|
1. |
Proceedings of the Chemical Society. October 1959 |
|
Proceedings of the Chemical Society ,
Volume 1,
Issue October,
1959,
Page 285-340
Preview
|
PDF (6220KB)
|
|
摘要:
PROCEEDINGS OF THE CHEMICAL SOCIETY OCTOBER 1959 FAITH AND DOUBT THE THEORY OF STRUCTURE IN ORGANIC CHEMISTRY By W. V. FARRAR R. FARRAR and KATHLEEN [(K.R.F.) THE UNIVERSITY, MANCKESTER] WHAT is the significance of a structural formula his time “we shall be obliged to acquire a geo- HH metrical conception of (the atoms’s) relative II arrangement in all the three dimensions of solid such as H-C-C-OH? Most people would extension”. But Dalton’s pictorial formulae for I I ethanol (1835) as in Fig. 1 had no structural HH meaning and seems to have been dictated only reply that it is in some sense a map of the mole- by a vague feeling for symmetry. cule representing (if necessary to scale and in three dimensions) the relative positions of the The early nineteenth-century conception of a atoms in the molecule.This view which now molecule was of a roughly spherical collection of Methylalkohol Athylalkohol Dimethylcarbinol Trimethylc&binoL FIG.1 seems so natural was in fact slow to mature and atoms randomly arranged and held together by met with prolonged and bitter opposition. “affinity”. Moreover the increasingly popular In 1808 only a few months after the publica- kinetic theory supposed that these atoms were in tion of Dalton’s “New System of Chemical violent and unpredictable motion. However Philosophy” Wollaston wrote decades ahead of during the eighteen-thirties the work of Liebig 285 286 PROCEEDINGS and Wohler on the benzoyl group showed that some groups of atoms (“compound radicals”) could persist unchanged through a series of chemical reactions.At about the same time a few cases of isomerism were discovered but were ignored or regarded as doubtful; in any case the lack of methods for determining molecular weights made it impossible to distinguish be- tween monomers and polymers. So the clue which now seems to lead straight to the heart of the structure theory was of little value. Towards the end of the decade Dumas recog- nised that chemical properties were dependent upon order within the molecule. He was im- pressed by the resemblance between acetic acid and the chloroacetic acids and saw that “chem- ical character is dependent primarily upon the arrangement and number of the atoms”. But the situation in chemistry was too confused for effec- tive theorising.Because of the lack of agreement on atomic weights even empirical formulae were uncertain and many chemists Dumas among them despaired at times of the atomic theory it- self. Incorrect atomic weights were not a com- plete bar to the development of the structure theory as we shall see later; they merely in- creased the difficulty of systematising the rapidly accumulating data. Valency relationships (perhaps foreshadowed in the law of multiple proportions) were not formulated until 1852 when Frankland de- H H1 :lo H H HN veloped the idea as a result of his researches on organometallic compounds and applied it to nitrogen phosphorus arsenic and antimony. For the structure theory a knowledge of valency relations is essential; it gives a precision clarity and power which are completely lacking in the other theories of the time.Frankland’s ideas were not effectively applied to carbon however until 1857 when Kekulk realised that the valency of carbon must be four. In that year KekulC related a number of carbon compounds to the “marsh gas type.” Using 6 as the atomic weight of carbon he wrote this “type” as CgHHHH The significance of the word “type” in this context is worth discussing. KekulC stated that he was using the word in the same sense as Dumas and not that of Gerhardt. Dumas had spoken of “mechanical types” and “chemical types”. Compounds belonging to the same OCTOBER 1959 decompositions”. In plainer words type formulae are shorthand summaries of all the reactions of which a compound is capable.Thus it was pos- sible for a compound to belong to more than one type according to the reactions under considera- tion. So little reference had type formulz to molecular structure that Sterry Hunt an American adherent of the Type Theory was an opponent of the whole “atomic hypothesis”; clearly he felt this involved no contradiction. This attitude of “no further than the facts” seems to derive from the philosophy of Comte at that time a dominant figure in Parisian intel- lectual circles. Comte was teacher and friend to Williamson whose ether synthesis was respons- ible for the creation of the “water type”. Strangely enough it was Williamson who ex- pressed a point of view quite contrary to the spirit of Type Theory when he wrote in 1852 “(Form6lse) may be used as an actual image of what we rationally suppose to be the arrange- ment of the constituent atoms in a compound as an orrery is an image of what we conclude to be the arrangement of our planetary system”.This first clear expression of the modern view was not developed by Williamson who almost ceased original research in 1856. Two years later ap- peared the papers by KekulC and Couper. These two chemists are responsible almost entirely for the genesis of the Structure Theory. The claims of Butlerov advanced by Russian writers depend more on repetition than on docu- mentary evidence. The paper of KekulC is historically prior partly owing to the pro-crastination of Wurtz in whose school in Paris Couper and Butlerov were working.KekulC’s paper appeared eleven years after he entered the University of Giessen to study archi- tecture. He abandoned architecture for chemistry after hearing Liebig give evidence at an inquest on a hard-drinking countess whose death had been ascribed to spontaneous combustion. By 1858 he had been pupil and colleague to so many chemists that he no longer belonged to any school. In this classical paper (“On the constitu- tion and metamorphoses of chemical com-pounds and on the chemical nature of carbon”) he says “I regard it as necessary . . . to explain the properties of chemical compounds by going back to the elements themselves . . . I no longer regard it as the chief problem of the time .. . to refer compounds to a few types which can hardly have any significance beyond that of mere pat- tern formulae . . . we must ascertain the relation of the radicals to one another and from the nature of the elements deduce both the nature of the radicals and that of their compounds.’’ He went on to develop his ideas on the quadri- valency of carbon combined with the novel con- cept of “Verkettung” the linking together of carbon atoms in chains. Many years later he described how he first “saw” the chains of atoms in a sort of waking vision on a London omnibus one summer evening in 1856. He showed that carbon and hydrogen must form the radicals CH, CH, and CH which would be uni- bi- and ter-valent respectively. He also deduced the constitution of some of the radicals containing oxygen and suggested that radicals might some- times be linked through oxygen or nitrogen.This he thought might be the principle under- lying type formula. KekulC however could not free himself from the presuppositions underlying Type Theory. Not only did he express himself in terms of types and radicals but he thought that his formula were merely summaries of reactions ;like Gerhardt’s only better. He explained in a long footnote his reasons for not ascribing physical reality to his structures and ended his paper on an almost apologetic note “I myself believe considerations of this sort to have only subsidiary importance”. His attitude was remarkably similar to that of the famous preface to Copernicus’s book; that here was but a new hypothesis which would accommodate the known facts and perhaps lead to the discovery of new ones.Couper on the other hand was assured and masterly. At 27 with only three or four years’ study of chemistry behind him he understood clearly what he had done and how important it was. “The aim of chemistry is the erection of a theory” his paper began. As a proposition this is arguable but it is a fine opening to a classic paper. He wrote of Type Theory “it is based on an old but vicious principle which has already retarded science for centuries. It begins with a generalisation and from this generalisation de- duces all the particular instances”. He attacked Gerhardt for his restriction of chemistry to a discipline of mere classification and for his agnosticism about molecular structure.“Can such a view lead to the advancement of science? Would it not be rational in accepting this veto to renounce chemical research altogether ?” With Couper’s detailed destructive criticism of the Type Theory we need not now concern our- selves;constructively like Kekulk he concluded that the valency of carbon was four and that carbon atoms could combine with one another. With these two simple rules he was able to write the correct formula for a number of simple compounds. Carbon at first appeared as C, but was later changed to C; oxygen was always a double atom 0-0. Apart from this Couper’s theory does not suffer from his incorrect atomic weights (C = 6 0 = 8).Couper was also the first to indicate valency bonds by straight lines.* This simple device was of great importance in the development of the structure theory and gave his formulae a very modern look e.g. C (,,, H2 c{O-OH o2 I I C-H2 C-H2 I I C-H2 C-H3 I C-H3 Butan- 1-01 Propionic acid Apart from his convention of 0-0 for oxygen his formulae for the lower alcohols fatty acids ethers ethylene glycol and oxalic and tartaric acid were the same as those used today. He also formulated glucose as an open-chain hydrated aldehyde. Unlike KekulC Couper saw no need for equi- vocation or apology; there is no doubt that he thought of his formula quite naively as a representation of physical reality.Although trained in philosophy he had no leaning toward the scientific agnosticism of the time. Kekulb of course was considered by his con- temporaries to be the sole originator of the struc- ture theory. Couper’s more radical treatment was almost entirely forgotten until his memory was revived by Anschiitz in 1908. The reason is to be found in their very different life-stories. Couper a wealthy Scotsman and a wandering * See however Wheeler in Proc. Chem. SOC.,1959 221. PROCEEDINGS scholar was completely unknown as a chemist in 1858. After the delayed appearance of his paper (which he is said to have felt very deeply) he left Paris for Scotland. He obtained a Univer- sity post in Edinburgh but after a few months lost his reason and lived in retirement at his mother’s home in Kirkintilloch until his death in 1892.He was thus unable to defend his theory or to add to it. Moreover those who knew of his insanity would probably discount his paper for that reason. KekulC on the other hand went on to occupy chairs at Ghent and Heidelberg. He also wrote a text-book which helped to spread his ideas though with his curious lack of conviction he used type formula= throughout. Ten years after the Structure Theory he crowned his achieve- ment by his cyclic formula for benzene; it is un- just that this secondary success should be that for which he is most often remembered today. Himself a mediocre laboratory chemist he was fortunate in having as friend and confidant Adolf von Baeyer a superb practical worker who ac- cepted the Structure Theory from the beginning without being troubled by Kekulk’s philosophical scruples.He used it triumphantly in his classic work on indigo a problem which could not even have been formulated in terms of the Type Theory. His literal interpretation of the Struc- ture Theory is very obvious in his extension of it to strained ring systems. Korner a pupil of Kekulk accepted the enormous labour of working out the full implica- tions of the benzene theory. The final acceptance of the Structure Theory owes almost as much to Korner’s methodical plodding as to von Baeyer’s spectacular successes. There is also however a line of descent from Couper’s paper which is of very great import- ance.His one-time colleague Butlerov returned to Kazan wrote in 1859 a long criticism which damned Couper with faint praise and put for-ward his own version of the “marsh gas type”. By 1861 however he appeared at a conference at Speyer as an ardent advocate of the Structure Theory (he was the first to use the phrase “chem- ical structure”). His text-book was the first to use structural formulze throughout though they were so ingeniously written without valency bonds as to be capable of interpretation also as OCTOBER 1959 type formulae. Later he made his own brilliant contribution to the Structure Theory with his concept of tautomerism. Butlerov came to believe that chemical structures represented physical reality; to Soviet chemists this proves him a materialist and (by courtesy) a dialectical materialist.It is likely too that Couper influenced his contemporary at Edinburgh Crum Brown a man with a “life-long interest in knots and com- plicated systems of knitting”. We do not know what contact there was between the two young men in the few months between Couper’s return to Edinburgh and his final breakdown; but Crum Brown in his doctoral thesis of 1861 took up Couper’s device of the drawn valency bond and extended it to double bonds (1 864). He evolved a “graphic” system of far greater clarity than that of Kekulk (e.g. for propan-2-01 as illustrated) and was able to show that Kekulk’s clumsy symbols could actually lead to erroneous conclusions.a b Crum Brown’s formula for propan-2-01 Crum Brown descended from a long line of Scottish divines had a remarkably subtle mode of thought. He supposed that structural formulae represented reality in “chemical space” which was different from ordinary “physical space” ; this is curiously reminiscent of the scholastic doctrine of “double truth” that a proposition could be true in faith but not in reason and vice versa. At the same time he was attempting to formulate a mathematical theory of chemistry which would be independent of the truth or falsity of the “atomic hypothesis”. Crum Brown’s graphic formulae were an important factor in the success of the Structure Theory evolving as they did into a shorthand capable of manipulation by anyone after a little training.Frankland who began to use graphic formulz in 1866 wrote to Crum Brown the next year “There is a good deal of opposition to your formulz here but I am convinced that they are destined to introduce much more precision into our notions of chemical compounds. The water- type after doing good service is quite worn out”. As this quotation suggests acceptance of the Structure Theory was not immediate. Carey Foster in a report to the British Association on the progress of chemistry (1859) did not even mention it though he had clearly read Couper’s paper. More astonishing is the silence of the elder statesmen of chemistry; Wohler Liebig Dumas Bunsen and Hofmann were all active in the ’sixties and ’seventies but we have not found any comment of theirs favourable or not on the theory.The most curious reaction was that of Blomstrand who saw the new system as the con- firmation and fruition of the ideas of Berzelius; “like Columbus” wrote his obituarist “he stood on the shores of a new continent and thought he had reached the Indies”. In France Berthelot accepted neither the Structure Theory nor the Theory of Types but throughout his long career adhered to his own “synthetic formulae” which introduced no conjectural ideas. Wurtz on the other hand accepted the theory in Kekulk’s terms but went out of his way to disparage his ex-pupil Couper. “In general I find Couper’s formulae too arbitrary too far from experiment. We do not pretend to represent by our rational formulk the inner constitution of compounds.These formulae only represent metamorphoses that is to say the facts accessible to and demonstrated by experi- ment. That is their advantage. In Couper’s formulz on the contrary the place of every atom is indicated .. . It is too hypothetical and it is wrong to put it before us as the law and the prop he t s.” The most persistent opposition came from Hermann Kolbe. Like Berthelot he was an ag- nostic both in religion and chemistry and held firmly to the belief that the arrangement of atoms in a molecule was unknowable; chemical struc- tures were “Phantasiebilder” and an attempt to introduce spiritualism into science. He had a vitriolic pen a combative temperament and a mulish obstinacy; in addition he was editor of the important Journal fur praktische Chemie.In his early days the Quarterly Journal of the Chemical Society had heavily cut (for irrelevance) an article of his attacking some proto-structural ideas of Williamson. In his own journal he suffered no such indignity; not only did it offer a welcome to earnest bunglers who found more isomers than the Structure Theory allowed but the pages were strewn with Kolbe’s own charac- teristic comments giving these numbers a high entertainment value. Kolbe’s own system was a variant of the Type Theory in which there was only one type that of carbon dioxide. This he wrote as (C202)0 (C = 6 0 = 8) since he only belatedly adopted the correct atomic weights.The resulting formula were clumsy and were never used except by Kolbe and a few of his friends; though in his own hands they were sometimes capable of correct predic- tions as shown in the formulae reproduced here. Kolbe’sformula?. V Like Gerhardt he was attempting to formulate an “operational” theory of chemistry in which experimental results were manipulated in a pure- ly logical manner to produce predictions which could be verified. This approach is not neces- sarily wrong-headed though in chemistry it is usually unfruitful; but Kolbe’s system was a failure largely because he never properly defined the various brackets commas and “copulse” in his formulz and the correct use of it remained almost personal to himself.In 1875 van’t Hoff and Le Be1 (two more pupils of Wurtz) independently published their theory of stereoisomerism. This roused Kolbe to one of his highest flights of invective. “A certain PROCEEDINGS Dr. J. H. van’t Hoff an official of the Veterinary School at Utrecht has it seems no taste for exact chemical investigation. He has thought it more convenient to bestride Pegasus (evidently hired at the veterinary stables) and to proclaim in his ‘La chimie dans I’espace’ how during his bold flight to the top of the chemical Parnassus the atoms appeared to him to have grouped themselves throughout universal space.’’ From his own point of view he did well to object; for “chemistry in space” marks the end of agnos- ticism about structures.If you are superimposing three-dimensional formula upon their mirror images you have silently capitulated to the reality of these formula. Van’t Hoff understood this clearly. “When we arrive at a system of atomic mechanics the mole- cule will appear as a stable system of material points; . . .for what we are dealing with here is nothing else than the spatial-i.e. the real- positions of these points the atoms.” He also cleared up another troublesome aspect by sug-gesting that the atoms probably oscillated rapidly about their mean positions in the molecule; this reconciled the demands of the kinetic theory with the static role required of the atoms by the Structure Theory. From this time onwards doubts grew less and the doubters fell silent.Even Kolbe could not entirely exclude structural formula from his Journal though he tried to smother them with footnotes of editorial indignation. He kept up his rearguard action almost alone right up to his death in 1884. Soon afterwards the pages of the J. prakt. Chem. were full of the hated “Phantasiebilder” ;and no one has since arisen to take up his task of creating a “truly scientific chemistry” free from structures and superstition. SECOND INTERNATIONAL CONGRESS OF POLAROGRAPHY CAMBRIDGE AUGUST 24-29TH 1959 THISCongress the successor to the first one held in Prague in 1951 was attended by 215 delegates from 23 countries. It was organised by the Polarographic Society to honour the seventieth birthday of Profes- sor Heyrovsky.It was Heyrovsky’s work on electro- capillarity phenomena conducted largely in this country nearly half a century ago which lead him to develop the method of polarography. Since then he has continued its further development at the Charles University in Prague and more recently in his own Polarographic Institute in that city. It must be a rare event for one man actively with his own hands to do so much for a technique as important as polarography. It was the Polarographic Society’s intention to invite him to give the inaugural lecture and to pre- sent him with a silver medal. Unfortunately ill- health prevented his attendance. In his absence Pro-fessor Brdicka Heyrovsky’s colleague for many years accepted the silver medal on his behalf and OCTOBER 1959 read his message of good wishes.His paper was read for him by his son a close collaborator. His address was devoted to “Oscillographic Polarography”. This is a relatively recent develop- ment. It is a powerful new tool in the elucidation of electrode kinetics not only because of the calcula- tions that can be made from the results obtained with it but because the cathode-ray presentation enables one to make a qualitative assessment by inspection. The scientific programme of the Congress began with four Plenary Lectures. Professor Paul Delahay (Louisiaima U.S.A.) spoke on “Electrode Processes in Polarography”. These he divided into (1) deter-mination of kinetic parameters for charge-transfer processes and elucidation of reaction mechanisms ; (2) correlation between double-layer structure and polarographic processes; (3) study of electrode pro- cesses with preceding coupled chemical reaction ;and (4) adsorption phenomena.Dr. H. N. M. H. Irving (Oxford Great Britain) spoke on “The Stability of Metal Complexes and their Measurement Polarographically”. In this paper it was shown how the polarography of metal com- plexes had yielded new knowledge on their composi- tion and stability. At the same time complex formation was shown to be a valuable tool in the determination of metals in mixtures. Professor W. Kemula (Warsaw Poland) spoke on “The ‘Hanging Drop’ Methods”. This relatively recent technique has increased 1000-fold the sensi- tivity of the polarographic determination of a number of metals and at the same time shown the formation of some intermetallic compounds.When the methods of cyclic voltametry are applied this electrode can demonstrate the existence of intermediate transition radicals. Professor M. von Stackelberg (Bonn Germany) spoke on “Polarographic Maxima”. These maxima were one of the earliest phenomena observed in polarography. They are produced by a number of mechanisms and their elucidation has only recently been completed. The remaining contributions which numbered 29 I over a hundred were divided into six sections. In each section there was one invited lecturer. Dr. G. C. Barker (Harwell Great Britain) (Section 1 Instru-mentation) spoke on “Oscillographic Polarography and Radio Frequency Polarography”.He described the development and use of these new instruments. Professor 0.H. Miiller (Syracuse U.S.A.) (Section 2 Theory and Kinetics) spoke on “Rate Controlled Reactions as Illustrated by the Reduction of Pyruvic Acid”. He advanced a new hypothesis to explain the effect of pH on the two reduction waves. Dr. G. W. C. Milner (Harwell Great Britain) (Section 3 Analytical and Industrial Applications) described the use of the new highly sensitive polaro- graphs and recent methods of preventing the inter- ference of unwanted elements during analysis. Professor R. Brdicka (Prague Czechoslovakia) (Section 4 Kinetic Currents in Polarography) described the placing of chemical reactions in the de- polarisation scheme into antecedent parallel and subsequent reactions with respect to the electrode processes.Dr. I. S. Longmuir (London Great Britain) (Section 5 Biological and Medical Applications) discussed the principles of the application of polaro- graphy to biology and medicine with some illustra- tive examples. Dr. V. S. Griffiths (London Great Britain) (Section 6 Miscellaneous) chose as his topic “Educa- tion in Polarography”. He began with a survey of scientific education in this country from the age of eleven onwards and showed how various streams converge on the universities and colleges of tech-nology. He concluded by discussing where in the curriculum polarography should be taught and to what level.From the remaining papers which were of a high order it would be invidious to select any for special mention. They illustrated the extent of the applica- tion of the technique varying from atomic energy at one extreme to the diagnosis of cancer at the other. The full proceedings of the Congress and the discussions which followed contributions will be published by the Pergamon Press. I. S.LONGMUIR. TILDEN LECTURE* The Triplet State in Chemistry By GEORGE PORTER (UNIVERSITY OF SHEFFIELD) THEconcept of a biradical or a molecule having two used in chemistry. The spectroscopist has been separate unsatisfied valencies has been frequently equally familiar with the term “triplet state” used to * Delivered before the Chemical Society at Burlington House London on December 11 th 1958; at King’s College Newcastle upon Tyne on January 20th !959; at The University Southampton on January 30th; at The University Leicester on February 16th; and at Marischal College Aberdeen on March 10th.describe an atom or molecule having two electrons with parallel spin. Spectroscopic observations of the triplet state were until recently confined to atoms and diatoms whilst the most characteristic examples of biradicals are found in more complex organic molecules. One of the most important recent ad- vances of molecular spectroscopy has been the ob- servation of the triplet states of a wide variety of organic molecules and the determination of their energy levels. Equally important to the chemist is the development of methods which make possible the direct observation of triplet states in solution and in the gas phase and the study of their physical stability and chemical reactivity.It will be appropriate to consider first how the triplet state arises and why it is of importance in chemistry. Molecules having an even number of electrons i.e. nearly all chemically stable molecules have odd multiplicity and usually have a singlet ground state but excited configurations give rise to both singlet and triplet states. If the spins of electrons occur in antiparallel pairs the resultant spin angular momentum is zero giving a single or “singlet” level. I€the spins of two electrons are parallel the resultant spin angular momentum is + + Q = 1 in atomic units and this vector has three components +1 0 and -1 which results in a “triplet” of energy levels in an internal molecular or external field.The degree of splitting of the three components of the triplet in the absence of an external field is determined by the coupling between spin and orbital motions which in turn depends on the atomic masses. In light atoms such as those of which most organic molecules are composed the coupling is small and it will be suflicient to regard the triplet state in such molecules as a single energy level. According to the Pauli principle a single non-degenerate atomic or molecular orbital can accom- PROCEEDINGS configuration of lowest energy will usually be a singlet -ft-Singlet ground state + but each singly excited configuration may be sing1et or triplet; e-g.the lowest excited configuration -+ + + -+t + Triplet Singlet The singlet and the triplet state arising from the same configuration have different energies and the triplet state lies the lower. This has long been known for atoms in the empirical form of Hund‘s rule. The difference in energy between singlet and triplet levels of the same configuration does not arise from interactions between the spins of the electrons which are relatively quite weak but from the different spatial electron distributions which are dictated by the spin functions in accordance with the Pauli principle. This principle may be stated in the more general form “The total wave function of an atom or molecule is antisymmetric to the exchange of any two electrons.” Allowing for indistinguishability of the two electrons we obtain the following satisfactory wave functions for the ground and first excited configurations modate only one electron of each spin quantum Here $a and t,hb are the co-ordinate wave functions number.The maximum number of electrons in any for the orbitals u and b and a and /3 the spin wave orbital is therefore two and these must have anti- functions corresponding to spin 4 and -*. Other parallel spins SO that the ground state having the combinations of these spin and co-ordinate functions OCTOBER 1959 would not change sign on interchanging electrons 1 and 2 and would be contrary to the Pauli principle.The spin functions therefore determine the co-ordinate functions which are acceptable. When we used the “single electron” molecular orbitals the singlet and triplet states arising from the same configuration had exactly the same energy but coulombic repulsion between the two electrons was completely ignored. This repulsion will increase the energy by an amount which depends on the averaged relative position of the two electrons. Now the prob- ability that electron 1 is in a volume dT1 and at the same time electron 2 is in a volume element d7 is proportional to Y/y*d~,d~ where for the excited singlet state Ys = $mqLb(2) + $m$bW and for the triplet YT = $,(Q$am -$,GI $aW The two functions therefore lead to different probability density distributions and it has been shown by Dickens and Linnett for several typical cases that if the probability is plotted as a function of the co-ordinates the chance of the two electrons’ being at the same point is kite when the singlet wave function is used but is zero with the triplet wave function.Quite irrespectively of electron repul- sion the electrons are further apart in the triplet than in the singlet state. _____c__ Singlet Triplet When electron repulsion is considered the electrons will be further separated in both states but will re- main further apart in the triplet than in the singlet state. The coulombic repulsion will therefore be less in the triplet which will be of lower energy than the singlet.A second consequence of the different elec- tron distributions is that in so far as a “biradical” has two separate unpaired electrons the triplet state is more of a biradical than the corresponding singlet. A true biradical in which the two electrons would be so far separated that no exchange occurred would of course have degenerate singlet and triplet levels and be better regarded as two independent doublets. ,+ If two orbitals are degenerate or have nearly equal energies the lowest singlet and triplet states will have equal energies in the single electron approximation and after allowance for electron interaction the triplet state will be the lower. ++ ++ Triplet ground state This is the case in oxygen and also in some highly conjugated organic molecules where a near-degener- acy of orbital energies results in a triplet ground state.Triplet Levels in Atoms and Diatomic Molecules To illustrate how this works out in practice we may refer to some typical atomic and molecular energy levels. The levels of three atoms He Ca and Hg are shown in Fig. 1. Ca 4’P -/3’D -J3D4 -4=pz -4’s FIG.1. Lowest excited levels of some atoms. In a heavy atom such as mercury spin-orbit interaction results in a considerable energy difference between the three components of the triplet and also in strong intercombination lines. The levels of diatomic molecules are quite similar as shown in Fig. 2 for H, CO I, and GaCl. Again in the heavy molecule iodine the singlet-triplet transition is quite strong and is indeed responsible for its colour.All the states shown on this diagram are stable except for the triplet level of H which is entirely repulsive and is the state which arises from the antisymmetric combination of co-ordinate wave functions in the Heitler-London treatment. In molecular-orbital language one of the electrons is promoted from a bonding to an antibonding orbital so that the energy is greater than that of the two PROCEEDINGS 300 GaCl 3 200 s F % e 100 * 0- FIG.2. Lowest excited levels of some diatomic molecules normal 1s hydrogen atoms with which the state cor- relates. The corresponding singlet has an even higher energy with respect to two 1s hydrogen atoms but correlates with a 1s and a 2p atom and is therefore stable with respect to its dissociation products.But the instability of the first excited triplet state of hydrogen is by no means general. Some of the higher excited triplet states of H2 have deeper potential minima than the singlet states and the lowest triplet states of most diatomic molecules are quite stable. The calculations of Heitler and London have only been carried out for s states and apply of course only to two electrons other electrons being ignored. It is probable that the hydrogen situation also occurs in other molecules in which all electrons apart from those in very low energy levels are used in forming single bonds e.g. in saturated paraffins.On the other hand the excitation of non-bonding or antibonding electrons will not necessarily reduce the bond strength whilst the excitation of electrons in a multiple bond although it may reduce the bond energy will not usually result in an unstable state. Since few molecules apart from the paraffins have all outer electrons in single bonding orbitals we shall expect to find that most lower triplet states are stable. To summarise the lowest excited state of nearly all stable molecules is a triplet state and except in molecules whose electrons are all occupied in single- bond formation the lowest triplet state is probably stable. This state has two separate unpaired electrons and in so far as excited states play a part in chem- istry we shall expect that the lowest triplet level will be of prime importance.Triplet States in Polyatomic Molecules The positions of the triplet states of simple polyatomic molecules such as H,O CO, C2H, and C,H are not yet known. We may be quite certain that such states exist and it is probable that they are stable. In some of these molecules the triplet may lie near enough to the ground state to be significant in the determination of thermodynamic properties especially at high temperatures such as the heat capacities of water and carbon dioxide in combustion processes. The dearth of our knowledge about triplet states of polyatomic molecules arises from the fact that radiative transitions between states of different multiplicity e.g.singlet to triplet are forbidden and because the probability of radiationless processes of energy dissipation is high. Two special methods are now available for the direct study of the triplet states of polyatomic molecules. The first is observation of the phosphorescence of rigid solutions and the second is the study of absorption spectra of the triplet state after flash activation in solids liquids or gases. Phosphorescence of Rigid Solutions.-The phos-phorescence of solid solutions of aromatic molecules and dyes has been known for more than half a century. In 1935 Jablonski suggested that the pheno- menon involved a transition from a metastable level of the molecule which was reached by radiationless crossing from the first excited state as shown in Fig.3. Excitad It state Metastable state Phosphorescence Ground state FIG.3. Jablonski diagram In view of the previous discussion it is now natural to conclude that this metastable level is the lowest OCTOBER 1959 triplet level of the molecule although this interpreta- tion was not found until a decade later. In the meantime attempts were made to explain phos- phorescence and the Jablonski metastable state in terms of tautomeric or isomeric forms ,of the mole- cule. Interpretation in terms of the triplet state was first suggested by Lewis Lipkin and Magel and also states in solution by means of their absorption spectra. Strong absorption spectra to higher triplet levels were known to exist and had been detected in rigid media even before the triplet state theory was developed.It was not clear however whether the absence of phosphorescence in ordinary solutions meant that the triplet state was not generally formed by Terenin and was put in a clear convincing form or that it was deactivated in fluid media after forma- by Lewis and Kasha in 1944 since when extensive tion by some radiationless process which was very rapid compared with the radiative process of Iphosphorescence. -& The principle of our method is shown in Fig. 5. loor 80 -51 The photolysis flash cannot excite the triplet state -G 40 a[ -6 Naphtha/ene Ant hracene FIG.4. Lowest singlet and triplet levels OJ some organic compounds directly to any significant extent but it populates upper singlet states which may pass by radiationless conversion into the lowest triplet state as they are known to do in the rigid media.After a short time interval has elapsed a second spectroscopic monitor- ing flash is triggered; light from this passes through the solution and records the absorption spectrum of the lowest triplet level as it passes to an excited triplet state. The experiment was immediately successful and the first record of triplet-triplet absorption spectra in solution is shown in Plate 1 (facing p. 310). It was obtained from a 10-5~-solution of anthracene in hexane which had been thoroughly outgassed. Absorption of GROUND STATE 1 monitoring source \,conversion 4u 1 TRIPLET I Fluorescence Phosphorescence and Radiationless conversion FIG.5.Transitions involved in flash-photolysis studies of the triplet state work has established the lowest triplet energy levels and radiative lifetime of a large number of sub- stances. Some typical energy levels of polyatomic molecules are shown in Fig. 4. Observation of Triplet States by Flash Photolysis of Solutions and Vapours.-Important as phosphores- cence studies have been in the development of triplet-state theory and the measurement of the lowest triplet energy levels they are not normally applicable to the study of triplet states in solution or in the gas phase which are of course of most interest to the chemist. In 1952 Mr.Windsor and I decided to apply the method of flash photolysis which had been success- fully used in the detection of transient free radicals in the gas phase to attempt the observation of triplet The life of the triplet state is seen to be about 200 psec.which is short enough to explain why phos- phorescence is not observed under these conditions but long enough to be studied kinetically by flash techniques. Absorption spectra of many other triplet states were observed in this way in a variety of solvents and shortly after these experiments Dr. F. J. Wright using a longer reaction vessel surrounded by a furnace to give the required vapou pressure was able to record the triplet-triplet absorption spectra of several of the same molecules in the gas phase. One may ask how these absorption spectra are known to be those of triplet states; indeed other spectra are obtained under identical conditions which we have assigned to free-radical dissociation products.The assignments are based on several types of evidence. First the study of a series of related molecules usually makes possible the identification of spectra which arise from free radicals and similar transient products. Secondly the spectra assigned to triplet states are also observed in rigid solvents where they decay with the same rate as the phosphorescence whilst free radicals and similar products if formed at all in rigid solvents are usually trapped indefi- nitely. Recently theoretical calculations and absolute measurements of extinction coefficients have pro-vided a third type of evidence.In many of our flash-photolysis experiments the proportion of molecules d I I I I 7 2 3 4 .. Number of rings 60 h 40 “I 4 20 I I I 3 4 OL ; 2 Number of rings converted into the triplet state is so high that the singlet state is depopulated to a significant extent and it has therefore been possible to measure the absolute concentrations and extinction curves of the PROCEEDINGS triplet transitions. Semiempirical molecular-orbital calculations particularly those of Pariser have pre- dicted the positions of these levels for the linear polyacenes and a comparison of our measurements with Pariser’s calculations is shown in Figs. 6 and 7. The combined errors of the experimental estimations and the calculations are quite high but the general picture leaves little doubt as to the identity of the theoretically predicted triplet levels and those which we have measured.Triplet State Kinetics in Gases and Solutions These observations opened up the possibility of direct studies of the triplet state in chemical reactions. FIG. 6. Experimental triplet levels and oscillator strength of the linear polyacenes 1 5 FIG.7. Calculated triplet levels and oscilla- tor strengths for the linear polyacenes (Pariser) -2u I 5 But before chemical behaviour is investigated the more fundamental question of triplet state deactiva- tion in the absence of chemical change must be better understood. The outstanding problem was simply OCTOBER 1959 this “Why is phosphorescence of long duration ob- served only in solid or rigid media?” It is clear since we have found that the triplet state is readily formed in gases and liquids that the absence of phos-phorescence must be caused by a rapid radiationless process of some kind.What is this process and why does it not occur in rigid media? Our first experiments showed that the decay was predominantly of the first order and that the rate constant decreased markedly with increasing vis- cosity. The viscosity effect suggested some form of diffusion control and it therefore seemed obvious that we must look for collisional deactivation pro- cesses. Several of these were found e.g. oxygen deactivated the triplet state with nearly unit en-counter efficiency but in the absence of any known quenching molecules the decay was essentially of first order in triplet concentration and the lifetime was still very much shorter in solution than the radiation lifetime found in rigid media.Flash-spectroscopic methods involve measurement of photographic plate density which is not a very accurate procedure and subsequent kinetic work has been carried out mainly by photoelectric methods. Here a continuous monitoring source is used and the transient absorption is recorded at one wavelength throughout its formation and decay. Typical records of this kind obtained by Mr. M. R. Wright for triplet naphthalene in solution are shown in Plate 2.Measurement of these traces leads to very accurate kinetic data and we found that the decay was in fime (min.) FIG. 8. First-order plot of the decay of triplet naphthalene in hexane energy-transfer from an excited state which is apparently independent of collisions indeed the re- sults in the gas phase indicate that the decay rate is independent of gas pressure. The only reasonable explanation of these facts seems to be as follows (1) The rate-determining process of triplet-state deactivation in solution and in the gas phase in the absence of specific quenching by other molecules is radiationless conversion to an isoenergetic level of the ground state. (2) The transition is inhibited or prevented in viscous or rigid media for the following reason.If a TABLE 1. First-and second-order rate constants of triplet naphthalene decay in several solvents. (Concentrationof naphthalene = lo-%) Solvent Viscosity k (sec.-l) k (1. mole- sec.-l) (CP) n-Hexane 0-3 12.1 & 1.3 2.1 f0-6 Water 1.1 Ethylene glycol 21.1 Liquid paraffin I 33.0 Liquid paraffin I1 167 fact partly of the second order i.e. there was some contribution to the triplet decay from triplet-triplet encounters.A first-order plot from the data of Plate 2 is shown in Fig. 8. The conclusion that the decay was predominantly of the first order was however confirmed and it was possible to measure both second- and first-order decay constants and to show that not only the second-order decay but also the first-order decay was viscosity-dependent.The rate constants for naphthalene in various solvents are shown in Table 1. These findings are rather strange. We have a spontaneous unimolecular process which is apparent- ly diffusion-controlled. We are also dealing with 7.5 f0.6 4.1 f1.2 0.97 f0.1 0.22 f0-03 1.5 f0.1 0.39 f0.02 0.31 f0.03 0.08 f0.01 molecule is rigidly held in one particular nuclear configuration it will oscillate back and forth between the two electronic states but will not be able to lose energy since no conversion into kinetic energy can take place. It will therefore remain in this configura- tion until energy is lost by radiation. It is interesting that the appearance of phos-phorescence is then dependent not only on the spin-forbidden nature of the radiative transition but also on the configurational forbiddeness of the radia- tionless transition in rigid media.The early explana- tion of the metastability of this level in terms of an isomeric form is therefore in some respects rein- troduced. AIthough the rate of radiationless conversion is not generally very dependent on the properties of the solvent apart from its viscosity some molecules are able to quench the triplet state with high efficiency. Apart from those which quench by chemical reaction two other principal types of quenching molecule have been distinguished. (a) Pnrcrinagnetic molecules. Windsor showed that oxygen and nitric oxide quench the triplet state of anthracene and other aromatic molecules with nearly unit encounter efficiency.More recently we have found that paramagnetic ions of the first transition and rare-earth series also quench it but that the quenching eficiency is not related to the magnetic susceptibility as it is in the somewhat similar process of nuclear spin conversion in ortho- and para- hydrogen. The process probably involves complex- formation the function of the paramagnetic mole- cule being that the complex formed between triplet and paramagnetic quencher can dissociate into the ground singlet and unchanged quencher without violation of the spin conservation rules. (b) Molecules with lower energy triplet states. The transfer of electronic energy by the process A* (triplet)+ B (singlet)+A (singlet) and B* (triplet) is allowed by the spin-conservation rules and is pos- sible provided the triplet of B lies lower than that of A.Terenin showed by phosphorescence measure- ments in rigid glasses that transfer of this type does take place although with very low efficiency the rate constants being about 10 1. mole-1 sec.-l. Mr. Wilkinson has recently shown that in ordinary solu- tions this type of process can be very efficient rate constants greater than 1081. mole-l sec.-l being found. This high efficiency of transfer coupled with its long lifetime suggests an important role for the triplet state in radiation chemical and biological systems. The flash photolysis of chlorophyll solutions for example shows that extremely efficient conversion into a metastable-probably triplet-state occurs and it will be surprising if this state does not play an important part in the primary processes of photo- synthesis The Triplet State in Photochemistry We are now in a position to consider those collisional processes of the triplet state which result in chemical change.We have every reason to expect that the triplet state will play an important if not a predominant part in many photochemical reactions. It is formed in high yield from many molecules as exemplified by the following quantum efficiencies Benzene 18% Triphenylene 36% Anthracene 1304 Benzophenone 55 yo PROCEEDINGS Its electronic structure leads us to expect a high reactivity typical of a biradical. The excited singlet state which may also be regarded as a biradical in some respects has a shorter intrinsic lifetime than the lo-* sec.which we have found to be typical of most triplet states in solution so that the triplet has a greater probability of reaction. I have chosen three examples which have played an important part in the development of photochem- istry. These are anthracene the quinones and the aldehydes and ketones all of which are readily written in biradical form 0. Anthracene.-Anthracene solutions on irradia-tion yield dianthracene and if oxygen is present the transannular anthracene peroxide. The fact that the two reactions proceed independently and apparently require the postulate of two distinct excited states had already been indicated by the fluorescence and photochemical investigations by Bowen.Kinetic flash-photolysis studies showed us that the anthracene triplet state is deactivated by oxygen at nearly every encounter and that encounters with normal anthracene have no effect suggesting that the dimer is formed from the singlet state and the peroxide from the triplet state. Consideration of all the evidence from fluorescence and flash studies shows this view to be correct although peroxide is also formed via the singlet state at high oxygen con- centrations. The scheme of reactions is shown in Fig. 9. The oxidation reaction is really a photo- sensitised oxidation of anthracene by anthracene and a similar mechanism applies to photosensitised oxidation of other molecules by anthracene as well as the photo-oxidation of anthracene by other molecules.AO + A FIG.9. Scheme ofphoto chemical reaciions for anthraceite Here we have a rather well-established case of triplet-state participation and of the fact that singlet OCTOBER 1959 and triplet states may have quite distinct chemical be haviour . Quirtones.-Quinones and dyes of related structure are also able to photosensitise the oxidation of other molecules but the mechanism of photosensitisation is quite different from that of anthracene and indeed these two examples illustrate the two principal modes of photosensitised oxidation. In anthracene the pri- mary reaction of the excited state was with oxygen; in the quinones the primary reaction is with the substrate.The whole scheme of reactions both in the presence and in the absence of oxygen is shown in Fig. 10 without distinguishing for the moment between the importance of singlet and triplet states. Oxidation of the substrate occurs even in the absence of oxygen the quinone being simultaneously reduced to the semiquinone and subsequently to the quinol. That oxidation of alcohols involves hydrogen- abstraction by the excited state has been shown by the observation that the semiquinone neutral radical appears first even when the equilibrium form is the radical-ion. In other cases e.g. the oxidation of ferrous ion the primary process is probably electron- transfer. \\ Io2 -Q t HO, -FIG.10. Scheme of photochemical reactions of quinones.The question we are concerned with is whether the triplet or the singlet state is responsible for the oxidation reaction and the results on the two systems so far studied viz. duroquinone and Methylene Blue are somewhat surprising. Dr. Bridge studied the flash photolysis of duroquinone in alcohol solu- tions and was able to detect three separate short- lived transients which were assigned to the triplet the radical and the semiquinone radical ion. The spectra of triplet and radical in viscous paraffin solu- tion are shown in Plate 3. It was then possible to follow the triplet and the radical concentrations by recording their extinctions simultaneously. If the radical R were formed from triplet T then the relation d[R]/dt = k[T] would be valid the decay of the radical being rela-tively slow.That this is not the case is readily seen from the oscilloscope traces of the two species in Plate 4 where the rate of formation of radical is nearly zero at the time when the triplet concentration is a maximum. Detailed analysis shows that the rate of radical formation is proportional to light intensity F and we can conclude that the radical is formed from the singlet state and independently of the trip- let. On the other hand a preliminary study of the Methylene Blue-ferrous ion system by Parker sug- gests that the opposite is true in this case and that photo-oxidation occurs predominantly via the triplet state. I may be forgiven if I do not attempt to generalise from two examples which give opposite results.Ketones.-More attention has been paid to the photochemistry of ketones and aldehydes than to that of any other class of compounds. The greater part of this work has been concerned with photodis- sociation which is of relatively little importance in solution where the excited carbonyl compounds re- act as typical biradicals. Indeed this was pointed out R,,$-a+ H+ R\ R' FIG.1 1. Scheme of photochemical reactions of benzo-phenone and other ketones. in the early work of Backstrom before the triplet- state theory had been developed. Essentially the situa- tion is very similar to that found for the quinones but hydrogen-abstraction now gives ketyl radicals which dimerise as shown in Fig.11 for the particular case of benzophenone. The ketyl radicals which are formed have recently been detected in both their neutral and their ionised forms by Mr. Willcinson in flash-photolysis experiments very similar to those of Bridge on the quinones. Wilkinson has also been able to show that the abstraction of hydrogen by benzo- phenone to yield the ketyl radical occurs by reaction of the triplet state. In this case absorption by the triplet state is not observed and a more subtle method had to be used. As already mentioned trans- fer of energy can occur from an excited triplet to a second molecule with a lower energy triplet level. Benzophenone has a higher triplet state than PROCEEDINGS naphthalene but its excited singlet state lies above the lowest excited singlet of naphthalene so that if benzophenone is excited in the presence of naphtha- lene transfer between triplet states can occur but transfer between singlet states is energetically impos- sible.Three solutions were examined by flash photo- Iysis using wavelengths absorbed only by benzo- phenone. This is shown in Plate 5. The first (a) was benzophenone alone which showed the ketyl radical the second (b) was naphthalene alone which showed no transient spectra since no light was absorbed and the third solution (c) contained naphthalene and benzophenone at the same concentrations as in experiments (a) and (b) and showed the triplet of naphthalene and complete absence of the ketyl radical.This transfer of energy from the triplet of benzophenone has produced the triplet of naphtha- lene and has quenched the triplet of benzophenone before it reacted to give the ketyl radical showing that the triplet state was responsible for all ketyl radicals originally formed.* These few examples are sufficient to show that the triplet state is sometimes of predominant importance in photochemical reactions whilst in other cases the singlet state only is responsible even when triplet is present. There are several reasons why the first excited singlet and triplet states have different chemical pro- perties. In the first place even where they correspond to the same type of electronic transition the electron distribution in the two states is different as we have already seen.Secondly in some molecules it may happen that owing to different singlet-triplet split- ting of different types of transitions the lowest excited states of the triplet and singlet manifolds may arise from different types of electron transition resulting in excited states having quite distinct electron-distributions and chemical reactivities. The important development of the last few years is that this distinction has been recognised and it is now an essential part of the study of photochemistry to find the degree of conversion into singlet and triplet states and to separate their reactions. The Triplet State in Thermal Reactions Whilst it is clear that excited electronic states are of primary importance in photochemical and radia- tion chemical reactions their participation in thermal reactions is less apparent.Thermal population of the triplet state will only occur to a measurable extent at normal temperatures when the excitation energy is less than about 15 kcal. mole-l. This is quite rare. It is also rare to find bond-dissociation energies as small as this but bonds are broken nevertheless. In fact energetically the fission of single bonds and the formation of triplet states are quite comparable as is shown by the typical energies in the annexed Table. Triplet energy (kcal. mole-l) Benzene 85 Naphthalene 61 Anthracene 42 Pentacene 22 Naphthol 60 Anthraquinone 56 Benzop henone 68 Acetone 70 Biacetyl 55 Bond energy (kcal.mole-') C-H 80-100 c-c 60-85 Some reactions which follow dissociation into free radicals are compared with analogous reactions fol- lowing the formation of the biradical state in Table 2. The reactions may be divided into three classes which are distinguished by the energies required to dis- sociate a bond or to form the biradical state. TABLE 2. Comparison of radical and biradical reactions. Radical BiradicaZ Formation and equilibrium hv heat hv heat . AB-A+B AzB-A-B Transfer reactions A + RX + AX + R A -B + RX +A -BX + R Addition A+B=C+AB-k A -B + C=D+A -B-C-D Isomerisation or ABCD AB(1) -+ A + B -+ AB(2) A=B(l) -+ A -B -+A=B(2) Participation in the transition state X+AB+X.. A.. B+XA+B x + A=B +A = B -+XA -B * I am grateful to Mr.Wilkinson for these unpublished results.OCTOBER 1959 (i) Energies of 15 kcal. or less. Here we may compare molecules such as hexaphenylethane and Chichibabin’s hydrocarbon (I) which at room temperature are in equilibrium with a measurable proportion of radical and triplet state respectively. In the extreme case we have diphenylpicrylhydrazyl which is a free radical even in the solid state and Schlenck’s hydrocarbon (11) which has a triplet ground state. (ii) Energies between 15 kcal. and 40 kcal. In this range the equilibrium concentration of radical or bi- radical is too small to be observed directly but dis- sociation and activation occur at a significant rate at normal temperatures.Molecules with weak bonds such as the acyl peroxides yield free radicals and molecules with low triplet -st a te excitation energies such as the linear polyacenes react in a manner characteristic of biradicals. In high-temperature pyrolysis bond fission and triplet excitation may be important at energies above this range and the trip- let state may contribute to the complexities of some of these decompositions and the apparent occurrence of two separate transition states which has recently been found by Hinshelwood and his collaborators. (iii) Energies greater than 40 kcal. The limit of 40 kcal. is chosen rather arbitrarily to represent an energy which is greater than the activation energy of reaction so that direct thermal dissociation or excita- tion is precluded.But just as bond energies still play a part in determining the rate of radical reactions such as hydrogen abstraction the triplet state may be expected to be involved in a similar way in the transition state e.g. for radical addition at a multiple bond. Some evidence that this is so has been provided by Szwarc who found that the logarithm of the “methyl affinity” of aromatic hydrocarbons was closely related to their triplet-state energy and ex- plained this finding in exactly the same way as Polanyi explained the relation between activation energy and bond energy in metathetical reactions. Direct evidence of the participation of triplet states in thermal reactions is not available as it is for photo- chemical reactions but this is not surprising since evidence for free radicals in thermal reactions is also mainly indirect.In photochemistry triplet-state formation is now known to be comparable in im-portance with dissociation as a primary process and it will not be surprising if in thermal reactions as well the energy levels and reactivities of the triplet state become as much a part of chemistry as bond- dissociation energies and the reactions of free radicals. COMMUNICATIONS Perfluoroalkyl Boron Compounds J. J. LAGOWSKI and P. G. THOMPSON CHEMICAL LENSFIELD (UNIVERSITY LABORATORY ROAD CAMBRIDGE) IN spite of the extensive investigations of perfluoro- alkyl derivatives of both metals and non-metals during the past ten years no compounds containing a B-R (R = perfluoroalkyl) bond have been re-ported.The formation of these bonds is discussed briefly here and information is presented concerning the preparation of compounds containing such bonds. Boron alkyls are most readily prepared by the action of a Grignard reagent on boron halides but they have been classically prepared by the reaction of boron halides with zinc or mercury alkyls. We find however that reaction of bis(trifluoromethy1)- mercury with boron halides or substituted boron halides at or below room temperature yields boron trifiuoride or its derivatives as the major volatile product ; minor products include trifluoromethyl halides and a substance which yields tduoro- methane on hydrolysis. Bis(pentafluoroethy1)mercury does not react with boron trichloride below 190°,but boron trifluoride is formed above this temperature.The fluorination of boron halides could occur (a)by formation of a perfluoroalkylboron followed by its decomposition or (b)by direct fluorination involving the transfer of a fluorine atom from a perfiuoroalkyl group attached to the mercury atom. Nuclear mag- netic resonance and infrared data on the system Hg(C2F,),-BC13 suggest that the latter process does not occur to a significant extent at room tempera- ture. On the other hand several observations indicate that B-R bonds are formed initially and that com- pounds containing these bonds are unstable with respect to boron trifluoride. Thus fluoromethyl- boron difluoride is associated in the gaseous liquid and solid state and is unstable yielding boron tri- fluoride; infrared data indicate that the association occurs between a fluorine atom bound to a carbon atom and the tervalent boron atom of a second PROCEEDINGS mo1ecule.l Compounds containing B-CF bonds yielding a solid residue.Hydrolysis of this residue would be expected to have properties similar to those with aqueous potassium hydroxide at room tempera- of fluoromethylboron difluoride and the isolation of ture yielded heptduoropropane ; heptafluoro-n-a small amount of a substance which yielded tri- propyl iodideisnot hydrolyzedunder theseconditions. fluoromethane on hydrolysis supports the suggestion The infrared spectrum of the solid residue showed that unstable trifluoromethyl derivatives of boron bands corresponding to tri-B-methyltri-N-phenyl-are formed in the reaction of bis(tduoromethy1)-borazole (presumably formed by reaction of methyl-mercury with boron halides.From these considera- lithium with the trichlorotriphenylborazole) and tions it appears that stable compounds containing unidentified bands in the B-N borazole-ring and the B-R bonds can be prepared if the acceptor pro- C-F vibration region. These result< suggest that a perties of boron could be eliminated or substantially moderately stable borazole derivative containing a reduced e.g. by employing stable boron addition heptafluoro-n-propyl group was present in the re- compounds or borazole derivatives. action mixture. On the other hand heptafluoro- These suggestions are supported by the reactions propyl-lithium with boron trichloride or bromide of heptafluoro-n-propyl-lithium with tri-B-chloro- under the same conditions yielded hexafluoropropene tri-N-phenylborazole and boron halides.Ethereal as the only volatile product no heptafluoro-methyl-lithium was added to a solution of hepta- propane was obtained on hydrolysis of the solid fluoro-n-propyl iodide and tri-B-chlorotri-N-phenyl-residue or the ethereal solution but tetrafluoroborate borazole in the ether at -50" the whole was ions were present in the former. allowed to warm slowly to room temperature and then refluxed to decompose quantitatively any un- We thank Professors H. J. Emelius and A. W. changed hep t afluoro-n-prop yl-li thium the super- Laubengayer for their interest and encouragement natant liquid was decanted from the precipitated during this work.lithium salts and the solvent was removed in vacuo (Received,June 25th 1959.) Goubeau and Rohwedder Annalen 1957,604 168. Pierce McSee and Judd J. Amer. Chem. Suc. 1954 76 474. A Revised Structure for Gibberellic Acid JOHN FREDERICK J. S. MOFFATT, By B. E. CROSS GROVE,J. MACMILLAN T. P. C. MULHOLLAND, and J. C. SEATON (LMPERIAL LNDUSTRIES LIMITED LABORATORIES, CHEMICAL AKERSRESEARCH WELWYN, HERTS) and N. SHEPPARD (UNIVERSITY LABORATORY, CHEMICAL CAMBRIDGE) THE&basic acid C19H2,0,,H20 obtained by opening followed by methylation did not yield methyl gib- the lactone ring of gibberellic acid was shown by berellate but gave an isomer of m.p.174" also ob-0 tained directly from methyl gibberellate by the action of 0.01N-sodium hydroxide and believed to bs the 2-epimeric alcohol by analogy with the 2-epimerisa- tion of methyl a-dihydrogibberellate proved to take place1 under identical conditions. Further work has now shown that gibberellic acid has structure (I; R = H) and in agreement with the intuitive sugges- tion of Japanese workers2 (although not with their structural arguments3) a rearrangement occurs during the formation of the acid (11; R = H). Ozonolysis of methyl gibberellate gave a ketol Cross et aZ.l to have structure (11; R = Ii) arid it was C1gH2,07(IV) m.p. 229-231 O (decomp.). This with concluded that gibberellic acid had structure (ID; chromic oxide and the ester of the derived acid R = H).Relactonisation of the acid (11; R = H) C,,H2,0,,H,0 (V) with manganese dioxide in Cross Grove MacMillan Mulholland and Sheppard PI-oc. Chem. SOC., 1958 221. Takahashi Seta Kitamura and Sumiki Bull. Agric. Chem. SOC.Japan 1957 21 327. Idem ibid. 1958 22 432. OCTOBER 1959 0 Me (VIII) chloroform gave ketones C19H2007,Amas. 229 mp (E 7050) and C2,H2,O, Amax. 229 mp (E 7500) respectively which must from their filtraviolet spectra contain the grouping -CH =CHCO-. These results prompted a re-investigation of the oxidation of methyl gibberellate by manganese dioxide. Al- though a manganese dioxide preparation which oxidised the ester (II; R = Me) to the ketoll C2IH&7,H2O [Amax.240 mp (E 17,OOO)l left methyl gibberellate unchanged the latter compound was oxidised by a more active preparation to an @-unsaturated ketone C~oH2~06, m.p. 186-1 87" Amax. 228 mp (E 9700). Hydrogenation of this ketone gave a saturated ketone (VI) CZ0H26O6 m:p. 129-131" which was also obtained together with the previously described4 8-epimer m.p. 160.5-162.5" by chromic oxide oxidation of the mixed methyl tetrahydrogibberellates resulting from hydro- genation of methyl gibberellate. Therefore no re- arrangement takes place during the formation of the ketone C2,)H2@6 from methyl gibberellate which must be an allylic alcohol. This evidence together with the earlier work1 which established the position of the secondary hydroxyl group in ring A is quite inconsistent with the ring A structure (VII) advanced by the Japanese workers3 and leaves only two possible structures for gibberellic acid namely (I; R = H) and (VUI) in which the carbon atom of the y-lactone is attached to ring A at 4a.Structure (VIII) cannot accommodate the formation in dilute acid of gib- berellenic acid CI9H2,O, A,, 253 mp [E 21,4001 (cf. ref. 5) which must be the heteroannular diene (IX;R = H) from its ultraviolet absorption and from 303 the manganese dioxide oxidation of its ester (IX; R = Me) to a dienone C21H2406,m.p. 134-136" Amax. 309 mp (E 16,500). Gibberellic acid must there- fore have structure (I; R = H) and oc-dihydrogib- berellic acid the ring A structure (X). 0 ,c ,CCH,-C-C'c (xu) the I -0 C-~H-CH=CH (XI) Co2Me 0 (XIID Treatment with collidine of the toluene-p-sulphon- ate of the keto-ester1 C20H2606, m.p.226" obtained from methyl a-dihydrogibberellate with acid gave the anhydro-compound o, m.p. la" a reaction consistent with structure (X)for ring A of the keto- ester but impossible with the structure derived from (In;R = Me). The ester m.p. 174" (vmax. in chloroform 1768 1721 cm.-l) was not oxidised by manganese dioxide and is considered to have structure (HI;R = Me) previously put forward for methyl gibberellate. Strongly marked differences between the hydrogen nuclear magnetic resonance spectra of methyl gib- berellate and the isomer of m.p. 174" and of their monoacetyl derivatives also provided evidence that these esters are not epirners.The nuclear magnetic resonance spectrum of methyl gibberellate was used previously1 to show the presence of the grouping (XII) but at that time we were not able to explain in detail the pattern of lines in the olefinic region which are however immediately explicable in terms of the new structure (I;R = H); in particular the spectrum provides strong evidence for the presence of the three-hydrogen system (XLTI). (Received,June 18th 1959.) Cross Grove MacMillan and Mulholland Chem. and Znd. 1956 954. Gerzon Bird and Woolf Experientia 1957 13 487. Internuclear Distances in Gaseous Silver Halides By R. F. BARROW and C. V. WRIGHT E. MORGAN CHEMISTRY OXFORD (PHYSICAL LABORATORY UNIVERSITY) SOMEof the excited states of the gaseous silver a step towards their interpretation we have photo- halides appear to show unusual feat~res,l-~ and as graphed and analysed the rotational structure of Mulliken Phys.Rev. 1937 51 310. Metropolis ibid. 1939 55 636; Metropolis and Beutler ibid. p. 1113. Barrow and Mulcahy Proc. Phys. SOC.,1948 61 99. PROCEEDINGS several of the bands of the B-X systems of silver chloride and silver iodide. The spectra were photo- graphed in absorption in the fourth order of a 6.5 m. grating instrument; the resolving power was at least 400,000. The structure of the silver iodide bands was well resolved and showed four branches R and P of lo7AgI and loDAgI.The silver chloride bands are much more complex in appearance and in the cor- responding bands of silver bromide there is so much overlapping that a successful analysis probably re- quires the use of an isotopically enriched sample.The essential results are given in the Table. Molecule State Be ae 1080e re(A) 107A$5Cl Pl7,+ 0.1225 0-00160 11 2*2U3 XIC+ 0.1264 0.0007 8.5 2.249 lo9AgI B3n,+ 0.04028 0.00056 2 2.672 X1c+ 0.04442 0-00013 l. 2.544 Internuclear distances in the gaseous diatomic -halides do not obey any simple additivity principle?^^ even with an electronegativity correction but the distance in silver bromide can be estimated as follows. The ratio { r4MBr) -re(MCl)/{ re(MI) -re(MC1)) is a constant equal to 0.404 f. 0.0015 both for the alkali halides4 and for the monohalides of gallium indium and tha1li1.m.~ Thus we estimate re“(AgBr) = 2.36 A.It is thence possible to calculate the entropy of the gaseous silver halides and thus to determine from the vapour-pressure measurements their latent heats of sublimation by a third-law procedure. These quantities are all that are now required for the thermochemical determination of the dissociation limits for the gaseous molecules. (Received September loth 1959.) Honig Mandel Stitch and Tomes Phys. Rev.,1954 96 629. Barrett and Mandel ibid. 1958 109 1572. Heterolytic and Homolytic Fission of S-S and S-Cl Bonds By R.G. R.BACON,R. G. GUY R. S. IRWIN and T. A. ROBINSON DEPARTMENT UNIVERSITY NORTHERN (CHEMISTRY QUEEN’S BELFAST IRELAND) BATEMAN and his co-workers1 have discussed the conditions associated with heterolysis or homolysis of disulphide bonds and have concluded that forma- tion of radicals is restricted to reactions involving photo-initiation high temperature or attack by other radicals.Some results of our investigations with thiocyanogen NC-S-SCN are relevant to this discussion. The reagent contains an exceptionally reactive disulphide bond. It is best known as a nuclear thiocyanating agent for aromatic com-pounds,2 and the observed features of this reaction suggest that it is a heterolytic process like nuclear substitution by molecular halogens. On the other hand reports that some reactions of thiocyanogen are promoted by light3 suggest the possibility of homolysis. We have studied the behaviour of thiocyanogen with several olefins in organic solvents and find that this reaction is strongly promoted by light is in- fluenced by the presence of peroxides and may lead not only to a dithiocyanate (I) by addition but also to an isothiocyanate (11) by allylic substitution I II I I I (SCN) + -C C*CH- -+ -C(SCN)*C(SCN).CH-(1) I II or -C:C.C(NCS)-(It) + HNCS The relative proportions of products (I) and (11) vary with the structure of the olefin e.g.cyclohexene Bateman Moore and Porter J. 1958 2866. gives types (I) and (11) to an equal extent but 1 -methylcyclohexene gives largely type (II) and oct- l-ene gives almost exclusively type (I). The products are quickly formed at room temperature in yields of e.g.60-95%. We have also studied numerous examples of a reaction which readily occurs between thiocyanogen and aralkyl hydrocarbons in organic solvents with ultraviolet irradiation. Provided that an a-hydrogen atom is present it is replaced to give high yields of a benzyl-type thiocyanate (HI) or isothiocyanate (IV) I I (SCN) + ArCH +-Arc-SCN (111) I I I or Arc-NCS (IV) + HNCS I The tendency of compounds (IV) rather than (LII) to be produced is associated with secondary a-carbon atoms and still more sowith tertiary a-carbon atoms particularly when more than one aryl substituent is present at the reaction centres of the hydrocarbons. We suggest that these reactions with olefins and aromatic hydrocarbons involve the thiocyanate radical the relatively easy formation of which may be attributed as in the case of cyanoalkyl radicals to resonance stabilisation *S-C=N t-f S=C=N* Wood “Organic Reactions,” Wiley New York 1946 Vol.111 p. 240; Soderback,Acta Chem. Scand. 1954,8,1851. a E.g. Soderback,Annalen 1925 443 142. OCTOBER 1959 The behaviour of the reagent resembles that of N-bromosuccinimide or of irradiated chlorine or bromine and provides another example of the marked pseudohalide character of the thiocyanate group. Some resemblances are found between reactions of S-S and S-Cl bonds. Like disulphides sub phenyl chlorides are known to react by homolytic as well as heterolytic mechani~rns.~ In previous papers5 we have described some heterolytic reactions of the S-C1 bond of thiocyanogen monochloride Cl-SCN.Later experiments have shown that this compound also readily undergoes homolytic reactions in organic solvents when irradiated. Aralkyl hydro- carbons may thus be thiocyanated in a reaction analogous to that of thiocyanogen I I CI-SCN + ArCH -+ Arc-SCN (or .NCS) + HCI 1 1 However thiocyanogen chloride unlike thiocyano- gen is sufficiently reactive to cause thiocyanation of the nucleus of most aromatic hydrocarbons! The reaction with thiocyanogen chloride may be directed to side chain or nucleus by the right choice of condi-tions. For example 1-methylnaphthalene gives ex- clusively 1-methyl-4-thiocyanatonaphthalene in ace- tic acid in darkness gives exclusively 1-thiocyanato-methylnaphthalene in irradiated carbon tetra-chloride solution and gives mixtures of the two in varying proportions in daylight.Thiocyanogen trichloride for which we have suggested' the structure Cl.S.CCl:NCI is a more complicated case. Its numerous reactions include light-promoted substitution of toluene which we likewise ascribe to homolysis of the S-Cl bond. (Received July 8th 1959.) Kharasch,J. Chem. Educ. 1956 33 585. Angus and Bacon J. 1958 774; Bacon and Jrwin J. 1958 778. 1 Bacon and Guy unpublished results. Bacon Irwin Pollock and Pullin J. 1958 764. Oxidations by Argentic Picolinate By R. G. R. BACONand W. J. W. HANNA (QUEEN'SUNIVERSITY NORTHERN BELFAST IRELAND) SILVER nitrate reacts in aqueous solution with an alkali-metal persulphate and picolinic acid giving an orange-red crystalline precipitate of argentic picolinate,l (C5H4N.C02),Ag.Other argentic com-plexes are known e.g. AgBd2+S20,2-in which B is a heterocyclic base. We selected the picolinate for use in oxidation studies because the reactions of a bivalent silver compound may thus be examined without complications due to oxidations involving persulphate ion. We have used the reagent as an oxidant for organic compounds at 0-90" in a stirred aqueous suspen- sion. During reaction the insoluble orange argentic picolinate gives place to the white argentous com- pound. When the reaction is a dehydrogenation it can be formulated 2(C5H4N*COa2Ag+ XH + 2C5H4NC02Ag+ 2C,H,N-CO2H + X ie. 2Ag2++ XH2+ 2Ag+ + 2H+ + X In the presence of water alone the argentic salt is stable at room temperature and requires about 3 hr.at 85-90" for 50 % decomposition. It cannot be assumed that the oxidation rates observed for various compounds necessarily reflect their relative ease of oxidation by bivalent silver. The metal atom is stabilised by the picolinate ligand. groups and oxidation may be preceded by a rate- determining substitution involving dissociation of a metal-ligand bond or association of the metal with a new ligand.2 Amines possess strong ligand-forming properties and some are particularly vulnerable to oxidation giving aldehydes or ketones by hydrolysis of intermediate iminocompounds resulting from dehydrogenation RCHz.NHCH2R +RCH:N*CH,R -+ RCHO + NH,.CH,R NH2CH2R +NH CHR -+ RCHO + NH In cases so far examined yields have been higher for secondary than for primary amines.For example dibenzylamine gave 86% of the theoretical yield of benzaldehyde in 3 hr. at 20" with 2 mol. of argentic picolinate and di-s-butylamine gave 5 1% of butanone under similar conditions; corresponding figures for monobenzylamine and mono-s-butylamine were both -30%. Mercuric acetate a known dehydro-genating agent for other types of amines? was very much less effective than argentic picolinate for amines we examined. The relative ease of oxidation so far found for other types of compounds is phenols > glycols and Barbieri Atti R. Accad. Lincei 1933 17 1078. Cf. Basolo and Pearson "Mechanisms of Inorganic Reactions," Wiley and Sons,New York 1958 ch.4. Leonard and Hauck J. Amer. Chem. SOC.,1957,79,5279. PROCEEDINGS alcohols > carboxylic acids > aromatic hydrocar- bons. Both dehydrogenation and oxygenation effects are observed; the following are typical results (yield quantity of reagent and conditions are given in parentheses) generally attributed to the following reaction which A -+ B (30%; 2 mol.; 1 hr. at 20") provides two possible oxidising species Ag2+ and (CHPh-OH) -+ 2PhCHO SO4- (74%; 2 mol.; 1 hr. at 65") Ag+ + S202-+ Ag2++ SO4*-+ SO4,-PhCH2*OH -+ PhCHO (53%; 2 mol.; 1 hr. at 85") Comparative experiments with argentic complexes PhCH2-CO2H+-PhCHO should therefore be indicative of oxidising charao (30%; 4 mol.; 3 hr.at 85") teristics of bivalent silver alone. PhEt + PhCOMe We acknowledge financial support from the U.S. (14%; days at 85") Army through its European Research Office,and These observations are of interest in connection thank the Midland Tar Distillers Ltd. for a grft of with the well-known catalytic effect of silver nitrate picolinic acid. on oxidations by aqueous pers~lphate.~ This is (Received August loth 1959.) 'Bawn and Margerison Trans. Faraday Soc. 1955,51,925; Bacon Grime and Munro J. 1954,2275; Bacon et al. n the press. The Constitutions of Dammarenolic and Nyctanthic Acids By D. ARIGONI (ORGANISCH-CHEMISCHES DER EIDG.TECHNISCHEN HOCHSCHULE, LABORATORIUM ZURICH) D. H. R. BARTON and R. BERNASCONI (IMPERIAL LONDON,S.W.7) COLLEGE and CARL DJERASSI," J.S. MKLS and R. WOLFF DEPARTMENT STATE DETROIT) (CHEMISTRY WAYNE UNIVERSITY THEphotochemical cleavage of menthone (I) to the from the fact that the methyl ester was readily acid (II) was discovered by Ciamician and Si1ber.l hydrolysed [2% solution of potassium hydroxide in Irradiation of lanostanone (as 111) (1 g.) in aqueous dioxan-ethanol jl:7) for 30 min. under reflux] this acetic acid (9 1) under oxygen-free nitrogen in a acid is formulated as in (IV). J A (VI I) (VIII) Pyrex vessel with a bare mercury-arc lamp gave an This cleavage reaction has been applied to the acid (350 mg.) m.p. 186-188" [a] + 14" (methyl lactone (V) from hydroxydammarenone-I12ssto give ester m.p. 79-83' [aID+ 16").From analogyf and an acid lactone (as IV),m.p. 99-104" [aID + 43' * Present address :Department of Chemistry Stamford University California U.S.A. Ciamician and Silber Ber. 1907 40,2419; 1909 42 1510. Mills and Werner J. 1955 3132; Mills J. 1956 2196. See also Mills Chem. and Znd. 1956 189; Cosserat Ourisson and Takahashi ibid. p. 190; Godson King and King ibid. p. 190; Crab% Ourisson and Takahashi Tetrahedron 1958 3 279; Biellmann Crabb6 and Ourisson ibid. p. 303. OCTOBER which on methylation and reduction with lithium aluminium hydride in tetrahydrofuran afforded the nicely crystalline triol (VI) m.p. 150-152” [a],+ 48”. The same compound was obtained from dam- marenolic acid2 in the following way. Methyl dam- marenolate was oxidised with chromium trioxide in acetic acid to give inter al.the y-lactone (side chain as in V). Hydrogenation over platinum in acetic acid and then reduction with lithium aluminium hydrideas above afforded the same triol (VI) (identified by m.p. mixed m.p. rotation and infrared spectrum). Since the methyl ester lactone contains the grouping :C=CH (ozonolysis) dammarenolic acid must be formulated as (VII; R = H R’ = C3H,). In agree-ment the methyl ketone (VII; R = Me R’ = Ac) (nuclear magnetic resonance spectrum) was also formed in the chromic acid oxidation referred to above. The formulation of dammarenolic acid as (VJI; R = H R’ = C3H5)provides evidence for a new biogenetic mechanism in triterpenoids possibly of the type (VIII) -+ (VII) or equivalent where X is a suitable leaving group.We have found that nyctanthic acid4 is related in the same way to /3-amyrone (IX). Photochemical cleavage of B-amyrone as above gave dihydro- nyctanthic acid (identified by m.p. mixed m.p. rota- tion and infrared spectrum). Since nyctanthic acid contains4 the grouping :C=CH it must be repre- sented as (X; R = C3H5) and its dihydro-derivative Turnbull Vasistha Wilson and Woodger J. 1957 569. as (X; R = C3H7). Photochemical cleavage of a-amyrone gave an isomeric acid (XI) m.p. 170-172” [a],+ 73”. Rotations above refer to chloroform solutions. The compounds showed the expected infrared spectra. We thank Professor Guy Ourisson (Strasbourg) for a generous specimen of the lactone from hydroxydammarenone-I1 (dipter ocarpol) and Dr.J. H. Turnbull (Birmingham) for an authentic specimen of dihydronyctanthic acid. (Received Jury 30th 1959.) Trimerisation of Benzonitrile by Iron Carbonyls By S. F. A. KETTLE and L. E. ORGEL (DEPARTMENT CHEMJSTRY CHEMICAL OF THEORETICAL UNIVERSITY LABORATORY LENSFIELD ROAD,CAMEWDGE) THEreaction between a metallic carbonyl and an acetylenic compound can lead to the substitution,l polymerisation,2 or combination with carbonyl frag- ment~~ of the acetylene. It seems likely that com- parable reactions will occur when the acetylene is replaced by an organic cyanide. When either Fe(CO) or Fe,(CO) (1 part) is refluxed with benzonitrile (5 parts w/w) under nitrogen for several hours and then cooled pale brown crystals of impure triphenyltriazene m.p.233-234” (lit. 230-234”) the cyclic trimer of benzonitrile separate. In a typical experiment 15.6 g. of iron pentacarbonyl and an excess of benzonitrile gave 8.2 g. of the triazene. Replace-ment of the metal carbonyls by the iron carbonyl hydride-acetylene adduct FezClo05H4 first des- cribed by Reppe4 and prepared by the method of Sternberg et aZ.,5 also gives good yields of the trimer. This reaction may be compared with the trimerisation of acetylenes to benzenes in the pre- sence of metal carbonyls.2 We thank the Mond Nickel Co. Ltd. for a gift of iron pentacarbonyl and the Nuffield Foundation for financial support. (Received May 20th 1959.) Hiibel Braye Clauss Weiss Kriierke Brown King and Hoogzand J.Inorg. Nuclear Chem. 1959,9 204. a Raphael “Acetylenic compounds in Organic Synthesis,” Butterworths London 1955 p. 160. Copenhauer and Bigelow “Acetylene and Carbon Monoxide Chemistry,” Reinhold New York 1949 Chap. * Reppe and Vetter Annalen 1953 582 133. ’ Sternberg Markby and Wender J. Amer. Chem. Soc. 1956,78 3621. PROCEEDINGS Electron Magnetic Resonance Study of Free Phenoxy-radicals By J. K. BECCONSALL and GERALD S. CLOUGH SCOTT DEPARTMENT CHEMICAL LIMITED, (RESEARCH IMPERIAL INDUSTRIES DIVISION HOUSE MANCHESTER 9) DYESTUFFS HEXAGON BLACKLEY ADAMS BLOIS and SAND$have reported an 8-line 90 minutes dominated the spectrum. The identifica- electron magnetic resonance spectrum on oxidising tion of this radical (01 = 1.3 gauss) was not possible 4-methyl-2,6-di-t-butylphenol with air in alkaline on the basis of the evidence.acetone. We have confirmed this and fbd that if the supply of oxygen is limited the 8-line spectrum decays and is replaced by a 1:2 1 triplet (CC,= 1.3 gauss). Attempts to identify this second free radical have so far failed. It has been established that it is neither of the radicals (I) and (II) either of which might be expected to be stable end-products on chemical grounds although both do give triplets (I aH = 1.5 gauss; 11 E = 1.6 gauss). FIG. 1. Derivative spectrum obtained on oxidation of 4-menthyl-2,6-di-t-butylphenol with lead dioxide in anhydrous acetone (1) (u After two days the original spectrum was replaced Oxidation of 4-methyl-2,6-di-t-butylphenolin by a ten-line spectrum (Fig.2) which consisted of a alkaline methanol gave initially a complex 15-line basic doublet-quintet splitting (a doublet = spectrum which could not be due to any single 5.9 gauss a quintet = 1.4 gauss) with some dis- radical species. This spectrum again decayed after tortion of one quintet this Wig almost certainly three days to the triplet described above. due to the presence of the above triplet. This radical In contrast to the complexity of the results ob-species is very stable and appears to be capable of tained in alkaline solution when 4-methyl-2,6-di-t- butylphenol was oxidised with lead dioxide in an- hydrous acetone or benzene four triplets were obtained with coupling constants ctH (quartet) = 10.7 gauss 01 (triplet) = 1.8 gauss.This result indi- cates a major interaction of the unpaired electron with three equivalent protons and a smaller inter- action with two. It seems likely that this is the parent phenoxy-radical (III) and this is supported by the fact that 2,4,6-tri-t-butylphenol in which the inter- a4 (Iv) FIG.2. Final derivative spectrum obtained as for Fig.1 acting p-methyl group has been replaced by a t-butyl existing for a long time in benzene in the absence of group gives only a triplet (cc = 1.8 gauss) under the oxygen. A probable interpretation is that it is due to same conditions. This is believed to be due to the the stable radical cv)which has been prepared as a radical (IV). solid by Coppinge? and J0shi3and has recently been The spectrum due to radical (III) changes rapidly reported by Miiller et aZ.4 to give a ten-line electron with time (Fig.1). Almost as soon as it was observed magnetic resonance spectrum. This radical might be a further triplet was seen growing in the centre of the expected to be much more stable than (IJJ since it spectrum at the expense of the quartet and this after cannot be further oxidised to a quinone. Its forrna- Adams Blois and Sands J. Chem. Phys. 1958,28 774. Coppinger J. Amer Chem. SOC.,1957 79 502. Joshi Chem. and Ind. 1957 525. 'Miiller Ley Scheffler,and Manat Chem. Ber. 1958 91 2682. OCTOBER 1959 tion must involve the elimination of a p-methyl group (cf. Cosgrove and Waters5). These results throw some light on the mechanism of the formation of alkyl-bridged compounds on oxidation of phenols.Cook et aL6 have questioned the view7 that the initial point of oxidation is the methyl group in a p-methyl-phenol. The observation of the phenoxy-radical itself would support their conclusion. Dibenzyl and diphenylmethane deriva- tives can be formed from the phenoxy-radical in the following ways :(1) by oxidation of another molecule of phenol by the phenoxy-radical to give a benzyl radical (VI) followed by dimerisation of the benzyl radical; (2) by intramolecular rearrangement of the (v1) phenoxy-radical to a benzyl radical followed by di- merisation; or (3) by simultaneous dimerisation and rearrangement of the phenoxy-radicals. If (1) or (2) were the mechanism involved then appreciable pro- portions of the alternative crossed combination of H* H-H radicals would be expected since phenoxy-radicals are always present in excess [no spectrum from 4-hydroxy-3,5-di-t-butylbenzyl (VI; R = R’ = Bu? has been observed].3w Me OH Me OH In fact compound (VII; R = R’ = But) has not been reported as an oxidation product of 4-methyl- 2,6-di-t-butylphenol and 75% yields of the stilbene- quinone (VIII; R = R’ = But) have been obtained in these laboratories. It is found from the splitting constants for the p-methyl-hydrogen atoms in the radical (TU) that 6.4% of the unpaired-electron density is on these hydrogen atoms by hyperconjugation (IIIa). It seems possible therefore that two phenoxy-radicals may react to form a complex from which all the possible stable end-products can be derived by rearrangement elimination etc.The measurements were made at room temperature with an X-band spectrometer at a microwave frequency at 9160 Mc./sec. The magnetic field sweep was calibrated by measurements on free radicals whose hyperfine splittings have been reported in the literature. The proton hyperfine splitting constants a are subject to error limits of f5%. (Received Jury 9th 1959.) Cosgrove and Waters J. 1951 1762. Cook Nash and Flanagan J. Amer. Chem. Suc. 1955,77 1783. Moore and Waters J. 1954 243. Structure of Diels and Alder’s Quinoline-Acetylenedicarboxylic Ester Adducts. A Contribution to the Cyclodecapentaene Problem By E.E. VAN TAMELEN and G. MILLER P. E. ALDRICH,P. BENDER (DEPARTMENT OF CHEMISTRY UNIVERSITY OF WISCONSIN MADISON U.S.A.) WISCONSIN AMONG the remaining classical problems in organic Alder.l These workers proffered for the initially chemistry is the structural nature of the 2 1 adducts formed “labile” adduct a zwitterionic formula,lc now of acetylenedicarboxylic ester and pyridine or its recognised as theoretically unsound; and for the analogues first prepared and studied by Diels and “stable” form obtained by heating the labile isomer (a) Diels and Alder Annalen 1932 498 16; (b) 1933 505 103; (c) 1934 510 87. the structure (Ia) a reasonable but not necessarily correct representation. In reinvestigating the prob- lem we measured the proton magnetic resonance frequencies of the two adducts (L and M) obtained from quinoline as well as one (N) of the pair derived from quinaldine.lgJc PROCEEDINGS other hydrogen atoms.These results together with the chemistry recorded for the adducts restrict the reasonable structures to the valency tautomers (Ia-b; R’ = H) for (L) and (IIa-b2) for (M)? Structure (IIb) more simply accommodates the nuclear magnetic resonance results in that (i) the Proton magnetic resonance frequencies of adducts in CDCI (40 mc. relative to H,O = 0) Adduct Chemical shift (c.P.s.) ; H ratio; assignment “Labile” yellow from -18 ;1 ;z quinoline (L) -52; 1; x or y -76; 1; x or y -100; 4; benzenoid “Stable” red from -87; 1; Z’ quinoline (M) -116; 1 ;x’ or y’ -152; 1 ;x‘ or y‘ “Labile” yellow from -70; 2; x y 3 proton ABX system JAB= 9.4 C.P.S.Jm = 2.2 C.P.S. JBx= 2.9 C.P.S. 2 proton AB system }Jm = 93. C.P.S. quinaldine (N)* -100; 4; benzenoid f131; 3; methyl (R’) In all the adducts listed in the Table nuclear magnetic resonance peaks ascribable to the hydrogen atoms of ester groups and the benzenoid rings appear in acceptable regions with the expected absorption intensities. The positions and the nature of the spin-spin splitting of the three remaining protons in adduct (L) indicate the presence of three non-equivalent hydrogens in proximity two of which are olefinic; however the splitting pattern of the three corresponding protons in adduct (M) suggests that one of them is in this case isolated from all the appearance of the six hydrogen atoms at -116 -152 -100 to -120 C.P.S.implies the presence of a quinoline ring system and (ii) the chemical shift (-87 c.P.s.) of the isolated proton is consistent with its presence in a system of maleic (or fumaric) ester type. The parallel nuclear magnetic resonance ultra- violet and infrared spectral behaviour of the quino- line adduct (L) and the quinaldine adduct (N) reveals that these substances fall in the same struc- tural class. Because of steric factors the quinaldine derivative (N) cannot possess a planar structure (Ib; R’ = Me) and therefore the corresponding planar representation (Ib; R’ = H) is ruled This conclusion is of some theoretical interest despite the opportunity the still unknown fully con- jugated planar cyclodecapentaene system for which aromatic character has been customarily anticipated does not make its appearance as a stable entity.The authors are grateful to Dr. J. N. Shoolery and his group (Varian Associates) for helpful discussions. (Received July 30th 1959.) *Despite Diels and Alder’s designation neither quinaldine adduct is convertible into the other.lc a Or the remaining dihydropyridine doublebond isomer. Three structures (Ia’ Ira’ and TIb’) corresponding to (Ia Ira and IIb) were entertained previously by R. Woodward and E. Kornfeld (Kornfeld Dissertation Harvard University 1945) for the pyridine adducts. These investigators proposed (Ia’) for the labile adduct while they interpreted the chemical evidence available at that time along with new experiments devised in their own work,as favouring (Ira’) for the stable isomer.An improbable alternative non-planar (Ib) can be ruled out by nuclear magnetic resonance data (Table and else- where) which we intend to discuss in a full publication. The Crystal Structure of Anhydrous Cupric Nitrate By S. C. WALLWORK (THEUNIVERSITY, NOTTINGHAM) THEpreparation and volatility of anhydrous cupric nitrate have already been rep0rted.l The presence of covalently bonded nitrate groups in both the solid and the vapour has been indicated by an infrared spectroscopic study,2 though the two phases showed significant structural differences. A crystallographic investigation of the solid structure is now nearing completion and the main features are clearly discernible.The crystals are orthorhombic space group PmnZ, with a unit cell of dimensions a = 11.12 b = 5.05 c = 8-28 A containing 4 molecules of &(NO&. Fourier projections along the a and the c axis have been refined to the stage where the disagreement factor R between observed and cal- culated OkZ and hkO structure factors is about 0.20 for each set. The structure so obtained is illustrated in the Figure. Eight oxygen atoms are linked to each copper atom two by bonds of length 1-9 A represented by full lines and six by weaker bonds of length about 2.5 A represented by broken lines. The 1.9 A bonds link alternate copper and nitrate groups into infinite chains parallel to the a axis.Three such chains are shown in full (for one repeat distance along a) and four more are represented for clarity by vertical full lines. The chains are arranged parallel to each other in a pseudo-hexagonal manner and are linked side- ways by the longer bonds between copper atoms and nitrate groups. Each of these nitrate groups is slightly inclined to the mean plane perpendicular to the chains and links three copper atoms by means of six bonds. Each copper atom is therefore surrounded in this plane by a distorted and puckered hexagon of oxygen atoms. The bonds of length 1 a9 A are typical of the shorter Cu-0 bonds in previously determined structures in which oxygen is co-ordinated to coppe? and so represent covalent bonds.The longer bonds form an arrangement of metal and nitrate groups similar to that in rubidium uranyl nitrate,4 which has been re- garded as being partly covalent and also in lead nitrate,6 which is ionic.2 In view of this and the rough agreement between the observed distance of 2-5 8 and the sum of the ionic radii the broken lines in the Figure can be taken as representing bonds which are at least largely ionic. The two types of nitrate group in the structure therefore fulfil quite distinct functions. Indeed the structure could well be formulated as [Cu(NO&Inn +,n(NOJ-. Replacement of the nitrate ions by perchlorate ions could account for the existence of a compound CU(NO,)(C~O,).~ This structure does not provide an immediate explanation of the volatility or the blue colour of the compound.It is unique in having nitrate groups forming strong covalent bridges between metal atoms and the volatility is presumably related to this feature but the structure in the vapour phase is needed before the solid-vapour transition can be fully interpreted. It has been generally accepted that the formation of four strong bonds to the copper atom is necessary for the appearance of a blue colour,* but in this case only two oxygen atoms are strongly bonded. Further detail on this structure and work on related structures will be published elsewhere. Addison and Hathaway Proc. Chem. SOC.,1957 19; 1958 3099. Addison and Gatehouse Chem. and Znd.1958,464. Orgel and Dunitz Nature 1957,179,462; Bnmton Steifink,and Beck Acta Cryst. 1958,11 169; Shibataand Sone, Bull. Chem. Soc. Japan 1956 29 852. Hoard and Stroupe Atomic Energy Project Report 1943 A 1229; Zachariasen Acta Cryst. 1954 7 795. Coulson and Lester J. 1956 3650. Hathaway Proc. Chem. SOC.,1958 344. Hamilton Acfa Cryst. 1957 10 103. * Gattow and Zemann Acta Cryst. 1958 11 866. 3~12 PROCEEDINGS Thanks are due to Dr. C. C. Addison for suggesting D. H. Thomas for assistance with the calculations this problem and for encouragement throughout the and to the Council of the Royal Society for a grant work to Dr. W. E. Addison for assistance in the towards the cost of apparatus. earlier stages of the structure determination to Mrs. (Received July 17th 1959.) A Convenient Method for the Preparation of Some Substituted Pyridines By P.F. G. PRAILL and A. L. WHITEAR (QUEEN ELIZABETH LONDON, COLLEGE W.8) Olefin Substituted pyridine M.p. of picrate ~~ 2 3 4 5 6 Iso bu tene __ Me -Me 157" -Me -Et 140 Mesityl oxide -Me -Me 157 -Me -Me 114 2-Methylbut-1 -ene (Mostly) Me Me -Me 107 (+some) -Et -Me 121-1 22 (Mostly) Me Me -Et 137 (+some) -Et -Et (Not isolated) 2-Methylbut-2-ene Me Me -Me 107 Me Me -Et 137 Pent-1-ene Et -Me 122 Pent-Zene Me -Me Me 177.5 a-Me thy 1s tyrene __ Ph -Me 234 (decomp.) DURING some experiments with tertiary butyl alcohol and acetic anhydride-perchloric acid we ob- served the formation of 2,4,6-trimethylpyrylium perch1orate.l Further work has shown that although other tertiary alcohols give similar pyrylium salts it is often better to use the corresponding olefin.Since 1I (R3.CO),0-HCI0 2 NH,.OH the pyrylium salts are readily converted into pyrid- comparable conditions the yields from these ines by aqueous ammonia2 this constitutes a simple materials are not so good. method of obtaining some substituted pyridines. Perchloric acid slowly added to an ice-cooled solution of the olefin in an excess of acid anhydride gives the crystalline pyrylium salt. If the latter is soluble in the reaction mixture it can be precipitated by the addition of ether. The perchlorate can then 1 (R3.CO),0-HCI0 2 NH,.OH be converted into the corresponding pyridine by The scope of the reaction is illustrated by the adding it to aqueous ammonia or better by refluxing Table.In those cases where the pyridine is known it with ammonium acetate in acetic acid. Traces of the products isolated have properties which are in perchlorate are removed by washing the pyridine good agreement with values in the literature. with potassium hydroxide solution. Nowhere is the More extensive studies on the nature and mechan- overall conversion of olefin into pure pyridine better ism of the reaction will be reported elsewhere. than 50% but it is hoped to improve upon this. The reaction appears to be general for suitably One of us (P.F.G.P.) thanks the University of constituted olefins and acid anhydrides. London for a grant from the Central Research Fund. From earlier work3 it was assumed that the reac- We are indebted to Professor H.Burton for his tion proceeded via the unsaturated ketone but under interest. (Received August 5th 1959.) Praill Ph.D. Thesis London 1954; Chem. and Znd. 1959 1123. Baeyer and Piccard Annalen 1911,384 208. Burton and Praill Chem. and Znd. 1954 75 see also Diels and Alder Ber. 1927 60 716. PLATE1. Flash-spectroscopic record of the triplet state of anthracene in hexane solution. [G. PORTER The Triplet State in Chemistry.] (See page 295.) PLATE 3. PLATE2. Duroquinone transients in viscous parafin. Decay of ti iplet naphthalene in (a>hexane (b) viscous parafin. [G. PORTER The Triplet State in Chemistry.] (See pages 297 and 299.) PLATE4.PLATE 5. Photoelectric record of the decay of triplet Demonstration of energy transjkr.from triplet and radical from duroquinone. benzophenone to naphthalene in benzene solution. [G. PORTER: The Triplet State in Chemistry.] (See pages 299 and 300.) Spcctriini ieferred to by J. H. Callomon in his Conimmication “Emission Spectrum of Ionised Nitrous Oxide N20+”(page 313). OCTOBER 1959 313 Emission Spectrum of Ionised Nitrous Oxide N20+ By J. H. CALLOMON RAMSAY LABORATORIES, (WILLIAM AND RALPHFORSTER COLLEGE, UNIVERSITY LONDON) has briefly described a new ultra- 2c-2n(i)case (a) of a linear molecule. The BROCKLEHURST~ violet band-system observed in emission from inter molecular and electronic constants are given in alia the negative glow of a discharge through Table 2.B and D are inertial and centrifugal distor- tion constants; y is a case (6) spin-splitting constant; gaseous nitrous oxide. The bands extend from 3400 to 4200 A and were ascribed to N20+,by analogy with isoelectronic C02+whose spectra are observed under similar conditions. I have been able to photograph bands of this system with high dispersion on a 6-metre Ebert grating spectrograph resolving lines about 0.09 cm.-l apart and to confirm that the emitter is N,O+. The hollow-cathode source and experimental procedure were the same as used previously2 with CS2+.Each band consists of two sub-bands separated by about 133 cm.-l. The positions of the strongest band-heads are as in Table 1.TABLE 1. Xair (A) vvnc (cm.-’) Intensity vl’ Itl’/ 3380-45 29,573.4 4 1 0 96.72 440.2 1345.5 3541.59 58-40 28,227.88 094.50 -I 9 O O 1 126.2 3688.76 706-98 27,101-76,968.4 -_ 1 10 0 1 1103.8 3845.38 25,997-9 3 0 2 64.17 871.5 They are assigned to short upper- and loclcr-state progressions in the stretching-frequency vl with the band of wavelengths 3541-3558 A as origin. There are in addition many other weaker bands which however do not readily fall into simple progressions or sequences. The bands show well-resolved rotational fine-structure (see Plate facing). There is unfortunately much interference from overlying bands of the second positive (Nd first negative (N2+)and 18 (NO) systems which are still strong even at the lowest dis- charge currents and it seemsworth trying to develop a better source before making a complete analysis.p and 4 are id-type doubling-constants; A is the case (a) spin(-orbit) doubling constant. TABLE 2. Rotational constants (mi.-l) Upper state B2C+ B,’ = 0.43287 fO-ooOo4 D,’ = 1-42 x 10-7 f0.1 x 10-7 y’ % 0 (<0*001) Electronic constants (cm.-l) 2L’-2171/2 VO = 28,229.82 2L’-2L!3/2 vo = 28,096-73 Avo = 133.09 A = -132.27 Yo = A,/B,- = -321.46 Too= 28,162-86 = 3.4916 ev dB”1,2 (obs.) = + 0.002619 f0400007 (calc.) = 2B,2/A0 = + 0.002560 Vibrational terms (crn.?) B2C AGi(100) = ’ -1345 X2L! 1119 rtr’ (cf. N20:XrC’ v1 = 1285) Moments of inertia I (c.g.s.u.) N,O+ B2C 64.659 f0.006 x lo-,’ xn 68.024 , 9 This note therefore presents the results of rotational (cf.neutral N20 X1c’ 66.798 1 99 analysis of the (0,O) band only.* The values of B show that removing the outermost The transition is unambiguously of type electron (n) from N20 expands the molecule by * Added in proof (30. ix. 59) The spectrum free from impuritks has recently been described by Horani and Leach4 in another source under conditions of controlled electron-impact. Brocklehurst,Nature 1958 182 1366. Callomon Proc. Roy. SOC.,1958 A 244 220. Tanaka Jursa and LeBlanc J. Chern.Phys. 1958,28 350. Horani and Leach Compt. rend. 1959 248 2196. 314 PROCEEDINGS 1-8% whereas removing instead the next most loose- ly held electron (0)contracts the molecule by 3.3%.This electron is thus considerably more antibonding than the analogous ones in CO and CS (for dis- cussion see ref. 2). This behaviour is similarly reflected in the vibrational frequencies vl. Both values of D are about 20% less than expected from the simple formula 4B3/v12. The value ofp” is reason- able and corresponds to a state in “pure precession” lying 70,000 cm.-l away. However q” is larger than required by simple “pure precession” theory which predicts p/q = Y = -321 compared with the ob- served value of -22. The value of A corresponds to spin-orbit coupling intermediate between that in atomic N and 0 or N+ and O+. The electronic transition energy of 3492 ev agrees well with the difference between the limits of the first two Rydberg series of N,O of 3-45 ev observed by Tanaka Jursa and LeBlan~,~ confirming the X2n state as ground-state of N,O+.The rotational analysis alone cannot decide between 2C+and zC-in the upper state; however if the latter p” for the 2Cstate would have to be negative which could not be in any way simply explained in a ground state. The spectrum appears therefore to be the exact analogue of the well-known B22,+-X217,(i) transitions of CO,+ and CS2+. as expected from electron con- figurations. (Received July Sth 1959.) The 1,4-Reduction of 7-Dehydrocholesterol with Diborane. A New Synthesis of Cholest-6-en-3P-01 MAZUR,MANASSE and FRANZ By YEHUDA NUSSIM SONDHEIMER (DANIEL INSTITUTE INSTITUTE SIEFFRESEARCH THE WEIZMANN OF SCIENCE REHOVOTH, ISRAEL) IT has been shown that diborane reacts readily in an ether with compounds containing isolated double bonds to give the di- or tri-alkylboranes,l which may be converted into the corresponding saturated hydro- carbons2 or a1cohols.l Monoethylenic steroids have been successfully subjected to this the diborane being generated most conveniently in situ by means of lithium aluminium hydride and boron trifluoride in ether.ld We have now found that the conjugated diene 7-dehydrocholesterol (Ia) under the above-mentioned conditi~nsl~ undergoes 1,4-reduction to 5 a-cholest-6-en-3/3-01 (IIa) the two new asymmetric centres being introduced in such a way as to give the thermo- dynamically more stable configurations.The reaction was carried out by adding lithium aluminium hydride H (1)a R=H @)a R= H b R = Tetrohydro-b R=Ac -2-pyrany l in ether to a solution of the cholesterol (Ia) and boron trifluoride in ether.The resulting alcohol (IIa) isolated in 20% yield with m.p. 116-117” [cc], -86” (in CHC13),3 gave the acetate (IIb) m.p. 104-105” [a] -88” (in CHC1.J,3 vmax. (in CS,) 770 738 730 and 702 cm.? (&double bond4). which was obtained in improved overall yield (35 yo) by reduction of 7-dehydrocholesterol tetrahydro- pyran-2-yl ether (Ib) and subsequent acetylation. The structures of the products (IIa) and (IIb) are based on the physical properties (cf. ref. 3) analyses and catalytic hydrogenation of the acetate (IIb) to 5a-cholestan-3P-yl acetate.That diborane is the reducing agent was confirmed by the fact that the acetate (Ub) was obtained though only in 10% yield when diborane (generated from sodium borohydride and boron trifiuoride-ethefl) was passed into an ethereal solution of the ether (Ib). This reduction could not be effected by lithium alumiilium hydride and aluminium chloride6 in ether. We are indebted to Dr. B. A. Hems Glaxo Ltd.. Greenford Middlesex for a gift of 7-dehydro-cholesterol. (Received July 14th 1959.) (a) Brown and Subba Rao J. Amer. Chem. Soc. 1956,78,5694; J. Org. Chem. 1957,22,1136,1137; (b)Brown and Zweifel J. Amer. Chem. SOC.,1959 81 247; (c) Wechter Chem. and Znd. 1959 294; (d) Wolfe Nussim Mar and Sondheimer J. Org. Chem. 1959,24 1034. a Brown and Murray,J.Amer. Chem. SOC.,1959,81,4108. Barton and Rosenfelder J. 1949 2459; Wintersteiner and Moore J. Amer. Chern. SOC.,1950 72 1923; James,, Rees and Shoppee J. 1955 1370. Henbest Meakins and Wood J. 1954 800. Brown and Tierney J. Amer. Chem. SOC.,1958 80 1552. Cf. Eliel and Rerick J. Org. Chem. 1958 23 1088. OCTOBER 1959 315 Catalytic Activity in the First Transition Series By D. D. ELEYand D. SHOOTER (UNIVERSITY OF NOTTINGHAM) THEcatalytic activity of transition metals such as palladium and nickel has been associated with the presence of holes in the electronic d-band of the metal.1.2 This activity is maintained in palladium- gold alloys until the last of the holes has been filled by s-electrons from the gold c0mponent.l Nothing is known about the way activity of the bulk metals (as distinct from metal oxides3) varies across a whole transition series and accordingly we have examined the metals titanium to zinc as evaporated metal films.The films were protected from grease by a liquid-air trap and conversion of para- into ortho- hydrogen was examined at 1.2 mm. pressure of hydrogen over a range of temperature. In the Figure we plot the logarithm of the velocity constant k (molecules cm.-2 sec.-l) at 293”~ as a function of atomic number and it is seen that the activity is remarkably constant apart from the small dip at manganese up to nickel and that it then drops abruptly to copper and zinc. In the same diagram we plot the logarithm of the cohesive energy E(M-M) of the metal and the parallel is quite striking.This result is in accordance with views expressed by Dowden and the significance may be as follows A 72t t I I I I a I I 1 l\P Ti V Cr Mn Fc CO Ni Cu Zn Vuriation of the velocity constant and cohesive energy (sublimation energy) with atomic number.. The activation energy of the parahydrogen con- version by any chemical mechanism (as here) will depend on the bond energy E(M-€3) of the surface hydrogen bond entering the reaction which can be approximated by Paulings equation :596 E(M-H) = +[E(M-M) + E(H-H)] + 23(XM-XJ2 where the electronegativity of the metal X can be estimated from its work function #M (X = 0-355&J.6 The work function for a given metal will vary considerably according to the lattice plane be- ing highest for the planes of closest packing e.g.for tungsten from 4.39ev (1 11) to 5.53 (1 lo) and will also increase with surface ~overage.’.~ Let us postulate that the catalytically most active M-H bonds are those of minimum ionic character. The ionic character of a bond will be zero when tl-e electronegativity of the surface metal atom equals 2.1 the value for the hydrogen atom.g According to the above equation this will require a work function & of 5.9 ev which is much larger than the published value for clean metal surfaces (averaged over the ex- posed planes).1° It may however be approached by the most closely packed planes particularly under conditions of high coverage by hydrogen atoms.If the parahydrogen conversion goes preferentially on this plane for each metal then the electronegativity term (X -X,j2 will be small and changes in the cohesive energy E(M-M) will closely parallel changes in E(M-H) explaining the result shown in the Figure. The apparent frequency factors (at 293”~ and 1.2 mm. pressure) for these evaporated films fall in the order Co Ni Cu (f.c.c.) 10l9 molecules cm.-* sec.-l; V Cr Fe (b.c.c.) Mn (cubic) Ti (hex) 10l6 and Zn (hex) 1Ol3. The difference in the first two groups may be partly due to a greater availability of close-packed planes of high work function in evaporated films of the face-centred cubic (f.c.c.) metals. Zinc is a special case as the very low fre- quency factor accompanies a negligible hydrogen adsorption and a small activation energy the cata- lytic sites being probably only about 1 in lo* of the surface.The idea that the para-ortho-conversion occurs preferentially on the closest-packed lattice plane may also explain the greater frequency factors for metal wires over evaporated films of the same metal.ll (Received August 28th 1959.) Couper and Eley Discuss. Faraday SOC.,1959 8 172. Dowden J. 1950 242. Dowden Mackenzie and Trapnell Adv. Catalysis 1957 9 65. Dowden “Chemisorption,” ed. W. E. Garner Butterworths London 1957 p. 3. Eley Discuss. Faraday Soc. 1950 8 34. Stevenson J. Chem. Phys. 1955 23 207. R. Suhrmann Adv. Catalysis 1955 7 303. Mignolet Rec. Trav. chim. 1955 74 685. Pauling “Nature of the Chemical Bond,” Cornell Univ.Press 1939. lo Michaelson J. Appl. Phys. 1950 21 236. Eley and Rossington “Chemisorption,” ed. W. E. Garner Butterworths London,p. 137. PROCEEDINGS The CrystalStructure of and the Hydrogen Bond in Sodium Hydrogen Diacetate By J. C. SPEAKMAN DEPARTMENT GLASGOW, (CHEMISTRY THEUNIVERSITY W.2) CRYSTALLINE acid salts are formed by many mono- Na -Na * Na triad. More significantly it liesso near carboxylic acids and in a number of them the acidic to a crystallographic two-fold axis that pairs of such hydrogen atom is involved in a short hydrogen bond atoms related by the axis must be participating in that is crystallographically symmetrica1.l The salts a hydrogen bond; and this bond must involve the known to possess this type of structure (but not other acidic hydrogen atom which (formally at least) lies acid salts) also give anomalous infrared spectra:2 the on the axis in the 24-fold position (d).As in some OH-stretching frequency does not appear near its other acid salts the two acid radicals in the formula normal position and instead there is a broad region are equivalent C2H302- and C2H302H cannot be of absorption culminating somewhere in the region differentiated though in this example there is the 700-1400 cm.-l. The crystal structure of sodium peculiarity that the pair are related by an axis instead hydrogen diacetate is of special interest not only of by a centre of symmetry. Each acidic hydrogen because this is almost the simplest of these acid salts atom is associated with two carboxylate groups; each but particularly because it shows the spectral carboxylate group takes part in one hydrogen bond anomaly in an extreme form.and makes contacts with two different triads of The compound crystallises in the cubic system sodium atoms; each triad of sodium atoms makes contacts with twelve different carboxylate groups. with a = 15.9 A and the unit cell belonging to the space group Iu3 (No. 206) contains 24 At the present stage of least-squares refinement NaH(C2&02) molecule^.^ The structure has now (R= 17%) the O..H..O distance is 2.40 A with been determined by a partial three-dimensional an estimated standard deviation of k0.03 A. This analysis (based on some 200 independent reflexions- is one of the shortest hydrogen bonds between nearly one-third of those accessible to copper radia- oxygen atoms to have been measured with any tion) fuller details of which will be published else- accuracy.The presence in the crystal of a high pro- where. portion of such bonds linked together as outlined above so that they may exert a co-operative effect on The 24 sodium atoms lie on three-fold axes parallel one another is presumably to be connected with the to the cube-diagonals 8 of them Na(l) are at remarkable spectrum. When the 0-0 distance in a centres of symmetry in positions (a) of “Inter-hydrogen bond is small enough the proton is likely national Tables,” and 16 of them Na(2) in posi- to be oscillating in a single potential-well centred at tions (c). These atoms thus occur as linear triads its rnid-p~int.~ Should the hydrogen bond in sodium with two Na(2) 3.29 A on either side of each Na(1).hydrogen diacetate be indeed as short as this analysis The other atoms except the acidic hydrogens are in indicates at present it comes into the range where general 48-fold positions (e). One oxygen atom is genuine (and not merely statistical) symmetry may situated 2-41 8 from Na(1) and 2.43 from Na(2) the be expected. trigonal axis producing equilateral triangles of these atoms with their centroids almost exactly midway [Note added in proof:] After further refinement between Na(1) and Na(2). The other oxygen atom of (R= 12$%) the distance O**Ois 2.41 30-018A. the carboxyl group is 2-44A from Na(2) of a different (Received July 27th 1959.) E.g.Skinner Speakman and Stewart J. 1954 180; Bryan and Speakman Acta Cryst. 1957 10 795. Hadii and Novak Nuovo Cimenfo 1955 (X) 2 715. Cf. Wyckoff 2. Krist. 1927 67 91. * Coulson Research 1957 10 156; see also Peterson and Levy J. Chem. Phys. 1958 29 948. Macrolide Antibiotics. Part 1X.l Filipin By BELIGBERKOZ and CARLDJERGSSI (WAYNE UNIVERSITY MICHIGAN, STATE DETROIT U.S.A.2) THErecent communication3 on a partial structure of Our latest analytical data (Found in three different the antibiotic lagosin4 prompts us to record some of laboratories C 63.9 63.9 64.0; H 9-0,9.0 8.9; our experimental results with the related fungicidal 0 26-7 27-0 26.7) are most consistent with a antibiotic fili~in.~ C32H540,0empirical formula (containing 3-4 Part VIII Djerassi Halpern Wilkinson and Eisenbraun Tetrahedron 1958 4 369.Present address :Department of Chemistry Stanford University Stanford California. Dhar Thaller Whiting Ryhage Stallberg-Stenhagen and Stenhagen Proc. Chem. SOC.,1959 154. OCTOBER 1959 C-methyl groups) rather than with the earlier C3oH,oO,o formulation5 (with 2 C-methyl groups). Catalytic hydrogenation under various conditions results in the consumption of 45-5 mol. of hydro- gen which is consistent with the reduction of a con- jugated pentaene chromophore whose presence was already indicated5 by the ultraviolet absorption spectrum. Filipin and perhydrofilipin are not attacked by sodium periodate showing the absence of ap-glycol functions in contrast to the behavioul.4 of lagosin.After opening of the lactone ring5 with alkali per- hydrofilipin consumes one equivalent of periodate with the liberation of acetaldehyde. These results require partial structures (A) or (B) in the antibiotic and this was confirmed by a positive iodoform reaction of perhydrofilipin. QH CH,-CH-F-O-s-(A) 0 I CH,-CH-&IIHO O-C-8 (81 CH;OHI-C-I C H,*OHI -CH -CH -I (C> CH;OH OH (D) Oxidation of perhydrofilipin by nitric acid yields oxalic acetic propionic butyric adipic hexanoic and 2-methylhendecanedioic acid. Structurally the last two acids are the most significant. Since similar oxidation of filipin itself furnishes the same acids with the exception of 2-methylhendecanedioic acid the latter must represent the site of the original pentaene chromophore.When this dibasic acid was not purified rigorously but submitted in the form of its dimethyl ester to mass spectrographf and gas chromatography it was found to consist mainly of dimethyl 2-methylhendecanedioate accompanied by some dimethyl 2-methyldodecanedioate and 2-methyldecanedioate. These fragments can be derived from structure (I) which must be considered solely a tentative expres- sion for filipin. The incorporation of moiety (B) rather than (A) is based purely on negative evidence in that the arrangement (A) might bz expected to yield unbranched C,, and Cll dicarboxylic acids which were not encountered. The polyene fragment is preferred in the location given in (I) because of its ultraviolet absorption spectrum rather than in an altei-native position directly conjugated with the carboxyl function of the lactone.The formation of both 2-methylhendecane- and 2-methyldodecane-dioic acid in the oxidation of perhydrofilipin by nitric acid is readily accommodated by assuming oxidative rupture on either side of the hydroxyl group marked*. Attempts to establish definitely the site of this substituent by manganese dioxide oxida- tion failed because of further attack on the molecule. The position of the hexyl fragment (yielding hexanoic acid) and that of a hydroxyl group in this hexyl side chain indicated in (I) are based on the isolation of hexanal (as the 2,4-dinitrophenylhydrazone)when filipin is treated with potassium or barium hydroxide the aldehyde being produced by retroaldolisation.Filipin contains between 7-8 acetylatable hydroxyl groups and since they are not vicinal the molecular formula requires alternate 1,3-glycol groupings. Structure (I) corresponds to C33H56Q10 while the analytical figures cited at the beginning of this communication are in better agreement with C,,H5,O,,. If the latter is correct then a possible structural feature may be the presence of a fragment such as (C) or (D) in lieu of one of the linear 1,3-glycolsof (I). OH YH I YH3 *I CH3-F-8-C=C-[C=C],-c-C-5 HO' PH YH ?H O=C-G-C-C-C-C-C-C-C-OH Further work is in progress to establish definitely the structure of filipin. We are greatly indebted to the National Heart Institute ofthe U.S.Public Health Service for financial assistance and to the Upjohn Company (Kalamazoo Michigan) for supplies of filipin and for microanalytical services. (Received July 9th 1959.) Dhar Thaller and Whiting Proc. Chem. SOC.,1958 148. Whitfield Brock Amrnann Gottlieb and Carter J. Amer. Chem. Soc. 1955 77 4799; Gottlieb Ammann and Carter Plant Disease Reporter 1955 39 219. We are greatly indebted to Professor Einar Stenhagen of the University of Uppsala for the mass-spectrographic and gas-chromatographic analyses. The Reaction of Disulphur Decafluoride with Chlorine By J. W. GEORGE and F. A. COTTON (MASSACHUSETTS OF TECHNOLOGY 39 MASS.,U.S.A.) INSTITUTE CAMBRXDGE THE literature contains little information on the water and aqueous bases its decomposition in con- chemical properties ofdisulphur decafluoride S,Fl0 tact with fused potassium hydroxide its oxidising except for observation of its lack of reactivity toward properties toward hot metals its reaction with red- 3 18 hot glass,l and its pyrolysis to sulphur tetra- and hexa-fluoride.2 Preliminary examination of the reaction between disulphur decafluoride and chlorine suggests the formation of a sulphur chlorofluoride a hitherto un-known type of compound.When a mixture of the pure gaseous decafluoride and pure chlorine (in excess) was passed through a Monel or Pyrex tube about 40 cm. long heated at various temperatures up to 350" the infrared spectra of the effluent gases showed several strong lines not assignable to any known compounds which might be formed e.g.sulphur fluorides chlorides oxyfluorides or oxy- chlorides or to volatile products of attack upon the Pyrex or Monel apparatus. The gaseous mixture was then passed through 20% aqueous sodium hydroxide to remove substances (e.g. SF, SiF, SOF, etc.) hydrolysable by base after which the new infrared lines remained while those of sulphur decafluoride were absent and those of sulphur hexafluoride were weak or absent. A vapour-density determination on a sample apparently containing no sulphur hexa- fluoride (according to the infrared spectrum) gave a molecular weight of 159 & 5 (Calc. for SF, 146. Calc. for SF,Cl 162.5. Calc. for SF,CI, 179). This datum together with the resistance of the new species to hydrolysis suggests an SF,_,Cl compound probably SF,Cl.Other evidence favouring the formula SF,Cl has been obtained. Mass spectra while very rich show no peaks assignable to fragments containing SCl but several peaks probably due to species containing SCl. The infrared spectrum down to 400 cm.-l has what appear to be fundamentals at 920(1_) 856(11) 706(_L or I\?) 600(1]) and -573(1_) cm.-l and weaker presumably combination and overtone bands at -1500 -1260 950 and 805 cm.-l. A normal co-ordinate analysis of an octahedral AB,C (C, symmetry) molecule has been carried out and the above data appear compatible with expectation PROCEEDINGS for SF,CI whereas it appears unlikely that they can be reconciled with either cis-or trans-SF,Cl,.trans-SF,CI would have only one infrared S-F stretching mode (incompatible with 920 and 856crn.-l bands which can scarcely be other than S-F stretches) while cis-SF,Cl (C2J would be expected to have a far richer spectrum than that observed. The P-Q-R structure of the 865 and 600 cm.-l bands agrees in respect to relative intensities and peak separations with expectation for SF,CI. The micro- wave spectrum is under study by Dr. P. H. Verdier Harvard University; absorption has been found in the region where the J = 5+6 transition would be expected for sulphur pentafluoride chloride. Some tentative discussion of the reaction mechan- km can be given in terms of the following possible schemes d (4 S,Fl* 3 SF + SF .. . (1) SF -+ CI2 -+ SFdCI2 * * . (2) (b) SZFm + CI2 + SFG + SF,CI . * . (3) (c) S,Flo + CI -+ 2SF,CI . . . (4) Mechanisms(a)and (b)are unlikely since the product appears to be the pentafluoride chloride and not the tetrafluoride dichloride. Moreover mechanism (a) may be ruled out since it is known3 that reaction (2) does not occur under the conditions used. Mechan- ism (b) is incompatible with the observation that sulphur hexafluoride does not appear to be produced in the required quantity; rather it appears in about the same molar quantity as sulphur tetrafluoride. Thus we suggest that the mechanism is of type (c) with the rate-determining bimolecular step indicated accompanied by reaction (1). We thank the Research Corporation for financial support and the Pennsylvania Salt Mfg.Co. and the U.S. Amy Chemical Center Maryland for kind gifts of disulphur decafluoride. (Received July 27th 1959.) Hazeldine Chem. SOC. Special Publ. No. 12 1958 p. 320. Trost and McIntosh Canad.J. Chem. 1951,29 508. Cotton and George J. Inorg. Nuclear Chem. 1958 7 397. The Abstraction Reactions of Methylene By H. M. FREY (CHEMISTRY UNIVERSITY DEPARTMENT OF SOUTHAMPTON) IT has seemed probable that in the addition of methylene to hydrocarbons in the gas phase as well as the direct insertion into the carbon-hydrogen bonds there is a radical component of the reaction. The production of ethane in the reactions of methyl-ene with propane and n-butane and of ethane and n-butane in the reactions of methylene with ethylene has been explained' on the basis of the abstraction OCTOBER 1959 of a hydrogen atom by methylene to yield methyl radicals.Recent work by Doering and Prinzbach2 on the reaction of methylene with 2-methyl [l-14C?]prop- 1-ene has shown in the examination of the distribu- tion of 14C in the 2-methylbut-1-ene formed that in the gas phase some "mixing" has taken place corresponding to a radical reaction. In order to determine quantitatively the extent of the abstraction reactions of methylene a detailed study of the reactions of methylene with propane and isobutane has been undertaken. The possible reac- tions in the former system are ic n-butane . . . (1) -+ isobutane .. . (2) 4-+ CH,. + *CH,.CH,CH . . . (3) -+ CH,. + CH,*CH*CH . . . (4) CH,. + -CH,-CH,-CH -+ n-butane . . . (5) CH,. + CH,*<H*CH -+ isobutane . . . (6) CH,. + CH,. -+ ethane . . . (7) 2 *CH2-CH2*CH -+ n-hexane . . . (8) 2 CH,.CHCH -+ 2.3-dimethylbutane (9) CH,.CH,CH + CH,.CHCH -+ 2-methylpentane (10) ProPYl + propyl -+ propene + propane (11 12 13) Methylene was produced by the photolysis of diazomethane and keten in the presence of a large excess of propane at 50". All the products predicted by equations (1)-( 13) were observed. Ethylene was also present formed by the reaction of methylene with diazomethane or keten. On use of a 20-fold excess of propane almost no propene was formed by the reaction of methylene with ethylene. From the analysis of the hexanes formed kI(,/(k&')% 1-9 close to the expected value of 2.By assuming k5/(k7fk8f) = k6/(k74k93)= 2 wewere able to calculate the relative rates of k and k to be 1 :2*35 which means that methylene abstracts a hydrogen atom 7 times more rapidly from a second- ary than from a primary carbon-hydrogen bond. This is similar to rates of abstraction of hydrogen by methyl. Further confirmation of the essential correct- ness of the scheme depicted by eqns. (1)-(13) was obtained from photolyses in the presence of small quantities of oxygen. Oxygen reacts rapidly with alkyl radicals and under these conditions no propene or hexane was observed and further the ratio of normal to isobutane was altered slightly in favour of the normal compound.In order to eliminate the possibility that reactions of the type R-+ CH,:CO -+ RCH,. + CO might occur in the system di-t-butylperoxide was pyrolysed in the presence of a large excess of keten at 180". Only traces of propane were formed. Hence this re- action is unimportant. (It cannot yet be ruled out when diazomethane is used as the precursor for methylene.) In the system methylene-isobutane it was expected that by far the most important radical reaction would be CH, + (CH,),CH -+ CH,. + (CH&C. The isobutyl radicals can add methyl add to each other or disproportionate. The analytical procedure did not permit the detection of octanes; however large quantities of isobutenewere found presumably formed by the disproportionation.This was con- firmed by photolyses in the presence of oxygen in which isobutene was not formed and the ratio of isopentane to neopentane increased considerably. The analytical procedure does not give very precise values for the relative rates of direct insertion and abstraction reactions. The approximate values found for propane are 78% to 22%. Doering and Prim- bach,2 using 2-methyl [1-14Cl]pr~p-l-ene found 8 % of 2-methyl [3-l4C1]but-1-ene which corresponds to some 16% of the reaction proceeding via a radical mechanism in the gas phase. This is a minimum figure since some radicals will be lost by dimerisation and addition to the oleh present. The author thanks the Royal Society for a grant. (Received August 19th 1959.) Frey and Kistiakowsky J.Amer. Chem. SOC.,1957 79 6373. Doering and Prinzbach Tetrahedron 1959 6 24. Existence of the Trifluoromethylmercaptide Ion By F. JELLINEK* (UNIVERSITY CAMBRIDGE) CHEMICALLABORATORY DURING an investigation of addition compounds of reactions in acetone solutions at concentrations of bis(trifluoromethy1thio)rnercuryl it was found that the order of 103~ has been undertaken by conduc- the complex ion Hg(SCF3),I- reacts with I-both in tivity measurements. the solid state and in acetone. A kinetic study of the When a solution of potassium iodide in acetone is * Present address:Laboratorium vaor anorganische chemie der Ryksuniversiteit,. Groningen Netherlands. Jellinek and Lagowski J.,submitted for publication. added to a solution of bis(trifluoromethy1thio)mer-cury in the same solvent the conductivity of the latter solution sharply increases until a molar ratio KI Hg(SCF,) of 1 :1 is reached corresponding to the formation of K+Hg(SCF,),I-in so1ution.l Further addition of potassium iodide however gives rise to irreversible reactions.Potassium fluoride is precipitated from the solution and the conductivity decreases with time. As long as the final atomic ratio 1:Hg in solution is between the limits 1 :1 and 3 1 the decrease of the conductivity with time corres- ponds to a reaction which is of first order with respect to the added quantity of I- in excess of a 1 :1 ratio and of zero order with respect to the concentra- tion of the mercury compound in solution.These observations can be accounted for by the following reaction scheme When the final ratio 1:Hg is between 1 :1 and 2 1 Hg(SCF,),I-+ I-+Hg(SCF,),I,*-. . . (la) Hg(SCF3)JZ2-+ Hg(SCF,)I,-+ (SCF,)-. . . (Ib) When the ratio I :Hg is between 2 1 and 3 3 Hg(SCF,)I,-fI-+ Hg(SCF,)132-. . -(24 Hg(SCF3)132-+ Hgl,-+ (SCF,)-. . . (26) Both reactions (lb) and (2b) are followed by the irreversible reaction (SCFJ -+CSFZ -+ F-. . . (3) Reaction (3) is responsible for the decrease of the conductivity as potassium fluoride is almost in- soluble in acetone and thiocarbonyl fluoride is a non-electrolyte. The observation that the final conduc- tivities of solutions with I :Hg ratios between 1 :1 and 3 1 are virtually independent of this ratio is in accordance with measurements showing the mobi- lities of Hg(SCF3)21- and HgI,- in acetone solution to be almost identical.Reactions (la) and (2a) must be very fast as the rate-determining reaction is independent of the con- centration of the mercury compound in solution. Furthermore it has been found under a variety of conditions that the speed of the rate-determining step in solutions with I Hg ratios between 1:1 and 2:l on the one hand and between 2 1 and 3 1 on the other is always the same within experimental error. This would be very coincidental if reactions (lb) and (B)were rate-determining. Therefore it seems justified to assume reactions (lb) and (2b) to be fast; it is possible that the reaction couples (la) + (lb) and (2u) + (2b) actually occur as single-step substitutions.The above reasoning leads to the conclusion that EmelCus and Lagowski J. 1959 1497. PROCEEDINGS reaction (3) is rate-determining and that the tri- fluoromethylmercaptide ion is present in the reacting solution in appreciable concentration. The half-life of the (SCF,)- ion is highly temperature-dependent (identification of reaction intermediates by infrared absorption studies of the reacting solution was attempted but the heat of the infrared beam was sufficient to increase the reaction rate about ten times) and it is also dependent on the purity of the solvent. When acetone with a specific conductivity of about is used as a solvent the observed half-life at 22"is about 3 hr.Conductivity measurements im- mediately after the addition of potassium iodide to the solution and extrapolated to zero time show the mobility of the trifluoromethylrnercaptide ion in acetone to be similar to that of the iodide ion. When further potassium iodide is added to a one- day-old solution prepared from acetone sohtions of KI and Hg(SCF,) in 3 :1 molar ratio the conducti- vity increases but again declines with time although much more slowly than in the region where I and Hg are in proportions between 1 :1 and 3 :1. At the same time the solution turns yellow owing to the formation of iodine which is probably produced by reaction of thiocarbonyl fluoride with free iodide ions. The final conductivities of solutions with a I :Hg ratio larger than 3 are not very reproducible; they are larger than the final conductivity of a solution with I :Hg = 3 :1 but smaller than would be expected for the addition of I-to a HgI,- solution.Reactions such as Hg(CF,),1,2-+ I-+ H,O -+ Hg(CF,)I3,-+ CHF + OH-. . (4) and Hg(CF,)la2-t I-t H,O -f HgI4*-+ CHF + OH-. . (5) have recently been observed in aqueous solution.2 It seems probable that these reactions are analogous to the substitution reactions of Hg(SCF,),I- with I-. Possibly the CF3- ion which may be expected to pick up a proton from the solvent is an intermediate in reactions (4)and (5). Therefore it is planned to study these reactions in non-proton-active solvents by essentially the same technique as was used to establish the existence of trifluoromethylmercaptide ions.I thank Professor H. J. Emeleus for his hospitality and the Netherlands Organisation for Pure Research (z.w.o.) for a stipend. (Received August 21st 1959.) OCTOBER 1959 32 1 The Structure of Tetraethyl Monothiopyrophosphate and Monoselenopyrophosphate By R. A. Y.JONES,A. R. KATRITZKY and J. MICHALSKI (THE UNIVERSITY LABORATORY and CHEMICAL ,CAMBRIDGE THEINSTITUTE POLITECHNIKA, OF TECHNOLOGY LODZ POLAND) syntheses of the two isomers (I; X = S) position of which agrees well with that (10.6) RATIONAL and (IT;X = S) of thiopyrophosphate esters all yield reported for pyrophosphoric acid.5 The monothio- and monoseleno-compounds each show two widely separated peaks which indicate the unsymmetrical structure (11; X = S Se); moreover the position of one peak agrees well with that for tetraethyl pyrophosphate and the position of the identical products (for summary see ref.1). Raman2 other is lowered by 67.8 and 69.6 chemical shifts and infrared spectra3 indicate the thione structure respectively. This lowering is in excellent agreement (TI; X = S) for thiopyrophosphate and infrared with the lowering of 69 and 71.9 respectively caused spectra the corresponding sellnone one (11; X = by the substitution of P:S and P:Se for P:O in Se) for selenopyrophosphate,4 but it was desirable to compounds (EtO),P X. Nuclear magnetic resonance shifs (p.p.m. relative ro phosphoric acid) 11; R = Et X = 0 f13.8 (Et0)3P0 +O*ga 11; R = Et X = S +14.8 -54.0 (Et0)SPS -68.1" 11; R = Et X = Se +14.7 -55.8 (EtO),PSe -71b a From ref.4. From ref. 6. obtain confirmatory evidence. This has been pro- Chemical shifts are measured in parts per million vided by phosphorus nuclear magnetic resonance relative to 85% phosphoric acid (cf. refs. 5 and 6). spectra. The nuclear magnetic resonance spectra of tetraethyl pyrophosphate and the monothio- and We thank Sir Alexander Todd F.R.S. for his monoseleno-compounds (IT; R = Et; X = 0,S Se) interest. We acknowledge a D.S.I.R. Research grant (measured as pure liquids) are given in the Table. (to R.A.Y.J.). Tetraethyl pyrophosphate shows a single peak the (Received,July 27th 1959.) Schrader Lorenz and Muhlmann Angew. Chem. 1958 70 690. a Michalski Mierzecki and Rurarz Roczniki Chem.1956 30 651. McIvor McCarthy and Grant Canad. J Chem. 1956 14 1819. Coe Perry and Brown J. 1957,3604. Muller Lauterbur and Goldenson J. Amer. Chem. SOC.,1956,78 3557. Van Wazer Callis Schoolery and R. C. Jones ibid. p. 5715. The Raman Spectrum and Structure of Disilyl [180]Ether By D. C. MCKEAN (CHEMISTRY OF ABERDEEN), DEPT. UNIVERSITY R. TAYLOR, and L. A. WOODWARD CHEMISTRY OXFORD (INORGANIC LABORATORY UNIVERSITY) THEvibrational spectrum of disilyl ether (SiH,),O indirectly by a force-field treatment on the assump- has been interpreted by Lord Robinson and tion that the force fields of disilyl ether and trisilyl- Schumbl on the basis of a structure in which the aminearesimilar. Si-0-Si angle a is 180". Ewing and Sutton,2 how- We have now obtained the Raman spectrum of an ever find by electron diffraction a value of about approximately 1:l mixture of (SiH3):80 and 141O for this angle.This lower value is supported (SiH3),la0 prepared from silyl iodide and water en- Lord Robinson and Schumb J. Amer. Chem. SOC. 1956 78 1327; see also Lord Mayo Opitz and Peake Spectrochim. Acta 1958 12 147. a Ewing and Sutton personal communication. a McKean Spectrochim. Acta 1958 13 38. riched in oxygen-18 and purified by shaking it with mercury and subsequent fractional condensation. For a linear Si-0-Si skeleton the frequency v1 of the symmetrical stretching mode would be unaffected by substitution of oxygen-18 for oxygen-16; but if the skeleton is bent there must be an isotopic displace- ment.A number of Raman spectra were obtained of the liquid mixture of (SiH,),l*O and (SiH3)260 excited by mercury 4358 A Toronto arc irradiation and a Hilger E.612 spectrograph (reciprocal dispersion approx. 100 cm.-l per 111111. in the region investigated) being used. They all show that the v1 line (unlike any of the others) is definitely broadened on the side of lower dv as compared with the corresponding line for pure (SiH3)21s0 in specti a photographed under identical conditions. By using a slit-width of about 33 cm.-l the v1 line of pure (SiH3),160 was observed to have a width of nearly 9 cm.-l. For the mixture of (SiH3),Is0 and (SiH3)260 no distinct resolution into two lines was obtained; but it being assumed that the observed extra broadening toward lower dv is due to the superposition of two similar lines one for (SiH3)280 and the other for (SiH3),160 their frequency difference can be estimated as approxi- mately 3-4 cm.-l.This observation provides convincing evidence that the Si-0-Si angle a is less than 180" in the liquid. Its value could be calculated if (a) the fre- quency v2 of the bending mode were established and McKean. unimblished work. PROCEEDINGS (b) the value of the stretch-bend interaction force constant F12were known. The sign of F12can how- ever be predicted to be p~sitive.~ By taking v1 = 606 and v2 = 117 cm.-l (McKean*) the maximum value of a consistent with an isotopic displacement of vl of 34 cm.-l is found to be about 160" for F, = 0.As F12is increased a becomes smaller. The sym- metry force constants for a = 140" are Fll (stretch) = 5.82 x lo5 dynes/cm. FZ2(bend) = 0.33 x ergs/radian2 and F, = 1.20 dyneslradian. These do not appear to be unreasonable values and so the value a = 141" found by electron diffraction cannot be excluded. A small isotopic displacement of v1 could occur for a strictly linear skeleton if the vibrations were sufficiently anharmonic. Rough estimates of the mag- nitudes of the relevant anharmonicity constants affecting vl can be made from the observed width of the v1 Raman line of (SiH3),160 and from the com- bination frequency at 171 1-5 cm.-l observed in the infrared ~pectrum.~ They are too small by a factor of 10 to account for the observed isotopic displacement.Although the Si-0-Si skeleton of disiIyl ether is thus definitely non-linear in the liquid it is remark- able that the departure from linearity for this mole- cule (unlike disilyl sulphide and selenide5) is so small that thevibrational selection rules for a linear skeleton still appear to be obeyed. (Received July 28th 1959.) Ebsworth Taylor and Woodward Trans. Faraday SOC.,1959,55 211. The Oxidation of Isostrychnic Acid By J. A. JOULEand G. F. SMITH (THEUNIVERSITY 13) MANCHESTER TEUBER recently showed1 that oxidation of isostrych- nic acid (I) with peroxides or potassium nitroso- disulphonate gives two products CzlH2003N2 and C21H,o0,N, to which he assigned structures (II) and (III) respectively.The alkylideneindolenine struc- tures were favoured mainly because they explained the catalytic hydrogenation in ethanol to bases (IV) and (V) which contain an aromatic indole nucleus. The final step in the formation of (11) and (III) from the acid (I) was visualised as a rearrangement of (VI) or (VII) involving a sirnultaneous transfer of hydride from C(17) to C(lsl and fission of the C(,)-C(,,,bond. Not only is such a rearrangement very unlikely but simple alkylideneindolenines are known only as salts. We have studied some of the reactions of the Teuber and Fahrbach Chem. Bey. 1958,91 713. Cook and Majer J. 1944 486. salt (a, previously prepared by Cook and hlajer.2 It is stable in acetic acid in which it shows h,,,.350 and 269 nzp (E 7550 and 18,700 respectively); it rt-acts very rapidly with water under controlled condi- tions to give sulphuric acid and 2-methyl-3-(1-hydroxy- 1 methylethy1)indole (prisms from ben-zene) m.p. 105-115" (decomp.) varying with the rate of heating Amax. 291 and 284 hmjn. 249 (E 6300 7100 and 2100 respectively); under uncon- trolled conditions the main product is polymeric ; reaction with methanol gives polymer with indole ultraviolet absorption. All attempts to produce a solution containing the free alkylideneindolenine base for spectral measurements have so far failed. It is thus clear that alkylideneindolenines are too sus- OCTOBER 1959 ceptible to nucleophilic addition and to polymerisa- tion to exist in aqueous or alcoholic solution.This is also Noland's concl~sion.~ We suggest structures (VI) and (VII) for the O3 and 0,bases respectively and explain the fission of the C(,+(161 bond in terms of an aldol type of equilibrium (VI) + (VIII). Such an equilibrium occurs in an ethanolic solution of the demethoxycar- bonylation product of ak~ammicine.~ The only other case of C(,)-C(le) bond fission is that of the conver- sion of $-strychnine into ~trychnone,~ which may well occur by an analogous mechanism. (I) Cop Me (LIlj,R=oH (IX! ,R=OH Structure (VI) for the 0 base is strongly supported by the ultraviolet absorption in dioxan which shows a simple broad band Amax. 269 mp (E 7060). The C=N frequency of 1562 crn.-l (Nujol mull) com- pares with 157 1 cm.-l in the akuammicine indolenine and with 1580 cm.-l in tetrahydro-1 l-methylcar- bazolenine.In ethanol the O3 base has a character- istic indole absorption (Amax. 292 and 285 Amin. 255 mp (E 7200 7700 and 4000 respectively) which shows that in that solvent the ring-opened form (VIIl) or its equivalent predominates. In 60% aqueous ethanol the spectrum changes to Amax. 283 Amin. 257 Ainfl 291 mp (E 7000 6000 and 6500 respectively). This corresponds to a mixture of (VI) and (VIII) for the absorption in the 260 mp region has risen considerably and that in the 290 mp region dropped. It is noteworthy that the akuammicine indolenine shows iiidolenine absorption in ethanol. This great difference between the two must be the result of steric effects in the strychnine derivatives the effect of the two extra rings the y-lactone and the ether ring must be to increase the stability of the indole form (VIII) relative to the indolenine form (W.In alcohol catalytic hydrogenation1 or reduction with potassium borohydride of the 0 base leads by way of attack of (VIII) at C(ls) to the formation of (IV). In dioxan reduction with lithium borohydride at room temperature leads to attack at C(*>to give an amorphous indoline base Amax. 292 Amin. 268 mp (E -3900 2000 respectively) in ethanol Amax. 269 263 256 mp (E -2000) in ethanolic hydrochloric acid. These reductions give chemical evidence for the existence of two readily intercon- vertible compounds one of which corresponds to Teuber's O3base and is the one that crystallises from solution.It is important to note that (a) the alkylideneindolenine structure (II) and a 17-hydroxy- indole base formed by the addition of the elements of water would not be in reversible equilibrium; (b) a 17-hydroxyindole base would not be reduced to (IV) by potassium borohydride;s (c) lithium boro- hydride reduction of (11) might be expected to go to (IV) but certainly not further to an indoline. The peroxidic oxidation of isostrychnic acid to the O3base may thus simply involve dehydrogenation at C(*)to give the indolenine (XI) which probably by way of the tautomeric enamine (XII) is oxidised to (XIIT)' which then lactonises to (VJ). :i (XiiIj HQ We are indebted to Dr.H.-J. Teuber for a specimen of the 0 base and to the D.S.I.R. for a grant (to J.J.) . (Received,August 17th 1959.) Noland personal communication October 1958. Smith and Wrobel in the press. Woodward Brehm and Nelson J. Amer. Cliem. SOC.,1947 69 2250. Cf. Thesing Chem. Ber. 1954 87 698. C' Witkop J. Amer. Clzem. SOC.,1950,72,1428 2312. PROCEEDINGS A Novel Relationship Between the Stability of Certain Metal Halides and the Absorption Spectra of their Complexes with Dithizone By H. IRVINGand J. J. Cox CHEMISTRY OXFORD (INORGANIC LABORATORY UNIVERSITY) THE comparatively selective analytical reagent cations with more than 18 electrons the tendency to dithizone (diphenylthiocarbazide ; 3-mercapto-l,5-form dithizone complexes (TIE > In111 & GaIIr; BirII $ diphenylformazan; HDz) dissolves in organic SbIII; Pb11 > SnII) is largest for those which favour He Ne A Kr Xe Rn Class (a) Class (b) Border region FIG.1.Acceptor atoms in their normal valency states. solvents to give bright green solutions which show ‘cu2+ two absorption bands in the visible region at \ 450 and 620 mp (CCl,). These may be identified with the absorption due to the so-called keto- and enol-forms Ph-NH-NHCS-N:NPh and Ph-NH-N:C(SH).N :NPh which are in mobile equi- 1ibrium.l On the other hand complexes formed with the cations Mn2+ Fe2+ Co2+ Ni2+ Cu2+ Zn2+ Cd2+ Hg2+ Pd2+ Pt2+ Pb2+ Sn2+,Ag+ Tl+ Bi3+ H& NI” ‘,-+\ \ In3+ and Au3+ are also highly coloured and soluble 0 in organic solvents but these solutions only possess L \ \\ a single strong absorption band located between 450 and 550 mp.We have recently found that a number of organometallic ions form dithizone complexes e.g.RHgDz R2TlDz R3SnDz R,SnDz, R,PbDz 209,QK and R,PbDz (R = alkyl or aryl). These complexes FIG. 2. The wavelength of maxiinurn absorption of are red or yellow and again exhibit a single absorp- certain metal-dithizone complexes in carbon tetra- tion band between 430 and 530 mp. chloride plotted against the stability of their 1:1 No successful attempt has yet been made to explain chloride complexes. why dithizone forms complexes with certain metals and with certain metals only Helmuth Fischer ob- a Transition-metal ion. 0 18-electron shell.served that the “dithizone elements” are either transi- 8 18 + 2 electrons. tion metals or belong to the B sub-groups of the For Ni2+ p1 is “very small.” Periodic Table with closed shells of 18 or 18 + 2 (Data from Spec. Publ. No. 7). electron^.^ However it is noteworthy that among K = [MBrn-l]/[Mn+]wr-]. Irving et a!. unpublished work. a Fischer Mikrochemie 1942 30 38. OCTOBER 1959 inert pairs of s-electrons. This may be an important factor in the recent observations that polonium will form a red dithizone complex PoODz soluble in chlorofom,3 and that TeIV will also form a complex atpH < 1.4 In seeking less qualitative relationships between the properties of a cation and those of its dithizone complex we noted that the small number of elements which form such complexes are also those for which the stability of their halide complexes increases in the order F-< C1-< Br-< I- and which show a greater tendency to combine with sulphur- than with oxygen-donors these have been referred to as Class (b) metals in recent discussion^.^ The more numerous Class (a) cations whose halides decrease in stability in the order F-$ Cl-> Br-> I- do not appear to form stable dithizone complexes those formed by borderline elements (see Fig.1) are usually of low stability. Baanall and Robertson. J.. 1957. 509. Since type (b) character is associated with dative n-bonding which should have a considerable and systematic effect on the electronic structure of any associated ligand it seems probable that any effect of the metal in changing the position of the absorp- tion maximum might be related to its tendency to form e.g.halide complexes. A somewhat similar correlation has been noted for various complexes of bivalent ions of the first transition series6 Fig. 2 shows that the shift in the absorption maximum of a metaldithizone complex increases with the strength of the metal-bromide bond. Similar plots are ob- tained for data for chloride or iodide complexes. The complex of HgIr does not conform to the general trend which suggests absorption at a shorter wave- length. The anomaly with NiII is very marked but here the spectrum is atypical in showing a second strong absorption band at 665 rnp (E = 19,200).(Received August 24th 1959.) Mibuchi Bull. Chem. Socl Japan 1956 29 842. Ahrland Chatt and Davies Quart. Reviews 1958 12 265; Carleson and Irving J. 1954 4390; Irving General Lecture at the Fourth International Conference on Co-ordination Chemistry Chem.SOC.Special Publ. No. 13 inthe press. Irving and Williams J. 1935 3192. A New Plant-growth Promoting Acid-Gibberellin A from the Seed of Phuseotus multifirus By J. MACMILLAN and P J. SUTER J. C. SEATON (AKERSRESEARCH THEFRYTHE,WELWYN, LABORATORY HERTS.) WEhave previously reported the isolation1 of gib- berellin A from immature bean seed (Phaseolus multiflorus). Using the same isolation procedure we have now found a new plant-growth promoting acid (1 mg. per kg. of seed) and as its chemical structure and biological properties are similar to those of the fungal gibberellins we suggest the name gibberellin A,.Analysis of gibberellin A (I) m.p. 260-261" [a]? -77",vmax. 3436 (hydroxyl) 1765 (y-lactone), 1734 (carboxylic acid) 1659 893 (>C=CH,) 1624 694 cm.-l ( .HC = CH. ) and of its methyl ester (11) m.p. 190-191" [a]* -75" vmax. 3670 1748 1724 1656 1624 898 700 cm.-l indicates the mole- cular formula C19H2,05 for the acid. (I). On hydrogenation the ester (11) absorbed two mols. of hydrogen and gave a tetrahydro-derivative (111) m.p. 205-207" [a]i5 + 30". The absence of hydrogenolysis products suggested that there was no double bond allylic to the lactone hydroxyl function as in gibberellic acid. MacMillan and Suter.Nafurwiss.. 1958. 45. 46. Treatment of the acid (I) with dilute hydrochloric acid gave a keto-acid (IV) m.p. 263-265" vmax. 1770 (y-lactone) 1740 (five-ring ketone) 1695 cm.-l (carboxylic acid) [methyl ester (V) m.p. 158-160"]. Similarly acid-catalysed rearrangement of a-di-hydrogibberellic acid (gibberellin A,) (VI) yields3 an analogous ketone (VII) indicating that the acids (I) and (VI) have the same C/D ring structure and locat- ing the hydroxyl group in (I). When the toluene-p-sulphonate of the methyl ester of (VII) was boiled with collidine an unsaturated ester C20H2405 m.p. 164" [a]b9-55" identical with the methyl ester (V) was obtained. Hydrolysis gave the acid (IV). Only structure (I) for gibberellin A is consistent with these results.West and Phinney4 recently reported the isolation of two growth-promoting substances bean factor I and bean factor 11 from immature bean seed (P.vulgaris) and thought that the former may be identical with gibberellin A,. Although bean factor Cross Grove MacMillan Moffait Muiholiand Seaton and Sheppard Proc. Chem.SOC., preceding communication. Cross Grove MacMilIan Mulholland and Sheppard Proc. Chem. SOC.,1958 221. West and Phinney J. Amer. Chem. SOC.,1959 81 2424. PROCEEDINGS /@-a!? Hc.,@$$?; co cop3 co cop3 R1 R2 R3 R1 R2 R3 I CH OH H VI CH OH H I1 CH OH Me VIT 0 Me H IV 0 Me Me V O MeH 11 m.p. 250-255 is incompletely characterised it O may be identical with gibberellin A as the two acids have similar melting points and R values (0.5 and 0.54 respectively in butan-l-ol-l~5~-ammo~um hydroxide).However while the infrared spectra of the two acids in potassium bromide discs show some similarities the carbonyl frequencies reported for bean factor I1 (1750 and 1717 cm.-l) differ from those (1769 and 1739 cm.-l) we find for gibberellin A,. All structures are supported by satisfactory analyses; infrared frequencies refer to Nujol mulls unless otherwise stated and optical rotations to (Received August 12th 1959.) Addition of Maleic Anhydride to Benzene By H. J. F. ANGUSand D. BRYCE-SMITH (THEUNIVERSTTY, READING) IT is widely stated that benzene does not undergo the Diels-Alder reaction. The use of unsubstituted aro- matic hydrocarbons as the diene component in this reaction has hitherto been confined to anthracene and more complex polycyclic hydrocarbons.Reac- tion between naphthalene and maleic anhydride has been only barely detected even when carried out under forced c0nditions.l The presence of vinyl sub- stituents greatly increases the reactivity of aromatic hydrocarbons towards dienophiles and the reactions of maleic anhydride with certain substituted styrenes (e.g. isosafrole 1,l-diphenylethylene) constitute the only previously known examples of Diels-Alder type addition to an isolated benzene The reaction between p-xylene and maleic anhydride in the presence of a peroxide catalyst has been shown to involve an a-substitution which leaves intact the aromatic ring.4 We found that benzene and maleic anhydride react to form a stable 1 :2 adduct at 60" under the influence of ultraviolet radiation.By using a water- cooled form of apparatus previously ca. 1 g. of this adduct can be prepared with an irradia- tion time of 18 hr. The adduct is a colourless solid having the remarkably high m.p. 356". It is sparingly soluble in benzene and other common organic solvents including camphor. Carbon and hydro- gen analyses were consistent with the formula C,H,,2C4H,03. The corresponding tetracarboxylic acid forms colourless crystals from water. On slow heating it is converted into the anhydride m.p. 356" without melting but on rapid heating it melts at ca. 280" with frothing resolidifies and finally melts at 330-340" (Found Equiv.77-7. Calc. 773). The ultraviolet spectrum of an aqueous solution was non- benzenoid in character (E;;;;~ = 55). The corres- ponding tetramethyl ester m.p. 134" was readily prepared from the acid methanol and a trace of concentrated sulphuric acid (Found M 356. Calc 366) and per-acid titration gave 1.1 ethylenic bonds. The anhydride decomposed in vacuo at 360" to give maleic anhydride and benzene as the sole volatile organic products (identified spectroscopically and by m.p.) some tar was also formed. The free acid and its tetramethyl ester both provided satisfactory analyses for carbon and hydrogen. We provisionally suggest that structure (I) for the adduct is best in accordance with our present results and that following electronic excitation of a benzene molecule there is an initial 1,Zaddition of maleic anhydride to give (11).This adduct which would normally be expected to revert rapidly to the starting Kloetzel Dayton and Herzog J. Amer. Chem. SOC., 1950 72 273. Hudson and Robinson J. 1941 715. Wagner-Jauregg Annalen 1931 491 1. Bickford Fisher DoIlear and Swift J. Amer. OilChemists' SOC.,1948 25 251. Blair and Bryce-Smith Proc. Chem. Soc. 1957 287. Blair Bryce-Smith and Pengilly J. in the press. OCTOBER 1959 materials we suppose to be stabilised by rapid addi- tion of a further molecule of maleic anhydride to the alicyclic diene system giving (I). Structure 011) would result from initial 1,4-addition and might at first sight be preferred to (11) on certain anaIogies and from considerations of relative ring strains.It is apparent however that an intermediate having structure (111) would much less readily add a further molecule of maleic anhydride. Among other possible considerations any greater steric strain in the forma- tion of 01) might be offset to some degree by the greater retention of conjugation. Examples of 1,2- addition leading to the production of cyclobutane derivatives are known e.g. photodimerisation of stilbene and cinnamic acid,’ and addition of ketens to dienes. The infrared spectra of the anhydride acid and tetramethyl ester have a fairly marked peak within the region 880-920 cm.-l (weakest in the ester). Many cyclobutane derivatives show absorption with- in this region but a few exceptions are known and the structural correlation is rather doubtful.* Benzene is partly isomerised to fdvene under irradiation conditions similar to those in the present work.5 The possibility that the present adduct is a derivative of fulvene is eliminated because its pyrolysis gives benzene totally free from fulvene and by other evidence to be detailed in a future publica- tion.No trace of the present adduct was formed when benzene was heated with maleic anhydride in the dark at 250-300” for 19 hr. carbon dioxide was produced and other products are under examination. We thank Dr. L. Crombie for comments on the infrared spectra and Messrs. Esso Research Ltd. for a maintenance grant (to H.J.F.A.).(Received August 12th 1959.) ’Mustafa Chem. Rev.,1952 51 1. Bellamy “The Infra-red Spectra of Complex Molecules,” Methuen 1958 p. 30. The Nuclear Magnetic Resonance Spectra of meso-and Racemic 2,3-Dibromobutane F. A. L. ANET (DEPARTMENT UNIVERSITY OF CHEMISTRY OF OTTAWA OTTAWA CANADA) IN symmetrical molecules (e.g.,symmetrically substi- tuted ethanes) it is not easy to obtain the coupling constants between protons with the same chemical shift. This difficulty can be overcome by the examina- tion of suitably deuterated derivatives but the preparation of such compounds is not easy. Cohen Sheppard and Turner1 have made use of the 13CH satellites but this method fails if the signal strength is spread over many lines which then become indis- tinguishable from the noise.If the two protons with the same chemical shift are differently coupled to at least one other magnetic nucleus it is known2 that the required coupling constant can be found directly from an analysis of the spectrum. This method de- serves wider application and can be applied with advantage to 2,3-disubstituted butanes but-Zenes and their derivatives etc. As an example we have examined the nuclear magnetic resonance spectra of nzesu-and racemic 2,3-dibromobutane at 60 Mc./sec. with a Varian Model V-4302 spectrometer. In each compound the methyl signal is well separated from that of the 2,3-hydrogen atoms and each signal is symmetrical about its centre except for a very slight skewness of the intensities towards the centre of the spectrum.The rnesu-isomer has 4 re-solved lines corresponding to the methyl groups and 12 for the 2,3-hydrogen atoms. For the racemic isomer these numbers are 6 and 12 respectively. The main problem here is to analyse the eight-spin system. A full treatment of the method used will be given elsewhere but the essence of the general pro- cedure used to simplify the problem is to consider the spectrum of the 2,3-protons as the sum of the spectra of AB or A systems (the usual terminology being used) depending on the total spin of each of the methyl groups. For example when each methyl group has a spin of +3/2 the 2,3-protons form an A system and give rise to a line displaced by 3/2(J12 + J13) from where the line would be if the coupling constants were zero.When one methyl has a spin of +3/2 and the other +1/2 the 2,3-protons form an AB system inwhich the protons have a rela- tive chemical shift of J12+ JI3and are coupled by J2,. The solution to this is well known3 to be a quartet and the positions and intensities of the lines are easily calculated. By taking into account all the spin states with their proper statistical weights the complete spectrum of the 2,3-protons is obtained and is found to consist of 28 lines. Two of these lines have a separation of J12 + J1 and account for 9/32 of the intensity. The spectrum of the methyl groups is found in a somewhat similar fashion. Here the problem is reduced to that of the sum of a number of ABX systems.An exact treatment of the latter Cohen Sheppard and Turner Proc. Cherrr. Sor. 1958 118. McConnell McLean and Reilley J. Chem. Phys. 1955 23 1152. Pwple Schneider. and Bernstein “High-resolution Nuclear Magnetic Resonance,” McGraw-Hill New York 1959. system has been given.4 The calculated methyl spec- trum contains 14 lines of which two (separated by J12+ J13) account for half the intensity six others always have appreciable intensities and most but not always all of the remaining six have negligible intensities. Only six of the lines of the entire spectrum are independent of J23. It is assumed that there is no 1:4 coupling and that the chemical shift (al2) is large compared with any of the coupling constants. The coupling constants were easily obtained by inspection of the observed spectra and of some typical calculated spectra.These parameters are given in the Table. The spectra calculated from these values agree with the observed spectra in every detail except for the previously mentioned slight skewness in in-tensities which is due to the fact that SI2 is not extremely large compared with the coupling con- stants as is assumed in the calculations. Coupling constants (J in c.P.s.) and the relative chemical sh$t (S, in c.P.s.) for 2,3-dibromobutane isomers. The signs of the J's are indeterminate. Isomer 6, J12 J23 meso 142 i1 6.5 + 0.1 7.7 i0.1 0.0 J1 f 3 0.2 Racemic 161 f1 6.7 i0.1 3.3 f0.1 0.0+ 0.2 PROCEEDINGS Thus J23 is much larger for the mexo than for the racemic material and this no doubt reflects the different populations of the various conformations of each isomer and the angular dependence of coupling constants in ethane derivatives.One would expect that the most stable conformation of the meso- and the racemic isemer would have the bromine atoms trans and hence the 2,3-hydrogen atoms rrans and gauche respectively. These coupling constants agree with those found by Cohen et a1.l for dioxan (Jt,,, 94; Jsauche 2.7) by Lemieux et aL5 for carbo- hydrate and cyclohexane derivatives (.It,,, -7; Jgauche -3) and with Karplus's calculations.6 Other 2,3-disubstituted butanes are being studied and the measurements are being extended to low tempera- tures. The spectra of cis- and trans-but-2-ene have been analysed by the above method but the results are not completely satisfactory because of appreci- able 1:4-coupling.Calculations taking this into account are being undertaken. However the present results strongly point to JI3being of differentsign to J23 for these butenes as in other olefinic corn- pound~.~J*~ (Received August 24th 1959.) Fessenden and Waugh J. Chem. Phys. 1959,30,944. Lemieux Kullnig Bemstein and Schneider J. Anwr. Chem. SOC.,1957 79 1005. Karplus J. Chem. Phys. 1959 30 11. Elvidge and Jackman Proc. Chern. SOC.,1959 89. Alexander J. Chem. Phys. 1958 28 358. Peroxides formed during the Slow Combustion of n-Heptane By R. LONGand J. R. TODD (DEPARTMENT THEUNIVERSITY BIRMINGHAM) OF CHEMICAL ENGINEERING CARTLIDGE and TIPPER^ recently reported the absence of free hydrogen peroxide from the products of oxidation of n-heptane over the range 250-270".Only heptyl hydroperoxides and a mixture of aldehyde-peroxide compounds (roughly half of the total peroxide) were detected. In work in this Department at slightly higher temperatures (above 290") hydrogen peroxide has been found to be formed. Approximately stoicheiometric mixtures of n-heptane and air were oxidised by a flow method in a Pyrex vessel at temperatures within the range 290-400". The peroxides which were almost en- tirely in the aqueous phase of the products were investigated p~larographically.~~~ Initial results in- dicated a free hydrogen peroxide content of 25 % of the total peroxide (determined iodometrically) at 340° rising to 43% at 400".These values appeared to be low since the polarograms did not reveal hydro- peroxides and since similar work with cyclohexane3 indicated that an appreciably higher proportion of the total peroxide was present as hydrogen peroxide in the products of oxidation of this hydrocarbon. However after the reaction vessel for heptane oxidation had been used for some time the total peroxide concentration decreased and hydrogen peroxide could no longer be detected suggesting that contamination of the surface of the vessel had oc-curred. There was a concomitant increase in the acetaldehyde concentration. The vessel was cleaned with hydrochloric acid (followed by distilled water) and higher peroxide values were then obtained with hydrogen peroxide constituting the only peroxide in the aqueous phase.A noticeable difference in the polarograms of the Cartlidge and Tipper Proc. Chem. Soc. 1959 190. Bruschweiler and Minkoff Analyt. Chim. Acta 1955 12 186. Allen Garner Long and Todd Combustion and Flame 1959 3 75. OCTOBER 1959 aqueous phase of the products before and after acid- treatment was the absence of a wave at E* -1.62 v (referred to the mercury-pool anode) in the latter case. It is suggested that this wave was due to an addition compound between hydrogen peroxide and an aldehyde (possibly butyraldehyde). This would explain why hydrogen peroxide constituted only a part of the total peroxide before treatment of the vessel (although no hydroperoxide appeared to be present) the rest being combined with an aldehyde.This addition compound did not appear in the pro- ducts after acid-treatment of the vessel hydrogen peroxide then constituting the whole of the peroxide present. The marked increase in the formation of hydrogen peroxide after acid-treatment of the vessel can be attributed to the preservation or reflection of HO,. radicals (together with the preservation of H,O,) at the surface as suggested by Cheaney et al.4 It is suggested that part of the hydrogen peroxide identified was formed by the further pyrolysis or oxidation of acetaldehyde; this explains why an in- crease in the concentration of the latter accompanied the absence of hydrogen peroxide.Semenov5 quotes a mechanism for the pyrolysis of acetaldehyde in the presence of traces of oxygen studied by Niclause and Letort CH CHO + 0 -f CH CO + HO CH CHO + HO -t CH CO + H,O the chain being propagated by decomposition of the acetyl radical. Hydrogen peroxide is probably formed from acetaldehyde by such steps during oxidation of heptane at the temperatures used in this work. Under certain conditions when the surface becomes con- taminated surface destruction of HO,. radicals with- out production of hydrogen peroxide (and of hydrogen peroxide itself) probably occurs. Under these conditions some inhibition of the further reaction of acetaldehyde also occurs. (Received September 7th 1959.) * Cheaney Davies Davis Hoare Protheroe and Walsh Seventh Symposium (International) on Combustion 1958 Butterworths London (1959).Sernenov “Some Problems of Chemical Kinetics and Reactivity,” Vol. I p. 262 Pergamon Press London 1958. 0- ERRATUM 7t!J+-CH ’7 Formula (VIII) on p. 170 of Proc. Chem. SOC. [?I5 LH-4+JH& 1959 should read 0- NEWS AND ANNOUNCEMENTS Appointment of Honorary Secretary.-The Council announces with regret that Professor M.J. S. Dewar has resigned as Honorary Secretary on taking up his appointment to a Chair of Chemistry at the Univer- sity of Chicago. Professor K. W.Sykes of Queen Mary College London has been nominated to fill this vacancy until the next Annual General Meeting. Perkin Centenary Trust.-The Trustees invite applications before May lst 1960 for the under- mentioned awards available for the academic year 1960/61.Enquiries should be addressed to The Secretary The Perkin Centenary Trust c/o The Chemical Society Burlington House London W.1 from whom forms of application may be obtained. The Perkin Centenary Fellowship.-This award is offered for one or two years to a graduate for the purpose of higher study of any subject related to the manufacture or the application of colouring matters. It has a value of not less than E600 per annum and is tenable from October 1960 at any university technical college or other institution approved by the Trustees. The Perkin Centenary Scholarship.-Two such awards are offered each for two years starting in October 1960 and renewable at the discretion of the Trustees for one further year to enable candidates employed in an industrial firm or other institution concerned with the manufacture or the application of colouring matters to receive an education at a university or technical college.Each award will have a value of E300 per annum. There is no means test for the award and a successful candidate is not de- barred from receiving the whole or a part of his normal salary from his employers during his tenure of the Scholarship. Perkin Centenary Trust Travel Grants.-Applica- tions are invited from teachers concerned with the study of any aspect of the manufacture or the ap- plication of colouring matters at a university tech- nical college or other institution.The purpose is primarily to assist those for whom grants are not readily available from other sources. Each applica- tion will be considered on its merits. The Trustees however expect that preference will be given to applications from lecturers senior lecturers and readers or the equivalent grades in other institutions wishing to gain experience at a similar institution or in industry overseas. The object of the intended visit must be clearly stated (e.g. study of special tech- niques apparatus or industrial processes ; to assist in some stated research project; or to study some specific educational method) and the applicant will be expected to devote sufficient of this time to that object so that effective study is possible.Grants will be available towards the cost of travel and main- tenance for periods related to the purpose of the visit of from one to three months. XXth Conference of the International Union of Pure and Applied Chemistry.-At the XXth Conference of the International Union of Pure and Applied Chem- istry which took place at the Technische Hochschule Munich on August 25-29th 1959 Professor W. A. Noyes Jr. University of Rochester New York State was elected President of the Union. Other officers of the Union include Sir Alexander Todd (Vice-president) and the following Presidents of Sections who ex oficio hold office as Vice-presidents Professor H. J. EmelCus (Inorganic Chemistry) Professor E. J. King (Biological Chemistry) Profes- sor R.Belcher (Analytical Chemistry) Dr. J. H. Bushill (Applied Chemistry). Sir Charles Dodds is Treasurer of the Union. Perhaps the most important decision taken by the Conference is contained in the resolution in respect of atomic weights “The Council is resolved that the Union tentatively adopt a new scale of atomic weights based on the whole number 12 as the atomic weight (nuclidic mass) of the dominant natural iso- tope of carbon carbon-12 in replacement of the currently used scale based on the whole number 16 as the atomic weight of natural oxygen provided that action is taken by the International Union of Pure and Applied Physics (I.U.P.A.P.) to recom- mend the adoption of the same scale in replacement of the scale of nuclidic masses currently used by physicists which is based on the whole number 16 as the nuclidic mass of the dominant natural isotope of oxygen oxygen- 16.If the I.U.P.A.P. takes such action at its General Conference in 1960 it is proposed that final approval be taken by I.U.P.A.C. at its XXI Conference in 1961.” The Conference devoted considerable time to improving the machinery for the conduct of its business and in this connection resolved that the Bureau of the Union should meet annually that the PROCEEDINGS number of the Commissions should be kept to an absolute minimum and that the membership of the majority of Commissions should be reduced. Various recommendations to increase the income of the Union were also adopted mainly by proposals to increase the scale of annual subscriptions from member countries and to seek support from chemical industry through National Committees.A number of proposals related to the re-organising of the Union to accommodate interests of chemical engineering proposed new divisions of the Applied Chemistry Section and the affiliation of bodies such as the World Petroleum Congress were considered and it was agreed that the Union should appoint a com- mittee to examine these proposals and to report to the next Conference in 1961. The Conference also decided that a journal of the Union should be insti- tuted to report on the activity of the Union and its Commissions as well as to provide a record of symposia sponsored by the Union.The next Conference is to take place in Montreal in Canada on August 2-5th 1961 Ciamician Medal.-At the opening session of the IVth International Meeting on Molecular Spectro- scopy held in Bologna (Italy) on September 7th 1959 the following were presented with the Ciamician Medal Professor R. Mecke (Germany) Professor J. Lecomte (France) Professor A. Mangini (Italy) Dr. H. W. Thompson (United Kingdom) Professor W. Kuhn (Switzerland) Professor A. N. Terenin (U.S.S.R.) Professor D. Marotta (Italy) Professor Cambi (Italy) Professor W. Ciusa (Italy) Professor Nasini (Italy). This Medal had been recent- ly instituted by the University of Bologna to com- memorate the distinguished Italian spectroscopist. The Wain Trust Fund.-The Agricultural Research Council has announced that by the generous bene- faction of Professor R.L. Wain an income from the patents on his discoveries in the field of selective herbicides has been made available to the Council and has been used to establish the Wain Trust Fund. The objects of the Fund are to facilitate the inter- change of scientists between universities in the United Kingdom and overseas; to provide equipment in universities in the United Kingdom for agricultural research; and to further other schemes of benefit to agricultural research for which the ordinary funds of the Council are not available. Election of New Fellows.-98 Candidates whose names were published in the Proceedings for August have been elected to the Fellowship.Deaths of Fellows.-The death on 14.9.59 was announced of Sir Ian Heilbron D.S.O. F.R.S. a distinguished Past President of the Society. A full obituary notice will be published in due course. We also regret to announce the deaths of the OCTOBER 1959 following Dr. Sydney Thomas Bowden (4.8.59) of University College Cardiff; Sir AIJi-ed Egerton F.R.S. (7.9.59) Professor of Chemical Technology in the Imperial College of Science until 1952; Mr. Johan Ernst Nyrop of Hellerup Copenhagen; Mr. Ernest Pedley (16.7.59) of the North Western Forensic Science Laboratory Preston ; Dr. Walter Edward Stein (September 1959) of Harrow; and Dr. Allan Leonard Whynes (2.6.59) of Felixstowe. Personal.-Dr. A. L. J. Buckle hq left British Insulated Callender’s Cables Ltd.Helsby to take up an appointment as Technical Director of Technicon Instruments Co. Ltd. Dr. J. R. Cross has been appointed Superintendent of the Chemical Laboratories at the University Col- lege of Swansea as from September 16th. Dr. R. N. Cunningham has been appointed a chemist in the State Laboratories Melbourne. Dr. R. W. M. D’Eye has been appointed Research Manager of the Chemistry Department U.K.A.E.A. Industrial Group Capenhurst Cheshire. Mr. L. Fletcher has retired from his post as Chief Chemist with William Younger & Co. Ltd. Edin- burgh after nearly 51 years’ service with the Company. Dr. K. J. Gallagher has been appointed Lecturer in Inorganic Chemistry at the University College of Swansea as from October 1st.Dr. R. F. Hudson of Queen Mary College London has been appointed a group director of one of the first research units at the Cyanamid European Research Institute in Geneva. Dr. L. A. Jordan relinquished his appointment as founder-director of the Research Association of British Paint Colour and Varnish Manufacturers on October 6th. To commemorate his 33 years of service it is proposed that extensions to the Paint Research Station Teddington which are now being built shall be called the Jordan Laboratory and a suitable ceremony will take place probably early in 1960. Dr. J. F. King has been appointed Assistant Professor in the Department of Chemistry Univer- sity of Western Ontario London Canada. Dr.P. de Mayo has been appointed Professor of Organic Chemistry in the University of Western Ontario London Canada. Dr. N. J. L. Megson has been awarded the Gold Medal of the Plastics Institute for “a paper of out-standing merit” on “Modern Ideas on Polymer Formation”. Dr. A. C. Monkhouse takes office this month as President of the Institute of Fuel. Dr. R. T. Parker who was recently appointed Head of the Banbury and Geneva offices of Alumin-33 1 ium Laboratories Ltd. has now been elected a Director and Vice-President of the Company. Dr. A. R. feacocke Senior Lecturer in Bio-physical Chemistry University of Birmingham has been appointed as a University Demonstrator in the Department of Biochemistry University of Oxford and as Lecturer in Chemistry at St.Peter’s Hall Oxford. Dr. J. H. Pryor has been re-elected Honorary Treasurer of the Scientific Film Association for 1959-60. Dr. D. M. C. Reilly Manager Technical Informa- tion and Advertising Chemicals and Plastics Divi- sion Food Machinery and Chemical Corporation has been appointed Manager of the newly created Advertising and Publicity Department. Dr. F. L. Rose (Research Manager at Imperial Chemical Industries Limited Pharmaceuticals Divi- sion) and Dr. J. Honeyman (Head of the Chemistry Section of the Shirley Institute) have been appointed Honorary Reader and Honorary Lecturer respective- ly in the Department of Chemistry The Manchester College of Science and Technology. Both will give a series of lectures to undergraduates and will be able to direct the activities of research students in the Department.Dr. J. B. Stothers has been appointed to a Lecture- ship in the Department of Chemistry University of Western Ontario London Canada. Dr. M. C. R. Symons has been awarded the degree of D.Sc. of the University of London. Dr. James Tay lor of Imperial Chemical Industries Limited is to be Chairman of the Board of the Imperial Aluminium Company formed recently to link Imperial Chemical Industries Limited with the Aluminium Company of America in the manufacture of wrought aluminium products. Professor F. L. Warren Head of the Department of Chemistry and Chemical Engineering at the Uni- versity of Natal Pietermaritzburg has been ap- pointed Vice-president of the South African Chemical Institute.Dr. D. M. S. Wheeler Assistant Professor of Chemistry University of Nebraska has been appointed to a similar position at the University of South Carolina as from September 1st. Dr. John Whetstone has been awarded the degree of D.Sc. of the University of Birmingham for his thesis on the effect on the crystal habit and structure of inorganic oxy-acid salts when small quantities of dyes are added. Dr. Whetstone is Head of the Organic Chemistry Section Research Department Imperial Chemical Industries Limited Nobel Division. Dr. G. W. Wood has resigned from his post as Senior Lecturer in Organic Chemistry at Hatfield Technical College to take up an appointment as Head of the Science Department Bolton Technical College.PROCEEDINGS Sir Walter Worboys who has served Imperial Chemical Industries Limited for 34 years retires from the Board at the end of this month in order to devote more time to his other activities. FORTHCOMING SCIENTIFIC MEETINGS London Thursday November 12th at 7.30 p.m. Meeting for the Reading of Original Papers. To be held in the Rooms of the Society Burlington House w.l. Thursday December loth at 7.30 p.m. Centenary Lecture “Some Recent Advances in Fluorocarbon Chemistry,” by Professor G. H. Cady. To be given in the Rooms of the Society Burlington House W.l. Aberdeen Monday November 16th at 8 p.m. Lecture “Stereochemistry of the Complex Halides of the Transition Metals,” by Professor R.S. Nyholm D.Sc. F.R.I.C. F.R.S. Joint Meeting with the Royal Institute of Chemistry and the Society of Chemical Industry to be held in the University Union. Thursday November 26th. Lecture “Research on Fish Reservation at the Torry Research Station,” by Dr. J. J. Connell. Joint meet- ing with the Royal Institute of Chemistry to be held at Dounreay. Thursday December loth at 8 p.m. Lecture “Chemical. Kinetics in Relation to Large- scale Production,” by Professor K. G. Denbigh. Joint Meeting with the Royal Institute of Chemistry and the Society of Chemical Industry to be held in the University Union. Birmingham Friday November 13th at 4.30 p.m. Lecture “Spectra and Reaction Kinetics of Some Organic Anions,” by Dr.E. Warhurst M.Sc. Joint Meeting with Birmingham University Chemical Society to be held in the Large Chemistry Lecture Theatre The University. Bristol (Meetings will be held in the Department of Chem- istry The University.) Thursday November 19th at 5.15 p.m. Lecture “Fungi as Molecular Architects,” by Dr. W.B. Whalley F.R.I.C. Joint Meeting with the Student Chemical Society. Thursday November 26th at 5.15 p.m. Lecture by Dr. M. F. Perutz. Joint Meeting with the Student Chemical Society. Thursday December 3rd at 6 p.m. Lecture “Beryllium :Production Properties Appli- cations,” by Dr. G. A. Wolstenholme. Joint Meeting with the Chemical Engineering Group the Institute of Metals the Royal Institute of Chemistry and the Society of Chemical Industry.To be followed by Dinner at 8 p.m. Thursday December loth at 5.15 p.m. Lecture “Rockets,” by Dr. J. Black. Joint Meeting with the Student Chemical Society. Cambridge (Meetings will be held in the University Chemical Laboratory Lensfield Road.) Monday November 2nd at 5 p.m. Lecture “Quadrupole Moments,” by Dr. D. Buckingham. Friday November 6th at 8.30 p.m. Lecture “Fast Halogenation Reactions in Solution,” by Mr. R. P. Bell M.A. F.R.S. Joint Meeting with the University Chemical Society. Monday November 16th at 5 p.m. Lecture “Maytenone a Tetraterpene ?” by Dr. T. J. King. Friday November 27th at 8.30 p.m. Lecture “Some Studies on Peptides of Cystine,” by Professor H. N. Rydon Ph.D. D.Phil. F.R.I.C. Joint Meeting with the University Chemical Society.Monday December 7th at 5 p.m. Lecture “Fluorescence of Organic Vapours,” by Dr. B. Stevens M.A. Cardiff Monday November 16th at 5.30 p.m. Lecture “Synthetic Studies in the Vitamin D Field,” by Professor B. Lythgoe Ph.D. F.R.I.C. F.R.S. Joint Meeting with the Student Chemical Society to be held in the Chemistry Department University College Cathays Park. Durham (Meetings will be held in the Science Laboratories The University.) OCTOBER 1959 Monday November 2nd at 5 p.m. Lecture “Recent Advances in Infrared Spectro-scopy,” by Dr. L. J. Bellamy. Joint Meeting with the Durham Colleges Chemical Society. Monday November 16th at 5 p.m. Lecture “Reaction Mechanisms,” by Professor E.D. Hughes F.R.I.C. F.R.S. Joint Meeting with the Durham Colleges Chemical Society. Monday November 30th at 5 p.m. Lecture “Acyl Trifluoroacetates and Related Corn-pounds,” by Professor E. J. Bourne Ph.D. F.R.I.C. Joint Meeting with the Durham Colleges Chemical Society. Correction. The Lecture “Reactions in Liquid Di- nitrogen Tetroxide,” by Dr. C. C. Addison F.Inst.P. F.R.I.C. will take place at 5 p.m. on Monday January 25th 1960 not on Tuesday January 26th as announced (Proceedings September). Edinburgh (Meetings will be held at the North British Station Hotel.) Thursday November 12th at 7.30 p.m. Lecture “Nitrogen Fixation,” by Dr. E. R. Roberts. Joint Meeting with the Royal Institute of Chemistry and the Society of Chemical Industry.Thursday December Sth at 7.30 p.m. Lecture “The Biosynthesis of Porphyrins,” by Pro- fessor A. w. Johnson Ph.D. Sc.D. A.R.C.S. Joint Meeting with the Royal Institute of Chemistry and the Society of Chemical Industry. Exeter Friday November 27th at 5 p.m. Lecture “The Acetylenic Approach to the Synthesis of Natural Products,” by Professor R. A. Raphael D.Sc. Ph.D. F.R.I.C. To be given in the Washing- ton Singer Laboratories Prince of Wales Road. Glasgow Friday November 20th at 4 p.m. Lecture “Reaction Mechanisms,” by Professor E. D. Hughes F.R.I.C. F.R.S. Joint Meeting with the Alchemists’ Club to be held in the Chemistry Department The University. Hull (Meetings will be held in the Organic Lecture Theatre Chemistry Department The University.) Thursday November 12th at 7.30 p.m Lecture “Liquid Crystals in Solutions of Polyeptides and Other Substances,” by Dr.Conmar Robinson. Joint Meet’ng with the Royal Institute of Chemistry. Tuesday December lst at 5 p.m. Lecture “Recent studies in the para-ortho-Hydrogen Conversion,” by Professor D. D. Eley Sc.D. Ph.D. Joint Meeting with the University Student Chemical Society. Irish Republic Friday November 13th at 7.45 p.m. Lecture “The Stereochemistry of Some Metal Ions,” by Dr. L. E. Orgel M.A. Joint Meeting with the Werner Society to be held in the University Chem- ical Laboratory Trinity College Dublin. Leeds Thursday November 26th at 6.30 p.m. Lecture “Chemical and Biogenetical Studies of Emetine,” by Dr.A. R. Battersby. Joint Meeting with the University of Leeds Union Chemical Society to be held in the Chemistry Lecture Theatre The University. Leicester (Meetings will be held at The University.) Monday November 2nd at 4.30 p.m. Lecture “The Organic Reactions of Molecular Oxygen,” by Dr. A. G. Davies. Joint Meeting with the University of Leicester Chemical Society. Monday November 30th at 4.30 p.m. Lecture “Pure Metals,” by Dr. J. C. Chaston. Joint Meeting with the University of Leicester Chemical Society. Liverpool Thursday November 26th at 5 p.m. Lecture “Experimental Methods in Chemical Kinetics,” by Professor J. C. Robb D.Sc. Ph.D. A.R.I.C. Joint Meeting with the University Chemical Society to be held in the Department of Inorganic and Physical Chemistry The University.Newcastle upon Tyne (Meetings will be held in the Chemistry Department King’s College.) Friday November 20th at 5.30 p.m. Bedson Club Lecture “Organic Semi-conductors,” by Professor D. D. Eley Ph.D. Friday December 4th at 5.30 p.m. Lecture “Bridged Rings,” by Professor R. C. Cookson. Northern Ireland Tuesday December lst at 7.45 p.m. Lecture “Oxidation of Organic Sulphides,” by Dr. L. Bateman. Joint Meeting with the Royal Institute of Chemistry and the Society of Chemical Industry to be held in the Department of Chemistry Queen’s University Belfast. Nottingham (Meetings will be held in the Chemistry Department The University unless otherwise stated.) Tuesday November IOth at 5 p.m.Lecture “Alkyl and Aryl Derivatives of Transition Metals,” by Dr. J. Chatt M.A. F.R.I.C. Joint Meetings with the University of Nottingham Chemical Society. Tuesday December 8th at 8 p.m. Lecture “The Study of Knock and Anti-Knock by the Method of Kinetic Spectroscopy,” by Professor R. G. W. Norrish Sc.D. Ph.D. F.R.I.C. F.R.S. Joint Meeting with the Society of Chemical Industry and the University of Nottingham Chemical Society. Thursday December loth at 7.30 p.m. Lecture “Infrared Spectroscopy,” by Dr. L. J. Bellamy. Joint Meeting with the Royal Institute of Chemistry to be held at Nottingham and District Techincal College. Oxford (Meetings will be held in the Inorganic Chemistry Laboratory.) Monday November 9th at 8.15 p.m.Lecture “Nuclear Magnetic Resonance,” by Dr. R. E. Richards M.A. F.R.S. Joint Meeting with the Alembic Club. Monday November 30th at 8.15 p.m. Lecture by Dr. E. Schlittler. Joint Meeting with the Alembic Club. St. Andrews and Dundee (Meetings will be held in the Chemistry Department St. Salvator’s College St. Andrews unless otherwise stated.) Friday November 13th at 5.15 p.m. Lecture “Some Recent Developments in the Por- phyrin Field,” by Professor G. W. Kenner Ph.D. Sc.D. Joint Meeting with the Unversity Chemical Society. Tuesday November 24th at 5 p.m. Lecture “Chemotherapy,” by Dr. F. L. Rose O.B.E. F.R.I.C. F.R.S. To be given in the Chem- istry Department Queen’s College Dundee.Friday November 27th at 5.15 p.m. Lecture “Aromatic Character and the Inorganic Aromatics,” by Professor D. P. Craig. Joint Meeting with the University Chemical Society. PROCEEDINGS Sheffield (Meetings will be held in the Chemistry Department The University.) Thursday November 19th at 7.30 p.m. Lecture “Chemistry in Fruit and Vegetable Can- ning,” by Dr. D. Dickinson. Joint Meeting with the Royal Institute of Chemistry and the University Chemical Society. Thursday November 26th at 4.30 p.m. Lecture “Infrared Investigation of the Structure of High Polymers including those of Biological Interest,” by Dr. G. B. B. M. Sutherland F.R.S. Joint Meeting with the Royal Institute of Chemistry and the University Chemical Society. Southampton (Meetings will be held in the Chemistry Department The University.) Friday November 13th at 5 p.m.Lecture “Chemical Applications of Oxygen-18,” by Dr. C. A. Bunton. Joint Meeting with the University Chemical Society. Friday November 27th at 5 p.m. Lecture “Solvent-extraction of Inorganic Sub-stances,” by Dr. A. G. Maddock. Joint Meeting with the University Chemical Society. Swansea (Meetings will be held in the Department of Chem- istry University College.) Monday November 9th at 5.15 p.m. Lecture “Interhalogen Compounds and Poly-halides,” by Dr. A. G.Sharpe M.A. F.R.I.C. Joint Meeting with the University College of Swansea Chemical Society. Friday December 4th at 3 p.m. Lecture “Fun with Free Radicals,” by Professor D.H. Hey D.Sc. Ph.D. F.R.I.C.,. F.R.S. Joint Meeting with the University College of Swansea Chemical Society. Tees-side Monday November 2nd at 8 p.m. Lecture “Recent Developments in the Chemistry of the Less Common Elements,” by Professor R. S. Nyholm D.Sc. F.R.I.C. F.R.S. Joint Meeting with the Society of Chemical Industry to be held at the William Newton School Norton Stockton-on-Tees. Wednesday December 9th at 8 p.m. Film Show. To be given at Spark’s Cafk High Street Stockt on- on-Tees. OCTOBER 1959 335 APPLICATIONS FOR FELLOWSHIP (Fellows wishing to lodge objections to the election of these candidates should communicate with the Honorary Secretaries within ten days of the publication of this issue of Proceedings.Such objections will be treated as confidential. The forms of application are available in the Rooms of the Society for inspection by Fellows.) Adey Joy Heather B.R. 47 Kirby Lane Kirby Muxloe Leices ter . Bliss Robert Hugh B.A. French Field Solent Avenue Lymington Hants. Bloom Harry MSc. Ph.D. A.N.Z.I.C. Chemistry Department University of Auckland P.O. Box 2553, Auckland New Zealand. Boulton Anthony John B.A. 82 Shaman Cross Road Solihull Warwickshire. Colegate Terence Dudley. Malabar Church Lane, Lexden Colchester Essex. Cooper Robin David Grey B.Sc. A.R.C.S. 90 Great Brickkiln Street Wolverhampton. Cork Celia Winifred. 286 Hungerford Road Crewe Cheshire. Goodland Frank Charles. 29 Tugela Road Uplands Bristol 3. Hart Gavin.13 Princess Street Pennington Park South Australia. Hill Charles Bradlaugh B.R. 51 Grosvenor Avenue North Harrow Middlesex. Koszuta Roman Michal B.Sc. 3 Emanuel Avenue London W.3. Kupchik Eugene John Ph.D. 282 Main Avenue, Wallington New Jersey USA. Lipman Roger David Arnold B.Sc. A.R.C.S. 55 Sutton Court Chiswick W.4. Loveluck Graham David Ph.D. A.R.I.C. Metcalf Research Laboratory Brown University Providence 12 R.I. USA. Mcparlane Neil Robertson. 63 Penylan Road Cardiff. Marshall Derek Lawrence B.Sc. A.R.I.C. 287 Grange Road Plaistow E.13. Merritt Ronald B.Sc. 6 Oakridge Avenue Menston nr. Ilkley Yorks. Molnarfi Alexander Andreas. 52a Buckingham Court Kensington Park Road London W.1 1. Morris Ivor Graham B.Sc. 9 Merlais Street Roath Park Cardiff.Mosher Harry Stone M.S. Ph.D. Department of Chem-istry Stanford University Stanford California U.S.A. Nash Brian Walter B.Sc. 100 Pennine Drive Hendon Way London N.W.2. Pelchowicz Zvi M.Sc. Ph.D. Israeli Institute for Biological Research Ness-Ziona Israel. Pusey David Frederick George B.Sc. 42 Hampden Street Nottingham. Perrin John Henstock B.Pharm. The Hill Shelford Nottingham. Rathinajamy Tholandi Kandasamy M.Sc. Department of Chemistry Annamalai University Annamalainagar South India. Rice Albert Edward MSc. 34 Bostock Avenue, Northampton. Roemmele Michael Cameron B.Sc. 64Kessington Road Bearsden Glasgow. Sheldon John Charles Ph.D. Department of Chemistry University College Gower Street London W.C. 1.Smith Geoffrey Harold B.Sc. 2 Faraday Street Burnley Lancs. Stewart Michael Charles BSc. A.R.C.S. 47 The High- way Sutton Surrey. Straughan Brian Penrose BSc. 44 Dorket Drive, Wollaton Park Nottingham. Svedres Edward Visvilla M.S. Ph.D. Plant Develop- ment Laboratories Smith Kline & French Laborator- ies 1530 Spring Garden Street Philadelphia 1, Pennsylvania U.S.A. Thomas Evan Gower M.S. Chemistry Department University of Maine Orono Maine USA. Walker David Malcolm B.Sc. 140 Woodsley Road Leeds 2. Wedegaertner Donald Keith B.S. 219 Noyes Laboratory University of Illinois Urbana Illinois U.S.A. Williams Leslie Pearce Ph.D. The Lodge Orchard Lea Boars Hill Oxford. Wilkinson John Arthur Elton B.A. 14 Manor Road Sidcup Kent.ADDITIONS TO THE LIBRARY Biographisch-literarisches Hand worterbuch der exak- ten Naturwissenschaften by J. C. Poggendorff. Vol. VIIa. Part 3 L-R. Part 4. Akademie-Verlag Berlin. 1959. High-resolution nuclear magnetic resonance. J. A. Pople W. G. Schneider and H. J. Bernstein. Pp. 501. McGraw-Hill Book Company Inc. New York. 1959. Acid-base titrations in non-aqueous solvents. J. S. Fritz. Pp.47. The G. Frederick Smith Chemical Company. Columbus Ohio. 1952. (Presented by the publishers.) Manual of physico-chemical symbols and terminology. J. A. Christiansen. Issued by the International Union of Pure and Applied Chemistry Physical Chemistry Section Commission on Physico-Chemical Symbols and Termino- logy. Pp. 27. Butterworths Scientific Publications.London. 1959. Physical properties of chemical compounds-11 a systematic tabular presentation of accurate data on the physical properties of 476 organic straight-chain com-pounds. Compiled by R. R. Dreisbach. (Advances in Chemistry Series. No. 22.) Pp.491. American Chemical Society. Washington. 1959. Sadtler Standard Spectra. Midget edition. Sadtler Spec- Finder. 1959. 1st Edition. A11 codings for Sadtler Standard Spectra from 1-14,650. Samuel P. Sadtler & Son Inc. Philadelphia. 1959 Sadtler Midget Edition Spectra 14,201-14,650. Sadtler Alphabetical Supplement 2nd quarter 1959. Sadtler Numerical Supplement 2nd quarter 1959.Sadtler Composite Molecular Formula Index. Sadtler Midget Edition Spectra. Correction Spectra covering Nos.1-3400. Samuel P. Sadtler & Son Inc. Philadelphia. 1959. (Presented by Sadtler.) Separation of the boron isotopes. Edited by G. M. Murphy. Issued by the United States Atomic Energy Cornmission Technical Information Service. (Report de- classified 1957.) Pp. 485. United States Atomic Energy Commission. Technical Information Service Extension. Oak Ridge Tennessee. 1952. Chemistry of carbon compounds a modern compre- hensive treatise. Edited by E. H. Rodd. Volume IV. Part B. Heterocyclic compounds. Pp.656. Elsevier Pub- lishing Company. Amsterdam. 1959. Elsevier’s Encyclopaedia of Organic Chemistry. Series 111. Carboisocyclic condensed compounds. Vol. 14. Supplement. Pp. 221 5s-2990s. Steroids; oxo-com-pounds. Edited by F.Radt. Springer-Verlag Berlin. 1959. Chemistry of Heterocyclic Compounds. Vol. 13. s-Triazines and derivatives. E. M. Smolin and L. Rapoport. Pp. 644. Interscience Publ. Inc. New York. 1959. Tutorial questions in organic chemistry. P. A. Ongley. Pp. 276. University of London Press Ltd. London. 1959. (Presented by the publishers.) Industrial microbiology. S. C. Prescott and C. G. Dunn. Pp.945. McGraw-Hill Book Company Inc. New York. 1959. DDT the insecticide dichlorodiphenyltrichloroethane and its significance; das Insektizid Dichlordiphenyltri- chlorathan und seine Bedeutung. Edited by P. Muller. Vol. 11. Human and veterinary medicine. Edited by S. W. Simmons. Birkhauser Verlag. Basle. 1959. (Presented by the publishers.) High Polymers.Vol. 12. Analytical chemistry of polymers. Part I. Analysis of monomers and polymeric materials; plastics; resins; rubbers; fibers. Edited by G. M. Kline. Interscience Publ. Inc. New York. 1959. Standard methods of analysis of iron steel and ferro- alloys as used by the Laboratories of the United Steel Companies Limited. Pp. 169. The United Steel Com- panies Ltd. Sheffield. 1959. Analytical Chemistry in Nuclear Reactor Technology 1st Conference Gatlinburg Tennessee 1957. Edited by C. D. Susano,H. S. House and M. A. Marler. Sponsored by the Oak Ridge National Laboratory. Pp. 256. U.S. Atomic Energy Commission. Technical Information Ser- vice Extension. Oak Ridge Tennessee. 1958. Proceedings of the Second United Nations International Conference on the Peaceful Uses of Atomic Energy held in Geneva 1958.Volume 17. Processing irradiated fuels and radioactive materials. Pp. 709. United Nations. Geneva. 1958. PROCEEDINGS 31st Congres International de Chimie Industrielle Liege 1958. Compte rendu. Vol. I. (Industrie Chimique Belge 1959. Supplement.) Pp. 809. Industrie Chimique Belge. Brussels. 1959. Theoretical organic chemistry papers presented to the KekulC Symposium London 1958 ; organised by the Chemical Society for the Section of Organic Chemistry of the International Union of Pure and Applied Chem- istry. Pp. 298. Butterworths Scientific Publications. London. 1959. (Presented by I.U.P.A.C. ; 2nd copy presented by the publishers.) Ozone chemistry and technology.Proceedings of the International Ozone Conference held in Chicago 1956. (Advances in Chemistry Series. No. 21.) Pp. 465. American Chemical Society. Washington. 1959. Proceedings of the Fourth International Congress of Biochemistry Vienna 1958; edited by 0. Hoffmann-Ostenhof. Issued by the International Union of Biochem- istry. Volume 14. Transactions of the plenary sessions. Edited by W. Auerswald and 0. Hoffinann-Ostenhof. Pp.299. Pergamon Press. Symposium Publications Divi- sion. London. 1959. (Presented by the publishers.) 7th Colloquium Spectroscopicum Internationale held at Liege 1958; organised by the Association des In- genieurs sortis de I’UniversitC de Lii3ge. (Revue Univer- selle des Mines.) Pp. 388. Association des Inghieurs sortis de 1’Universite de Liege.Liege. 1959. NEW JOURNALS Annali Stazione Chimico-Agraria Sperhentale di Roma 1959 l(1). Electrochimica Acta. An International Journal of Pure and Applied Electrochemistry. 1959,l (1). London. Plastics Today. 1959 No. 1. (Quarterly.) Presented by Imperial Chemical Industries Limited Plastics Division. ICSU Review. 1951,1(1). (Quarterly.) Presented. OBITUARY NOTICES ROBERT BENJAMIN FORSTER 1879-1 959 PROFESSOR died on February loth R. B. FORSTER 1959 at his home at West Moors near Bourne-mouth where he had lived in retirement since 1949. Born in Galway Ireland in 1879 he studied in Dublin where he attended successively the High School St. Andrews College and the Royal College of Science. His main interest became electrical en-gineering and he obtained the A.R.C.Sc.1.and worked for five years in the electrical engineering industry before returning to the Royal College of Science Dublin for further study. After a short period of lecturing at Kingstown and Blackrqk Technical Colleges Professor Forster spent three valuable years on research at the University of Berlin where his work on naphthylamines published in Berichte led to the award of the Degree of Ph.D. in 1911. In the same year he was appointed Senior Demonstrator in Chemistry at University College Galway in which capacity he published three papers in the Journal of the Chemical Society on phototropy and thermotropy relating to aromatic amines. In 1916 Professor Forster joined the staff of British Dyestuffs Corporation Ltd.as research chemist but returned to academic life as Lecturer in Colour Chemistry at Leeds University. In the latter position he published numerous papers including important work on the identification ofnaphthalene-sulphonic acids through their arylamine salts which has proved of great value to students of colour chem- istry; his versatility is shown in publications relating to filtration plants and liquid-extraction apparatus and to identification of dyes on the fibre. At Leeds University Professor Forster is remem-bered as a kindly man with a keen sense of humour ; he was always willing to help fellow members of the staff and numerous undergraduates are grateful for his patience and guidance in the laboratories; his lectures delivered with clarity were punctuated by numerous anecdotes from his wide experience.His sensitivity to chlorodinitrobenzene was so marked OCTOBER 1959 that he could tell if a bottle were open in the labora- tory and he immediately would order its transfer to the fume-cupboard. He had a wide knowledge of a variety of subjects and could often be found in the engineering workshop carrying out precision work or in the dark room testing a new developer; in the Photographic Journal he published papers on the relation of developing function to constitution of certain naphthalene derivatives. His experience in chemical engineering contributed considerably to his great success in establishingan entirely new Department of Chemical Technology at the University of Bombay where he was appointed Professor in 1933; laboratories of his design remain as a concrete example of his vision and industry in this field.He quickly established himself as a first-class organiser but found time to carry out research on a variety of topics. As a successful ambassador for Britain and working with his Indian colleagues Professor Forster published fifteen papers several in collaboration with Professor K. Venkataraman. This profitable range of work included studies on oxycel- lulose and hydrocellulose Naphtol AS derivatives colour measurement on natural cotton mildew on cotton wetting agents and other textile auxiliaries technical gas analysis and fastness standards for coloured cements.After leaving Bombay in 1938 and returning to England Professor Forster was Examiner in the Aeronautical Inspection Department and Technical Adviser in the Textile industry and retired unwilling- ly in 1949. Professor Forster is survived by two daughters Sheelah and Mary. A. T. PETERS. LEONARD JAMES SPENCER I a70-1959 DR. L. J. SPENCER, F.R.S. formerly Keeper of Minerals at the British Museum (Natural History from 1927 to 1935 who died on April 14th aged 88) was a born collector and an indefatigable and inces- sant worker whose curatorial work on the national collections in his department was remarkable for its thoroughness and exactitude. His scientific research was of a high order but his most outstanding achievement was his complete and detailed mastery of the whole of mineralogical literature which he had abstracted and indexed since 1894.He devoted all his life to his work taking few holidays and working excessively long hours. Leonard James Spencer was born at Worcester on July 7th 1870 the eldest son of the late James Spencer for many years headmaster of the day school department of the Bradford Technical Col-lege. After a very successful career at the Royal College of Science Dublin and then at Sidney Sussex College Cambridge he was appointed to an assistant keepership in the department of mineralogy in the British Museum in 1893. Spencer’s scientific papers number over 100 he translated from the German two large quarto works Max Bauer’s Precious Stones (1904) and Reinhard Brauns’s The Mineral Kingdom (1908-1912) and he wrote two books on minerals and on gem-stones which had a wide circulation The World’s Minerals (1 9 1 1) and A Key to Precious Stones (1 936).He con- tributed the articles on minerals to the Encyclopaedia Britannica (eleventh to fourteenth editions) and to Thorpe’s Dictionary of Applied Chemistry. He had abstracted papers on mineral chemistry for the Chemical Society since 1894 catalogued the minera- logical papers for the period 1883-1900 for the Royal Society Catalogue and for 1901-1914 for the International Catalogue of Scient8c Literature. The Mineralogical Magazine was edited by him from 1901 to 1955. His friends contributed articles to an editorial jubilee number of the magazine and a dinner was given in his honour to celebrate his completion of 50 years as editor in November 1950.Minera-logical Abstracts made its first appearance in 1920 and Spencer wrote most of the abstracts and com- piled all the indexes until 1955. His lists of new mineral names which he had published triennially since 1897 were in constant use by mineralogists all over the world by 1958 when the twenty-first list was published they included 3,363 names and synonyms. When he succeeded Prior as Keeper of Minerals in 1927 he carried on his predecessors’ work on the great collection of meteorites making a special study of meteorite craters and of the glass formed by the fusion of the desert sand in the vicinity of the newly discovered craters at Henbury in Central Australia (1931) and Wabar in Arabia (1932).This work led him to investigate other forms of silica glass of sup- posed meteoritic origin and in 1934 he joined an expedition to the Libyan Desert in an endeavour to discover the source of the remarkable masses of yellow silica glass discovered there by P. A. Clayton in 1932. His work received frequent recognition both at home and abroad. He was elected a Fellow of the Royal Society in 1925 became a Fellow of the Chemical Society in 1926 and was awarded the C.B.E. in 1934. He was an honorary member of the Mineralogical Societies of America and Germany and president of the Mineralogical Society of Great Britain from 1936 to 1939.PROCEEDINGS He married in 1899 Edith Mary daughter of lslip J. Close and had one son and two daughters. Mrs. Spencer died in 1954. (Reprinted with permission from The Times.) HARRY DUGALD KEITH DREW 1886-1958 H. D. K. DREWwas born on June 29th 1886. He was educated at Queen Elizabeth’s School Barnet and later at Birkbeck College London. In 1910 at the age of 24 he passed the General B.Sc. degree and thereafter stayed at Birkbeck College until the 1914-1918 war started. For some time he was in the R.N.A.S. and worked on war gases at Stratford. During the period 1915-1918 he published three papers with G. Senter (then Head of the Chemistry Department at Birkbeck College) on the influence of solvents on the Walden inversion which was then one of the great chemical conundrums.After the war Drew became Research Assistant to G. T. Morgan (afterwards Sir Gilbert) at the Univer- sity of Birmingham and whilst in this position (1919-1922) he was joint author of a number of papers on residual affinity and co-ordination dealing mainly with selenium and tellurium derivatives of acetylacetone such for example as 1 l-dichlorocyclo- telluripentane-3,5-dione.During this period Drew and Morgan also put forward possible mechanisms for the conversion of ethyl a-bromoacetoacetate into ethyl y-bromoacetoacetate in presence of trace of hydrogen bromide. From 1922-1930 Drew was a Lecturer at Birmingham and for a time worked with Morgan on Claisen condensations between ketones and esters it was shown that at low temperatures normal pro- ducts were formed but that at higher temperatures alcoholysis supervened.When W. N. Haworth (later Sir Norman) became Professor at Birmingham Drew became interested in sugar chemistry and in 1926 was conjoint author of the very important paper in which the significance of Hudson’s rule was made evident. In the following year in association with Goodyear and Haworth Drew was a joint author of an outstanding paper in which a comparative study of ten methylated lactones was described. It was shown that the rate of hydrolysis of five lactones from normal sugars (k,6-lactones) was much higher than that of five lactones derived from y-sugars (k,y-lactones).This work made it very easy to distinguish which was 8 and which was y of a pair of related lactones and is a well-recognised milestone in sugar chemistry. It was in 1926 that Drew and Haworth proposed the modem representation of the configurational formula of glucose. The sugars were not Drew’s sole interest at this time however. In 1926-1928 he carried out some very important investigations on heterocyclic com- pounds of tellurium not to be thought of though as a sequel to the Drew-Morgan work. It was shown that diphenyl ether condensed with tellurium tetra- chloride under mild conditions to give p-phenoxy- phenyltellurium trichloride which under the action of heat was converted into 10,l O-dichlorophenoxtel- lurin (phenoxtellurin dichloride).Reduction of the latter with potassium metabisulphite gave phenox- tellurin which had the surprising property of reacting with selenium or with sulphur with formation of phenoxselenin or phenoxthionin. It was noted that selenium tetrachloride chlorinates diphenyl ether without formation of phenoxselenin. Moreover phenoxselenin dibromide acts as a brominating agent. The chemistry of the phenoxtellurins was investigated very fully and virtually the whole of the experi- mental work was done by Drew himself. Incidentally he was probably the first person to introduce Pregl’s methods of microanalysis into this country. He in- variably did his own microanalyses and in 1928 (Drew and Porter) published notes on practical microanalytical methods of determining carbon and hydrogen.Running alongside of this work were some determinations (Drew and Haworth) of the molecular weight of inulin in water from which it was concluded that the molecules contained at least 24 anhydrofructose units. One of the most elegant pieces of experimental work1 ever witnessed was that done by Vernon just after the First War. It will be remembered that he had described the isolation of cis-and trans-forms of dimethyltelluronium di-iodide and dibromide corresponding to planar molecules. The evidence seemed excellent but in 1929 Drew showed that the halides of the a-series were non-polar with tetra- hedral molecules those of the p-series being salt-like complexes such as [Me,Te]+ [TeMeX&. It is very clear that Drew was less interested in “getting on” than in being associated with the living OCTOBER 1959 chemical problems with which he happened to come into contact.Thus in 1930 he was publishing papers with Wardlaw Pinkard Cox and others on metal co-ordination complexes. In 1930 it was concluded that Blomstrand‘s two isomeric complexes of platinous chloride with diethyl sulphide formerly thought to be cis-and trans-planar isomerides were most probably related as below. Later is was con- cluded that the supposed cis-trans (planar) isomerism of PdAzX was non-existent. The pink compound was probably [PdA JPdX,. e.g. [Pd(NHd,] [PdC14] and the yellow one [PdA,X,]. ,/SEt ...CI EtzS Ipt / + pt Et,Sf \Cl \SEt ...CI a-DichIoride p-Di chI oride In 1930 before the results of this work had been published Drew had become University Reader in Organic Chemistry at East London College (later Queen Mary College) and had for some time held the Ph.D.degree of Birmingham and the D.Sc. degree of London. He continued his investigations of metal complexes and showed for example that the a-di- sulphinesof platinum are trans-forms and correspond stereochemically and structurally with the a-diam- mines and not with the kdiammines as previously supposed. Drew also worked on copper complexes showing among other things that the compound written as [CuCl,,en] is better expressed as [Cu,2en]CuC14. He later made an extensive study of copper derivatives of o-hydroxyazo-compounds one simple result of this work being the discovery that o-hydroxyazobenzene forms an anhydrous (and non- solvating) co-ordination complex containing one copper atom with two residues of the azo-compound.Another considerable investigation was of the metal- lic “lakes” of alizarin. Drew also made notable dis- coveries of chemiluminescence among hydrazides of phthalic acid and added greatly to our knowledge of heterocyclic compounds and related substances obtainable from phthalonitrile. Few chemists can have covered so wide a field as Drew nor made such a variety of interesting observations. Drew who was a bachelor resigned from his University post at the age of 61 and went on living quietly in his rooms at Chislehurst. E.E. TURNER. CHRISTMAS COMPETITION THEEditor is indebted to Dr. G. M. Dyson for the following mnemonic on the Periodic Table dating from his student days Ha ! He Lies Between Bed- Clothes Not Over Fresh; Nevertheless A Nasty Knife Magnus Always Si.. .. .P.. . . .S Clover. Can Scrape Tins Very Creditably Minutely ; Cu . . ..Zn* Gabriel Gets Astonishing Sea - Feeds Cod Nicely! Breezes. Kruger Robs Agnes Serbia yet Cried In Ezra Nabs Sentimental More. Sobbing Tears Z Ruin Rhoda’s Pound!? Xer . .. .CesS Bans Large Cereals Takes Wine U!so Irritatingly Persistent. Golden Haig Tills Public Bills; Emma Rather Thrills U. * Cousin. 7 A bit of a “let down” this. $ In those days I couldn’t spell Xerxes. A prize (book token €2 2s.) is offered by The Chemical Society for a modern version of at least a substantial portion of the Periodic Table.Competitors may cover the Table horizontally or vertically and may divide it into groups as they wish; they may also handle (or omit) the transition elements and the rare earths etc. at their discretion. Entries must be received by the Editor not later than first post on December 31st. The author’s name and address and if desired a pseudonym for publication should be given. It is hoped to publish a report in Proceedings for January 1960. The Editor’s decision will be final.
ISSN:0369-8718
DOI:10.1039/PS9590000285
出版商:RSC
年代:1959
数据来源: RSC
|
|