|
1. |
Proceedings of the Chemical Society. June 1964 |
|
Proceedings of the Chemical Society ,
Volume 1,
Issue June,
1964,
Page 161-200
Preview
|
PDF (4250KB)
|
|
摘要:
PROCEEDINGS OF THE CHEMICAL SOCIETY JUNE 1964 THE MEASUREMENT OF THERMODYNAMIC PROPERTIES AT THE NATIONAL CHEMICAL LABORATORY By E. F. G. HERINGTON (HEADOF THE CHEMICAL DIVISION) PHYSJCS MAJORcontributions to thermodynamic theory were made in the eighteenth and nineteenth centuries but it was not until the present century that thermodynamic methods were widely ap- plied to chemistry. Willard Gibbs (1839-1903) developed the theory systematically but with the exception of Haber’s work (c. 1905) on the ammonia synthesis little use of thermodynamic calculations was made by chemists until Lewis and Randall published (1923) their book “Thermodynamics and the Free Energy of Chemical Substances.” As a result of this publication chemists particularly in America began to appreciate the practical value of the subject but it was quickly realised that useful results could be obtained only if an adequate supply of basic data was available.Collections of existing data were therefore made; examples of such compilations are the work of Kharasch (1929) summarising combustion data the tables due to Kelley who started to review inorganic thermodynamics in the early 1930s and Bichow- sky and Rossini’s (1936) book on thermo-chemistry. However it soon became clear that thermal measurements would have to be made with the utmost accuracy before chemists could use the full power of thermodynamic methods. An appreciation of this fact by the American Petroleum Institute led to the setting up of Project 44 for the measurement of thermo-dynamic properties of hydrocarbons.Following the success of this large undertaking the work was extended to other classes of substances; sulphur compounds for example are now being investigated under Project 48. In 1946 a modest programme similar to the American project was begun at Teddington for the study of compounds derived from coal tar. This programme has ex- panded steadily and work on other compounds has been introduced; the Chemical Thermo- dynamics Group of the National Chemical Laboratory is now well equipped to supply thermodynamic data to British industry. The present article describes the objectives and achievements of this Group. It may be remarked that there is a concensus of opinion that work of this kind is more suitable for a national labora- tory than for a university because the project requires a continuity of effort on an integrated programme unsuited to the needs of students studying for higher degrees.Some Uses of Thermodynamic Data in Chem-161 istry.-Modern chemical technology is dependent upon accurate thermodynamic data; thus re-action heats latent heats heat capacities saturated-vapour pressures and critical tempera- tures and volumes are used by chemical engineers designing large-scale equipment. The disastrous consequences of using poor data can be illus- trated by a recent report from the U.S.A. that two 38-million-dollar chemical plants for the production of boron compounds had to be shut down prematurely mainly because erroneous and incomplete thermodynamic data had been used for their design.Some idea of the feasibility of a proposed industrial synthesis can be gained without resort to direct observations if the free-energy change AGO of the reaction can be calculated. In rough terms it may be said that if the calculated free- energy change is greater than say 10kcal. mole-l the reaction is unlikely to be worth experimental study. If dG" lies between +10 and -10 kcal. mole-l the reaction might be feasible but if dGo is found to be negative and numerically greater than 10 kcal. mole-l then from a thermo-dynamic point of view the reaction is possible and an attempt to find the correct conditions for the reaction to proceed at a favourable rate is worthwhile.The value of AGO at a given temperature can be obtained from values of the free energies of formation of all the reactants and products at this temperature. The free energies of formation can in turn be calculated from measured heats of combustion and heat capaci- ties. A high precision in these experimentally measured quantities is essential in order to avoid large errors in dGoarising from the accumulation of experimental errors. This is particularly im- portant if an accurate value for an equilibrium constant K is to be calculated from AGO by means of the equation -AGO = RTlnK; for example an error of only 0-06 kcal. mole-1 in dG" corresponds to a 10% error in the equi- librium constant at 298.16"~.The best modern measurements of heats of combustion are accu- rate to k0.02 % and entropies (at 298.16"~)can be determined within *0-2% in favourable instances so that standard free energies of forma- tion can be found to about 0.2 kcal.mole-l. Accuracy of measurement is of course of no avail unless highly purified samples are used. An alternative route to values of the thermo- dynamic functions for a substance in the ideal- PROCEEDINGS gas state is sometimes offered by the methods of statistical thermodynamics. By these methods moreover the heat capacity and entropy of an ideal vapour can be computed from spectroscopic observations and a comparison of values so obtained with those derived from heat measure- ments will often provide additional information about the structure of the molecule.For example agreement between the values obtained by the two methods may confirm a suggested molecular structure or a disparity may reveal the magni- tude of an energy barrier to restricted internal rotation. Information on chemical linkages can be obtained from bond-energy values deduced from heats of combustion and these bond-energy values can often be used to estimate the heats of formation of compounds on which heat measure- ments have not been made. Direct observation of suitably chosen gaseous equilibria provide a searching test of data obtained by thermal measurements and can also supply information of direct industrial importance. Other thermodynamic properties of interest to analysts are freezing points and triple points which can be used to establish the purity of samples and electrolytic dissociation constants which are used in analytical and separation processes.Choice of Substancesfor Study.-Much thermo-dynamic information can be obtained on organic compounds with economy of time and effort if a programme is devised to make a wide range of measurements on highly purified samples of carefully selected compounds. For example reliable thermodynamic and physico- chemical data for a series of organic homologues can often be obtained by interpolation and extra- polation of the measured properties of shrewdly selected members of the series. Members of the pyridine series were first chosen for study at Teddington because they make a useful contribution to the profit of tar distillation even though they are not produced in large tonnage.Many of the simple properties for example boiling points and densities were not accurately known before the work was started. Indeed commercial specifications for "pure" pyridine at that time would have ex-cluded chemically pure pyridine because the density of the pure base is outside the range that was specified. Thermodynamic and physico- JUNE 1964 chemical measurements were made on a sufficient number of lower members of the series (viz. pyridine picolines and lutidines) to enable values for many higher members to be estimated with confidence.A series of phenol homologues (phenol cresols xylenols and ethylphenols) was next chosen for study and an extensive set of measure- ments was made. Again there was much uncer- tainty in the values of some simple physical constants before the work commenced. For example in tables of collected density values for phenols at 25" there was considerable confusion because the compilers had not realised that many of the early values referred to impure liquids whereas the pure compounds are solids at this temperature. Assistance in the choice of substances for study and of properties for measurement is received from this laboratory's Chemical Ther- modynamics Committee which comprises mem- bers from industry and the universities working under the chairmanship of Professor D.H. Everett. At the request of this Committee the Association of British Chemical Manufacturers asked their members what additional thermo- dynamic data would be most useful. Their replies and a subsequent literature survey re- vealed the need for data on organic oxygen- and fluorine-containing compounds and on certain inorganic phosphorus compounds. The Group is now measuring the thermodynamic functions of these materials. In addition some work on uranium compounds is being undertaken in co-operation with Atomic Energy Research Establishment Harwell. Properties Measured 0rganic Compounds.-T he range of measurements on organic compounds includes heat of combustion low-temperature heat capacities vapour heat capacities latent heats of vaporisation and fusion vapour pres- sures gas-imperfection data and the derivation of thermodynamic functions from molecular spectra.Electrolytic dissociation constants and gas -phase equilibrium constants are also measured. The heats of combustion of organic com-pounds which yield heats of formation are determined in a static-bomb or rotating-bomb calorimeter. The combustion products are care- fully analysed to establish the extent of reaction; for example carbon dioxide is determined gravi- metrically to an accuracy of 1 part in 10,OOO. Desk calculation of the standard heat of forma- tion from the combustion data is tedious and a programme for the application of the necessary correction terms which may number as many as 100,has been written for an electronic computer.The static-bomb calorimeter incorporates an electrical heater so that the energy equivalent of the equipment can be determined in absolute joules per degree by the use of standards of electromotive force resistance and time certified by the National Physical Laboratory. In the rotating-bomb calorimeter a chemically and thermodynamically-defined final state is ob- tained when water or other suitable solvent pre- viously placed within the bomb is caused to wash the walls of the vessel after combustion. This equipment is particularly valuable when the re- action products are a complex mixture or are highly corrosive or have a high heat of dilution. The energy equivalent of this calorimeter is found by the use of thermochemical-standard benzoic acid certified with the aid of the static- bomb calorimeter.At present the rotating-bomb calorimeter is being used to measure the heats of combustion of aromatic fluorine-containing compounds. Values for the heats of solution of carbon dioxide in hydrofluoric acid previously obtained in this laboratory by a temperature- solubility study of this system are used in the calculation of the standard heat of formation. One object of the work on fluoro-compounds is to establish bond-energy relations and it is in- teresting to note that hexafluorobenzene is about 40 kcal. mole-1 less stable than was anticipated. Another object of this study is to provide second- ary standards for combustion calorimetry.Measurements have confirmed an American suggestion that p-fluorobenzoic acid is a suitable standard and have also shown that pentafluoro- benzoic acid is a good standard of high fluorine content. Three cryostats are in use for heat-capacity measurements in the temperature range 12-330"~.Entropies calculated from these observa- tions are combined with heats of formation to obtain free energies of formation. Heat capacity measurements also give information on solid- state transitions while values for the temperature of the triple point heat of fusion and purity of material are obtained if the substance melts in the temperature range studied. Each calorimeter is surrounded by an adiabatic shield the temper- ature of which is maintained within &0.002" of the temperature of the calorimeter by automatic controls and occasional manual adjustment.Cooling of the calorimeter and shield is effected by conduction through a thermal switch to a refrigerant tank containing liquid helium or nitrogen. Measured amounts of energy are then introduced by means of an electrical heater. The temperatures of the calorimeter before and after the energy is introduced are measured by a platinum-resistance thermometer enclosed in a platinum capsule. Heat-capacity values derived from these observations are plotted against temperature and values of entropy and enthalpy are obtained by means of a digital computer from values read from the graph.The per- formance of the equipment has been tested by the use of heptane and synthetic sapphire. Heat capacities of vapours and the latent heats of vaporisation are measured at temperatures up to 230" and in the pressure range 0-25-1 atm. by means of a flow calorimeter. Heat capacities of the vapour as an ideal gas obtained by extra- polation from the experimental data can often be compared with the results of statistical thermodynamic calculation. Comparison of the directly measured enthalpies of vaporisation with those calculated from vapour-pressure and gas-imperfection data provide an overall assess- ment of the accuracy of measurements of these properties. Saturated vapour pressures in the range 150-1000 mm. Hg are measured ebullio-metrically.In this technique the boiling point of the compound is measured to &0.001" at each of a series of pressures. The pressure of the sample vapour is calculated from the boiling point of a standard substance (e.g. water) under the same pressure by the use of accurate vapour- pressure data for the standard substance. Temperatures are measured with platinum-resistance thermometers. An Antoine equation log,,P =A -B/(C + t) where A B and C are constants is fitted by a computer to the data. This type of equation is convenient for the calculation of pressures from temperature and vice-versa and usually fits the experimental data well. Vapour pressures of certain substances in the range 0-01-1 mm. Hg (temperatures 0-5O"c) are determined by a transpiration method in which a measured volume of nitrogen at a known PROCEEDINGS near-atmospheric pressure is equilibrated with the compound at each of a series of tempera- tures.The vaporised compound is recovered from the saturated carrier gas and the amount measured by a suitable analytical procedure. To measure critical pressures the sample is confined in a glass tube over mercury which serves to transmit the pressure to a dead-weight piston gauge. The upper end of the glass tube is heated in an oven provided with a window for observation and conditions are arranged so that the vapour-liquid meniscus is visible at a temperature just below the critical. The tempera- ture is allowed to oscillate slowly through the critical value so that the meniscus disappears and reappears and the pressure is measured to an accuracy of &O.Ol atm.Critical temperatures are measured by observation of the temperatures of disappearance and reappearance of the meniscus in a sealed glass tube heated in an oven consisting of a massive bronze block. Critical densities are obtained from measurements of orthobaric densities made over a range of temperatures with a series of sealed tubes each containing a different amount of material. The movement of the menisci in the tubes as their temperature is raised allows the densities of the liquid and vapour phases to be calculated in terms of the liquid density at room temperature. The deviations of organic vapours from ideal- gas behaviour at pressures below 1 atm.are studied by measurement of the pressure of a known volume of vapour confined over mercury at a closely controlled temperature. The vapour is compressed to another known volume and the pressure again measured. The observations are repeated with various amounts of vapour. Pres- sures are determined with the aid of a null-point diaphragm separating the vapour from a gas whose pressure is measured manometrically. Facilities for the study of infrared spectra include a diffraction grating spectrometer for measurements in the range 8850-375 cm.-l and a spectrometer with a caesium iodide prism for the range 420-200 cm.-l while measurements down to 20 cm.-l have been made by the use of a Michelson interferometer as part of a joint programme of work with Basic Physics Division National Physical Laboratory.For the study of Raman spectra equipment is available with photographic and photoelectric recording; de- polarisation measurements can be made by the JUNE 1964 latter recording technique. The fundamental vibration frequencies are found from a careful study of spectral data on a compound and on related materials supported in some instances by observations on suitably deuterated specimens and by force-field calculations. The calculation of thermodynamic properties from the fundamental frequencies and molecular- structure data provides a unification of the experimentally measured quantities and extends the information to a much wider range of temperatures.The usual practice is to calculate the following functions for the ideal-gas state free energy function (G"-H;)/T heat-content function (H"-H,")/T heat content (H"-Hi) entropy So and heat capacity C,"; H," is the heat content at O'K. Values of these functions are usually tabulated for 273.16" and 298.16"~ and for the range 300-1000"~ at intervals of 100 deg. K. Electrolytic-dissociation constants at 25Oc of bases (e.g. pyridine homologues) and of acids (e.g. phenols) are determined by a spectroscopic technique using buffer solutions of accurately known pH. Gas-phase equilibrium constants are being measured to provide a check on the accuracy of thermally derived values of free-energy heat and entropy changes of chemical reactions.Some time ago the pyridine-hydrogen-piperidine equi- libria were investigated and at present ketone- hydrogen-secondary alcohol equilibria are under investigation. In these experiments a liquid mix-ture of a ketone and the corresponding second- ary alcohol is vaporised and the mixed vapour is passed through a heated glass tube containing a solid catalyst to yield an equi1ibr;ium mixture which is quantitatively analysed by gas-liquid chromatography. Care is taken that equilibrium is attained and by-product formation is care- fully controlled and measured. The experiments are carried out with the catalyst at different temperatures. Properties Measured Inorganic Compounds.-The programme of thermodynamic measure-ments on inorganic compounds is at present much smaller than the organic project.In addi- tion to the present study of phosphorus and uranium compounds by low-temperature calori- metry work is proposed on heat-capacity measurements to a high temperature and on the heats of formation of selected materials. Preparation of Highly PuriJied Samples of Organic Compounds.-The measurement pro-ject is supported by a preparative and purifica- tion programme. Preparative organic chemistry finds application in the synthesis of compounds unavailable commercially but required to com- plete a series (e.g. m-cresol). Occasionally pre- parative chemistry is used to make a derivative for purification when the parent compound is difficult to purify.The desired compound is subsequently regenerated from the purified derivative. Samples with a purity better than 99.9 moles per cent are obtained by physical purification methods such as distillation sublimation zone melting and fractional freezing. Small samples sufficient for the study of infrared spectra are prepared by the use of a gas-chromatographic column. Gas chromatography is often employed to follow the course of a purification and is used to assess the purity of a final product when freezing-point or melting-point procedures are inapplicable. Pure samples surplus to the requirements of the measurement programme and pure materials purchased in bulk are sold as standard samples for use in industry for the calibration of analytical equipment such as spectrometers employed in the analysis of plant-stream products.N.C.L. Standard Samples are also employed in univer- sity research. For example specimens have been used for the investigation of the physical pro- perties of solutions infrared and vacuum ultra- violet spectra densities of gases andvapours com- pressibilities and transport properties. Samples have been used for the study of the mechanisms of chemical reactions induced thermally and by radiolysis and photolysis. Publications.-Thermodynamic values deter-mined in this project together with descriptions of techniques developed for purification purity control etc. are made available by publication in the journals and transactions of scientific societies or as books.Approximately 100 papers and three books have been published in the last decade. Ultimately of course the measurements appear in data compilations. Thus for example N.C.L. measurements of critical temperatures of 40 compounds have been included as a table in the 44th Edition of "Handbook of Chemistry and Physics" (The Chemical Rubber Publishing Co. Cleveland Ohio 1962-3). Vapour pres- PROCEEDINGS sures of phenols cresols and xylenols have ap- peared as selected values in the “Manufacturing Chemists Association Research Project” and certain physical properties have been presented in “The Coal Tar Data Book” (The Coal Tar Research Association). Large compilations such as those of Landolt-Bornstein also include the results of N.C.L.measurements. Facilities Available to Industry and Guest Workers.-From this account it can be seen that the National Chemical Laboratory has accepted responsibility for the provision to industry of first-class thermodynamic data. In some in-stances it is possible to arrange to have measure- ments made on a payment basis; an example is the determination of the calorific value of highly purified specimens of benzoic acid. Indeed the calibration of nearly all the combustion calori- meters in Britain depends upon the use of samples of benzoic acid certified by the National Chemical Laboratory and issued by chemical supply houses. Investigations can often be carried out as a joint project as exemplified by the study of uranium carbides as a collaborative effort with the United Kingdom Atomic Energy Authority which is responsible for the supply and charac- terisation of the specimens employed in the work.Scientists from industry can use the facilities at Teddington in a joint programme for example a guest worker has been seconded to collaborate in the study of phosphorus com-pounds. Industry can also expedite measure- ments by the supply of purified samples for measurement; such provision (for example a recent donation of hexafluorobenzene) will be acknowledged in publications. Students from universities can be accom-modated and thus a small team from Marseilles have studied cryoscopic methods for purity control. Any person interested in utilising the equip- ment or in making measurements on important compounds or who wishes to have a list of publications of the Chemical Thermodynamics Group should write to the Director National Chemical Laboratory Teddington Middlesex.LIVERSIDGE LECTURE* Some Basic Problems of Solid-state Chemistry By J. S. ANDERSON CHEMISTRY UN~VERSITY (INORGANIC LABORATORY OF OXFORD) MUCHfruitful work in the chemistry of solids during the past two decades has been strongly influenced if not directly stimulated. by the concept of inherent imperfections in crystals. The idea that thermo- dynamic equilibrium corresponds to the presence of a certain equilibrium concentration of point defects goes back 30 years; in particular to a fundamentally important analysis of the statistical thermodynamics of crystals by Schottky and Wagner in 1931.l In this original formulation of the ideas and in the im- mediately succeeding years the role of point defects in transport processes within solids in chemical re- actions and as a basis for the stoicheiometric vari- ability of numerous inorganic and intermetallic compounds was sketched out.Although the ideas passed only slowly into the general currency of chemical thought intensive interest has developed within the last few years in the theory and in those aspects of solid state chemistry to which it most obviously applies. Just as these ideas have become the regular back- ground to thought about the inorganic and physical chemistry of solid materials we find ourselves on closer examination confronted with a real dilemma.The concept and theory of lattice imperfections is so straightforward and so free from arbitrary assump- tions that it cannot but be valid. The very chemical systems to which 10 years ago it appeared to supply the interpretation now prove to be outside the scope of the defect solid theory at least in its present form. The aim of this lecture is to examine where we stand in relation to one intriguing problem of inorganic chemistry the stoicheiometry of solid compounds and to suggest the way in which theory will probably develop. We are concerned with point defects unoccupied lattice sites atoms in positions inappropriate for the particular crystal structure or atoms on wrong lat- tice sites.In a stoicheiometrically perfect crystal point defects necessarily have to be created in corn- * Delivered before The Society at The Royal Institution London on December 12th 1963; at the University College I North Wales on January 30th 1964; and at the Marischal College Aberdeen on February 13th 1964. Schottky and Wagner 2.phys. Chem. 1931 By 11 163. JUNE 1964 plementary sets. Their formation involves an endo- thermic change in the electrostatic energy of an ionic crystal (or the bond energy of a covalent crystal) partially offset by the relaxation of the crystal lattice around the point defect. The entropy change is also positive being dominated by the configurational entropy term that arises from a random distribution of the point defects within the crystal.There must be some change in vibrational entropy also but this is generally assumed to be unimportant. On balance the equilibrium constant for internal disorder of a crystal is in general dominated by the large endo- thermicity of creating point defects. It follows that the concentration of point defects in thermal equilibrium is steeply temperature dependent. Reliable information about the concentrations of point defects in thermal equilibrium in crystalline compounds is scanty but the evidence is that even at temperatures approaching the melting point the concentration in most types of crystals is very small. There are however a few compounds in which the equilibrium concentration of point defects in the stoicheiometric crystal is extremely high.Indeed it looks as if the known materials fall rather sharply into two groups those with the low defect concentra- tions conformable with Schottky-Wagner theory and those with anomalously high concentrations of point defects. Inherent in the theory of the imperfect crystal is the further notion that the stoicheiometrically per- fect crystal loses its unique significance. In it the different modes of point defects must be exactly complementary but in principle the statistical theory would permit a surplus of atoms or ions of one or other kind thereby creating an excess of one type of point defect together with some kind of valence anomaly. This can be variously described in chemical terms as due to ions of lower or higher valence or in terns of band theory as electrons or positive holes trapped at or in the neighbourhood of point defects.Within the limits of tolerance of the crystal lattice (and in this form the simple theory gives no hint as to what determines that tolerance) the composition of a crystal depends on the chem- ical potential of its components in the system. The equilibrium conditions can be formulated in terms of statistical thermodynamics and like the internal equilibrium are dominated by an enthalpy term- the work of creating an additional point defect of effecting a valence change in one or other kind of atom and the resulting change in lattice energy- which can be taken as subsuming those chemical factors which differentiate one system from another.The compounds of many of the elements apparently conform perfectly to the law of constant composition while others (for example the oxides of the transition elements) are notably non-stoicheiometric. This elaboration of the statistical thermodynamic theory is again so general that it ought to be uni- versally applicable. If we knew the necessary energy quantities it should be possible to explain in terms of it all the known facts of solid-state inorganic chemistry. One necessary proviso is that no account is taken of the interaction of point defects with one another; the model is strictly applicable only for very low concentrations of point defects or very small departures from ideal composition.However it has proved a powerful stimulus to thought about chemical systems where this last proviso is certainly not applicable. These are the properties that are used and con- trolled in the chemistry of the compound semi- conductors. A stoicheiometric excess of the metallic component introduces excess electrons while a stoicheiometric excess of the non-metallic component leaves the crystal with a deficiency of electrons i.e. with positive holes. In compound semi-conductors electrical measurements of the concentration of cur- rent carriers can be used to investigate how the stoicheiometry of the crystal is shifted by a change in the external variables and we owe what little we know about the thermodynamics of slightly non- stoicheiometric crystals to work of this kind.As an example lead sulphide exists over a com- position range on either side of the ideal composi- tion.2 Above about 550" (below which diffusion is too slow for equilibrium to be reached) the com- position of a crystal will adjust itself to equilibrium with the prevailing partial pressure of sulphur. The composition range is at its broadest between about 700" and 1050" extending from about PbS,.,,, to PbS1. ooo25. The evidence3 indicates that the stoi- cheiometric variation is due to the presence of an excess of interstitial lead atoms or lead atom vacancies in the metal-excess and sulphur-excess ranges respectively and although the point defects are dilute it is necessary to take account of a strong association between the point defects and the excess electrons or positive holes.For most solid compounds the stoicheiometric range is very small and any deviations from stoi- cheiometry escaper detection. Nevertheless because of the highly endothermic enthalpy change involved defect equilibria shift markedly with temperature. Non-stoicheiometric behaviour therefore becomes of real importance both scientifically and techno-logically in highly refractory compounds at the highest temperatures. For example the general chem- istry of thorium shows no disposition for variability of valency; thorium dioxide is the most stable of Bloem Thesis Utrecht 1956; Bloem and Kroger 2.phys. Chern. (Frankfurt)1956,7 1.Simcovich and Wagner jun. J. Chern.Phys. 1963 38 1368. oxides showing no electronic conductivity or other evidence of stoicheiometric imperfection at any temperatures ordinarily attained in experimentation. Nevertheless the work of Thorn and Ackermann on the vaporisation of metallic oxides at temperatures above 2400" has shown that under these extreme conditions the expectation of the Schottky-Wagner theory appears to be fulfilled. Thoria loses oxygen and the congruently vaporising oxide in equilibrium with vapour4 is not ThO,.,, but ThOl.96. Other similar examples have been discovered and the stoicheiometry of refractory solids is currently a matter both of technological importance and of very active fundamental research especially in the United States.The foregoing is concerned with systems that conform fairly well to the uncomplicated theory. The Schottky-Wagner ideas were stimulating however because they provided an interpretation of the con- siderable number of known inorganic compounds- notably the oxides sulphides etc. of the transition metals and of the lanthanide and actinide elements- which show gross deviations from stoicheiometry. By a straightforward development of the theory it seemed possible to bring much of their chemistry within a rational framework. In fact however they pose some severe problems which have been exposed only as a result of the work of the past few years. In these compounds the stoicheiometric unbalance is so great as to connote the presence not of a few point defects per ten thousand lattice sites but of a substantial percentage of vacant sites or of inter- stitial positions occupied.In some cases the stoi- cheiometric compound is incapable of existence even though the stoicheiometric range may be rela- tively narrow. In other cases there is unequivocal evidence that a substantial percentage of lattice sites of each kind is unoccupied in the stoicheiometric compound itself. The example of the high temperature form of Ti0 can profitably be pursued further. Superficially it is an example of NaC1-type structure with 6:6 co-ordination having unbalanced Schottky defects that confer a very wide stoicheiometric range. In fact the extraordinarily high concentration of lattice vacancies would if randomly distributed imply that only a minority of Ti or 0 atoms over the whole composition range can have the proper co-ordina- tion environment (Fig.1. See p. 177). It is indeed not clear how far such a random model could properly be described as having the NaCl structure or why that structure should remain the stable configuration. Nevertheless the experimental facts are clear above 1100"c this highly defective structure statistically described as of NaCl type is the one formed. Ackermann Rauh Thorn and Cannon J. Php. Chem., Bertaut Acta Cuyst. 1953 6 557. PROCEEDINGS We need to enquire more deeply into the stability of non-stoicheiometric phases and the scale of departure from ideal composition that can be ex- pected.It could be argued in terms of the Third Law that the configurational entropy must trend to zero by the unmixing of any non-stoicheiometric compound into two phases of definite composition and perfect order. It would then follow that the non- stoicheiometric phases were unstable at low tempera- tures and that the entropy term was dominant in determining their stability at higher temperatures. In this connection it is relevant that Bertaut5 has attempted to calculate the electrostatic lattice energy for one composition in the iron-deficient iron sulphide phase FeS,., or Fe,S, for several alterna- tive models (i) for a completely disordered structure with higher valent cations and vacant iron sites dis- tributed at random the typical defect solid model; and (ii) for structures with the vacancies ordered into alternate cation sheets of the NiAs layer struc- ture but various different hypotheses about the dis- tribution of the Fe3+ cations.He concluded that although the best ordered arrangement of the Fe3+ ions contributed some additional stabilisation energy the energy gained by ordering the vacancies alone is very large (around 320 kilo cal. per mole). It must in. fact swamp the entropy term in the free energy of the crystal. The implication is that the ordered crystal should be of rational formula enormously more stable than the defective one at all tempera- tures. Non-stoicheiometric pyrrhotite ought to be metastable and should anneal into a mixture of dis- crete compounds of fixed formula.Comparable lat- tice energy calculations are not available for other examples but the general conclusion might well be the same; namely that thermodynamically permis- sible deviations from ideal compositions are very small under all conditions. Such a theoretical con- clusion is at variance with a large body of experimental evidence. Nevertheless work of the last few years has clearly established that in a large number of instances compounds which had been reported as non-stoicheiometric undergo disproportionation when they are annealed under conditions which allow slow solid state diffusion and re-arrangement processes to occur. They give place to successions of phases of fixed composition and sometimes rather complex formula the series as a whole constituting the in- organic equivalent of a homologous series.The inter- mediate phases in many cases have large unit cells implying the operation of ordering forces over long distances within the crystal lattice. The term homo- logous series is apt because the members of any homologous series are related through the operation 1963 67 762. JUNE 1964 of a common structural principle which replaces a random structure with vacancies or interstitial atoms by an ordered structure with no more than the normal level of point defects. The crystal structure evidence is that a surplus of point defects is elimi- nated from the crystal lattice in one or other of two ways? by superstructure ordering or by a re-arrangement of the co-ordination around a propor- tion of the ions in the crystal lattice.The first way of passing from a random defective structure to an ordered and stoicheiometrically de- fined compound is exemplified (Table 1. See p. 177) by the chromium sulphides. The higher sulphides can be regarded as being derived from the NiAs-type CrS by the omission of up to 33% of the metal atoms. According to Haraldsen’s pre-war investigations the range between CrS,. and Cr,S3 (or Cro.67S) was covered by series of non-stoicheiometric sulphides with broad composition ranges. The true equilibrium condition however appears to correspond to the existence of a homologous series of well defined ~ulphides,~ conforming to the general formula Cr,S,+, formed by an ordering of the metal atom vacancies.This ordering takes place at two levels of intensity. Segregation of the vacancies into alternate sheets of cations is strongly favoured energetically in accordance with Bertaut’s calculations already referred to. Some recent work by Benard on the TiS-TiS system also based on the NiAs-CdI structure topotaxy suggests that the driving force behind this level of ordering is so strong that it determines the broad structure from the outset. Within the metal-deficient sheets the vacant cation sites can then undergo a further ordering process and the discrete phases of the homologous series correspond to the fractions of vacant sites which can be ordered rationally in a hexagonal network of cations as shown purely schematically in Fig.2 (See p. 177). The chemistry of the lanthanide oxides is also determined by this kind of relationship between non- stoicheiometric and defined ordered phases. It has been very largely clarified in the past few years chiefly by the work of Brauer Eyring and Bevan.8 The cubic sesquioxides can themselves be formally regarded as oxygen-deficient fluorite structures with one quarter of the oxygen positions vacant but the vacancies are strictly organised to give a new struc- tural pattern and new symmetry with all the metal atoms present in [MO,] co-ordination polyhedra of two kinds. The higher oxides of cerium praseo- dymium and terbium have long been undefined but it is now clear that the equilibrium state corresponds to the existence of a whole homologous series of oxides Mn02n-2, derived in concept either by putting oxygen atoms into the vacant sites of the type C sesquioxide or by creating vacancies in the perfect fluorite MO structure; but once again the remaining vacant sites are strictly ordered.The most thorough work of this kind (Fig. 3. See p. 178) by Bevan Hyde and Eyring on the Pr,O,-Pro system is important because it shows two things that these well-defined stoicheiometric phases of the homologous series themselves display minor variations of stoicheio- metric composition depending on the chemical potential of oxygen in the system; and that at sufficiently high temperatures these ordered struc- tures unquestionably pass over-either by a first- order phase reaction or by an order-disorder transi- tion-into a true non-stoicheiometric phase.This can be described as derived from the fluorite struc- ture by the random introduction of up to about 15 % of vacant anion sites but we may still be left with the problem of defining the term “random.” The second way of eliminating point defects is exemplified by the so-called shear structures found by Magneli in the homologous series of transition metal oxides; for example the molybdenum and tungsten oxide series Mn03n-m,or the titanium and vanadium oxides Mn02n.-l.These consist essentially of slabs of perfect structure-the ReO structure the MOO layer structure or the rutile structure in the cases cited-joined together by an abnormal linkage of co-ordination polyhedra that economises in the use of oxygen atoms.s For example the ReO struc- ture consists of [MO,] octahedra linked by sharing corners.Blocks of this structure are linked in the intermediate oxide phases by grouped columns of edge-sharing octahedra (Fig. 4. Seep. 178).In therutiie structure [MO,] octahedra share edges; in the lower intermediate oxides slabs of rutile structure are linked through groups of octahedra sharing faces. A real gap in our knowledge is that the chemistry and the thermodynamics of these interesting oxides is totally unknown. In principle they must be formed by the abstraction of oxygen from the perfect higher oxide structure and one can envisage three stages of organisation of their complex ordering (i) the creation of oxygen vacancies (ii) the segregation of oxygen vacancies into suitably spaced arrays (iii) a co-operative re-arrangement along planes of shear whereby the oxygen vacancies are eliminated A.D. Wadsley Adv. Chem. 1963 No. 39 Non-Stoicheiometric Compounds p. 23. ’Jellinek Thesis Utrecht 1957; Haraldsen 2.anorg. Chem. 1937 234 337 372. Brauer and Gradinger 2. anorg. Chem. 1954 277 89; Brauer Gingerich and Holtschmidt J. Inorg. Nuclear Chem. 1960 16 77; Brauer and Gingerich ibid. p. 87; Bevan J. Inorg. Nuclear Chem. 1955 1 49; Bevan Hyde and Eyring 3rd Rare Earth Conference 1963 Clearwater Florida U.S.A. Cf. Wadsley ref. 6; Kihlborg ibid. p. 37. (Figs.5a and 5b. See p. 179). The remarkable feature of the structures is the repeat of the shear planes at perfectly regular intervals rather than at more or less random spacings thus creating stacking faults. In the latter case non-stoicheiometric oxide phases would be formed such as have been found by SatolO in the reduction of U,O,. The size of the repeating unit in these shear structures is clear evidence that perturbation of the parent structure extends well beyond the distorted co-ordination polyhedra and the sites of the alter-valent ions. Kihlborg,ll in very careful work has shown that oxygen atoms are dis- placed from their idealised positions across the whole width of the approximately unperturbed slabs so that the change in co-ordination and the localisa- tion of alter-valent ions does in fact exert long- range influence.What is not known is whether these Magneli structures display any stoicheiometric vari- ability at all and whether they undergo reaction or transition into non-stoicheiometric phases at elevated temperatures. In the light of the foregoing considerations we have to reconsider the constitution and the origin of the stability of non-stoicheiometric compounds with broad composition ranges. In doing this we should re-examine the factor explicitly left out of the original model the interaction between lattice defects especially when they are no longer very dilute. Such interactions prove to be important. As perturbations of the regular pattern of charge distribution in the crystal lattice point defects bear a virtual charge.This implies that complementary intrinsic defects tend to associate or cluster. Inter- action is stronger with the alter-valent ions of a non- stoicheiometric phase so that point defects and valence anomalies mutually trap each other. The magnitude of the interaction energies can be ob- tained from suitable thermodynamic measurements but such evidence is available so far in very few cases. In non-stoicheiometric oxides and similar systems the interaction energies are quite large and the association between point defects and alter- valent ions is tight. Contrary to what might have been intuitively guessed the interaction between like defects is also attractive (cf.Table 2. See p. 182). Conceptually this must be so since it is the accumulation of lattice defects (and their complementary ions of abnormal valency) in some element of volume that is respons- ible for the nucleation of a new structure and a new phase when reactions take place in the solid state. In a dilute system the defects may be randomly dis- tributed; as their concentration exceeds a certain value they tend to segregate into defect-poor and defect-rich regions which can co-exist. At any temperature the concentration at which this occurs PROCEEDINGS depends on the nearest-neighbour attractions and the magnitude of the attractive interaction energy defines some critical temperature above which no segregation of defects occurs. The statistical thermo- dynamic model of the defect solid was extended a number of years ago to include this idea which gives a fair account of phase ranges.A good deal of attention has recently been given to interpreting a number of experimental systems in the light of this formal theoretical model. It is becoming clear that interaction between lattice defects may involve more than this statistical clustering. There may be a strong association of point defects and valence defects in new configura- tions involving and in some cases displacing atoms from their “normal” positions in the crystal lattice. The formal description of lattice defects in terms of vacant sites and of interstitial atoms is too crude; we need more specific information about their real nature.Such information is now becoming available especially as neutron diffraction is applied to the location of the light atoms in oxides and hydrides and the location of magnetically distinguishable alter-valent atoms. Three pieces of evidence may be cited in this connection. They provide also a link with the prob- lem of the structural and thermodynamic relation between non-stoicheiometric phases and the ordered intermediate compounds. Ferrous oxide exists only at high temperature as an iron deficient phase Fe,-,O. On the basis of the evidence hitherto obtain- able it has been described in terms of the presence of vacant cation sites and a corresponding propor- tion of Fef3 atoms with all the iron atoms on the octahedral sites of the NaCl type structure.In 1960 from a neutron diffraction study of Feo.860 Roth12 concluded that this model was wrong and that the iron deficient oxide involved the presence of defect complexes (O+),(Fel A) in which an Fef3 atom adjacent to a cation vacancy had moved into an inter- stitial position thereby creating a second cation va- cancy (Fig. 6. See p. 180).Within a cubic close-packed oxygen lattice we thus have Fe+2 ions on octahedral sites surrounding this defect complex of an Fe+3 atom on a tetrahedral co-ordination site. This cor- responds to a minute element of volume out of the spinel structure of Fe304. It is defect complexes of this type which are randomly disposed throughout the crystal lattice of the ferrous oxide.If they tend to aggregate then a larger element of volume could be described as having essentially the Fe,04 struc- ture. There is a sense in which the germ of the next higher phase is already latent in the non-stoicheio- metric oxide. A comparable situation has been found in the lo Sato Doi Ishii and Uchikoshi Acta Cryst. 1961 14 763. l1 Kihlborg Zoc. cit. ref. 9 and personal communication. l2 Roth Acta Cryst. 1960 13 140. JUNE 1964 uranium oxide system where at high temperatures there exist both the random non-stoicheiometric phase U02+aand the U40gphase (which may be oxygen deficient). Both are formally described as arising from the insertion of interstitial oxygen atoms into the fluorite type cell of UO,. That description is valid in the sense that up to one additional oxygen atom per unit cell is introduced without a gross perturbation of the structure.The additional oxygen atom is however not just an interstitial oxygen in the traditional sense but is present in a defect com- plex which is of the same type in the random non- stoicheiometric oxide and in the ordered phase of nearly ideal composition.13 For each interstitial oxygen that enters an oxygen atom is displaced from a regular lattice site and there is formed a defect complex which could be described as two interstitial oxygen atoms on either side of a vacant anion posi- tion,(U-)(O2- IA),(Fig.7. See p. 181).Geometrically this displacement process makes room for the addi- tional oxygen without making any interatomic distan- ces too small.The pair of oxygens involved in the defect complex form a bridge between a pair of uran- ium atoms which may be the favoured sites for the pair of (mobile) positive holes created in the oxidation. In the u4og phase there is one such defect complex in every unit cell and the orientations of the com- plexes in different fluorite cells are so correlated as to build up the large super-structure cell of the com- plex U40 structure. In the UO,+ phase the defect complexes will presumably be distributed and oriented quite at random as long as x is small. How- ever the phase diagram shows that at high tempera- tures the composition of the random UO,, phase can rise up to UO,., or above that value so that there is statistically one or rather more than one such defect complex in every fluorite unit cell.The question then arises how random is such a con- centrated system of defect complexes ? In this particu- lar case we have some direct thermodynamic evidence. From the extensive free energy measure- ments we may extract the information that the entropy of the UO,+%phase is higher than that of the U409-yphase so that in the oxidation of the random non-stoicheiometric phase to the ordered nearly perfect phase there is a large entropy decrease as long as the phases differ substantially in composition. At increasing temperatures as the composition difference between the two co-existing phases diminishes so also does the entropy of reaction.A short extrapolation shows that at 1127" where the phases are related by something like an order-disorder transition the conversion of the ordered intermediate phase to the random non-stoicheio- metric phase involves only a small entropy change.l* The UO,, phase has the higher entropy; it retains an element of randomness; it has a variable composi- tion. But its entropy is considerably lower than it would be for a completely or highly random distribu- tion of the defect complex over all orientations and all possible positions throughout the structure. As a third illustration we may take the structural and thermodynamic relation between the a-and /%phases of the tantalum-hydrogen system. The former is regarded as a true random solid solution of hydrogen in tantalum the /?-phase is nominally Ta,H though almost certainly having a range of composition.The X-ray evidence shows that the @phase transforms to the a-phase at temperatures below 56" depending on the particular p-phase com- position (Fig. 8. Seep. 18 1). The thermodynamics show that the crystallographic transformation does not co- incide with the main entropy increase but that in fact there are three successive A-points within the stability range of the ,&phase leaving only a small amount of configurational entropy to be gained at the stage where long range order disappears. More- over if the a-phase is truly a disordered solid solu- tion the total entropy change is smaller than would correspond to the dissolution of a fully ordered Ta,H structure.There persists an element of randomness in the so-called ordered intermediate hydride Ta,H. From the combination of X-ray diffraction which locates the ordering of heavy atoms neutron diffrac- tion which locates the ordering of hydrogens and the thermodynamic evidence Wallace and his co- workers concluded that this system must be described in terms of the ordering of hydrogens around the metal atoms in what he termed basic structurd units.15 These correspond to the defect complexes already mentioned. Because only two hydrogen atoms per tantalum metal cell can be accommodated only half the hydrogen positions of the basic struc- tural unit are occupied and there is a residual ran- domness in the /3-phase structure (Fig.9. Seep. 182). It is the relation of these basic structural units to one another that determines the entropy changes and that progressively leads from the most highly organised long-range order to the break-down of all order in the /%+a phase transition. Even below the lowest A-point there is not perfect long-range order but the basic structural units achieve parallel alignment over small regions or microdomains. The concept is related to that which has been invoked by Aston to explain the classical non-stoicheiometric palladium- hydrogen system. The situation this far may be summarised as follows non-stoicheiometric phases undoubtedly l3 Willis Proc. Roy. Soc. 1963 A 274 132 134; A.E.R.E. R 4414 (1963). l4 Roberts and Walter J.Inorg. Nuclear Chem. 1962,22 213; Roberts Adv. Chem. 1963 No. 39 p. 66. l5 Wallace et al. J. Chem. Phys. 1956,24 634; 1961 35 2148,2156. exist in true equilibrium with a degree of imper- fection that is structurally and thermodynamically surprising. We cannot leave out of account the inter- actions of defects with one another and with the sur- rounding crystal lattice. ‘In many cases this leads to ordering or elimination of defects with the formation of intermediate phases but where direct thermo- dynamic information exists the difference between ordered phase of definite stoicheiometry and random non-stoicheiometric phase is smaller than would be expected. Where direct structural information exists we find that this interaction between defect and crystal lattice leads to a local rearrangement and the creation of a new defect cluster or in Wallace’s terminology a basic structural unit.When their con- centration is high these basic structural units must further interact and build up a region displaying more extended but still short-range order. We arrive in this way at a concept which has come independently to a number of people but was prob- ably first stated in more or less explicit terms by Ariya.lG This is that the grossly non-stoicheiometric compounds are inherently characterised by what Ariya called microhefevogeneity laying the emphasis on fluctuations of structure and composition from point to point or what others have referred to as microdomains laying the emphasis on short-range ordering of defect clusters or basic structural units.Ariya’s view was based on two main considera- tions (1) the integral heats of formation of a non- stoicheiometric compound throughout its composi- tion range are in many cases (for example the FeO VO and Ti0 phases) linear functions of cornposition between the values for the next higher and next lower stoicheiometric phases ;(2) the magnetic pro- perties are also linear functions of composition but may vary differently on either side of the ideal com- position. By themselves Ariya’s arguments are not conclusive. The integral enthalpies are not very sensi- tive indices of structure. The average energy gained per atom of oxygen combined (for example in the iron and vanadium oxides) changes only gradually throughout the series of compounds even though the successive phases differ completely in crystal struc- ture.The relevant point is that the biggest contribu- tion to electrostatic energy comes from nearest neighbour interactions and Ariya may well be justi- fied in arguing that the V3+ions in the VO1+ com-position range have on the average the same fully tenanted octahedral co-ordination environment as the cations of V,O,. We do know however that the partial molal thermodynamic properties are strong functions of composition throughout the range of most non-stoicheiometric compounds and any valid Ariya and Morozova Zhur. obshchei Khim. 1958 28 PROCEEDINGS theory has to reproduce that behaviour satisfactorily.The magnetic properties of transition metal ions are sensitive to the crystal field of their immediate nearest neighbour co-ordination environment and their evidence leads to permissible deductions about the immediate short-range order around the ions of two valency states in these non-stoicheiometric com- pounds. On this basis Ariya’s description of the Ti0 and VO phases is that they are sub-microhetero- geneous in structure; in these crystals are regions with the composition MOO.,,containing MI+ atoms within which half of the oxygen positions are vacant and regions with the composition MO,., containing M3+,within which one third of the metal atom posi- tions are vacant (Fig. 10. See p. 183). In accordance with the data from X-ray investigations the occupied lattice sites in all the regions are those appropriate to a structure of the NaCl type.In terms of our previous discussion such compounds involve defect clusters or basic structural units of two distinct kinds which aggregate into microdomains with a new short-range order. Within each microdomain the ordering may be nearly perfect. How big are these microdomains? Ariya has sought to estimate this from the consideration that the observed concentrations of anion vacancies and cation vacancies have to be accommodated in micro- domains of two kinds which fit together by boundary regions of nearly perfect order. He concluded that over the stoicheiometric range of the VO and Ti0 phases the size of the microdomains would range from about two to ten unit cells in linear dimensions.The idea is not that of a mixture of two phases. The boundary regions are inherently common to domains of two kinds and the whole situation is dynamic. The atom movements of self diffusion lead continually to the transfer of any element of volume from one type of microdomain to another so that the position and the size of the microdomains fluctuate at tempera- tures where inner equilibrium is maintained by self- diffusion processes. This is a qualitative picture. To serve a useful purpose it will have to be translated into a self- consistent statistical thermodynamic form. The model is essentially that of a single macroscopic phase with wide fluctuations approximating to two favoured configurations markedly different from the average structure and composition of the whole assembly of the crystal lattice.The thermodynamics of fluctuations has been treated notably by Smoluchowski in relation to nucleation and the kinetics of phase transformations in metallic systems. It has been shown that the probability of substantial fluctuations depends strongly on the shape of the free energy-composition curve in the neighbourhood 2617; Ariya and Popov ibid. 1962 32 2077. JUNE 1964 of the idealised stoicheiometric composition. For solid solutions in which substitutions have occurred this can be treated fairly simply in terms of nearest neighbour interactions. For binary compounds be-tween elements of markedly different polar character and with vacant sites or interstitial atoms as solute species formulation of the problem is less simple.An important point is that we are concerned with an equilibrium condition not with the role of micro- domains as germs for growth nuclei. At any tempera- ture within the stability range of a non-stoicheio- metric phase the probability that a microdomain will grow beyond some mean size characteristic of the temperature falls rapidly to zero. It is precisely here that the size of the microdomains is important as is the protean nature of their boundaries and their mutual overlap. It is not only the local lattice energy but the energy of these boundary regions that deter- mines and dominates the free energy of the whole system.Their thermodynamics is the thermo-dynamics of small systems,17 but it differs from cases hitherto treated in that we do not deal with small solute particles within a solvent medium or potenti- ally growing protonuclei within a matrix. There is no continuous phase; both types of domain are micro- domains and the fluctuations of size and material transfer affect both kinds of micro-species present in the assembly. This is very much a problem for contemporary thermodynamics and the tools for treating it are only now being developed. l7 Cf. T. L. Hill J. Chern. Phys. 1962,36 3182; “Thermodynamics of Small Systems,” W. A. Benjamin Inc. New York 1963. COMMUNICATIONS A Wave Function and Chemical Formula for Benzene :A Correction By P.B. EMPEDOCLES and J. W. LINNETT* (Proceedings 1963 303) THEfirst line of the Table should read “Hiickel k = 3 leads to an energy 2-9 ev (67 kcal./mole) molecular orbital function Energy 115-4 ev.” The below that obtained by using the Hiickel function.” sentence beginning immediately above the Table The general conclusion of the communication is should then read “The non-pairing function with unaffected. * Inorganic Chemistry Laboratory Oxford. A Lead Hydride of High Stability By B. R. WELLSand M. W. ROBERTS* WE have investigated the interaction of atomic hydrogen with lead films formed by the evaporation of pure lead supported in an electrically heated tungsten basket the pressure during evaporation being < mm.and the temperature -195”. The films (between 4 and 40 mg.) were sintered at 23” in V~CUOand are known to have surface areas close to the geometric area of the reaction vessel (-100 cm.2). At 0” lead films absorbed large quantities of atomic hydrogen at a very fast rate and the volume absorbed is directly proportional to the weights of the films. The hydrogen is therefore distributed throughout the metal lattice and not confined to a region close to the surface. There is definite evidence for the formation of a metastable supersaturated hydride at 0”which when the hydrogen content is increased beyond a critical limit decomposes spontaneously to give a hydride of slightly lower hydrogen content. Subsequent reaction with atomic hydrogen is slow and we have taken the maximum hydrogen uptake for each film to corres- pond to this point.From the slope of the plot of hydrogen uptake against film weight we derive the formula PbH,., for the hydride formed at this stage. When the supply of atoms was interrupted a slow desorption of molecular hydrogen occurred at 0”,the extent of which increased with increasing hydrogen content but was never more than about 8% of the hydrogen uptake. This desorption cannot be accounted for by only recombination of surface * Department of Chemistry The Queen’s University of Belfast. adatoms and must be due to the presence of a small proportion of a phase which is only stable at 0” in the presence of a high concentration of hydrogen atoms.When the hydride (PbH,.,,) was heated to various temperatures below 160” hydrogen was evolved stepwise and when the rate of desorption had become negligible at 140” only about 25 % of the hydrogen had been desorbed. The activation energy of desorption for the hydrogen evolved up to 100” was between 16 and 20 kcal./mole-l. The desorption data therefore indicate that as the H:Pb ratio de- creases the stability increases suggesting the presence of discrete structural domains. Since desorption will be related to the degree of hydrogen mobility within the lattice it is likely that the activation energy for PROCEEDINGS atom mobility at low H:Pb ratios (-0-1) is much greater than 25 kcal./mole-l and may approach the Pb-H bond energy which is at least 52 kcal./mole.The hydrogen atoms are therefore localised and there is little overlap of the potential wells associated with the binding sites. This is compatible with the small H:Pb ratio. The hydride of composition PbHo.19 is therefore a stable state intermediate between the highly unstable PbH4 and the metal. This however does not imply that we have a single phase of the above composition on the contrary our data suggest that the lead-hydrogen system at 0”is composed of a number of quasi-independent domains probably varying in stoicheiometry. (Received May 6th 1964.) Temperature Dependence of the Secondary Isotope Effect on Aqueous Alkaline Ester Hydrolysis By E. A. HALEVIand (Mrs.) ZAFRAMARGOLIN* BENDER reported that aqueous alkaline and FENG~ hydrolysis of ethyl trideuteroacetate is 10 % faster than that of ethyl acetate at 25”.Since as one of us has stressed,2asb it is inadvisable to draw mechanistic conclusions from secondary isotope effects observed at one temperature in water we have undertaken an investigation of the temperature and solvent depend- ence of this isotope effect. Our results to date all in water at 0.0 35.0 and 65.0” along with Bender and Feng’s result at 25*0”,are Temp. 0” 25O 35” 65” kK/kD1.00 lfO.01 0.90f0.01 0.93 fO.01 1.15 A0.09 Although our use of Bishop and Short’s3 dif- ferential potentismetric technique allows us to work at lower concentrations and thus to deal with faster runs than do ordinary titrimetric methods the rates at 65” are still uncomfortably high and the figure quoted must be regarded as tentative.However a conventional statistical analysis of the fourteen runs carried out at this temperature confirms well beyond 99 % confidence limits that k > kD. The observed temperature dependence is un-precedented for secondary isotope effects and-to our knowledge-also for primary isotope effects and offers a serigus challenge to the application of general isotope effect theory to secondary isotope effect^.^ Since (E,D -E,“) [which is by definition -RPdln(kH/kD)/dT] is zero near 25 O Bender and Feng’s maximal inverse effect is due to entropy of activation and not to internal zero-point energy differences as they assumed. Moreover the tempera- ture dependence between 0” and 25” requires that EaD> EaH so that if the activation-energy difference at lower temperatures is ascribed to zero-point energy shifts these must be taken as favouring the protio- transition state and hence being in the “inductive” direction rather than in the “hyperconjugative,” direction as Bender and Feng supposed.We suggest however that the occurrence of an extremum near 25 O just where the dissociation con- stants of acetic and other similarly weak aliphatic acids such as propionic and butyric take on maxi- mum values 2a9596 is evidence that small differences in energy and entropy of solvation resulting from isotopic differences in average charge distribution largely determine the secondary isotope effects in this and other highly aqueous systems.Considera- tions similar to those applied to secondary isotope effects on aqueous solvolysis of alkyl halides and sulph~nates,~c~d reproduce the qualitative features of the temperature dependence satisfactorily. (Received March 31st 1964.) * Department of Chemistry Israel Institute of Technology Haifa. Bender and Feng J. Amer. Gem. Soc. 1960,82,6318. Halevi “Secondary Isotope Effects” in “Progress in Physical Organic Chemistry,” ed. Cohen Streitwieser and Taft Vol. 1 Interscience 1963 (a) p. 129 (b) p. 188 (c) p. 174 (d) p. 191. Bishop and Short Analyst 1962 87 467. 4 Bigeleisen and Wolfsberg in “Advances in Chemical Physics,” ed. Prigogine Vol. I Interscience 1958 p. 63. 5 Leffler and Griinwald “Rates and Equilibria of Organic Reactions,” Wiley 1963 p.48. 6 Conway “Electrochemical Data,” Elsevier Amsterdam 1952 p. 184. JUNE 1964 175 The Acidolysis of Dialkylzincs by p-Toluidine By M. H. ABRAHAM and J. A. €€ILL* THE only investigations so far reported of the Additionally we have shown that in the unsym- protolysis of the more reactive alkylmetallic com- metrical dialkylzincs EtZnPrn and PrnZnBun the pounds in which a sequence of alkyl groups has been relative rates of cleavage of the alkyl groups are and Et (100) Prn(52) and Bun (41) in di-isopropyl ether studied are those in which Grignard reagent~l-~ dialkylmagnesi~ms~ have been allowed to react with at 68". the weak acids indene? the chloromercury salt of Although the sequences of relative rates reported phenylacetic acid2 and hex-l-~ne.,,~ In all of these here are similar to those found for several electro- cases the relative reactivities of the alkylmetals were philic substitutions of tetra-alkyltins5 and dialkyl- in the order Pr' > Et > Prn > Me.mercurys6 in non-polar solvents our results do not We have now established the relative rates of allow a definite choice to be made between the acido- acidolysis of dialkylzincs by means of competitive lysis mechanism in which electrophilic attack at the experiments in which two dialkylzincs were treated a-carbon atom of the alkyl group preceeds co-ordina- with a deficiency of p-toluidine (which functions as a tion of nitrogen to zinc and the mechanism in which monobasic acid) co-ordination of nitrogen to zinc is the initial step.-f R,Zn + ArNH RH + RZn-NHAr At present though we favour the former mechan- Our results (Table) are expressed relative to Et,Zn ism and interpret our results following Gielen and Nasielski as due to an inductive effect (Pri > Et > Me2Zn Et2Zn Prn2Zn Pri2Zn Me) superimposed upon the usual steric sequence. 100 42 61 in Et20 at 35" -100 33 67 in Pri20 at 68" (Received April 9th 1964.) * Chemistry Department Battersea College of Technology London S.W.11. Ivanoff and Abdouloff Compt. rend. 1933 196,491; Ivanoff and Ibdulov Ann. univ. Sofia 11 FucuZte'phys.-muth., 1934 30 53. Ivanoff and Spassoff BUZZ.SOC.chim. France 1932 51 619. Wotiz Hollingsworth and Dessy J. Amer. Chem. SOC.,1955 77 103. Podder Smalley and Hollingsworth J.Org. Chem. 1963 28 1435. Gielen and Nasielski J. OrgunometuZZic Chem. 1963 1 173. Dessy Reynolds and Kim J. Amer. Chem. SOC.,1959 81,2683; Dessy and Kim ibid. 1%0,82 686. Some Compounds Isostructural with Manganese Carbonyl T. G. DUNNE, By F. A. COTTON B. F. G. JOHNSON,and J. S. WOOD* THE factors which are decisive in determining the measurements at room temperature show that it is structures of binuclear metal carbonyls are not well almost completely dissociated to Fe(CO),I under understood. The results of two investigations in- these conditions. These vapours show extremely tended to provide data bearing on these questions intense absorption rising into the ultraviolet are reported here. region. The new compound Fe,(CO),I has been prepared Sacco and Freni2 have reported that by reaction of iodine with Fe,(CO),,.It is a white Co(CNCH,),(C104) can be obtained in an unstable solid melting at -5 O to a red liquid and it dissolves paramagnetic blue form and a red stable diamag- in various organic media to give pale red solutions. netic form. Since MnO and CoI1 are isoelectronic it The compound in solution is diamagnetic within seemed plausible that the red form might be a dimer experimental error and exhibits only two equally isostructural with Mn,(CO),,. A three-dimensional intense CO stretching bands at 2020 and 2000 cm.-l. X-ray diffraction study has shown this to be the Thus the structure is almost certainly analogous to case. Space group P2,22 (No. 18); a = 13-17 that1 of Mn,(CO)lo [with which Fe,(CO),I is iso- b = 12.49 c = 12.47 A z = 4.The [(CH3NC),Co- electronic] the two axial CO groups being replaced Co(CNCH,),I2+ ions have D4d symmetry within the by iodine atoms. The red colour is attributed to uncertainties. The Co-Co distance is 2.74 f0.01 8 minute amounts of the Fe(CO)41 monomer. The [cf. Mn-Mn in Mn,(CO)lo of 2.93 A] and the Co-C compound is rather volatile and vapour-density distances are in the range 1.82-1-92 A. Very large * Department of Chemistry Massachusetts Institute of Technology Cambridge Mass. U.S.A. Dahl and Rundle Actu Cryst. 1963 16 419. 'L Sacco and Freni Guzzettu 1959 89 1800. temperature factors are found for the perchlorate oxygen atoms indicative of hindered rotation or librational motions with very high amplitudes.Because of this the standard deviations in the dimensions of the cations are somewhat higher than might have been expected and no definite information PROCEEDINGS was obtained as to the linearity of Co-C-N-C chains. These results support the view that the Mn,(CO), type of structure is generally preferred for molecules isoelectronic with Mn,(CO),,. (Received April lst 1964.) Triplet-triplet Polarisation Measurements in Mixed Crystals By D. P. CRAIG and GADFISCHER* EARLY studies of triplet-triplet absorption in complex molecules1y2 were made in rigid-glass matrices in which the molecular orientations are random. Knowledge of the polarisation of the transitions depends on an indirect method of observing polarisa- tions in molecules which on account of favourable orientation are “photoselected” in the glass medium on excitation by polarised light and internally con- verted into the lowest triplet ~tate.39~ Triplet-triplet measurements in mixed crystals in benzophenone are inconclusive as to polarisation because of incomplete crystal structure data.5 We have now observed polarised triplet-triplet absorption in several mixed crystals including in most detail that of naphthalene in durene leading to unambiguous directional assignments of the transi- tions.The method depends on knowledge of the orientation of the guest molecules in the host crystal. For naphthalene-durene the molecular arrangement is known from both optical and spin-resonance studies6 to be that the molecular planes of host and guest are parallel with the long naphthalene axis parallel to the long durene axis.The naphthalene molecules are internally converted into their lowest triplet states after excitation with intense unpolarised light from a mercury or xenon arc and the polarised triplet-triplet spectrum measured in the way familiar in optical transitions of pure crystals. A typical set of spectra is reproduced showing as well as the triplet spectra traces of the well-known polarised fluorescence spectrum of the first singlet system‘ of naphthalene made simultaneously as an internal check on the triplet polarisation. We find the polarisation of the first triplet-triplet system to be the same as that of the first singlet system thus con- firming it as parallel to the long molecular axis as found by photo~election.~ Its assignment is thus 3B3gt3Blu.The absorption wavelengths agree closely with those in the rigid glass, characteristic of the naphthalene triplet spectrum. 29,000 30.000 3l.000 ni 6-axis &oxis I 1 1 * William Ramsay and Ralph Forster Laboratories University College London. McClure J. Chem. Phys. 1951 19 670. Craig and Ross J. 1954 1589. Albrecht J. Mol. Spectroscopy 1961 6 58. El-Sayed and Pavlopoulos J. Chem. Phys. 1963 39 834. Hochstrasser and Lower J. Chem. Phys. 1964 40,1041. Hutchison and Mangum J. Chem. Phys. 1961 34 908. ’McClure J. Chem. Phys. 1956 24 1. JUNE1963 177 s 1 ! I I I I I :o:o:o..0:o010 :o:o:o' 0:o:0:o 0.0.0.0.0.0.0.00.0.0.0 0.8 x 0.9 in TiO 1.0 1.1 1.2 1.3 -s 1/30 2/30S l0Oj Cr S 1/3cr 2hOS cr,s ul0 V c v .-?,O 60-40- .o*o,o-0 0 0'0.0.0.0 ooooooo.nn n cn .)r..- a 20- 0 0- b 08 0.9 1.0 1.1 1.2 1.3 Site filling 'in a random TiO structure FIG.1. (a) Concentrations of titanium and oxygen FIG. 2. Notional ordering of cation vacancies in vacancies in TiO. (6) Statistics of ideal and incomplete alternate sheets of the homologous chromium sulphides co-ordina tion en vironmen ts-oxygen about titanium titanium about titanium and titanium about oxygen- for randomly distributed vacancies in the Ti0 phase. TABLE 1. Intermediate phases formed by ordering of defects NiAs-CdI Completely random Non-stoicheiometric layer structures M atoms or S vacancies segregated in alternate Non-stoicheiometric sheets ;random distribution within layers M atoms or S vacancies segregated in alternate Discrete intermediate sheets; ordered array in incomplete sheets phases Fluorite oxides Completely random 0 vacancies Non-stoicheiometric or-CeO,.MO2-2 Vacancies ordered in linear arrays of MO (high temp.) co-ordination polyhedra. Linear arrays randomly disposed Non-stoicheiometric Linear arrays regularly disposed Discrete intermediate MnO2 n-2 phases M = Ce,Pr Tb Fluorite oxides Completely random interstitial sites occupied Non-stoicheiometric MO2+x Ordered arrangement of inters ti tials Discrete intermed iate phases PROCEEDINGS 1200 I I 0" h FIG.3.Probable equilibria between narrow-range phases of the homologous series Pr,n02n-2and broad-range non-stoicheiomet r ic phases. [Reproduced with the authors' permissioil (ref. 81. b FIG.4. @-Tungsten oxide WzoOBs 1 slabs of corner-sharing [WOd Octahedra linked by colunzns of'edge-sharing octahedra along lines of shear. (Reproduced with permission from A. Magneli Arkiv Kemi 1950 1 513.) JUNE 1964 179 Co-operative of 0 vacancies rearrangement of 0 vacancies in arrays or in blocks of dislocation lines perfect structure Reduction of higher oxides; formation of Magneli shear quasi-ReO type structure phases oxides with M:O < 1 :3 FIG.5a Schematic sequence for formation of Magneli shear structures.FIG.5b In Mo,,O,, strips of the MOO layer structure are spliced together by displacement (along the dotted line) and tilt so that edges are shared between [MOO,]octahedra in adjacent rows along the zigzag line of shear. (Reproduced with permission from Kihlborg Adv. Chem. 1963 No. 39 p. 37.) PROCEEDINGS Fe 0 Fe 0 Fe 0 Fe 0 Fe 0 Fe 0 Fe* 0 0 Fe 0 Fe 0 0 Fe 0 Fe* 0 Fe 0 Fe* 0 Fe 0 Fe 0 Fe 0 (a) Fe 0 Fe 0 Fe* 0 Fe 0 Fe 0 Fe* 0 Fe 0 0 Fe 0 Fe 0 Fe 0 Fe 0 Fe* 0 Fe Fe 0 Fe* 0 Fe I 0 Fe /-0 0 I Fe 0'"' Fe* \\o Fe*0 Fe 0 Fe 0 (* \ 0 Fe 0 Fe',O Fe OJ Fe 0 Fe* 0 Fe\ 00' 0 Fe 'Fe* 0 Fe 0 Fe 0 0 Fe 0 Fe 0 Fe* 0 Fe 0 Fe 0 Fe 0 Fe 0 0 Fe 0 Fe Fe 0 Fe 0 0 Fe 0 0 Fe Fe Fe 0 0 Fe 0 0 Fe Fe Fe 0 0 Fe 0 0 Fe 0 (C) Fe Fe 0 Fe 0 0 Fe 0 Fe Fe 0 Fe r) 0 Fe 0 0 Fe Fe Fe 0 0 Fe 0 0 Fe FIG.6.Fe,-,O. (a) The structure as hitherto supposed (b)Roth's model (1960) with defect complexes con-stituting micro-domains from the spinel structure of Fe,O (c). JUNE 1964 181 "Interstitial" oxygen in UO~+~ Formation of [OlA 1 [Ol 01defect complex YYYT FIG.7. Defect complexes in U02+z. s cat./%eg. OI :onf igurational entropy Ta2H 3000 -360' d+ 8 -3 40' -320' -300°K PI 10 20 30 40 Atom 70 H fully fully ordered random FIG.8. Phase equilibria and configurational entropy-temperature relations in the Ta-H system. PROCEEDINGS -+-I-0 0 0 A-w w-a -0 0 0 0 -++-Basic s_tructure unit 0 ,I -.*s 0 * 0 0 c. :0 0 0 '4 ' lo lo I ::.: Microdomains in jj,-Ta,H FIG.9. Microdomains in &Ta,H. based on short-range ordering of basic structural units. In each basic structural unit two of the four hydrogen positions are occupied at random. TABLE 2. Interaction energies and homogeneity ranges of hydrides H ydride E Max. % of Vacant kcal./mole Sites Uranium hydride 4.3 0.8 at 450"c. 3.1 at 650"c. Zirconium hydride 1.5 9 at 500-600"~. (Tetrag.) Hafnium hydride 1.0 15 at 250"c. to (Te tr ag .) 8 at 320"c. Palladium hydride 0.35 30 at 160"c. to 50 at 290"c. -.*....... .-.-.-.-MO M O M O M MO M 0 M 0 M 0 :Mi0 M**. M 07 . I OM0 0 OM OM M 0 :O 'M 0 M'.. M; :\ I MnO Microdomain MOMOMOMO O M 0 M M 0 M 0 :M d\M 0 i**.a/ 01 OMOMOMOMOMOM 0 M 0 M 0 i0 M O'.M 0 M'* .*. ...-... \ 1' MO OMOMOMOMO M 0 ,*O 0 M 0 M O',M 0;; licrodomain OMOMO OM MOM 0 M 0 :M 0 M 0 M M MOM MOMOMOMO M 0 :o M 0 M /@M\\O MO$ OMOM M OM M 0 :M *.. M MOMOMOMOMO M MO -. . ..... Microdomain 8 OMOM OMOMOM OM M 0 IM 0 M 0 M.r,! MOMOMOM OM0 OM0 10 OMOM01 I I O M 0 OMOMOMOM OMOMO\MOMO OM) I5 Rancforii defects in each sub-lattice. 15 o< Defects in each sub-lattice ordered in niicro-ckomaiiis FIG.10. Schematic representation of random structures (a) and of microdomains (b) in the highly defective Ti0 and VO phases. In M,O microdomains oxygen vacancies may probably be clustered in discs in alternate (11 1) sheets. In M,O microdomains there is a superstructure ordering of an M,.,,O structure.PROCEEDINGS FIG.1. Autoradiograph of iron oxide specimen. (Magnificalion x 9.) FIG.2. Photograph of iron oxide specimen. (Magnification x 9.) JUNE 1964 185 Selective Adsorption of Promethium on to Surfaces from Aqueous Solution C. E. MELLISH and J. A. PAW* THE adsorption of radioactive ions from aqueous solutions on to surfaces has been studied both as a possible metallographic tool1 and as an important phenomenon in the operation of nuclear reactors containing water-cooling circuits.2 The interpreta- tion of the data obtained which have come from experiments with univalent and bivalent ions has been in terms of the ion-exchange properties of surfaces and of the action of local electrochemical cells on surfaces.Experiments with multivalent ions such as yttrium and the rare earths have been prevented by the formation in solutions of salts of such elements of “radiocolloids.” These are highly radioactive aggregates of colloidal size which settle out on surfaces to give irregular distributions of radio-activity on the surfaces. Autoradiographs of such surfaces show a random pattern of radioactive spots. For this reason it has been stated for instance that promethium-147 is not suitable for adsorption experiments on metal surfaces. We have prepared solutions of promethium-147 as chloride by the methods reported* for making particle-free solutions of yttrium ;such solutions are made by dialysis of appropriate chemicals into ultra- pure water and handled always in polystyrene vessels.In this way we have been able to obtain uniform adsorption on to surfaces of spectroscopi- cally standardised metals and selective adsorption on to constituents of multicomponent surfaces with- out interference from radioactive particulate matter. An example of selective adsorption is shown in Fig. 1 which is an autoradiograph of a piece of iron oxide taken after adsorption of 14’Pm from a solution of promethium chloride at pH 3.8; Fig. 2 is a photo- graph of the specimen. The specimen of iron oxide was made by completely oxidising mild steel at 800”c in air. The radioactivity which shows as white areas on Fig. 1 follows exactly the pattern of the grain boundaries in the iron oxide.It is probably adsorbed on to some impurity in the iron oxide as the effect is not obtained with iron oxide made from pure iron. (Figs. 1 and 2 appear on p. 184.) Selective adsorption of this kind has also been obtained under alkaline conditions ; adsorption of promethium at pH 11.7 on to the surface of a polished specimen of gun metal (Sn 5 % Pb 5 % Zn 5% Ni 1 % Cu balance) follows in detail the den- dritic structure revealed by etching. In these experi- ments it has been checked that the pattern shown by the autoradiograph is due to radioactivity and not chemical reaction between the film and the specimen. The mechanism by which this adsorption takes place is not at present known; neither is the chemical form of promethium in these solutions.The interest of these results lies in the two discoveries that ter- valent elements can be used for adsorption and auto- radiographic studies without interference from parti- culate matter in solution and that their adsorption from solution can in certain circumstances be highly selective. (Received,March 26th 1964.) * Wantage Research Laboratory (A.E.R.E.) Wantage Berks. Simnad “Properties of Metallic Surfaces,” Inst. of Metals 1963 p. 23. Proc. Conf. on Transport of Materials in Pressurised-Water Systems Atomic Energy of Canada Ltd. Report A.E.C.L. 1265 June 1961. Junghahn Werkstofe Korrosion 1962 13 474. * Mellish Payne and Worrall Radiuchem. Acta in the Press. Some Reactions of Tetracyanoethylene Oxide By P.BROWNand R. C. COOK SON^ THE ability of tetracyanoethylene oxide (TCNEO) to add to ethylene acetylene and benzene has been reported,l and kinetic evidence2 indicates initial rate- determining promotion to an activated species (TCNEO)*. Stereospecific products formed with cisltrans-isomeric olefins a relative reactivity sequence of olefins > acetylenes > aromatics and the small effect on rate of addition attendant on solvent change2 are strongly reminiscent of 2 + 3 -+ 5 1,3-dipolar cycloaddition reaction^,^ with (TCNEO)* depicted as a 1,3-dipole with internal octet stabilisation. We report here the addition of TCNEO to a wide variety of dipolarophiles (mainly di- and poly-enes) in all cases except one (anthracene) a tetracyano- tetrahydrofuran adduct resulting from 1,3-addition of the oxide across one unsaturated linkage of the dipolarophile.The monoadduct of benzene1 will add a second molecule of TCNEO to give the cis-trans- cis bisadduct m.p. 226-227” (decornp.). t Department of Chemistry The University Southampton. Linn Webster and Benson J. Amer. Chem. Suc. 1963 85 2032. Benson and Linn Amer. Chem. Sac. 146th Meeting Abstracts 1964 p. 13c. Huisgen Angew. Chem. Internat. Edn. 1963 2 565. D ipolarophile Representative adducts* But-1-ene Cyclopentene C yclopentadiene (I and 11) Indene Furan (111) Thiophen (IV) Cycloheptatriene 01 Cyclo-octatetraene (W Bicyclo [2,2,1 Iheptene Bicyclo [2,2,1 Iheptadiene (Vm Bicyclo [2,2,2]octadiene Benzene (ref.1) Durene (VIII) Naphthalene (Ix) Anthracene (XI Phenan t hr ene (XI) * Where none is given the adduct was shown to have the expected structure. Nc$5+) CN NC CN (I ,X=CH,) (m x= 0) (IV x= s) CN 0JCN YCN CN However many of the reactions also yielded oxidation products of the dipolarophile and either tetracyanoethylene (TCNE) or the TCNE-dipolaro- phile adduct. Thus cyclopentadiene cyclohepta- triene and bicyclo [2,2,1 lheptadiene all gave TCNE adducts; durene gave duroquinone and TCNE ; indene gave TCNE adduct and 2-indanone ; whilst anthracene gave TCNE adduct anthrone anthra- quinone and bianthronyl. All new compounds have satisfactory elemental analyses and the structures assigned are supported by i.r.U.V. and n.m.r. spectra and by their chemical transformations. For example adduct (X) has n.m.r. resonances at T 2.33 (8-proton multiplet) and at Huisgen Angew. Chem. Innternat. Edn. 1963 2 633. PROCEEDINGS T 3-55and 4-42 (each 1-proton singlet) in acetone with tetramethylsilane as internal reference. The U.V. absorption of (X) in ethanol is almost identical with that of the known TCNE-anthracene adduct. Adduct (IX) on treatment with aqueous ethanol containing catalytic quantities of acid or base affords 2-dicyanomethylnaphthalene which in turn is converted into 2-naphthoic acid by hot dilute nitric acid. These results are consistent with the operation of two distinct mechanisms (1) Preliminary opening of TCNEO followed by 1,3-dipolar addition of the type 2 + 3 --f 5 involving delocalisation of six electrons in a cyclic transition state4 (A).(2) Electrophilic attack of the electron-deficient oxygen atom of TCNEO on the n-system of the di-polarophile to give an intermediate (XII) which can yozz NC CN (XI I readily collapse to TCNE and the oxidation product of the dipolarophile (B). With triphenylphosphine as nucleophile TCNE was generated very rapidly and almost quantitatively from TCNEO at room temperature being trapped and characterised as the cyclopentadiene adduct. In the case of anthracene any 1,3-dipolar cycloaddition is energetically unfavoured. The formation of an intermediate of the type (XII) however will be rela- tively easy due to the high n-basic character of anthracene and stabilisation of the subsequent car- bonium ion.Ring closure of (XII) as an alternative to the loss of TCNE leads to the unsymmetric adduct (X). (Received May 11thy 1964.) JUNE 1964 187 ~ ~~ The Structure of an Organocobalt Intermediary in the Synthesis of ortho-Substituted t-Butylbenzenes By 0.S. MILLSand G. ROBINSON* THE preparation of an ortho-substituted t-butyl- benzene 1,2,4-tri-t-butylbenzene, was first achieved by the bromine decomposition of an organometallic complex Co,(CO),(C,HBut), for which a structure involving a heterocyclic seven-membered ring thought to be non-planar had been suggested.l We have determined the structure of a related complex Co2(CO),(C,HBut),(C,H,) from which by similar decomposition 1,2-di-t-butylbenzene has been obtained.This complex occurs as violet crystals with crystal data a = 16.75 f0-02 b = 12-52 f0.02 c = 17.69 f0.02 A 18 = 90” 50’ f 1’; I/ = 3710 f15A3;dabs = 1.49 g. Id.-’ dcalc = 1.50 g. ml.-l; 2 = 8. Systematic absences are con- sistent with the spacegroup 12/a. The analysis is based on the intensities of 1734 reflexions which were measured by scintillation counter and Mo-Ka radia-tion from a stationary crystal by the stationary counter method. No correction for absorption has been applied and after refinement of the atomic positions and isotropic temperature parameters of all atoms (excluding hydrogen) the R factor is 8-7%. The molecular structure (see Figure) contains a non-crystallographic axis of two-fold symmetry (Ca and hence the two cobalt atoms are equivalent.Al- though it is possible to consider the six-carbon chain formed through fusion of the three alkyne groups as a bridging group which supplies eight electrons a better description in our view is that of a hetero- cyclic eight-ring twisted into the shape of a figure “8.” It should be noted that the central carbon- carbon bond of this chain is not perpendicular to the cobalt-cobalt bond and therefore is not a bridging group in the sense that tolan bridges the metal atoms in CO,(CO),C,P~,~ and in Ni2(C5H5),C2Ph2.* The angle is only 28”so that a better description of the arrangement is a “flyover” rather than a “bridge.” The nietakarbon distances show that each cobalt atom is involved in bonding with four of these six carbon atoms.We interpret this arrangement in terms of a diallyl structure. In accord with this view each allyl moiety is regarded as trisubstituted as well as n-bonded viz. n-l-syn-t-butyl-anti-a-cobalt-3-flyover is 1.50 A. The average Co-C(ally1) separa- tion is 2-04 A and Co-C(alky1) is 2.01 A. It is inevitable since each half of the ring acts bifunction- ally that some distortion from idealised geometry occurs. As a result the angles within the allyl groups are contracted from the theoretical 120” by up to seven degrees and some of the substituent atoms are forced out of the plane of the allyl carbons. The planes of the two allyl groups are mutually per- pendicular.b The terminal carbon atoms of the trialkyne residue nearly coplanar with the cobalt atoms and substituted with t-butyl groups are 3.2 A apart. Thus the t-butyl groups only occupy adjacent positions as a result of ring closure between these carbon atoms. This ring closure may be facilitated by a distortion from the observed planar arrangement by addition of bromine and cleavage of the Co-Co bond. The formation of 1,2,4-tri-t-butylbenzene from the tri-t- butyl-acetylene complex can be readily envisaged. The proton magnetic resonance spectrum of the disubstituted complex is very simple and shows three bands at 1.5 (sharp methyls) 3.8 and 5.7 p.p.m. (doublets) relative to SiMe,. The trisubstituted com- plex shows an additional absorption at 1.1 ppm.(methyl) and a reduced absorption at 3.8 p.p.m. together with further splitting especially of the band anti-allylallyl. The average C-C distance within each at 5.7 p.p.m. The absorption at 5.7 p.p.m. is thus allyl group is 1.41 A whereas the C-C separation associated with the H atom. between the two allyl groups the central span of the (Received April lst 1964.) * Department of Chemistry University of Manchester Manchester 13. Kriierke Hoogzand and Hubel Chem. Ber. 1961,94,2817. Hoogzand and Hubel personal communication. Sternberg Greenfield Friedel Wotiz Markby and Wender J. Amer. Chem. SOC.,1954 76 1457; Greenfield Sternberg Friedel Wotiz Markby and Wender ibid. 1956 78 120; Sly ibid. 1959 81 18.Tilney-Bassett and Mills J. Ameu. Chem. Soc. 1959 81 4757; Tilney-Bassett J. 1961 577; Mills and Shaw unpublished results. PROCEEDINGS Direct Determination of the Molecular Structure of Trichodermin By SIXTEN and Bo NILSSON* ABRAHAMSSON work1 has related the antifungal meta- graphic programmes developed at this In~titute.~ CHEMICAL All bolite trichodermin (C1,H&4) to trichothecin. The Fourier series have been scanned for peaks in the formula based on the suggested trichothecin struc- ture2 was however in conflict with certain reactions. In order to establish the molecular structure we have performed a single-crystal X-ray analysis on tricho- dermol p-bromobenzoate (C22H2504Br). Good single crystals of the derivative were kindly provided by Dr.W. 0. Godtfredsen. The crystals were ortho- '2 rhombic with a = 10.69 b = 9.24 and c = 20.70 A; space group is P212121. The structure was solved from bromine-phased three-dimensional electron-density series without use of any chemical informa- tion. The parameters of the molecule are being re- fined by least-squares treatment. The R-value is at present 14 % for the 900 observed reflexions. The result of the X-ray analysis is illustrated in the Figure. The molecular formula of trichodermol should accordingly be (I). The calculations have been performed on a Saab D21 computer using the integrated set of crystallo- computer making possible a short cycle time. In the present case one round of structure factors and electron-density calculation with peak scanning of the series took 20 minutes.0Br oc 00 Scale drawing of the p-bromobenzoate of tri-chodermol. The absolute configuration has not been determined. (Received March 25th 1964.) *Crystallography Group Institute of Medical Biochemistry University of Goteborg Sweden. Godtfredsen and Vangedal following communication. Fishman Jones Lowe and Whiting J. 1960 3948. Abrahamsson Aleby Larsson Nilsson Selin and Westerdahl to be published. Trichodermin a New Antibiotic Related to Trichothecin By W. 0. GODTFREDSEN and S. VANGEDAL* A NEW antifungal metabolite trichodermin C1,H,,O, m.p. 45-46' [a]2,0-ll" has recently been isolated from a strain of Tvich0derma.l Alkaline hydrolysis of trichodermin gave a sesquiterpene alcohol trichodermo1,t C15H2203 m.p.11 8" [a]g -34" which on acetylation was reconverted into trichodermin in high yield. The chemical and spectro- scopic properties of trichodermin3 suggested a rela- tionship to trichothecin4 to which structure (Ia) has been and this was confirmed by the fact that trichodermin on vigorous oxidation gave tri- chothecolon acetate (Ib) albeit in poor yield. Trichodermin and trichodermol should therefore be represented by formula (Ic) and (Id) respectively but certain reactions are incompatible with these structures. Thus reduction of trichodermol with lithium aluminium hydride gave a diol C1,H2,0, m.p. 148" which according to the n.m.r. spectrum contains four quaternary methyl groups and which on acetylation and oxidation gave a monoacetate C,,H2604 m.p.1 17-1 18" and a monoketone C15H2203 m.p. 175-1 76" respectively. This indi- * Leo Pharmaceutical Products Ballerup Denmark. t The m.p. optical rotation and i.r.-spectrum of roridin C a metabolite of Myrothecium roridum,2 show that this compound is identical with trichodermol. Belg. Patent 642,118. Harri Leoffler Sigg Stahelin Stoll Tamm and Wiesinger Helv. Chim.Acta 1962 45 839. Godtfredsen and Vangedal in preparation. Freeman Gill and Waring J. 1959 1105. Fishman Jones Lowe and Whiting J. 1960 3948. JUNE 1964 189 cates the presence of a tertiary hydroxyl group in the diol and suggests that trichodermin contains the grouping (9. CH -0 (x> y X-Ray crystallographic analysis of trichodermol p-bromobenzoate m.p.161-163" has now con- firmed this and rigorously established (apart from absolute configuration) the structure as (IIa).s Tri- choderniin and trichodermol are consequently represented by (IIb) and (IIc) respectively. The conversion of (IIb) into trichothecolon acetate implies that the structure of the latter should be revised into (IId) and that of trichothecin into (Ire). The identity between roridin C and trichodermol raises the question of the correctness of formula (111) recently suggested' for the sesquiterpene ver- rucarol obtained by alkaline hydrolysis of the anti- biotic verrucarin A, which is formed by the same micro-organism as roridin C. The argument against the presence of an epoxide ring in verrucarol (stability of a tetrol obtained on acid hydrolysis of verrucarol to periodate and lead tetra-acetate) is invalidated by the fact that the triol C1,H,,O, m.p.144" prepared analogously from trichodermol is also resistant to these oxidising agents. In view of this the data reported by Gutzwiller and Tamm' are consistent with formula (IV). fie0 R' (I)a R'=O,R%socrotonyl b;R'=O R2=0Ac c ;R'=~I-(&OAC d ,R'=2H,R2= H (Received March 25th 1964.) Abrahamsson and Nilsson preceding communication. Gutzwiller and Tamm Helv. Chim. Acta 1963 46 1786. Halogeno-anions of Tervalent Rhenium By J. E. FERGUSSON and B. H. ROBINSON* CHEMICAL and crystal structure studies of the Re,Cl,;-and Re,ClIl2- anions have now been extended to bromo-compounds.In addition to iso- lating Re,Br,;- and Re,Brll2- anions we have pre- pared Re,Br,,- with a number of cations such as Ph3PH+,Ph3 Benzyl O+ which are for the chloro- anions associated with Re,Cl,;- and Re,Cll12-respectively. In every case the compounds were pre- pared by adding the appropriate cation as the bromide salt to solutions of rhrenium tribromide in hydrobromic acid or hydrobromic acid-ethanol. It is suggested that solutions of these compounds probably contain as the stable entity Re,X, which can under various conditions of ionic size and crystal packing be forced to form anions in the solid ranging from Re,X,:- to Re,X1,,-. The ionic-size effect is supported by the common occurrence of the Re,Br,,- anion whereas for the chlorides the most common species is Re,Cl,?-.The larger bromine atom is therefore favouring the formation of the Re,Brlo- anion. Treatment of the compound [Ph,PH]Re,Br, with more than 3 moles of pyridine produces the complex Re,Br 9(pyridine),. However if less pyridine is used evidence is obtained for a compound [Ph,PH]Re,Br,,(pyridine),. This type of system is being further investigated to obtain additional in- formation as to the reasons for the formation of the various anions and their constitution in solution. The anion Re,Brl0- may be for example represented more accurately as Re,Br,o(solvent)2~. Preliminary X-ray diffraction investigation of CsRe3BrlO3 shows that the crystals are orthorhombic with a = 9-52 b = 16-08,c = 13.72A and probable space group Prima (0;;).The observed density is 4.8g. cm.-l showing that there are 4 formula units per cell. This suggests that here as in CsReCI, the rhenium atoms are in bonded groups of three with one of them lying on a mirror plane. This evidence substantiates the conclusion deduced from the close similarity of the visible absorption spectra of the chlorides and bromides. The compounds were characterised by analysis molecular weight and conductivity measurements. For example [Ph,PH]Re,Br, is a 1-1 electrolyte in nitrobenzene and has half the calculated molecular weight. (Received March 9th 1964.) * Chemistry Department University of Canterbury Christchurch New Zealand. l Robinson Fergusson and Penfold Proc.Chem. Soc. 1963 116; Bertrand Cotton and Dollase Inorg. Chem., 1963 2 1166. Fergusson Penfold and Robinson Nature 1964 201 181. Penfold and Elder personal communication. PROCEEDINGS Evidence for The Saddle Conformation of Cyclo-octane By JOHANNES ILONALASZLO RULAND* DALE and WILHELM ITis generally assumedl that the basic conformation of cyclo-octane is that of a symmetrical or stretched crown (Ia) although no really compelling evidence has so far been produced. Some arguments for a different conformation (Ib) whose ideal carbon skeleton disregarding hydrogen interactions would follow the diamond lattice has already been given by one of us. We have now collected some additional evidence for this latter conformation for which we suggest the descriptive name “saddle”.The first evidence can be extracted from a dipole- moment study by Henniger Wapenaar and Havinga3 of cis-and trans- 1,2-dihalogenocyclo- hexanes -heptanes and -octanes. The 6- and 7- membered trans-isomers show appreciably lower moments than the cis-isomers which can be ascribed to the presence of some of the trans-diaxial form (zero moment) together with the trans-diequatorial form (moment as large as the cis-isomer); in con- trast the 8-membered trans-derivatives show essenti- ally no reduction. This is easily understandable on the basis of the saddle conformation (Ib) as the “diaxial” form would necessarily have one bulky halogen atom pointing inward in collision with a hydrogen atom from the opposite side of the ring; the “diequatorial” form remains therefore the only possibility.Also the CBr-stretching infrared band4 for the diaxial form at 663 and 655 cm.-l in the 6- and 7-membered trans- 1 ,Zdibromides is absent in the 8-membered ring and similarly the axial CC1-stretching band,* at 692 and 687 cm.-l in the 4.\ \ ,_-..* 6- and 7-membered trans-l,2-dichlorides is lacking in the 8-membered derivative. It is not easy to see how these observations can be explained on the basis ./’...-\ ** -7 h 1 of the crown conformation. 23-\ ’*** One of the strongest arguments for the stretched- \ k**\-~--,.--.._.-*-crown conformation (Ia) has been that three methylene scissoring bands are observed in the infra- red spectrum at 1450 1470 and 1477 cm.-l.These have been interpreted5 as representing three different types of methylene which would exclude the sym- S=L x metrical crown having one type and conformations Wide-angle X-ray scatterinp of liquid cyclo-octane. having two types such as the saddle (Ib). As a Full line observed intensity; broken line calculated model substance we have now studied bicyclo [3,3,1]- for the saddle (Ib); dotted line for the symmetrical nonane whose unique conformation has recently crown. * Union Carbide European Research Associates Brussels 18 Belgium. See for example Allinger Jindal and DaRooge J. Amer. Chem. SOC. 1962,27,4290 for a complete list of references. Dale J. Chem. SOC. 1963 93. Henniger Wapenaar and Havinga Rec.Trav. chim. 1962 81 1053. Cf. Klaeboe Lothe and Lunde Acta Chem. Scand. 1957 11 1677. Chiurdoglu Doehard and Tursch Bull. SOC. chim. France 1960 1322. 13 Pumphrey and Robinson I.U.P.A.C. Congress London 1963; cf. Abstracts A3-19 p. 135; Brown Eglinton Martin Parker and Sim Proc. Chem. SOC.,1964 57. Meerwein and Schiirmann Annalen 1913 398 196; Meerwein Kiel Klosgen and Schoch J. prakt. Chem. 1922, 104 161. JUNE1964 191 ~~ ~ ~~~~~ ~~ ~~~~ may well represent just one type of methylene. Thus solutions of cyclo-octane-adamantane were also in- it can at least be concluded that the infrared spectrum vestigated by X-rays (powder diagram) at room cannot be used as an argument against the saddle temperature and it was found that the cubic unit conformation.cell of adamantane was preserved even with respect Space-filling molecular models show the cyclo- octane saddle to have an almost spherical shape resembling closely in form and dimensions both bicyclo[3,3,1 Inonane (11) derived formally by inser- tion of one methylene bridge and adamantane (111) derived by insertion of two. It would therefore be expected* that cyclo-octane may form solid solutions with these compounds considering in particular that all three of them form plastic crystalsg with “free rotation” in an expanded “tolerant” crystal lattice above their solid transition points. On the other hand the crown-conformation (Ia) is a rather flat and wide disc and would not be expected to give solid solutions with (11) and (111).As an example of a “non-globular” model substance cyclodecane with the established conformation (IV)l0 was used. The observed phase diagrams show that (11) and (111) do indeed form solid solutions with cyclo-octane where- as with cyclo-decane they give eutectics. It was further verified that (11) and (111) give solid solutions and that cyclo-octane and cyclo-decane do not. Solid to cell parameter (9.45 A) for mixtures containing 22 and 48 % of cyclo-octane. Thus cyclo-octane mole- cules can replace adamantane molecules without any distortion of the crystal lattice. Finally liquid cyclo-octane was examined at room temperature by wide-angle X-ray scattering using Cu-radiation and Ross filters for monochromatisa- tion. The observed intensity distribution is compared in the Figure with molecular scattering functions cal- culated for the symmetrical crown conformation and for the idealised saddle (Ib) disregarding the “flatten- ing” of the molecule which undoubtedly occurs because of the two transannular repulsions.Such a comparison seems permissible since the wide-angle scattering of non-polar organic liquids is pre-dominantly determined by intramolecular inter- ference functions. No thermal motion was taken into account but it is quite clear that the experimental curve fits much better with the scattering function of the saddle than with that of the symmetrical crown. (Received March 9th 1964.) Kitaigorodsky Kristallografiia 1957 2 4. 51 Proc. Symp. on Plastic CrystaIs and Rotation in the Solid State Oxford 1960; J.Phys. and Chem. Solids 1961,18,l. lo Huber-Buser Dunitz and Venkatesan Proc. Chem. SOC.,1961 463. Reactivities of n-z* and Charge-transfer States in the Photo-pinacolisation of Ketones By G. PORTER and P. SUPPAN* THE well-known photochemical pinacolisation of benzophenone and other ketones and aldehydes in solution involves hydrogen abstraction from the solvent as the first step Ph&O + RH = Ph2COH + R-Benzophenone itself abstracts hydrogen with unit quantum yield from many organic solvents including alcohols and paraffins.l Studies of a variety of sub-stituted benzophenones in isopropyl alcohol have revealed that certain substituents such as amino- and hydroxy-groups have a profound effect on the re- activity and that the pinacolisation reaction is almost completely ~uppressed.~*~ The explanation of this effect which is similar to the effect of substituents on the photochemical reduction of quinones? has been given in terms of the relative energies and reactivities of nlr* and T-T* states.n-n* excitation results in an increase of positive charge at the oxygen atom of the carbonyl group whereas T-T* excitation which as * Department of Chemistry The University Sheffield 10. Beckett and Porter Trans. Faraday SOC.,1963 59 2038. Pitts Johnson and Kuwana Symp. Photochemical Processes Duke University 1962 195. Beckett and Porter Trans. Furaday SOC.,1963 59 2051. Bridge and Porter Proc. Roy SOC.,1958 A 244,259. PROCEEDINGS we shall show elsewhere is better described in these molecules as an intramolecular charge transfer from the electron-donating group to the carbonyl group results in an increased negative charge on the car- bonyl.Hydrogen and electron abstraction by a car-bonyl group will therefore be strongly favoured by n-n* excitation and inhibited by charge-transfer excitation. It is known that the reaction occurs exclu- sively from the lowest triplet level and that in benzo- phenone this is of n-n* type.5 The absence of reaction in benzophenone containing electron-donating sub- stituents is then explained if the effect of the substitu- ent is to lower the energy of the charge-transfer triplet states so as to bring them below the n-n* triplet.Spectroscopic and photochemical studies of these compounds which we have just completed fully substantiate these views. These observations provide comparisons of re-activity of the excited states of different molecules and it would clearly be more satisfactory if a com- parison could be made between the two different electronic states of the same molecule. This can be effected in a rather striking manner by utilisation of the different solvent shifts on the two electronic states. We have found that p-aminobenzophenone and p-hydroxybenzophenone although photochem- ically (3650 A) quite unreactive in isopropyl alcohol undergo a pinacolisation reaction in pure cyclo- hexane with quantum yields of 20% and loo% respectively. That any species should abstract hydro- gen from a paraffin and not from isopropyl alcohol is quite the opposite of chemical expectation and we must conclude that the excited states are different in the two solvents.This is clearly seen to be the case when the absorption and phosphorescence spectra are investigated and is in accordance with the known effect of hydroxylic solvents on n-n* transitions.6 In paraffinic solvents the reactive n-n* triplet state is the lowest excited level but in isopropyl alcohol not only is the n-n* promotion energy raised but the energy of the charge-transfer state is considerably reduced by solvation so that this unreactive state becomes the one of lowest energy. The completely different chem- ical reactivity in the two solvents illustrates the very specific relationship between electron distribution and reactivity in otherwise identical and isoenergetic molecules.(Received March 26th 1964.) Porter and Wilkinson Trans. Furaday SOC.,1961 57 467. Kasha Discuss. Furaday Soc. 1950 9 14. Reactions of Diphenylcarbene By D. BETHELL and D. WHITTAKER* J. D. CALLISTER THE thermal decomposition of diphenyldiazo-methane in an inert solvent and an inert atmosphere yields benzophenone azine and tetrapheny1ethane.l In hydroxylic solvents thermal2 and photochemical3 decomposition gives the appropriate derivative of diphenylmethanol. Kirmse3 suggested that the photo- lysis product arises by abstraction of the hydroxyl proton by the intermediate (nucleophilic) diphenyl- carbene forming diphenylmethyl cation.Alterna-tively both azine and ether may be formed by com- peting attack of (electrophilic) diphenylcarbene on the diazo-compound and hydroxylic species. In either case the carbene is envisaged as having its non- bonded electrons paired although the azine could be formed by reaction of the diradical form of the carbene with diphenyldia~omethane.~ We have found that thermal decomposition of diphenyldiazomethane in aqueous acetonitrile under nitrogen is kinetically of the first order in diphenyl- diazomethane. At water concentrations between IM and 4h.1the reaction yields exclusively benzophenone azine and diphenylmethanol at lower water con- centrations some bisdiphenylmethyl ether is formed and at higher concentrations a little tetraphenyl- ethylene.Since the ratio of these products is in- fluenced by the walls of the glass reaction vessels all decompositions have been carried out in glass tubes coated internally with polytetrafluoroethylene.Under these conditions the product ratio is controlled by the concentrations of diazo-compound and water. Formation of free diphenylmethyl cations is im- probable since reaction with unchanged diphenyldi- azomethane should give exclusively tetraphenyl- eth~lene.~ A more conclusive test of the proton- abstraction mechanism is provided by replacement of water in our system by deuterium oxide. If proton removal competes with attack on the terminal diazo- nitrogen in the product-determining step then deuteron abstraction should be slower and result in a decreased yield of diphenylmethanol.* The Robert Robinson Laboratories University of Liverpool. Parham and Hasek J. Amer. Chem. SOC.,1954 76 935. Staudinger Anthes and Pfenninger Ber. 1916 49 1928. Kirmse Annulen 1963 666 9. Reimlinger Chem. Ber. 1964 97 339. Bethel1 and Callister J. 1963 3801 3808. JUNE 1964 Rate constants and products of the thermal decomposi- tion of diphenyldiazomethane in aqueous acetonitrile at 85.0" [Ph2CN2] (MI0.0668 water] (M) - 1O2kob8 (min.-l) 0.97 % Azine 24.9 rH (ref. 5) - 0.0668 1.02 0.76 4-5 0.75 0.0668 0.126 1.05 (D20)2.16 0.70 0-84 4-2 3.0 0.74 0.55 0-126 2.16 (D20) 0.79 2.8 0.52 The results in the Table which are the averages of a number of runs indicate that replacement of water by deuterium oxide results in a slight increase in the yield of diphenylmethanol.This rules out the proton abstraction and insertion6 mechanisms and suggests that the process competing with reaction of the carbene with unchanged diazo-compound is attack on the hydroxyl oxygen atom.The absence of pro-ducts of radical-like reactions (e.g. tetraphenyl-ethane and succinonitrile) when hydroxylic com- pounds are present leads us to favour interpretation of our results in terms of competing reactions of singlet diphenylcarbene. (Received April 14th 1964.) Simon and Rabinovitch J. Arner. Chern. Sac. 1963 85 1023; Goldstein and Baum ibid. p. 1885; Chesick and Willcott J. Phys. Chem. 1963 67 2850.The Crystal Structures of Complexes between Trimethylplatinum(1v) and (a) Salicylaldehyde and (b) 8-Quinolinol By J. E. LYDON,MARYR. TRUTER, and R. C. WATLING* IN all the complexes of known structure containing trimethylplatinum(Iv) the platinum atom is six-co- ordinated whatever co-ordination number might be deduced from the empirical formula. We now report the structures of two more complexes having em- pirical formula corresponding to five-co-ordination ; they are trimethyl(salicylaldehydato)platinum Me3-C7H,02Pt and trimethyl-(8-quinolinolato)-platinum Me,C,H,NO.Pt. Details of their prepara- tion will be published separate1y.l FIG.1.Thestructure ofa trimethyl(salicyla1dehydato)-platinum molecule. Three-dimensional X-ray crystal-structure analysis was carried out on each compound observations being made on crystals cooled to 110-120"~.In trimethyl(salicylaldehydato)platinum there are two crystallographically independent centrosymmetrical dimeric molecules in a triclinic unit cell having a = 10.828,b = 11.455,c = 9.147A 01-= 74.8",p = 110.8",y = 92.5"and space group P 1. One molecule is shown in Fig. 1.There are four dimeric molecules * Department of Inorganic and Structural Chemistry The Kite and Truter in preparation. Swallow and Truter Proc. Roy. Soc. 1960 A 254 205. of trimethyl-(8-quinolinolato)platinumin a unit cell c having a = 15-25,b = 16-10 = 10.00A and space group P 2,2,2,; one molecule is shown in Fig. 2. FIG.2.The structure of a trirnethyl-(8-quinolinolato)-platinum molecule.In both compounds there is a symmetrical 0 Pt<*>Pt bridge the six-fold co-ordination of platinum being maintained by the sharing of oxygen atoms. A quite different stereochemistry from that found2 in the 13-dicarbonyl complexes [Me,(RCOCHCO.R')Pt],. Although the chelated ring of the salicylaldehyde bears a superficial resemblance to the mono-en01 form of a chelated /3-diketone the lack of a n-bond to the platinum from the former suggests that the "active methylene" carbon atom to platinum bond has considerable o-bond character. Quantitative discussion must be deferred until refinement of the bond lengths is complete. (Received March 24th 1964.) 1University Leeds 2. PROCEEDINGS Stereochemistry of The Stannic Fluoride-Ethanol Complex 0.RAGSDALE By RONALD and BURCH B. STEWART* IT was found by means of a fluorine-19 nuclear magnetic resonance (n.m.r.) study that tin tetra-fluoride in ethanol forms both the cis- and the trans- isomer. This is in contrast to the report1 that the fluorine-19 n.m.r. spectra of tin tetrafluoride in organic bases consisted of two triplets of equal in- tensity. The two triplets would come from a structure in which two molecules of solvent occupy cis- positions in the octahedron. Stannic halide di-adducts (especially SnCl,) have been thought to form the cis-configurations based upon some dipole moment studies.2 Infrared studies3 have also shown that stannic chloride complexes have cis-configura- tions.Our work now shows that both isomers of stannic fluoride di-adducts can be formed and are at equilibrium in solution. EtOH F I I I El OH Fluorine-19 n.m.r. spectrum of SnF in ethanol at -42" (56.4 Mc./sec.; chemical shift in j9.p.m. with CFCl as external reference). We have studied stannic fluoride in ethanol and found that the n.m.r. spectrum does not resemble two triplets as shown in the Figure. It is concluded that both the cis- and the trans-isomer are present in ethanol solution. The multiplet is an A2B2system and it is assigned to the cis-complex. A complex multiplet was ob- tained since the chemical shift difference for the fluorines was of the same order of magnitude as the F-F coupling constants. Six lines in the multiplet were resolved and the observed intensities agree well with the theoretical intensities for an A2B2spectrum given by Wiberg and Ni~t.~ The single-line resonance was assigned to the trans-isomer.Since exchange collapse of a multiplet will also lead to a single resonance it is usually not possible to assign structural parameters to a single line. The trans-structure was assigned to the single line for the following reasons The single line is not associated with the complex multiplet. The single line is so large that it cannot be assigned to an impurity (the SnF solution was freshly prepared from anhydrous ethanol and SnF found by analysis to contain 97.3% of SnF,) . Since the complex multiplet was resolved a fast exchange process at -40" between SnF4-2C,H,.OH and ethanol can be elimin- ated.Hexafluorostannate was prepared from the SnF,-ethanol solution in quantitative yield by addition of fluoride ion. (Only SnF62- was detected in the n.m.r. spectrum.) -50" the cis- and the trans-isomer are approxi- mately of the same concentration. As the solution is warmed towards room temperature the concentra- tion of trans-isomer decreases; at room temperature the cis-complex predominates. The cis-trans conver- sion is reversible. The preliminary results were treated by a least-squares analysis and a free-energy value of -0.5 kcal./mole was obtained for the trans to cis conversion. The enthalpy change for this reaction is 2 kcal./mole and the entropy change 8 cal./deg.For every ethanol ligand position there is statistically one chance of a second trans-ligand and four of cis-ligands. If exp(dS/R) came out numerical- ly close to 4/1 this could help confirm that the second sharp resonance comes from a trans-complex but for this quantity to be meaningful (in this work) more accurate thermochemical data are needed. The average l"Sn-F and llgSn-F coupling was 1800 c./sec. for the trans-complex and 1850 c./sec. for the cis-complex. These compare with an average coupling of 1550 c./sec. for hexafluorostannate. (Received November 26th 1963.) * General Chemical Research Laboratory Allied Chemical Corporation Morristown New Jersey. Present address (R.O.R.) Department of Chemistry University of Utah Salt Lake City Utah U.S.A.Muetterties J. Amer. Chem. SOC.,1960 82 1082. Osipov J. Gen. Chem. U.S.S.R. 1959,26 343; Osipov Artemova and Bedarev ibid. p. 957. a Beattie et al. J. Chem. Soc. 1963 1514. Wiberg and Nist "Interpretation of NMR Spectra," W. A. Benjamin Inc. New York 1962. JUNE 1964 195 The Effect of Light on a Piels-Alder System By J. A. BARLTROP and B.HESP* THE absorption spectrum of 1,4-benzoquinone in product a 1 :l-adduct of 2,3-dimethylbutadiene and cyclohexane exhibits in addition to short-wave benzoquinone was deduced to have the spiro-pyran absorption a weak absorption (E -20) in the region structure (11) from its n.m.r. spectrum (Table) and 400-500 mp which may be attributed to a "forbid- i.r. spectrum. den" n -+ n* carbonyl transition.The ultraviolet spectrum of 1,4-benzoquinone in 2,3-dimethylbuta-Protons Line Relative 1,3-diene shows in addition a broad band (Amax. positions peak areas 297 mp) which probably arises from excitation of a benzoquinone-dimethylbutadiene charge-transfer r = 3.10 J = 10-2c./s~c. 4 complex. This charge-transfer band decays rapidly at room temperature owing to formation of the Diels- c2 5.92-2 Alder adduct (I). The rate of formation of the thermal c5 7-90 2 adduct is three-four times slower at 0" than at 20". C,,,-methyls 8.32 6 The compound (11) with boron trifluoride in bMe Meb0 acetic anhydride gave a mixture from which quinol Me diacetate could be isolated in 25 % yield. 0- (> U (I) ' In order to define more precisely the electronic ' ' 6- (11) transition responsible for the formation of the pyran (11) the quinone-diene system was irradiated under similar conditions but with light of wavelength > 400 mp.A cleaner product was obtained from which (11) was isolated in an overall yield of 33%. This experiment shows that the spiro-pyran is pro- duced by excitation of the n -w* transition of benzoquinone. (Received March 31st 1964.) Illumination for 1-5 hr. at a temperature below 0" of a deoxygenated solution of 1,4-benzoquinone in 2,3-dimethylbuta-l,3-diene,by a 5OOw "Hanovia" medium-pressure mercury arc in conjunction with a filter transmitting only light of wavelength > 300 mp gave after distillation a fraction separable by preparative vapour phase or silica-gel chromato- graphy and which consisted mainly of a pale yellow solid (m.p.3942O; A,,,. 225 mp E 13,500). The * Dyson Perrins Laboratory Oxford University. The Structure of CephaIochromin By G. TERTZAKIAN G. P. SLATER, R. H. HASKINS and L. R. NESBITT* THE fungus Cephalosporium sp. (P.R.L. 2070) produces in copious amounts a yellow pigment which on continuous extraction with n-hexane yields orange crystals C28H22010, m.p. > 300" [aJD= 510" (in CHCl,). We suggest the name cephalo- chromin and structure (I) for this compound. CephaIochromin has six hydroxyl groups [cf. (11)] two of which are strongly hydrogen bonded to car- bonyl groups [tetra-acetate (111),tetrabenzoate (IV)]. Its i.r. spectrum has strong bands at 3100 (free OH) 2900 (broad chelated OH) 1648 1636 and 1590 cm.-l the last three due to a hydrogen-bonded car- bonyl group and aromatic frequencies found in bi- naphthyl systems.The compound easily yields a di-phenylhydrazone and the carbonyl functions can be reduced to methylenes with zinc and hydrochloric acid. It cannot be hydrogenated at room temperature and pressure over Raney nickel or platinum oxide. The U.V. spectrum of the compound [Amax. (in EtOH) 234 270 295 329 and 418 mp] is very similar to that of 4,5,4',5'-tetrahydroxy- 1,l '-di- naphthy1,l and almost identical with that of ustil-aginoidin A.2 The presence of a binaphthyl skeleton is further confirmed by the detection of small amounts of perylene on fusion of the pigment with zinc and zinc ~hloride.~s* The n.m.r.spectrum (tetramethylsilane as stand- ard; 8 = 0) of cephalochromin exhibits signals at 1.32 p.p.m. (8) (doublet 6 protons 2 methyl groups) 2-64 p.p.m. (doublet 4 protons at positions 3) 4.5 p.p.m. (multiplet 2 protons at positions 2) 5.90 and 6.48 p.p.m. (singlets 2 aromatic protons each) *National Research Council of Canada Saskatoon Sask. Canada. Allport and Bu'Lock J. 1958 4090. * Shibata Ogihara and Ohta Chem. Pharm. Bull. Tokyo,1963 11 1179. Read Shu Vining and Haskins Canad.J. Chem. 1959 37 731. Read and Vining Chem. und Znd. 1963 1239. and 9.28 and 14.44 p.p.m. (singlets 2 protons each from the phenolic hydroxyls at positions 6 and 5 respectively). Superimposed on the 6.48 p.p.m.signal was a broader signal (2 protons) centred at near 6.5 p.p.m. due to the phenolic protons at positions 8. The signals due to the phenolic protons were located by deuteration methylation and acetylation experi- ments. The signal at 14.44 p.p.m. is due to a very strongly hydrogen-bonded proton similar for example to the phenolic proton in 6-acetyl-2-chloro- 3,5-dimetho~yphenol,~ whereas that at 9.3 p.p.m. is also due to a chelated proton but not as strongly hydrogen bonded as the former. The third type of phenolic proton appears to be almost free of hydrogen bonding and gives a similar signal to that obtained from 2,2’-dihydroxy-l,1 ’-binaphthyl. The signals due to the aromatic protons are compatible only with a structure where these are in different environments.The above considerations can lead to only two possible structures (I) and (V) for cephalochromin. Structure (I) is indicated since ustilaginoidin A2 (VI) is obtained on boiling cephalochromin under reflux with iodine and potassium acetate in acetic acid.6 Also the “Birch acetate rule”’ can apply only to (I). r L PROCEEDINGS TI-1 04 (v0 Whereas Allport and Bu’Locks have obtained 5-hydroxy-2-methylchromanone from a fungus we believe this to be the first instance where the more complex dinaphthodihydropyranone system has been described. Further its behaviour towards oxidising agents to be later described more fully is entirely compatible with the results and views of the above workers.(Received April 6th 1964.) Arison Wendler Taub Hoffsommer Kuo Slates and Trenner J. Amer. Chem. SOC.,1963 85 627. Geissman Austral. J. Chem. 1958 11 376. ’Birch and Donovan Austral. J. Chem. 1953 6 373. Allport and Bu’Lock J. 1960 654. NOTICE References cited in the Society’s Publications IT has been decided that authors’ initials will be included in all references cited in any of the Society’s publications. All manuscripts submitted henceforth should conform with this change. June 1964. EDITOR. NEWS AND ANNOUNCEMENTS Library.-The Library will close at 6 p.m. on Friday July 31st and will re-open at 9.30 a.m. on Wednesday August 5th 1964. Election of New Fellows.-95 Candidates were elected to the Fellowship in May 1964. Deaths.-We regret to announce the deaths of the following:Professor S.Bezzi (22.4.64) of the Wni- versity of Padova; Mr. E. T. H. Bucknell (5.3.64) former Science Master; Dr. H. Greene (26.4.64) Adviser on Tropical Soils at Rothamsted Experi- mental Station; Mu. R. I. E. Hall (2.5.64) formerly with Scott and Bowne Limited; Mr. H. T. Lester (10.5.64) of the Ministry of Aviation; and Dr. G. W. Monier- Williams (20.5.64) formerly of the Ministry of Health. The University of East Anglia.-The University of East Anglia has announced the following appoint- ments in the School of Chemical Sciences :Professor of Chemistry Dr. S. F. Mason; Lecturers Dr. A. J. Boulton Dr. M. A. A. Clyne. Dr. R. K. Harris Dr. K. J. Packer and Dr. F. Wilkinson. University of Hull.-The following appointments have been announced :Dr.P. J. Chapman as Lecturer in the Department of Biochemistry Mr. W. J. Wilkinson as Lecturer in the Department of Educa-tion (Physics) Mr. A. F. J. Cox and Mr. J. E. Nicholls as Assistant Lecturers in the Department of Physics Dr. S. G. Wilkinson as Assistant Lecturer in the Department of Chemistry and Dr. D. Clarke as Senior Research Fellow in the Department of Applied Physics. University of Sussex.-The following appoint- JUNE 1964 ments in the Department of Chemistry have been announced to take effect from October 1st Dr. M. F. Luppert of the Manchester College of Science and Technology has been appointed Reader; Mr. R. W. Bott of Leicester University Mr. J. R. Hanson of Imperial College London and Dr.J. B. Pedley of Manchester College of Science and Technology have been appointed Lecturers; and Dr. D. R. M. Wultun of the University of Sussex has been appointed Temporary Assistant Lecturer. Meldola Medal.-The Council of the Royal Institute of Chemistry with the concurrence of the Society of Maccabaeans has awarded the Meldola Medal for 1963 to Dr. A. Carrington for his work in the field of physical and theoretical chemistry with special reference to the measurement and interpreta- tion of electron spin resonance spectra. Van’t Hoff Fund.-The Committee of the Van’t Hoff Fund for the endowment of investigations in the field of pure and applied chemistry invites applications for grants from the fund. The amount available for next year is about Hfl.1750. Applications should be sent by registered post to Het Bestuur der Koninklijke Nederlandse Akademie van Wetenschappen bestemd voor de Commissie van het “van’t Hoff-Fonds” Trippen- huis Kloveniersburgwal 29 Amsterdam before December 1 st 1964. The purpose for which the grant is required the reasons for the application and the amount desired must be stated. Grants from the Fund for 1964 were awarded to Dr. J-C. Maire (Marseilles) and Dr. E. Zbiral (Vienna). Symposium on Molecular Relaxation Processes.- The Chemical Society announces that it is arranging a symposium on “Molecular Relaxation Processes” to be held at the Edward Davies Chemical Labora- tory Aberystwyth from July 14-16th 1965.The techniques which will form the basis of the sym- posium will include Dielectrics Ultrasonics Nuclear Magnetic Resonance and Electron Spin Resonance. These topics will be introduced and reviewed in invited papers by experts in the respective fields. A strictly limited number of short contributed papers can also be accepted for presentation during the meeting. Two copies of abstracts of approximately 300 words should be sent before December 31st 1964 to Dr. Manse1 Davies Edward Davies Chem- ical Laboratory Aberystwyth Wales. Accommodation will be available at University Halls of Residence in Aberystwyth. Advance registration for attendance at the symposium is neces- sary and application forms will be available in April 1965. These will be sent only to those who have applied to the General Secretary The Chemical Society Burlington House W.1. Symposia etc.-The Tenth International Sym-posium on Combustion sponsored by the Combus- tion Institute Pittsburgh Pennsylvania will be held in Cambridge on August 17th-21st 1964. Further enquiries should be addressed to Dr. W. G. Parker Secretary The Combustion Institute College of Advanced Technology Birmingham 4. The Second International Conference on Water Pollution Research will be held in Tokyo on August 24-28th 1964 Further enquiries should be ad- dressed to Ralph E. Fuhrman Executive Secretary Water Pollution Control Federation 4435 Wiscon- sin Av. N.W. Washington D.C. U.S. A. Third European Conference on Electron Micro- scopy sponsored by the International Federation of Societies for Electron Microscopy will be held in Prague on August 26th-September 3rd 1964.Further enquiries should be addressed to the Organising Committee Albertov 4 Prague 2 Czechoslovakia. The Ninth International Conference on Low Temperature Physics sponsored by the I.U.P.A.P. will be held in Columbus Ohio on August 31st- September 4th 1964. Further enquiries should be addressed to Professor J. G. Daunt Chairman Organising Committee c/o Meldenhall Laboratory of Physics Ohio State University 174 W. 18th Avenue Columbus Ohio U.S.A. A Symposium on Corrosion from Heavy Oil Combustion Products sponsored by the German Society for Petroleum Science and Coal Chemistry in connection with its Annual Meeting will be held in Cologne on October 6-7th 1964.Further enquiries should be addressed to Deutsche Gesell- schaft fur Mineralolwissenschaft und Kohlechemie Mittelweg 180 Hamburg 13 Germany (Fed. Rep.). The Third International Synthetic Rubber Sym- posium will be held in London on October 13-l5th7 1964. Further enquiries should be addressed to Rubber and Plastics Age Gaywood House Great Peter Street London S.W.l. An International Conference on Correlations of Particles Emitted in Nuclear Reactions sponsored by the American Physical Society will be held in Gatlinburg Tennessee on October 15-17th 1964. Further enquiries should be addressed to Mr. Alexander Zucker Oak Ridge National Laboratory P.O. Box X Oak Ridge Tennessee 37830 U.S.A.A Symposium on Radiochemical Methods of Analysis will be held in Salzburg on October 19th- 23rd 1964. Further enquiries should be addressed to International Atomic Energy Agency 1 1 Karntner- ring Vienna 1 Austria. The Eighteenth International Symposium on Modern Methods of Analytical Chemistry will be held at Baton Rouge Louisiana on January 25- 28th 1965. Further enquiries should be addressed to Professor P. W. West Department of Chemistry Louisiana State University Baton Rouge La. U.S.A. A Conference on Atomic Spectra and Radiation Processes will be held at Oxford on April 12-14th 1965. Attendance will be by invitation only and further details may be obtained from the Administra- tion Assistant of The Institute of Physics and The Physical Society 47 Belgrave Square London S.W.1.The Thirty-sixth International Congress of In- dustrial Chemistry will be held in Paris on April 21st to May 2nd 1965. Further enquiries should be addressed to SociCtC de Chimie Industrielle 28 rue Saint-Dominique Paris 715,France. Personal.-Dr. W. E. Addison Senior Lecturer in the Department of Chemistry at the University of Nottingham has been appointed Warden of Rutland Hall from October lst 1964. Dr. G. de W. Anderson has been appointed Director of Organic Research at Minnesota 3M Research Limited. Professor S.J. Angyal has been awarded the degree of D.Sc. by the University of New South Wales for his research on the Chemistry of Tnositols and Related Compounds.This is the first award of a D.Sc. by the University. Dr. F. X. Aylward is now Director of the Food and Nutrition Programme sponsored by the U.N. Food and Agricultural Organisation and U.N. Special Fund in co-operation with the Polish government. Dr. V. C. Barry has been elected Treasurer of the Royal Irish Academy. He is a member of the science section of the council. Dr. W. R. Boon has been appointed Director of Imperial Chemical Industries Agricultural Division’s Research Station at Jealott’s Hill Berkshire. Dr. B. E. Conway of the University of Ottawa has been awarded the 1964 Noranda Lectureship which is administered by the Chemical Institute of Canada and sponsored by Noranda Mines Limited. Dr. F. M. Dean of the University of Liverpool has been appointed Visiting Fellow in the School of Chemistry at the University of New South Wales he also plans to visit several other Australian Univer- sities.Dr. G. R. DeZpierre has been awarded an Alexander Brown Coxe Memorial Fellowship at the Yale University School of Medicine where he will undertake biochemical research. Professor C. Eaborn University of Sussex has been awarded the Kipping Award for 1964 for his outstanding contributions in the field of organo-silicon chemistry. He is the first non-U.S. citizen to receive this award which is sponsored by Dow Corning Corporation and which was established in 1962 in honour of Frederic Stanley Kipping. The award was presented to Professor Eaborn at a PROCEEDINGS Meeting of the American Chemical Society held in Philadelphia on April 6th.Mr. E. W. M. Fawcett has been appointed new projects adviser to Laporte Industries Limited. Dr. J. 0. Harris Group Chief Chemist of the Scottish and Newcastle Breweries Limited has been appointed to the Board of Scottish Brewers Limited. Dr. D. E. Hoare and Dr. E. T. Stewart have been appointed Senior Lecturers in Chemistry at Queen’s College the University of St. Andrews. Dr. H. V. R. Iengar formerly a research chemist with Imperial Chemical Industries (India) Private Limited is now at the Central Laboratory of the Alkali and Chemical Corporation of India Rishra. Dr. S. Jacobs formerly of the National Institute for Medical Research has taken up a post in the Biochemistry Department of the University of Ibadan Nigeria.Dr. R. W.L. Kirnber formerly Research Associate at Duke University North Carolina U.S.A. has been appointed a Research Officer in the C.S.I.R.O. Division of Soils Adelaide Australia. Dr. J. A. McCleverty has been appointed Assistant Lecturer in Chemistry at the University of Sheffield. Dr. M. L. McGlashan Reader in Chemistry at Reading University has been appointed to the Chair of Physical Chemistry in the Department of Chem- istry at the University of Exeter from October 1st. Dr. A. McL. Mathieson of the Chemical Physics Section of C.S.I.R.O. has been awarded the H. G. Smith Medal for 1963 by the Royal Australian Chemical Institute. Dr. D. E. Moore Research Fellow in Leeds University has been appointed to a Lectureship in Pharmaceutical Chemistry at the University of Sydney.Dr. R. G. Neville has been appointed Director of Research at the Epoxylite Corporation South El Monte California. Professor H. B. Nisbet of the Heriot-Watt College Edinburgh. has had the Decoration of Commander of the Royal Norwegian Order of St. Olav conferred on him by King Olav of Norway. Mr. R. C. Odarns has been appointed a Vice-chairman of the British Chemical Plant Manu-facturers Association. Dr. J. H. O’Donnell formerly of the University of Leeds and recently a Research Fellow in polymer chemistry at the Polytechnic Institute of Brooklyn has taken up a Lectureship in Chemistry at the University of Queensland.Dr. A. M. Pomrner has been appointed Clinical Assistant Professor of Pediatrics (Nutrition) at Georgetown University Maryland U.S.A. Dr. J. F. Richardson has been appointed Head of the Department of Physics at the Napier Technical College Edinburgh. JUNE 1964 Dr. J. C. Roberts has been appointed Reader in the Department of Chemistry at the University of Nottingham. Dr. R.A. Ross formerly of the Royal College of Science and Technology Glasgow has been ap- pointed Head of the Department of Chemistry at the College of Technology Belfast. Dr. D. R. Rosseinsky has been appointed Lecturer in Physical Chemistry at the University of Exeter from October 1st. Dr. W. D.Scott formerly with the British Tyre and Rubber Company has joined Air Products Limited.Dr. J. Sheridan formerly Reader in Chemistry at the University of Birmingham has been appointed to the Chair of Physical and Inorganic Chemistry at the University College of North Wales. Dr. A. 0.W.Stretton St. Catharine's College Cambridge has been elected into the Stringer Fellowship. Dr. J. G. Tillett of Bedford College London has been appointed Lecturer in Chemistry at the University of Essex. Dr. D. P. TJzornhiZZ formerly of the University of Saskatchewan is now working in the Department of Pharmacology and Therapeutics in the University of Manitoba Medical School Winnipeg. Dr. R. J. Turner has been appointed Director of Research and Commercial Development at the David and Geck Division of Cyanamid International.ADDITIONS TO THE LIBRARY Landolt-Bornstein Zahlenwerte und Funktionen aus Physik Chemie Astronomie Geophysik und Technik. Band 2 Teil2c. 6th edn. Edited by J. Bartels et al. Pp. 731. Springer. Berlin. 1964. Mass spectrometry. Edited by C. A. McDowell. Pp. 639. McGraw-Hill Book Company Inc. New York. 1963. Interpretation of mass spectra of organic compounds. H. Budzikiewicz C. D-jerassi and D. H. Williams. Pp. 271. Holden-Day Inc. San Francisco. 1964. Nuclear chemistry and its applications. M. Haissinsky. (Translated from the French.) Pp. 834. Addison-Wesley Publishing Co. Inc. Reading Massachusetts. 1964. Photosynthesis J. Leggett Bailey. (Royal Institute of Chemistry Lecture Series 1963 no. 5.) Pp. 30. Royal Institute of Chemistry.London. 1963. (Presented by the publisher.) Azeotropy and polyazeotropy. W. Swietoslawski. Translated and edited by K. Ridgway. Pergamon Press. Oxford. 1963. (Presented by the publisher.) Fluidised particles. J. F. Davidson and D. Harrison. Pp. 155. University Press. Cambridge. 1963. Mechanisms of oxidation of organic compounds. W. A. Waters. Pp. 152. Methuen. London. 1964. (Pre-sented by the author.) Technique of inorganic chemistry. Edited by H. B. Jonassen and A. Weissberger. Vol. 1. J. D. Corbett et al. Pp. 268. Vol. 2. N. R. Johnson E. Eichler and G. Davis O'Kelley. Pp. 202. Vol. 3. C. J. Barton et al. Pp. 345. Interscience. New York. 1963. Handbook of preparative inorganic chemistry. Edited by G. Brauer. Vol. 1 2nd edn.(Translated from the German.) Pp. 1002. Academic Press. New York. 1963. Noble-gas compounds. Edited by H. H. Hyman. Pp. 404. University of Chicago Press. Chicago. 1963. Non-stoichiometric compounds. Edited by L. Mandel- corn. Pp. 674. Academic Press. New York. 1964. Organoboron chemistry. Vol. 1. H. Steinberg. Pp. 950. Interscience. New York. 1964. The chemistry of beryllium. D. A. Everest. (Topics in inorganic chemistry. Edited by P. L. Robinson. Vol. 1.) Pp. 15 1. Elsevier. Amsterdam. 1964. Fluorine chemistry. Edited by J. H. Simons. Vol. 3. Pp. 240. Academic Press. New York. 1963. The inorganic chemistry of nitrogen. W. L. Jolly. Pp. 124. W. A. Benjamin. New York. 1964. Studies in the organic chemistry of vanadium. H. J. de Liefde Meijer M.J. Janssen and G. J. M. van der Kerk. Pp. 152. Institute for Organic Chemistry T.N.O. Utrecht. 1963. (Presented by the publisher.) Friedel-Crafts and related reactions. Edited by G. A. Olah. Vol. 2 part 1. Pp. 658. Part 2. P. 659-1362. , Interscience. New York. 1964. Polymerization of aldehydes and oxides. J. Furukawa and T. Saegusa. (Polymer Reviews Vol. 3.) Pp. 482. Interscience. New York. 1963. The cyanine dyes and related compounds. F. M. Hamer. (Chemistry of Heterocyclic Compounds. Vol. 18.) Pp. 790. Interscience. New York. 1964. (Presented by the author.) Comparative biochemistry a comprehensive treatise ; Edited by M. Florkin and H. S. Mason. Vol. 6. Pp. 561. Academic Press. New York. 1964. Chemical analysis of soils.K. K. Gedroits. (Translated from the Russian.) Pp. 602. Israel Program for Scientific Translations. Jerusalem. 1963. Techniques for the use of radioisotopes in analysis. A laboratory manual. Pp. 135. Spon. London. 1964. (Presented by the publisher.) Titrimetric organic analysis. Part 1 Direct methods. M. R. F. Ashworth. (Chemical Analysis. Vol. 15.) Pp. 501. Interscience. New York. 1964. Methods of quantitative inorganic analysis an encyclopedia of gravimetric titrimetric and colorimetric methods. K. Kodama. Pp. 507. Interscience. New York. 1963. Inorganic ultramicroanalysis. I. P. Alimarin and M. N. Petrikova. (Translated from the Russian.) Pp. 151. Pergamon. Oxford. 1964. Fibre structure. Edited by J. W. S. Hearle and R. H. Peters.Pp. 667. Textile Institute and Butterworths. Manchester. 1963. Cellulosic plastics cellulose acetate ; cellulose ethers ; regenerated cellulose; cellulose nitrate. V. E. Yarsley et a!. Pp. 220. Iliffe Books Ltd. for the Plastics Institute. London. 1964. Polyurethanes chemistry technology and properties. L. N. Phillips and D. B. V. Parker. Pp. 129. Iliffe Books Ltd. for the Plastics Institute. London. 1964. Wool its chemistry and physics. P. Alexander and R. F. Hudson. 2nd edn. C. Earland. Pp. 417. Chapman and Hall. London. 1963. Aluminium its applications in the chemical and food industries. P. Juniere and M. Sigwalt. (Translated from the French.) Pp. 267. Crosby Lockwood. London. 1964. (Presented by the publisher.) Nickel plating from sulphamate solutions.R. A. F. Hammond. Pp. 24. International Nickel Co. (Mond) Ltd. London. 1964. (Presented by the publisher.) Physical chemistry of petroleum solvents. W. W. Reynolds. Pp. 211. Reinhold. New York. 1963. An introduction to metallic corrosion. U. R. Evans. 2nd edn. Pp. 253. Edward Arnold. London. 1963. Quality control theory and applications. B. L. Hansen. Pp. 498. Prentice-Hall. Englewood Cliffs. N. J. 1963. Manual on industrial water and industrial waste water sponsored by ASTM Committee D-19 on Industrial Water. 2nd edn. Pp. 757. 1962 printing. ASTM. Phila- delphia. 1963. Elements of chemical reactor design and operation. H. Kramers and K. R. Westerterp. Pp. 245. Chapman and Hall. London. 1963. Chemical engineers’ handbook.Edited by J. H. Perry. 4th edn. R. H. Perry C. H. Chilton and S. D. Kirk-patrick. McGraw-Hill Book Co. Inc. New York. 1963. Chemical engineering. J. M. Coulson and J. F. Richard-son. Vol. 1. 2nd edn. Pp. 492. Pergamon Press. Oxford. 1964. Developments in applied spectroscopy. Vol. 2. Edited by J. R. Ferraro and J. S. Ziomek Proceedings of the 13th Annual Symposium on Spectroscopy Chicago 1962. Sponsored by the Society for Applied Spectroscopy Chicago Section. Pp. 438. Plenum Press. New York. 1963. Theory and practice of ion exchange. Proceedings of the Republican conference Alma-Ata 1962. (In Russian.) Pp. 186. Akademiia Nauk Kazakhskoi SSR. Alma-Ata. 1963. (Presented by the publisher.) The structure and function of the membranes and surfaces of cells.Symposium held in London 1962. Organized by J. K. Grant. Edited by D. J. Bell and J. K. Grant. (Biochemical Society Symposium no. 22.) Pp. 172. University Press. Cambridge. 1963. (Presented by the Society.) Catalytic reactions in the liquid phase. Proceedings of the All-Union Conference held in Alma-Ata 1962. (In Russian.) Pp. 459. Akademiia Nauk Kazakhskoi SSR. Alma-Ata. 1963. (Presented by the publisher.) Methods of separation of subcellular structural com- ponents. Symposium held jointly by the Biochemical Society of Great Britain and the SociCtC Belge de Bio- chimie Louvain 1962. Organised by C. de Duve. Edited by J. K. Grant. (Biochemical Society Symposium no. 23.) Pp. 157. University Press. Cambridge.1963. (Presented by the Biochemical Society.) The control of lipid metabolism Biochemical Society Symposium held at Oxford 1963. Organised by G. Popjak. Edited by J. K. Grant. (Biochemical Society Symposium no. 24.) Pp. 191. Academic Press. London. 1963. (Presented by the Biochemical Society.) Fluidisation. Papers read at a joint symposium organised by the Chemical Engineering Group of the Society of Chemical Industry the Institution of Chemical Engineers (South Eastern Branch) and the Institute of Petroleum London 1963 and the discussions that followed. Pp. 110.Society of Chemical Industry. London. 1964. (Presented by the publisher.)
ISSN:0369-8718
DOI:10.1039/PS9640000161
出版商:RSC
年代:1964
数据来源: RSC
|
|