIC POLYMERIZATIONS ‘1 0 0 1.5 3D 4.5 E(kV/cm) FIG. 6.-Dependence of the free ion rate constant of polystyryl salt on field intensity in various solvents. 0, Li salt ; A, Na salt ; 13, K salts ; x , Cs salt. again that the k; value of (12+2) x lo4 M-l s-l is the rate constant of the desolvated free anion of living polystyrene. Furthermore, from the field effects observed in various systems, the order of coordination ability of various ethereal solvents to free carbanions decreases along the series DME > THF-MTHF > THF. This view seems to be supported by the fact that the reactivity of polystyryl free anions decreased along the series THF > MTHFETHP > DME in the absence of an electric field.HAMMETT’S RELATIONS The relative rate constants and dissociation constants of living polymers of sub- stituted styrene should obey the Hammett relations. This is demonstrated in fig. 7. The results for homopolymerization are plotted. Therefore, it should be noted that the reactivities of not only monomers but also of living active ends are reflected in the results, although it is found that the reactivity of very similar active ends is not significantly affected by structure.17 The data for both caesium and sodium salts are plotted in fig. 7 for comparison. Szwarc et al. have studied l8 addition reactions of a series of analogous monomers to sodium polystyryl in THF to give a reactivity constant p = 5.0 for the apparent propagation rate constant k,. The value of k , is not, however, very significant, since it is a function of kk, k$ and K.Accord- ingly we studied Hammett’s relations for &, and K, separately.H. HIROHARA, M . NAKAYAMA AND N . ISE 65 For the apparent rate constant, p = 1.7 was obtained for caesium salts in THF and p = 4.2 for sodium salts in MTHF. The latter value is in agreement with Szwarc’s data in spite of the difference in solvent. For k; and K, also, the p values were found to depend on the gegenion. It should be noted that the gegenion de- pendence of k, is a consequence of the gegenion dependences for kb and K. For k;, P = 0.5 -2.0 for caesium salts and 3.8 for sodium salts.For K, p = 0 for caesium A I I n 1 4 -0.5 -0.3 -Q2 -0.1 D 0 I (I FIG. 7.-Hammett plots for living poly(substituted styrene) at 25°C. x , sodium salt in MTHF. A, apparent rate constant ([LEI = 1 x 0, caesium salt in THF; M) ; B, ion-pair rate con- stant ; C, dissociation constant ; D, free ion rate constant. salts and 4.2 for sodium salts. These results show that the solvation state and inter- ionic distance of contact ion pairs of caesium salts are not affected by anions having the electron donating substituent group because of the large Pauling ionic radius of Csf ion. This explanation, however, might be questionable. As a matter of fact, poly(0-methylstyry1)caesium was much less dissociated in MTHF than polystyryl- caesium, although in THF the former salt was dissociated to the same extent as the latter.12 On the other hand, Na+ ion has a small Pauling ionic radius, so that the interionic distance of contact ion pairs will be small, and in the formation of solvent- separated ion pair its interionic distance will be large since Na+ ion is coordinated strongly with the ethereal solvents.Consequently, the solvation state and interionic distance of sodium ion pairs vary greatly with the anion. The relative reactivity of the free ions fell on the same line of a slope of 1.7 ir- respective of the solvent and the gegenion. This might be reasonable since k: should be independent of gegenion. The identity of the values implies that the relative66 HAMMETT'S RELATIONS IN ANIONIC POLYMERIZATIONS reactivity of free carbanions in THF is equivalent to that in MTHF although the magnitudes of k; are different in both solvents as shown in table 3. Finally, it is safe to conclude that, in the case of sodium salts, the k, values of various living polymers reflect the dissociation state of these polymers, which is determined by solvation state of the ion pairs. On the contrary, for caesium salts, the most important factor influencing the k, value is the reactivity of the free ions. e.g., N. he, Ad. Polymer Sci., 1969, 6, 347, in which previous work is referred to. N. Ise, H. Hirohara, T. Makino and I. Sakurada, J. Phys. Chem., 1968,72,4543. N. Ise, H. Hirohara, I. Makino, K. Takaya and M. Nakayama, J. Phys. Chem., 1970,74,606. H. Hirohara, M. Nakayama, K. Takaya and N. Ise, Trans. Favaday SOC., 1970,66,1165. H. Hirohara, K. Talcaya, M. Nakayama and N. be, Tram. Faraday Soc., 1970,66,3163. H. Hirohara, K. Takaya and N. Ise, Macromolecules, 1971,4,288. D. J. Worsfold and S . Bywater, J. Chem. SOC., 1960, 5234. D. N. Bhattacharyya, C. L. Lee, J. Smid and M. Szwarc, J. Phys. Chem., 1965,69,612. l o H. Hostalka and G. V. Schulz, 2. phys. Chern., 1965,45,286. K. Takaya, H. Hirohara, M. Nakayama and N. Ise, Trans. Faraby SOC., 1971,67, 119. l2 H. Hirohara, M. Nakayama, R. Kawabata and N. Ise, J.C.S. Fczraday I, 1972, 68, 51. l3 (a) R. M. Fuoss, J. Amer. Chem. Soc., 1935,57,488. ' K. Ziegler and H. Dislich, Ber., 1957,90, 1107. (b) R. M. Fuoss and F. Accascina, Electrolytic Conductance (John Wiley and Son Inc., New York, 1959), chap. 17. l4 J. Shimomura, K. J. Tolle, J. Smid and M. Szwarc, J. Amer. Chem. Soc., 1967,89, 796. l 5 B. J. Schmitt and G. V. Schulz, Makrornol Chem., 1971,142,325. l6 G. Natta, F. Danusso and D. Sianesi, Makromol. Chem., 1959,30,238. l' M. Szwarc, Carbanions, Living Polymers and Electron-Transfer Process (John Wiley and Son Inc., New York, 1968). M. Shima, D. N. Bhattacharyya, J. Smid and M. Szwarc, J. Amer. Chem. SOC., 1963,85,1306.

ISSN:0300-9599    

DOI:10.1039/F19726800058

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Interpretation of the Effects of Ionic Scavengers at High L.E.T. in Irradiated Cyclohexane, Based on Ambipolar Diffusion BY W. G. BURNS AND C. R. V. REED Chemistry Division, A.E.R.E., Harwell Received 17th June, 1971 Some results on the effects of ionic scavengers in cyclohexane irradiated at high L.E.T. are re- interpreted on the basis of ambipolar diffusion. It is shown that at high L.E.T., 200 keV pm-l, the density of ions is great enough for ion atmospheres to be set up, so that the diffusion of oppositely charged ions is mutually influenced (the Debye length is much smaller than the track radius), through- out the significant lifetime of the track. Expressions are derived for the relevant ambipolar diffusion constants with scavenging occurring, which change as scavenging and neutralization of ions take place.These, applied to a diffusion model of the track, yield curves of G(scavenged ions) against scavenger concentration which agree well with experiment. These concepts should be helpful in interpreting the scavenging of ions produced by the high L.E.T. component of y and fast electron irradiation. In a previous paper on the comparison of the effects of ionic scavengers at high and low L.E.T., it was suggested that the high L.E.T. case might be considered in terms of ambipolar diffusion. We present here a development of this concept, and a brief rCsumC of the considerations at low L.E.T. in order to facilitate com- parisons. The concept of ambipolar diffusion, introduced by Schottky,2 is well known in the field of densely ionized gases (ion concentration > lo8 ~ m - ~ ) .The density of positive ions n+ and that of electrons n- at any point not within one Debye length of the boundary become equal, since any imbalance results in an increased electric field due to the space charge separation, which restores the balance. The diffusion of elec- trons (diffusion constant 0,) is retarded by the slower diffusing positive ions (diffusion constant D+), which are caused to diffuse faster than at low ionic concentrations. Ions of both charges diffuse together with a common diffusion constant, the ambipolar diffusion constant. Its value with one kind each of positive and negative ions is obtained by equating the current densities of positive and negative particles, I?.+ and I?, respectively : where E is the space charge field and p+ and 1-1- the respective mobilities.With the conditions r+ = r- = I?*, and n+ = n- = nk, the elimination of E results in The ambipolar diffusion constant, from Fick's first law of diffusion, is therefore (D+pe + De,u+)/@+ +pe). In the low L.E.T. case we are concerned with isolated ion pairs which have a distribution in separation. This is obtained from calculations by Mozumder and Magee of the yield of track entities as a function of energy for a 1 MeV electron (table TI, ref. (5)) making allowance for the formation of &electrons by using their table 111, and converting from energy to separation at thermalization by using their table IV. The final distribution is given in table 1, and for each r+ = - D+ grad n+ + n+p+E, re = - De grad ne - YtePeE, r k = I- (D+pe + DeP+) grad n*110.1- + P+).6768 IONIC SCAVENGING separation the probability of scavenging is obtained, using the method of Hummel,6 from a treatment of Monchick,' which gives P(q, rt) = 2 exp ( - qr,)[ 1 - exp (- rc/r,)]/ { 1 + exp ( - 2qri) + 2qrPr; l[ 1 - exp (- rc/ri)]>, where P is the probability of ion recom- bination; q = (kscs/D')3, where k, is the scavenging rate-constant, c, the scavenger concentration, D' is the sum of diffusion constants for positive and negative ions; r, = e2/Kli.T is the Onsager critical escape distance. For a given q the G value for mean energy eV 12.5 17.5 22.5 27.5 32.5 37.5 42.5 47.5 52.5 57.5 62.5 67.5 72.5 77.5 82.5 87.5 92.5 97.5 125 175 225 275 325 375 425 475 750 1250 1750 2250 2750 3250 3750 4500 TABLE 1 thermalization separation, ri ( nm r, = 31.2 nm 7.25 7.60 7.95 8.20 8.45 8.65 8.80 9.10 9.35 9.55 9.80 10.00 10.25 10.45 10.65 10.83 10.93 1 1 .oo 11.16 12.30 13.20 14.37 15.52 16.97 18.50 20.22 32.50 95.00 140.0 215.0 303.9 406.1 521 .O 720.4 Y value including 6 electrons 1.115 0.7290 0.4123 0.1946 0.1264 0.0799 0.0573 0.0425 0.0350 0.0260 0.0193 0.0170 0.0144 0.0093 0.0089 0.0084 0.0076 0.0054 0.0374 0.0222 0.0132 0.01 18 0.0075 0.0020 0.0015 0.0012 0.0055 0.0014 O.OOO9 0.0008 O.OOO4 0.0003 0.0003 0.0003 scavenged ions is given by G(s.i) = CG,(l -P(q,r,)].Fig. 1 shows a graph of i G(s.i) against q2 = ac, = (k,/D')c, obtained in this way. The experimental points are for the decrease in G(H2) with nitrous oxide present, (negative ion scavenging), and for G(HD) in the presence of ND, (positive ion scavenging) as functions of c,.They were fitted to the curve by adjusting a and are plotted for a = 5.75 x cm for the positive ion scavenger ND3, and a = 0.83 x cm for the negative ion scavenger N20.W. G . BURNS AND C. R. V . REED 69 In the densely ionizing cases (mean L.E.T. 200keV/pm) the density of ions precludes their consideration as isolated pairs. The initial ionization will cause a separation of ions and electrons, but the resulting space charge will promote inter- mingling which will be followed by neutralization and outward diffusion. The criterion for ambip~larity,~'~ that the Debye length A is a small fraction of the qz = UC, in cm-2 FIG. 1.-G(scavenged ions) against q2 = ac, for lightly ionizing radiation. Curve calculated, see text.0, AG(H2) against 0.8 x lo-' [N20] ; 0, G(HD) against 5.75 x [NDJ. Concentrations in molecules CM-~. physical dimensions of the diffusing assembly, is fulfilled initially, since IE = (KkT/ 47re2n)*, where n is the ionic concentration so that A/r = (Kkt/4e2N)*, N being the number of ion pairs per unit length and r the track radius ; with W = 30 eV per ion pair and T = 300 K, A/r z 0.03. The condition would also be fulfilled initially even at T = lo00 K and is also fulfilled at T = 300 K when N has decreased to - 0.1 of its original value. Wilhelm * has treated ambipolar recombination and diffusion in the gas phase and has warned against the application of ambipolar diffusion in condensed media, chiefly on the ground that the positional uncertainty of the electrons is of the same order of magnitude as its nearest neighbour separation.Although recent experimentsg with very pure substances show very high mobilities for the negative charge carrier (up to - lo3 times that of the positive ion) other workers lo have found this concept difficult to incorporate into charge-scavenging kinetics, and earlier work l 1 suggested a charge carrier of molecular dimensions. If we assume that the negative species behaves as an ion the difficulty of positional un- certainty is reduced, and we proceed, with reservations, on this assumption. Ambipolar diffusion including ion scavenging has been considered by Oskam. The effective diffusion constants become time-dependent, as they depend on the relative concentrations of the different ions present, which change with time.The reactions which cause these changes are as follows for negative ion scavenging Mf + e-product 1 e+S+S- M+ + S-+product 2,70 IONIC SCAVENGING and for positive ion scavenging the last two reactions are replaced by M++S+S+ S+ + e4product 2, where M+ and e are the original positive and negative ion and S the scavenging molecule. In Oskam’s treatment, as in that with only one kind of positive and negative ion, the total positive and negative current densities were equated. Oskam simplified his equations by assuming that (grad n)/n is equal for all ions present, which was appropriate for his case. In our case, infinite outward cylindrical diffusion, an appropriate assumption is that of “ prescribed diffusion ” which gives (grad n)/n = 2r/b2, where bZ =r:+4Dt, yo being the initial radius, so that at a given t and r, the ratios of (grad n)/n for any pair of ions are the inverse ratios of their values of b2.Using subscripts +, e for the original positive and negative ions, and s for the ion produced on scavenging, the three ambipolar diffusion constants become : positive ion scavenging negative ion scavenging Da,e De-PePQ De - PePQ Da,, Ds + P~@%~)PQ D S - PS(b,2/b,2)PQ D,+ D+ + ~ + ( b : l b ; ) P Q D+ + ~ + ( b : / b , ” ) p Q P Q CCl -Y)P+ + P e + w s l - W+r)P+ +Pe+YPsl-l, De -Ds~(b,”/bs2> - D + (1 - y>(b,2/b”, D, + ~s~(b,2ib,2> - ~ + ( 1 + ~>(bf/b2,> where D and p represent diffusion constant and mobility, and y = n,/ne.We have combined these ambipolar diffusion constants with the extension of the Ganguly- Magee spur/track model to entities with differing diffusion constants, developed by Schwarz.13 The parameters were chosen as follows. The number of eV per spur k,c, in s-l FIG. Z.--G(scavenged ions) against kscs, electron scavenging for 100 eV spurs, 2 x lo7 spurs cm-l, G(ion pairs) = 3, recombination rate constants = 2 x cm3 ion-l s-l. Curve 1, yo = 10 nm, D+ = D, = D, = 2 X cm2 s-’, p+ = ps = pe ; curve l’, ro = 5 nm, other constants as for curve 1 . Curves 2 and 2’, D+ = D, = 0.2 De = 2x cm2 s-’, p+ = ps = 0.2 pe, otherwise as curves 1 and 1’ respectively.W. G . BURNS AND C . R . V . REED 71 was taken as 100, giving the number of spurs per cm, 2 = 2 x lo7, and G(ion pairs) was taken as 3.The recombination rate-constants were taken as 2 x cm3 ion-1 s-l, and the diffusion constants for the original positive ion and ions formed by scavenging as 2 x cm2 s-l ; these rate and diffusion constants are based on values measured in conductivity experiments." The mobilities of the original positive ion, and ions formed by scavenging were taken as equal. Variations were made in the ratio of the mobility of the original negative ion to that of the other ions, Be = p,/p+, in D,, and in k,c,. The initial track radii for all ionic entities were taken as equal, and varied between 5 and 10nm which represent reasonable limits for electron thermalization lengths. time in s FIG. 3.-Numbers of ions reacted or remaining per 100 eV, as marked on curves, functions of time.Conditions as for curve 2, fig. 2, with kscs = 10" s-l. Typical calculated curves for G(s . i) against kbcS are given in fig. 2. The following points are noteworthy; (a) curves for positive and negative ion scavenging are approximately the same for the same parameters and become identical when all diffusion constants are equal and p, = 1 ; (b) the effect of decreasing ro from 10 to 5 nm is to move the curve to values of kscs higher by a factor -4.5 ; (c) starting with all diffusion constants and mobilities equal, and increasing D, and Be in the same ratio the change shown in fig. 2 occurs at low values of G(s.i) for a increase in D, and be by a factor 5, but further change is small even when the factor becomes lo3. Fig. 3 shows the progress in time of electron scavenging at high k,c,.The fit obtained with experiment is shown in fig. 4. The points for electron scavenging, which are at high values of G(s.i) are fitted by all values of Be above 1 and D, above D+, and at ro = 10 nm require k, = 1.4 x cm3 molecule-l s-l. The points for positive ion scavenging fit best the curve for values of Be and Dell)+ of 5 and over, and give k, = 1.2 x 10-l2 cm3 molecule-l s-l at ro = 10 nm. It may not be fortuitous that the ratio of values of a for low L.E.T. positive and negative ion scavenging, viz., - 14, is approximately equal to the ratio of values of k, for high L.E.T. scavenging, viz., 12. Hummel has suggested that these ratios72 IONIC SCAVENGING of a are fairly large ( N lo), because for scavenger molecules of similar radius they are given by (D, + D,)/(D,+ D+), assuming diffusion-controlled reactions, and D, is somewhat larger than D+.Since at high L.E.T. for diffusion-controlled scavenger reactions a similar ratio holds and k,(negative)/k,(positive) is high, a high value of L c - 109 ton Id' 10-3 I 1 1 1 1 1 1 1 ' I 1 1 1 1 1 1 1 ' 1 1 1 1 1 1 1 1 ' I 1 1 1 1 1 1 1 ' 1 f i 1 1 1 1 1 1 ' I I I I I ' I 0' 106 10' 10' kscs in s-l FIG. 4.-0, AG(H,) against 1 . 4 ~ 10-I' [N,O] for 2 MeV 4He ions, curve as curve 2, fig. 2 ; curve 1 in this region not significantly different ; 0, G(HD) against 1.2 x 10-l2 [ND3] for 3 MeV 4He ions ; curve as curve 2, fig. 2 ; concentrations in molecules ~ r n - ~ . D, is implied rather than Da,e, which in general, like the case with no scavenging is not greater than 2D+.This suggests that although the collective diffusion of the assembly of ions is ambipolar, the ion atmosphere restriction does not influence the microscopic diffusive approach of the original ion and a neutral molecule. The concept of ambipolar diffusion in high L.E.T. scavenging appears to be useful, and may help the interpretation of the low L.E.T. case at high scavenger concentra- tions at which the ions of the higher L.E.T. component of low L.E.T. radiation are scavenged. W. G. Burns and C. R. V. Reed, in Uses of Cyclotrons in Chemistry, Metallurgy and Biology, ed. C. B. Amphlett (Butterworths, London, 1970), p. 61. W. Schottky, Phys. Z., 1924,25, 342. E. W. McDaniel, CoZZision Phenomena in Ionized Gases (John Wiley and Sons, Inc., New York, 1964), p.512. ref. (3), p. 693. A. Mozumder and J. L. Magee, J . Chem. Phys., 1967,47,939. A. Hummel, J. Chem. Phys., 1968,49,4840. ' L. Monchick, J. Chem. Phys., 1956,24, 381. H. E. Wilhelm, J . Chem. Phys., 1967,47,4356. W. F. Schmidt and A. 0. Allen, J . Chem. Phys., 1970,52,4788. S . J. Rzad, P. P. Infelta, J. M. Warman and R. H. Schuler, J . Chem. Phys., 1970,52, 3971. A. Hummcl, Thesis (Free University, Amsterdam, 1967). l2 H. J. Oskam, PhirIips Res. Reports, 1958, 13, 358. l3 H. A. Schwarz, J. Phys. Chem., 1969,73, 1928. Interpretation of the Effects of Ionic Scavengers at High L.E.T. in Irradiated Cyclohexane, Based on Ambipolar Diffusion BY W. G. BURNS AND C. R. V. REED Chemistry Division, A.E.R.E., Harwell Received 17th June, 1971 Some results on the effects of ionic scavengers in cyclohexane irradiated at high L.E.T. are re- interpreted on the basis of ambipolar diffusion.It is shown that at high L.E.T., 200 keV pm-l, the density of ions is great enough for ion atmospheres to be set up, so that the diffusion of oppositely charged ions is mutually influenced (the Debye length is much smaller than the track radius), through- out the significant lifetime of the track. Expressions are derived for the relevant ambipolar diffusion constants with scavenging occurring, which change as scavenging and neutralization of ions take place. These, applied to a diffusion model of the track, yield curves of G(scavenged ions) against scavenger concentration which agree well with experiment.These concepts should be helpful in interpreting the scavenging of ions produced by the high L.E.T. component of y and fast electron irradiation. In a previous paper on the comparison of the effects of ionic scavengers at high and low L.E.T., it was suggested that the high L.E.T. case might be considered in terms of ambipolar diffusion. We present here a development of this concept, and a brief rCsumC of the considerations at low L.E.T. in order to facilitate com- parisons. The concept of ambipolar diffusion, introduced by Schottky,2 is well known in the field of densely ionized gases (ion concentration > lo8 ~ m - ~ ) . The density of positive ions n+ and that of electrons n- at any point not within one Debye length of the boundary become equal, since any imbalance results in an increased electric field due to the space charge separation, which restores the balance.The diffusion of elec- trons (diffusion constant 0,) is retarded by the slower diffusing positive ions (diffusion constant D+), which are caused to diffuse faster than at low ionic concentrations. Ions of both charges diffuse together with a common diffusion constant, the ambipolar diffusion constant. Its value with one kind each of positive and negative ions is obtained by equating the current densities of positive and negative particles, I?.+ and I?, respectively : where E is the space charge field and p+ and 1-1- the respective mobilities. With the conditions r+ = r- = I?*, and n+ = n- = nk, the elimination of E results in The ambipolar diffusion constant, from Fick's first law of diffusion, is therefore (D+pe + De,u+)/@+ +pe).In the low L.E.T. case we are concerned with isolated ion pairs which have a distribution in separation. This is obtained from calculations by Mozumder and Magee of the yield of track entities as a function of energy for a 1 MeV electron (table TI, ref. (5)) making allowance for the formation of &electrons by using their table 111, and converting from energy to separation at thermalization by using their table IV. The final distribution is given in table 1, and for each r+ = - D+ grad n+ + n+p+E, re = - De grad ne - YtePeE, r k = I- (D+pe + DeP+) grad n*110.1- + P+). 6768 IONIC SCAVENGING separation the probability of scavenging is obtained, using the method of Hummel,6 from a treatment of Monchick,' which gives P(q, rt) = 2 exp ( - qr,)[ 1 - exp (- rc/r,)]/ { 1 + exp ( - 2qri) + 2qrPr; l[ 1 - exp (- rc/ri)]>, where P is the probability of ion recom- bination; q = (kscs/D')3, where k, is the scavenging rate-constant, c, the scavenger concentration, D' is the sum of diffusion constants for positive and negative ions; r, = e2/Kli.T is the Onsager critical escape distance.For a given q the G value for mean energy eV 12.5 17.5 22.5 27.5 32.5 37.5 42.5 47.5 52.5 57.5 62.5 67.5 72.5 77.5 82.5 87.5 92.5 97.5 125 175 225 275 325 375 425 475 750 1250 1750 2250 2750 3250 3750 4500 TABLE 1 thermalization separation, ri ( nm r, = 31.2 nm 7.25 7.60 7.95 8.20 8.45 8.65 8.80 9.10 9.35 9.55 9.80 10.00 10.25 10.45 10.65 10.83 10.93 1 1 .oo 11.16 12.30 13.20 14.37 15.52 16.97 18.50 20.22 32.50 95.00 140.0 215.0 303.9 406.1 521 .O 720.4 Y value including 6 electrons 1.115 0.7290 0.4123 0.1946 0.1264 0.0799 0.0573 0.0425 0.0350 0.0260 0.0193 0.0170 0.0144 0.0093 0.0089 0.0084 0.0076 0.0054 0.0374 0.0222 0.0132 0.01 18 0.0075 0.0020 0.0015 0.0012 0.0055 0.0014 O.OOO9 0.0008 O.OOO4 0.0003 0.0003 0.0003 scavenged ions is given by G(s.i) = CG,(l -P(q,r,)].Fig. 1 shows a graph of i G(s.i) against q2 = ac, = (k,/D')c, obtained in this way. The experimental points are for the decrease in G(H2) with nitrous oxide present, (negative ion scavenging), and for G(HD) in the presence of ND, (positive ion scavenging) as functions of c,. They were fitted to the curve by adjusting a and are plotted for a = 5.75 x cm for the positive ion scavenger ND3, and a = 0.83 x cm for the negative ion scavenger N20.W.G . BURNS AND C. R. V . REED 69 In the densely ionizing cases (mean L.E.T. 200keV/pm) the density of ions precludes their consideration as isolated pairs. The initial ionization will cause a separation of ions and electrons, but the resulting space charge will promote inter- mingling which will be followed by neutralization and outward diffusion. The criterion for ambip~larity,~'~ that the Debye length A is a small fraction of the qz = UC, in cm-2 FIG. 1.-G(scavenged ions) against q2 = ac, for lightly ionizing radiation. Curve calculated, see text. 0, AG(H2) against 0.8 x lo-' [N20] ; 0, G(HD) against 5.75 x [NDJ. Concentrations in molecules CM-~. physical dimensions of the diffusing assembly, is fulfilled initially, since IE = (KkT/ 47re2n)*, where n is the ionic concentration so that A/r = (Kkt/4e2N)*, N being the number of ion pairs per unit length and r the track radius ; with W = 30 eV per ion pair and T = 300 K, A/r z 0.03.The condition would also be fulfilled initially even at T = lo00 K and is also fulfilled at T = 300 K when N has decreased to - 0.1 of its original value. Wilhelm * has treated ambipolar recombination and diffusion in the gas phase and has warned against the application of ambipolar diffusion in condensed media, chiefly on the ground that the positional uncertainty of the electrons is of the same order of magnitude as its nearest neighbour separation. Although recent experimentsg with very pure substances show very high mobilities for the negative charge carrier (up to - lo3 times that of the positive ion) other workers lo have found this concept difficult to incorporate into charge-scavenging kinetics, and earlier work l 1 suggested a charge carrier of molecular dimensions.If we assume that the negative species behaves as an ion the difficulty of positional un- certainty is reduced, and we proceed, with reservations, on this assumption. Ambipolar diffusion including ion scavenging has been considered by Oskam. The effective diffusion constants become time-dependent, as they depend on the relative concentrations of the different ions present, which change with time. The reactions which cause these changes are as follows for negative ion scavenging Mf + e-product 1 e+S+S- M+ + S-+product 2,70 IONIC SCAVENGING and for positive ion scavenging the last two reactions are replaced by M++S+S+ S+ + e4product 2, where M+ and e are the original positive and negative ion and S the scavenging molecule.In Oskam’s treatment, as in that with only one kind of positive and negative ion, the total positive and negative current densities were equated. Oskam simplified his equations by assuming that (grad n)/n is equal for all ions present, which was appropriate for his case. In our case, infinite outward cylindrical diffusion, an appropriate assumption is that of “ prescribed diffusion ” which gives (grad n)/n = 2r/b2, where bZ =r:+4Dt, yo being the initial radius, so that at a given t and r, the ratios of (grad n)/n for any pair of ions are the inverse ratios of their values of b2.Using subscripts +, e for the original positive and negative ions, and s for the ion produced on scavenging, the three ambipolar diffusion constants become : positive ion scavenging negative ion scavenging Da,e De-PePQ De - PePQ Da,, Ds + P~@%~)PQ D S - PS(b,2/b,2)PQ D,+ D+ + ~ + ( b : l b ; ) P Q D+ + ~ + ( b : / b , ” ) p Q P Q CCl -Y)P+ + P e + w s l - W+r)P+ +Pe+YPsl-l, De -Ds~(b,”/bs2> - D + (1 - y>(b,2/b”, D, + ~s~(b,2ib,2> - ~ + ( 1 + ~>(bf/b2,> where D and p represent diffusion constant and mobility, and y = n,/ne. We have combined these ambipolar diffusion constants with the extension of the Ganguly- Magee spur/track model to entities with differing diffusion constants, developed by Schwarz.13 The parameters were chosen as follows.The number of eV per spur k,c, in s-l FIG. Z.--G(scavenged ions) against kscs, electron scavenging for 100 eV spurs, 2 x lo7 spurs cm-l, G(ion pairs) = 3, recombination rate constants = 2 x cm3 ion-l s-l. Curve 1, yo = 10 nm, D+ = D, = D, = 2 X cm2 s-’, p+ = ps = pe ; curve l’, ro = 5 nm, other constants as for curve 1 . Curves 2 and 2’, D+ = D, = 0.2 De = 2x cm2 s-’, p+ = ps = 0.2 pe, otherwise as curves 1 and 1’ respectively.W. G . BURNS AND C . R . V . REED 71 was taken as 100, giving the number of spurs per cm, 2 = 2 x lo7, and G(ion pairs) was taken as 3. The recombination rate-constants were taken as 2 x cm3 ion-1 s-l, and the diffusion constants for the original positive ion and ions formed by scavenging as 2 x cm2 s-l ; these rate and diffusion constants are based on values measured in conductivity experiments." The mobilities of the original positive ion, and ions formed by scavenging were taken as equal.Variations were made in the ratio of the mobility of the original negative ion to that of the other ions, Be = p,/p+, in D,, and in k,c,. The initial track radii for all ionic entities were taken as equal, and varied between 5 and 10nm which represent reasonable limits for electron thermalization lengths. time in s FIG. 3.-Numbers of ions reacted or remaining per 100 eV, as marked on curves, functions of time. Conditions as for curve 2, fig. 2, with kscs = 10" s-l. Typical calculated curves for G(s . i) against kbcS are given in fig. 2. The following points are noteworthy; (a) curves for positive and negative ion scavenging are approximately the same for the same parameters and become identical when all diffusion constants are equal and p, = 1 ; (b) the effect of decreasing ro from 10 to 5 nm is to move the curve to values of kscs higher by a factor -4.5 ; (c) starting with all diffusion constants and mobilities equal, and increasing D, and Be in the same ratio the change shown in fig.2 occurs at low values of G(s.i) for a increase in D, and be by a factor 5, but further change is small even when the factor becomes lo3. Fig. 3 shows the progress in time of electron scavenging at high k,c,. The fit obtained with experiment is shown in fig. 4. The points for electron scavenging, which are at high values of G(s.i) are fitted by all values of Be above 1 and D, above D+, and at ro = 10 nm require k, = 1.4 x cm3 molecule-l s-l.The points for positive ion scavenging fit best the curve for values of Be and Dell)+ of 5 and over, and give k, = 1.2 x 10-l2 cm3 molecule-l s-l at ro = 10 nm. It may not be fortuitous that the ratio of values of a for low L.E.T. positive and negative ion scavenging, viz., - 14, is approximately equal to the ratio of values of k, for high L.E.T. scavenging, viz., 12. Hummel has suggested that these ratios72 IONIC SCAVENGING of a are fairly large ( N lo), because for scavenger molecules of similar radius they are given by (D, + D,)/(D,+ D+), assuming diffusion-controlled reactions, and D, is somewhat larger than D+. Since at high L.E.T.for diffusion-controlled scavenger reactions a similar ratio holds and k,(negative)/k,(positive) is high, a high value of L c - 109 ton Id' 10-3 I 1 1 1 1 1 1 1 ' I 1 1 1 1 1 1 1 ' 1 1 1 1 1 1 1 1 ' I 1 1 1 1 1 1 1 ' 1 f i 1 1 1 1 1 1 ' I I I I I ' I 0' 106 10' 10' kscs in s-l FIG. 4.-0, AG(H,) against 1 . 4 ~ 10-I' [N,O] for 2 MeV 4He ions, curve as curve 2, fig. 2 ; curve 1 in this region not significantly different ; 0, G(HD) against 1.2 x 10-l2 [ND3] for 3 MeV 4He ions ; curve as curve 2, fig. 2 ; concentrations in molecules ~ r n - ~ . D, is implied rather than Da,e, which in general, like the case with no scavenging is not greater than 2D+. This suggests that although the collective diffusion of the assembly of ions is ambipolar, the ion atmosphere restriction does not influence the microscopic diffusive approach of the original ion and a neutral molecule. The concept of ambipolar diffusion in high L.E.T. scavenging appears to be useful, and may help the interpretation of the low L.E.T. case at high scavenger concentra- tions at which the ions of the higher L.E.T. component of low L.E.T. radiation are scavenged. W. G. Burns and C. R. V. Reed, in Uses of Cyclotrons in Chemistry, Metallurgy and Biology, ed. C. B. Amphlett (Butterworths, London, 1970), p. 61. W. Schottky, Phys. Z., 1924,25, 342. E. W. McDaniel, CoZZision Phenomena in Ionized Gases (John Wiley and Sons, Inc., New York, 1964), p. 512. ref. (3), p. 693. A. Mozumder and J. L. Magee, J . Chem. Phys., 1967,47,939. A. Hummel, J. Chem. Phys., 1968,49,4840. ' L. Monchick, J. Chem. Phys., 1956,24, 381. H. E. Wilhelm, J . Chem. Phys., 1967,47,4356. W. F. Schmidt and A. 0. Allen, J . Chem. Phys., 1970,52,4788. S . J. Rzad, P. P. Infelta, J. M. Warman and R. H. Schuler, J . Chem. Phys., 1970,52, 3971. A. Hummcl, Thesis (Free University, Amsterdam, 1967). l2 H. J. Oskam, PhirIips Res. Reports, 1958, 13, 358. l3 H. A. Schwarz, J. Phys. Chem., 1969,73, 1928.

ISSN:0300-9599    

DOI:10.1039/F19726800067

出版商:RSC    

年代:1972

数据来源: RSC

摘要:

Ion Exchange Involving Several Groups of Homogeneous Sites BY R. M. BARRER AND J. KLINOWSKI Physical Chemistry Laboratories, Chemistry Dept., Imperial College, London S. W.7 Received 22nd June, 197 1 Crystalline exchangers such as zeolites may contain more than one group of different exchange sites. The theoretical basis of these multi-site exchangers has therefore been examined. Overall exchange isotherms and selectivities have been expressed in terms of the properties o f j different kinds of co-existent site. The resulting plots of log K;1 against equivalent cation fraction of the entering ion (Kielland plots) are complex, showing maxima, minima, inflection points and sigmoid shapes according to the proportions and properties of the several groups of sites. It is concluded that linear Kielland plots in such circumstances can at best only be approximations. Even where each group of sites behaves ideally the overall exchange isotherm can be strongly non-ideal.Where Kielland coefficients C1 are not zero for the overall isotherm, this isotherm may show maxima and minima (and hence limited solid solubility of the end members of the exchange), the criterion for which is a positive value of C1 which varies according to the valence of the exchanging ions. By an appropriate choice of quantities characterizing each of two groups of sites Kielland plots and isotherms were found, as an example, for chabazite which closely agreed with those determined experimentally. It has been established by X-ray crystallographic studies that in zeolite ion exchangers the exchangeable ions may be present in each of several different crystallo- graphic environments.' 9 This can result in distinct groups of sites, each group being homogeneous but different from the others. There has been no full examination of the effects of this upon the properties of the overall exchange isotherms and selectivities, although several limited approaches to the problem have been made. Barrer and Meier successfully analyzed the sigmoid Ag+Na exchange isotherm in zeolite A in terms of two kinds of ion site, 12 of one kind to every one of the other. Successful analyses of a number of isotherms in zeolite K-F have been made by Barrer and Munday.' In this case the ratio of the numbers of sites of each kind was about 2 : 1. When the logarithm of the corrected selectivity quotient, K, (expressed as in eqn (3)) is plotted against the equivalent cation fraction of one of the competing ions, curves of considerable complexity are often found, including some with clear maxima (e.g., NaST1 and all uni-divalent exchanges investigated in chabazite ; and uni- divalent exchanges in zeolites A,7 Y and X 9).It is possible, though unlikely, that the maxima arise from side reactions such as hydrolysis of the zeolite phase. Whether or not hydrolysis can be a factor it is interesting to find whether and under what conditions both maxima and minima can arise as a result purely of the presence in a zeolite of crystallographically different kinds of exchange site. For the above reasons, an examination was made of the theoretical basis of multi-site ion exchange.THE OVERALL EXCHANGE EQUILIBRIUM Exchange equilibrium must first be formulated by considering all kinds of exchange sites together. The exchanger is non-ideal in the sense that the activity coefficients of the ions in the exchanger are not unity and depend upon the cation composition of the exchanger. The general exchange reaction is74 ION EXCHANGE INVOLVING HOMOGENEOUS SITES where ions of species A have a charge ZA+ and of B having a charge ZB t are involved. The subscripts Z and S refer to zeolite and solutions, respectively. The rational thermodynamic equilibrium constant K, is then where A , and Bz are equivalent cation fractions of A and B in the zeolite and m t , mA, itre these concentrations in the solution in g ions per kg of solvent.fk, fg, yA and yB are the corresponding activity coefficients. We now define a quotient K,, related to ion selectivity, by K , = Ap(mF)zAyZA ---.!.- - AZ,"(aF)ZA BgA( mc)zByp - BgA(a$)ZB' (3) It is then possible to express fA andf, in terms of K,, which can be measured. The general relation between thefand K, are given later (eqn (1 6)). If we replace molalities in solution by equivalent fractions As and B,, then m,A = AS -(Z,m,A+Z,m~) ZA BS ZB m: = -(ZAme+ZBm,B), and so Accordingly, where (4) EXCHANGE EQUILIBRIUM INVOLVING SEVERAL HOMOGENEOUS GROUPS OF SITES The exchanger containsj types of site, the sites within each group being equivalent, but differing from sites in another group. The overall exchange isotherm as given in the previous section is to be related to the isotherms on the individual sites.Io Let quantities relating to the ith type of site bear the subscript i.For exchange equilibrium on these sites, where Ki is the thermodynamic equilibrium constant. From eqn (2) and (7) the condition for equilibrium is K,B;A EA? where E = ( f %>zB/(f3zA; (8) K ~ B ~ K,B;A E,A;B E ~ A ~ Z B ' -- ... - - --- ... - - Ei = (f f E j = (f f (9)R. M. BARRER AND J . KLINOWSKI 75 In terms of the quotient K, defined by eqn (3), the condition of eqn (8) becomes (10) -- (Kc) jBjzA ... - -- - ... --- - (Kc)iBZA - K , B , Z ~ A? AZ” A? ’ THE KIELLAND SELECTIVITY QUOTIENT The quotient K, of eqn (3), which is that often measured experimentally, will be considered further. The relationships between AZ and the A i and between Bz and the Bi are : is the ratio of total equivalents of cations in sites of type i to total equivalents in the zeolite.rnf and rn? are molalities of ions A and B in sites of type i (mol kg-l of zeolite). One may substitute eqn (1 1) in eqn (3), use eqn (7) to obtain Al in terms of Bi and so re-write eqn (3) in the form j j K , = [C piB~AizB(Kc)f’ZB]ZB/[~ piBilzA 1 1 or, in terms of Al, (13) Eqn (13) forms the basis for the subsequent calculations of the shapes of the Kielland plots for the overall exchange (i.e., plots of log K, against AZ or Bz), assuming different physically reasonable values of the pi and the Ki together with simple concentration dependences of (fr)zA/CfA)ZB which accord with previous experience l 1 and with a theoretical treatment.The expressions (13) and (14) for K, are ratios of two polynomials. The result of this division is another polynomial with an infinite number of terms in Ai or Bi. Since the Al and Bi are functions of AZ and Bz (eqn (8)) this means that even when all sites are ideal (El = E2 = . . . = Ej = 1) the plot of log K, against AZ or Bz cannot in general be a horizontal or even a straight line. For a zeolite composed of one type of ideal site, K, = Kl = K,. This situation is seldom encountered, but was closely approached for some exchanges in sodalite hydrate. l2 This zeolite provides only one kind of cation position. In the most general case, the value of K, for the overall reaction will depend on the properties of the term E = ( f , A > Z B / ( f g ) z ~ .Gaines and Thomas l3 obtained for this ratio the expression log E = 0.4343 (ZB-ZJ- log Kc+ log K , dAZ. (15) s: This equation omits a term involving the change in activity of zeolitic water during exchange since in the analogous case of the clay minerals this term was estimated to76 ION EXCHANGE INVOLVING HOMOGENEOUS SITES L 1 = exp {(&-2,)-2.3 5 .c.> ' L, = exp (ZB-ZA)+2.3 C,}. be small. To evaluate E, log K, must be known as a function of AZ. Whatever this dependence is, it can be represented as a power series of the type log K , = Co+2C1Az+3C2Ag+ . . . +(n+l)C,A;, (16) where the C are constant coefficients. Substituting eqn (16) in eqn (15) gives E = exp ((ZB-ZA)+2.3[cl(1-2AZ)$C,(1-3A%)f . +cn(l-(n+l)A",]]. (17) Further, since K,E = K,, then while, from Gaines and Thomas's relations for single ion activity coefficients, log K, = 0.4343(2, -2J + Co + C1+ .. . + C,, (18) log (f,")zB = 0.4343(2B-ZA)BZ-B,(2C,AZ+ . . +(n+ l)CnA;)+ 1 (19) Cl(l-A;)+ . . . +C,(l-A;+'), 1% (.f,")"* = - o.4343(zB - ZA)Az + Az(2C1A, + . . . +(n + l)C,A;) - n AZ,Ai+1 0 [$ ZA 0 IZA -2.3 lim K , = exp 2.3 1 (n + l)C, = pi exp - 1 (n + 1)& lim K , = exp 2.3c0 = pi exp -Co,i , ; (21) MAXIMA A N D MINIMA I N PLOTS OF log K, AGAINST AZ OR Bz From the relationship K, = K,E-l the condition for extrema in plots of K, or log K, against AZ is dK,/dAZ = Ka d(E-')/dAZ = 0.R. M. BARRER AND J . KLINOWSKI 77 But E according to eqn (17) is an exponential function of AZ, i.e., E = exp q5(AZ), and so the above condition becomes d4/dAZ = 0.(22) By considering increasing numbers of coefficients C in the exponential of eqn (17) the conditions associated with extrema can be found. These can first appear for C1, C, # 0 but C3, ..., C, = 0, since d#/dAZ = -2.3 (2C1 +6C2Az) and is zero when AZ = -C,/3C2. Because O<A,< 1 the ratio -Cl/3C, must also lie within these limits and so CI and C, must have opposite signs. Similarly, when C1, C2, C3 # 0 but C,, . . . , C,, = 0 the conditions for extrema are with Ci <:CiC3. Plots of log K, against AZ are shown in fig. 1 for typical choices of the coefficients C. A, FIG. 1.-Plots of log,, KC against A, for various values of three coefficients C,. The thermodynamic equilibrium constant Ka = 10. (a) (1) Cz = C2 = C3 = 0; (2) C, = 0.3, c, = c3 = 0; (3) c1 = -0.3, Cz = Cs = 0.-0.3, Cz = C, = 0.3. (b) (1) C1 = -0.3, C2 = 0.3, Cs = 0; (2) C1 = Cz = 0.3, C3 == 0 ; (3) C1 = (c) (1) C1 = Cz = 0.3, C3 = -0.3; (2) C1 = 0.3, Cz = -0.3, C3 = 0. If we consider the zeolite specifically in terms of its j homogeneous groups of sites, the condition for extrema can be found from eqn (13) and (10). After some manipulation, the condition is = A,, $ piAidlnBz ZA dlnB, d In Bi ( +A d In (KJi where By substituting in eqn (23) and bearing in mind that In = -2.3 Bi n(n+ l)Cn,iA!n-l), d In Bi 178 ION EXCHANGE INVOLVING HOMOGENEOUS SITES j C PiAi . we get the condition for extrema as I A; 1 ZAAZ +ZBBZ (24) - - 2.3B. Ai 1 - 2 n(n + l)Cn,iAy- ( z A 1 ZAAi+ZBBi-2.3BiAi C n(n+ l)Cn,iA:-' F o r j groups of ideal sites all Cn,i are zero and eqn (24) simplifies to give the condition for extrema as It can be formally demonstrated (appendix 1) that this equation leads, for 2, = ZB, to the condition A l = A2 = .. . A j , which is never fulfilled since ideal isotherms can never cross. Hence, no extrema are possible for the j groups of ideal sites, however large j is and however different the Ki. EXTREMA AND CROSSING POINTS I N ISOTHERMS Eqn (5) for the ion-exchange isotherm can be re-arranged to F(AZ, As) = KL(1- AZ)'*A2 - A 2 ( 1 - As)ZAEI' = 0. (25) The condition for extrema in the isotherm contour and hence for immiscibility gaps (appendix 2) in solid solutions of the end members of the exchange (pure A- and B-forms of the zeolite, respectively) is -- dAs - --(W~AZ)A, = 0. dAZ (aFfaAS)A, Since (aF/aA& is finite this condition requires that (aF/dA,),, = 0.eqn (25) and (17), one obtains after dfferentiation of each and some rearrangement Then, from n ZAA,+Z,(1-Az)-2.3(1-AZ)~ n(n+l)CnAz = 0. (26) 1 This relation cannot be satisfied for the ideal case where all coefficients C1 to C, are zero, so that here, as expected, no extrema are possible. The next simplest case is that where C, # 0 but C, = . . . = C, = 0 (linear Kielland plots). Then for ZA # Z B , with 18.4zBc1/[4.6cl +(ZB-z2,)]' < 1 (28) for real roots. The maximum corresponds with the smaller value of AZ, the minimum with the larger. When 2, # ZB, the maximum and minimum are equidistant from AZ = 3- (ZB+ZJ/9.2C,, and for 2, = 2, they are equidistant from AZ = 4 in both cases by distances dAZ = (1 /9.2C1) ,/[4.6C1 + (ZB-zA)]" - 18.4ZBC1.The necessary in- equality (28) can be rearranged to the form c1 > (ZA+ZBf J42~&)/4.6. The inequality associated with the negative sign appears to arise from the quadratic procedure involving taking a square root, since in the case ZA = 2, quadratic pro-R. M . BARRER AND J . KLINOWSKI 79 cedure can be avoided and only the inequality associated with the positive sign then appears. The requirements for C1 are given for the positive sign in table 1 for several values of ZA and 2,. TABLE 1.-vALUES WHICH c1 MUST EXCEED FOR EXTREMA TO APPEAR ZA ZB c1 1 1 > 0.87 1 2 >(3+ 2/8)/4.6 = 1.27 2 1 1 3 >(4+ 2/12)/4.6 = 1.62 3 1 2 2 > 1.74 2 3 >(5+ 1/24)/4.6 = 2.16 3 2 3 3 >2.61 Fig. 2 shows isotherm contours calculated from eqn (5) with r = 1 for different ZA and 2, and for several values of C1.In these diagrams, all isotherms for which 2, and 2, and the Ki are the same pass through a common point. The requirement for this to happen, whether 2, and 2, are equal or not, is that at the crossing point A, FIG. 2.-Exchange isotherms for ions of valence ZA with ions of valence ZB with C1 # 0; Cz, Ct, . . . , c n = 0 ; and Ka = 1. (a) ZB = ZA = 1 ; (b) ZB = 1, ZA = 2 ; (C) ZB = 1, ZA = 3 ; (d) ZB = 2, ZA = 3. For curves 1, C1 = - 1 ; 2, C1 = 0; 3, C1 = 1 ; 4, C1 = 2; 5 , C1 = 3. Isotherms showing maxima and minima have ranges of composition A z over which there is metastability. Here there is limited miscibility of the end members of the exchange.80 ION EXCHANGE INVOLVING HOMOGENEOUS SITES E must be independent of all the coefficients C involved in eqn (17) and equal to exp (ZB-ZA).No general condition can be laid down for such behaviour but examples in particular cases are given in table 2. The value of As at the crossing point is to be found from that for AZ and the relation For example, when ZB = 2, = 1 ; C1 # 0 ; Cz to C, = 0, As = 1/(1 +Ki/r), since AZ = 3 (table 2). Some consequences of immiscibility gaps on the isotherms are discussed in appendix 2. TABLE 2.-REQUIREMENTS FOR E TO BECOME INDEPENDENT OF COEFFICIENTS c conditions for C value of A Z for E = exp (ZB-ZA) subsidiary conditions - c1 # 0 ; c2 to c, = 0 C, # 0; C1 and C3 to C, = 0 c3 # O ; C1, C2 and C4 to C, = 0 c1 # 0, c2 # 0; c3 to c, = 0 1 12 1 / 4 3 1 /w - C1 + 4 C + 3 C2(Cl + C2) 3cz - - (i) Cl /C2 > - 1 (ii) the values of C1 and C, can vary but the ratio Cl/C2 must be constant.DISCUSSION Eqn (13) and (14) relate the overall corrected selectivity coefficient of the zeolite to the properties of individual sites. If we specify the thermodynamic equilibrium constants K/ for each type of site and the coefficients Cn,i in the polynomial #(A,) describing the non-ideality of each type of site, then the values of Ai are roots of the set of j implicit equations (where j = number of kinds of site) : Kf(1- Ai)'*(AS)'" -(Ai>'"(1- A,)ZAT exp {+(Ai)) = 0. It was assumed throughout that I' = 1. This assumption is reasonable for solutions of low ionic strength. The solution of the above equation very much depends on the coefficients Cn,r in #(Ad), i.e., on the type of polynomial used.Many calculations were made for C1, = 0 and C2, i , . . . ? C,, = 0 for each type of site, which corresponds with the linear dependence of the log (K& on the A t . It does not follow, however, that the overall exchange for the zeolite can then be treated with C1 # 0 and C, . . . , c, = 0. The numerical calculations were made for uni-univalent and mi-divalent exchanges and different numbers of types of site. All calculations were made on the Imperial College IBM 6600 computer using a Fortran IV programme SITE written for this purpose. The programme solves numerically the implicit eqn (29) with any required degree of accuracy. In the present work, all values of A , were obtained with an error of less than K, was obtained from eqn (13).Fig. 3 shows the ion-exchange isotherms and the plots of log K, against AZ for one, two, three and four types of site, each type behaving ideally (all coefficients Cn,i zero except Co,r). The Kielland plots for the overall exchange isotherm are sigmoid in shape and do not show extrema. The larger the difference of the effective rational thermodynamic equilibrium constants K/ for individual sites, the moreR. M. BARRER AND J . KLINOWSKI 81 sigmoid the plots. Conversely, for close values of K/ the plots of log K, against AZ approach a straight line. The plot of log K, f o r j = 1 is horizontal but for j = 2,3, 4 . . . these plots are curves, despite the ideality of all groups of sites. Thus, in the multi-site theory of ion exchange in zeolites the behaviour of the zeolite as a whole can never be considered ideal.This conclusion was anticipated from the algebraic discussion of eqn (13) and eqn (14). Furthermore, it is difficult to distinguish between isotherms and plots of log K, drawn for different number of types of site. We cannot clearly decide how many groups of sites contribute to the exchange by examina- tion of the plots if the number of groups of sites is greater than one. 1 0 - 2 0.4 0.6 0.8 A, FIG. 3 . 4 ~ 1 ) Ion-exchange isotherms and (b) plots of loglo Kc against A z for exchangers containing respectively 1,2, 3 and 4 different types of ideal site in a uni-univalent exchange. Curve la, K; = 0.1; lb, K: = 100; 2, KL = 0.1, K i = 100, p i =p2 =0.5; 3, Ki =0.1, Ki = 50, K: = 100, P ~ = P Z = P ~ = ~ ; 4, K i = 0 .1 , Ki=33.33,Kg=66.66, K ~ = l O O , p ~ = ~ ~ = p 3 = p ~ = O . 2 5 . For two-site zeolite models and solid phases treated as non-ideal solid solutions (Kielland coefficients Cl,i # 0 for all site types), some of the plots of log K, against AZ show extrema (fig. 4). Extrema occur in cases when at least one of the Kielland constants is positive. For example, when C1,l GO, C1,2 >O we obtain minima for Ki Ki and maxima for Ki < K2/. Maxima and minima are the more pronounced, the more different the values of Ki and Ki and of C1,l and C1,2.82 ION EXCHANGE INVOLVING HOMOGENEOUS SITES For exchanges other than uni-univalent the value of K, depends on the value of Q (eqn. (6)). Q is proportional to the (ZB-ZA)th power of the concentration in equivalents of the aqueous solution and can serve as a normalizing factor required for comparison of results obtained for different aqueous concentrations.Fig. 4 shows a selection of curves calculated for uni-univalent exchanges and different values of constants Ki and C1,i assuming two kinds of site. The experimental Kielland plots obtained by Barrer, Davies and Rees ti together with their corresponding isotherms can be obtained by choosing the rational thermodynamic constants Kf and the Kielland constants Clni. This is shown for the Kielland plots in fig. 5a-d. Eqn (13) gives smooth curves of K,, but the authors obtained each experimental point at a different concentration of the aqueous solution. In uni-divalent exchanges, where 2, = Zs, the value of Q is different for each of these concentrations and so influences the shape of the Kielland plot.The procedure was as follows. For each value of Q a curve of log K, against AZ was computed, assuming two types of site with thermodynamic equilibrium constants Ki and Ki and Kielland coefficients C1,l and C1,2. From each of the curves so obtained, the point corresponding with the AZ and the aqueous concentration quoted in ref. (6) was marked. A new curve was then drawn through these marked points. For approprite values of Ki, K;, C1,l and C1,2, as given in the caption to fig. 5, both the Kielland plots and the isotherms are similar to the experimental ones. The scatter of points reflects the experimental scatter of the original work.6 -0.3 /3 I- 1 1 I I I I I I 1 ’ 1 0 2 0.4 0.6 0 8 0.2 0.4 0 6 0.8 A, FIG. 4.-Plots of Ioglo Kc against A z for mi-univalent exchanges in a zeolite having two types of ion-exchange site.For (a)-(e) Kielland-type non-ideality is assumed for each type of site (Cl,l and Cl,2 # 0). In each case for curves 1, p 1 = 0.2 (and so p2 = 0.8); for curves 2, p1 = 0.5; andforcurves3,pl-0.8. (a)K’=l,K;=lO,C1,l= -0.3,C1,z=0.3; ( b ) K : = l O , K ’ = l , C1,l = C1,2 = -0.3; (e) Ki = 1, Ki = 10, C1,l = Cl,z = 0.3; (f) K[ = 100, K i = 0.05, CI,1 = -0.3, C1,2 == 0.3; (c)K: = 10, KL = 10, C1,1 = -0.3, C1,2 = 0.3; ( d ) K : = 1, Ki = 10, C1J = c1,2 = 0.R . M. BARRER AND J . KLINOWSKI 83 The above problem of correcting for different aqueous concentrations for the isotherm points does not exist for uni-univalent exchange since Q = 1 for all such concentrations. An isotherm and a Kielland plot resembling those of the NagT1 exchange in chabazite were obtained by taking Ki = 30, C1,l = -0.2, Ki = 150, CIe2 = 0.3 and the relative population of site 1 as 0.4 of the total (fig.54. Baz TI, FIG. 5.--Comparison of experimental points for exchanges in chabazite with calculated plots of loglo Kc = AZ for (a) 2Na++Ca2+, Ki = 0.002, Ki = 2000, C1 = -0.4, C, = 0.4, p 1 = 0.15; (b) 2Na++Sr2+, K: = 0.01, K’ = 2000, C1 = -0.5, C2 = 0.5, p 1 = 0.15; (c) 2Na+-+Ba2+, Ki = 0.01, K; = 2000, Cl = -0.5, C2 = 0 . 5 , ~ ~ = 0.15; (d)Na++Tl+, Ki = 30, Ki = 150, Cz = -0.2, C2 = 0 . 3 , ~ ~ = 0.4. Crosses denote the experimental points,6 circles the theoretical points obtained for the aqueous concentrations quoted by the authors of the paper and assuming the above values of all constants.Two final comments may be made. First, a single zeolite with j groups of homo- geneous sites, each group differing from the others, is identical, in the application of the foregoing theory, with a physical mixture of different exchangers each of which has only one kind of exchange site. Secondly, the assumption is made that the p i do not change with composition of the exchanger. In zeolites, one often finds that there are in toto more crystallographic sites in t h e j groups than the number of cations needed to neutralize the framework charge. In this case, the actual total populations of each group and hence the p i could indeed change with the composition of the crystal.This situation is not easy to quantify and so has not been considered in the present treatment.84 ION EXCHANGE INVOLVING HOMOGENEOUS SITES APPENDIX I For j types of site, all ideal, and ZB = ZA, the condition for extrema (eqn (24)) reduces to Hence This is a quadratic with respect to one equation with respect to all Ai. chosen equivalent cation fraction A,. One may treat eqn (2) as a quadratic One obtains i i i j p r ( l - p ~ ) A ~ - 2 p r A r C p i A i + C p i ( l - P i ) A : - c C 2pipkAiAk = O. (3) i = 1 i = 1 k = 2 i = l i f r i + r k > i k # r i # r In the expression for the roots of this quadratic equation, the quantity under the 1/ sign is A where i 2 r k> i i # r k#r After simplifying the expression in braces, bearing in mind that k # r we obtain f j i Eqn (3) has real roots only when A>O.However, the r.h.s. of (5) is always negative, except when the sum in braces is zero. In the latter case P1P2(A1--A2)2+P1P3(A1--3)2+PZP3(A2-A3)2+. - 0 = 0. (6) As the sum of positive terms cannot be equal to zero, we must have A l = Az = . . . = A,, and therefore Kl = K2 = . . . K i . This is contrary to the assumption of j different groups of ideal sites and therefore no extrema can exist for this case. APPENDIX 2 TWO PHASE EQUILIBRIUM FOR A KIELLAND ISOTHERM Two conditions which must be satisfied for two-phase equilibrium between exchanger rich in ion A and exchanger rich in ion €3 are : (i) since each phase is in equilbrium with the same aqueous solution, the two phases, indicated by point 1 and 2 on fig. 6 4 must be on a line parallel with the abscissa.(ii) The free energy of the two phases at points 1 and 2 must be equal. Thus, GI = Gz, or dG = 0. 1:: For a fixed amount of zeolite and at constant temperature and pressure dG is given by dG = pAdAz+pBUz = ( P A - ~ B ) ~ Z - (1)R. M. BARRER AND J . KLINOWSKI 85 If A is the entering ion, the treatment of Barrer and Falconer,12 which leads to the Kielland relationship, gives for the chemical potentials and ,UB ,UA = RTln N A - R T ~ ~ ~ A ( T ) - E A + ( ~ N A W / ~ ) (2) PB = RTln NB-RThJ’B(T)-EB, where W is an energy term,* ~ A ( T ) and J’B(T) are the partition functions of ions A and B in the crystals, EA and EB are the energies of these ions in the crystals relative to their energies in a convenient reference state and NA, NB and N are respectively the numbers of ions A and B and the total number of available cation sites in the zeolite.By making the substitutions (3) Az = NA/(NA+Nd; Bz = NRI(NA+Nd W(NA + N,)/NRT = - 2.3C, and on integration one obtains G-const. = RT In [R‘AzA$zBgZ] -2.3CRTAz (4) where Fig. 6b shows the value of (G-const.)/RT as a function of AZ for C = 1.5 and two values of R’. The Gibbs free energy of the zeolite goes through two minima and a maximum. By comparing fig. 6a and 6b it is seen that points on the isotherm joined by horizontal lines, corresponding to the same aqueous concentrations, do not generally have the same free energy. We seek the pair of values of Az which satisfy the requirement GI = G2.Condition (1) gives the relation between Az, and Az, for any horizontal line as where Azl, Bz, are the cation fractions at points 1 and Az2, Bz2 are these fractions at points 2. The second condition for two-phase equilibrium (G, = Gz) gives from eqn (4), There are many pairs A z , and Az, which satisfy eqn (6) but only one pair which, for a given value of R’, will satisfy both eqn (6) and (7). This pair determines the two-phases equili- brium line for that value of R’. From the implicit eqn (6), for any given Az, the value of Azz was determined using a computer programme. This pair of values was introduced into eqn (7) together with the values of C and R’, and the operation repeated for a succession of such pairs until the pair was found that satisfied both eqn (6) and (7).The choice of R’ is not arbitrary. If a tangent is drawn to the minimum and to the maximum in the curve of fig. 6a, the values of Az to correspond require R’ to be 30.17 and 34.45 respectively, in order to satisfy eqn (7). Depending on the choice of R’, the value of the Gibbs free energy in the two phases at equi- librium varies. The particular value of R’ which gives the lowest equilibrium line in fig. 66 is 31.663, the equilibrium line then being a tangent to both minima (which are not of equal depth) in the free energy diagram. This lowest line in fig. 6b corresponds in fig. 6a with the line bisecting the isotherm into parts of equal area. If the system takes the con- figuration of minimum free energy, the line 5-6 would be considered as defining the actual two-phase equilibrium.* Win eqn (2) is Now of ref. (12), where NO is Avogadro’s number. Also, the entering ion in ref. (12) was ion B, and in the present paper this ion is taken as A, so that the term W now refers to this ion.86 ION EXCHANGE INVOLVING HOMOGENEOUS SITES At the turning points we must have which, from eqn (5) gives as the condition for the extrema In R’ = 4.6C(&&- In(Az/B&+ la) I I I I I I I . 0.2 0.4 0.6 0!8 A, FIG. 6 . 4 4 Ion-exchange isotherm; (b) plot of (G-const)/RT against A z for K’ = 1, C = 1.5 and two values of R : 31.66 (solid line) and 31.0 (dotted line) in a uni-univalent exchange. The dotted horizontal lines connecting the pairs of points 3-4 and 5-6 correspond respectively in fig. 6a and b, and they, and 1-2 represent the two-phase equilibrium for the given values of R’. When this relation is used to eliminate R’ in eqn (4) one obtains rF)ex = 4.6CAz-2.3CAg+ In BZ We require this free energy to be equal at both extrema, i.e., for both Azl and Az,, so that 1R .M . BARRER A N D J . KLINOWSKI 87 The above equation is fulfilled only when A z l + A z , = 1, which can be shown by putting A z , = l - A z t in the above. There is only one situation where Az1+Az2 = 1, which corresponds with the horizontal line of equal areas (fig. 6a), passing through the point The values of K,, R’ and C in Barrer and Falconer’s treatment are not independent but Az = 0.5, As = l/(K,+ 1). are related by the expression, where the p* are standard chemical potentials of the aqueous ions A and B.Solving eqn (7) for In R’ and introducing the condition Azl + Az, = 1, we obtain for the equilibrium line In R’ = 2.3C, (10) which, introduced in eqn (8), gives Where K, = 1 this leads to Substituting eqn (10) in eqn (8) we obtain a condition for the minima on the free. energy plot, which contains and C only : In (A) -2.3~[(2(~,),,- 11 = 0. 1 - A Z ex Thus, the line giving two equal areas in the isotherm corresponds in the case considered with (a) the curve of (G-const.)/RTagainst Az, having two equal minima ; (b) the lowest values of GI and Gz for the two-phase equilibrium line, this line being the tangent to the minima; (c) standard chemical potentials given by eqn (1 1). J. V. Smith, 2ndZeolite Symp. (Worcester, U.S.A., 1970), p. 401. R. M.Barrer and W. M. Meier, J. Inorg. Nucl. Chem., 1966, 28,629. R. M. Barrer and J. W. Baynham, J. Chem. Soc., 1956,2882. R. M. Barrer and B. M. Munday, J. Chem. SOC. A, 1971,2914. R. M. Barrer, J. A. Davies and L. V. C. Rees, J. Inorg. Nucl. Chem., 1969,31,219. ’ R. M. Barrer, L. V. C. Rees and D. J. Ward, Proc. Roy. SOC. A, 1963,237,180. * R. M. Barrer, J. A. Davies and L. V. C. Rees, J. Innorg. Nucl. Chem., 1968,30, 3333. lo R. M. Barrer, J. A. Davies and L. V. C. Rees, J . Inorg. Nucl. Chem., 1969,31,2599. l1 J. Kielland, J. Sac. Chem. Ind., l935,54,232T. l2 R. M. Barrer and J. D. Falconer, Proc. Roy. SOC. A, 1956,236,227. l3 G. L. Gaines and H. C. Thomas, J. Chem. Phys., 1953,21,714. * R. M. Barrer and H. Villiger, 2. Krist., 1969, 128, 352. R. M. Barrer, L. V.C. Rees and M. Shamsuzzoha, J. Inorg. Nucl. Chem., 1966,28,629. Ion Exchange Involving Several Groups of Homogeneous Sites BY R. M. BARRER AND J. KLINOWSKI Physical Chemistry Laboratories, Chemistry Dept., Imperial College, London S. W.7 Received 22nd June, 197 1 Crystalline exchangers such as zeolites may contain more than one group of different exchange sites. The theoretical basis of these multi-site exchangers has therefore been examined. Overall exchange isotherms and selectivities have been expressed in terms of the properties o f j different kinds of co-existent site. The resulting plots of log K;1 against equivalent cation fraction of the entering ion (Kielland plots) are complex, showing maxima, minima, inflection points and sigmoid shapes according to the proportions and properties of the several groups of sites.It is concluded that linear Kielland plots in such circumstances can at best only be approximations. Even where each group of sites behaves ideally the overall exchange isotherm can be strongly non-ideal. Where Kielland coefficients C1 are not zero for the overall isotherm, this isotherm may show maxima and minima (and hence limited solid solubility of the end members of the exchange), the criterion for which is a positive value of C1 which varies according to the valence of the exchanging ions. By an appropriate choice of quantities characterizing each of two groups of sites Kielland plots and isotherms were found, as an example, for chabazite which closely agreed with those determined experimentally.It has been established by X-ray crystallographic studies that in zeolite ion exchangers the exchangeable ions may be present in each of several different crystallo- graphic environments. ' 9 This can result in distinct groups of sites, each group being homogeneous but different from the others. There has been no full examination of the effects of this upon the properties of the overall exchange isotherms and selectivities, although several limited approaches to the problem have been made. Barrer and Meier successfully analyzed the sigmoid Ag+Na exchange isotherm in zeolite A in terms of two kinds of ion site, 12 of one kind to every one of the other. Successful analyses of a number of isotherms in zeolite K-F have been made by Barrer and Munday.' In this case the ratio of the numbers of sites of each kind was about 2 : 1.When the logarithm of the corrected selectivity quotient, K, (expressed as in eqn (3)) is plotted against the equivalent cation fraction of one of the competing ions, curves of considerable complexity are often found, including some with clear maxima (e.g., NaST1 and all uni-divalent exchanges investigated in chabazite ; and uni- divalent exchanges in zeolites A,7 Y and X 9). It is possible, though unlikely, that the maxima arise from side reactions such as hydrolysis of the zeolite phase. Whether or not hydrolysis can be a factor it is interesting to find whether and under what conditions both maxima and minima can arise as a result purely of the presence in a zeolite of crystallographically different kinds of exchange site.For the above reasons, an examination was made of the theoretical basis of multi-site ion exchange. THE OVERALL EXCHANGE EQUILIBRIUM Exchange equilibrium must first be formulated by considering all kinds of exchange sites together. The exchanger is non-ideal in the sense that the activity coefficients of the ions in the exchanger are not unity and depend upon the cation composition of the exchanger. The general exchange reaction is74 ION EXCHANGE INVOLVING HOMOGENEOUS SITES where ions of species A have a charge ZA+ and of B having a charge ZB t are involved. The subscripts Z and S refer to zeolite and solutions, respectively. The rational thermodynamic equilibrium constant K, is then where A , and Bz are equivalent cation fractions of A and B in the zeolite and m t , mA, itre these concentrations in the solution in g ions per kg of solvent.fk, fg, yA and yB are the corresponding activity coefficients. We now define a quotient K,, related to ion selectivity, by K , = Ap(mF)zAyZA ---.!.- - AZ,"(aF)ZA BgA( mc)zByp - BgA(a$)ZB' (3) It is then possible to express fA andf, in terms of K,, which can be measured. The general relation between thefand K, are given later (eqn (1 6)). If we replace molalities in solution by equivalent fractions As and B,, then m,A = AS -(Z,m,A+Z,m~) ZA BS ZB m: = -(ZAme+ZBm,B), and so Accordingly, where (4) EXCHANGE EQUILIBRIUM INVOLVING SEVERAL HOMOGENEOUS GROUPS OF SITES The exchanger containsj types of site, the sites within each group being equivalent, but differing from sites in another group.The overall exchange isotherm as given in the previous section is to be related to the isotherms on the individual sites.Io Let quantities relating to the ith type of site bear the subscript i. For exchange equilibrium on these sites, where Ki is the thermodynamic equilibrium constant. From eqn (2) and (7) the condition for equilibrium is K,B;A EA? where E = ( f %>zB/(f3zA; (8) K ~ B ~ K,B;A E,A;B E ~ A ~ Z B ' -- ... - - --- ... - - Ei = (f f E j = (f f (9)R. M. BARRER AND J . KLINOWSKI 75 In terms of the quotient K, defined by eqn (3), the condition of eqn (8) becomes (10) -- (Kc) jBjzA ... - -- - ... --- - (Kc)iBZA - K , B , Z ~ A? AZ” A? ’ THE KIELLAND SELECTIVITY QUOTIENT The quotient K, of eqn (3), which is that often measured experimentally, will be considered further.The relationships between AZ and the A i and between Bz and the Bi are : is the ratio of total equivalents of cations in sites of type i to total equivalents in the zeolite. rnf and rn? are molalities of ions A and B in sites of type i (mol kg-l of zeolite). One may substitute eqn (1 1) in eqn (3), use eqn (7) to obtain Al in terms of Bi and so re-write eqn (3) in the form j j K , = [C piB~AizB(Kc)f’ZB]ZB/[~ piBilzA 1 1 or, in terms of Al, (13) Eqn (13) forms the basis for the subsequent calculations of the shapes of the Kielland plots for the overall exchange (i.e., plots of log K, against AZ or Bz), assuming different physically reasonable values of the pi and the Ki together with simple concentration dependences of (fr)zA/CfA)ZB which accord with previous experience l 1 and with a theoretical treatment.The expressions (13) and (14) for K, are ratios of two polynomials. The result of this division is another polynomial with an infinite number of terms in Ai or Bi. Since the Al and Bi are functions of AZ and Bz (eqn (8)) this means that even when all sites are ideal (El = E2 = . . . = Ej = 1) the plot of log K, against AZ or Bz cannot in general be a horizontal or even a straight line. For a zeolite composed of one type of ideal site, K, = Kl = K,. This situation is seldom encountered, but was closely approached for some exchanges in sodalite hydrate. l2 This zeolite provides only one kind of cation position. In the most general case, the value of K, for the overall reaction will depend on the properties of the term E = ( f , A > Z B / ( f g ) z ~ .Gaines and Thomas l3 obtained for this ratio the expression log E = 0.4343 (ZB-ZJ- log Kc+ log K , dAZ. (15) s: This equation omits a term involving the change in activity of zeolitic water during exchange since in the analogous case of the clay minerals this term was estimated to76 ION EXCHANGE INVOLVING HOMOGENEOUS SITES L 1 = exp {(&-2,)-2.3 5 .c.> ' L, = exp (ZB-ZA)+2.3 C,}. be small. To evaluate E, log K, must be known as a function of AZ. Whatever this dependence is, it can be represented as a power series of the type log K , = Co+2C1Az+3C2Ag+ . . . +(n+l)C,A;, (16) where the C are constant coefficients. Substituting eqn (16) in eqn (15) gives E = exp ((ZB-ZA)+2.3[cl(1-2AZ)$C,(1-3A%)f .+cn(l-(n+l)A",]]. (17) Further, since K,E = K,, then while, from Gaines and Thomas's relations for single ion activity coefficients, log K, = 0.4343(2, -2J + Co + C1+ . . . + C,, (18) log (f,")zB = 0.4343(2B-ZA)BZ-B,(2C,AZ+ . . +(n+ l)CnA;)+ 1 (19) Cl(l-A;)+ . . . +C,(l-A;+'), 1% (.f,")"* = - o.4343(zB - ZA)Az + Az(2C1A, + . . . +(n + l)C,A;) - n AZ,Ai+1 0 [$ ZA 0 IZA -2.3 lim K , = exp 2.3 1 (n + l)C, = pi exp - 1 (n + 1)& lim K , = exp 2.3c0 = pi exp -Co,i , ; (21) MAXIMA A N D MINIMA I N PLOTS OF log K, AGAINST AZ OR Bz From the relationship K, = K,E-l the condition for extrema in plots of K, or log K, against AZ is dK,/dAZ = Ka d(E-')/dAZ = 0.R. M. BARRER AND J . KLINOWSKI 77 But E according to eqn (17) is an exponential function of AZ, i.e., E = exp q5(AZ), and so the above condition becomes d4/dAZ = 0.(22) By considering increasing numbers of coefficients C in the exponential of eqn (17) the conditions associated with extrema can be found. These can first appear for C1, C, # 0 but C3, ..., C, = 0, since d#/dAZ = -2.3 (2C1 +6C2Az) and is zero when AZ = -C,/3C2. Because O<A,< 1 the ratio -Cl/3C, must also lie within these limits and so CI and C, must have opposite signs. Similarly, when C1, C2, C3 # 0 but C,, . . . , C,, = 0 the conditions for extrema are with Ci <:CiC3. Plots of log K, against AZ are shown in fig. 1 for typical choices of the coefficients C. A, FIG. 1.-Plots of log,, KC against A, for various values of three coefficients C,. The thermodynamic equilibrium constant Ka = 10.(a) (1) Cz = C2 = C3 = 0; (2) C, = 0.3, c, = c3 = 0; (3) c1 = -0.3, Cz = Cs = 0. -0.3, Cz = C, = 0.3. (b) (1) C1 = -0.3, C2 = 0.3, Cs = 0; (2) C1 = Cz = 0.3, C3 == 0 ; (3) C1 = (c) (1) C1 = Cz = 0.3, C3 = -0.3; (2) C1 = 0.3, Cz = -0.3, C3 = 0. If we consider the zeolite specifically in terms of its j homogeneous groups of sites, the condition for extrema can be found from eqn (13) and (10). After some manipulation, the condition is = A,, $ piAidlnBz ZA dlnB, d In Bi ( +A d In (KJi where By substituting in eqn (23) and bearing in mind that In = -2.3 Bi n(n+ l)Cn,iA!n-l), d In Bi 178 ION EXCHANGE INVOLVING HOMOGENEOUS SITES j C PiAi . we get the condition for extrema as I A; 1 ZAAZ +ZBBZ (24) - - 2.3B. Ai 1 - 2 n(n + l)Cn,iAy- ( z A 1 ZAAi+ZBBi-2.3BiAi C n(n+ l)Cn,iA:-' F o r j groups of ideal sites all Cn,i are zero and eqn (24) simplifies to give the condition for extrema as It can be formally demonstrated (appendix 1) that this equation leads, for 2, = ZB, to the condition A l = A2 = .. . A j , which is never fulfilled since ideal isotherms can never cross. Hence, no extrema are possible for the j groups of ideal sites, however large j is and however different the Ki. EXTREMA AND CROSSING POINTS I N ISOTHERMS Eqn (5) for the ion-exchange isotherm can be re-arranged to F(AZ, As) = KL(1- AZ)'*A2 - A 2 ( 1 - As)ZAEI' = 0. (25) The condition for extrema in the isotherm contour and hence for immiscibility gaps (appendix 2) in solid solutions of the end members of the exchange (pure A- and B-forms of the zeolite, respectively) is -- dAs - --(W~AZ)A, = 0.dAZ (aFfaAS)A, Since (aF/aA& is finite this condition requires that (aF/dA,),, = 0. eqn (25) and (17), one obtains after dfferentiation of each and some rearrangement Then, from n ZAA,+Z,(1-Az)-2.3(1-AZ)~ n(n+l)CnAz = 0. (26) 1 This relation cannot be satisfied for the ideal case where all coefficients C1 to C, are zero, so that here, as expected, no extrema are possible. The next simplest case is that where C, # 0 but C, = . . . = C, = 0 (linear Kielland plots). Then for ZA # Z B , with 18.4zBc1/[4.6cl +(ZB-z2,)]' < 1 (28) for real roots. The maximum corresponds with the smaller value of AZ, the minimum with the larger. When 2, # ZB, the maximum and minimum are equidistant from AZ = 3- (ZB+ZJ/9.2C,, and for 2, = 2, they are equidistant from AZ = 4 in both cases by distances dAZ = (1 /9.2C1) ,/[4.6C1 + (ZB-zA)]" - 18.4ZBC1.The necessary in- equality (28) can be rearranged to the form c1 > (ZA+ZBf J42~&)/4.6. The inequality associated with the negative sign appears to arise from the quadratic procedure involving taking a square root, since in the case ZA = 2, quadratic pro-R. M . BARRER AND J . KLINOWSKI 79 cedure can be avoided and only the inequality associated with the positive sign then appears. The requirements for C1 are given for the positive sign in table 1 for several values of ZA and 2,. TABLE 1.-vALUES WHICH c1 MUST EXCEED FOR EXTREMA TO APPEAR ZA ZB c1 1 1 > 0.87 1 2 >(3+ 2/8)/4.6 = 1.27 2 1 1 3 >(4+ 2/12)/4.6 = 1.62 3 1 2 2 > 1.74 2 3 >(5+ 1/24)/4.6 = 2.16 3 2 3 3 >2.61 Fig.2 shows isotherm contours calculated from eqn (5) with r = 1 for different ZA and 2, and for several values of C1. In these diagrams, all isotherms for which 2, and 2, and the Ki are the same pass through a common point. The requirement for this to happen, whether 2, and 2, are equal or not, is that at the crossing point A, FIG. 2.-Exchange isotherms for ions of valence ZA with ions of valence ZB with C1 # 0; Cz, Ct, . . . , c n = 0 ; and Ka = 1. (a) ZB = ZA = 1 ; (b) ZB = 1, ZA = 2 ; (C) ZB = 1, ZA = 3 ; (d) ZB = 2, ZA = 3. For curves 1, C1 = - 1 ; 2, C1 = 0; 3, C1 = 1 ; 4, C1 = 2; 5 , C1 = 3. Isotherms showing maxima and minima have ranges of composition A z over which there is metastability.Here there is limited miscibility of the end members of the exchange.80 ION EXCHANGE INVOLVING HOMOGENEOUS SITES E must be independent of all the coefficients C involved in eqn (17) and equal to exp (ZB-ZA). No general condition can be laid down for such behaviour but examples in particular cases are given in table 2. The value of As at the crossing point is to be found from that for AZ and the relation For example, when ZB = 2, = 1 ; C1 # 0 ; Cz to C, = 0, As = 1/(1 +Ki/r), since AZ = 3 (table 2). Some consequences of immiscibility gaps on the isotherms are discussed in appendix 2. TABLE 2.-REQUIREMENTS FOR E TO BECOME INDEPENDENT OF COEFFICIENTS c conditions for C value of A Z for E = exp (ZB-ZA) subsidiary conditions - c1 # 0 ; c2 to c, = 0 C, # 0; C1 and C3 to C, = 0 c3 # O ; C1, C2 and C4 to C, = 0 c1 # 0, c2 # 0; c3 to c, = 0 1 12 1 / 4 3 1 /w - C1 + 4 C + 3 C2(Cl + C2) 3cz - - (i) Cl /C2 > - 1 (ii) the values of C1 and C, can vary but the ratio Cl/C2 must be constant.DISCUSSION Eqn (13) and (14) relate the overall corrected selectivity coefficient of the zeolite to the properties of individual sites. If we specify the thermodynamic equilibrium constants K/ for each type of site and the coefficients Cn,i in the polynomial #(A,) describing the non-ideality of each type of site, then the values of Ai are roots of the set of j implicit equations (where j = number of kinds of site) : Kf(1- Ai)'*(AS)'" -(Ai>'"(1- A,)ZAT exp {+(Ai)) = 0. It was assumed throughout that I' = 1. This assumption is reasonable for solutions of low ionic strength.The solution of the above equation very much depends on the coefficients Cn,r in #(Ad), i.e., on the type of polynomial used. Many calculations were made for C1, = 0 and C2, i , . . . ? C,, = 0 for each type of site, which corresponds with the linear dependence of the log (K& on the A t . It does not follow, however, that the overall exchange for the zeolite can then be treated with C1 # 0 and C, . . . , c, = 0. The numerical calculations were made for uni-univalent and mi-divalent exchanges and different numbers of types of site. All calculations were made on the Imperial College IBM 6600 computer using a Fortran IV programme SITE written for this purpose. The programme solves numerically the implicit eqn (29) with any required degree of accuracy.In the present work, all values of A , were obtained with an error of less than K, was obtained from eqn (13). Fig. 3 shows the ion-exchange isotherms and the plots of log K, against AZ for one, two, three and four types of site, each type behaving ideally (all coefficients Cn,i zero except Co,r). The Kielland plots for the overall exchange isotherm are sigmoid in shape and do not show extrema. The larger the difference of the effective rational thermodynamic equilibrium constants K/ for individual sites, the moreR. M. BARRER AND J . KLINOWSKI 81 sigmoid the plots. Conversely, for close values of K/ the plots of log K, against AZ approach a straight line. The plot of log K, f o r j = 1 is horizontal but for j = 2,3, 4 . . . these plots are curves, despite the ideality of all groups of sites.Thus, in the multi-site theory of ion exchange in zeolites the behaviour of the zeolite as a whole can never be considered ideal. This conclusion was anticipated from the algebraic discussion of eqn (13) and eqn (14). Furthermore, it is difficult to distinguish between isotherms and plots of log K, drawn for different number of types of site. We cannot clearly decide how many groups of sites contribute to the exchange by examina- tion of the plots if the number of groups of sites is greater than one. 1 0 - 2 0.4 0.6 0.8 A, FIG. 3 . 4 ~ 1 ) Ion-exchange isotherms and (b) plots of loglo Kc against A z for exchangers containing respectively 1,2, 3 and 4 different types of ideal site in a uni-univalent exchange.Curve la, K; = 0.1; lb, K: = 100; 2, KL = 0.1, K i = 100, p i =p2 =0.5; 3, Ki =0.1, Ki = 50, K: = 100, P ~ = P Z = P ~ = ~ ; 4, K i = 0 . 1 , Ki=33.33,Kg=66.66, K ~ = l O O , p ~ = ~ ~ = p 3 = p ~ = O . 2 5 . For two-site zeolite models and solid phases treated as non-ideal solid solutions (Kielland coefficients Cl,i # 0 for all site types), some of the plots of log K, against AZ show extrema (fig. 4). Extrema occur in cases when at least one of the Kielland constants is positive. For example, when C1,l GO, C1,2 >O we obtain minima for Ki Ki and maxima for Ki < K2/. Maxima and minima are the more pronounced, the more different the values of Ki and Ki and of C1,l and C1,2.82 ION EXCHANGE INVOLVING HOMOGENEOUS SITES For exchanges other than uni-univalent the value of K, depends on the value of Q (eqn.(6)). Q is proportional to the (ZB-ZA)th power of the concentration in equivalents of the aqueous solution and can serve as a normalizing factor required for comparison of results obtained for different aqueous concentrations. Fig. 4 shows a selection of curves calculated for uni-univalent exchanges and different values of constants Ki and C1,i assuming two kinds of site. The experimental Kielland plots obtained by Barrer, Davies and Rees ti together with their corresponding isotherms can be obtained by choosing the rational thermodynamic constants Kf and the Kielland constants Clni. This is shown for the Kielland plots in fig. 5a-d. Eqn (13) gives smooth curves of K,, but the authors obtained each experimental point at a different concentration of the aqueous solution.In uni-divalent exchanges, where 2, = Zs, the value of Q is different for each of these concentrations and so influences the shape of the Kielland plot. The procedure was as follows. For each value of Q a curve of log K, against AZ was computed, assuming two types of site with thermodynamic equilibrium constants Ki and Ki and Kielland coefficients C1,l and C1,2. From each of the curves so obtained, the point corresponding with the AZ and the aqueous concentration quoted in ref. (6) was marked. A new curve was then drawn through these marked points. For approprite values of Ki, K;, C1,l and C1,2, as given in the caption to fig. 5, both the Kielland plots and the isotherms are similar to the experimental ones.The scatter of points reflects the experimental scatter of the original work.6 -0.3 /3 I- 1 1 I I I I I I 1 ’ 1 0 2 0.4 0.6 0 8 0.2 0.4 0 6 0.8 A, FIG. 4.-Plots of Ioglo Kc against A z for mi-univalent exchanges in a zeolite having two types of ion-exchange site. For (a)-(e) Kielland-type non-ideality is assumed for each type of site (Cl,l and Cl,2 # 0). In each case for curves 1, p 1 = 0.2 (and so p2 = 0.8); for curves 2, p1 = 0.5; andforcurves3,pl-0.8. (a)K’=l,K;=lO,C1,l= -0.3,C1,z=0.3; ( b ) K : = l O , K ’ = l , C1,l = C1,2 = -0.3; (e) Ki = 1, Ki = 10, C1,l = Cl,z = 0.3; (f) K[ = 100, K i = 0.05, CI,1 = -0.3, C1,2 == 0.3; (c)K: = 10, KL = 10, C1,1 = -0.3, C1,2 = 0.3; ( d ) K : = 1, Ki = 10, C1J = c1,2 = 0.R . M. BARRER AND J . KLINOWSKI 83 The above problem of correcting for different aqueous concentrations for the isotherm points does not exist for uni-univalent exchange since Q = 1 for all such concentrations.An isotherm and a Kielland plot resembling those of the NagT1 exchange in chabazite were obtained by taking Ki = 30, C1,l = -0.2, Ki = 150, CIe2 = 0.3 and the relative population of site 1 as 0.4 of the total (fig. 54. Baz TI, FIG. 5.--Comparison of experimental points for exchanges in chabazite with calculated plots of loglo Kc = AZ for (a) 2Na++Ca2+, Ki = 0.002, Ki = 2000, C1 = -0.4, C, = 0.4, p 1 = 0.15; (b) 2Na++Sr2+, K: = 0.01, K’ = 2000, C1 = -0.5, C2 = 0.5, p 1 = 0.15; (c) 2Na+-+Ba2+, Ki = 0.01, K; = 2000, Cl = -0.5, C2 = 0 . 5 , ~ ~ = 0.15; (d)Na++Tl+, Ki = 30, Ki = 150, Cz = -0.2, C2 = 0 .3 , ~ ~ = 0.4. Crosses denote the experimental points,6 circles the theoretical points obtained for the aqueous concentrations quoted by the authors of the paper and assuming the above values of all constants. Two final comments may be made. First, a single zeolite with j groups of homo- geneous sites, each group differing from the others, is identical, in the application of the foregoing theory, with a physical mixture of different exchangers each of which has only one kind of exchange site. Secondly, the assumption is made that the p i do not change with composition of the exchanger. In zeolites, one often finds that there are in toto more crystallographic sites in t h e j groups than the number of cations needed to neutralize the framework charge. In this case, the actual total populations of each group and hence the p i could indeed change with the composition of the crystal.This situation is not easy to quantify and so has not been considered in the present treatment.84 ION EXCHANGE INVOLVING HOMOGENEOUS SITES APPENDIX I For j types of site, all ideal, and ZB = ZA, the condition for extrema (eqn (24)) reduces to Hence This is a quadratic with respect to one equation with respect to all Ai. chosen equivalent cation fraction A,. One may treat eqn (2) as a quadratic One obtains i i i j p r ( l - p ~ ) A ~ - 2 p r A r C p i A i + C p i ( l - P i ) A : - c C 2pipkAiAk = O. (3) i = 1 i = 1 k = 2 i = l i f r i + r k > i k # r i # r In the expression for the roots of this quadratic equation, the quantity under the 1/ sign is A where i 2 r k> i i # r k#r After simplifying the expression in braces, bearing in mind that k # r we obtain f j i Eqn (3) has real roots only when A>O.However, the r.h.s. of (5) is always negative, except when the sum in braces is zero. In the latter case P1P2(A1--A2)2+P1P3(A1--3)2+PZP3(A2-A3)2+. - 0 = 0. (6) As the sum of positive terms cannot be equal to zero, we must have A l = Az = . . . = A,, and therefore Kl = K2 = . . . K i . This is contrary to the assumption of j different groups of ideal sites and therefore no extrema can exist for this case. APPENDIX 2 TWO PHASE EQUILIBRIUM FOR A KIELLAND ISOTHERM Two conditions which must be satisfied for two-phase equilibrium between exchanger rich in ion A and exchanger rich in ion €3 are : (i) since each phase is in equilbrium with the same aqueous solution, the two phases, indicated by point 1 and 2 on fig.6 4 must be on a line parallel with the abscissa. (ii) The free energy of the two phases at points 1 and 2 must be equal. Thus, GI = Gz, or dG = 0. 1:: For a fixed amount of zeolite and at constant temperature and pressure dG is given by dG = pAdAz+pBUz = ( P A - ~ B ) ~ Z - (1)R. M. BARRER AND J . KLINOWSKI 85 If A is the entering ion, the treatment of Barrer and Falconer,12 which leads to the Kielland relationship, gives for the chemical potentials and ,UB ,UA = RTln N A - R T ~ ~ ~ A ( T ) - E A + ( ~ N A W / ~ ) (2) PB = RTln NB-RThJ’B(T)-EB, where W is an energy term,* ~ A ( T ) and J’B(T) are the partition functions of ions A and B in the crystals, EA and EB are the energies of these ions in the crystals relative to their energies in a convenient reference state and NA, NB and N are respectively the numbers of ions A and B and the total number of available cation sites in the zeolite.By making the substitutions (3) Az = NA/(NA+Nd; Bz = NRI(NA+Nd W(NA + N,)/NRT = - 2.3C, and on integration one obtains G-const. = RT In [R‘AzA$zBgZ] -2.3CRTAz (4) where Fig. 6b shows the value of (G-const.)/RT as a function of AZ for C = 1.5 and two values of R’. The Gibbs free energy of the zeolite goes through two minima and a maximum. By comparing fig. 6a and 6b it is seen that points on the isotherm joined by horizontal lines, corresponding to the same aqueous concentrations, do not generally have the same free energy.We seek the pair of values of Az which satisfy the requirement GI = G2. Condition (1) gives the relation between Az, and Az, for any horizontal line as where Azl, Bz, are the cation fractions at points 1 and Az2, Bz2 are these fractions at points 2. The second condition for two-phase equilibrium (G, = Gz) gives from eqn (4), There are many pairs A z , and Az, which satisfy eqn (6) but only one pair which, for a given value of R’, will satisfy both eqn (6) and (7). This pair determines the two-phases equili- brium line for that value of R’. From the implicit eqn (6), for any given Az, the value of Azz was determined using a computer programme. This pair of values was introduced into eqn (7) together with the values of C and R’, and the operation repeated for a succession of such pairs until the pair was found that satisfied both eqn (6) and (7).The choice of R’ is not arbitrary. If a tangent is drawn to the minimum and to the maximum in the curve of fig. 6a, the values of Az to correspond require R’ to be 30.17 and 34.45 respectively, in order to satisfy eqn (7). Depending on the choice of R’, the value of the Gibbs free energy in the two phases at equi- librium varies. The particular value of R’ which gives the lowest equilibrium line in fig. 66 is 31.663, the equilibrium line then being a tangent to both minima (which are not of equal depth) in the free energy diagram. This lowest line in fig. 6b corresponds in fig. 6a with the line bisecting the isotherm into parts of equal area. If the system takes the con- figuration of minimum free energy, the line 5-6 would be considered as defining the actual two-phase equilibrium.* Win eqn (2) is Now of ref. (12), where NO is Avogadro’s number. Also, the entering ion in ref. (12) was ion B, and in the present paper this ion is taken as A, so that the term W now refers to this ion.86 ION EXCHANGE INVOLVING HOMOGENEOUS SITES At the turning points we must have which, from eqn (5) gives as the condition for the extrema In R’ = 4.6C(&&- In(Az/B&+ la) I I I I I I I . 0.2 0.4 0.6 0!8 A, FIG. 6 . 4 4 Ion-exchange isotherm; (b) plot of (G-const)/RT against A z for K’ = 1, C = 1.5 and two values of R : 31.66 (solid line) and 31.0 (dotted line) in a uni-univalent exchange. The dotted horizontal lines connecting the pairs of points 3-4 and 5-6 correspond respectively in fig. 6a and b, and they, and 1-2 represent the two-phase equilibrium for the given values of R’. When this relation is used to eliminate R’ in eqn (4) one obtains rF)ex = 4.6CAz-2.3CAg+ In BZ We require this free energy to be equal at both extrema, i.e., for both Azl and Az,, so that 1R . M . BARRER A N D J . KLINOWSKI 87 The above equation is fulfilled only when A z l + A z , = 1, which can be shown by putting A z , = l - A z t in the above. There is only one situation where Az1+Az2 = 1, which corresponds with the horizontal line of equal areas (fig. 6a), passing through the point The values of K,, R’ and C in Barrer and Falconer’s treatment are not independent but Az = 0.5, As = l/(K,+ 1). are related by the expression, where the p* are standard chemical potentials of the aqueous ions A and B. Solving eqn (7) for In R’ and introducing the condition Azl + Az, = 1, we obtain for the equilibrium line In R’ = 2.3C, (10) which, introduced in eqn (8), gives Where K, = 1 this leads to Substituting eqn (10) in eqn (8) we obtain a condition for the minima on the free. energy plot, which contains and C only : In (A) -2.3~[(2(~,),,- 11 = 0. 1 - A Z ex Thus, the line giving two equal areas in the isotherm corresponds in the case considered with (a) the curve of (G-const.)/RTagainst Az, having two equal minima ; (b) the lowest values of GI and Gz for the two-phase equilibrium line, this line being the tangent to the minima; (c) standard chemical potentials given by eqn (1 1). J. V. Smith, 2ndZeolite Symp. (Worcester, U.S.A., 1970), p. 401. R. M. Barrer and W. M. Meier, J. Inorg. Nucl. Chem., 1966, 28,629. R. M. Barrer and J. W. Baynham, J. Chem. Soc., 1956,2882. R. M. Barrer and B. M. Munday, J. Chem. SOC. A, 1971,2914. R. M. Barrer, J. A. Davies and L. V. C. Rees, J. Inorg. Nucl. Chem., 1969,31,219. ’ R. M. Barrer, L. V. C. Rees and D. J. Ward, Proc. Roy. SOC. A, 1963,237,180. * R. M. Barrer, J. A. Davies and L. V. C. Rees, J. Innorg. Nucl. Chem., 1968,30, 3333. lo R. M. Barrer, J. A. Davies and L. V. C. Rees, J . Inorg. Nucl. Chem., 1969,31,2599. l1 J. Kielland, J. Sac. Chem. Ind., l935,54,232T. l2 R. M. Barrer and J. D. Falconer, Proc. Roy. SOC. A, 1956,236,227. l3 G. L. Gaines and H. C. Thomas, J. Chem. Phys., 1953,21,714. * R. M. Barrer and H. Villiger, 2. Krist., 1969, 128, 352. R. M. Barrer, L. V. C. Rees and M. Shamsuzzoha, J. Inorg. Nucl. Chem., 1966,28,629.

ISSN:0300-9599    

DOI:10.1039/F19726800073

出版商:RSC    

年代:1972

数据来源: RSC