摘要:

QUARTERLY REVIEWS THE PRODUCTION DETECTION AND ESTIMATION OF ATOMS IN THE GASEOUS PHASE By K. R. JENNINGS (DEPARTMENT OF CHEMISTRY THE UNIVERSITY SHEFFIELD 10) WITH the exception of the atoms of the inert gases and of certain metallic vapours free atoms possess high chemical reactivity. Atomic reactions have been of interest to chemists since the early part of the present century and progress in this field has been determined largely by the development of experimental methods both to produce atoms and to follow the course of their reactions. It is the purpose of this Review to consider the scope and limitations of the different techniques now in use for the study of such reactions. In 191 1 Struttl gave an extensive description of the properties of atomic oxygen and atomic nitrogen produced by a high-frequency electrical discharge and in 1920 Wood2 produced high concentrations of hydrogen atoms by using a low-frequency electrode discharge.About the same time Langmuir3 showed that hydrogen could be dissociated on a hot tungsten filament and a few years later photochemical methods of producing atoms were in use.4 In recent years the microwave discharge flash photolysis and shock tubes have been added to the methods available for the produc- tion of atoms and these developments are responsible to some extent for an increased interest in atomic reactions. Whereas there is a number of methods which may be used to demonstrate the presence of atoms in a reaction system relatively few yield reliable quantitative information. Of these even fewer measure absolute atom concentrations which are necessary for the evaluation of absolute rate constants of atomic reactions.A comprehensive review of studies of atomic reactions up to 1953 has been given by Stea~ie.~ References to much of the more recent work with particular reference to the stabilisation of energetic species at low temperatures have been given by Bass and Broida.6 In addition to their intrinsic interest atomic reactions are important l Strutt Proc. Roy. SOC. 1911 A 85 219. Wood Proc. Roy. SOC. 1920 A 97 455. Langmuir J. Amer. Chern. SOC. 1912 34 1310; 1915 37 417. Noyes and Leighton “The Photochemistry of Gases” Reinhold Publ. Corp. Steacie “Atomic and Free Radical Reactions” Reinhold Publ. Corp. New York Bass and Broida “Formation and Trapping of Free Radicals” Academic Press New York 1941.1954. London 1960. 237 1 238 QUARTERLY REVIEWS in many branches of chemistry. Atoms are known to be reactive inter- mediates in many complex reactions e.g. flames and explosions and a knowledge of their reactions is valuable in elucidating the overall reaction mechanisms. With the advent of high-altitude high-speed rockets the physics and chemistry of the free atoms present in the upper atmosphere have assumed greater importance and at lower altitudes atomic reactions are thought to play a significant role in the formation of “smog”. The Production of Atoms Atoms are produced from stable molecules by supplying sufficient energy to the molecules to cause them to dissociate. The energy may be supplied in a number of different ways e.g. heat (hot filament) light (resonance lamp) collision with high-speed electrons (discharges) etc.Ideally it would be desirable to produce a given concentration of atoms in a single known electronic state in the absence of reactive by-products. As will be seen in the following sections this ideal is very rarely achieved. Reactive by-products may be formed along with the atoms or may arise from reactions of the atoms. For example hydrogen atoms may be pro- duced by the photosensitised decomposition of an alkane but for each atom produced an alkyl radical is also produced. Details of the different methods of producing atoms are given in the appropriate sections below but the following generalisations may be made at this stage (a) With the exception of the shock-tube technique thermal methods are used to produce low concentrations of atoms from weakly bound diatomic molecules at low pressures in static or flow systems.(b) Photolysis is the most readily controlled method and may be used in static or circulating systems at any pressure to produce low concentrations of atoms in either the ground state or an electronically excited state. ( c ) Radiolysis is rarely used since it produces very low yields of atoms together with other reactive ionic and radical species. (d) Electrical discharges have been used to obtain relatively high concentrations of hydrogen oxygen and nitrogen atoms in fast flow systems at low pressures. (e) So-called “hot” atoms can be produced by bombardment of a gas with a beam of neutrons. These atoms are quite different from normal atoms in their reactions since they possess a very considerable excess of energy.All the above methods produce a more or less steady supply of atoms. The use of flash-photolysis and shock-tube techniques enables one to obtain very high transient concentrations of atoms in static systems. Thermal Methods.-When thermal methods are employed it is usual to use the elements as sources of atoms. In the laboratory temperatures in JENNINGS ATOMS IN THE GASEOUS PHASE 239 excess of 2000"~ are difficult to attain and consequently it is necessary for a molecule to have a fairly low dissociation energy if appreciable yields are to be obtained. The method has therefore been limited very largely to the production of atoms from H, Cl, Br, I, and metallic sodium the corresponding dissociation energies being 103 57 45 35 and 17 kcal./mole respectively.Three methods of supplying the heat have been used (i) The reaction vessel is enclosed in a furnace; (ii) an electrically heated filament maintained at a high temperature is present in the gas; (iii) the gas is subjected to a shock wave. Although the fast flow of a substance through a tube surrounded by a furnace has been used to produce free radicals it has found little applica- tion in the production of atoms. Hydrogen and deuterium atoms have been obtained in a static system at about 1000"~ in the study of the conversion of para-hydrogen into ortho-hydrogen and of ortho-deuterium into para- deuterium and of the exchange reactions between hydrogen and deu- terium.' In general other reactants cannot be mixed with hydrogen in such a system since they would decompose at the high temperatures which are necessary.More recently Scheer,8 working at very low pressures has produced a beam of iodine atoms by pumping them from a quartz tube containing iodine at 1100"~. An attempt to stabilise the atoms at very low temperatures in a solid matrix was unsuccessful. Owing to the low dissociation energy of the Na molecule sodium vapour in equilibrium with the molten metal is very largely monatomic. A stream of sodium atoms can therefore be obtained by passing an inert carrier gas over the molten metal. If this is then passed into an atmosphere of a substance such as an alkyl halide the sodium atoms will abstract the halogen atoms to form a sodium salt and an alkyl radical. Work in this field up to 1951 has been reviewed by Warhur~t.~ Langmuir3 was the first to show that hydrogen is dissociated on a heated tungsten filament at very high temperatures.At low pressures the atoms diffused to the walls and reduced metallic oxides. More recently Tollefson and LeRoylO produced a stream of hydrogen atoms at low pressures by passing the gas very rapidly over a heated tungsten filament electrically maintained at about 1 7 0 0 " ~ . The atom concentration measured by an isothermal calorimeter was in the region of 0.1 %. Klein and Scheerll have also used this method to produce hydrogen atoms in a static system at very low pressures. The atoms are allowed to diffuse to the walls where they react with a thin layer of solid olefin at 7 7 " ~ . Other atoms have not been obtained by the filament method probably because many of them would attack the filament.Farkas "Ortho-hydrogen Para-hydrogen and Heavy Hydrogen" Cambridge Scheer unpublished work quoted on page 22 of reference 6. Klein and Scheer J. Phys. Chem. 1958 62 1011. Univ. Press 1935. ' Warhurst Quart. Rev. 1951 5 44. lo Tollefson and LeRoy J. Chem. Phys. 1948 16 1057. 240 QUARTERLY REVIEWS During the last fifteen years the shock-tube technique has been increas- ingly applied to the study of chemical problems. When a shock wave passes through a gas the gas is compressed adiabatically and a rapid rise in temperature results. With monatomic gases temperatures as high as 18,000"~ may be obtained but the higher heat capacities of diatomic and polyatomic gases together with the heat absorbed in dissociation will lead to a much smaller rise in temperature. The temperature attained is controlled in practice by the strength of the shock and for a given gas the stronger the shock the higher the temperature.Such a shock wave may be generated in the laboratory by the sudden bursting of a diaphragm in a brass tube. On the high-pressure side of the diaphragm is an inert driver gas at a pressure of a few atmospheres. When the diaphragm bursts the driver gas expands rapidly into the region of low pressure in which is the gas to be studied. The main application of this technique to the production of atoms has been in the study of the recombination of bromine and iodine atoms in the presence of various third bodies. A number of other simple systems involving atoms has been investigated and references to this work are to be found in a recent review of this technique by Pritchard.12 Photochemical Methods.-The use of light as a source of energy is the basis of some of the most widely used methods of producing atoms.The two most important reasons for the popularity of photochemical methods are that (1) the amount of energy per photon and hence the amount of energy absorbed per molecule can be controlled by varying the wave- length of the incident radiation and (2) the rate of input of energy can be controlled by adjusting the intensity of the source. For radiation of con- stant intensity over long periods of time the most stable source is a resonance lamp such as a medium-pressure mercury lamp which emits radiation at several wavelengths. Light of the required wavelength is obtained by interposing a filter between the source and the reaction cell.The cell contains the absorbing molecule which is the source of atoms together with the other reactant which should not absorb the incident radiation. Since the output of even the most powerful resonance lamp is relatively low the rate of production of atoms is also low and con- sequently photochemical reactions are usually studied in stat ic or circulat- ing systems. Although the steady-state concentration of atoms is too low to be measured it is usually unnecessary to know this value. By making use of competing reactions relative values of rate constants pre-exponen- tial factors and energies of activation may be determined from product analyses. Since photochemistry is concerned with the study of atomic and free- radical reactions only a few representative methods of producing atoms can be mentioned in a review of this nature.For a fuller discussion the reader is referred to standard text-b~oks.~ r 5 Photochemical methods of producing atoms can be conveniently clas- l2 Pritchard Quart. Rev. 1960 14 46. JENNINGS ATOMS IN THE GASEOUS PHASE 241 sified under three main headings (a) direct photolysis (b) photosensitised decomposition and (c) flash photolysis. These will be considered in turn. (a) When a diatomic molecule absorbs a quantum of radiation in the visible or ultraviolet region of the spectrum it is excited to a higher electronic state. According to the Franck-Condon principle the most probable transitions are those in which the internuclear distance remains constant. The absorption spectrum of the molecule is determined by this principle together with the relative positions of the potential-energy curves and the requirements of spectroscopic selection rules.If the mini- mum of the potential-energy curve of the upper state is at a slightly greater internuclear distance than that of the ground state the absorption spectrum consists of bands at the longer-wavelength end converging to a continuum at a well-defined wavelength at shorter wavelengths. The onset of the continuum occurs at the maximum wavelength (minimum energy) necessary to dissociate the molecule into atoms one of which is usually in an excited state. The spectra of oxygen chlorine bromine and iodine all correspond to this description and in each case one normal and one excited atom are produced on dissociation. If the upper state is purely repulsive as in the case of hydrogen no bands are observed at the long- wavelength end of the continuum.For direct photolysis to be a useful source of atoms it is necessary that the region of continuous absorption should occur in an easily accessible part of the spectrum. Continuous absorption by hydrogen and nitrogen occurs only in the extreme vacuum-ultraviolet region where experimental difficulties are too great for the method to be of use. The oxygen conti- nuum begins at 1759 A however and by.using a combined lamp and reaction-cell unit separated by a thin quartz window the direct photolysis of oxygen could be used as a source of atoms. Chlorine bromine and iodine on the other hand are all dissociated by light of wavelength shorter than about 4800 A a much more readily accessible region of the spectrum.Consequently direct photolysis is by far the most important source of halogen atoms. The excess of energy of the excited atom is 2.5 10.5 and 22.7 kcal./mole. for C1 Br and I atoms respectively so that whereas the method is excellent for the production of chlorine atoms it is somewhat less than ideal for iodine atoms. Although the direct photolysis of hydrogen and oxygen is unsuitable as a source of atoms both H and 0 atoms have been obtained by photo- lysis of other molecules. The primary step in the photolysis of formal- dehyde is the formation of a hydrogen atom and a CHO radi~a1.l~ The photolysis of dideuteroformaldehyde has recently been used as a source of D atoms to study their reactions with hydrogen methane and deuterated f0rma1dehyde.l~ Darwent and Roberts15 have produced deuterium atoms l3 CaIvert and Steacie J.Chern. Phys. 1951 19 176. l4 Klein McNesby Scheer and Schoen J. Chern. Phys. 1959,30 58. l5 Darwent and Roberts Discuss. Faraday SOC. 1953 14 55. 242 QUARTERLY REVIEWS by the photolysis of deuterium sulphide D2S in order to study their reactions with several hydrocarbons. Oxygen atoms may be obtained by the photolysis of NO2 at different wavelengths. Photolysis in which radiation in the region of 3360 is used results in the formation of ground state (") atoms whereas the use of the cadmium resonance line at 2288 A almost certainly produces atoms in the l D state.16 The reactions of oxygen atoms in different electronic states can therefore be studied although the reactivity of NO2 itself limits the usefulness of the method.(b) The dissociation energy of hydrogen is about 103 kcal./mole which corresponds to the energy available from radiation of about 2750 A. However since hydrogen does not absorb in this region no dissociation occurs under these conditions. This difficulty may be overcome by addition tion of a sensitiser the function of which is to absorb the radiation and to transfer the energy to the hydrogen by collision. It is clearly desirable that the sensitiser should absorb the incident radiation very strongly and that it should not interfere with subsequent reaction of the atoms. Since the mercury resonance lamp is the most widely used source of radiation mercury vapour is an obvious choice as a sensitiser. It will absorb mercury resonance radiation very strongly and is relatively inert.In order that reversal of the resonance radiation should not occur in the lamp low- pressure lamps are used. If a mixture of hydrogen and mercury vapour is irradiated with the 2537 A mercury resonance radiation the mercury atoms will be excited to the 63P1 state 112 kcal./mole above the 6lS ground state. In the absence of the hydrogen the radiation would be re-emitted but this is quenched by collision with the hydrogen. Since the hydrogen acquires energy in excess of its dissociation energy the molecule dissociates and the excess energy appears as kinetic energy of the atoms. The presence of hydrogen atoms in such a system was first demonstrated by the fact that they would reduce metallic oxides present in the reaction cel1.l' Gaseous-phase reactions of the atoms may be studied by the addition of a second gas but a large excess of hydrogen is necessary to minimise quenching of the excited mercury atoms by the other reactant.Hydrogen atoms have also been obtained by the mercury-photosensi- tised decomposition of paraffin hydrocarbons in which the initial reaction is RH -+ R + H followed by H + RH -+ R + H2. If a small amount of an olefin is added to the system the rate of addition to the olefin relative to the rate of abstraction from the paraffin may be determined by measur- ing the yield of hydrogen under different conditions.ls Mercury-photosensitised decomposition of nitrous oxide has proved to be a very convenient source of oxygen atoms.19 If no other reactant is l6 Sat0 and Cvetanovic Canad. J. Chem. 1958 36 1668. l7 Cario and Franck 2. Phys. 1922 11 161.l8 Jennings and Cvetanovic unpublished work. Cvetanovic J. Chem. Phys. 1955,23,1203; Canad. J . Chem. 1960,38 1678. JENNINGS ATOMS IN THE GASEOUS PHASE 243 present oxides of nitrogen and mercuric oxide are formed. In the presence of a very small concentration of a substance such as an olefin which will rapidly remove the oxygen atoms the reaction is very clean and the rate of production of oxygen atoms can be found by measuring the rate of pro- duction of nitrogen. This technique has been used to study the reactions of oxygen atoms with a large number of olefins. In order to vary the energy input vapours of other metals have also been used as sensitisers the most useful being zinc and cadmium. The rare gases krypton and xenon can also be used as sensitisers in the vacuum- ultraviolet region of the spectrum and Croth2O has recently obtained evidence which suggests that nitrogen atoms can be obtained by the krypton-photosensitised decomposition of nitrogen.(c) In normal photochemical studies a very low steady-state concen- tration of atoms is maintained by continuous irradiation. In flash photo- 1 ysis however very high concentrations are produced momentarily by subjecting the reaction mixture to a single intense flash obtained by the discharge of a bank of condensers.21 The appearance and decay of reactive intermediates can be observed by following the intensity of their absorp- tion spectra over a period of several hundred milliseconds. Since atoms do not absorb in the accessible region of the spectrum their reactions can only be inferred from changes in the intensity of other absorption spectra.To avoid the complication of a sharp rise in temperature the reactants are diluted with a large excess of an inert substance such as nitrogen or argon. Flash photolysis has been widely used in the study of the recom- bination of halogen atoms in the presence of various third bodies and in the study of combustion processes.22 In the latter studies it is frequently necessary to introduce a sensitiser such as nitrogen dioxide in order to initiate the reaction the presence of which may in certain cases interfere with subsequent reactions. Radio1ysis.-This technique has been very little used for the gaseous- phase production of atoms for two main reasons. First the transfer of energy from a beam of y-rays to the gas is very inefficient resulting in extremely low conversions even when a reaction time of several hours is used.Secondly the very high energy radiation produces many reactive species in addition to atoms. Whereas in the mercury-photosensitised decomposition of pentane a hydrogen atom and a pentyl radical are the initial products in the radiolysis of pentane y-rays cause the emission of secondary electrons from the surface of the reaction vessel leading to additional reactions23 such as It is therefore very difficult to follow the reactions of the atoms alone and a 2o Groth 2. phys. Chem. (Frankfurt) 1954 1 300. 'L1 Porter Proc. Roy. Soc. 1950 A 200 284. 22 Norrish and Thrush Quart. Rev. 1956 10 149. 23 Futrell J. Phys. Chem. 1960 64 1634. 244 QUARTERLY REVIEWS large variety of products is found.The addition of a scavenger such as an olefin suppresses the formation of hydrogen by a radical mechanism and indicates that as much as one-third of the total hydrogen is produced by a molecular mechani~m.~~ Electrical Discharges.-Electrical discharges have been used for many years to produce a stream of atoms in fast flow systems at low pressures. They have the advantage that high concentrations of atoms are produced so that a flow system may be used and the atoms may be generated before being mixed with the other reactant. There are three main types of discharge which are useful to the chemist in the study of atomic reactions (a) the low-frequency electrode discharge or Wood’s tube (b) the radiofrequency or electrodeless discharge operat- ing on a frequency of a few Mc./sec.and ( c ) the microwave discharge which operates on a frequency of 2500-3000 Mc./sec. A review of electrical discharge techniques has recently been given by Shaw in chapter 3 of reference 6. In each case energy is supplied by accelerating electrons under the influence of an applied electric field. The high-velocity electrons collide with molecules leading to dissociation either by excitation to an unstable electronic state or by a mechanism involving ions. For the discharge to be maintained the rate of production of electrons in ionising collisions must equal the rate of removal of electrons at the walls or by ion-electron recombination. Consequently optimum working conditions vary both with the gas to be dissociated and with the type of discharge employed. The low-frequency discharge was first described by Wood2 and was later used extensively by Bonhoeffer for the production of hydrogen atoms.25 It consists of a U-tube 1-2 metres long with an internal diameter of 2-3 cm.normally of Pyrex glass. The gas is admitted close to the alumi- nium electrodes at the end of each arm and is withdrawn from the middle of the tube. A mains-frequency voltage of about 2 kv is applied to the electrodes; a steady discharge is normally used but a pulsed discharge appears to give better results with nitrogen.26 The main disadvantage of the low-frequency of discharge is the risk of contamination with electrode materials. For oxygen Linnett and Marsden found evidence to suggest that aluminium oxide was carried along in the gas The yield of atoms varies widely with different experimental conditions but for hydrogen atoms Poo1e28 recommends the use of high flow rates at a pressure of 0.6 mm.Hg. The presence of traces of impurities such as water vapour and the poisoning of the walls greatly increase the yield and these effects are discussed in a later section. This type of discharge tube may be used in the pressure range 0.5-5 mm. Hg and the addition 24 Back J. Phys. Chern. 1960 64 124. 25 Bonhoeffer Z. phys. Chem. 1924 113 199. 26 Evans Freeman and Winkler Canad. J. Chem. 1956,34 1271. 37 Linnett and Marsden Proc. Roy. SOC. 1956 A 234 489. Poole Proc. Roy. SOC. 1937 A 163,404. JENNINGS ATOMS IN THE GASEOUS PHASE 245 of an inert carrier gas such as neon or helium extends the upper limit to about 20 mm. Hg. The radio-frequency discharge has the advantage that no special dis- charge tube is necessary and that no electrodes come into contact with the flowing gas.The power is supplied by a radio-frequency oscillator with an output of a few hundred watts in the region 1-20 Mc./sec. The most common form of coupling is a coil of several turns of copper wire wound round a quartz discharge tube with a clearance of about 5 mm. A variable air-condenser across the coil enables one to tune the load to the frequency of the oscillator. In order to avoid the spread of the discharge an earthed metal screen may be placed between the discharge tub2 and the All leads and electrical instruments must be carefully screened from stray radio-frequency fields. Since this discharge can be maintained at much lower pressures than are possible for electrode or microwave discharges it is particularly useful in the study of surface reactions of In recent years the use of the microwave discharge has become in- creasingly widespread for the production of hydrogen oxygen and nitrogen atoms.The usual source of power is a magnetron operating in the frequency range 2500-3000 Mc./sec. with an output of several hundred watts. The output is fed into a standard 3” x 1” waveguide which is in turn coupled to a tunable resonant cavity through which the quartz discharge tube passes (see Figure). It is usually necessary to trigger the Appratus for microwave discharge. M Magnetron. A Water attenuator. T Tuning screws. S Slotted diaphragm. D Discharge tube. P Movable plunger for tuning. discharge with a Tesla coil. An attenuator is necessary to absorb the power before the discharge is struck and may take the form of a glass or Poly- thene tubz through which water is passed.The tubz is partly lowered into a longitudinal slit in the broad face of the waveguide and is removed after the discharge has been struck. Impedance matching between the magne- tron and the discharge is accomplished by means of a movable plunger and a screw tuner. The Figure shows one of many possible arrangements. The main advantages of the microwave discharge over the radio- frequency discharge are that it is a very intense localised discharge with no tendency to spread and stray electrical fields are eliminated by the use of the waveguide. It is also possible to maintain a microwave discharge at pressures of several cm. Hg but the radio-frequmcy discharge is to be preferred at very low pressures.Jennings and Linnett Nature 1958 182 597. ao Greaves and Linnett Trans. Faraday Soc. 1959 55 1338. 246 QUARTERLY REVIEWS It has long been known that the presence of very low concentrations of impurities greatly increases the percentage dissociation of hydrogen oxygen and nitrogen in electric discharges. The presence of as little as 0.05% of water vapour may increase the yield of atoms by a factor of ten or more. Hydrogen which has been purified by passage through a Deoxo unit and silica gel was appreciably dissociated by a microwave discharge but when it was further dried in a liquid-nitrogen trap the yield dropped to less than one-tenth of the former value.31 Wood2 explained the effect of water vapour in terms of poisoning of the walls of the discharge tube i.e.prevention of recombination of atoms on the walls. However the insertion of a liquid-nitrogen trap in the hydrogen line results in a rapid drop in hydrogen atom concentration sug- gesting a gaseous-phase rather than a surface effect. Kaufman and K e l ~ o ~ ~ find rapid and reversible catalytic effects when 0.01 % of nitrogen nitrous oxide or nitric oxide is added to oxygen which is dissociated in a micro- wave discharge. Although a surface effect may also be present it seems that the traces of impurities may facilitate dissociation of the gas in the dis- charge perhaps by the initiation of chain reactions. Surface removal of atoms can be minimised by pre-treating the surface of the discharge tube and reaction vessel with acids such as ortho- phosphoric acid but a trace of water vapour is still necessary if high yields of atoms are required.More recently a mixture of methyltrichlorosilane and dimethyldichlorosilane known as “Drifilm” has been but water vapour was again necessary to obtain high yields. The presence of traces of impurities in the gases passing through the discharge is not normally a serious drawback. A more troublesome feature is the production of other potentially reactive products in addition to ground-state atoms. There is no evidence of other reactive species in the products from a hydrogen discharge but in the cases of oxygen and nitrogen there is considerable evidence of the presence of excited molecules. This evidence is discussed in later sections. It is difficult to compare the efficiency of the different types of discharge since the yield of atoms depends on the power input as well as the presence of traces of impurities.The products obtained on passing hydrogen or oxygen through a discharge do not appear to depend on the type of discharge but as will be seen below there is evidence to suggest that a powerful condensed electrode discharge through nitrogen produces a second reactive species which is possibly not formed in a micro-wave discharge. The electrical discharge is seen to be a convenient source of atoms at low pressures and so is well suited to the study of surface reactions and very simple homogeneous reactions such as 0 + NO N + NO etc. but the necessity of working at low pressures makes the method less suitable for the study of more complex reactions. Shaw J. Chem. Phys.1959.31 1142. 82 Kaufman and Kelso J. Chem. Phys. 1960,32,301. JENNINGS ATOMS IN THE GASEOUS PHASE 247 “Hot Atoms.”-A “hot” atom is one which is produced with a great excess of energy and is therefore not in thermal equilibrium with its surroundings. If the atom reacts before attaining thermal equilibrium the reaction is termed a “hot” reaction. Such reactions normally bear little relation to the reactions of normal atoms and may be distinguished from the latter on addition of a scavenger which will not suppress the reactions of the hot atoms. Hot atoms rapidly lose their excess energy by collision so that unless they are produced with a vast excess of energy (several hundred kcal./mole) or the “steric factor” is high hot reactions are not observed. Hot atoms are usually produced by irradiation of the gas with a beam of neutrons.When chlorine gas is irradiated in this way atoms of the radio- active isotope 3sCl are produced with very high energy and their reactions can be followed by tracer techniques. Hot bromine and iodine atoms may be produced in a similar manner and recently hot tritium atoms have been produced from 3He.33 A review of hot-atom chemistry up to 1955 has been given by Willard.34 The Detection and Estimation of Atoms With few exceptions atoms are always found in systems which contain other species in addition to the atoms and consequently the method chosen to detect or estimate the atoms is to some extent governed by the nature of these other species. Although it is possible to demonstrate the presence of atoms in complicated reacting systems (such as flames) quantitative measurements in such systems are difficult to make and have therefore been confined very largely to simple systems in which fairly high concentrations of atoms were present.Almost any property in which the atom differs from the other species present may be used as a basis for detection. Methods based on such differences as mass spectra and chemical reactivity have been widely used and although some of these methods are of qualitative or semi- quantitative value only several techniques are now available for the measurement of reZative atom concentrations with reasonable accuracy. Unfortunately reliable values of absolute atom concentrations are more difficult to obtain. Several different methods were used to measure the absolute atom concentration in the effluent gas from an oxygen discharge and the values obtained differed by as much as 25%.35 Possible errors involved in the use of the different methods are considered in the ap- propriate sections below.For quantitative work it is desirable that the act of measuring the atom concentration should interfere as little as possible with the system under investigation. Alternatively the atom concentration may be measured in 83 Henchman Urch and Wolfgang Canad. J. Chem. 1960,38 1722. s6 Elias Ogryzlo and Sch8 Canad. J. Chem. 1959,37,1680. Willard Ann. Rev. Phys. Chem. 1955,6 141. 248 QUARTERLY REVIEWS the absence of the second reactant after which the reaction is carried out under as nearly identical conditions as possible. In addition the method employed should be specific for the atoms concerned and should not be affected by the presence of other species in the system.In certain cases it is desirable to measure the atom concentration at a point or at a given position along a tube rather than to obtain an average over a fairly large volume. The following generalisdons may be made concerning the more common methods in use. (a) The mass spectrometer and the Wrede gauge are the two methods in general use that are dependent on mass difference. Both affect the system to a negligible extent the Wrede gauge giving absolute values in simple systems at very low pressures the mass spectrometer being more specific. (b) Calorimetric methods are based on the heat liberated by the recom- bination of the atoms. They have a greater effect on the system than the above methods and are unreliable in the presence of other excited species.(c) Emission spectra give information about excited species only and since most atoms do not absorb in readily accessible regions of the spec- trum indirect methods are necessary. Spectroscopic methods usually yield values over an extended zone rather than at a point. (d) Electron spin resonance is just beginning to be used for the gaseous phase. It is very sensitive has no effect on the system and has been used both to detect atoms and to obtain relative concentrations. (e) Chemical reactions of atoms may be used to measure atom con- centrations if the reaction is simple and very fast and has a known mechanism. Chemical methods are probably the most reliable methods of estimating oxygen and nitrogen atoms generated by discharges.Methods Dependent upon Mass Difference.-(a) The Mass Spectrometer. Since the application of the mass spectrometer to the study of free radicals has recently been reviewed,36 the reader is referred to that article for an account of the experimental procedure. Since the mass spectrometer can distinguish between atoms of different mass it may be used for detection of more than one type of atom in a complex reaction. In addition it is possible to distinguish between different electronic states of an atom since these have different appearance potentials. Although absolute atom concentrations cannot be measured because of recombination on the walls of the instrument reliable relative concentrations can be obtained. Mass-spectrometric analyses have been carried out on the products obtained when oxygen3' and nitrogen3* are subjected to an electrical discharge.In the case of oxygen ground-state atoms were found to be the most abundant reactive species but electronically excited oxygen 36 Cuthbert Quart. Rev. 1959 13 215. s7 (a) Foner and Hudson. J . Clzem. Phys. 1956 25 602; (b) Herron and Schiff ** (a) Jackson and Schiff J. Chem. Phys. 1955 23 2333; (6) Berkowitz Chupka Canad. J. Chem. 1958,36,1159. and Kistiakowsky J. Chem. Phys. 1956,25,457. JENNINGS ATOMS IN THE GASEOUS PHASE 249 molecules in the A state were present in concentrations in th? region of 10-20%. In the case of nitrogen ground-state atoms were the only reactive species found when a microwave discharge was When an electrode discharge tube was used however a second excited species was found with an appearance potential of 16.1 e~~~~ and suggests the presence of metastable nitrogen molecules in an unknown state of excitation.The absence of this appearance potential when a microwave discharge was used may have been caused by unfavourable sampling conditions or may imply that the second species is not formed in a microwave discharge. (b) The Wrede Gauge. This was first described by W ~ e d e ~ ~ and has been used to measure absolute atom concentrations in simple systems at low pressures. A detailed account of its construction and use has recently been given.30 The gauge consists of a hole in the wall of the tube along which the gas is flowing. Behind the hole a catalytic surface is enclosed in a small volume together with a pressure gauge. Providing that the mean free path in the gas A is at least ten times the diameter d of the hole the flow through the hole is purely molecular.Under these conditions the recombination of atoms on the catalytic surface behind the hole causes a pressure differential to be set up across the hole. If p is the pressure in the tube Ap the pressure differential and a the volume percentage of atoms (no. of atoms/total no. of particles) then at equilibrium a = 3.41 Ap/p and a may be determined by measuring p and Ap. Since .dp must be known very accurately it is best measured directly by means of a differen- tial pressure gauge. The time necessary to attain equilibrium is inversely proportional to the area of the hole and directly proportional to the volume behind the hole. Because of the A/d requirement a hole of diameter 0.1 mm.cannot be used in hydrogen at pressures exceeding 0.1 mm. Hg and in oxygen and nitrogen the maximum pressure would be about 0.07 mm. Hg at room temperature. The use of sintered discs to decrease the equilibration time at higher pressures may lead to unreliable results since the large number of small holes in close proximity may be little better than the use of a single larger hole and will also cause more atoms to be abstracted from the system.30 Care must also be taken to avoid temperature differentials across the hole which would give rise to spurious readings because of thermal molecular-pressure effects. Similar effects complicate its use at elevated temperatures. Since Ap is usually measured by means of a thermal-conductivity gauge the Wrede gauge is suitable for use with a pure gas only.In a reacting system changes in conductivity may arise from variations in pressure or in composition. If the use of the gauge is confined to very low pressures in pure gases at room temperature measurements of absolute concentrations should be quite reliable since the readings are not affected by the presence of excited species. 88 Wrede 2. Phys. 1929 54 53 250 QUARTERLY REVIEWS Calorimetric Methods.-Since the dissociation energies of hydrogen oxygen and nitrogen are 103 117 and 225 kcal./mole respectively considerable heat is liberated when atoms of these elements recombine as may be seen from the “hot spots” which are occasionally observed on the walls of fast-flow systems. B~nhoeffer~~ coated the bulbs of thermo- meters with catalytic surfaces and observed appreciable rises in temperature when the thermometers were exposed to a stream of hydrogen atoms.The rise in temperature is only very approximately proportional to the atom concentration for several reasons. The catalytic activity of the surface is usually temperature-dependent and may decrease slowly owing to “poisoning” by traces of imp~rities.~~ In addition any excited atoms or molecules present in the system will also liberate heat on striking the catalytic surface and heat losses will be difficult to estimate. Furthermore considerable numbers of atoms will be removed from the system in order to produce an appreciable temperature rise. Two modifications of Bon- hoeffer’s original method have been developed in an attempt to overcome some of these objections.(a) The Catalytic Probe. This usually consists of a very small piece of metal such as platinum or silver attached to a thermocouple the leads of which are encased in silica. The area of the metal is in the region of 1 mm.% so that comparatively few atoms are removed from the system. Since small increases in temperature can be measured accurately cooling corrections and variations of the catalytic activity with temperature are minimised. Providing that surface poisons as well as excited atoms and molecules are absent the catalytic probe may be used to obtain relative atom concentrations. Greaves and Linnett were able to correlate the temperature rise of a catalytic probe with the absolute concentrations obtained from a Wrede gauge suggesting that these assumptions were justified under their experimental condition^.^^ (6) The Isothermal Calorimeter.Whereas the catalytic probe is de- signed to abstract as few atoms as possible from the system the iso- thermal calorimeter aims to abstract all atoms from a stream of gas thereby allowing absolute concentrations to be determined. The calori- meter is electrically maintained at a given temperature and the fall in the input of electrical energy recorded when the discharge is in operation is a measure of the heat supplied to the calorimeter from the gaseous phase. If the flow-rate of the gas and the heat of recombination of the atoms are known and on the assumption that this is the only source of heat absolute concentrations may be calculated. Since the process is isothermal heat losses are constant and the catalytic activity of the surface remains much more nearly constant.This method was used by Tollefson and LeRoy to measure hydrogen atom concentrations,1° the detector being a coil of platinum wire. Since 40 Fox Smith and Smith Proc. Phys. SOC. 1959,73,533. JENNINGS ATOMS IN THE GASEOUS PHASE 25 1 there is no evidence of the presence of excited atoms or molecules in the effluent gas from a discharge through hydrogen this method is likely to be very reliable generally for the measurement of absolute concentrations of hydrogen atoms in flow systems. Concentrations of deuterium atoms determined in this way agree very closely with values obtained by chemical methods.35 In the case of oxygen its use is complicated by the presence of excited oxygen molecules.35 The platinum coil was silver-plated.On exposure to the atoms the silver was rapidly converted into silver oxide which is an excellent catalyst for the recombination of oxygen atoms. When a little nitric oxide was added to the gas stream the green air-afterglow W ~ S completely quenched by the coil indicating the complete removal of atoms. This was however no longer true at temperatures above 10Ooc presumably owing to a decrease in the activity of the silver oxide surface. It was found that chemical methods and Wrede-gauge measurements gave consistently lower readings than those obtained from the isothermal calorimeter and the discrepancy increased when the detector was coated with cobalt oxide. When all the atoms were removed by a mercury mirror a considerable amount of heat was still liberated at the cobalt oxide detector presumably owing to deactivation of excited molecules.It was found that a freshly prepared silver oxide surface removed all the atoms but deactivated very few excited molecules and this was used to measure atom concentrations. After several runs however increasing numbers of excited molecules were also recorded and a fresh detector was then necessary. Little is known about the catdytic activity of different surfaces in effecting recombination of nitrogen atoms and in view of the presence of other reactive species in “active nitrogen” calorimetric methods are of little use in the estimation of nitrogen atoms. Spectroscopic Methods.-Since the ground-state atoms of most non- metallic elements do not absorb in easily accessible regions of the spectrum absorption spectroscopy has not been widely used to detect the presence of atoms.However Tanaka et aL41 were able to show that the concentration of excited atoms in active nitrogen is very low but they observed strong absorption at 1200 A which indicated a much higher concentration of ground-state atoms. Absorption spectroscopy is widely used in the technique of flash photolysis to detect unstable radicals but the presence of atoms has to be inferred by indirect methods. In the study of the recombination of iodine atoms the decay of the atoms is followed by observing the increase in the intensity of the absorption spectrum of molecular iodine and the ap- pearance of the hydroxyl radical spectrum in a system is usually due to hydrogen-abstraction by oxygen atoms. ‘l Tanaka Jursa and LeBlanc “The Threshold of Space” Pergamon London 1957 p.89. 252 QUARTERLY REVIEWS A quantitative estimate of oxygen and nitrogen atom concentrations has recently been made by using the absorption spectrum of nitric oxide.42 When just enough NO2 is added to quench the faint oxygen afterglow a molecule of nitric oxide is produced for every oxygen atom present (see below under chemical methods) and the concentration of nitric oxide is then measured spectroscopically. In the case of nitrogen the afterglow is just quenched by the addition of nitric oxide and the discharge is then switched off and the nitric acid concentration is again found spectro- scopically. In each case the atomic concentration is equivalent to the nitric oxide concentration. Emission spectra have been little used in quantitative work since it is necessary for the species to be present in an excited state for such spectra to be observed.As is to be expected atomic lines are often observed in discharges and the emission of radiation by metal atoms is the basis of the familiar flame tests of inorganic qualitative analysis. Hartel and P01anyi~~ measured the concentration of ground-state sodium atoms in a diffusion flame by illuminating the flame with the radiation of a sodium resonance lamp and measuring the intensity of the resonance radiation emitted. Other methods of measuring the concen- tration of ground-state atoms based on emission spectra have been indirect. When a little nitric oxide is introduced into a flame containing oxygen atoms a yellow-green continuum is observed due to the occurrence of the reaction 0 + NO 3 NOz + hv and this has been used in a q~alitative~~ and ~emiquantitative~~ manner to demonstrate the presence of oxygen atoms in hydrogen flames.Further use of this reaction is described in the section below on chemical methods. Sugden and his collaborator^^^ have recently described a method of measuring the concentration of hydrogen atoms in flames. Small equal amounts of sodium and lithium are introduced into the burner gases as salt sprays and the intensities of the resonance lines of each metal are compared the relative transition probabilities being known. Since sodium atoms do not form compounds the sodium resonance line is used as a reference. Some lithium is removed as the very stable lithium hydroxide by the reversible reaction Li + H20 + LiOH + H.The weakness of the lithium resonance line readily shows up and from its intensity and from the calculated equilibrium constant of the above reaction the concentra- tion of hydrogen atoms may be determined. The concentrations of lithium and sodium are kept very low so as to avoid self-reversal. The temperature of the flame is measured by the sodium-line reversal method. It is esti- mated that absolute values can be measured to within &20% and relative values to within &-5 %. 42 Broida Schiff and Sugden Nature 1960 185 760. 43 Hartel and Polanyi Z. phjts. Chm. 1930 B 11 97. 44 Gaydon Trans. Faradcy SOC. 1946,42,292. 45 Jarres and Sugden Nature 1955 175,252. 46 Bulewicz James and Sugden Proc. Roy. Soc. 1956 A 235 89. JENNINGS ATOMS IN THE GASEOUS PHASE 253 Electron-spin Resonance.-One of the intrinsic properties of atoms with the exception of the rare gases is that they all possess one or more unpaired electrons.If a strong magnetic field is applied to such atoms the normal system of energy levels is disturbed and splitting of the levels occurs. Since the electron has a magnetic moment associated with it it will align its spin and moment either parallel or antiparallel to the field if placed in a d.c. magnetic field. The energy splitting between the two levels is given by gPH where g is the splitting factor of 2.0023 fi is the Bohr magneton and H is the strength of the magnetic field. By choosing a suitable value for H the atom can be made to absorb in the microwave region the frequency being given by hv = gPH. The sensitivity is much greater at microwave frequencies than at radio- frequencies and a field of about 3000 gauss is usually used in conjunction with radiation of a frequency in the region of 9000 Mc./sec.When this method is used to detect atoms in the gaseous phase the stream of atoms is passed through a quartz tube between the poles of the magnet and at right angles to the broad face of the waveguide. The power is provided by a magnetron or klystron and a crystal detector may be used to measure the absorption of radiation. Details of practical systems are given by 111gram.~‘ Since the intensity of the absorption will depend on the number of atoms present the method is ideally suited to the measuring of relative atom concentrations. Shaw used this technique to determine the effect of water vapour on the dissociation of hydrogen in a microwave discharge,31 and it has also been used to demonstrate the presence of ground-state atoms in active nitrogen.48 The relative concentrations of oxygen atoms have been determined by electron-spin resonance spectroscopy in a study of the iecombination of the atoms at surfaces.49 The method can be made very sensitive detecting less than 10l2 atoms per C.C.(less than 0.1 % atoms at 0.1 mm. Hg) but absolute measurements can only be made with the help of a suitable reference measurement. In addition to its high sensitivity the method has the advantages that it may be used in complex systems to follow relative concentrations of different species simultaneously without in any way affecting the chemical processes and it also gives an instantaneous reading over a small region of the tube.Application of electron-spin resonance to the gaseous phase is a very recent development and it is probable that its use will be extended in the near future. Its high inherent sensitivity opens up a prospect that one may be able to measure steady-state concentrations of reactive intermediates during the course of a reaction under favourable conditions Chemical Methods.-The high chemical reactivity of atoms has fre- 47 Ingram “Free Radicals as studied by Electron Spin Resonance” Butterworths 4B Strandberg and Krongelb J. Chem. Phys. 1959 31 1196. London 1958. Heald and Beringer Phys. Rev. 1954 96 645. 254 QUARTERLY REVIEWS quently served as a basis for a method of detecting their presence. It is the only method which can be used in photochemical systems since the steady-state concentrations of the atoms are too low to be measured by physical methods.In complex systems however unless the mechanism is well understood product analysis may not provide unequivocal evidence for the presence of atoms. One of the earliest examples of the use of chemical reactivity to demon- strate the presence of atoms in a photochemical system was the reduction of metallic oxides by hydrogen atoms produced by the mercury-photo- sensitised decomposition of hydr0gen.l' Melville and Robb recently used molybdenum trioxide for quantitative this oxide is a yellow powder which becomes blue on exposure to hydrogen atoms. By measuring the rate at which the oxide becomes blue in the presence and the absence of an olefin and by calibrating the apparatus by using the rate of conversion of para-hydrogen into ortho-hydrogen absolute rate constants were obtained for the reaction of hydrogen atoms with a number of olefins.Pearson Robinson and Stoddart51 used the Paneth mirror technique to detect the presence of hydrogen atoms. An antimony mirror on the walls of the tube is slowly removed by the formation of a volatile hydride. A lead mirror which is not affected by the atoms is used to remove alkyl radicals before they reach the antimony mirror. This method has not been used for oxygen or nitrogen atoms since the mirrors would rapidly be coated with a non-volatile oxide or nitride. Even when it is used with hydrogen atoms the results are not very reproducible because of poison- ing by traces of impurities and gaseous-phase reactions are preferable for quantitative work.For such a reaction to be of use two general requirements must be fulfilled (1) the reaction must be simple and the mechanism known and (2) the reaction must be much faster than any other reaction which removes atoms. A reliable method of following the extent of the reaction is also necessary and this may be accomplished photometrically if the reaction is chemiluminescent or alternatively by product analysis. Iodine has been used as a scavenger for hydrogen atoms and free radicals in the photolysis of a number of corn pound^.^^^ Iodine reacts rapidly with the species formed in the primary photochemical act forming compounds such as hydrogen iodide and alkyl iodides. This allows one to infer the nature of the species without the complications of secondary processes.Although the concentration of hydrogen atoms produced in a discharge may be estimated by physical methods it has been seen that reliable values for oxygen and nitrogen atom concentrations are more difficult to obtain and gaseous-phase titration of these atoms has become increasingly popular. The various methods used will be dealt with under types of atom. 50 Melville and Robb Proc. Roy. Soc. 1949 A 196,445. I1 Pearson Robinson and Stoddart Proc. Roy. SOC. 1933 A 142 275. JENNINGS ATOMS IN THE GASEOUS PHASE 255 (a) Deuterium. If deuterium atoms are produced in a fast-flow system by an electrical discharge or by a heated filament and are then mixed with an excess of ethylene the reaction D + C2H4 -+ Products is very rapid and effectively all the deuterium atoms are removed in this way.The products of the reaction are deuterated methanes ethanes and ethylenes and if it is assumed that all atoms present initially are to be found in these products an analysis of the products for deuterium will enable one to calculate the deuterium atom flow rate. Results obtained in this way agree to within 0.5% with those obtained by means of an iso- thermal cal~rimeter.~~ (b) Oxygen. Atomic oxygen reacts very rapidly with nitrogen dioxide according to the equation the rate constant for this reaction being52 in the region of 3 x ~ m . ~ molecules-l sec.-l. Spealman and R ~ d e b u s h ~ ~ added an excess of nitrogen dioxide to a stream of atomic oxygen and froze out the products at liquid- nitrogen temperature. The amount of nitric oxide formed in a given time was then determined enabling them to calculate the flow-rate of atomic oxygen.More recently Kaufman5* has used the above reaction in conjunction with the reaction 0 + NO + 0 + NO . . .(1) O+NO+NO,+hv ...( 2) If a little NOz is added to the stream of atomic oxygen nitric oxide is formed by reaction (1). This in turn reacts as in reaction (2) to produce the yellow-green air afterglow. If the flow rate of nitrogen dioxide is slowly increased more nitric oxide is formed but fewer atoms remain to react with the nitric oxide until finally a point is reached where the glow is just extinguished. At this point the flow-rate of the nitrogen dioxide is equal to that of the oxygen atoms. The results obtained by this method agree within experimental error with those obtained by using the method of Spealman and Rodebush and since the gaseous-phase titration is simpler it has generally been preferred.At pressures in the region of 1 mm. Hg this is probably the most reliable method of determining absolute con- centrations of oxygen atoms since excited molecules interfere with calori- metric methods and unless a sintered disc is used the pressure is too high for the use of a Wrede gauge. Relative atom concentrations down a tube can be determined photo- metrically by making use of the green glow emitted in reaction (2). If a little nitric oxide is added to a stream of atomic oxygen nitrogen dioxide is formed but since reaction (1) is faster than reaction (2) or the correspond- ing three-body reaction 0 + NO + M -+ NO2 + M the nitric oxide is s2 Ford and Endow J. Chem. Phys.1957,27 1156. 63 Spealrnan and Rodebush J. Amer. Chem. SOC. 1935,57 1474. 64 Kaufman Proc. Roy. Soc. 1958 A 247 123. 256 QUARTERLY REVIEWS immediately regenerated and its concentration remains constant down the tube. Hence the intensity of the glow is proportional to the concentration of oxygen atoms at any point and relative concentrations are readily found. These can be converted into absolute readings by calibrating the system by the nitrogen dioxide titration method or by means of the isothermal calorimeter. (c) Nitrogen. Two quite different chemical methods have been used to determine absolute concentrations of nitrogen atoms. The first of these makes use of the reaction N + N O + N + O . . .(3) in a way which is entirely analogous to Kaufman’s use of reaction (I) for oxygen atoms.If a little nitric oxide is added it is all destroyed by reaction (3) so that the system contains both oxygen and nitrogen atoms. Excited nitric oxide molecules may then be formed by slow combination of these atoms in the presence of a third body. Emission from these molecules accounts for the range of colours yellow-pink-blue as the nitric oxide content increases. If sufficient nitric oxide is added all the nitrogen atoms will be consumed and there will be no afterglow. At this point the flow- rate of nitric oxide gas is equal to that of nitrogen atoms and the stream of gas consists only of nitrogen molecules and oxygen If an excess of nitric oxide is added the green 0 + NO afterglow is produced (reaction 2). Nitrogen atoms were estimated in this manner in a recent study of the rate of recombination of nitrogen atoms by three-body collision^.^^ In an earlier investigation the effluent gas was continuously sampled mass-spectrometrically as the nitric oxide flow-rate was in- crea~ed.~’ At the “equivalence point” no nitrogen atoms or nitric oxide molecules could be detected and the flow-rate of nitric oxide was equal to that of the nitrogen atoms.The addition of nitric oxide to a stream of active nitrogen is a useful way of producing a stream of oxygen atoms in the absence of oxygen molecules and has recently been used for this purpose. 58 The above method of estimating nitrogen atom concentrations assumes that nitric oxide is destroyed only by the atoms and not by any other excited species. This has been disputed by Verbeke and Winkler who determined the atom concentration by measuring the rate of production of hydrogen cymide in the reaction with ethylene at 400”c.Although the mechanism of this reaction is not well-established the reaction is fast and products containing nitrogen other than hydrogen cyanide are formed in very small amounts and it is assumed that all nitrogen atoms produce hydrogen cyanide in the presence of an excess of ethylene.59 Concentra- 65 Kaufman and Kelso J. Chem. Phys. 1957,27,1209; 1958,28,992. 56 Harteck Reeves and Manella J. Chem. Phyr. 1958 29 608. b7 Kistiakowsky and Volpi J . Chm. Phys. 1957 27 1141. 6R Morgan Elias and Schiff J. Chem. Phys. 1960,33 930. Is Verbeke and Winkler J. Phys. Chern. 1960,64 319. JENNINGS ATOMS IN THE GASEOUS PHASE 257 tions determined in this way were compared with results obtained by the nitric oxide titration method at pressures of 1-16 mm.Hg. At all pres- sures it was found that the latter method gave much higher values ranging from 40% higher at 1 mm. Hg to 140% higher at 16 mm. Hg. Verbeke and Winkler suggest that the nitric oxide may be destroyed either by excited nitrogen molecules in the A state or perhaps by some sort of chain reaction. Since the reaction of nitrogen atoms with ethylene is not simple it seems possible that not all the nitrogen atoms react to form hydrogen cyanide thereby leading to a low result. However a number of different hydro- carbons differing widely in structure gave the same maximum rate of hydrogen cyanide production. This suggests either that all the hydro- carbons were equally efficient catalysts in causing nitrogen atoms to recombine which is most unlikely or that the rate of production of hydrogen cyanide is a true measure of the concentration of the atoms.It is possible that the different types of discharge used may explain the discrepancy. Verbeke and Winkler used a condensed electrode discharge which is known to produce a reactive species in addition to ground state The microwave discharge which is more generally employed does not appear to produce such molecule^.^^^ Ammonia is thought to react with the second excited species,Go but the concentration measured by the nitric oxide titration technique was unaffected when ammonia was injected between the microwave discharge and the nitric oxide jet,G1 again suggesting that the microwave discharge does not produce appreciable concentrations of the second excited species.In addition although ammonia is decomposed to some extent by the products of a condensed discharge through nitrogen,60 no reaction was found with the products from a microwave discharge.62 Until a little more is known about the second excited species it seems reasonable to suppose that the nitric oxide titration method is probably quite reliable when used with a microwave discharge but that the N + C2H reaction at 400"c is the more reliable method when a condensed discharge is used. Conclusion Although considerable progress has been made in our understanding of atomic reactions there is as yet no substantial body of quantitative results. The energies of activation of many reactions can be estimated only by assuming a steric factor of 0.1 and experience has shown that this is often in error by several factors of ten. Reliable rate constants have been obtained for many reactions at room temperature but much still remains to be done on the effect of temperature and surfaces and the presence of 6o Freeman and Winkler J. Phys. Chem. 1955,59 371. 41 Herron Franklin Bradt and Dibeler J. Chem. Phys. 1959 30 879. 62 Kistiakowsky and Volpi J. Chem. Phys. 1958 28 665. 258 QUARTERLY REVIEWS different third bodies in many cases. In addition little is known about the reactions of excited atoms which are certainly present in some of the more energetic reactions. The author thanks Dr. J. W. Linnett for reading the manuscript of this Review.

ISSN:0009-2681    

DOI:10.1039/QR9611500237

出版商:RSC    

年代:1961

数据来源: RSC

摘要:

ALKALOID BIOSYNTHESIS By A. R. BATTERSBY (THE UNIVERSITY BRISTOL) SINCE the early days of biogenetic theory there has been much thought and discussion about the way in which plants build up the rich variety of alkaloid structures. The resultant speculations have largely been based upon structural similarities within the alkaloid series and also upon the relations of alkaloids to simpler natural products. Obviously such an approach cannot prove and is not claimed to prove that plants follow the suggested biosynthetic schemes. The great value of the proposals lies in the help they can give to structural studies on new alkaloids and in prompting experiments on living plants. Further progress depends upon researches in vivo and the present survey will cover the results gained so far from tracer experiments on alkaloid biosynthesis.The studies which predate the use of radioactive tracers were often beset with difficulties and the results obtained are fully reviewed elsewhere;ly2 they will therefore not be discussed here. Nor is it intended to give an account of the present state of biogenetic theory but rather to bring theory and experimental result together in those cases where the living plant has been studied. The usual technique for biosynthetic studies of alkaloids has been to administer labelled precursors to the plants and after a suitable period of growth to isolate the alkaloids. These are then degraded in a controlled way to determine the positions of the labelled atoms. It must be emphasised that negative results should be treated with great caution. Thus they may simply mean that the plant is not synthesising the alkaloid of interest at the time of the experiment or again that the precursor used has not reached the site of synthesis.Even when incorporation has been achieved the results must be carefully interpreted. It must always be borne in mind that incorporation of a substance which is not normally involved in the biosynthesis of an alkaloid could occur by transformation into one of the intermediates on the actual biosynthetic route. However with proper care in execution and interpretation the tracer technique is extremely powerful in this field. Pyridine and Piperidine Alkaloids.-The biosynthesis of nicotine (l) the major alkaloid of many Nicotiana species has been extensively studied by tracer methods. This alkaloid also occurs in club mosses3 (Lycopodium) in "horse- tail^"^ (Equisetum) and in many other genera,* so that it is R.F. Dawson Adv. Enzymol. 1948 8 203; W. 0. James “The Alkaloids,” ed. R. H. F. Manske and H. L. Holmes Academic Press New York 1950 Vol. I p. 16. K. Mothes Handbuch Pflanz. physiol. 1958 8 989. a Reviewed by R. H. F. Manske “The Alkaloids,” ed. R. H. F. Manske and H. L. Holmes Academic Press New York 1955 Vol. V p. 295; R. H. F. Manske and L. Marion Canad. J. Res. 1942 20 B 88. K. Mothes J. Pharm. Pharmacol. 1959 11 193. 259 260 QUARTERLY REVIEWS remarkably widespread in Nature. When [2-14C]ornithine (4) was fed to N. tabacum and N. rustica plants it was found5s6 that this amino-acid is a good precursor of nicotine. Degradation of the radioactive alkaloid by oxidation with nitric acid gave nicotinic acid (2) and the nitropyrazole (3).The carboxyl group5p6 of the acid (2) and the nitropyrazole7 (3) each contained within the limits of experimental error,? half the activity of the original nicotine; the N-methyl group of the alkaloid was inactive. Thus the activity is equally divided between positions 2’ and 5’ of nicotine and this has been confirmed by a further degradation8 which allowed the isolation of the carbon atom from position 5’. Clearly these results mean that a symmetrical intermediate is being formed and Leekg suggests putrescine (8) or the mesomeric anion (7) as possibilities. It may be significant that a plant oxidase is known which will convert putrescine into 1-pyrroline (9); this enzyme has been used in very significant in vitro experirnents.l0 n Nicotine (I) Support for the above scheme comes from the incorporation of labelled putrescine (8) proline and glutamic acid into the pyrrolidine ring of nicotine though compared with ornithine these substances are less efficient precursor^.^ In the case of glutamic acid however the percentage incorporation will be markedly affected by the large poolll of the free glutamic acid present in tobacco plants.t To avoid tedious repetition in the sequel the words “within the limits of experi- cental error” should be understood in all cases where tracer results are reported. E. Leete Chem. and Ind. 1955 537; L. J. Dewey R. U. Byerrum and C. D. Ball Biochim. Biophys. Acta 1955 18 141. E. Leete J. Amer. Chem. Soc. 1956 78 3520; 1958 80 4393. ’ E. Leete and K. Siegfried J. Amw. Chem.Soc. 1957 79 4529. * B. L. Lamberts L. J. Dewey and R. U. Byerrum Biochim. Biophys. Acta 1959,33 22. E. Leete J. Amer. Chem. SOC. 1958,80,2162; B. L. Lamberts and R. U. Byerrurn J. Biol. Chem. 1958 233 939. A. J. Clark and P. J. G. Mann Biochem. J. 1959 71 596; cf. H. Tuppy and M. S. Faltaous Monatsh. 1960 91 167. lo P. J. G. Mann and W. R. Smithies Biochern. J. 1955 61 89. l1 B. Commoner and N. Varda J Gen. Physiol. 1953,36,791; E. A. H. Robeits and D. N. Wood Arch. Biochem. Biophys. 1951 33 299. BATTERSBY ALKALOID BIOSYNTHESIS 261 Similar feeding experiments6 with Nicotiana gfauca have established that lysine (10) and cadaverine (1,5-diaminopentane) can serve as precursors of the piperidine ring in anabasine (1 1). In contrast to the case of ornithine above the biosynthesis from lysine does not involve a symmetrical inter- mediate since over 90 % of the radioactivity in the alkaloid biosynthesised n (1 0) 0 I> (I 2) 0 3) from [2-14C]lysine (lo) was located at position 2’ of the piperidine ring.Probably both lysine and cadaverine are converted into dl-piperideine (13) with lysine undergoing oxidation first at the a-position to give the acid (12) ; in this way the specific labelling of one position would be preserved. This certainly seems to be the main route from lysine to pipecolic acid (piperidine-2-carboxylic acid) in the rat12 and in Neurospora crassa.13 The fact that the pyridine ring of anabasine was not labelled in the above experiments with lysine and cadaverine shows that the aromatic system is not derived directly from either of these precursors.These results are in agreement with the isolation6 of inactive nicotine from N . tabacum plants fed with [2-14C]lysine; also nicotine of very low activity was obtained after the roots of the plants had taken up generally labelled 1y~ine.l~ Fortunately positive information concerning the origin and mode of incorporation of the pyridine ring is rapidly accumulating. Nicotine labelled in the pyridine ring was isolated from plants fed with ring-labelled nicotinic acid,15 but when carboxyl-labelled nicotinic acid was used16 inactive nicotine resulted. Thus the plant can decarboxylate nicotinic acid at some stage in the incorporation of the preformed pyridine ring into nicotine. These experi- ments have been extended by feeding 2- 4- 5- and 6-tritium-labelled nicotinic acid and nicotinic acid generally labelled with tritium to sterile tobacco root c~1tures.l~ The results show that the hydrogen atoms at positions 2 4 and 5 are unaffected in the conversion of nicotinic acid into nicotine.The hydrogen at position 6 is lost and a reasonable explana- tion of this would be the intermediate formation of the pyridone (14). However when this labelled precursor was fed to such cultures,17 there was virtually no incorporation into nicotine. A similar result17 was obtained by feeding the pyridone (1 5 ) labelled with tritium at position 2. l2 M. Rothstein and L. L. Miller J. Amer. Chem. SOC. 1954 76 1459. l3 R. S. Schweet J. T. Holden and P. H. Low J. Biol. Chem. 1954,211 517. l4 A. A. Bothner-By R. F. Dawson and D. R. Christman Experientia 1956,12 151. l5 R. F.Dawson D. R. Christman R. C. Anderson M. L. Solt A. F. D’Adamo and U. Weiss J. Amer. Chem. SOC. 1956 78 2645; T. C. Tso and R. N. Jeffrey Arch. Biochem. Biophys. 1959 80,46. l6 R. F. Dawson D. R. Christman and R. C. Anderson J. AmPr. Chem. SOC. 1953 75 5114. l7 R. F. Dawson D. R. Christman A. F. D’Adamo M. L. Solt and A. P. Wolf Chem. andInd. 1958,100; J. Amer. Chem. SOC. 1960,82,2628. 262 QUARTERLY REVIEWS flC0.,, hz (15) fPH O ; (I 4) The above researches deal with a preformed pyridine ring so that the way in which this ring is constructed still remains to be discussed. Two possible precursors ring-labelled tryptophanl8Jg (16) and ring-labelled anthranilic acid,20 yielded inactive nicotine ; both amino-acids were fed to intact N. tabacum plants. Thus it seems that the route used in animals and Neurospora21 from tryptophan (1 6) -+ kynurenine -+ 3-hydroxyanthranilic acid (17) + nicotinic acid (1 8 ; R = OH) is not followed in the higher plants.There is perhaps further evidence for this suggestion in the failure of ring- labelled tryptophan and tritium-labelled 3-hydroxyanthranilic acid (17) to act as precursors of nicotinamide (18; R = NH,) in maize.19 Also (19) (2 0) NHMe Me0 Me [3-14C]tryptophan was not incorporated into damascenine (19) in NigeZZa damascena or trigonelline (20) in pea seedlings,22 nor would soya-bean leaves convert ring-labelled 3-hydroxyanthranilic acid (1 7) into trigonel- line.23 Damascenine (19) and trigonelline are obviously closely related structurally to the amino-acids (1 7) and (1 8) respectively. These negative results are now being followed by positive ones and at the time of writing the field is in a state of rapid development.Whereas [ lJ4C]acetate was incorporated by tobacco plants almost entirely into the pyrrolidine ring of nicotine2* (2 l) [2J4C]acetate yielded nicotine with about 40 % of the activity in the pyridine nucleus.24; cf.25,26 [l-14C]Propion- ate gave almost inactive nicotine but [2-14C]propionate led to active alkaloid with 39% of the activity in the pyridine The highest incorporation into nicotine was achieved with [ 1 ,3-14C]glycerol as the precursor and here the isolated alkaloid had 57% of its activity in the l9 L. M. Henderson J. F. Someroski D. R. Rao P.-H. L. Wu T. Griffith and R. U. 2o J. Grimshaw and L. Marion Nature 1958 181 112. 21 C. E. Dalgliesh Quart. Rev. 1951 5 227; A.H. Mehler in “Amino Acid Meta- 22 E. Leete L. Marion and I. D. Spenser Cunud. J. Chem. 1955,33,405. 23 S . Aronoff Plant Physiol. 1956 31 355. 24 T. Griffith and R. U. Byerrum Science 1959 129 1485. 25 G. S. Win Doklady Akud. Nuuk S.S.S.R. 1958 119 544. E. Leete Chem. undInd. 1958 1477. T. Griffith K. P. Hellman and R. U. Byerrum J. Biol. Chem. 1960 235 800. E. Leete Chem. undInd. 1957 1270. Byerrum J. Biol. Chem. 1959 234 94. bolism,” ed. W. D. McElroy and B. Glass Johns Hopkins Press Baltimore 1955. BATTERSBY ALKALOID BIOSYNTHESIS 263 aromatic nucleus.27 Degradation of the nicotine from the [2-14C]acetate and [2-14C]propionate experiments was also carried out to yield hygric acid (22) as shown; this allowed the level of activity at position 3 of the aromatic ring to be examined.27 The present interpretation of all the results obtained is that glycerol is involved in the construction of the C-4 C-5 C-6 system of nicotine (21) and that C-2 and C-3 arise respectively from C-3 and C-2 of propionate.Clearly the carboxyl group of propionate is eliminated in the process. These suggestions are to some extent tentative27 but have an analogy in the recent demonstration that glycerol is capable of supplying all the carbon atoms of nicotinic acid in Escherichia coli2* Further developments will be watched with interest. The labelling patterns in the pyrrolidine ring of the various nicotine samples derived from the simple precursors just discussed strongly suggest that the precursors enter the tricarboxylic acid cycle and are converted into glutamic acid before incorporation.A related experiment in which [2-14C]acetate was fed to Nicotiana glauca plants26 afforded radioactive anabasine (23) with over 90% of the activity concentrated in the aromatic ring. It is not surprising that in contrast to the result with nicotine (21) the incorporation into the reduced ring is low. It has been mentioned that acetate can readily pass into the ornithine-proline-glutamic acid group by way of the tricarboxylic acid cycle whereas acetate is metabolically much further removed from l y ~ i n e ~ ~ the precursor of the piperidine ring of anabasine. When the feeding experiment was repeated26 with a large quantity of inactive nicotinic acid together with the [2-14C]acetate almost inactive anabasine was isolated and this suggests that acetate is converted into nicotinic acid before incorporation.There has been considerable interest in the source of the N-methyl group of nicotine and here tracer experiments show that it can arise from choline30 or rnethi~nine.~~ In the latter case it is established that the methyl group is transferred intact that is a true transmethylation occurs.32 Separate feeding experiments with [2-14C]gly~ine,33 [14C]formate,31 [14C]formalde- E. Bilinski and W. B. McConnell Canad. J. Biochem. Physiol. 1957 35 357. 28 M. V. Ortega and G. M. Brown J. Amer. Chem. SOC. 1959,81,4437. 30 R. U. Byerrum and R. E. Wing J. Biol. Chem. 1953 205 637. 31 S. A. Brown and R. U. Byerrurn J. Amer. Chem. SOC. 1952,74 1523. 38 L. J. Dewey R. U. Byerrum and C. D. Ball J. Amer. Chem. Suc. 1954,76,3997. 33 R. U. Byerrurn R. L. Hamill and C.D. Ball J. Biol. Chem. 1954,210,645. 264 QUARTERLY REVIEWS I ! h ~ d e ~ ~ [3J4C]serine and [2-14C]g1y~ollate35 all gave radioactive nicotine with most or all of the activity in the N-methyl group; [l-14C]glycine however yielded inactive nicotine. This array of precursors is not surpris- ing when it is recalled that some of them for example glycine serine and glycollate are interconvertible in living systems. The fact that the N-methyl group of nicotine is labelled after feeding methyl-labelled methionine could be interpreted as a net synthesis from nornicotine or some other unmethylated precursor or simply as a transfer of methyl groups between nicotine and methionine. It is that administration of [Me-14C]- nicotine to intact tobacco plants resulted in the formation of [Me-14C]- choline so that methyl transfer from the alkaloid to acceptors can occur.The results of the many tracer experiments outlined above can now be summarised in the biosynthetic scheme below. It seems probable that the conversion A does not involve the 6-pyridone of nicotinic acid. In this scheme and in those which appear later in the Review a broken circle with attached arrow has been used to indicate that the arrowed group is derived from the ringed group. Thus here the N-methyl group of nicotine arises by transfer of the S-methyl group from methionine. A n I Ana basine CO,H *,’ and other I Y precursors Nicotine Only ricinine (25) among the other pyridine alkaloids has so far been studied by tracer methods. It was isolated with its cyano-group labelled from castor-bean plants (Ricinus communis) which had been fed with carboxyl-labelled nicotinic acid ;37 the incorporation was 0-76 %.More- over [ l -14C]acetate and [2-14C]acetate are both incorporated into ricinine.26 The two methyl groups can be derived from methionine but not from formate choline or hydrogen carbonate under the conditions used in the feeding experiment^.^^ In contrast to the results with nicotine it is 34 R. U. Byerrurn R. L. Ringler and R. L. Hamill Fed. Proc. 1955 14 188. 35 R. U. Byerrum L. J. Dewey R. L. Hamill and C. D. Ball J. Biol. Chem. 1956 36 E. Leete and V. M. Bell J. Arner. Chem. Soc. 1959 81 4358. 37 E. Leete and F. H. B. Leitz Chem. and Ind. 1957 1572. 38 M. Dubeck and S. Kirkwood J. Biol. Chem. 1952 199 307. 219 345. BATTERSBY ALKALOID BIOSYNTHESIS 265 reported that [2-14C]lysine (24) can serve39 as a precursor of the pyridine nucleus to give ricinine labelled at position 6 though the incorporation was very low (0.01 %).Indeed other workers4* state that lysine is not used for ricinine formation. When a-amino [~-~~C]adipic acid (26) was fed to the ricinine equally labelled at positions 2 and 6 was isolated with incorporation of 0.13 %. The fact that different labelling patterns result from the unsymmetrical precursors (24) and (26) demands two routes to *\-.I* N ricinine and the ones suggested in the chart will rationalise the results.39 All one can say at present is that the various findings are hard to reconcile and moreover are rather disturbing as they stand since they indicate a lack of specificity in the biosynthetic activities of castor-bean seedlings.The notorious hemlock (Conium maculatum) has recently been the subject of tracer experiments. These plants used generally labelled lysine (27) to yield radioactive coniine (28) but the alkaloid was not degraded.41 Further work will therefore be necessary to distinguish beyond all doubt between the two possible biosynthetic schemes below based on acetate and ammonia,42 and lysine and respectively. Analogy with anabasine suggests that the route from lysine will turn out to be the correct one. Pyrrolidine and Tropane Alkaloids.-The tropane alkaloids are based upon the skeleton (29) which is obviously closely related to the simple 3e H. Tamir and D. Ginsburg J. 1959 2921. 40 E. J. Reist and L. Marion quoted in ref. 20. 41 U. Schiedt and H. G. Hoss 2. Naturforsch.1958 13b 691. Cf. K. Biemann G. Buchi and B. H. Walker J. Amer. Chem. Soc. 1957 79 43 Sir R. Robinson “The Structural Relations of Natural Products” Clarendon 5558. Press Oxford 1955. 266 QUARTERLY REVIEWS pyrrolidine bases such as hygrine (30). Indeed it is possible43 that one of the stages in tropane formation is cyclisation of the pyrroline (30A). Such a linking of the pyrrolidine and tropane types is attractive because of the occurrence of hygrine (30) with cocaine (32) in Erythroxylon truxillense and of a related pyrrolidine base with hyoscyamine (33) in Atropa bellu- dona. 44 has shown that when [2-14C]- ornithine is fed to five-months old Datura stramonium plants it is in- corporated into hyoscyamine (33). Hydrolysis of the alkaloid gave The work of Marion and his (32) (36) .. . . . . . . . . . . H02C’ * Ph i + Y..... (35) CH;OH Ph + 2Ph-?02H CO H [‘:,,Me 3 co tropine (34) which was degraded as illustrated. The imide and the diphenyl- pyrrole both had the same specific activity as the original tropine so that carbon atoms 2 3 and 4 must have been inactive. Moreover since the benzoic acid had half the specific activity of the tropine it follows that the original activity of the hyoscyamine is located at position 1 or 5 or is distributed between these two positions. From this experiment one cannot say which of these possibilities is correct because of the symmetry of the degradation products. Nevertheless it is clear that ornithine is incorporated in a specific manner so that the biosynthetic scheme on p. 267 is a probable one. It should be mentioned that there is as yet no evidence that the nitrogen atom of the alkaloid is one of those originally present in ornithine and Wenkert46 has suggested that the true precursor of the pyrrolidine ring in 44 P.R. van Haga Nature 1954 174 833. 45 E. Leete L. Marion and I. D. Spenser Nature 1954 174 650; Canad. J. Chern. 46 E. Wenkert Experientia 1959 15 165. 1954,32 1 116. BATTERSBY ALKALOID BIOSYNTHESIS 267 the tropane series is D-erythrose 4-phosphate (37). One attraction of his theory is that it can give a plausible account of the oxygenated alkaloids such as hyoscine (31). However there is now evidence4' that young D. -G NMe CO J CO H ,CO - (31) and (33) stramonium plants can bring about the oxidation of hyoscyamine (33) to hyoscine (3 1). Three carbon atoms of the tropane system remain to be accounted for and it is found that Datura metel roots will incorporate labelled acetate into these alkaloid^.^^ The active hyoscyamine (33) produced was hydrolysed to give tropine (34) which retained most of the activity of the original alkaloid.Oxidation of the tropine to N-methylsuccinimide then allowed the amount of activity concentrated in carbon atoms 2 3 and 4 to be estimated. When [1-l4C]-acetate was fed 85 % of the activity of the tropine was located somewhere in this three-carbon chain and when [2-14C]acetate was used the figure was 75%. These results are consistent with the biosynthetic schemes given above. The hyoscine (31) which was also isolated from the D . stramonium plants fed with radioactive ornithine was to be completely inactive.At first this result was a surprising one but it was later found that incorporation occurred only into the N-methyl group of hyoscyamine (33) when [MeJ4C]methionine was fed to D. stramonium plants of the same age as used in the foregoing experiment^.^^ This is good evidence that hyoscine (31) is not being synthesised at this stage in the plants' development and there are other indications50 that the hyoscine present in D. stramonium is produced by the young plants whereas adult plants synthesise hyoscyamine exclusively. It is also established by tracer experi- ments that plants two months old are synthesising both alkaloids.5f This case illustrates how little trust can be placed in completely negative results. The biosynthesis of the tropic acid residue (35) in the alkaloids (31) and (33) has been studied by feeding [3J4C]phenylalanine (36) to D.stra- inonium plants;51 the acid (35) was then obtained by hydrolysis of the isolated alkaloids. Suitable degradation showed that the radioactive tropic acid had the label located almost entirely at the starred position. 47 A. Romeike and G. Fodor Tetrahedron Letters 1960 22 I . '* J. Kaczkowski H. R. Schiitte and K. Mothes Naturwiss. 1960,13 304. 4B L. Marion and A. F. Thomas Canad. J. Chem. 1955,33 1853. 5 0 E. M. Trautner Austral. Chem. Inst. J. Proc. 1947,14,411. j1 E. Leete J. Amer. Chem. SOC. 1960 82 612. 268 QUARTERLY REVIEWS Thus the C6-C1 fragment of phenylalanine can provide the boxed part of the molecule in formula (35); the carboxyl group may or may not have been derived from C-2 of the original phenylalanine.Surprisingly neither formate nor formaldehyde was incorporated into the hydroxymethyl group of tropic acid (35) and the origin of this feature remains to be elucidated. As in the case of nicotine formate was used in the synthesis of hyoscyamine almost entirely as a precursor of the N-methyl Stachydrine (40) from alfalfa is the only simple pyrrolidine alkaloid to have been studied by modern methods. It now appears that the synthesis is straightforward in that both [~arboxy-~~C]proline~~ (38) and [carboxy- 14C]hygric acid53 (39) when fed separately to alfalfa plants gave rise to radioactive stachydrine (40) though the former amino-acid is only in- corporated when mature plants are used. It thus seems that the enzyme system necessary for the conversion of proline into hygric acid is absent in young plants.The methyl groups of stachydrine were labelled after [Me-14C]methionine had been Incorporation of proline leads one to expect that ornithine and glutamic acid would also be used by mature plants to form stachydrine and it would be interesting to re-examine the earlier which was carried out on very young plants. Hexahydropyrrolizine and Octahydroquinolizine Alkaloids.-Tracer studies have yet to be reported for simple alkaloids based on the hexa- hydropyrrolizine (pyrrolizidine) system (4 1). For the octahydroquinolizines (quinolizidines) the published knowledge is that [2-14C]lysine and [ 1 ,5-14C]- cadaverine (cadaverine = 1,5-diaminopentane) are good precursors of lupinine (42) and sparteine (45) in Lupinus l ~ t e u s ; ~ ~ the alkaloids were not degraded.These results are in keeping with but do not as yet establish the proposal43 that lupinine is derived from a precursor such as the dialdehyde (43) formed in turn from two molecules of lysine. A laboratory synthesis of (&)-epilupinine (epimeric with lupinine at C-1) by this approach has recently been achieved.56 It has been proposed43 that sparteine biosynthesis involves the inter- mediate (44) which could be formed from two lysine residues one aceto- acetate residue (or its equivalent) and two formaldehyde equivalents 52 L. Marion and D. J. McCaldin personal communication. 53 A. V. Robertson and L. Marion Canad. J. Chem. 1960,38,396. E. Leete L. Marion and I. D. Spenser J. Biol. Chem. 1955 214 71; A. Morgan and L. Marion Canad. J . Chem. 1956 34 1704; G. Wiehlei and L.Marion J. Biol. Chem. 1958 231 799; A. V. Robertson and L. Marion Canad. J. Chem. 1959 37. 1197. 55 H. R. Schutte and E. Nowacki Naturwiss. 1959 46,493. 56 E. E. van Tamelen and R. L. Foltz J. Amer. Chern. Soc. 1960. 82 502; cf. N. J. Leonard and S . W. Blum ibid. p. 503. BATTERSBY ALKALOID BIOSYNTHESIS 269 though there are other possibilities ;43 clearly the amino-aldehyde (44) could be in the carbinolamine form (46) rather than the open form. This scheme forms the basis of a recent neat synthesis of sparteine (45) in vitm6' In addition tetracyclic lupin alkaloids are appearing58 which are oxygen- ated at position 8 thus giving some support to the idea that an intermediate is used which is oxidised at this position. An alternative biosynthetic scheme46 which at present is equally attractive involves the dialdehyde (47); this could be derived from lysine or possibly from shikimic acid.C H,*OH c1$ c+j 612) (43) %?' COzH NH2 G o (41) pp NH2 __c QN Mtcy CH20 NfJ .c- %PJ C - O P 5 - J - N\H \OH (Y&yJ- H 0,C / @ 4) Steroidal Alkaloids.-Various Solanum species produce steroidal bases in the form of glycosides which can be hydrolysed to the correspond- ing aglycones a typical example being solanidine (48). This base was H found to be radioactive after [I J4C]acetate had been fed59 to sprouting potatoes (Solanurn tuberosum). The same precursor was used by tomato plants (Lycopersicon pimpinellifolium) in the construction of tomatine,gO a glycoside based upon the aglycone tomatidine (49). These results suggest that the acetate-mevalonate pathway is being used for the biosynthesis of these bases.b7 E. E. van Tamelen and R. L'. Foltz J . Amer. Chem. SOC. 1960,82,2400. 58 M. Carmack quoted by E. E. van Tamelen in ref. 57. 5 0 A. R. Guseva and V. A. Paseshnichenko Biokhimiya 1958,23,412. 6o H. Sander and H. Grisebach 2. Naiurforsch. 1958 13b 755. 2 270 QUARTERLY REVIEWS Phenethylamine and Isoquinoline Alkaloids.-One of the most valuable contributions of biogenetic theory has been the proposal of reasonable routes to the vast number of alkaloids in this group starting with the same relatively simple precursors. It was sugge~ted~l,~~ before the advent of tracers that the main building stones are the aromatic amino-acids phenyl- alanine (SO) tyrosine (5 l) and 3,4dihydroxyphenylalanine (54). De- carboxylation oxidation and methylation which are well known in living systems are then required in order to convert these precursors into such simple alkaloids as hordenine (55; R = R‘ = Me) and the hallucinogen mescaline (53).There is now sound evidence that plants do in fact use this route.62 Thus phenylalanine (50) tyrosine (51) and tyramine (52) all labelled with 14C at the starred position were found in separate experiments to be incorporated by sprouting barley into N-methyltyramine (55; R = H; R’ = Me) and hordenine (55; R = R’ = Me). Suitable degradation of the alkaloids showed that the label was located entirely at the starred position and the quantitative results suggested that hordenine (55; R = R‘ = Me) is formed from tyramine (55; R = H; R‘ = Me) in a stepwise methylation process. This was supported by the results from a study of the methylation by l*C-tracer methods in which it was shown that ~-methionine~~ (56) and b e t a i ~ ~ e ~ ~ (57) can serve as sources of the methyl groups in these alkaloids; formate was a poor methyl precursor and choline was ineffect- The biosynthesis of the barley alkaloids can thus be summarised in HO v ; R p \ (55)*..+ CH2 + .- &sj.M.e).NH . . . f . . . . . . . . -. . I - * * - - - * - * ..... (,)-*2=,/% (56) H02 the annexed scheme. However the first step A is not necessarily on the normal pathway; for just as it is probable that in wheat and buckwheat phenylalanine and tyrosine arise from prephenic acid by separate routes,65 as they do in Escherichia co1i,66 so it may well be that the major route in 61 E. Winterstein and G. Trier “Die Alkaloide” Borntraeger Berlin 1910.62 E. Leete S . Kirkwood and L. Marion C a d . J. Chem. 1952,30,749; E. Leete and L. Marion ibid. 1953,31 126; J. Massicot and L. Marion ibid. 1957,35 1 . 63 S. Kirkwood and L. Marion Canad. J. Chem. 1951 29 30; T. J. Matchett L. Marion and S. Kirkwood ibid. 1953,31,488; E. Leete and L. Marion ibid. 1954,32 646. G4 M. Sribney and S. Kirkwood Canad. J. Chem. 1954,32 918. 65 0. L. Gamborg and A. C. Neish Canad. J. Biochem. Physiol. 1959,37,1277. G6 1. Schwink and E. Adams Biochim. Biophys. Acta. 1959,36 102 and refs. therein. B. D. Davies Arch Biochem Biophys. 1958 78 497 and refs. therein. BATTERSBY ALKALOID BIOSYNTHESIS 27 1 barley is shikimic acid -+ prephenic acid -+ tyrosine and then on through the further stages as shown in the scheme. A similar study carried out with the cactus Lophophora williamsii has dem~nstrated~~ the conversion of [2-l4C]tyrosine into mescaline (53) with at least 99 % of the radioactivity at the starred position.Turning now to (-)ephedrine (62) and (+)-norpseudoephedrine (63) we find two new features the benzylic hydroxyl group and the C-methyl group. Clearly these bases are similar to noradrenaline (59) the biosyn- thesis of which in animal systems is generally believed6* to involve hydro- xylation of dopamine (58) derived in turn from tyrosine. When 15N-labelled phenylalanine and 15N-labelled alanine were fed separately to Ephedra distachya the important result was obtained that nitrogen-1 5 was incorporated into ephedrine from the former but hardly at all from the latter.69 This double experiment overcomes an objection to the use of 15N-labelled amino-acids that nitrogen may be transferred into and out of the general nitrogen pool of the plant by transamination.Thus in the ephedrine case it seems that the actual amino-acid is being .4 H H (62) &o*e OH H used rather than phenylpyruvic acid for example. The N-methyl group of ephedrine can be derived from methionine or from formate and formate also serves as a precursor of the C-methyl There have been complementary researche~'~ on (+)-no rpseudo- 67 E. Leete Chem. and Znd. 1959 604. 68 S. Udenfriend and J. B. Wyngaarden Biochim. Biophys. Acta 1956 20 48; M. Goodall and N. Kirshner J. Biol. Chem. 1957 226 213 821; G. Rosenfeld L. C. Leeper and S. Udenfriend Arch. Biochem. Biophys. 1958 74,. 252; H. Blaschko Brit. Med.Bull. 1957,13 162; S. Senoh C. R. Creveling S. Udenfnend and B. Witkop J . Amer. Chem. SOC. 1959,81,6236. S. Shibata and 1. Imaseki Pharm. Bull. (Japan) 1956,4,277; S . Shibata I. Imaseki and M. Yamazaki ibid. 1957 5 71 594. 'O E. Leete Chem. and Ind. 1958 1088. 272 QUARTERLY REVIEWS ephedrine (63) in which this alkaloid was obtained labelled at the starred carbon atom after [3J4C]phenylalanine had been taken up by a shoot of Catha edulis. Since in addition 14C-labelled w-aminoacetophenone (6 1) has been shown to act as a precursor of ephedrine (62) in E. distachya without randomisation of the labe1,'l the annexed tentative scheme can be written for the biosynthesis of these alkaloids. At present other schemes are equally acceptable in which for example decarboxylation or N- methylation occurs later in the sequence of events.The next stage in complexity in this group of alkaloids brings one to the simple isoquinolines such as pellotine (63A). This can be derived in theory43 from tyrosine and acetic acid or some equivalents of these substances and tracer experiment^'^ on Lophophora williamsii suggest that at least one part of the speculation is correct. When [2-14C]tyrosine was fed to the cactus radioactive pellotine (63A) could be isolated ; degradation of this material will yield information about the biosynthesis of one of the simplest isoquinoline alkaloids. HO As early as 1910 it was suggested6' that the benzylisoquinoline system known as norlaudanosoline (68) is formed from two molecules of di- hydroxyphenylalanine (64) by the sort of scheme illustrated.Methylation at some stage could give laudanosine (69) and dehydrogenation of the heterocyclic system which might possibly occur by way of the N-~xide,'~ 71 I. Imaseki S. Shibata and M. Yamazaki Chern. and Ind. 1958 1625. '* A. R. Battersby and S. Garratt unpublished work. 73 E. Wenkert Experientia 1954 10 346. BATTERSBY ALKALOID BIOSYNTHESIS 273 could lead to papaverine (66). Laudanosine and papaverine occur together in the opium poppy (Papaver somngerum). In addition to these two alkaloids many other isoquinoline bases have been regarded in biogenetic t h e ~ r y ~ ~ ~ ~ ~ ~ as being derived from norlaud- anosoline (68) as the key intermediate. Condensation with formaldehyde or some equivalent one-carbon unit could account for the biosynthesis of canadine (71) which is found in many CorydaZis species.Dehydrogenation of canadine could then afford berberine (72) whereas oxidative modifica- tion of the canadine skeleton (71) would give one possible route to the phthalide isoquinoline alkaloids such as hydrastine (67) which is found alongside berberine (72) in Hydrastis canadensis. Narcotine (70; R = Me) and narcotoline (70; R = H) would on this theory require an oxygenated canadine skeleton as the precursor. It is p o s ~ i b l e ~ ~ ~ ~ ~ ~ to draw a large number of other alkaloids into the sort of scheme illustrated by formula (64) to (72). to the phthalide isoquinoline bases involves the conversion of prephenic acid (73) into the hydrated derivative (74). This may undergo an acid-catalysed rearrangement and further A second hypothetical transformations as illustrated.Reaction of the hydroxylated o-carboxy- phenylpyruvic acid (75) with a phenethylamine could then give the phthalide-isoquinoline skeleton such as structure (70). Before going further with biogenetic theory in the isoquinoline series the experimental evidence for the cases considered so far should be examined. If the scheme tyrosine -+ (64) -+ (68) -+ (66) is in fact followed by the plant it can be seen that specifically labelled papaverine (66) should be obtained from the uptake of specifically labelled tyrosine. In fact it was found76 that when [2J4C]tyrosine is fed to Papaver somni- ferum plants it is incorporated into papaverine. The radioactive alkaloid was methylated and reduced to laudanosine (69) which was then degraded as shown to give the key fragment (76).The starred atoms were isolated separately from this by ozonolysis and by decarboxylation and each carried half of the activity of the original papaverine. The veratric acid (77) was inactive. The absence of labelling at positions other than the 74 R. B. Turner and R. B. Woodward “The Alkaloids,” ed. R. H. F. Manske and H. L. Holmes Academic Press New York 1953 Vol. 111 p. 54. 75 R. H. F. Manske J. 1954,2987; R. H. F. Manske “The Alkaloids,” ed. R. H. F. Manske and H. L. Holmes Academic Press New York 1954 Vol. IV p. 1. 76 A. R. Battersby and B. J. T. Harper Proc. Chem. SOC. 1959 152. 274 QUARTERLY REVIEWS starred ones is convincing evidence against the possibility that the ad- ministered tyrosine is degraded to small fragments in the plant and that these are then used in the biosynthesis of the alkaloid.This result thus establishes the biosynthesis of papaverine (66) from two tyrosine molecules. Me0 M e O S o M e / \ OMe 1. Oxidation. (69) Hofmann_ 2. Hofmann. It should be stressed that “biosynthesis from tyrosine” must at present be understood to include all the close biological equivalents of tyrosine such as tyramine (52) dihydroxyphenylalanine (64) and 3,4-dihydroxy- phenylpyruvic acid. It is known7’ that radioactive berberine (72) can be isolated from plants fed with [2-14C]phenylalanine but so far the necessary controlled degrada- tion of the alkaloid has not been reported. Tracer experiments have also been carried out on narcotoline (70; R = H); here 14C-generally labelled tyrosine was fed78 to Papaver somniferum plants. The radioactive narcoto- line so produced was then cleaved to cotarnoline (78) and meconine (79) (~o;R=H) - (78) OH + (79) which had activities consistent with the use of the carbon skeletons of two tyrosine molecules for the synthesis of narcotoline (70; R = H).However because the original tyrosine was generally labelled other interpretations of this interesting result are still possible. The suggestions involved in the scheme (64) -+ (72) and extensions of this scheme represent one line of thought in biogenetic theory for the isoquinoline alkaloids; a second line involves the proposal that the versatile norlaudanosoline (80; R = can undergo oxidative coupling. This could lead to a wide variety of alkaloids and the biosynthesis of one group of these has been extensively studied in the living plant.Accordingly this group containing morphine (86) codeine (87) and thebaine (83) will be discussed in the present survey. It was first recognised by Gulland and Robinson79 that if norlaudano- 77 J. L. Beal and E. Ramstad Naturwiss. 1960 47 206. 78 G. Kleinschmidt and K. Mothes 2. Naturforsch. 1959 14b 52. 7n J. M. Gulland and R. Robinson Mem. Proc. Manchester Lit. Phil. SOC. 1925 69 79. B A m B Y ALKALOID BIOSYNTHESIS 275 soline (80; R = H) [which is structure (68) re-written] undergoes oxidative coupling of the two aromatic rings then the skeleton of morphine (86) can in theory be obtained. There have been many suggestions about the mechanism of the coupling The most satisfying scheme proposed by Barton and Cohen,sl is illustrated here and is based upon knowledge gained in the elucidation of the structure of Pummerer’s ketone.82 The R groups may be methyl or they may represent part of an enzyme surface which is controlling the direction of the coupling process.The exact state of the nitrogen is left unspecified. Oxidation by some one- electron transfer system could generate radicals which if coupled would yield the dienone (81). The oxide bridge could then be formed by addition as shown to the enone system so leading to the base (82) which by obvious (86 R=H) (87 R =Me) changes could afford thebaine (83) codeine (87) morphine (86) and neopine (85). If an oxide bridge is not formed from the intermediate (81) then only simple changes need be postulated to account for sinomenine It is fortunate that tracer experiments with the opium poppy have been 81 D.H. R. Barton and T. Cohen “Festschrift A. Stoll” Birkhauser Basle 1957 p. 8a D. H. R. Barton A. M. Deflorin and 0. E. Edwards J. 1956 530. (84). C . Schopf Naturwiss. 1952,39,241; K. W. Bentley Experientia 1956,12,251. 117. 276 QUARTERLY REVIEWS very fruitful and a fairly complete picture of the biosynthesis of the alka- loids is emerging. When [2-14C]tyrosine is fed to Papaver somniferum plants it is i n c ~ r p o r a t e d ~ ~ ~ ~ ~ into morphine (86) codeines3 (87) and thebaines3 (83); the formation of labelled papaverine (66) in the same feeding experiment has already been discussed. Degradations3 of the labelled morphine by several stagess5 gave the perhydrophenanthrene (88) from which the terminal methylene group was isolated by suitable oxida- tion. The formaldehyde and the major fragment (89) each contained half of the activity of the original morphine (86) so that half the alkaloid's activity is located at position 16.A second degradationss converted Me + ?H,O (8 8) (89) the radioactive morphine into the phenanthrene (90) which was oxidised to the diphenic acid (91). The required carboxyl group was then selected by decarboxylation of the derived coumarin (92) followed by ring opening tw - !zzB* __c ?Zog2H ~ oc* M;% / / p,H (921 \ (9 0 \ w \ I CO + (94) (96) (95) to give the biphenyl-acid (93). Decarboxylation of this product gave carbon dioxide which carried half the activity of the original morphine and the biphenyl residue (94) was almost inactive. Thus the radioactive morphine (86) is labelled equally at positions 9 and 16. Independent degradative works4 has demonstrated that the active morphine derived from [2-14C]- tyrosine yields the phenanthrene (90) which was oxidised to phthalic 83 A.R. Rnttersby and B. J. T. Harper Chem. and Ind. 1958 364; A. R. Battersby 84 E. Leete Chem. and h d . 1958,977; J . Arner. Chem. Soc. 1959,81,3948. a5 R. Pschorr and F. Dickauser Ber. 191 I 44,2633; H. Rapoport and G. B. Payne 86 A. R. Battersby R. Binks and D. J. Le Count Proc. Chern. Soc. 1960 287. and B. J. T. Harper Tetrahedron Letters 1960 No. 27 21. J. Amer. Chcm. SOC. 1952,74 2630 and refs. therein. BATTERSBY ALKALOID BIOSYNTHESIS 277 acid. The phenanthrene (90) and the phthalic acid (95) each contained half the activity of the original morphine whereas anthranilic acid (96) prepared as shown contained half the activity of the phthalic acid.Be- cause of the equivalence of the carboxyl groups of phthalic acid these results mean that half the activity of the original morphine is located at position 9 or position 12 (or is spread between them) and it is argued8* that the first possibility is the correct one; the other half of the activity is in the eliminated ethanamine chain. The two degradative studies are thus in full agreement. The above results establish that two molecules of tyrosine or a close biological equivalent are built into morphine in the biosynthesis. More- over since two molecules of [2-14C]tyrosine would give a norlaudano- soline system (80) labelled equally at positions 1 and 3 (p. 275) the labelling pattern of morphine is exactly in agreement with the scheme on p.275. It seems probable that tyrosine is converted first into 3,4-dihydroxy- phenylalanine. Both groups of worker^^^^^^ found that [2-14C]phenylalanine can serve as a precursor of morphine but is much less efficiently incorporated than in tyrosine. Here again it may well be that the main route used is shikimic acid + prephenic acid + tyrosine -+ morphine and that phenylalanine enters only by conversion into tyrosine. The N-methyl group of morphine (86) and the N- and the 0-methyl groups of codeine (87) and thebaine (83) can be derived from methio- nine;88 formate is a poor precursor of these groups,88 and choline is ineffecti~e.~' One of the stages in the proposed biosynthesis of morphine which has attracted considerable interest is the ring-closure of norlaudanosoline (80; R = €3) or some protected derivative (80).Many unsuccessful attempts have been made to carry out this conversion in the laboratory but it has now been demonstrated in the plant.89 For this norlaudano- soline (80 ; R = H) labelled with 14C at position 1 was synthesised and fed to mature Papaver somngerum plants. The isolated morphine was highly radioactive. Indeed norlaudanosoline is the most efficient precursor of morphine so far used. Degradation of the morphine showedgo that all the activity is at position 9. Thus it is almost certain that the bond between positions 12 and 13 of morphine (86) is formed as was suggested in the scheme (p. 275) by a coupling reaction between two aromatic rings. It was further showng0 that the plants fed with labelled norlaudanosoline (80; R = H) yielded radioactive papaverine (66) which was proved by degrada- tion to be specifically labelled at position 1.Aromatisation of the reduced intermediate (80; R = €€) is thus demonstrated. The rate at which radioactivity is incorporated into morphine codeine A. R. Battersby and B. J. T. Harper unpublished work. A. R. Battersby and B. J. T. Harper Chem. andZnd. 1958,365. A. R. Battersby R. Binks and G. V. Parry unpublished work. 8 9 A. R. Battersby and R. Binks Proc. Chem. Soc. 1960,360. 278 QUARTERLY REVIEWS and thebaine from carbon dioxideg1 and [2-14C]tyr~~ine92 has been studied and it is found that in both cases the activity moves rapidly into the thebaine fraction and then on through codeine to morphine. The simplest interpretationg1sg2 of these results is that the biosynthesis runs first to thebaine (83) and then on through codeine (87) to morphine (86).This has been neatly confirmedg3 by feeding generally labelled thebaine to poppy plants codeine and morphine both labelled in their skeletons being isolated. Similarly generally labelled codeine was used by the plants to give labelled morphine but not labelled thebaine whereas no activity from generally labelled morphine passed into the skeletons of codeine and thebaine. These results show the importance of demethylation as a biosynthetic process. All the findings are thus falling nicely into place in the opium alkaloids. The use of two C&z units to provide the skeletons of morphine and papaverine and the conversion of norlaudanosoline into morphine and papaverine have been demonstrated in the plant. The order of formation of the hydrophenanthrene alkaloids is known and the use of thebaine as a precursor of morphine and codeine suggests strongly that phenolic coupling of the intermediate (80) is partly controlled by having R = Me.In this way many ring-closures which might occur when R = H are then blocked. It can be seen from the above biosynthetic scheme that phenolic coupling of the molecule (80; R = Me) leads readily to thebaine (83). The Amaryllidaceae Alkaloids.-The intensive search among the AmaryZZidaceae over the last few years has yielded a large number of alkaloids. Barton and CohenS1 suggest that their biosynthesis involves a precursor of the type (97; R = H or OH) possibly with some of the hydroxyl groups protected. This could give rise to the various types of AmaryZZidaceae alkaloid by suitable phenolic coupling.For example the product (98; R = OH) of ortho-para-coupling could be further modi- fied as shown to give lycorine (100). Norpluviine (101) could be formed in a similar way from the precursor (97; R = H). Significantly a simple derivative of the proposed precursor has emerged as belladine (99) in AmaryZiis If a simple derivative of the same precursor (97; R = €3) is re-written as (102) to be coupled as shown then the structures of narwedine (103) galanthamine (104) and narcissamine (105) can be built up. Equally satisfying routes to the other alkaloids in this group can be given.*l Wenke~-t,~~ on the other hand suggests the use of reduced precursors based upon shikimic and prephenic acid and the intermediate (106) derived from them could by reasonable steps be converted into lycorine (100).Here again the theory can embrace the other types of Amarylli- dizceae alkaloids. H. Rapoport F. R. Stermitz and D. R. Baker J. Amer. Chem. SOC. 1960,82,2765. O2 A. R. Battersby and B. J. T. Harper Tetraltedron Letters 1960 No. 27 21. O3 F. R. Sterrritz and H. Rapoport. NaturP 1961 189 310. O4 E. Warnhoff Chem. and I d . 1957 1385. BATTERSBY ALKALOID BIOSYNTHESIS 279 [2-14C]Tyrosine and the phenol (102) labelled with 14C on the N-methyl group have been showng5 to be incorporated by King Alfred daffodils into galanthamine (104). Good incorporations of [2-14C]tyrosine were also achieved into lycorine (100) when Twinkg6 and King Alfredg5tQ7 daffodils t H-0 ($ / Me MeO\ 0 Me0 o@ \ (I04 R=Me) {OS R= H) were used. In addition radioactive norpluviine (101) was isolated from the Twink plantsg6 The radioactive lycorine was subjectedg6 to Hofmann degradation and the terminal methylene was isolated from the methine (107) by glycol formation and periodate fission.The formaldehyde carried all the activity A. R. Battersby R. Binks and W. C. Wildman Proc. Chem. Soc. 1960,410. 95 D. H. R. Barton and G. W. Kirby Proc. Chem. Soc. 1960 392. 97 Personal communication from Professor D. H. R. Barton. 280 QUARTERLY REVIEWS of the original lycorine and further oxidation of the major fragment gave the lactam acid (108) which was inactive. This interlocking evidence establishes that the activity in the lycorine is located at the starred position. Tyrosine will thus provide that part of the lycorine molecule (100) drawn with heavy bonds and if one accepts that the conversion of prephenic acid into tyrosine is irreversible as seems probableYg8 thin these results support Barton and Cohen’s theory.Colchicine.-Colchicine (1 lo) which occurs in the autumn crocus presents particularly attractive biosynthetic problems because of its tropolone ring system and also because of the unusual position of the Ho2caco2H HO 0 11) MeO&fc02 nitrogen atom placed three carbons away from the aromatic ring A. [3-l4C]Pheny1alanine (109) was incorporatedg9 into colchicine (1 10) synthesised by Colchicum byzantinum and oxidation of the active alkaloid to the anhydride (1 15) followed by specific decarboxylation gave the benzoic acid (1 16). This still retained in its carboxyl group 93 % of the activity of the original colchicine.Thus the colchicine (110) derived from the precursor (109) was specifically labelled at position 5. An independent studyloo showed that [2-14C]tyrosine is incorporated into colchicine by Colchicum autumnale. If the C,-C system of tyrosine 98 I. Schwinck and E. Adams Biochim. Biophys. Ada 1959,36 102; 0. L. Gamborg lo0 A. R. Battersby and J. J. Reynolds Proc. Chem. Soc. 1960 346. and A. C. Neish Canad. J. Biochem. Physiol. 1959 37 1277. E. Leete and P. E. Nemeth J . Amer. Chem. SOC. 1960 82 6055. BAITEMBY ALKALOID BIOSYNTHESIS 28 1 provides ring A and atoms 5 6 and 7 of the alkaloid (1 lo) then in this experiment position 6 should be specifically labelled. The active colchicine was converted by way of allocolchiceine (1 14) and the base (1 13) into the neutral product (117).Fission of the double bond and cyclisation of the resultant dialdehyde then gave the phenanthrene (118; R = CHO); this was oxidised and the acid (118; R = C02H) was decarboxylated to give the carbon atom from position 6 of colchicine (1 10) as carbon dioxide. The results showed that the N-acetyl group of the active colchicine carried half of the original activity and that the carbon atom at position 6 was virtually inactive. The most reasonable interpretation of the above two sets of results is that the side chains of the aromatic amino-acids are degraded to the C6-Cl state before incorporation. The appearance of radioactivity in the N-acetyl group is consistent with this proposal since presumably a labelled two-carbon fragment is available; an analogy will be met later in the biosynthesis of gramine from tryptophan.L- [Me-l*C J M e t h i ~ n i n e ~ ~ ~ ~ ~ ~ and formateloo were a h incorporated into colchicine the former being used to provide the methyl groups of the alkaloid.lo1 Formate was a poor precursor.loO Both groups of workerslOOJO1 fed [l-l*C]acetate to the autumn crocus and obtained active colchicine. It was shown that alllo' or almost all100 of the activity was present in the acetyl group and Substance F (119) isolated from the same plants was inactive.99 A small amount of activity was associated in one case with ring A and its attached atoms,loO but ring c was isolated as trimellitic acid (1 12) and found to be inactive.lo0 These results lead to the conclusion that the tropolone ring of cokhicine (110) is not formed from acetate in contrast to the mould tropolones.102 All the present theorieslo3J04 for the biosynthesis of colchicine now require some modification.It seemslo0 that the non-tropolone part of the molecule is built up from the fragments shown in (120; R = H or OH) where R' may be H to give a glycine or an alanine equivalent or perhaps a carbon chain or ring to provide part or all of the tropolone system. The origin of the tropolone part of the molecule is unknown and more feeding experiments are in progress to solve this fascinating problem. 101 Personal communication from Professor E. Leete. 108 R. Bentley Biochim. Biophys. Actu. 1958 29 666; J. H. Richards and L. D. 103 F.'A. L. Anet and R. Robinson Nature 1950 166 924; B. Belleau Ejcperientiu lo4 A. I. Scott Nature 1960,186,556. Ferretti Biochem.Biophys. Res. Corn. 1960,2 107. 1953,9 178; E. Wenkert W d 1959,15 165. 282 QUARTERLY REVIEWS On the theoretical side there has been the attractive proposallo4 that the bond between rings A and c of colchicine (1 10) is formed by a phenolic coupling process as illustrated in formula (121). Such a step would be analogous to the ring closures which give the morphine skeleton (p. 275) and the lycorine and galanthamine systems (p. 279). Zndole AZkaZoids.-Gramine (1 23) is the simplest indole alkaloid to be examined by tracer methods. When barley seedlings were fed with [P-l4C]- tryptophan (122) gramine (123) was formed with labelling only at the starred position.lo5 Moreover tryptophan doubly labelled at the 2- and the ,&position i.e. [P,2-14C]tryptophan gave radioactive gramine labelled at the starred position and position 2 and (122) the ratio of the activities at the (123) eco2H (125) two positions was the same as in the original tryptophan.105 This proves that the arrowed bond in tryptophan (122) remains intact during the formation of the alkaloid.Here then is another example of cleavage of the side-chain of an aromatic amino-acid between the a- and the ,8- position by a process which apparently operates on a three-carbon side- chain. Thus 3-indolyl [j3-14C]pyruvic acid (1 24) and 3-indolyl [/3-14C]acrylic acid (125) also gave rise in sprouting barley to gramine (123) labelled only at the side-chain methylene group.lo6 However when 3-indolylacetic acid and its amide 3-indolylglyoxylic acid and 3-formylindole all radio- active were fed separately all gave inactive gramine ;loS these four pre- cursors were labelled in the side-chain at the carbon atom adjacent to the indole ring.The pathway used for the removal of the two carbon atoms from tryptophan does not therefore pass through these simple indoles and further tracer experiments here will be followed with interest. Seven biogenetic schemes have been published which might in theory account for the structure of lysergic acid (131) the indolic component of a number of the ergot alkaloids. Some theories which made use of 5-hydroxytryptophan (cf. 130) have been eliminated by the isolation of inactive ergot alkaloids from Claviceps purpurea cultures fed with 5- hydroxy [P-14C]trytophan.107 As will be seen below other precursors were satisfactorily incorporated under the same conditions.However on the positive side several independent researches are now giving strong lo5 K. Bowden and L. Marion Canad. J. Chem. 1951,29 1037 1045; E. Leete and lo* A. Breccia and L. Marion Canad. J . Chem. 1959,37 1066. lo7 R. M. Baxter S. I. Kandel and A. Okany Chem. and Znd. 1960 266. L. Marion ibid. 1953 31 1195. BATTERSBY ALKALOID BIOSYNTHESIS 283 support to one biogenetic scheme.lo8 This like most of the others calls upon tryptophan (130) to provide the main part of the molecule but it differs from the rest in suggesting the use of an isopentenyl residue to build up the piperidine ring as indicated in (130); the structure of chano- clavinefos (127) accords well with this proposal. Experimental support has come from feeding C. purpurea growing both parasitically and saprophytically with [P-l4C]tryptophan (1 30).The former techniquelo* gave a significant incorporation into the ergot alkaloids and the latter a good indeed in the German work with saprophytic cultures incorporations as high as 47 % were achieved.l12 All the many ergot alkaloids isolated in one experimentlll had approxi- mately the same specific activity including those based upon relatives of lysergic acid such as agroclavine (128) and elymoclavine (129) a result which indicates a common biosynthetic pathway to the various bases [carbo~y-~~CITryptophan fed under the conditions which gave the high incorporation already mentioned did not yield radioactive ergot alkaloids so that the carboxyl group must be eliminated during the biosynthesis.l12 So far no degradation of the active lysergic acid (1 3 l) agroclavine (1 28) or elymoclavine (129) from these tracer experiments has been reported and it is to be hoped that the expected location of the label entirely at the starred position will be rigorously established.This would satisfy critics who may argue that the tryptophan side-chain could have been degraded to small fragments before incorporation. However it has been shown that the indolic part of tryptophan is used in the biosynthesis since tryptophan labelled with deuterium in the heterocycle was used by Pennisetum typhoideum to form deuterated agroclavine (128) and elymoclavine113 (129). High incorporations into agroclavine (1 28) and elymoclavine (1 29) have been achieved from mevalonolactone (126) labelled at position 2 with carbon-14 or tritium and also from [4-3H]mevalonolactone ;l12 saprophytic cultures were used.Other workers114 observed lower incorpora- tions but here the active alkaloids were degraded. Both agroclavine (128) and elymoclavine (1 29) were examined and the results obtained114 were very similar for the two alkaloids. The bases derived from [2-14C]mevalono- lactone carried approximately half their activity at position 17 and no lO8 K. Mothes F. Weygand D. Groger and H. Grisebach Z. Naturforsch. 1958 13b 41 A. J. Birch and H. Smith “CIBA Foundation Symposium on Amino Acids and Peptides with Antimetabolic Activity” ed. Wolstenholme and O’Connor Churchill London 1958 p. 247. loo A. Hofnann R. Brunner H. Kobk and A. Brack Hefv. Chim. Acta 1957 40 1358. 110 D. Groger H. J. Wendt K. Mothes and F. Weygand 2.Naturforsch. 1959,14b 355. ll1 W. A. Taber and L. C. Vining Chem. undZnd. 1959 1218. 112 D. Groger K. Mothes H. Simon H.-G. Floss and F. Weygand 2. Naturforsch. 1960. 15b 141 E. H. Taylor and E. Ramstad Nature 1960,188,494. 113 H. Plieninger R. Fischer W. Lwowski A. Brack H. Kobel and A. Hofmann Angew. Chm. 1959 71 383. 114 A. J. Birch B. J. McLoughIin and H. Smith Tetrahedron Letters 1960 No. 7 1. 284 QUARTERLY REVIEWS activity at position 8 those from [l-14C]acetate had about half their activity at position 8 and none at position 17 and those from [2-14C]- acetate were appreciably labelled (ca. 10-25 of the total) at position 17 and only weakly labelled at position 8. If these precursors are built into the ergot alkaloids as they are into terpenes and steroids one would CH,OH LMC H (128) ; R = Me (129) ; R = CH,.OH expect incorporation from [2-14C]meva~on~~actone into position 7 or 17 or both.Similarly [l-l%]acetate ought to pass its label into positions 8 and 10 and [2J4C]acetate into positions 7 9 and 17. It will be seen that the results are consistent with these expectations and the combined evidence from all the researches just reviewed suggests strongly that the ergot alkaloids are formed from an isopentenyl residue and tryptophan. There is as yet little experimental information about the biosynthesis of the vast complex of inter-related indole alkaloids which includes such bases as yohimbine (135) strychnine (133) and the calabash curare alka10ids.l~~ Current interest is centred on the origin of ring E of yohimbine (135) and the analogous carbon atoms in the many other alkaloids.Thus one t h e ~ r y ~ ~ s ~ ~ ~ - ~ ~ ~ makes use of tryptophan a dihydroxyphenylalanine residue and a formaldehyde equivalent to build up yohimbine (135) by way of (1 32) as illustrated. A valuable and ingenious alternative scheme has been elaborated by Wenkert and Bringi,ll* who propose that hydrated prephenic acid (134) may be one precursor as shown. Cleavage of ring E first suggestedllQ for the biosynthesis of strychnine (133) and known as “Woodward fission” is then invoked in both theories to accommodate 116 K. Bernauer Fortschr. Chem. org. Nuturstofe 1959 17 183; A. R. Battersby and H. F. Hodson Quart. Rev. 1960 77. 116 G. Barger and C. Scholz Helv. Chim. Acta 1933 16 1343. 11’ R. B. Woodward Angew. Chem. 1956,68 13. 118 E. Wenkert and N.V. Bringi J . Amer. Chem. SOC. 1959 81 1474. 110 R. B. Woodward Nature 1948 162 155. BATTERSBY ALKALOID BIOSYNTHESXS 285 such alkaloids as serpentine (136). Further ring closures after the fission step are necessary to account117 for ajmaline (1 38). The difference between the two theories is that in one ring E of yohimbine (135) for example is thought to be formed from an aromatic precursor and in the other the intermediates never become aromatic. The important point must be made that Wenkert and Bringi's theory leads to the correct absolute stereo- (136) chemistry at the arrowed position (see 133-138) of the various indole alkaloids and this configuration is constant for all the other bases in this / m H 2 - mCHo+ iH,O Et Me ISle group with the sole exception of #-akuammicine.lz0 Though it is not possible here to give the details these theories can embrace a large number of indole alkaloids.P. N. Edwards and G. F. Smith Proc. Chem. SOC. 1960,215. 286 QUARTERLY REVIEWS Recently it has been proved121 that Rauwolfia serpentina plants will build [2J4C]tryptophan into ajmaline (138) reserpine (1 37) and serpentine (136) and the first was degraded by alkali fusion to yield ind-&methyl- harman (1 39). The further degradations shown then allowed the isolation of the starred carbon atom as formaldehyde and it carried all the activity of the original ajmaline. Thus the alkaloid (138) is labelled solely at the starred position and there can be little doubt that tryptophan is a direct precursor of ajmaline. In conclusion it can be said that the present state of research on alkaloid biosynthesis is that the breakthrough has been made. Not only has it been demonstrated that plants will incorporate amino-acids acetate mevalonolactone and some still simpler materials into alkaloids but also it is known that the large intermediates can be successfully introduced into the plant’s biosynthetic system. The direction in which future re- searches should be aimed is clear and a rapid increase in knowledge of the way alkaloids are made can be predicted with confidence. E. Leete Chern. andlnd. 1960,692.

ISSN:0009-2681    

DOI:10.1039/QR9611500259

出版商:RSC    

年代:1961

数据来源: RSC

摘要:

MOLECULAR ELECTRONIC ABSORPTION SPECTRA 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. By S. F. MASON (DEPARTMENT OF CHEMISTRY THE UNIVERSITY EXETER) Introduction . . .. .. .. .. Factors determining the absorption of light . . Selection rules . . .. .. .. .. Saturated compounds . . . . . . .. Olefins and dienes .. .. . . . . Triple-bonded and cumulative systems . . Carbonyl compounds . . . . . . .. Aromatic systems.. .. .. . . . . Non-alternant hydrocarbons . . .. . . Heteroaromatic systems . . .. . . . . Carbanions and carbonium ions .. .. Linear and cyclic conjugated systems . . . . Charge transfer spectra . . .. .. .. Steric effects . . .. .. . . .. Environmental effects . . .. . . . . Unsaturated sulphur and nitrogen compounds .. .. .. . . . . .. .. . . . . . . .. .. . . . ... . . 0 . .. .. .. . . . . .. .. .. . I .. .. .. . . * . .. 287 289 29 1 295 298 301 305 314 317 327 330 335 339 353 363 367 1. Introduction THE development of visible and ultraviolet absorption spectroscopy falls broadly into two main periods characterised by the current levels of theory and instrumentati0n.l The measurement of absorption wavelengths using photographic spectrographs was of primary interest in the early period the results being interpreted in terms of classical vibration theory. During the second period dating from the mid-l940’s the use of photoelectric instruments became more widely spread facilitating the measurement of absorption intensities which had assumed a greater diagnostic and theoretical importance whilst quantum mechanics became the more informative of the interpretative procedures.A series of reports2 from a committee of the British Association appointed to investigate the relation between absorption spectra and chemical constitution summarises the pioneer work to which the chair- man of the committee Hartley made a notable contribution. Acquiring Mason “Main Currents of Scientific Thought A History of the Sciences,” Routledge and Kegan Paul London 1953. Report Brit. Assoc. Adv. Sci. London 1899 p. 316; 1900 p. 151; 1901 p. 225; 1902 p. 99; and 1903 p. 126. 287 288 QUARTERLY REVIEWS in 1872 a quartz spectrograph constructed by Miller who had reported earlier that he could find no connection between the “diactinic power” of matter and its chemical n a t ~ r e ~ Hartley over a period of some thirty years arrived at the following conclusions:2 The absorption of visible and ultraviolet light is a constitutive property of a molecule as opposed to infrared absorption which appears to be at least in part an additive property of the groups within the molecule.Aromatic substances display selective banded absorption whilst aliphatic compounds give rise to weak and general absorption.2 Hartley ascribed the absorption of light to sub-molecular vibrations synchronous with the frequency of the light waves.2 Drude showed4 that the sub-molecular vibrations corresponding to visible and ultraviolet frequencies were due to particles with the mass and the charge of the electron whilst those corresponding to infrared frequencies derived from the motions of particles which were positively charged and had atomic masses.The electron vibration theory of light absorption was extensively developed but its application to the spectra of complex molecules described in several reviews,”‘ was not remarkably fruitful. A number of empirical regularities were successfully inter~reted,~ but apart from the linear polyenes and cyanines the theory yielded few detailed predictions stimulating experimental investigation. An important achievement of the early period not always appreciated was the empirical classification of molecular absorption band types. Aliphatic compounds were found to give two types of absorption band termed R- (Radikal) and K- (Konjugiert) bands by Burawoy.8 The low- intensity R-bands are given by groups such as carbonyl or nitroso. The conjugation of these groups with unsaturated residues produces a small shift of the R-bands to longer wavelengths but electron-releasing sub- stituents shift the bands to shorter wavelengths.A change from a non-polar to a polar solvent also gives rise to a blue shift of the R-bands and in acid solution they often disappear. The high-intensity K-bands are given by conjugated systems and on substitution by either electron-attracting or electron-releasing groups these bands undergo large shifts to longer wavelengths. The K-bands are relatively insensitive to environmental conditions but they show small red shifts as the refractive index or the dielectric constant of the solvent used is increased. The spectra of aromatic compounds were classified in the same period by the annellation method introduced by Clar,9 who distinguished three main types of aromatic absorption termed the a-,p- and @-bands.lo The a-bands Miller Phil.Trans. 1862 152 861. * Drude Ann. Physik 1904,14,677,936. ti Lewis and Calvin Chem. Reviews 1939,25,273. Maccoll Quart. Reviews 1947 1 16. Ferguson Chem. Revicws 1948 43. 385. * Burawoy Ber. 1930 63 3 155 et seq.; summary in J. 1939 1177. a Clar Ber. 1936 69 607. lo Clar “Aromatische Kohlenwasserstoffe,” Springer Verlag Berlin 1952. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 289 are weak comparable in intensity to the aliphatic R-bands but they show no marked solvent effects. The p-bands are of moderate intensity and they shift strongly to longer wavelengths with linear annellation in the polyacene series but angular annellation produces only small shifts. The p-bands are of high intensity and like the &-bands they move moderately to longer wavelengths with both linear and angular annellation.The usual wave- length order is 01 > p > /? but the a-band disappears under the p-band and then emerges between the p- and the /3-bands in the polyacene series. In the recent period both the aliphatic“ and the aromatic l2 classifica- tion have been “rediscovered” albeit with quantum-mechanical inter- pretations. Orbital theory was applied early to the elucidation of the spectra of atornsl3 and of diatomic systems14 the latter constituting a rather unique group of molecule^.^^ The application of the theory to the light- absorption properties of complex molecules is however recent and it has stimulated not only investigations enlarging the scope of classical molecular spectroscopy which tended to centre upon the chemistry of colouring matters but also studies leading to the development of new branches of the subject.Amongst the latter are the investigation of singlet-triplet spectra derived from the work of Lewis and Kasha,le the study of acceptor-donor complexes stimulated by the theory of Mulliken,17 the application of ligand field theory to the spectra of the transition-metal ions due notably to Van Vleck,18 and the renewed interest in the spectra of crystals stem- ming from the work of Davydov.lg The last two topics have been reviewed r e ~ e n t l y ~ ~ - ~ ~ and they will not be subjects of special mention here. The present Review after a preliminary account of the factors governing the absorption of light will be confined mainly to a discussion of the absorption spectra of organic compounds and non-metallic inorganic molecules.2. Factors Determining the Absorption of Light When a molecule absorbs visible or ultraviolet light of frequency Y an electron undergoes a transition from a lower to a higher energy level l1 Kasha Discuss. Faraday SOC. 1950 9 15. la Platt J. Chem. Phys. 1949 17 484. l3 Condon and Shortley “Theory of Atomic Spectra” Cambridge University Press 1935. lS “Diatomic molecules are peculiar because they only have two ends and these ends are very close together” Ruedenberg quoted by Coulson Rev. Mod. Phys. 1960,32 170. l6 Lewis and Kasha J. Amer. Chem. SOC. 1944,56 2100. l7 Mulliken J. Amer. Chem. SOC. 1950 72 600 4493. l9 Davydov J . Exp. Theor. Phys. (U.S.S.R.) 1948,18 210. 2o McClure Solid State Phys.1959 8 1. a1 Wolf Solid State Phys. 1959 9 1. a2 McClure Solid State Phys. 1959 9 400. 23 Dunn in “Modern Coordination Chemistry” ed. Lewis and Wilkins Interscience Herzberg “Spectra of Diatomic Molecules” van Nostrand New York 1950. Van Vleck J. Chem. Phys. 1940 8 787. New York 1960. 290 QUARTERLY REVIEWS E in the molecule the energy-level difference LIE being given by the well- known relationship d E=hv. In general non-bonding lone-pair electrons are the least strongly bound in a molecule and in the bonding levels n-electrons have higher energies than corresponding a-electrons whilst in the antibonding levels that order is reversed (Fig. 1). Thus in the spectra Antibonding T l (Tr* ) Non-bonding (n) Bonding TI' (To Bonding a- (4 Level Symbol FIG. 1. of simple molecules the bands due to n -+ T* transitions generally lie at longer wavelength than those arising from n -+ T* excitations whilst the (T -+ a* absorptions usually lie in the far ultraviolet region.According to the Beer-Bouguer-Lambert law,24 the integrated fraction of light absorbed by an assembly of molecules is proportional to the number of absorbing systems in the light path namely log, (I,/I) = ECZ . . . . . . . . . (1) where I and Z are respectively the intensities of the entrant and the emergent light with a path length I of absorbing species at a concentra- tion of c mole/l. The constant E is then the decadic molar extinction coefficient in 1. mole-lcm.-l. If a molecule with a cross-section area a absorbs a quantum of light incident upon it with a probability p the extinction coefficient may be shown25 to have the approximate value Emax = lo4 pa with a in A2.Thus absorption bands in the spectra of polyatomic molecules should have25 extinction coefficients greater than The maximum extinction coefficient of an absorption band may be affected markedly by a change of phase or of solvent but the band area A = JEdv frequently remains constant affording a better measure of absorption intensity. The band area is related to the oscillator strength 104. *' Beer Ann. Phys. Chem. 1852 163 (Dritte Reihe 86) 78. 25 Braude J. 1950 379. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 29 1 f which in the classical vibration theory is the number of oscillators each of mass m and charge e giving rise to the absorption band f = (1 O3mc210g 1 O/ne2N) Jedv . . . . . .. (2) In the case of smooth and symmetrical absorption bands the oscillator strength may be obtained in good approximation from the experimental data by . . . . . . f = 2.2 x 1 0 - 9 ~ ~ ~ ~ dY (3) where AY is the band width at half maximum extinction (in cm.-l). In the quantum theory26 the oscillator strength of an absorption band depends upon the absorption frequency Y and the electric transition moment length Q . . . . . . f = 1.085 x 10-5vQ2 (4) with v in cm.-l and Q in A. The transition moment length is a measure of the mean displacement of the promoted electron during the transition and it is given by Q = J+n (Ziri)+mdr . . . . . . . where +n and +m are the respective wave functions of the ground and the excited state and ri is a vector defining the position of the i-th electron.For the transition of an electron from an occupied to an unoccupied orbital with wave functions t,bo and #u respectively equation (5) becomes . . . . . . Q = d2Jt,b0(r)t,budr (6) the factor 4 2 allowing for the promotion of either of the two electrons in the occupied orbital. Electronic displacements of 1-3 A give oscillator strengths of the order of unity corresponding to maximum extinction coefficients measured in solution of -lo5. 3. Selection Rules The selection rules are the conditions for which the oscillator strength and thus the transition moment are non-zero. The integrals of equations (5) and (6) vanish if there is a change of electron spin on excitation and singlet-multiplet transitions occur only to the degree that the nominal singlet and multiplet states are mixtures of singlet and multiplet configura- tions so that both states have components with the same spin.The electric dipole transition moment is a real physical property of a molecule and as such it must be invariant with respect to any rotation reflection or inversion of the molecular symmetry axes. The vectors r have the symmetry properties of a translatory motion along one or more of the molecular axes and so the integrals of equations (5) and (6) are aa Mulliken J. Chem. Phys. 1939 7 14 20. 292 QUARTERLY REVIEWS non-zero only if the direct product of the symmetries of the lower and the upper state wave functions +,& or of the orbital functions t,hot,hu contains a term with the symmetry properties of a translatory motion.27 The direction of that motion gives the orientation of the electric dipole transition moment.For the magnetic dipole or electric multipole inter- action of radiation and matter different criteria obtain; in the former case a rotational term is essential in the direct product +n+,, or $ot,hu. However such interactions are weak allowed magnetic dipole and electric quadru- pole transitions having oscillator strengths of the order of and respectively so that they are usually termed “forbidden” transitions. Most absorption bands even the very weak are due to electric dipole transitions and the approximate rule can be formulated that absorption of high intensity arises generally from an electronic promotion to an upper orbital with one more nodal plane or surface than the lower orbital from which the excitation originates.28 In electronic transitions from a nodeless orbital the transition moment is directed perpendicular to the new nodal plane and in excitations from multinodal orbitals the resultant moment is the vector sum of the component transition moments at the points of nodal change.Hydrogen for example has a band system at 1109 A with P and R rotational branches. The absence of a Q rotational branch indicates that the transition moment lies along the molecular axis and that the upper orbital is cut by a nodal plane perpendicular to the bond. The transition is14 thus qs + qs* (Fig. 2). The hydrogen system at 1002 A in contrast has P Q and R branches the transition being,14 ulS + 7 ~ ~ ~ . The nodal plane of each component of the doubly degenerate upper orbital now contains the bond axis the transition moment being perpendicular to the bond (Fig.2). “p‘“ FIG. 2. The molecular orbital changes involved in the 1002 and 1109 8 band systems of hydrogen. The broken lines represent nodal planes and the broken arrows the orientation of the transition moment. 27 Sponer and Teller Rev. Mod. Phys. 1941 13 75. 28 Bowen Quart. Reviews 1950 4 236. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 293 Weak absorption results from electronic transitions to upper orbitals with more than one node additional to the number of nodes in the lower orbital since the component transition moments tend to have opposed directions and partly cancel. The cancellation is particularly complete in molecules with a centre of symmetry if the nodes change by an even number on excitation (parity forbidden). In the transition of an electron from the highest occupied to the highest unoccupied w-orbital of butadiene (Fig.3a) the component transition moments which have the instantaneous Fig. 3a CH,=C\H CH=CH c- Fig. 3b Fig. 3c RG. 3a. The nodal changes involved in the transition of an electron from the highest bonding to the highest antibonding n-electron molecular orbital of butadiene. The broken lines represent nodal planes. The instantaneous orientations of the component transition moments cfull arrows) in trans-butadiene. RG. 3b. FIG. 3c. The same for cis-butadiene. directions given (Fig. 3b and 3c) cancel completely in the trans-isomer which has a centre of symmetry but give a net resultant in the cis-is~mer.~~ The selection rules for a polyatomic molecule have an analogy to those for the atoms of which it is composed.30 Approximately an allowed atomic transition in the lighter elements involves no change in the elec- tronic spin angular momentum and a change of one atomic unit in the electronic orbital angular momentum corresponding to a unit difference between the lower and upper atomic orbitals in the number of angular nodes.Transitions requiring a change of orbital angular momentum by an even number of units are strongly prohibited (Laporte rule). Linear or angular orbital momentum is a constant of the electronic motions only in linear molecules or in abstract models of non-linear polyatomic molecules such as the free-electron model of the linear or cyclic polyenes but the nodal properties associated with orbital momentum in atoms persist in polyatomic molecules and characterise the molecular orbitals.Spin and orbital angular momentum are not independent quantities in the heavier 29 Mulliken J. Chem. Phys. 1939 7 121. Hatt J. Opt. SOC. Amer. 1953 43 252. 294 QUARTERLY REVIEWS atoms and compounds containing such atoms give absorptions due to nominal singlet-multiplet transitions of relatively high intensity. Prohibited transitions occur in polyatomic molecules with a finite probability owing to perturbations which mix in a component of an allowed transition with the formally forbidden transition. The perturbation may be intermolecular or intramolecular. Singlet-triplet bands have enhanced absorption intensities when measured in solvents containing a heavy such as ethyl iodide or more particularly in media containing dissolved paramagnetic substance^,^^ such as nitric oxide or oxygen.The singlet-triplet absorption of benzene for example cannot be de- t e ~ t e d ~ ~ with a 22-metre path length of the liquid free from oxygen show- ing that the unperturbed oscillator strength of the transition is less than 1O-l1 but the absorption is easily measured with a 10-cm. path length of the liquid saturated with oxygen at the partial pressure of the atmosphere giving an apparent oscillator strength of 4 x 10-6. The approximate selection rules are based on the assumption that a molecule has a fixed shape which remains unchanged on excitation. However there are always some vibrations of a polyatomic system which periodically change the shape of the molecule such as the bending vibra- tions of linear molecules. The direct product of the symmetries of the vibrational and electronic ground and excited state wave functions may then contain a term with the symmetry properties of a translatory motion whilst the product of the electronic functions alone does not.The transi- tion then becomes allowed to the small degree that the perturbing vibra- tion changes the electronic wave function. Moreover the equilibrium shapes of a molecule in the ground and in the excited state may differ so that a nominally forbidden transition becomes allowed on account of the change in symmetry. The change of electronic wave function with a gross change in molecular shape is large but the absence of gross nuclear motion during an electronic transition by the Franck-Condon principle ensures that the resultant light absorption is usually small.Changes of molecular shape and size on excitation and the character of perturbing vibrations can be determined from the vibrational and rotational structure of an electronic band. In an allowed electronic band the totally symmetric vibrations of the molecule those which do not change the molecular shape appear as progressions in the excited state frequency starting from the zero vibrational level of the electronic ground state with weaker temperature-sensitive bands on the long-wavelength side due to transitions from the thermally populated higher vibrational levels of the ground state. The relative intensities of the vibrational bands in a progression give by the Franck-Condon principle a measure of the change in molecular size on excitation and they may indicate a change of shape.81 Kasha J. Chem. Phvs. 1952 20 71. 82 Evans Nature 1956 178 534. 33 Craig Hollas and King J . Chem. Phys. 1958 29 974. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 295 In the case of forbidden transitions made allowed by a bending or other non-totally symmetric vibration the absorption is marked by the absence of a band origin arising from the electronic energy change without accompanying vibrational energy changes. Progressions in the totally symmetric vibrations build up on bands due to the absorption of the electronic transition energy plus or minus one quantum or an odd number of quanta of the energy of the perturbing vibration. The resultant absorp- tion pattern lacking a band origin is characteristic of bands due to forbidden transition^.^' The rotational structure of an electronic absorption band is determined by the direction of the transition moment and by the relative values of the molecular moments of inertia.The direction of the transition moment is given by the overall appearance of the rotational structure namely the presence or absence of a Q branch in the case of linear molecules or the predominance of J- or K-structure in symmetric and slightly asym- metric tops. The degradation of the rotational structure to the red region indicates a probable increase in one or more of the moments of inertia on excitation and conversely the less frequently observed degradation of the structure to the blue region suggests a decrease in molecular dimensions. The values of one or more of the molecular moments of inertia in the ground and the excited state can be obtained from the frequency spacings in the rotational structure by the methods used in the analysis of rotational- vibrational bands.34 The direction of the transition moment can be determined also by measuring the spectrum of the orientated molecule with polarized light as a molecule absorbs only that part of the incident radiation which has an electric vector parallel to the transition moment.4. Saturated Compounds The electronic absorption bands of saturated compounds with no lone- pair electrons lie in the far ultraviolet region (Table 1). The bands are due to the promotion of an electron from a bonding (T orbital to a correspond- ing antibonding a* orbital (Fig. l) or to a level composed of atomic orbitals with a main quantum number greater than that of the valency shell (Rydberg bands).In polyatomic molecules with several a bonds of a similar kind such as the paraffins the valency electrons are only approximately localised in individual bonds and they move in a group orbitals of different energies over the molecule as a whole. The energy of the highest occupied group orbital in a paraffin increases with the number of methylene groups in the chain and more particularly with the degree of chain branching as is shown by the values of the ionisation potentials of the saturated hydrocarbon^.^^ There is a parallel decrease in the energy s4 Herzberg “Infrared and Raman Spectra of Polyatomic Molecules,” van Nostrand New York 1945. a5 Hall Trans. Furadzy Soc. 1953 49. 113; 1954 50 319 773. 296 QUARTERLY REVIEWS of the lowest unoccupied group orbital so that the paraffins absorb at wavelengths which increase with the length and the degree of branching in the hydrocarbon chain.Thus saturated aliphatic molecules containing quaternary carbon atoms absorb in cyclohexane solution at 1760 with an extinction coefficient of -lo3 per quaternary atom allowing the number of such atoms in a molecule to be e~timated.~~ In saturated compounds with lone-pair electrons the longest-wave- length absorption arises from the promotion of a nonbonding electron to an antibonding O* group orbital namely a n -+ O* transition (Fig. 1). Trimethylamine in the vapour and in a number of solvents absorbs at 2273 A but in acid solution the band disappears owing to the lowering of the energy of the lone-pair electrons upon the formation of a N - H bond.37 In general bands derived from n + u* transitions have lower extinction coefficients than those due to o -+ u* transitions (Table l) particularly in the case of the compounds containing Group VI and VII elements in which the lone-pair orbital of highest energy is a p-orbital.The antibonding O* level is composed largely of a p-orbital in these com- pounds so that the n -+ u* transition has the character of a forbidden p2,y -+ pa atomic transition. A molecule containing two atoms with overlapping lone-pair orbitals as hydrazine the halogens and the methylene dihalides absorbs at longer wavelengths than the corresponding monoatomic compound (Table 1). The overlap of lone-pair atomic orbitals on different atoms gives rise to a low-energy bonding and a high-energy antibonding lone-pair molecular orbital both of which are filled with electrons.Thus the p z y -+ ups* transition of the methyl halides becomes a nz,y* -+ aPz* transition in the halogens (Fig. 1). Taking place between two adjacent antibonding levels the latter transition is of lower energy accounting for the absorption of the halogens in the visible region.38 Trimethylene disulphide absorbs at longer wavelengths than the open-chain disulphides as the latter compounds assume a staggered conformation due to repulsion between the non- bonding electrons and the overlap between the lone-pair orbitals on the adjacent sulphur atoms is small.3g~*o The intensity of the visible absorption band increases with the mole- cular weight in the halogen series (Table 11 the selection rules formulated for the light atoms applying the less rigorously the heaviei the element.Electronic spin and orbital momentum are no longer quantised separately in the heavier halogens owing to magnetic interactions between the electrons and the atomic nucleus. Spin-orbital coupling which is pro- portional to the square of the nuclear charge is strong when the electron is travelling with relativistic velocities close to the atomic nucleus.41 36 Turner Chem. and Ind. 1958 626. 37 Tannenbaum Coffin and Harrison J. Chem. Phys. 1953,21,311. 38 Mulliken Phys. Review 1940 57 500. 88 Barltrop Hayes and Calvin J. Amer. Chem. SOC. 1954,76,4348. (* Bergson Arkiv Kemi Min. Geol. 1958 12 233. 41 McClure J. Chem. Phys. 1949 17 665. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA TABLE 1. Saturated compounds Compound B2H6 B10H14 CH4 C2H6 NH3 Me,CCHzCHMe2 MeNH Me,NH Me3N “HMe) H2O MeOH Me20 Me,S Prns.sPrn s- [CH,],.S EtS4Et S8 CH3C1 CH,Br CH,I CH2I2 CHI F2 c12 Br2 12 IC1 IBr C1,O ClOEt ClOZ Br20 sc1 PI3 Ad3 Solvent* V CH V V V V V V V V V V V E E E E E V V PE PE HX V V V V FA FA cr CT V CT CT ET PE hlax (hi 1820 1350 2720 1219 1350 1540 1942 1515 2150 1737 2200 1905 2273 1990 2450 1667 1835 1838 2290 2100 2520 3340 2900 2750 1725 2040 2575 2915 3490 2845 3300 4200 5200 4500 4870 2600 3100 3600 3200 3040 3600 3780 E r n a x t 20 10,Ooo 3200 strong strong 10,Ooo 5600 strong 600 2200 100 3300 900 3950 lo00 1480 150 2520 140 1020 475 160 2400 8000 weak 200 365 1320 230 6 66 200 950 121 308 600 30 1300 250 1150 8800 1600 297 Ref.42 43 44 45 46 36 37 37 37 37 49 50 50 51 52 39 39 52 53 54 55 56 56 57 58 58 58 58 59 60 61 61 62 61 63 64 65 *The following abbreviations are used in the present and succeeding tables for the solvents listed V vapour S unspecified solvent E ethanol CH cyclohexane PE light petroleum ET diethyl ether CT carbon tetrachloride HX hexane FA trifluoroacetic acid 1 0 iso-octane IP isopentane HP heptane W Water.D dioxan CL chloroform B benzene T toluene ME methanol SA concentrated sulphuric acid HF hydrofluoric acid TF tetrahydrofuran. ?Values in italics refer to shoulders or inflexions. 298 QUARTERLY REVIEWS 5. Olefins and Dienes region due to the promotion of an electron from the bonding to the antibonding v-orbital without a change of ~pin.~~-'l At room temperature the band has on the long-wavelength side a pronounced (Fig. 4) which Olefins possess a strong absorption band in the 1600-2000 30,000 40,000 50.000 v (cmrl) F'IG.4. Schematic olefin absorption at 77" K (based on the spectra recorded in refs. 69,70,71,72 and 73). at room temperature - - - - Price J. Chem. Phys. 1948,16,894. 43 Blum and Herzberg J Phys. Chem. 1937,41 91. 44 Pimentel and Pitzer J. Chem. Phys. 1949 17 882. 45 Moe and Duncan J. Amer. Chem. SOC. 1952,74,3140. 46 Price Phys. Review 1935 47 444. 47 Sun and Weissler J. Chem. Phys. 1955 23 1160. 48 Duncan Phys. Review 1935 47 822 1936 50 700. 4g Kay and Taylor J. Chem. Phys. 1942 10,497. so Harrison Cederholm and Terwilliger J . Chem. Phys. 1959,30,355. Harrison and Price J. Chem. Phys. 1959 30 357. s2 Fehnel and Carmack J. Amer. Chem. SOC. 1949,71,84. 63 Baer and Carmack J. Amer. Chem.SOC. 1949 71 1215. 64 Zobel and Duncan J. Amer. Chem. SOC. 1955,77 2611. 66 Porret and Goodeve Proc. Roy. SOC. 1938 A 165 31. 66 Haszeldine J. 1953 1764. 67 Hunter Qureishy and Samuel J. 1936 1574. Steunenberg and Vogel J. Amer. Chem. SOC. 1956,78 901. Buckles and Mills J. Amer. Chem. SOC. 1954 76 3717. 6o Buckles and Mills J. Amer. Chem. SOC. 1954,76 6021. 61 Anbar and Dostrovsky J. 1954 1105. *2 Buser and Hanisch Helv. Chim. Acta 1952 35 2547. 63 Bacon Irwin Pollock and Pullin J. 1958 764. 64 Bennett Emelkus and Haszeldine J. 1953 1566. 66 Walsh Trans. Faraday SOC. 1945 41 35. Emelkus Haszeldine and Walaschewski J. 1953 1552. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 299 is much reduced in intensity at liquid nitrogen temperatures (-77°K). With long path lengths (2.5 m.) of liquid ethylene,72 or with gaseous ethylene in the presence of 100 atmospheres of oxygen,73 further weak absorption is found in the 2600-3500 A region due to a transition between the same orbitals with a change of electron spin.For some years it has been predicted7* that the planes of the methylene groups at equilibrium in the first excited state of ethylene are mutually perpendicular and more recently it has been that those groups are bent and staggered. An examination of the 1620 A band of ethylene and tetradeuteroethylene in the v a p ~ u r ~ ~ and the measurement of the absorption of alkenes in a hydrocarbon glass at low temperat~re,~~ support a non-planar excited state for the olefins. In the vapour absorption a progression is observed in the upper-state ethylene carbon-carbon stretching vibration and superimposed upon it are diffuse bands due to the methylene twisting vibration.The maximum intensity occurs at a frequency corresponding to the absorption of the electronic energy and some twenty quanta of the stretching vibration indicating a lengthening of the carbon-carbon bond from 1.353 to 1-69 A and the prominence of the twisting mode in the vibrational structure suggests a change to a non- planar excited In the electronic ground state of ethylene not only the zero but also the first methylene twisting vibration level with torsional amplitudes of 1 1 O and 19" respectively are occupied at room temperature and an electronic transition from a vibrationally excited molecule requires an energy less by some 3000-4000 cm.-l than the corresponding transition in a molecule with zero-point twisting vibrational energy (Fig.5). Transitions from the vibrationally excited ethylene molecules give the absorption shoulder observed on the long-wavelength side of the main singlet band (Fig. 4) for at low temperatures all of the molecules are in the zero twisting vibration level in the electronic ground state and the shoulder disappear^.^^ In the higher alkenes the frequency of the twisting vibration is reduced and a correspondingly larger proportion of the molecules occupy the first and higher twisting vibration levels in the electronic ground state at room temperature so that the long-wavelength shoulder is more pronounced than in the case of ethylene.69 Alkyl-substituted ethylenes absorb at longer wavelengths than the parent c o m p ~ u n d ~ ~ - ~ l (Table 2) owing largely to the hyperconjugative effect of the substituents.Correlations for structural analysis have been 67 Semenow Harrison and Carr J. Chem. Phys. 1954,22 635. Gary and Pickett J . Chem. Phys. 1954 22 599. 69 Potts J. Chem. Phys. 1955 23 65. 70 Wilkinson and Mulliken J. Chem. Phys. 1955,23 1895. 71 Jones and Taylor Analyt. Chem. 1955 27 228. 72 Reid J. Chem. Phys. 1950 18 1299. 73 Evans J . 1960 1735. 74 MuIliken J. Chem. Phys. 1935 3 517. 76 Walsh J. 1953 2325. 300 QUARTEXLY REVIEWS based on these shifts employing either the wavelength and extinction of the band maximum for use with vacuum instrument^,^^^^^ or the value of the extinction coefficient at a given position on the long-wavelength absorption edge for use with silica prism ~pectr~photometers.~~-~~ The latter correlations are less reliable than the former not only because of stray light and the steep gradient of band edges but also because of the side band absorption of quaternary carbon atoms,36 and of the shoulder which is dependent on the temperature and torsional vibrational frequency in the olefin spectrum.TABLE 2. OleJns Compound Ethylene 1 -alkenes cis-Zalkenes trans-Zalkenes 2-alkyl- 1 -alkenes 2-substituents 3-substituents 4substituents Solvent V V V V V HX HX HX Amax (A) 1 620 Acyclic olefins 1750 f 25 1765 f 15 1790 f 10 1875 f 15 1850 f 30 1905 f 25 1980 f 20 Polycyclic olefins Emax -10,Ooo 11,800 f 1200 12,300 f 200 11,700 f 800 8900 f 1200 Ref. 70 71 71 71 71 76 76 76 The rules,80,81 relating the absorption and the structure of conjugated dienes are now well established.Relative to the high-intensity absorption band of butadiene in hexane solutiona2 at 2170 A a red shift of 4 0 8 is conferred by the addition of an alkyl substituent or by the exocyclic disposition of a double bond (Table 3). In polycyclic six-membered ring systems the trans-diene chromophore (double bonds in different rings) absorbs some 200 A to shorter wavelengths but with greater intensity than the cis-diene chromophore (double bonds in the same ring).80The cis-and trans-dienes have the same orbital energies but differential changes in the electron repulsion energy confer the larger transition energy upon the trans-isomer which has also the longer transition moment length.26 In the series of monocyclic dienes the absorption of cyclohexadiene is maximal in wavelength and intensity.83 The formally single bond in the diene conjugated system is subject to steric torsion in the larger rings and hyperconjugation through the methylene group of cyclopentadiene gives 76 Turner J.1959 30. 77 Bladon Henbest and Wood J. 1952 2737. 78 Stich Rotzler and Reichstein Helv. Chim. Ada 1959,42 1480. 7s Ellington and Meakins J. 1960 697. 8o Booker Evans and Gillam J. 1940 1453. Woodward J Amer. Chem. SOC. 1941,63,1123; 1942,64,72,76. 8a Smakula Angew. Chem. 1934 47 657. 83 Fawcett and Harris J. 1954 2673. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA TABLE 3. Conjugated dienes Compound Solvent Amax (A) Emax Butadiene V ;A;:) -20,000 (b) ring exocyclic S 2170 + j 35,000 2035 C yclopen t adiene HX 2385 3400 Cyclohexadiene HX 2565 SO00 Cycloheptadiene I 0 2480 7500 Acyclic dienes with subs tituent s (a) acyclic S 2170 + 1 50 per group 5000- to double bond 150 per ring Polycyclic dienes with conjugation (a) homoannular S 2710 f 110 -9000 (6) heteroannular S 2420 f 70 -1 6,000 30 1 Ref.71 84 85 86 81 80 80 the chromophore a quasi-cyclic character. The energy levels of a cyclic polyene are more widely spaced than those of the corresponding linear molecule and inter-level transition energies are consequently larger in the former. 6. Triple-bonded and Cumulative Systems In molecules with triple and cumulative bonds there are two sets of n-orbitals which with the z-direction lying along the molecular axis may be designated the n and the nY orbitals. A d.~* excited configuration arises from two pairs of electronic transitions namely vZ -+ nE* ny + ny* and nTT -+ nu* ny -+ T,*.In acetylenes7 the in-phase and out-of-phase combination of the former pair give rise to a high- and a low-energy excited state with symmetries of Cu+ and Cu- respectively. The latter pair remain degenerate giving an excited state of intermediate energy with the symmetry d,.Transitions from the ground (Cg+) to the Cu+ state are allowed but transitions to the other states are prohibited unless the mole- cule becomes non-linear on excitation.8s With a change of symmetry notation a similar situation obtains in other triple bonded and in cumula- tive molecules but the energies of the lower excited states may be inter- ~hanged.~ 84 Scheibe Ber. 1926 59 1333. 85 Henri and Pickett J. Chem. Phys. 1939 7 439. 86 Pesch and Friess J. Amer.Chem. SOC. 1950 72 5756. Ross Trans. Faraday SOC. 1952 48 973. Ingold and King J. 1953 2702 2704 2708 2725 2745. 89 Mulliken Canad. J. Chern. 1958 36 10. 3 302 QUARTERLY REVIEWS Vibrational and rotational analyses of the weak acetylene absorption between 2100 and 2500 A ~ h o ~ ~ ~ ~ ~ that the molecule is trans-bent at equilibrium in the first singlet excited state with a carbon-carbon bond lengthened from 1.205 to 1.383 A and a CCH bond angle of 120.2". The lowest energy Cg+ -+ Cu- transition forbidden in the linear molecule becomes allowed to a trans-bent excited state but the resultant absorption intensity is weak owing to the absence of nuclear motion during the transition.** The main progression is in the upper-state frequency of the trans-carbon-hydrogen bending vibration and from the first to the fifth band in the progression the absorption intensity increases 300-fold as the linear ground-state configuration progressively approximates to the turning point configurations of the higher bending vibration levels in the excited state.8s A similar analysis has been made of the weak absorption of hydrogen cyanide.91 Two overlapping band systems have been distinguished one at 1600-2000 A due to the C+ -+ 2- transition and the other at 1550- 1700 A due to the C+ -+ LI transition. The molecule is bent at equilibrium in the upper state of both transitions with bond angles of 125" and 114" respectively. 91 Compound HC = CH A1k.C CH AlkC = CMe Me.[C E C],.Me Me.[C = C],.Me MeCN NCC f CCN TABLE 4. Acetylenes and nitriles Solvent hllax (A) Emax V 2200 weak 1820 moderate 1520 strong CH 1865 450 CH 1905 850 E 2750 -1 2500 160 V 1650 strong E 3300 -1 V 1670 weak 3060; 2070 120; 135,000 I 0 2800; 2680 8-4; 17.1 Ref.88 92 93,94 95 95 96 97 482 96 98 101 102 In the spectra of the polyacetylenes and the acetylene-dinitriles three main regions of absorption have been foundg6-lo5 (Table 4) the weak Innes J. Chem. Phys. 1954 22 863. 91 Herzberg and Innes Canad. J. Phys. 1957 35 842. 92 Rose Z. Physik 1933 8 85. 93 Price Phys. Review 1935 47 444. 94 Moe and Duncan J. Amer. Chem. SOC. 1952,74,3136. 96 Wojtkowiak and Romanet Compt. rend. 1960 250 2865 3305. O6 Beer J. Chem. Phys. 1956 25 745. 97 Armitage Cook Entwistle Jones and Whiting J. 1952 1998. 98 Armitage Cook Jones and Whiting J. 1952 2010. 99 Armitage and Whiting J. 1952 2005. loo Armitage Entwistle Jones and Whiting J.1954 147. lol Herzberg and Scheibe Z. phys. Chem. 1930,57 390. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 303 moderate and very high intensity bands at progressively shorter wave- lengths being assignedg6 to transitions to the I&- Id, and lZu+ upper states respectively. The spin-forbidden orbitally-allowed triplet upper state 3Cu+ of the lower polyacetylenes has been detected in emissiong6 and in ab~orption.'~ In solution the moderate and very high intensity bands of the polyacetylenes and their nitrile derivatives appear with marked vibrational structure consisting in each case of a progression of some five members in the upper state triple bond stretching vibration frequency (-2000 cm.-l). In the classification due to Jones and his c o - w o r k e r ~ ~ ~ - ~ ~ ~ the vibrational bands of the moderate and very high intensity systems are designated to shorter wavelengths A B C .. . and L M N . . . respectively. In solution the weak intensity system appears as a single bandg6t102 close to the moderate intensity band system with which it merges in the higher poly- acetylenes. 96 The absorption intensity of the moderate band system of acetylene is very sensitive to substitution effect^.^^^^^ On passing from the 1- to the 2-alkynes the extinction coefficient of this band is doubled (Table 4) owing probably to the admixture of a small component of the allowed (T -> (T* transition of the alkyl group with the forbidden lCg+ + ld acetylene transition. In diacetylene the replacement of hydrogen by an alkyl group brings about a smaller intensity increase but p- and more particularly a-halogen -oxygen or -nitrogen atoms give rise to large intensity enhance- m e n t ~ ~ ~ the largest being comparable to the effect of a conjugative substituent (Table 5).The shift of the Id band to longer wavelengths is considerable also the wavelength displacement being roughly parallel to the intensity increase (Table 5). TABLE 5. TJze absorption spectra of the diacetylenes RCEC*C_C.R. Values are quoted only for the most intense band (the B-band) where vibra- tional$ne structure occurs R H Me Et ClCH,CH BrCH,CH MeOCH Me,NCH ClCH BrCH ICH Ph Solvent I 0 E E E E E E E E E E b l a x (-4 2349 2360 2385 2410 2420 24-40 2440 2530 2600 2800 3060 Emax 270 330 340 490 895 440 1600 2100 4500 7400 31,000 Ref. 105 99 99 99 99 99 99 99 99 99 100 Io2 Miller and Hannan Spectrochim.Ada 1958 12 321. Io3 Bohlrnann and Mannhardt Ber. 1956 89,2268. Io4 Bohlmann Ber. 1953 86 63. Io5 Georgieff and Richard Canad. J. Chem. 1958,36 1280. 304 QUARTERLY REVIEWS The n -+ u* transitions of the hetero-atom substituents are too weak (Table 1) to account for more than a small part of the intensity borrowing and it is probable that the relatively large intensities are due to a mixture of the forbidden lCg+ -+ ld diacetylene transition with an allowed charge- transfer transition of a lone-pair electron from the substituent hetero-atom to an unoccupied diacetylene n-orbital. For any marked bathochromic and hyperchromic effect it appears to be sterically necessary that the lone- pair orbital of the substituent hetero-atom shall overlap the n-orbitals of the diacetylene group as /?-substituents are less effective than those in the a-position and the acetylene-nitriles in which the lone-pair and rr-orbitals have zero overlap give only weak-moderate Cg+ -+ ld bands.The degree of mixing between the diacetylene and the charge-transfer transitions depends upon the ionisation potential of the lone-pair electrons of the substituent hetero-atom and the magnitude of the overlap between the lone-pair and n-orbitals. Although the ionisation potential of trimethyl- aminelo6 (7.86 ev) is smaller than that of methyl iodidefo6 (9.54 ev) the iodomethylene group has a larger bathochromic and hyperchromic effect upon the ld diacetylene absorption than has the dimethylaminomethyl- ene groups owing to the larger size of the iodine atom and the more effective overlap between the iodine lone-pair and the diacetylene n-orbitals.FIG. 5. The variation of the potential energy of ethylene with the angle of twist (0) about the carbon-carbon bond.for the ground ( N ) singlet ( S ) and triplet ( T ) T.T* excited states (based on the spectra recorded in refs. 69,70 and 73). The vertical arrows represent transitions from the zero (1 1") and the first (19") methylene twisting vibrational lev el of the electronic ground state. The broken portions of the curves are uncertain experi- mentally but the~retically~~ the curves for the ground and triplet excited state can cross at large angles of twist. The spectra of molecules containing cumulative bonds show in general one or two weak bands at longer wavelengths followed by an intense band lo6 Watanabe J.Chern. Phys. 1957 26 542. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 305 at short wavelengths (Table 6). Few of these band systems have been at all completely characterised. In the cumulative compounds containing hetero- atoms n -+ T* transitions may be responsible for one or both of the weak absorption bands and the set of upper states arising from the v3.71* excited configuration may not have the same relative energies as in the case of the triply-bonded molecules. The weak-moderate carbon dioxide TABLE 6. Cumulative systems Compound Solvent EtCH= C= CH HX V V MeCH=C=CHMe V CH2= C= C= CH CH,=C=O V o=c=o V EtN3 E Et02CCH = N E EtN= C=NEt HX s=c=s CT Anax (4 2250 I880 1810 2000;1940;1830 3100; 2400 3300; 2200 1700 1475 1332; 1121 2870 2220 3775 2490 2700; 2300 3180 Emax Ref.500 118 4000 71 20,000 71 weak; 4000; 20,000 71 250; 20,300 10s 12; 80 109 strong 110 300; 600; 80,OOO 111 20; 150 113 16; 10,050 114 25; 200 115 108 116 band at 1475 A for example is thought to be due to the lCg+ + ld transition and the intense 1121 A band to a Rydberg excitation the 1332 A band being of uncertain provenan~e.~~ The weak absorption of carbon disulphide at 3200 A results from a transition to a bent upper state (equilibrium bond angle -140") deriving from the ld component of the T~.T* excited configuration of the linear The band contains both singlet and triplet components the long but not the short wavelength part exhibiting a Zeeman effect.lo7 7. Carbonyl Compounds Simple carbonyl compounds give four main regions of absorption below the Rydberg bands in the far ultraviolet region.A very weak absorption (Emax near 4000 A and a weak band (Emax -10) near 2900 A are due respectively to the triplet and the singlet transitions of an lo' Douglas unpublished quoted in ref. 89. lo8 Schubert Liddicoet and Lanka J. Amer. Chem. SOC. 1952 74 569. I o 9 b o x Norrish and Porter J. 1952 1477. 110 Price Teegan and Walsh J. 1951; 920. ll1 Watanabe Inn and Zelikoff J. Cnem. Phys. 1953 21 1648. 112 Wilkinson and Johnstone J. Chem. Phys. 1950 18 190. 113 Sheinker Doklady Akad. Nauk S.S.S.R. 1951 77 1043. 11* Wolf Z. phys. Chem. 1932 B 17 46. 115 Lardy J. Chim. phys. 1924,21 281 and 353. 116 Doran and Gillam J. SOC. Chem. Znd. 1928 47 259. 11' Kuhn and Martin Z. phys. Chem. 1933 B 21,93. 118 Burr and Miller Chem.Reviews 1941 29 419. 306 QUARTERLY REVIEWS oxygen 2p lone-pair electron to the carbonyl antibonding ~ - o r b i t a l l ~ ~ - l ~ ~ (n + n*). A n -+ o* band of moderate-strong intensity lies near to 1800 A and an intense T -+ T* absorption at shorter wavelength (Table 7). The excitation of an oxygen 2p lone-pair electron is forbidden to the carbonyl nTT,* orbital but it is partly allowed to the carbonyl uZ* orbital the permitted atomic orbital components being 2py(0) -+ 2pz 2s(c). Calculations by McMurry,121 based on the theory of Mulliken,26 indicated that the 1800 A band of the carbonyl group had the intensity required of a n -+ o* transi- tion and the shorter wavelength absorption that of a T -+ 7 ~ * transition suggesting by exclusion that the 2900 A band had a n -+ T* character.Compound H&=O MeCHO COMe MeC0,Et MeCOCl MeCOBr MeCOCH,Br MeCOCH,SEt TABLE 7. Solvent IP V HX V CH CHW) V W HP HP HX E The carbonyl chromophore )cmax (4 Emax Ref. 3100 5 119 1749; 1555 18,000; 23,000 122 2935 12 123 1816; 1600 10,000; 20,000 124 2750 22 78 1900 1100; (9800) 78,124 1660; 1500 3000; 30,000 124 2040 60 145 2400 34 145 2500 93 146 3105 83 147 2990; 2430 257; 363 52 MeCOCH,CH,.SEt E 2780; 2350; 2100 32 170 1400 52 With modifications this suggestion is supported by analyses of the rotational and vibrational structure126-12s of the weak formaldehyde system near 3000 A. Formaldehyde is planar in the ground state but becomes pyramidal at equilibrium in the excited state with an angle of -20" between the plane of the methylene group and the carbonyl axis. In the excited state there is a barrier of some 650 cm.-l to the umbrella inversion mode and the carbonyl bond length is increased from 1.22 to 1.32 A.The T and T* orbitals of planar formaldehyde become two lone- pair orbitals in the pyramidal molecule one on the carbon and the other on the oxygen atom and these orbitals contain three electrons in the excited state. In absorption the positive and the negative levels of the 119 Cohen and Reid J. Chem. Phys. 1956 24 85. 120 Ford Spectvochim. Ada 1958 12 394 McMurry J . Chem. Phys. 1941 9 231 and 241. 122 Fleming. Anderson Harrison and Pickett J. Chem. Phys. 1959 30 351. lZ3 Herold and Wolf Z . phys. Chem. 1931 B 21 165. 124 Lake and Harrison J . Chem. Phys. 1959 30 361. Sidman Chem. Reviews 1958 58 689. lZ8 Brand J . 1956 858. lZ7 Robinson Canad.J . Phys. 1956,34,699. 12* Dieke and Kistiakowsky Phys. Review 1934 45 4. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 307 inversion mode are the origins of progressions respectively weak and strong in upper-state symmetric vibration frequencies. The bands of the strong progression have transition moments in the molecular plane perpendicular to the carbonyl axis and they arise from the mixing of the forbidden n -+ T* transition with a high-energy allowed transition induced by the out-of-plane bending ~ibrati0n.l~~ The weaker bands which include the band origin have transition moments parallel to the carbonyl bond and since they cannot be due to a vibrational perturbation they have been ascribed129 to electronic interaction with rotation about the carbonyl axis or to a magnetic dipole tran~iti0n.l~~ The latter explanation is the more probable as a nppy -+ n* transition has a magnetic transition moment of about one Bohr magneton along the required z-direction,13* and the n -+ n* absorption of optically active carbonyl compounds has a large rotational strength131 for which a magnetic as well as an electric transition moment is necessary.130 The equilibrium excited state of the n -+ n* triplet absorption of formaldehyde at 3900 A is also ~yramida1.l~~ The transition moment lies along the carbonyl bond axis as required by the that spin- orbital coupling with the singlet 7~ -+ n* transition confers a non-zero probability upon this otherwise doubly forbidden transition.The conjugation of a carbonyl with a vinyl group gives four n-electron energy levels of which the highest occupied and the lowest unoccupied have higher and lower energies respectively than the corresponding orbitals of the carbonyl group.The lone-pair and 0 orbitals of the carbonyl group to a good approximation are unchanged so that the n -+ n* and more particularly the 7~ + 7 ~ * carbonyl absorptions shift to longer wave- lengths in the +unsaturated derivatives (Table 8) whilst the n -+ u* TABLE 8. Solvent I0 { ECH E E E CH CH E CH E CH Unsaturated carbonyl compounds hIlax(A) 3300 2120 3370; 2290 3085 2380 3185; 2290 3050 2430 3325 2530 2900 2020; 1900 3020; 2600 2850; 2140 4600;2710;2380 3100 3500;3090;2280 Emax 17; 4600 37; 16,500 66; 15,400 37; 15,600 90 1400 150 18,600 110; 3000; 3000 72- 423 24 1500 150; 1750; 2600 121 17; 35; 220 Ref. 148 137 137 137 149 150 143 151 153 142 152 193 lZ9 Pople and Sidman J.Chem. Phys. 1957 27 1270. 130 Kauzmann Walter and Eyring Chem. Reviews 1940 26 339. lS1 Djerassi “Optical Rotatory Dispersion” McGraw-Hill New York 1960. 133 Giorgio and Robinson J. Chem. Phys. 1959 31 1678. Sidman J. Chern. Phys. 1958,29 644. 308 QUARTERLY REVIEWS band remains at the same energy probably lying beneath the strong n -+ n* absorption. Alkyl groups in the a- or P-position of such unsaturated carbonyl compounds shift the n -+ n* absorption to longer wavelengths by -100 A per substituent or by -250 8 for cyclic /3/3-dis~bstitution.~~s~~~ However the shifts may be up to 60 A larger for 18- than for a-~ubstituents,~~~ and the cyclopentenones are absorbing at wavelengths up to 150 A shorter than predicted by the empirical r ~ l e ~ .~ ~ s ~ ~ ~ Particularly when carrying a positive charge electronegative hetero- atoms transannular to the ap-unsaturated carbonyl chromophore (e.g. 1) induce a red shift of the n + n* band and a blue shift of the 7 -+ n* band.136J37 A change from a hydroxylic to a non-polar solvent produces similar shifts in the band positions8 (Table 8). The solvent and substitution effects suggest that the excitation giving rise to the n -+ n* band of a/3- unsaturated ketones has a charge-transfer character an electron under- going a transition from a largely ethylene 7-orbital to a predominantly chromophore (2) has a hybrid character but (2a) and (2b) separately contribute with the greater weight to the ground and the excited state respectively. An electronegatively or positively charged group adjacent to the /3-carbon atom destabilises the excited state and raises the T -+ T* transition energy.Conversely a polar solvent lowers the energy of the upper state and produces a bathochromic shift of the n -+ T* band. In the excited state of the n -+ n* transition the n-electron system carries an excess of electronic charge and that state is stabilised by vicinal electronegative or positively charged groups reducing the transition energy. Hydrogen bonding in hydroxylic solvents lowers the energy of the oxygen lone-pair electrons resulting in a blue shift of the n + n* band. In contrast to the effect of unsaturated residues alkyl groups and monoatomic substituents with lone-pair electrons replacing an aldehydic hydrogen atom shift the carbonyl n + n* absorption to shorter wavelengths 134 Evans and Gillam J.1943 565 and 815. 135 Gillam and West J. 1942 483 and 486. 137 Kosower and Remy Tetrahedron 1959,5 281. Clarke and Pinder J. 1958 1967. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 309 (Table 7). Since the n-orbitals of the carbonyl group contain an extra electron in the excited state of the n -+ n* transition electron-releasing groups raise the energy of that state. A compensating energy change arises from the inductive effect of substituents with a large electronegativity but inductive interactions also lower the energy of the oxygen lone-pair electrons in the ground state. More striking are the effects of substituents both conjugated and formally non-conjugated upon the intensity of the n -+ n* absorption of carbonyl compounds.Progressive substitution by methyl in formalde- hyde doubles and then quadruples the extinction coefficient of the n +- n* band and grosser changes result from the substitution of unsaturated groups or atoms bearing lone-pair electrons at the carbonyl group or at vicinal positions along a saturated chain (Tables 7 and 8). In the case of the diketone (3) the intensity enhancement is especially large the extinction coefficients of the long-and the short-wavelength carbonyl absorption being 980 and 2890 respectively13* (Fig. 6). . . 1 I I I I 3000 4- 5ooQ Wavelength (A) FIG. 6. The electronic spectra in cyclohexane solution of -~ 9,lO-dihydro- 1 1,12-dioxo-9,1O-ethanoanthracene ( 3 ) of - - - - 9,lO-dihydro-9,1O-ethano- anthracene and of . . . . . . . . . . . . bicyclo [2,2,2]octane-2,3-dione (reproduced with permission from ref.138). 138 Cookson and Lewin Chem. and Znd. 1956 984; Lewin Thesis London 1958. 139 Cookson J . 1954 282. lr10 Cookson and Dandegaonker J. 1955 352 and 1651. Bird Cookson and Dandegaonker J. 1956 3675. 141 Cookson and Warizer J. 1956 2302. lr12 Birnbaum Cookson and Lewin J. 1961 1224. 3 10 QUARTERLY REVIEWS The spectra of carbonyl compounds containing formally non-con- jugated substituents are often marked by the presence of absorption bands which do not belong to either of the chromophores separately. The 01- and the @-ethylthio-ketones give such bands at 2430 and 2350 A respect- ively (Table 7) and the Byunsaturated ketones absorb with variable intensity in the 2000-2600 A region (Table 8). In that region occur the high-intensity bands of aryl and a/3-unsaturated ketones due to a 72 -+ T* transition with charge-transfer character (see above) and such transitions may occur with a reduced probability in Byand even &-unsaturated ketones if the v-orbitals of the carbonyl and the aryl or vinyl groups overlap appreciably.A detailed treatment143 of the ketone (4) indicates that the overlap between the carbonyl and the vinyl group is about one quarter of that between the corresponding groups of an unhindered @-unsaturated ketone. The calculated interaction between the carbonyl and vinyl groups accounts for the moderate intensity band (emax 3000) at 2020 A in the spectrum of the ketone (4) as an analogue of the high-intensity 7~ -+ v* band (emax - 15,000) near 2300 A in the spectra of +maturated ketones.143 Owing to the non-coplanar arrangement of the unsaturated groups in the ketone (4) the oxygen 2p lone-pair orbital overlaps with the @-carbon 2pn orbital the value of the overlap being 0.0021.Consequently the n -+ 7 ~ * and the 7~ 3 7 ~ * charge-transfer transitions mix the former assuming about one hundredth part of the oscillator strength of the 1atter.la3 A similar mixing accounts for the enhanced n -+ v* band intensi- ties of sterically hindered ap-unsaturated ketones (e.g. 5 ) and the 86- unsaturated ketone (6) (Table 8). The steric requirements in Byunsaturated ketones for the appearance of a charge transfer T -f v* band and for the perturbation of the n + 7 ~ * absorption differ.142 A charge-transfer v -f 7 ~ * absorption is observed in the spectrum of the ketone (7) in which there is attenuated conjugation be- tween the vinyl and carbonyl groups but these groups are coplanar and there is no overlap between the lone-pair and 72-orbitals so that the n + v* band intensity is normal (Table 8).In the case of propargyl- aldehyde the oxygen 2p lone-pair orbital overlaps the my orbital of the acetylene group but the charge-transfer transition takes place from the 7 ~ orbital of that group and there is no co-mixture with the n -+ 7 ~ * transition (Table 8). The oscillator strength .fn+n. sequestered by a n + r* transition in an unsaturated ketone depends upon144 the nominal oscillator strength of the charge transfer 7~ -f v* transition the energy separation LIE between the two unperturbed transitions and the overlap integral S between the oxygen lone-pair and the v-orbitals of the unsaturated group namely 143 Labhart and Wagniere Hdv.Chim. Ada 1959 42 2219. Herzberg and Teller 2. phys. Chem. 1933 B 21,410. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 311 where P is the mixing energy for unit overlap. Thus in the cases of the unsaturated a-diket~nesl~~ (3) and (8) the longer wavelength n + n* absorption appropriates a smaller share of the borrowed intensity than the shorter wave carbonyl band since it is further displaced in energy from the charge-transfer n -+ n* band. Both of the n -+ n* bands of the aryl diketone (3) have larger extinction coefficients than the corresponding bands of the vinyl diketone (8) for although the energy separation between the bands is similar in the two cases the charge-transfer n -37~* absorption of the diketone (3) has the greater intensity (Fig.6 and Table 8). In a series of benzo C2.2.21 bicyclo-octenones it has been shown142 that progressive reduction in the ionisation potential of the aromatic ring increases the intensity of the carbonyl n -+ n* band. The reduction in ionisation poten- tial is attended by an increase in the wavelength and the intensity of the charge-transfer n -+ n* absorption both factors favouring intensity borrowing by the n -+ n* tran~iti0n.l~~ The relatively small enhancement of the carbonyl n -+ n* absorption intensity on replacement of an aldehydic hydrogen atom by a methyl group (Table 7) probably arises from a small admixture of the n -+ n* transition with the CT + o* transitions of the alkyl groups but the larger effect of atoms with lone-pair electrons a- or p- to the carbonyl group may be ascribed to the sequestration of charge-transfer intensity.Bands due to the transition of sulphur lone-pair electron to the carbonyl n*-orbital are observed52 in the spectra of a- and p-ethylthio-ketones at 2430 and 2350 8, respectively (Table 7) but the corresponding bands of a- or 18- hydroxy- or halogeno-ketones have not been found as yet for they probably lie near to the wavelength limit of quartz instruments. In the cyclohexanone series axial a-halogen or -oxygen substituents shift the carbonyl n -+ n* band some 40-300 8 to longer wavelengths and give rise to a larger intensity increase of the band than the correspond- ing equatorial substituent which shifts the absorption to shorter wave- lengths by 40-125 A.1397140 Similarly these substituents in the y-position of a cyclic ap-unsaturated ketone produce a bathochromic shift of the n + n* absorption of 20-50 A if equatorial but 70-200 A if axia1.140 The important interaction between a carbonyl group and an equatorial a-halogen atom is conjugative giving changes in the n -+ T* band position and intensity which are smaller than.but similar in kind to those produced by the halogen atom in the acetyl halides (Table 7). Charge-transfer inter- action is more important between an axial a-halogen atom and the carbonyl group producing a larger increase in the intensity of the n -+ n* band and the bathochromic shift which arises from that form of inter- action (Tables 7 and 8). The n + n* absorption intensity is less affected by the mutual perturba- tion of two carbonyl groups than by the interaction of a vinyl and a carbonyl group.The series of a-diketones with a range of angles of twist about the inter-carbonyl bond have uniformly weaker n -+ n* bands 312 QUARTERLY REVIEWS (Table 9) than the sterically hindered @-unsaturated ketone (5). The 7~ -+ T* bands lie near 1750 in the a/3-unsaturated ketones (Table S) so that the former are the further separated in energy from the n -+ m* bands and there is a smaller mixing of the transitions. The spectra of the /3-diketones with orthogonal (9) and with coplanar (10) carbonyl groups (Table 8) show however that the mixing of the carbonyl n -+ T* and n -f T* transitions in the former is not negligible. in the a-diketones,121 compared with 2300 In the a-diketones there are two lone-pair molecular orbitals and two antibonding n*-orbitals each pair being formed by the symmetric and the antisymmetric combination of the respective orbitals of the separate carbonyl groups.Of each pair the symmetric orbital has the lower energy for all angles of twist about the inter-carbonyl bond except go" where the two levels of each type are degenerate. The maximum energy separation between the two lone-pair orbitals obtains in the cis-configuration of the a-diketone group where the oxygen 2p lone-pair orbitals overlap relatively strongly but the energy difference between the T*-orbitals is a maximum for both the cis- and the trans-configuration (Fig. 7). The lower and the - 8 0 90' 180' FIG. 7. The variation of the energy of the lone-pair (n' and n 3 and the antibonding n-orbitals (T*' and T*") in the a-diketones with the angle of twist (8) about the inter- carbonyl bond.The vertical arrows represent the electronic transitions allowed by the molecular symmetry. upper n*-orbitals have an antinode and a node respectively in the inter- carbonyl region and so the energies of these orbitals are respectively sensitive and relatively insensitive to twist about the inter-carbonyl bond. Of the four possible n -+ n* transitions in or-diketones only those between MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 313 the two symmetric (n' + z-*') or the two antisymmetric orbitals (n" -+T*") are allowed by molecular symmetry (Fig. 7) though they are all atomically forbidden as in monocarbonyl compounds involving the promotion py -+ p z at the oxygen atom.154y157 Two bands with intensities comparable to that of the n + n* absorption of monocarbonyl compounds are ob- served in the spectra of a-diketone~.l~*-l~~ The band at longer wave- lengths due to the n1 -+ n*' transition varies with the angle of twist about the inter-carbonyl bond over a range of 1290 A but the band at shorter wavelengths due to the nrr -+ T*" transition lies close to the position of the monocarbonyl absorption the range of displacement being only 200 A (Table 9).However camphorquinone with a nearly cis-configura- TABLE 9. ethanol solution) and the angle of twist about the COCO bond (8) Camp horquinone 0-10" 4660 31 156 2800 23 154 The relation between the absorption spectra of a-diketones (in Compound e Amax (A) Emax Ref. MQ \ / Me n=2 0-60" 3800; 2975 11; 29 156 /'\$-o n=3 90-1100 3370; 2990 34; 35 156 n=4 100-140" 3430; 2955 21; 43 156 Ma Me n=14 100-180" 3840; 2865 22; 59 156 Me,CCOCOCMe 90-1 80" 3650; 2850 21; 53 156 MeCOCOMe -180" 4200; 2800 10; 20 155 (CH ) \'" c=o C' / \ tion of the carbonyl groups does not absorb at shorter wavelengths than diacetyl as the energy-level scheme requires (Fig.7) suggesting that the latter compound may be twisted from the tvans-planar configuration. The prevalence of rotational isomerism in the related compound oxalyl chloride is indicated by the marked variation of the spectrum of the molecule with temperature.15* 146 Ley and Arhends 2. phys. Chem. 1932 B 17 177. 146 Saksena and Kagarise J. Chem. Phys. 1951 19 994. 14' Herold Z. phys. Chem. 1932 B 18 265. Howe and Goldstein J. Amer. Chem. SOC. 1958 80 4846.149 Henbest and Woods J. 1952 1150. 150 Jones J. Amer. Chem. SOC. 1945 67 2127. 151 Leonard and Owens J. Amer. Chem. SOC. 1958 80 6039. 152 Cram and Steinberg J . Amer. Chem. SOC. 1954 76 2753. 153 Caserio and Roberts J. Amer. Chem. SOC. 1958 80 5837. 154 Ford and Parry Spectrochim. Acta 1958 12 78. 155 Hoiman Lundberg and Burr;J. Amer. Chem. Sac. 1945 67 1669. 156 Leonard and Mader J. Amer. Chem. SOC. 1950,72 5388. 157 Sidman and McClure J. Amer. Cliem. SOC. 1955 77 6461. 156 Sidman J. Amer. Chem. SOC. 1956 78 1527. 159 Forster J . Chem. Phys. 1957 26 1761. 160 Alder Schafer Esser Krieger and Reubke Annalen 1955 593 23. 314 QUARTERLY REVIEWS The allowed n -+ v* transitions of a-dicarbonyl compounds should have moments directed perpendicular to the molecular plane. Analyses of the vibrational and rotational structure of the longer-wavelength carbonyl band of glyoxal indicate1e1e1G2 that the transition is polarized perpendicular to the molecular plane and that the molecule is trans- planar in both the ground and the excited state.The <CCO bond angle opens by about 3" on excitation and whilst the CC bond is lengthened it is probable that the CO bonds are slightly shortened in the excited state.le2 Triplet absorptions have been reported for b i a ~ e t y l l ~ ~ ~ l ~ ~ and camphor- q ~ i n o n e l ~ ~ at 5050 and 5500 A respectively. The bands have rather large extinction coefficients 0.07 and 0.12 respectively for triplet transitions but the o b s e r ~ a t i o n ~ ~ ~ ~ ~ ~ of enhanced absorption intensity in ethyl iodide solution supports the triplet assignment.8. Unsaturated Sulphur and Nitrogen Compounds The colour of the thiocarbonyl compounds is due to an absorption band with an intensity comparable to that of the It -f v* band in the spectrum of the corresponding carbonyl derivative (Table 10). From this observation it was early that the visible absorption of thioketones is due to the transition of a sulphur lone-pair electron to the thiocarbonyl v*-orbital. At shorter wavelengths the thiocarbonates and the thiocarboxylic acid derivatives give an intense band of v -+ T* origin (Table 10). The corresponding band of the carbonates and the carboxylic TABLE 10. Unsaturated sulphur derivatives Compound [CH,] > C= S Ph,CS M eCS-SEt MeCSOEt M eCSN Me2 (MeS),CS (E tO),CS (Me,N),CS so EtNCS EtSCN ClSCN (SCN) Solvent CH ET E ET E ET E CH CH I 0 CH CH I 0 HX D HX CT CT h l a x (A) 5040 6050 3150 5925 3535 5730 4335 4600; 3060 3770; 2410 3650; 2720 4290; 3030 3030; 2280 3300; 2620 3600; 2900 2450 2500 3 720 2950 Emax 1-10 66 17,000 360 30,900 850 43,800 18; 12,000 19; 8300 41; 15,000 28; 16,000 12; 4800 230; 15,000 0.05; 342 720 50 15 140 Ref.164 8 8 8 164 164 164 164 1 64 164 167 168 169 170 1 70 161 Brand Trans. Faraday Sue. 1954,50,431. 162 King J. 1957 5054. McConnell J . Chem. Phys. 1952 20 700. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 3 15 esters lies in the vacuum-ultraviolet region indicating that the energy separation between the n- and the n*-orbitals is smaller in the thiocarbonyl than in the carbonyl group. This conclusion is consistent with the general chemical instability of the carbon-sulphur double bond.The ionisation potential of sulphur lone-pair electrons is smaller than that of oxygen lone-pair electrons,loB and so in the thiocarbonyl group the n-level is of high energy and the n*-level of low energy accounting for the long- wavelength position of the n + n* absorption. Conjugation of the thiocarbonyl group with monatomic substituents bearing lone-pair electrons shifts the n -+ n* absorption to shorter wave- lengths (Table lo) since the energy of the lowest unoccupied n* orbital is raised by the addition of two n-electrons but only one p n atomic orbital to the conjugated system. An aromatic nucleus on the other hand con- tains as many p n atomic orbitals as n-electrons and on conjugation with the thiocarbonyl group a -*-orbital of lower energy is introduced shifting the n -+ n* band to longer wavelengths.At the same time an intense n -+ n* band appears at shorter wavelengths (Table 10). Electron-donating substituents in the aromatic nuclei of thiobenzophenoiie lower the energy of the n + n* excitation but raise that of the n + n* transition as in the latter an electron is added to the n-system and repulsive interactions increasing the energy of the excited state become greater the larger the electron-donating capacity of the substituent. As the n + n* and the n +- n* transitions of the thiobenzophenones approach one another in energy the greater becomes the degree of mixing between them so that the extinction coefficient of the thiocarbonyl n -+ n* band increases from a maximum value of 10 in thiocyclohexanone the precise value being un- certain because of tautomerism to the thiol,ls4 to 850 in the bis-p-dimethyl- amino-derivative of thiobenzophenone (Table 10).There must be some departure from coplanarity in the aromatic thioketones in order to allow the partial mixing of the otherwise orthogonal n -+ n* and n -+ n* transitions. The two absorption bands of sulphur dioxide in the near-ultraviolet region are due probably to the triplet and the singlet transitions of a sulphur lone-pair electron to the n*-orbital of the molecule. The 3600 A band shows a Zeeman effect,lo7 due to a multiplet upper state and the 2900 A band has the intensity of a singlet n -+ 7 ~ * absorption. The rotational and vibrational structure of the bands indicate that the sulphur-oxygen bonds are lengthened more considerably in the upper state of the 2900 8 absorption than in that of the 3600 A system and that whilst the bond angle is increased in the latter it may be decreased in the former.165 An increase of bond angle is expected in the upper state of a n 3 n* transition and a decrease of angle for that of a n + n* transition.166 However the intensification of the 2900 A band of sulphur dioxide on charge-transfer lg4 Janssen.Rec. Trav. d i m . . 1960. 79.454.464. lg6 Metropolis P.hys. Review 1941 60 283; 295. 166 Walsh J. 1953 2266. 316 QUARTERLY REVIEWS complex formation167 (see below) suggests that the transition moment is perpendicular to the molecular plane and this together with the weak normal intensity of the band is consistent with the n -+ n* but not the n -+ n* assignment. The weak absorption bands of the doubly-bonded nitrogen chromo- phores-the imino- azo- nitroso- and nitro-groups-have the same general properties as the weak carbonyl and thiocarbonyl absorptions as was noted by Burawoy* in his classification of these absorptions as R- bands and the longer wavelength bands of the nitrogen chromophores can be similarly ascribed to n -+ n* transitions.ll The nitroso-group gives a nitrogen lone-pair n -+ n* band in the visible region and an oxygen lone-pair n -+ T* absorption in the ultraviolet region at a wavelength close to that of the corresponding absorption of the nitro-group which possesses only oxygen lone-pair electrons (Table 11).The lower energy of the nitrogen TABLE 1 1. Unsaturated compounds of nitrogen and phosphorus Compound MeN=NMe PhN=NMe cis-PhN= NPh [CH2],> CH.N= CHPri Ph2C = NH PhP= PPh CF3*N=N*CF trans- PhN = NPh { [CHz]5>C=N- Solvent E V CH E E -12 CH CH CL CL Almx (4 3400 3600; 2670 4035; 2595 4432; 3196 4327; 2806 3080 2450 3400 3200; 2450 Emax 4.5 1.6; 1-8 87; 7800 510; 21,300 1518; 5260 89 74 125 2300; 25,000 Ref.187 189 190 191 191 192 192 8 193 n -+ v* transition in nitroso-compounds is due to the smaller ionisation potential of the nitrogen lone-pair electrons. In the series of simple nitrogen chromophores the n~ -f n* bands lie at wavelengths in the order R-NO > R-N=N-R > R,R=NH (Tables 11 and 12). The Wavelength order reflects mainly the decreasing energy of the antibonding n*-orbital arising from the increasing nuclear potential of the atom double bonded to nitrogen in the series. For the nitroso-compounds X-NO the n N -+ n* band has wavelengths in the order R3C-NO > RS-NO > Cl-NO > R2N-NO > RO-NO > F-NO (Table 11).The different wavelengths of absorption are an expression of both the inductive stabilisation by the electronegative atom X of the nitrogen lone-pair electrons in the ground state and the raising of the energy of the anti- bonding n*-orbital by the conjugation of the nitroso-group with the atom X in the cases where that atom has lone-pair electrons. In the latter respect the elements of the first short period are the more effective. In contrast to the effect of monatomic substituents with lone-pair 16' Booth Dainton and Tvin Trans. Faraday SOC. 1959 55 1293. 168 Svatek Zahradnik and Kjaer Acta Chem. Scand. 1959 13 442. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 317 electrons the conjugation of an aromatic nucleus with the nitrogen chromophores shifts the n -+ 7 ~ * bands to longer wavelengths and produces relatively large intensity increases (Tables 11 and 12).The intensity change is most marked in the case of cis-azobenzene which has an n -+ n* band with an extinction coefficient of 1518 compared with the value of 4.5 for the corresponding band of azomethane. In cis-azobenzene the benzene TABLE 1 2. Unsaturated nitrogen-oxygen compounds Compound ButNO CF3*N0 PhNO Me2N.N0 BunO.NO ButS.NO NOF NOCl N2°3 N2°4 MeNO Ph*NO Me,N*NO Bu'O- NO MeN=NMe 3. 3 . 3 . 0 MeN=NMe 0 0 Solvent ET V B E P P HP V CT T HX E HX D E E W Amax (A) 6650; 3000 6925; 2695 7547 2800 3610; 2320 3560; 2220 5988;3390;2290 31 10 4600; 2580 6620 3430; 2700 2700; 2010 3300; 2800 2400 2700 2740; 2170 2760 Emax Ref.20; 100 172 24; 1.8 173 53 174 9000 125; 5900 175 87; 1700 175 15; 980; 10,500 176 weak 177 13; 420 178 20 179 233; 240 180 20; 5000 181 125; loo0 182 6300 183 17 184 43; 7250 185 10,700 186 rings are rotated about the C-N bonds by some 50" out of the C-N=N-C plane,171 the distortion reducing the intensity of the T -f 7 ~ * band but allowing the n + 7 ~ * and the 7~ -f n* transitions to mix extensively. trans-Azobenzene is planar in the crystal,171 but torsional vibrations of the benzene rings about the C-N bonds allow the n -f T* transition to appro- priate a smaller fraction of the n -+ 7 ~ * oscillator strength. The two absorption bands of hexafluoroazomethane may be due to the cis- and the trans-form of the molecule or to n -f 7 ~ * transitions from the two levels formed by the symmetric and the antisymmetric combination of the lone- pair orbitals in the more stable isomzr.9. Aromatic Systems The classification due to Clar,lo of the three band types in the spectra of aromatic hydrocarbons is based primarily upon absorption intensity the lag Pestemer and Litschauer Monatsh. 1935 65 239. 171 Robertson J. 1939 232. Bacon and Irwin J. 1958 778. 318 QUARTERLY REVIEWS a- p- and /3-bands having extinction coefficients (and oscillator strengths) of -lo2 (f - -lo4 (f - lO-l) and -lo5 (f - l) respectively. The general wavelength order is a > p > 18 but in the spectrum of pentacene the a-band lies between thep- and the P-bands and in those of the lower polyacenes the a-band is not observed (Table 13).A second very intense absorption @'-band) is foundlo on the short-wavelength side of the a-p-/I-band system in the spectra of the larger hydrocarbons and a weak triplet absorption (t-band) has been observed on the long-wavelength side of the system in the spectra of a wide range of aromatic C O ~ ~ O U ~ ~ S . The a- and the p-bands appear to be related and so do the p - and the t-bands for the members of each pair undergo similar wavelength shifts on passing from benzene to the larger hydrocarbons but the two pairs differ in behaviour. In general the a- and the ,&bands shift moderately to longer wavelengths with progressive annellation whether linear or angular but thep- and the t-bands undergo shifts which are large and bathochromic if the annellation is linear but small and either bathochromic or less commonly hypsochromic if the annellation is angular (Fig.8 and Table 13). It has been observedlO that the ratio of the frequencies of the 18- and the a-bands is nearly constant in the series of aromatic hydrocarbons ( vj/va = 1-35) and the ratio of the t- and the p-band frequencies is roughly constant ( v t / v - 0-6) in the hydrocarbons with angular annellation (Table 13). 172 BaIy and Desch J. 1908 93 1755. 173 Mason (Banus) J. 1957 3904. 174 Keussler and Luttke 2. Elektrochem. 1959,63 614. 176 Haszeldine and Jander J. 1954 691. 176 Kresze and Uhlich Ber. 1959 92 1048. 177 Johnston and Bertin J. Mol. Spectroscopy 1959 3 683. 178 Collis Gintz Goddard Hebdon and Minkoff J. 1958 438. 17g Mason (Banus) J. 1959 1288. 180 Addison and Sheldon J.1958 3142. Nagakura Mol. Physics 1960,3 152. 18a Wolf and Herold Z. phys. Clzem. 1931 B 13 201. lS3 Jones and Thorn Canud. J. Res. 1949 B 27 828. lS4 Ungnade and Smiley. J. Org. Chem. 1956 21 993. lS5 Langley Lythgoe and Rayner J. 1952 4191. lS6 Gowenlock and Trotman J. 1956 1670. Ramsperger J. Amer. Chem. SOC. 1928 50 123. lB8 Dacey and Young J. Chem. Phys. 1955,23 1302. Rurawoy J. 1937 1865. lgO Birnbaum Linford and Style Trans. Furuday SOC. 1953 49 735. Grammaticakis Bdl. SOC. chim. France 1948 15 975. Is2 Weil Prijs and Erlenmeyer Helv. Chim. Acta 1952 35 616. lg3 Mason unpublished. lg4 McClure J . Chem. Phys. 1949 17 905. lg5 McClure J . Chem. Ph-vs. 1951 19 670. lg6 Shull J . Chem. Phys.. 1949 17 295. lg7 Ferguson Iredale and Taylor J. 1954 3160. lS8 McGlynn Padhye and Kasha J.Chem. Phys. 1955 23 593. log McGlynn Padhye and Kasha J . Chem. Phys. 1956,24 588. *01 Bowen and Brockehurst J. 1955 4320. ao2 Evans J . 1957 1351 ; 1959 2753. 203 Porter Proc. Chem. Soc. 1959,291. Kanda and Shimada Spectrochim. Arta 1959 13 211. F .' 5 E s Hydrocarbon Po PI0 cclo t v,dva V t l V p Ref. E m Benzene 1830 2068 2640 3400 1.44 0.61 196 G Naphthalene 2210 2885 3145 4695 1 *42 0.61 197 Ant hracene 2545 3785 - 6700 - 0.57 199 cl 5;1 Te t racene 2740 4710 - 9755 - 0.48 198 3 Pen tacene 3100 5755 4280 13000 1.38 0.44 203 Phenan t hrene 2547 2945 3445 4630 1-35 0.64 200 cl Chrysene 2670 3190 3600 5050 1-35 0.63 195 Picene 2865 3285 3760 - 1-31 - 1 ,ZBenzanthracene 2900 3590 3850 6060 1-33 0.59 195 1,2 5,6-Dibenzanthracene 3000 3510 3950 5460 1.32 0.64 195 Triphenylene 2595 2870 3425 4200 1.32 0.68 195 TABLE 13.The wavelengths (A) of the origins (a- p- and t-bands) and the maxima (13-bands) of aromatic hvdrocarbons 5 z and the frequency ratios of the /I- and a- bands (vp/va) and of the t- and p-bands (vt/vD,). Data for alcohol-glass solutions at low temperature are given where available c) (triplets) 0 k c Coronene 3050 3415 4100 5154 1 034 0.66 201 m 8 2 Z m cd cl 320 QUARTERLY REVIEWS The ratio vp/vd falls in the overcrowded aromatic hydrocarbons having the value of 1.27 in the case of 3,4:5,6-dibenzo~henanthrene,~O* and the ratio v,/v? falls progressively in the polyacene series (Table 13). The origin of the various bands in the spectra of aromatic hydrocarbons FIG. 8. The variafion of the frequencies of the t- a- p - and 8-bands of aromatic hydro- carbons with linear and angular annellation.has been widely investigated t h e ~ r e t i c a l l y ~ ~ ~ - ~ ~ ~ two particularly simple and informative treatments being Platt's free-electron and the 204 Clar and Stewart J. Amer. Chem. SOC. 1952 74 6235. 205 Platt J. Chem. Phys. 1949 17 484. 206 Platt J. Chem. Phys. 1954 22 1448. 207 Dewar and Longuet-Higgins Proc. Phys. Soc. 1954 67 A 795. 208 Coulson Proc. Phys. Soc. 1948. 60 A 257. 209 Sklar J. Chem. Phys. 1937 5 669. 210 Goeppert-Mayer and Sklar J. Chem. Phys. 1938 6 645. 211 Parr and Crawford J. Chem. Phys. 1948 16 1049; 1949 17 726. 212 Roothaan and Pam J. Chvm. Phys. 1949 17 1001. 213 Niira J . Chem. Phys. 1952 20 1498. 214 Parr Craig and Ross J. Chem. Phys. 1950 18 1561. 215 Craig Proc. Roy. SOC. 1950 A 200 401.216 Coulson Craig and Jacobs Proc. Roy. SOC. 1951 A 206 297. 217 Pariser and Parr J. Chem. Phys. 1953 21 466 767. 218 Pople. Proc. Phys. SOC. 1955 68 A 81. 219 Pariser J. Chem. Phys. 1956 24 250. 220 Moffit J . Chem. Phys. 1954 22 320. 221 Ham and Ruedenberg J. Chem. Phvs. 1956 25 13. 222 Forster Z. phys. Chem. 1938 B 41 287. 223 Jacobs Proc. Phys. Soc. 1949,62 A 710. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 32 1 E molecular orbital approach due to Dewar and Long~et-Higgins.~~~ In the free-electron model the cross links of the cata-condensed hydrocarbons which contain no internal carbon atoms are neglected and the n-electrons move in a circular path of uniform potential with a circumference defined by the perimeter length of the hydrocarbon. The energy levels of an elec- tron moving in a circular path are doubly degenerate corresponding to a clockwise and an anticlockwise electron motion apart from the lowest level which is non-degenerate the electron being stationary.In accord with the Huckel rule of aromaticity (4j + 2) n-electrons are necessary to form a closed shell j the quantum number of the electronic angular momentum having the values 0 &l &2 . . . etc. For a system with (4j + 2) n-electrons the lowest-energy electronic transitions are those between the levels with quantum numbers j- and & ( j + 1) (Fig. 9a). E 0 I -I - 0 I tij t i ) +.i +I *(m+2) *(m+ I ) * m '(m - I) . .... FIG. 9. The electronic transitions responsible .for the a- p- and p-bands of aromatic hydrocarbons (a) according ro the cyclic free-electron model and (b) according to the alternant molecular orbital model.The transitions,j -+ ( j + 1) and -j -+ - ( j + l) involve a change of one unit of electronic angular momentum and are allowed whilst the transi- tions -j -+ ( j + 1) and j 3 -(j + l) involve changes of (2j + 1) units of electronic angular momentum and are forbidden. The upper states of each pair of transitions are degenerate if the circular path has a uniform potential but in the alternating potential due to the atomic nuclei of a 322 QUARTERLY REVIEWS cyclic polyene the degeneracy of the forbidden pair is split into a lower- energy upper state with (2. + 1) nodal planes passing through the atoms and a higher-energy state with (2j + 1) nodal planes bisecting bonds. These are the upper states associated with the a- and thep-band absorp- tions respectively.The allowed pair of transitions remains degenerate in the cyclic polyene giving an excited state of still higher energy with a single nodal plane which bisects the molecule through atoms in one component and through bonds in the other. If a cyclic polyene is cross-linked to form a polycyclic aromatic hydrocarbon the degeneracy of the allowed transition is broken and the two components absorb at different wavelengths giving the ,f3- and the /3’- ban ds. The free-electron treatment of peri-condensed hydrocarbons206 is less satisfactory but the molecular orbital model applies generally to both cata-condensed and peri-condensed alternant hydrocarbons. In all alternant n-electron systems the energy levels are paired corresponding bonding and antibonding levels having respectively equal energies below and above a centre of zero binding energy (Fig.9b). If the levels in order of increasing energy are #J2 . . . $m $tm + 1) . . . $2m the highest occupied level is t,hrn. The transition t,bm -+ y!qm + 1) is unique giving rise to the p-band in the spectra of aromatic hydrocarbons but the transitions $(m - 1) -+ #J(m + 1) and z+hm -+ $(m + 2) are degenerate and electronic interactions mix the resulting configurations the antisymmetric combina- tion giving a low-energy forbidden upper state that of the a-band absorp- tion and the symmetric combination a high-energy allowed excited state that of the ,%band absorption (Fig. 9b). The a- and the ,&bands thus arise from transitions between the same set of energy levels accounting for the similar trends in the shifts of these bands with annellation in the series of aromatic hydrocarbons.In general electronic interactions become smaller the larger the hydrocarbon and so do the energy separations between corresponding occupied and un- occupied molecular orbitals so that the approximate constancy of the ratio vB/v, may be qualitatively understood if not quantitatively explained. The molecular orbital treatment accounts for the different trends in the energy of the p-band on linear and angular annellation. It has been ~ h o ~ n ~ ~ ~ ~ ~ ~ ~ that the upper and the lower levels of the p-band absorption 224 Coulson and Rushbrooke Proc. Roy. SOC. Edinburgh 1943-49 62 A 350. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 323 the orbitals z,bm and t,b(m + 1) have energies converging to a common zero of binding energy in the linear polyacene series (11) as n -+ co whilst they converge to a finite energy separation in the skew hydrocarbon series (12).In both series however the levels which are active in the absorption giving the 01- and the /?-bands converge to a finite energy ~ e p a r a t i o n ~ ~ ~ accounting for the moderate increase in the wavelength positions of the a- and the j3- bands with either linear or angular annellation and for the rapid bathochromic shift of the p-band with linear annellation in the polyacene series (1 1). The t- and the p-bands of aromatic hydrocarbons undergo parallel shifts on annellation as the t-band is the triplet counterpart of the singlet p-band.lg4 Electronic motions are so correlated that the average distance between electrons with the same spin is larger than that between cor- responding electrons with opposed spins the latter having the higher energy owing to the larger repulsive forces.A triplet transition accordingly requires a smaller energy than the corresponding singlet transition but in a series of related compounds such electron repulsion effects diminish as the size of the molecule increases resulting in a progressively smaller singlet- triplet energy separation. With angular annellation the decreases in the singlet-triplet splitting and in the energy separation between the orbitals t,bm and $(m + 1) are roughly commensurate giving the ratio v t / v p an approximate constancy but with linear annellation the orbital energy separation diminishes the more rapidly giving a falling vt/vD ratio in the polyacene series (1 1).The simple models have been elaborated in a number of more detailed treatments,210-221 which support most of the qualitative conclusions of those models and give quantitative theoretical transition energies and probabili- ties for benzene and the polyacenes in rather variable agreement with experiment. Simple treatments may give results in quite good accord with experiment owing to the cancellation of neglected factors and it is not infrequently found that the agreement becomes poorer when some of those factors are considered in an elaboration. However there are qualitative divergences of interpretation which can be checked experimentally. The free-electron model applies to benzene when the quantum number j is unity and it requires that the 01- p- and 16- bands of benzene are due to transitions from the ground state with a symmetry A (13) in the hexagonal symmetry group DGh to upper states with the symmetries Bzu (14) B, (15) and El (16) severally the t-band transition also having a B1 upper state symmetry.All other interpretations suggested hitherto give the same upper state symmetries for the cases of the 01- ,8- and t-band transitions but two valence-bond t r e a t m e n t ~ ~ ~ ~ y ~ ~ ~ and an orbital indicate that the p-band transition has an upper state with Ezs (17) symmetry. In the free electron model the Ezff upper state arises from two-quanta transitions e.g. j = 0 -f j = &2. A third possibility that the p-band is due to a Rydberg transition with an upper 324 QUARTERLY REVIEWS state involving carbon 3s orbitals,22e would require a transition moment perpendicular to the plane of the ring but it is known from the measure- ment of the spectrum of crystalline hexamethylbenzene with polarized light,227-231 that the a- p- and @-bands have transition moments lying almost entirely in the molecular plane.In the spectrum of benzene vapour232,233 the a-band system has vibra- tional structure with a missing band origin indicating that the absorption is due to a forbidden transition. Relative to the missing origin a band is observed 608 cm.-l to lower frequencies and another 521 cm.-l to higher frequencies representing one quantum of the ring angle bending vibration with a symmetry Ezs (1 8) in the ground and the excited state. The distortion of the benzene ring due to this vibration allows the a-band system to appear weakly by borrowing intensity from the and the particular symmetry of the vibration establishes that the upper state of the a-band transition has either B, or B, symmetry.Studies of the frequencies and intensities of the vibrational bands in the a-band ~ y ~ t e m ~ ~ ~ * ~ ~ indicate that benzene is planar and hexagonal at equilibrium in the upper state with the carbon-carbon bonds lengthened from 1.39 to 1.43 A. The a-system of the benzene spectrum is similarly characterised by the effects of substituents upon the absorption intensity of the band.236-238 Transitions to the B, and B, upper states of benzene are forbidden because the component transition moments cancel. The instantaneous 226 Dickens and Linnett Quart. Reviews 1957 11 291.226 Hammond Price Teegan and Walsh Discuss. Faraday SOC. 1950 9 53. 227 Scheibe Hartwig and Muller 2. Elektrochon. 1943 49 372. 228 Broude J. Expt. Theor. Phys. U.S.S.R. 1952 22 600. 229 Craig and Lyons Nature 1952 169 1102. 230 Nelson and Simpson J. Chem. Phys. 1955 23 1146. 231 Schnepp and McClure J. Chem. Phys. 1957,26 83. 232 Sponer Nordheim Sklar and Teller J. Chem. Phys. 1939 7 207. e33 Garforth Ingold and Poole J. 1948 406. 234 Craig J. 1950 59. 235 Craig J. 1950 2146. 836 Sklar J. Chem. Phys. 1942 10 135; Rev. Mod. Phys. 1942 14 232. 237 Forster Z. Naturforsch. 1947 2 A 149. 238 Platt J. Chem. Phys. 1951 19 263. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 325 directions and magnitudes of the component moments derived from the symmetries of the upper state wave functions (14) and (15) are (19) and (20) for the B, and the B1 states respectively.In a monosubstituted benzene the component moment at the position of substitution is changed resulting in an enhanced absorption intensity. The same intensity increase is given by ortho- or rneta-homo-disubstitution as the vector sum of the component moment changes equals the change due to monosubstitution. With para-homo-disubstitution the intensity increase is fourfold as the component moment change is doubled and the additional absorption intensity is proportional to the square of the net resultant transition moment. No intensity increase results from 1,2,3- 1,3,5- or hexa-homo- substitution owing to the cancellation of the component moment changes. If the conjugating effect of the substituent is not too large the predicted intensity changes are found experimentally supporting a B, or a B, symmetry for the upper state of the a-band transition.The effects of polysubstitution with different groups on the absorption intensity of the a-band of benzene and other aromatic hydrocarbons are similarly additive according to the vector diagrams (19) or (20) and their analogues for the larger Electron-withdrawing and electron-donating groups produce opposed changes in the component transition moments so that 4-fluorobenzotrifluoride for example has a weaker a-band than either fluorobenzene or benzotrifl~oride.~~~ Tables of “spectroscopic moments” which are measures of the component transition moment changes have been drawn up for a wide range of sub~tituents,~~~ but the values appear to vary somewhat from compound to compound the methoxy-group in particular changing the magnitude and even the sign of the halogen “spectroscopic moments”.239 Evidence that the a-band transition of benzene has a B, and not a B, upper state is provided by the spectrum of crystalline benzene measured with polarized light.240y241 Owing to the perturbation of the crystal field the band origin appears in the spectrum of the a-band system and it is a doublet the two components having opposite polarisation properties.This result is in accord with a B, upper state the B1 state requiring that the band origin should appear as a singlet.242 The gross polarisations of the unresolved a- and p-bands of 1,4-dirneth- oxybenzene 1,4-bisdimethylarninobenzene and the radical ion of the latter have been determined for the crystal or for rigid-glass In crystalline 1,4-dimethoxybenzene the a-band transition moment is predominantly in-plane polarised with a direction perpendicular to the 239 Goodman and Frolen J .Chem. Phys. 1959,30 1361. 240 Broude Medvedev and Prikhotjko J. Expt. Theor. Phys. U.S.S.R. 1951,21 665. 241 Fox and Schnepp J . Chem. Phys. 1955,23 767. 242 Davydov J. Expt. Theor. Phys. U.S.S.R. 1951 21 677. 243 Albrecht and Simpson J. Chem. Phys. 1955 23 1480. 244 Albrecht and Simpson J. Amer. Chem. SOC. 1955 77 4454. 246 Albrecht J. Amer. Chem. SOC. 1960 82 3813. 326 QUARTERLY REVIEWS line joining the methoxy-groups i.e. y-polarised (21) whilst the moment of thep-band transition lies largely along that line i.e.,is x-polarised (21).243 The para-methoxy-groups give rise to a net transition moment composed of the vector sum of the component moment changes due to substitution which is y-polarised for a B, upper state (19) x-polarised for a B1 excited state (20) and zero for a E2 upper state (22) The observed polar- ised spectrum of 1,4-dimethoxybenzene thus confirms a Bzu upper state for the benzene a-band transition and supports the B1 assignment for the p-band excited state.However the E, upper state for the benzene p-band transition is not eliminated. The amount of intensity borrowed by a forbidden from an allowed transition is inversely proportional to the square of the transition energy separation (e.g. eqn. 7). The energy interval between the benzene p- and ,&bands is particularly small (-6000 cm.-I) and the degeneracy of the allowed transition to the El state responsible for the ,&band is split by para-disubstitution the component with the x-polarisation (1 6a) having the lower transition energy.Given commensurate vibrational couplings an adjacent forbidden transition should borrow intensity predominantly from the lower energy component and should assume mainly the same x-polarisation sufficient to account for the 4:l ratio for the x:y polarisation of the p-band in the crystal spectrum of 1,4-dirnethoxybenzene. In contrast to the large intensity variations induced on the benzene a-band by substituents the intensity changes of the benzene p-band due to substitution are small. They are not large enough to allow a reliable distinction to be made experimentally between the possible p-band transition upper states B, and Ezs from the different intensity changes following di- or poly-substitution predicted by the polarisation diagrams (20) and (22) respectively.The insensitivity of the p-band intensity to substituent perturbations suggests that the intensity borrowing from the /3-band is due primarily to vibrational coupling. In the spectrum of benzene vapour the vibrational structure of the p-band system differs markedly from that of the a-system though a similarity would be expected for the B1 upper state. In detail the vibrational structure of the p-band system is consistent with the E, upper state and does not support the B1 assignment.246 Temperature-sensitive bands due to transitions from excited vibrational levels of the electronic ground state are few and very weak in the vapour spectrum of the p-system.The B1 assignment requires that the intensity ratio of the temperature-sensitive to the temperature-invariant bands should be as much as 1 15 but a ratio of less than 1 100 is expected for the E, upper state. The observation that each member of the main band progression is a doublet suggests that the upper state has a degeneracy which is split 846 Dunn and Ingold Nature 1955 176 65. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 327 owing to the Jahn-Teller effect,247 into two states of slightly different energy. Although the forbidden transition to a B1 upper state would borrow intensity from the allowed degenerate El transition and reflect in the vibrational structure of the p-system the Jahn-Teller splitting of the El state the reflection would be weak and the splitting of the vibrational doublets small.However the state is itself degenerate and it should give the observed relatively large vibrational doublet frequency separa- t i o n ~ . ~ ~ ~ The vibrational structure of the benzenep-band system is diffuse owing primarily to the multiplicity of totally symmetric vibrations in the lower symmetry excited states derived from the splitting of the Ezs state. Part of the diffuseness may arise from the presence of the B1 as well as the Ezs transition in the p-band absorption but whilst identifying features of the latter have been observed characteristics of the former have not been detected as yet. The transition-moment directions and upper-state symmetries of the a- p - and /3-bands in the spectra of the larger hydrocarbons have been ascertained generally from the polarised spectra of the crystals or of solid solutions in crystals of related molecules.20.21 The a-band transitions have been characterised by the intensity changes due to sub~titution,~~~-~~O and more recently the shapes of the rotational band contours in the a-band system of the spectrum of naphthalene vapour have been used to elucidate further the nature of the lowest singlet transition of naphthalene.251 10.Non-alternant Hydrocarbons The non-alternant hydrocarbons containing a Huckel number of con- jugated atoms (4j + 2) where j is an integer usually give a spectrum consisting of a weak a moderate and a strong band system at progres- sively shorter wavelengths similar to the a- p - and ,f3-band system of the alternant aromatic hydrocarbons (Table 14).A similar band pattern is not so generally apparent in the spectra of non-alternant hydrocarbons containing 4j conjugated atoms but the absorption may resemble that of the vinyl derivative of the corresponding hydrocarbon with (4j - 2) conjugated atoms as in the case of acenaphthylene which has a l-vinyl- naphthalene type of spectrum. Both classes of non-alternant hydrocarbon generally absorb at longer wavelengths than their alternant isomers though this property is the more marked in thecompounds of (4j + 2) conjugated atoms (Table 14). The free electron model can be used to interpret the spectra of non- alternant as well as alternant cyclic compounds provided that the molecule 247 Jahn and Teller Proc. Roy. SOC. 1937 A 161 220. 248 Platt J . Chern.Phys. 1951 19 1418. 248 McConnell. J . Chern. Phys. 1952 20 1978. 260 McConnell and McClure J. Chern. Phys. 1953 21 1296. 251 Craig Hollas Redies and Wait Proc. Chern. SOC. 1959 361. 328 QUARTERLY REVIEWS contains (4j + 2) n-electrons and so gives a closed shell configuration in the ground state of the cyclic polyene composed of the perimeter conjugated Cross links may be added to a cyclic polyene in two distinct ways. Either atoms of opposed parity one starred and the other unstarred can be linked to form an alternant hydrocarbon or atoms of the same parity both starred or both unstarred may be joined to give a non- alternant hydrocarbon. In each case the upper states of the a- p- and TABLE 14. Non-alternant hydrocarbons Compound Fulvene Azulene Acenaphthylene Pleiadiene Acepleiadylene 4,6-Dimethyl-l,8- pentenoazulene 2,4-Dimethyl- 1,lO- pen ten o hep t alene 1,2 4,SDibenzo- pentalene Amax (%L> 3730; 2420 6560;3470;2700 3380; 3240;2650 5500;3500;2500 5500;3300;2500 4760; 3830; 3310; 2630 7930; 3930; 2500 4150; 2810 Emax 280; 14,000 300; 4000; 47,000 4000; 9700; 2000 100 6300; 25,000 1500; 35,000 20,000 870; 2100; 6300; 44,700 128; 14,800; 50,000 15,000; 69,000 Ref.253 254 256 258 258 259 259 260 &band transitions largely retain their identity and undergo common additive energy shifts but the relative energies of the states are altered in characteristically different ways.22o When atoms of opposed parity are joined the energies of the a- and the ,&band upper states are relatively unchanged but the p-band upper state moves to lower energies and that of the ,@-band which is degenerate with the P-band in the cyclic polyene to higher energies.The linking of atoms with the same parity lowers the energy of the a-band upper state and raises that of one component of the cyclic polyene P-upper state leaving the energies of the other component and the p-upper state relatively unaltered.220 In a non-alternant hydro- carbon the energy intervals between the 01- and the /I-band and between the p- and the P-band are thus expected to be larger and smaller respect- ively than the corresponding energy separations in an isomeric alternant Me (23) hydrocarbon and the latter in general should absorb at shorter wave- lengths. The reported frequency intervals between the a- and the /?-bands and between the p - and the P-bands are respectively 19,400 and 6300 cm.-l for azulene and 13,950 and 10,250 cm.-l for naphthalene.252 252 Mann Platt and Klevens J.Chem. Phys. 1949 17 481. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 329 However the p-/3-band separations of pleiadiene and the dimethyl- hydrocarbon (23) are comparable to that of anthracene and larger than that of phenanthrene though the a-/%band intervals of the non-alternant tricyclic hydrocarbons are the larger (Tables 13 and 14). The spectra of all alternant hydrocarbons show a bathochromic shift following alkyl substitution but depending upon the particular absorption band and the position of substitution the spectra of non-alternant hydrocarbons give either blue or red shifts. The substitution of a methyl group at the exocyclic position of fulvene shifts the bands at 3730 and 2420 A to 3500 and 2550 A respectively.253 The a-band of azulene under- goes bathochromic shifts of 360 and 150 A on methyl substitution in the 1- and the 5-position respectively and hypsochromic shifts of 200 and 150 A for methyl substitution in the 2- and in the 4- or 6-position respect- i ~ e l y .~ ~ ~ In the polymethylazulenes these shifts of the a-band are additive.z55 The azulene p-band shifts are all bathochromic those for 2- 4- or 6- methyl substitution being larger than the displacements induced by methyl groups in the 1- or the 5-po~ition.~~~ The hyperconjugative effect of an alkyl group reduces the energies of electronic transitions in both alternant and non-alternant hydrocarbons by extending the conjugated ~ y s t e m . ~ ~ l - ~ ~ ~ Owing to the pairing property of corresponding bonding and antibonding 77-orbitals in alternant systems the inductive effect of alkyl groups does not produce a shift of the a- p- and P-bands in the spectra of aromatic hydrocarbons and the hyper- conjugative effect is d ~ m i n a n t .~ ~ ~ - ~ ~ ~ However bonding and antibonding n-orbitals are not paired in non-alternant systems and the inductive effect of an alkyl group gives rise to a blue shift of an absorption band if the one- electron charge density at the position of substitution is larger in the upper than in the lower level of the transition involved in the absorption or a red shift if there is a converse inequality in the charge densities. The Huckel theory gives one-electron charge-density distributions for the lower and the upper energy levels of the transitions active in the long wavelength absorption of f ~ l v e n e ~ ~ ~ and which account semi-quantita- tively for the shifts observed on methyl substitution.The observation255 that the wavelength displacement of the azulene a-band following poly- methyl substitution equals the sum of the shifts induced by each methyl 253 Thiec and Wiemann Bull. SOC. chim. France 1956 177; 1957 102. 254 Plattner and Heilbronner Helv. Chim. Acta 1947 30 910; 1948 31 804. 255 Gordon Chem. Reviews. 1952 50 157. 256 Craig Jacobs and Lavin J. Biol. Chem. 1941 39 277. 257 Allen and Van Allen J. Org. Chem. 1953 18 882. 258 Roekelheide and Vick J. Amer. Chem. SOC. 1956 78 653. 259 Hafner and Schneider Annalen 1959 624 37. 260 Blood and Linstead J. 1952 2263. 261 Pullman Mayot and Berthier J.Chern. Phys. 1950. 18 257. 262 Longuet-Higgins and Sowden J. 1952 1404. 263 Coulson Proc. Phys. Soc. 1952 65 A 933. 264 Peters J. 1957 646 1993 4182. 265 Berthier J. Chem. Phys. 1953 21 954. 330 QUARTERLY REVIEWS group singly is explained262 by the additive property of small inductive and hyperconjugative effects. 11. Heteroaramatic Systems The most extensively investigated heteroaromatic compounds are those containing nitrogen atoms. The spectra of N-aromatic systems containing only six-membered rings show three regions of absorption which may be related to the a- p - and /I-band systems of the parent aromatic hydro- carbons.266 In polyaza-compounds the intensity of the long-wavelength a-band is given by transition-moment vector addition (19) or its analogue for the larger hydrocarbons up to tetra-aza-derivatives when the predicted intensities become markedly too large267 (Table 15).To a limited degree the vector-addition treatment can be extended to account for the observed a-band absorption intensities of the substituted azine~.~~' The wavelength shifts of the monocyclic azine a-bands relative to that of benzene have been explained in studies of the perturbing influence of the nitrogen atoms upon the benzene energy TABLE 15. The monocyclic azines in cyclohexane solution278 Compound Pyridine Pyridazine Pyrimidine Pyrazine sym-Triazine 3,5,6-Trimethyl- 1,2,4-triazine sym-Te t razine n+r* bands 2700 450 3400 315 2980 326 3280 1040 2720 890 3840 520 5420 829 Amax(& emax r -+ r* bands Ansx(A) Emax 2510 2000 2460 1300 2430 2030 2600 5600 2220 150 2640 5100 2520 2150 On the long-wavelength side of the a- p- ,&band system in the spectrum of many N-aromatic compounds lies a weak-moderate absorption with the solvent sensitivity of a n -+ r* band.11s270-278 The n -f 7r* absorption undergoes bathochromic and hypsochromic shifts following substitution of the azine with electron-withdrawing and electron-releasing groups respectively whilst the 7~ + T* a-bands move to longer wavelengths with Badger Pearce and Pettit J.1951 3199. 267 Mason J. 1959 1247. e68 Murrell Mol. Physics 1958 1 384. 269 McWeeny and Peacock Proc. Phys. Soc. 1957 70 A 41 270 Halverson and Hirt J . Chem. Phys. 1949 17 1165. 272 Hirt Halverson and Schmitt J. Chem. Phys. 1954 22 1148. 274 Hirt Spectrochim. Acta 1958 12 114. 278 Ito Shimada Kuraishi and Mizushima J.Chern. Phys. 1957 26 150s. 276 Uber J . Chem. Phys. 1941 9 777. 277 Stephenson J. Chem. Phys. 1954 22 1077. Haiverson and Hirt J. Chem. Pnys. 1951 19 711. Hirt King and Cavagnol J. Chem. Phys. 1956 25 574. Mason J. 1959 1204. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 33 1 both types of substituent. In the monocyclic azines the n -+ n* band substituent shifts can be accounted for semiquantitatively on the assump- tion that an electron is promoted from a nitrogen lone-pair orbital to an unoccupied n*-orbital with a benzene-like one-electron charge distribu- tion.267,278,279 At positions where the charge density in the n*-orbital is substantial the wavelength displacement of the n -+ n* band on substitu- tion is large whilst the effect of substitution at nodal positions in the n*-orbital is small.The n -+ 7 ~ * absorptions of N-aromatic compounds are of only moderate intensity as a transition is allowed only from the 2s component of the 2s,2p2 hybrid lone-pair orbital of the nuclear nitrogen atom to a n*-orbital of the ring. Vibrational analyses show that the n -+ n* azine absorption is due to an allowed transition since the band origin appears strongly and that a change of bond angles is general in the upper states of these transi- tions as the strongest progression is in the upper-state frequency of a ring-angle bending vibration with the form (1 8).274-2769280 The rotational structure .of the monocyclic azine n -+ n* band origins indicates that the transition moment has the expected orientation per- pendicular to the molecular plane.280-283 At equilibrium in the upper state of the n -+ 7 ~ * transition the monocyclic azines are planar or nearly so the diazines having a larger average in-plane moment of inertia than in the ground ~ t a t e .~ ~ ~ * ~ ~ ~ However the average in-plane moment of inertia of sym-tetrazine which has a n -+ n* absorption of particularly low energy decreases on excitation indicating a shortening of the ring bonds.281 Wavelength displacements rather than intensity changes are the more marked feature of the effect of substituents upon the 7~ -f 7 ~ * absorption bands of the azines (Table 16) in contrast to the case of the a-band of benzene. In the pyridine derivatives (24) where X = -SR -OR or -NR2 the a-band wavelength covers a range of 1260 A corresponding to a frequency range of 15,000 cm.-l the position of absorption varying with the substituent the position of substitution and the ionic species of the compound.In general the pyridine compounds (24) and the corres- ponding polyazines exist under the appropriate conditions in any one of four spectrally different charged species namely the neutral form (N) e.g. (25) the cation (C) e.g. (26) the anion (A) e.g. (27) or the zwitterion 279 Goodman and Harrell J. Chern. Phys. 1959,30 1131. 280 Mason J. 1959 1263. 281 Mason J. 1959 1269. 282 Innes; Mason; Handbook of the 4th Meeting on Molecular Spectroscopy Bologna 1959. 283 Innes Merritt Tincher and Tilford Nature 1960 187 500. 332 QUARTERLY REVIEWS (Z) e.g. (28). Experimentally the following wavelength orders are observed for the a-band position of the pyridine derivatives (24),-2 > A > C > N for a given substituent and position of substitution,-3 > 2 > 4 for a particular substituent and a given charged species and X = S > N > 0 for a given position of substitution and a particular charged species284 (Table 16).TABLE 16. Substituted pyridine derivatives Pyridine Species Amad (A) 3-Methoxy- N 2760; 2160 3-Amino- N 2880; 2310 3-Methylthio- N 2940; 2530 2-Methylthio- N 2920; 2470 4-Methylthio- N 2630; 2140 3100; 2550 3130; 2680 3610; 2900 3 - Mercap to- Emax 3960 ; 8 320 3000; 8200 2500; 8700 4200; 8700 12,500; 9500 3200; 7800 2600; 13,500 2300; 12,000 Ref. 284 284 285 285 285 285 285 285 These observations can be accommodated by a carbanion model in which it is assumed that the a-band of the pyridine derivatives (24) is due to a transition between energy levels which are akin to the highest occupied and the lowest unoccupied n-orbitals of the benzyl anion.284 The one- electron charge distributions of these orbitals given by the Hiickel theory afford a measure of the change in the transition energy brought about by the replacement of a nuclear carbon atom in the benzyl anion by a neutral or a positively charged nitrogen atom and by the replace- ment of the exocyclic carbon atom of the benzyl anion by a neutral or a negatively charged sulphur oxygen or nitrogen atom.The electro- negativity orders oxygen > nitrogen > sulphur and positively charged atom > neutral atom > negatively charged atom together with the particular values of the one-electron charge densities at the various positions in the two n-orbitals of the benzyl anion then give the observed wavelength orders of the a-band position in the spectra of the pyridine compounds (24).The treatment can be extended to the corresponding diazine derivatives and to the wavelength displacements of the second n -f 7 ~ * absorption the band.^^^ The pairing property of the bonding and the antibonding n-orbitals in an alternant aromatic hydrocarbon is lost in the corresponding hetero- aromatic molecules and alkyl substitution may give rise to either batho- chromic or hypsochromic shifts of the band positions in the spectra of N-aromatic compounds. The a-band of 2- and 3-picoline for example lies 70 and 65 A respectively to the red of the corresponding absorption maximum of pyridine but in 4-picoline the band is displaced 30 A to the 284 Mason J.1959 1253; 1960 219. 286 Albert and Barlin J. 1959 2384. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 333 blue region.286 These shifts can be accounted for semi-empirically by the m e t h o d ~ ~ ~ ~ s ~ ~ ~ used for the treatment of the analogous phenomenon observed in the spectra of the non-alternant hydrocarbon^.^^^ Both n -+ n* and n -+ n* triplet absorptions have been observed in the spectra of N-aromatic C O ~ ~ O U ~ ~ S . ~ ~ ~ ~ ~ ~ ~ The n -+ T* singlet- triplet splitting is rather small (2000-5000 cm.-l) since an electron in a nitrogen lone-pair orbital and another in a ring n-orbital move in different regions of space and the electron repulsion energy difference between the singlet and the triplet configurations is not large.279 The most extensively studied series of N-aromatic compounds containing five-membered rings is that of the porphins chlorins and the tetrahy- d r o p o r p h i n ~ .~ ~ ~ - ~ ~ ~ The spectra of the porphins consist in general of four moderate intensity bands in the visible (bands I-IV at progressively shorter wavelengths) and a very high intensity absorption (Soret band Emax -lo5) near 4000 A. The intensities of bands I and 111 are particularly sensitive to the nature position and number of substituents in the porphin nucleus and so is that of the longer-wavelength band in the spectra of the cation and the metal complexes. Empirical rules based on the observed intensity changes were early established to distinguish the various types of sl.ibstituted porphin and c h l ~ r i n .~ * ~ The porphin spectrum has been the subject of many theoretical the molecular orbital free-electron model of Platt296 provid- ing the most comprehensive treatment of the experimental data. The chromophore common to porphin and its di- and tetra-hydro-derivat ives is an 18-membered ring system (29 heavy outline). In the free-electron model of this cyclic system the highest occupied 7-electron level is de- generate with a quantum number j = 4. There are degenerate pairs of forbidden transitions and allowed transitions to the lowest unoccupied n-electron levels for which j = & 5 and these transitions give rise to the visible and the Soret band systems respectively. The degeneracy of the 286 Andon Cox and Herrington Trans. Faraday SOC. 1954 50 918. 287 Chandra and Basu J. 1959 1623.288 Goodman and Kasha J. Mol. Spectroscopy 1958 2 58. 289 Stem and Wenderlein 2. phys. Chem. 1936 A 175 405; 1936 A 176 81. 290 Stern Wenderlein and Molvig 2. phys. Chem. 1936 A 177 40. 291 Baguley France Linstead and Whalley J. 1955 3521. 292 Eisner Linstead Parkes and Stephen J. 1956 1655. 293 Eisner and Linstead J. 1955 3742 3749. 294 Rimington Mason and Kennard Spectrochim. Acta 1958 12 65. 295 Rabinowitch Rev. Mod. Phys. 1944 16 226. 296 Platt “Radiation Biology,” Vol. 111 ed. Hollaender McGraw-Hill New York 297 Kuhn J . Chern. Phys. 1949 17 1198. 298 Simpson J. Chem. Phys. 1949 17 1218. 2 9 9 Longuet-Higgins Rector and Platt J. Chem. Phys. 1950 18 1174. 300 Nakajima and Kon J. Chem. Phys. 1952 20 750. 301 Matlow J. Chem. Phys. 1955 23 673. 302 Barnard and Jackman J . 1956 1172.303 Seely J. Chem. Phys. 1957 27 125. 304 Gouterman J . Chem. Phys. 1959 30 1139. 1956. 4 334 QUARTERLY REVIEWS forbidden transition is broken in the case of porphin bands I and 111 being the band origins of the two components. Bands I1 and IV are vibra- tional absorptions further bands of the same kind being detected under higher resolution particularly at low temperatures.294 In the spectra of the cation and of the metal complexes the degeneracy of the forbidden transi- tions is not split the longer wavelength band representing the band origin of both components. t t ' (29) o* 4- I 4- -c I t ' (31) The addition of atoms and bonds to the 18-membered ring system (29) giving the porphin nucleus introduces perturbations which give the for- bidden transitions an allowed character but the visible absorption of porphin is only moderate in intensity owing to the extensive cancellation of the component transition moments.Substitution in the porphin nucleus produces changes in those moments resulting in an enhanced absorption intensity but the absorption strengths of bands I and I11 increase in different ways which are characteristic of the substitution pattern. 1,2,5,6-Tetra-alkyl-substitution enhances the intensity of band 111 but not TABLE 17. Porphins; absorption in the visible region Compound Porphin 1,5-Dimethyl- 2,6-diethyl- 2,3-Dimethyl- 1,4-diethyl- Octamethyl- Mono-aza-ztio- @-Diaza-ztio- Octamethyltetra-aza- h a x (A) 6165; 5635; 5195; 4895 6190; 5605; 5300; 4960 6150;5620;5210;4920 6240;5700;5330;4990 6100;5600;5320;500 6200;5690;5430;5150 6270; 5970; 5560 Emax Ref.860; 4980; 2640; 15,750 293 990; 7020; 8720; 13,130 289 2310; 5850; 5680; 14,820 289 5100; 6400; 9300; 13,100 292 27,130; 9010; 25,100; 7970 290 47,050; 7580; 27,800; 9000 290 72,400; 8700; 45,700 29 1 that of band I and the addition of another four pyrrolic alkyl groups increases the intensity of band I but leaves that of band I11 unchanged (Table 17). Thus the component transition moments at the 1,2,5,6- and the 3,4,7,8-positions are not equivalent for either band I or band 111 and the two absorptions must be polarised at right angles. The spectrum of the porphin crystal measured with polarised light confirms that bands I and I11 have different polarisation directions.294 MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 335 The effect of substituents upon the intensities of the visible absorption bands of porphin indicates that the component transition moments have the directions (30) and (31) for bands I and 111 respectively.296 These directions can be derived from the forms of the highest occupied and the lowest unoccupied r-orbitals of p ~ r p h i n ~ ~ ~ apart from a uniform re- orientation of n/4 necessitated by the imperfections of the Tables have been drawn up for the component transition moment changes induced by a wide variety of ~ubstituents,~~~ allowing the visible absorption spectrum to be used diagnostically in conjunction with the polarisation diagrams (30) and (3 I) for structure determinations in the porphin and chlorin series.The intensity increase of the visible absorption resulting from the reduction of porphin to chlorin or to tetrahydroporphin can be treated as a substituent effect unifying the interpretation of the spectra of the series as a 12.Carbanions and Carbonium Ions A number of carbon ions long postulated as intermediates in organic reactions have been characterised recently by their visible and ultraviolet absorption ~ p e c t r a . ~ * ~ - ~ ~ ~ The ions containing conjugated ring systems are of three main types odd-alternant systems notably the arylmethyl ions the even-alternant ions formed by the oxidation or reduction of aromatic hydrocarbons and non-alternant systems such as the tropylium and cyclopropenyl cation. Both H i i ~ k e l ’ s ~ ~ ~ and the self-consistent field molecular orbital theory327 indicate that the anion and the cation of an alternant system whether even or odd should have the same long-wave- Vidal Kohn and Matsen J.Chem. Phys. 1956,25 180. Lavrushin and Verkhovod J. Gen. Chem. U.S.S.R. 1956 26 3005. Rosenbaum and Symons (a) Mol. Physics 1960 3 205; (b) J. 1961 1. Leal and Pettit J. Amer. Chem. SOC. 1959 81 3160. Deno Jaruzelski and Schriesheim J. Org. Chem. 1954,19 155; J. Amer. Chem. 300 Winstein and Ordronveau J. Amer. Chem. Soc. 1960,82 2084. 311 Deno Groves Jaruzelski and Lugasch J. Amer. Chem. Soc. 1960 82,4719. 312 Grace and Symons J. 1959,958. 313 Grinter and Mason unpublished. *14 Gold and Tye J. 1952 2173 2181 and 2184. 316 Dauben Gadecki Harmon and Pearson J. Amer. Chem. SOC. 1957,79,4557. alo Naville Strauss and Heilbronner Helv. Chim. Actu 1960 43 1221. 317 Doering Saunders Boyton Earheart Wadley Edwards and Laber Tetrahedron 318 Reid J.Amer. Chem. SOC. 1954,76 3264. 310 Pettit J. Amer. Chem. Soc. 1960 82 1972. 320 Farrell Mason and McCaffery unpublished. sB1 Breslow and Hover J. Amer. Chem. Soc. 1960 $2 2644. 3*1 Dalinga Mackor and Verrijn Stuart Mol. Physics 1958,1 123. 3a3 Balk Hoijtink and Schreurs Rec. Truv. chim. 1957,76 813; Hoijtink and We& 3a4 Paul Lipkin and Weissman J. Amer. Chem. Sac. 1956,78 116. 326 Aalbersberg Hoijtink Mackor and Weijland J. 1959 3049 and 3055. sf? Longuet-Higgins and Pople Proc. Phys. Sac. 1955,68 A 591. Soc. 1955,77 3044. 1958 4 178. land ibid. 1957,76 837. Hush and Rowlands J. Chem. Phys. 1956,25 1076. 4* 336 QUARTERLY REVIEWS length absorption. The bands of an aromatic hydrocarbon mono-negative and -positive ion have remarkably similar wavelength positions and absorption intensities and those of an arylmethyl anion and cation do not differ widely (Table 18).The measurement of the spectra of the anions requires the use of solvents with a low dielectric constant and it is found that the band positions vary with the solvent and with the particular alkali- metal g e g e n i ~ n . ~ ~ ~ Ion-pairing is probable in these solvents but the absorption wavelength of the free anion in the particular solvent employed can be obtained by extrapolating the measured band positions to the limit of infinite gegenion A number of aromatic hydrocarbons give both a mono- and a di-anion which are characterised by marked spectral differences,323 and in acid solution form a praton a d d ~ c t ~ l ~ e.g. (32) which can be oxidised to the The position of the hydrocarbon at which the proton adds can be found by comparing the absorption of the adduct with that of a model odd alternant system e.g.anthracene adds a proton at the 9-position since the spectrum of the adduct resembles closely that of the diphenyl- methyl cation314 (Table 18). TABLE 18. Carbanions and carbonium ions Ion Solvent Di- and tri-alkyl carbonium SA (33) SA (34) SA Dialkylphenyl carbonium SA _ _ . Diphenylmethyl { zy: SA ET Triphenylmethyl { :?& SA proton adduct cation anion i dianion Anthracene Perinaphthenium cation Tropylium cation Cyclo-octatetraene dianion Hep tamethylbenzenonium cation Ally1 cation Cinnamoyi cation HF SA TF TF SA SA ET SA SA SA Amax (A) 2950 3550; 2900 3500 4000 4410 4340 4410; 4340 4750; 4100 4450 7100 7140 6135 4000; 2260 2740; 2170 3400; 2750 3970; 2820 2730 4400 Emax Ref.5000 305-307~ 600; 8000 308 5000 309 10,000 310-312 42,800 3 14 22,000 313 37,800; 36,200 310 15,000; 7500 313 37,000 3 14 9600 322 9Ooo 323 27,000 323 47,000; 32,000 319 4500; 41,000 316 3000; 10,000 320 8500; 6800 317 4700 307b 4Ooo 320 Carter. McClelland. and Warhurst. Truns. Farudav SOL 1960. 56.455. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 337 In an arylmethyl ion bonding and antibonding n-orbitals are paired as in all alternant systems but with an odd number of conjugated atoms one orbital is left after the pairing and it remains a non-bonding level bisecting the array of paired levels at the centre of gravity of their energies329 (Fig. 10). In a symmetrical di- or tri-arylmethyl ion the highest bonding and the lowest antibonding levels have the same energies in the Huckel Energy 10 “on - bonding H ig hest bond i ng ArH Ar3C Ar,CH FIG.10. The relative energies of the highest bonding non-bonding and lowest anti- bonding orbitals in a di- (AraCH) and a tri-arylmethyl system (&&) derived from an aromatic hydrocarbon (Arm. approximation as the respective levels in the aromatic hydrocarbon from which the ion is schematically derived by the substitution of the ap- propriately charged methyl Within the limitations imposed by the neglect of ion-pairing of solvent effects of steric hindrance to conjuga- tion in the arylmethyl ions and of the contribution of electron repulsion changes to the transition energy it is to be expected that corresponding di- and tri-arylmethyl ions independently of the position of substitution of the methyl group in the aromatic nucleus and of whether the ion is positively or negatively charged should absorb at a wavelength approxi- mately twice that of the aromatic hydrocarbon from which the ion is derived.330 The longest wavelength band of biphenyl lieslO at 2500 A and it is found330 that the visible absorption bands of the tris-p-biphenylyl- tris-m-biphenylyl- and bis-p-biphenylyl-methyl carbonium ions lie at 5100 5000 and 5150 A respectively though the corresponding anions absorb313 at 5780,4850 and 5900 A respectively.The di- and tri-phenylmethyl ions absorb at wavelengths not greatly more than double that of the p-band of benzene or what is equivalent that of the centre of gravity of the 01- p- and /3-band system (Tables 13 and 18) but there are added uncertainties in this comparision.The benzene p-band is generally ascribed to a forbidden transition but the lowest energy transitions of the di- and tri-phenylmethyl ions are allowed and a38 Longuet-Wggins J. Chem. Phys. 1950 18 265. aao Grinter and Mason in “Steric Eff- in Conjugated Systems,” ed. Gray Buttcr- worths London 1958 p. 52. 338 QUARTERLY REVIEWS degenerate threefold in the former and fivefold in the latter ion. Inter- action between the accidentally degenerate configurations arising from these transitions is probably responsible for the observed splitting of the triphenylmethyl-ion absorptions (Table 18) and it may give first excited states with energies lower than is indicated by the simple theory. The mono- di- and tri-phenylmethyl carbonium ions together with the perinaphthenium cinnamoyl and the pentadienyl (benzenonium) cations have the same transition energies in the Huckel theory one unit of the empirical resonance energy integral 18.Although factors other than the energies of the highest occupied and the lowest unoccupied n-orbitals must contribute to the transition energies it is observed that these ions absorb in the region of 4000 8 (Table 18) a wavelength approximately 4 2 times as large as that of the ally1 cation absorption and double that of the p-band absorption of benzene the Huckel transition energies of the latter species being 42/3 and 2/? respectively. Bands are observed in the spectrum of the tropylium ion at 2740 and 2170 8, which on intensity grounds can be compared respectively with thep- and /3-bands of benzene.The cyclic free electron model applies to the tropylium ion as to benzene when the quantum number j is unity giving twofold degenerate pairs of forbidden and allowed transitions which involve quantum number changes of three and one respectively. The upper states of the forbidden transitions are cut by three trigonally disposed nodal planes perpendicular to the plane of the ion but in contrast to the case of benzene these planes cannot bisect bonds or cut through atomic positions alone in each component owing to the hepta- gonal arrangement of the nuclei in the tropylium structure. The nodal planes in the upper states of the forbidden transitions cut the tropylium ion in equivalent ways and so the transitions remain degenerate when the alternating potential due to the atomic nuclei is introduced into the free electron model.The allowed transitions also remain degenerate and only a moderate- and a high-intensity absorption band at longer and shorter wavelengths respectively are expected in the near ultraviolet spectrum of the. tropylium ion. A self-consistent field calculation331 gives wave- lengths of 2865 and 1960 8 for these two bands. Similar considerations apply in the case of the cyclo-octatetraene dianion for which the ring quantum number has the value j = 2. The moderate- and high-intensity absorption bands are expected to appear as is observed (Table IS) at longer wavelengths than the corresponding bands of the tropylium ion a self-consistent field treatment331 placing them at 3380 and 2140 A. The cyclopropenyl cation for which j = 0 should have an absorption maximum in the vacuum ultraviolet region as it has a Hiickel transition energy of 3p a value some one and a half times as large as that of the centre of gravity of the benzene a- p- p-band system.831 Murrell and Longuet-Higgins J. Chem. Phys. 1955 23 2347 Longuet-Higgins and McEwan ibid. 1957 26 719. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 339 The di-n-propylcyclopropenyl cation gives321 only rising end absorption at 2000 A. Secondary and tertiary aliphatic alcohols halides and alkenes in strong acid solution give an absorption band near 2950 A which is attributed to the corresponding secondary or tertiary carbonium ion.305-307 A similar band is given by alkanes in strong oxidising Secondary and tertiary aliphatic carbonium ions are relatively stable entities in the gas and it is probable that the recorded spectra refer to these species in solution.Aliphatic carbonium ions are electron-deficient systems in which strong isovalent hyperconjugation should obtain so that all of the valency orbitals are utilised by the available valency e l e c t r ~ n s . ~ ~ ~ ~ ~ ~ The isopropyl and t-butyl cations can be regarded as quasi-7r-electron systems in which two and three methyl groups respectively are conjugated with a central positively charged carbon atom,336 the n-electron levels antisymmetric with respect to the plane of the ions being analogous to those of the di- and tri-vinylmethyl system respectively. Owing to the symmetries of the ions the highest occupied quasi-7r- electron levels in the isopropyl and t-butyl cations have the same energies to a good approximation and the lowest unoccupied levels do not differ greatly in energy being analogous to the non-bonding levels of the di- and the tri-vinylmethyl system respectively.336 Corresponding di- and tri-alkyl carbonium ions like the di- and tri-arylmethyl ions should thus absorb at similar wavelengths as is observed (Table 18).An electron in the highest occupied quasi-orbital of the isopropyl or t-butyl cation is confined to the methyl groups but in the lowest unoccupied level a substantial portion of the electronic charge is found on the central carbon atom so that the transitions responsible for the absorp- tion of these ions have a charge-transfer character. The transitions giving rise to the absorption of the olefinic carbonium ions (33) and (34) probably have a similar character these ions absorbing at longer wavelengths than the alkyl carbonium ions as the vinyl group is a better electron donor than the methyl group.The spectrum of the ion (33) shows two bands one corresponding in position to the olefin carbonium ion absorption of the norbornadienyl cation (34) and the other to the di- and tri-alkyl carbonium ion absorption (Table 18) sug- gesting the concurrence of both the olefin and the alkyl charge-transfer transitions in this case. A (34) 13. Linear and Cyclic Conjugated Systems The light-absorption properties of linear and cyclic conjugated com- pounds are of special spectroscopic interest as both the classical and the 332 Lavrushin Kursanov and Setkina Doklady Akad. Nauk S.S.S.R. 1954,97,265. 333 Franklin and Lumpkin J.Chem. Phys. 1952,20 745. 334 Losstng Ann. N . Y. Acad. Sci. 1957 67 499. 336 Mulldcen Tetrahedron 1959 5 253. 336 Muller and Mulliken J. Arner. Chem. Soc. 1958 80 3489. 4- 340 QUARTERLY REVIEWS quantum theories can be applied in a particularly simple and general form to account for the spectral variations in a series of molecules differing successively by one conjugated unit. Recently there have been notable experimental advances in this fie1d.337-348 The light absorption of all- trans-dimethylpolyenes (35) and of di-t-butylpolyacetylenes (36) with up Me* [CH=CH] Me Me3C.[C=C];CMe3 (p=;:6 (3) (36) Me Me Me (37) Me O=CH*[CH=CH] SO- Me,N-[CH=CH] *CH=O Me;N”[CH=W] CH=6Me (38) (39) (400) MelN=CH* [CH=CH];NMe (40b) to ten conjugated units have been measured (Table 19) together with the spectra of compounds (37) with up to seven cumulative double bonds (Table 20).The simple hetero-atom series of the oxonols (38) the mero- cyanines (39) and the cyanines (40) have been spectral data now being recorded for compounds with up to six ethylene groups (Tables 20 and 21). In addition the absorption of two new types of closed chain chromophore has been studied the large cyclic polyenes (Table 22) and the cyclic phosphonitrilic All interpretations of the spectra of polyenes whether wave mechanical or classical have required that there should be an alternation of double and single bonds down the polyene chain. Early molecular orbital treatments of the ground-state energies of the p o l y e n e ~ ~ ~ ~ ~ ~ ~ indicated that the alternation should progressively diminish the longer the polyene in apparent contradiction to the spectroscopic interpretations.More recent 337 Nayler and Whiting J. 1955 3037. 338 Bohlmann and Mannhardt Ber. 1956 89 1307; Bohlmann ibid. 1952 85 386. 339 Bohlmann Ber. 1953 86 63 657. 340 Jones Lee and Whiting J. 1960 3483. 341 Bohlmann and Kieslich Ber. 1954 87 1363. 342 Malhotra and Whiting J. 1960 3812. 343 Gaoni and Sondheimer J. Amer. Chem. SOC. 1960 82 5765. 344 Sondheimer and Wolovsky Tetrahedron Letters 1959 No. 3 3. 346 Sondheimer and Wolovsky J. Amer. Chem. SOC. 1959 81,4755. 346 Sondheimer and Wolovsky J. Amer. Chem. SOC. 1960 82 754. 347 Lund Paddock Proctor and Searle J. 1960 2542. 348 Foster Mayor Warsop and Walsh Chem. and Znd. 1960 1445. s49 Lennard-Jones Proc. Roy. Soc. 1937 A 158 280. 860 Coulson Proc.Roy. SOC. 1938 A 164,383; 1939 A 169,413. n 2 3 4 5 6 7 8 9 10 TABLE 19. Polyenes and poly-ynes [band origin wavelengths and integrated intensities (oscillator strengths f) above 2000 A] Bd.[C-C ];But ME 2525 0.288 ME 3090 weak 21 30 140 ME 3600 0.100 2395 348 ME 3940 0.156 2650 442 ME 4300 0.103 2885 500 ET 4530 0.290 3105 527 HX 4755 0.150 3295 705 Solvent Amax (A) Emax x 10-3 HX 5130 0.100 3625 850 fexp- Ref. 339 339 2.08 339 3.13 339 4-11 339 5.14 339 339 12-0* 340 14.3* 340 Solvent ME ME CL ME CL CL ME CL CL ET B ET B * Includes strong absorption due to transitions below 2500 %i other than '&+. t Value for dotriacontadecaene. Me- [CH = CH],.Me (all-trans) Amax(& fexp fcalc. Ref. 2260 0.32 0-67 338 2740 0.98 0.94 338 2790 0.74 337 3140 1.1 1 1.21 338 3165 1.11 337 3430 1-37 1 -47 338 3490 1.66 337 3700 (1 *78) 1 *74 338 3800 2.17 337 4010 2.00 337 4200 2.52 2.28 4315 338 4405 2.55 338 4760t 2.80 338 342 QUARTERLY REVIEWS studies of the balance between 0 and T bond energies in the ground state of the have shown that all linear polyenes should possess an alternating bond character and that the same alternation should obtain TABLE 20.Long-wavelength absorption maxima (A) in the series of n 1 2 3 4 5 6 Ref. compounh (37),-(38) (39) and (41) (37) (38) (39) 2715 2675 2830 3390 3625 3615 4005 4550 4215 4650 5475 4625 (6440) 491 5 5125 341 342 342 TABLE 21. arw-Bisdimethylaminopolyene ions Me$N=CH-[CH=CH],*NMe,342 1 3125 645 2275 10 5 2 3 4160 1195 2540 22 2285 26 6 5190 2070 3090 28 2665 28 2350 24 ~ m a x (4 6250 3635 3050 241 5 2275 7345 41 80 3445 2720 2470 8480 4700 4365 3845 3165 301 5 271 5 TABLE 22.Cyclic polyenes < [-CH=CH.],> n Solvent Amax (A) %i3X 4 CH 2900 250 7 I 0 3740; 3140 5700; 69,000 (41) 2520 2800 3000 3100 3180 359 10-2Emax 2950 48 47 48 49 3530 73 70 60 78 (2200) (1 29) (78) (102) (88) (87) (98) Ref. 356 343 9 I 0 4480;4080;3690 21,800; 7500; 30,300 344 12 I 0 5120; 3630; 3500 1740; 201,OOO; 195,000 345 15 D 4280; 3290 144,000; 44,Ooo 346 361 Labhart J. Chem. Phys. 1957 27 957 963. 363 Tsuji Huzinaga and Hasino Rev. Mod. Phys. 1960 32 425. 363 Ooshika J. Phys. SOC. Japan 1957 12 1238 1246; 1959 14 747. SM Longuet-Higgins and Salem Pvoc. Roy. Soc. 1959 A 251 172; 1960 A 257,445. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 343 in the larger cyclic polyenes with an odd number of double bonds begin- ning perhaps with the compound C18H18.354 Extensions of the treatment indicate the probability of bond alternation in the longer cyanines (40) with twenty352 or thirty351 double bonds.Owing to the Jahn-Teller bond alternation may well begin earlier in the cyanine series the fixed bond structures (40a) and (40b) representing stable configurations separated by a potential barrier when the zero-order electronic transition energy becomes comparable to that of a vibrational transition. The cyclic polyenes with an even number of double bonds also suffer from a Jahn-Teller distortion the dianion of cyclo-octatetraene being planar and quasi- for example whilst the neutral hydrocarbon is non-planar with alternant single and double bonds. For practical purposes the known series of linearly conjugated com- pounds can be divided into two main types those with bond alternation e.g.( 3 9 and those without e.g. (40) the two types having different light-absorption properties. Hausser and his c o - w ~ r k e r s ~ ~ ~ showed that the wavelength maximum of the first absorption band in various polyenic series was approximately proportional to the square root of the chain length and subsequently Karrer and E ~ g s t e r ~ ~ ~ with a more extended series of carotenoid polyenes found that the wavelength maximum tended to converge to a limit. Such a convergence was earlier apparent in the of the polyphenyls (41). In contrast the first absorption wave- length of the symmetrical odd-atom dyes e.g. (40) increases by -1000 A for each additional ethylene group successive increases being non- In both types of series the intensity of absorption is approximately proportional to the chain length.357,360 These regularities have been accommodated by a number of classical vibration theory treatments.Lewis and Calvin5 accounted for the variation of absorption wavelength in the polyene series on the rather unsatisfactory assumption that all of the absorption electrons behaved as a single harmonic oscillator subject to a constant restoring force. Subsequent classical treatments were based on the more satisfactory premise that the conjugated units in a polyene or polyphenyl were interacting oscillators. W. Kuhn using vibration coupling theory the expression A2 = X$/(l - a cos[m/(n + l)]) . . . . . where A is the absorption wavelength of the individual unit ethylene in the polyene and benzene in the polyphenyl series n is the number of 356 Katz J.Amer. Chem. SOC. 1960 82 3784. 366 Cope and Bailey J. Amer. Chem. SOC. 1948 70 2305. 367 Hausser Kuhn Smakula and Kreuchen 2. phys. Chem. 1935 B 29 363; Hausser Kuhn Smakula and Hoffer 2. phys. Chem. 1935 B 29 371 378; Hausser Kuhn and Smakula ibid. p. 384. 358 Karrer and Eugster Helv. Chim. Ada 1951 34 1805. 360 Gillam and Hey J. 1939 1170. 360 Brooker Rev. Mod. Phys. 1942 14 275. asl Kuhn Helv. Chim. Acta 1948 31 1780. 344 QUARTERLY REVIEWS conjugated units a is a constant dependent upon the strength of the coupling and s is the order of the harmonic of vibration i.e. s = 1 for the fundamental s = 2 for the first overtone etc. Kuhn also obtained an expression for the oscillator strength f of absorption in the polyene series .. . . . . . . f = Snfo/n2s2 (9) where fo is the oscillator strength of ethylene. Equation (9) holds only for odd values of s; for even values f = 0. Other classical wavelength f ~ r m u l a e ~ ~ ~ ~ ~ ~ ~ based on the assumption of a power series interaction between the unit oscillators have the form . . . . . . . . X2 = A - BC“ (10) where A B and C are constants for a given series. Equations (8) and (10) both account for the observed convergence of the polyenic and polyphenyl series equation (8) being particularly successful if A is treated as a disposable parameter. Kuhn’s treatment361 also explains the relative oscillator strengths of the “fundamental” absorption bands of the polyenes though not of the “overtones”. A number of quantum mechanical treatments of the polyene and polyphenyl spectra have been based on the model that these compounds consist of weakly interacting units.Simpson has shown364 that if the n-electrons of the polyenes are localised the ethylene units interacting only by electron repulsion changes on excitation the frequency of the first absorption band Y is given by v = vo - 2/3cos[.rr/(n + l)] . . . . (1 1) where vo is the absorption frequency of a single unit i.e. ethylene and /I is the interaction energy between the n units. A formula of the same form as equation (1 1) is given if partial electron delocalisation is assumed in the ground and the excited states. The absorption frequencies of the p-polyphenyls (Table 20) give notably good agreement with equation (1 l) namely Y (cm.-l) = 50,600 - 21,200 cos[.rr/(n + l)] .. . (12) suggesting that the polyphenyl long-wavelength absorption derives from the benzene p-band at 50,000 cm.-l. More detailed t r e a t r n e n t ~ ~ ~ ~ indicate that the polyphenyl absorption originates from both the p- and the /I-band transitions of the individual benzene rings these transitions 362 Hirayam-a J. Amer. Chem. SOC. 1955 77 373. m3 Panouse Bull. SOC. chim. France 1956 1568. s64 Simpson J. Amer. Chem. SOC. 1951,73 5363; 1955,77 6164. sg6 Sirnpson J. Amer. Chem. SOC. 1953 75 597. Murrell and Longuet-Higgins J . 1955 2552. 367 Murrell and Longuet-Higgins Proc. Phi’s. Soc. 1955 68 A 601. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 345 being mixed by electron-repulsion interactions between the upper states. The first absorption frequencies of the polyenes do not fit equation (1 1) closely probably because the interaction between the ethylene units is not weak as is shown by the relatively slow convergence of the polyene absorption.The absorption frequencies of cyclic systems composed of quasi- independent units interacting weakly on excitation are given by v = v o - 2 p . . . . . . (1 3) Thus cyclic systems whatever the ring size should absorb at the con- vergence limit of the corresponding linear series. The spectra of the cyclic polyenes (Table 22) do not support equation (13) but equally they are not in accord with the cyclic polyene model used to interpret the spectra of cata-condensed aromatic hydrocarbons. The latter model in which complete delocalisation of the n-electrons is assumed requires that a cyclic polyene with an odd number of double bonds should have an aromatic spectrum giving weak moderate and strong absorption at progressively shorter wavelengths.Specifically cyclo-octadecanonene C18Hl8 should show weak and strong absorption in the red and the blue regions respectively,368 but the first absorption band is of high intensity and lies in the violet region (Table 22) suggesting that bond alternation obtains in this compound. The larger cyclic polyene C30H30 which should be ~train-free,~~~ has a first absorption of much higher intensity at even shorter wavelengths (Table 22) indicating a now more substantial bond fixation. (NPCI Jny is strikingly constant in the known series n = 3-13 in accord with equation (13) if the observed band sides at -2200 A are due to 7r -+ 7r* or to chlorine lone-pair n -+ n* transitions.The absorption of the trimer is unchanged in concentrated sulphuric indicating that nitrogen lone-pair n -+ n* or n -+ (T* transitions are not active. However chlorine lone-pair n -+ a* transitions may be responsible for the absorption and this possibility would account equally for the observed lack of variation of absorption with ring size. The quantum interpretation of the spectra of the main linearly conju- gated series requires the use of models in which substantial interaction between the individual units is assumed. The models are of two types based on the valence bond and the molecular orbital procedures the formzr being particularly useful for the interpretation of the spectra of the un- symmetrical odd-atom dyes and the latter for the treatment of the sym- metrical dye and polyene spectra.A comprehensive review of the spectra of dyes and polyenes based mainly on the valence bond mzthod has been given recently by Platt.369 The absorption of the cyclic phosphonitrilic Gouterman and Wagniere Tetrahedron Letters 1960 No. 11 22. 86g Platt J. Chem. Phys. 1956 25 80. 346 QUARTERLY REVIEWS In a simple orbital treatment Dewar has shown that if two odd alternant radicals are joined to form an even alternant polyene the non-bonding molecular orbitals of the radicals are transformed into the highest occupied and the lowest unoccupied n-orbitals of the p01yene.~~* The energy separa- tion between those orbitals is given by the coefficients of the radical non- bonding molecular orbitals at the joined atoms and an electronic transition between the polyene orbitals bond alternation being assumed requires the absorption of light with a wavelength = D [1 - a(2n + 2)] .. . . . (14) where n is the number of double bonds in the polyene and D the con- vergence limit and a the bond alternation constant are adjustable parameters. The non-bonding molecular orbital method with bond alternation has been extended to account for the spectra of the phenyl- and the diphenyl-polyenes and of the m- andp-polyphenyls the theoretical absorption wavelengths showing generally good agreement with experi- ment.370 Bond alternation is not required to account for the p-band absorption of aromatic hydrocarbons by the same method,371 supporting a recent of the ground states of the polyacenes in which bond fixation was found to be absent.The most general of the simple orbital treatments is the free-electron m 0 d e 1 ~ ~ ~ - ~ ~ ~ in which it is assumed as a first approximation that the r-electrons of a conjugated system move in a container of uniform potential with dimensions determined by the size of the molecule bounded by infinitely high potential walls. In a series of conjugated molecules the cross-sectional dimensions are constant and they are small compared with the length. The quantisation of the lower electronic energy levels is determined by the length of the conjugated system L and they have the energies . . . . . . . . E = h2j2/8mL2 (15) An electronic transition between the levels with quantum numbers j and ( j + l) requires the absorption of light with a frequency .. . . . . Y = h (2j + 1)/8rnL2 (16) The potential in a conjugated molecule without bond alternation has minima at the atomic positions and maxima at the bond centres and it has 370 Dewar J. 1952 3544. 371 Dewar J. 1952 3532. 378 Longuet-Higgins and Salem Proc. Roy. SOC. 1960 A 265,435. 573 Bayliss J. Chem. Phys. 1948 16 287. 374 Baylise Quart. Reviews 1952 6 319. 376 Kuhn J. Chern. Phys. 1948,16,840 1949 17 1198. 376 Kuhn “Progress in the Chemistry of Organic Natural Products,” Vol. 16 ed. Zechmeister Springer Verlag Vienna 1958 ; Bar Huber Handschig Martin and Kuhn J. Chem. Phys. 1960,32,470. 377 Simpson J. Chern. Phys. 1948,16,1124; 1949,17 1218. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 347 the effect in a system of N conjugated atoms of separating the electronic energy levels into bands each of N levels (Fig.11). The energy intervals between the levels in the lowest band are reduced but the levels retain the same relative energies so that equation (16) remains roughly valid for the main transition of interest namely for the case o f j = N/2. FIG. 11. The energy levels of a six rr-electron six conjugated atom system according to (a) the simple free-electron model (b) the model with atom-bond potential alternation and (c) the same with bond-bondpotential alternation. In a conjugated molecule with alternating single and double bonds there is an additional potential variation of longer period and an energy gap appears between the lowest band of the N/2 occupied levels and the next band consisting of N/2 empty levels (Fig. 11). An electronic transition for the case.j = N/2 then requires a larger energy and this energy is finite for the infinitely long conjugated system.Allowing for the potential variation due to the alternation of double and single bonds H. Kuhn has that for a linearly conjugated molecule containing N T-electrons equation (16) becomes Y = vz (1 - 1/N) + h ( N + 1)/8 m L2 . . . . (17) where vz the frequency of the convergence limit of the series equals the amplitude of the bond potential alternation. The effective length of a polyene or a cyanine may be set at one or two 348 QUARTERLY REVIEWS bond distances beyond the terminal cmjugated atoms so that equation (17) gives with the insertion of the values of the fundamental constants (m.-l) = vz + (3 x 105p - v ~ ) / ~ t . . . . . (18) where r is the average bond distance in A and ",approximately equal to N is the effective number of n-electrons or the effective number of bonds increasing by two units for the addition of each ethylene or each acetylene group.In general N' differs from N by an additive constant which allows for the effects of formally non-conjugated substituents or of partial localisation of the n-electrons. Terminal alkyl groups have an effect comparable to that of an additional double bond upon the diene singlet absorption wavelength and the singlet-triplet frequency splitting (Table 23). TABLE 23. Wavelengths of corresponding singlet-singlet and singlet- triplet absorptions and their frequency separation ( A v). Band maxima are quoted for the olefins and band origins for the acetylenes. The allowed singlet origin of thepolyacetylenes is taken to be the 'L' band R H Me Me Et n 2 2 3 4 Singlet Triplet 73 N& R.[CH= CH];R 1 620 2700 21 70 3850 2385 4000 2565 4350 2675 4800 R. [C_C];R 1630 3700 1650 3580 2070 45 10 2380 5320 Av (cm-l) 25,000 20,000 17,000 16,000 17,000 34,300 32,700 26,000 23,200 Ref. (singlet) 70 82 84 85 48 1 482 482 98 48 3 Equation (18) specifies that (he frequencies of the first absorption bands in the spectra of both the cyanine and the polyene type of conjugated series should be linearly related to I/" the former converging to zero frequency and the latter to a finite frequency the slopes of the relation- ships being of the order of lo5 cm.-l. Such relationships are found (Fig. 12) and the slopes give physically reasonable average bond length values r for the cases of the cumulenes and the ld band of the polyacetylenes and for the cyanines oxonols and the polyenes if that length represents an average unit of the rectilinear distance between the terminal conjugated atoms separated by an all-trans-chain (Table 24).The anomalous value for the lCU+ transitions of the polyacetylenes indicates that the agreement for the other cases is due to the partial cancellation of neglected factors. The predicted frequencies are the means of corresponding singlet and triplet absorptions but the experimental MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 349 values refer to the singlets alone. Experimental values of the corresponding triplet absorptions are available for a few members of the polyene and the polyacetylene series (Table 23). These values show that the 1Cu+-3Cu+ splittings of the polyacetylenes are particularly large and that the singlet- 10,000 20,000 30,000 40,000 U v (cm?) The relations between 1/N' (Eqn.18) and the frequency of absorption (v) in the linearly conjugated series the cyanines (40) V the cumulenes (37) 0 the polyenes ( 3 3 and x the 1 A band and A the Zw+ band of the polyacetylenes (36). FIG. 12. triplet frequency separations decrease in magnitude with a progressive increase of chain length in both series. Thus in the absence of compensat- ing factors the correlation of the experimental singlet absorption frequen- cies with If" should give a slope larger than that indicated by equation (1 8) the effect being pronounced when the singlet-triplet splitting is large. Apart from the lCU+ bands of the polyacetylenes this effect appears to be compensated in Iarge measure by the consequence of the alternating atom-bond potential along the conjugated chains.This potential neglected TABLE 24. The frequency of the convergence limit ( v J and the slope (AvfAN') of the relations (FIG. 10) based on eqn. (18) for linearly conjugated series together with the average bond Iength (r) and the eflective number of rr-electrons (N') in terms of the formal number ( N ) Series vr'cm.-l) A v/ A N' (cm.-l) r (A) N Cyanines (40) 0 190,000 1 -25 N Oxonols (38) 0 210,000 1 -20 N Cumulenes (37) 7000 180,Ooo 1-29 N + 2 Polyenes (35) 13,500 193,000 1-25 N + 2 12,000 180,000 1-29 N + 2 Polyacetylenes (36) { ygy 16,000 259,000 1.08 N + 2 Merocyanines (39) 13,000 90,Ooo 1.82 N - 2 350 QUARTERLY REVIEWS in the derivation of equation (18) crowds together the N levels of the first energy band thereby reducing the energy variation of the transition N/2 -+ (N/2 + 1) in a given series.The merocyanines follow equation (18) only if it is assumed that the effective number of n-electrons or of bonds is less than that indicated by the formulae (39) i.e. N' = (N - 2) (Table 24). Dipole-moment measurements that there is little charge-transfer from the amine to the carbonyl group in the ground state of the longer member of this series suggesting that the lone-pair electrons of the amino-group become progressively localised upon the nitrogen atom. From spectroscopic data it appears probable that the localisation persists to some degree in the lowest excited state and that the lowest energy transitions arise increasingly from the polyene moiety of the merocyanines as the chain is lengthened for the merocyanine absorption frequencies converge to the polyene limit (Table 24).The oxonols and the cyanines also require in equation (18) a smaller effective number of n-electrons than the polyenes (Table 24) owing to the partial localisation of the r-electrons upon the terminal hetero-atoms. In the limit of infinitely electronegative terminal atoms the cyanines would be hybrids of structures (42a) and (42b) with a shorter and an electron-depleted n-electron system. The absorption frequencies of the cumulene hydrocarbons (37) appear to converge to a low but finite limit although a uniformity of bonding along the conjugated chain might be expected in these compounds. The position of the limit is not too reliable as data are available for only four members of the cumulene series,341 but its non-zero value may indicate a small bond-bond alternation in these compounds arising from the circumstance that there are N/2 double bonds in the rz chain and (N/2 - 1) in the rY chain of a N-atom cumulene if N is even.The free-electron model accounts not only for the frequencies but also for the intensities of the main absorption bands in the spectra of linearly conjugated systems. From equations (4) and ( 5 ) the oscillator strength of absorption in either the cyanine or the polyene type of series is given by347 f = 0*134(N + 1)/3 . where s the difference between the values of the quantum numbers in the lower and the upper levels of the transition is an odd integer and N is the number of n-electrons.For the lowest-frequency absorption band s is unity and for even values of s the oscillator strength is zero. The calculated oscillator strengths for the polyene series are in satisfac- tory agreement with those measured experimentally (Table 19) and the maximum extinction coefficient of the main absorption band in the oxonol and cyanine series is proportional to the number of n-electrons in the molecule (Table 21) as equation (19) requires given a band half width which is a constant or linearly varying in the series. The merocyanine 378 Hutchinson and Sutton J. 1958 4382. MASON ; MOLECULAR ELECTRONIC ABSORPTION SPECTRA 35 1 absorption intensities however show no such proportionality to N. The intensities of the forbidden ld bands remain constant in the poly- acetylene series but that of the allowed lCu+ band increases linearly with the number of acetylene residues (Table 19).The oscillator strengths to which the transition is virtually the sole contributing excitation in the members below the octa-acetylene are more than twice as large as those of the corresponding polyenes (Table 19) an effect ascribed339 to the greater overall distance between the terminal conjugated atoms in the polyacetylenes. The cyanine and the polyene series give absorption bands of markedly different shape the cyanine bands being smooth with a half width of -1000 cm.-l whilst the polyene and polyacetylene absorption consists of a progression with three to five members in an upper state carbon-carbon stretching vibration covering -5000 and -8000 cm.-l r e s p e c t i ~ e l y .~ ~ ~ - ~ In the polyenes and the polyacetylenes the formally single and double bonds assume respectively a larger and a smaller double-bond character on excitation and equilibrium in the excited state is achieved only after a substantial readjustment of the nuclear positions. The ground state con- figuration corresponds to the turning points of'the higher vibrational levels in the upper state so that a strong vibrational progression appears in the electronic absorption. The structures (40a) and (40b) contribute equally to both the ground and the excited state of the cyanines and in each state all the bonds have the same double-bond character. On excitation there are no gross changes in the nuclear positions and absorption occurs pre- dominantly at the band origin frequency giving a band of small half- In addition to the main absorption band in the spectra of the cyanines and the polyenes a number of weaker bands appear at shorter wavelength.The weaker bands are ascribed to overtone electron vibrations in the classical theory,3s1 and to electronic transitions requiring quantum number changes greater than unity in the quantum treatments. For a quantum number change of s units the resultant absorption frequencies are according to the free-electron model v - vg = Cs"' . . . . . . . . (20) where C is a constant given by equation (18) for the case of s = 1. From equation (20) the frequency separation between successive electronic bands in the spectrum of a given molecule should equal the frequency interval between the main band and the convergence limit and the constant C should have the appropriate value (Table 24) for the series to which the molecule belongs.These relations are supported by the data for the higher polyenes (Table 25) but they are at variance with the cyanine data (Table 21). Equation (20) and the experimental data (Table 25) suggest that the first subsidiary band in the spectra of the higher polyenes arises from a 352 QUARTERLY REVIEWS transition between levels with quantum numbers differing by two. All transitions for which s is an even number are forbidden in the zero-order free-electron treatment of the all-trans-polyenes. However transitions in linearly conjugated systems involving a change in quantum number of s units are s-fold degenerate and the transition N/2 -+ (N/2 + 2) has the same energy or nearly so as the transition fN/2 - 1) -+ (N/2 + 1).TABLE 25. The wavelengths (A) relative intensities (Erel) and frequency separation from the convergence limit (v - v J of the main and subsidiary bands in the spectra of the higher all-trans-polyenes Me. [CH =CH],*Me n 8 9 10 A(A)33* 2395 2960 4200 3130 4405 3410 4760 3 1 14 1 7.4 1 8 v - vz(cm.- 28,300 20,200 10,200 18,400 9200 15,800 7500 The symmetric and antisymmetric combinations of the configurations arising from these zero-order excitations give upper states to which transi- tions from the ground state are allowed and forbidden respectively. The forbidden upper state has a nodal pattern with the same symmetry as the alternating single-double bond charge-density distribution of the zero- order ground state and interaction between them gives first-order ground and forbidden upper states of lower and higher energy (Fig.13). Other forbidden excited states with the symmetry of the single- double bond alternation of the zero-order ground state have the same effect enhancing the potential alternation in the improved ground state. The allowed upper states with different symmetries are unaffected and they have energies given approximately by the simple free-electron treatment accounting for the agreement between equation (20) and the experimental data for the polyenes (Table 25). In the cyanine series with a zero-frequency convergence limit the frequency of the main absorption band is always larger than the frequency interval between the main and the first subsidiary band though they con- verge the latter falling progressively from 15,300 to 9500 cm.-l and the former from 24,000 to 1 1,800 cm.-l between n = 2 and n = 6 respectively in the series (40).These differences and trends in the cyanine series are due in part to the large electronegativity of the terminal nitrogen atoms which adds to the zero-order band frequencies given by equation (20) a first- order Coulombic increment that is larger for the main than for the first subsidiary band both increments progressively decreasing as the chain length is extended. In part equation (20) is not in accord with the observed MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 353 spectra of the cyanines owing to interactions between configurations of the same symmetry. For the series (40) n = 0-3 it has been that with configuration interaction both transitions for which s = 2 are weakly allowed absorbing at different frequencies.The three transitions for which E t -- / s=2 / s=2 -- s= I 5= I G FIG. 13. The relative energies of the ground (c) and the excited states arising from quantum number changes (s) of 1 2 and 3 in the linear polyenes according to (a) the simple free-electron model and (b) the model with bond alternation and interaction between conJ5gurations with the same symmetry. The full and the broken lines refer to states to which transitions from the ground state are allowed and forbidden respectively. s = 3 behave similarly forming a complex set of weakly allowed higher- energy excited states which help to account for the small and apparently irregular frequency intervals between successive subsidiary bands in the spectra of the cyanines (Table 21).14. Charge-transfer Spectra All electronic transitions result in the redistribution of charge and awide variety of excitations involve the transfer of a substantial portion of an 379 McGlynn and Simpson J. Chem. Phys. 1958,28,297. 354 QUARTERLY REVIEWS electronic charge from a donor to an acceptor g r o ~ p . ~ O - ~ ~ Intramolecular charge-transfer absorptions described above e.g. those of the unsaturated ketones have their intermolecular analogues e.g. those of the aromatic hydrocarbon-carbonyl complexes and they do not differ in essentials. In both cases it appears to be necessary and sufficient for charge-transfer absorption that the donor shall have a high-energy filled orbital and the acceptor a low-energy vacant orbital and that these orbitals shall over- lap.381 The overlap may be sterically determined e.g.(3) or derived either from complex formation or from chance collisions between the com- p o n e n t ~ . ~ ~ ~ There are three ionically distinct types of acceptor-donor pairs with charge-transfer absorptions characterised by different environmental effects. Electron transfer between the components of an ion pair e.g. a halide ion and either a metallic or organic cation gives an excited state less polar than the ground state and large blue shifts of the charge-transfer absorption result on changing from a less to a more polar environment. A shift from 4489 A in chloroform to 3311 A in 7:3-ethanol-water is for the charge-transfer band of 4-methoxycarbonylpyridine ethiodide whilst sodium iodide absorbs at 3250 A in the vapouflgl and at 2200 A in the representing transition energy increases of 21 and 42 kcal./mole respectively.The changes of the charge-transfer transition energy with solvent polarity in the case of the pyridinium iodides parallel closely the solvent shifts of n -+ T* absorptions and the variations in the free energy of activation with solvent of reactions involving a redistribu- tion of charge.390 Electron transfer between the components of a neutral complex e.g. an unsaturated hydrocarbon and a nitro-compound gives an excited state more polar than the ground state and reverse solvent effects are observed. Such effects are usually small since the solvent molecules require a finite time to reorientate round the dipolar excited state. The lowering of the energy of the excited state by solvation results in a frequency separation between the charge-transfer absorption and fluorescence bands of complexes derived from neutral components e.g.hexamethylbenzene and tetrachlorophthalic anhydride which increases with the dielectric constant and refractivity of the 380 Rabinowitch Rev. Mod. Phys. 1942 14 112. 381 Mulliken J. Amer. Chem. SOC. 1950,72 600,4493; 1952 74 811. 382 Mulliken J. Phys. Chem. 1952 56 801. 383 Mulliken J. Chem. Phys. 1951 19 514; 1955 23 397. 384 Mulliken Rec. Trav. chim. 1956 75 845. 386 Orgel Quart. Reviews 1954 8 422. 386 Orgel and Mulliken J . Amer. Chem. SOC. 1957 79 4839. 387 Murrell J. Amer. Chem. SOC. 1959 81 5037. 388 McGlynn Chem. Reviews 1958 58 11 13. 389 Briegleb and Czekalla Angew. Chem.1960,72 401. 390 Kosower J. Amer. Chem. SOC. 1958 80 3253 3261 3267. 3v1 Franck Kuhn and Rollefson Z. Physik 1927 43 155. 392 Przibram Z. Physik 1923 20 196. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 355 Charge transfer between an ionic and a neutral component e.g. the transition-metal ions complexed with neutral ligands or the smaller non- metallic anions in polar solvents gives absorption bands upon which the effect of change of solvent is variable in magnitude and direction. The charge-transfer-to-solvent absorption of the iodide ion undergoes small blue shifts in hydroxylic solvents but larger red shifts in non-hydroxylic solvents as the polarity of the medium is reduced.393 Anion-to-solvent charge-transfer absorption occurs at wavelengths which appear to depend in part at least upon the electron affinity of the solvent molecules and in solvents of low polarity upon specific cation The firit charge-transfer bands to be characterised were those of the halide ions.4oo Alkali-metal bromides and iodides in solution or in the vapour or solid state give two absorption bands of high intensity separated approximately by the 2P3,2 - 2P1/2 energy interval of the corresponding halogen atom.385,3ge A similar interval is observed between the two longer- wavelength absorption bands of the tri-iodide and of the pyridinium iodides.403 The upper states of these transitions appear to consist of a halide atom and an electron which has been transferred respectively to a cavity bounded by solvent mo1ecules,3e3-396~401 to an orbital of the alkali- metal ion or ions,397 to an iodine molecule and to the lowest unoccupied pyridinium n - ~ r b i t a l .~ ~ ~ ~ ~ ~ ~ The halide atoms and anions form the transient dihalide ion the spectrum of this species being detected in flash-photolysis experiments,405 and the ultimate overall products are the trihalide ion and Many non-metallic inorganic ions in solution give intense absorption bands in the ultraviolet region (Table 26) but only the absorption of the monatomic or quasi-monatomic ions e.g. hydroxide can be ascribed with certainty to anion-solvent charge-transfer transitions. The intense absorption of the polyatomic ions is less sensitive to environmental effects,395 and it is accompanied by weak absorption at longer wavelength 39s Smith and Symons Discuss. Faraday SOC. 1957 24 206; Trans. Furaday SOC.394 Griffiths and Symons Mol. Physics 1960 3 90; Trans. Faraday SOC. 1960 56 1958 54 338 346. 1125. Stein and Treinin Trans. Furaday SOC. 1959,55 1086 1091. 396 Jortner Raz and Stein Trans. Faraday SOC. 1960 56 1274. a97 Gourary and Adrian Solid State Phys. 1960 10 127. 8g8 Kosower Martin and Meloche J. Chem. Phys. 1957,26 1353. 399 R m e n s Handbook of the 4th Meeting on Molecular Spectroscopy Bologna 400 Franck and Scheibe 2. ghys. Chem. 1928 A 139 22. 401 Platnan and Franck 2. Physik 1954 138 41 1. 402 Awtry and Connick J. Amer. Chem. SOC. 1951 73 1842. 4oa Kosower Skorcz Schwan and Patton J. Amer. Chem. SOC. 1960 82 2188. 404 Mason J. 1960,2437. 406 Edgecornbe and Nomsh Proc. Roy. SOC. 1959 A 253 154. 406 Rigg and Weiss J. 1952,4198; J. Chem. Phys. 1952,20 1194. 1959. 356 QUARTERLY REVIEWS (Table 26) of the It -+ u* or n + n* type.407-411 The latter bands indicate the presence of low-energy u* and n* orbitals which would serve as the upper levels of strongly absorbing u + u* and n -+ n* transitions in the polyatomic ions.The 2100 A absorption of the nitrite ion has a transition moment in the plane of the supporting a n -+ n* but not a charge- transfer assignment as the latter would require a moment perpendicular to the plane. TABLE 26. The absorption of non-meta€lic anions in aqueous or alcoholic solution Ion c1- Br- I- Br3- J3- ClO- c10,- BrO- OH- SH- S203'- s,0,2- NO2- NO3- N2022- Amax (A) 1810 1995; 1900 2260; 1940 2700 3530; 2875 2920; 1900 2600; 1900 3330; 1900 1870; 2300 2200 2540 3570; 2980;2110 3025; 1936 2480 Emax -104 11,000; 12,000 12,600; 12,600 11,500 26,400; 40,OOO 330; strong 150; strong 170; strong 5000 8000 4000 22 23; 8; 6O00 7; 8800 4Ooo Ref.41 3 41 3 413 414 402 81 82 81 412 41 5 402 41 7 412 412 416 The visible and ultraviolet spectra of organic complexes have been extensively investigated following the suggestion of M~lliken~~l that the light absorption of these complexes is due to photochemical charge transfer. Wavelength maxima and apparent extinction coefficients and dissociation constants have been obtained for a number of series of donors mainly aromatic hydrocarbons but also olefins and saturated compounds with a variety of acceptors notably the halogens and carbonyl cyano- and nitro-compounds.ass~389 The results provide good evidence for Mulliken's theory particularly for its later form^.^^^^^^^ To a good approximation the frequency of a photon V required to 407 Trawick and Eberhardt J.Chem. Phys. 1954,22 1462. 408 McGlynn and Kasha J. Chem. Phys. 1956,24,481. 40s Sidman J. Amer. Chem. SOC. 1957 79 2669. 410 Sayre J. Chem. Phys. 1959 31 73. 411 Friend and Lyons J. 1959 1572. 418 Friedman J. Chem. Phys. 1953,21 319. 4lS Scheibe 2. phys. Chem. 1929 B 5,355. 414 hnner J. Amer. Chem. SOC. 1952 74 5078. 416 Ellis and Golding J. 1959 127. 417 Heidt J. Chem. Phys. 1942 10,297. Addison Gamlen and Thompson J. 1952 338. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 357 transfer an electron from a donor (D) with an ionisation potential ID to an acceptor (A) with an electron affinity EA is v = ID - EA - A . . . . . . . (21) where A assumed to be constant in a series of related compounds is the difference between the binding energies of the components in the ground and the excited state.With a given acceptor the frequency of the first charge-transfer band is proportional to the ionisation potential of the donor in a number of series,418e428 as equation (21) requires. Similarly for a particular donor the frequency of the charge transfer absorption is pro- portional to the electron affinity of the acceptor in related series.4o4 However the slopes of the relationships do not always have the expected value of unity. The observed value of dv/dZ~ for the case of the acceptor iodine and a series of saturated olefinic and aromatic donors418 is 0.67 and for tetracyanoethylene with a series of methylbenzene~,~~~ the value is 0.49. Apart from uncertainties in the values of the ionisation potentials steric factors are important in these series.In complexes composed of a donor and an acceptor which are both aromatic the components lie in parallel planes.429 Bromine in its complex with benzene lies on the six- fold symmetry axis of the benzene molecule perpendicular to the aromatic ring and in the amine-iodine complexes the 3N - - - 1-1 group is linear.43o The distance between the donor and the acceptor groups is generally smaller than the sum of the van der Waals radii,429,430 and complex formation is sensitive to steric effects. Hexaethylbenzene for example invariably forms weaker complexes than he~amethylbenzene.~~~ Steric factors are not so important in the series of unsubstituted polycyclic aromatic hydrocarbons which give with a number of acceptors relations based on equation (21) with slopes near to unity.Few experi- mental ionisation potentials of the polycyclic aromatic hydrocarbons are available and the two main sets of values both measured by the electron impact method are somewhat at variance with one another (Table 27). The theoretical ionisation potentials of the hydrocarbons namely the 418 McConnell Ham and Platt J. Chem. Phys. 1953 21 66. 41D Hastings Franklin Schiller and Matsen J. Amer. Chem. SOC. 1953,75,2900. u0 Bier Rec. Trav. chim. 1956,75 866. 4a1 Peticolas J. Chem. Phys. 1957 26 429. Merrifield and Phillips J. Amer. Chem. SOC. 1958 80 2778. 423 Foster Nature 1958 181 337; 1959,183 1253; Tetrahedron 1960,10,96. a4 Booth Dainton and Ivin Trans. Faraday SOC. 1959,55 1293. 426 Bhattachary and Basu Trans.Farahy Soc. 1958,54 1286. 4es Chowdhury and Basu J. 1959 3085. 4a1 Chowdhury and Basu Trans. Faraday Soc. 1960,46 335 428 Briegleb and Czekalla 2. phys. Chern. (Frankfurt) 1960,24 37. 42B Powell Huse and Cooks J. 1943 153; Powell and Huse J. 1943,435. 4s0 Hassel Moi. Physics 1958 1 241. 43L Andrews Chem. Reviews 1954,54 713. 438 Wacks and Dibeler J. Chem. Phys. 1959,31 1557. 483 Stevenson unpublished quoted in ref. 432. TABLE 27. Charge-transfer band energies (ev) of the complexes formed between aromatic hydrocarbons and chloranil (CA) tetracyanoethylene (TCE) tetrachlorophthalic anhydride (TPA) I ,3,5-trinitrobenzene (TNB) and iodine (I2). The Hiickel energy of the highest occupied molecular orbital ( E ) and the ionisation potential (I.P.) of the aromatic hydro- carbons and the apparent maximum extinction coeficients (cmaX> and equilibrium constants ( K ) of the iodine complexes Hydrocarbon EP-l I.P.433 I.P.432 CA428 TCE422 TPA427 TNB420 K425 lo Benzene 1.OOo 9.57 9-38 3.69 3.22 - 4.36 4.18 16400 Biphenyl 0.705 - - 2.8 5 2.48 3.62 - 3.65 4000 0.37 i4 0-15 ?i z Naphthalene 0-618 8.60 8.2 2.59 2-25 3-50 3.40 3.45 2395 0.62 E Phenanthrene 0.605 8-62 8.0 2.66 - 3.38 3.35 3-29 1492 1.06 8 e m 2.96 161 36.49 Pyrene 0.445 - - 2.07 1.71 2.92 - 2.86 2.70 2.89 112 52.35 Anthracene 0.414 8-40 7.5 1.99 - B (ev) 2.3 3.1 2.9 2.7 2.6 2.6 2-3 MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 359 energies of the highest occupied n-electron orbital can be fitted to the two sets of experimental ionisation potentials to give values of 2.3 and 3.1 ev for /3 the empirical resonance integral of the Hiickel theory (Table 27).The frequencies of the bands due to charge transfer between the series of polycyclic aromatic hydrocarbons and a variety of acceptors give linear relations with the theoretical energies of the highest occupied n-electron orbital and the slopes of the relations lie between 2.3 and 2-9 ev (Table 27) agreeing with equation (21) within the limits of uncertainty of the experimental ionisation potentials. Although assumed to be constant in equation (21) the energy term A contains components which change with both the donor and the acceptor and the range of ,8 values observed in the aromatic hydrocarbon series (Table 27) probably arise from differential variations of A within the series from one acceptor to another.The most important components of A are the Coulombjc energy Ec of the ion-pair or dative bond state (D+A-) and the resonance energy ER arising from the mixing of the ion-pair and the non-bonded state (D.A) (Fig. 14). I FIG. 14. The relations between the energies of the ground (N) non-bonded @.A) ion-pair (D+A-) and excited ( E ) states and the distance (R) between the donor (D) and acceptor (A) molecules in a charge-transfer complex. For point charges at van der Waals distances the Coulombic energy Ec has a value of -4 ev and the energy A is found commonly to have a value of this order. In general the positive and negative charges of the 360 QUARTERLY REVIEWS ion-pair state are delocalised over the T-electron system of the donor and the acceptor respectively and EC should become smaller with the progres- sive annellation of the hydrocarbon.The mixing of the non-bonded a h ( ~ . ~ ) and the ion-pair &D+A-) wave functions gives a ground & and an excited state function $E (23) . . . . . where a and b with a 9 b are the mixing coefficients related by the normalisation condition a2 + b2 = 1. Values of the coefficients can be derived from the dipole moment,3s1 the electronic spectrum,41g or the heat of formation of the complex.434 For the benzene-iodine complex435 calculated values of b range from 0.13 to 0.37. Owing to mixing the energies of the ground and the excited state relative to those of the non-bonded and the ion-pair state respectively are separated by a wider interval each changing by the energy increment ER (Fig. 14). The resonance mixing energy ER is given roughly by the heat of formation of the complex AH though in general ER should be somewhat larger than AH as the distance between the components in the ground state of the complex is usually smaller than the sum of the van der Waals radii so that the repulsion energy of the non-bonded electrons exceeds the balancing attractive energy of the dispersion forces.The heats of formation of charge-transfer complexes3sg are of the order of 0-2 ev and in a series of donors with a common acceptor the stability of the complex as measured by AH or more generally by the free energy dG progressively increases as the frequency of the charge-transfer absorption or what is equivalent the ionisation potential of the donor de- c r e a s e ~ . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Thus both of the important components Ec and ER contributing to the energy term A vary in a manner which should tend to give relations based upon equation (21) a slope somewhat greater than unity.This conclusion is not yet established experimentally as the available ionisation potentials for aromatic hydrocarbons lack the re- quired precision. Numerous empirical regularities have been observed in the spectra of organic charge-transfer complexes. The iodine418 and the ~hloranil~~l complexes give charge-transfer bands with a frequency proportional to that of the lowest singlet absorption of the donor. Trinitrobenzene or chloranil complexed with a variety of donors gives charge-transfer bands with widths at half maximum extinction proportional to the frequency of maximum indicating that the difference between the ground and the excited state dimensions is greater the larger is the transition 434 Ketelaar J.Phys. Radium 1954 15 197. 435 Ferguson and Matsen J. Amer. Gem. SOC. 1960 82 3268. 436 Foster Hammick and Parsons J. 1956 555. 437 Ham J. Chem. Phys. 1953 21 66,756. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 361 energy. The trend of charge-transfer absorption band intensities in a series of complexes is variable. Benzene with a number of acceptors gives charge- transfer bands with intensities directly proportional to the wavelength of maximum absorption,418 but more generally charge-transfer bands have intensities which are inversely proportional to the wavelength of absorption or to the stability of the complex389 (Table 27). For complexes of moderate strength it is expected theoretically381 that the intensity of the charge-transfer absorption should be proportional to the degree of mixing between the non-bonded (D.A) and ion-pair (D+A-) states the intensity increasing progressively with an increase in the stability of the complex.The effects of pressure on charge-transfer spectra support this view the absorption bands of quinhydrone and the chloranil-hexa- methylbenzene complex shifting to lower frequencies by -2500 ern? and increasing in intensity by a factor of 1.2 and 1.7 respectively under a pressure of 50,000 atmospheres.438 On compression the increase in the repulsion between the outer shell electrons is larger in the non-bonded than in the ion-pair state (Fig. 14). The energy interval between the two states is reduced resulting in a lower transition energy and a greater mixing of the two states which entails in turn a higher absorption in- tensity.However charge-transfer absorption is observed in cases such as iodine and n - h e ~ t a n e ~ ~ ~ in which complex formation in any sense other than statistical collision pairing is dubious. In a chance collision encounter between a donor and an acceptor there is no minimum in the ground-state potential-energy curve (Fig. 12) that is no complex is formed but the orbitals of the components may overlap adequately to give a mixing of the non-bonded and the ion-pair states with a substantial transition In the cases where a definite complex can be isolated collision charge transfer as well as complex charge transfer contributes to the total observed absorption intensity.In a series of related complexes the portion of the intensity arising from the complexed donor-acceptor pairs progressively increases and that contributed by the collision pairs de- creases as the complexes become the more stable. With assumptions as to the extinction coefficients of the contact and the complex charge transfer absorption the observed general trend of a decrease in absorption intensity with an increase in the stability of the complex can be explained.386 The calculations suggest that as much as three quarters of the observed iodine- benzene charge transfer intensity can be derived from collision pairs.386 A mixing of the charge-transfer transition with the transitions of the individual component molecules particularly those of the donor can contribute additional intensity to the charge-transfer a b ~ o r p t i o n .~ ~ In complexes composed of a donor and an acceptor which are both aromatic such mixing is not possible for the polarised spectra of crystalline com- 438 Stephens and Drickamer J. Chew. Phys. 1959,30 1518. 439 Evans J. Chern. Phys. 1954 23 1436. 5 362 QUARTERLY REVIEWS plexes of this type indicate that the charge-transfer transition moment is directed perpendicular to the molecular as expected for an electron transfer between the parallel aromatic rings. The lower energy transitions of aromatic hydrocarbons are polarised in the molecular plane and they cannot mix with perpendicular transitions. In solution collision pairs particularly those of the halogens with aromatic hydro- carbons are unlikely to have the symmetrical strustures of the solid complexes,43o and contact charge-transfer absorption may be intensified by a contribution from a transition of the When both the donor and the acceptor have extended v-electron systems the planes of the components can remain parallel in solution particularly at low temperatures.Hexamethylbenzene and tetracyanoethylene both absorb at 2650 A the former weakly and the latter with high intensity. In the spectrum of the complex dissolved in a glass at low temperature the hexamethylbenzene absorption is shifted towards the red and the tetracyanoethylene absorption towards the blue region whilst the total intensity is shared equally between the two bands (Fig. 15) indicating a I u ' 16 20 2(8 32 36 4b 44x10' v (cm-I) FIG. 15. The absorption spectra in dipropyl ether-methylcyclohexane (4 1) at 9 3 " ~ of' tetracyanoethylene (I) hexamethylbenzene (11) and their complex (111) (reproduced with permission from refs.428 and 441). complete mixing of the two original The original transitions forbidden and allowed respectively are polarised in the planes of the individual component molecules and such extensive mixing of the transi- tions would not be possible if the planes of the components were not parallel in the complex. 440 Nakomoto J. Amer. Chem. Soc. 1952 74 1739. 441 Czekalla 2. Elektrochem. 1959 63 1157. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 363 The weak absorption (emax 342) of sulphur dioxide at 2900 8 is con- siderably enhanced by complex formation with olefins cyclopentene and sulphur dioxide giving bands at 27 10 and 2970 A with extinction coefficient maxima of 4170 and 4190 due to charge-transfer and enhanced sulphur dioxide absorption respectively.424 Thus the weak sulphur dioxide absorption is due to a transition polarised perpendicular to the plane of the molecule as otherwise the charge transfer and sulphur dioxide absorptions would not mix to the observed degree.The visible absorption band of iodine shifts to higher energies on complex formation by increments which are a little larger than the heat of formation of the The shift is due to the raising of the energy of the lowest unoccupied orbital of iodine on complex formation this orbital serving as both the upper level of the iodine absorption in the visible and the acceptor level of the charge-transfer absorption in the ultraviolet region.Complex formation raises the energy of the lowest unoccupied orbital of iodine by the resonance energy increment ER and lowers the ground-state energy by the same quantity of which the heat of formation of the complex is an approximate measure (Fig. 14). The intensity of the iodine absorption in the visible region is unchanged on complex formation indicating together with the observed energy shifts,387 that the transition responsible for that absorption does not mix with the charge-transfer transition for any mixing would enhance the intensity of the iodine band in the visible region and shift it to longer wavelengths. 15. Steric Effects Stereoisomeric compounds have in general different spectroscopic properties. Optical isomers absorb at the same wavelength but the extinc- tion coefficients may differ by as much as 10 % e.g.the n -+ 7 ~ * absorption of when their spectra are measured with circularly polarised light. C~nformational~~~ and geometrical446 isomers commonly show larger spectral differences. The absorption due to an allowed transition is invariably more intense in the trans- than in the cis-isomer of a conjugated compound owing to the greater transition-moment length but the cis- isomer because of its lower symmetry may absorb relatively strongly in regions where the trans-absorption is weak. All-trans- but not the cis- polyenes have a centre of symmetry and the first “overtone” transition for which s = 2 (see above) is forbidden in the zero-order approximation for the all-trans-compounds owing to the cancellation of the component transition moments.First-order configuration interaction renders the transi- tion weakly allowed in the all-trans-polyenes but the transition is allowed 442 Reid and Mulliken J. Amer. Chem. SOC. 1954 76 3869. 443 Tames Virzi and Searles J. Amer. Chem. SOC. 1953 75 4358. 444 Kuhn and Gore 2. phys. Chem. 1931 B 12,389. 446 Barton and Cookson Quart. Reviews 1956 10,44. 446 Crombie Quart. Reviews 1952 6 101. 364 QUARTERLY REVIEWS in the zero-order approximation for the cis-isomers as the component moments have a non-zero resultant and the first “overtone” transition appears as the relatively strong “cis-peak” absorption. In the non-coplanar polyenes the “cis-peak” may have its origin in a partial chromophore localised by the steric hindrance to full conjugation. Absorption wavelength differences between cis- and trans-isomers in the absence of steric hindrances to coplanarity in the cis-isomer depend upon detailed differences in the contribution of changes in electron repulsion energy to transition energy and they follow no general rule.cis-Polyenes often absorb at shorter wavelengths447 than their trans- isomers though exceptions have been re~orded,4~~ and in the case of the polycyclic dienes the converse is true (Table 3). The steric twisting or displacement of a bond which is part of a con- jugated system almost invariably reduces the intensity of bands due to allowed transitions. However the structural rearrangement may permit the mixing of an allowed and a forbidden transition enhancing the intensity of the absorption due primarily to the latter as in the notable case of cis-azobenzene (see above).Wavelength changes due to steric effects can be bathochromic or hypsochromic the latter being the more usual if one particular bond in a molecule is twisted e.g. in the 2-substituted biphenyls and the former if the steric strain is distributed uniformly throughout the molecule as in the helicenes e.g. (43) or the paracyclo- phanes (44). The twisting of a bond or the change of a bond angle results in a reduc- tion of w-electron delocalisation as measured by the resonance integral p of the bond or bonds. In an alternant hydrocarbon the highest occupied and the lowest unoccupied orbitals are paired with energies E = 2 k/3 so that the transition energy is 2kp for the p-band of an aromatic hydro- carbon or the long-wavelength absorption of a polyene.A distortion which is uniform or nearly so throughout a conjugated system reduces the average value of 18 giving a smaller transition energy and a bathochromic displacement of the absorption.449 In the paracyclophane (44; n = rn = 2) the benzene rings have a boat configuration with the substituted para- carbon atoms displaced some 0.26 A from the plane of the unsubstituted 447 Zechmeister Experientia 1954 10 1. 448 Holme Jones and Whiting Chem. and Ind. 1956 928. 449 Coulson “Steric Effects in Conjugated Systems,” ed. Gray Butterworths London 1958 p. 8. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 365 benzene carbon atom~,4~O and the resultant changes in the resonance integral 18 satisfactorily for the red shifts of the benzene a- and p-bands (Table 28). aZZ-cis-Deca-2,4,6,8-tetraene which probably has a helical structure with the strain distributed throughout the molecule absorbs at 3100 A whilst the all-trans-isomer absorbs at 3040 A.448 In contrast polyenes which are neither all-cis nor all-trans in which any strain is taken up by one or a small number of bonds generally absorb at a shorter wavelength than the all-trans-is~mer.~~~ TABLE 28.Direrences in band wavelength (A) between the paracyclo- phanes (44) and bis-p-ethyZphenyZb~tane~~~ Paracyclophane n=m=2 n=m=3 n=m=4 p-Band (223081) 210 230 30 a-Band (273081) 290 210 10 If only one bond in a conjugated system is twisted the wavelength displacements of the light absorption depend upon the sign and the magnitude of the change in the double-bond character of the distorted bond on excitation.449~452 Hypsochromic shifts arise if the twisted bond has a larger bond order in the excited than the ground state and conversely bathochromic shifts result from a decrease of bond order during the transition.In general the change in transition energy AE produced by the twist of a bond between the atoms r and s through an angle 6 is given by449945z A E = ~~O(COS 6 - I)APTs . . . . . (24) where APT is the change in the order of the bond between the atoms r and s and /lo a negative energy is the resonance integral of the un- distorted bond. The formally single bonds in the polyenes undergo an increase of bond order on excitation and a steric effect which distorts one or a small pro- portion of such bonds in the molecule produces a blue shift relative to the absorption of the all-trans-i~omer.~~~ In biphenyl styrene and stilbene the formally single bonds linking the unsaturated residues undergo changes of APT on excitation which are large and positive,452 resulting in considerable hypsochromic shifts of the long-wavelength absorption following ortho- substitution in these m01ecules~~~-~~~ (Table 29).The formally single bonds of a/3-unsaturated ketones and aryl ketones are subject to very small changes of bond order on excitation,45z and the wavelength shifts induced 450 Brown J. 1953 3265 3270 3278. 451 Cram Allinger and Steinberg J. Amer. Chem. SOC. 1954 76 6132. 452 Heilbronner and Gerdil Helv. Chim. Ada 1956 39 1996. 453 Beaven “Steric Effects in Conjugated Systems,” ed. Gray Butterworths London 454 Calvin and Alter J. Chem. Phys. 1951 19 765. 455 Beale and Roe J .Amer. Chem. SOC. 1952 74 2303. 456 Overberger and Tanner J . Amer. Chem. SOC. 1955 77 371. 1958 p. 22. 366 QUARTERLY REVIEWS by hindering substituents are relatively sma11,457,458 arising as much from the electronic as the steric perturbation of the mloecule (Table 29). The bond order of the link between the carbonyl group and the nucleus in l-acetylazulene decreases substantially on excitation and substituents hindering a coplanar arrangement of the chromophore produce consider- able red shifts of the absorption due to the transition of an electron from the azulene ring to the carbonyl group452 (Table 29). TABLE 29. Compound Diphenyl 4- Methyl- 2-Met hyl- 2 6-Dime t hyl- Acetophenone 4-Methyl- 2- Methyl- 2,6-Dimethyl- 1 -Acetylazulene 2-Methyl- 3-Acetylguaiazulene Amax (4 2480 2515 2365 23 10 2430 2520 2420 2510 2190 2400 2750 Emax 17,000 19,000 10,250 5600 13,200 15,100 8700 5600 20,000 17,800 13,500 Ref.453 453 453 453 457 457 457 457 452 452 452 In a compound consisting of two groups linked by a bond subject to steric distortion the absorptions due to the two main types of transition local excitations within the individual groups and charge-transfer transi- tions between the groups are modified in characteristically different ways.459 Probabilities of charge-transfer transition in such cases are particularly sensitive to steric distortions of the bond between the two groups resulting in large changes in intensity and displacements in wave- length dependent upon the change of the bond order of the link on excitation. The absorptions listed in Table 29 are largely of this kind.The bands due to local excitations within either of the two groups have in- tensities which are relatively insensitive to steric perturbations of the bond between the groups and they undergo wavelength displacements dependent upon changes in the contribution of electron repulsion energy terms to the transition energy.459v460 The 19-band of biphenyl at 2000 A which arises from the El transitions of the two benzene rings has an intensity Emax -40,000 but is little affected by progressive substitution in the ortho-po~itions.~~~ The aromatic a-band absorption of benzaldehyde and acetophenone undergoes on alkyl substitution intensity increases as well as The intensity changes are roughly additive according to the vector diagram (1 9) and they are due largely to the electronic perturbation of the component transition 467 Braude and Sondheimer J.1955 3754. 4s* Braude and Timmons J. 1955 3766. 46B Murrell J. 1956 3779. 480 McRae and Goodman J. Mol. Spectroscopy 1958 2 464. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 367 moments by the substituents steric effects entering only to the degree that they change the “spectroscopic moment” of the carbonyl group. Steric hindrance to coplanarity generally increases the intensity of the carbony1 absorption in unsaturated ketones owing to the mixing of the n -+ T* and the charge-transfer transitions e.g. (5). Few absorption bands of compounds consisting of two conjugated groups are due to transitions with a pure charge-transfer or a pure local- excitation character but complete mixing of the types is relatively un- common occurring only when the two transitions have the same energy and parallel moments.16. Environmental Effects In the presence of matter and radiation the energy levels of a molecule are only approximately true “stationary states” all interactions introduc- ing some uncertainty as to the exact values of the energies. Vapours at low pressures where the levels are well defined give electronic spectra with rotational and vibrational structure but at higher pressures the bands broaden and the rotational structure disappears owing to the breakdown of rotational quantisation when the collision and the rotational frequencies are of similar orders of magnitude. In the liquid phase the vibrational structure broadens and it is frequently lost if the absorption intensity is large.In the crystal the N molecules per unit cell interact on excitation giving for each allowed molecular transition N distinct allowed crystal transitions which can absorb at widely different frequencies. These strong interactions between the molecules in the crystal and probably also in the liquid phase break down the vibrational quantisation and in general only structureless absorption bands are observed considerably displaced from the wavelengths appropriate to the vapour phase. Forbidden molecular transitions interact only weakly in the crystal and their absorp- tions generally appear with well-resolved vibrational structure showing modifications relative to the spectrum of the vapour phase which can be diagnostic of the transition involved. The high-intensity S-band of anthracene appearing at 2500 A in solution is split into broad structureless components at -2700 and -2000 A in the crystal where there are two molecules per unit ce11.461-4s7 The moderate- intensity p-band of anthracene appears in the crystal with considerable vibrational structure and the splitting between the components is only 461 Craig and Hobbins J.1955 539 2309. 46a Craig J. 1955 2302. 463 Bree and Lyons J. 1956 2662. 464 Lyons and Morris J. 1959 1551. 466 Sidman Phys. Review 1956 102,96. 466 Broude Pakhomova and Prikhotjko Optics and Spectroscopy U S S R . 1957 467 Ferguson and Schneider J. Chem. Phys. 1958,28 761. 2 323. 368 QUARTERLY REVIEWS -30 cm.-l compared with some 16,000 cm.-l for the /3-band.461-467 These observations establish that the p - and the /3-bands of anthracene are polarised along the short and the long axis of the molecule respect- ively .461 A non-polar molecule in dilute solution or with a high pressure of inert gas in the vapour phase gives a spectrum shifted to longer wave- lengths relative to that of the vapour at low pressure.The wavelength displacement has been ascribed semi-classically to the polarisation of the electrons in the solvent molecules by the transition moment of the The shift to lower frequencies represents the reduction in the energy required to produce the transition dipole in a dielectric medium relative to the gas phase at low pressure and as such it should be pro- portional to the oscillator strength of the transition and the square of the refractive index of the High-intensity bands usually undergo larger solvent shifts than weak bands but the displacements of the latter are not inconsiderable.Between paraffin and benzene solution,470 the a- p- and ,&bands of aromatic hydrocarbons shift to lower frequencies by -1 50 -300 and -450 cm.-l respectively yet the transition moments of the 6-bands are some lo6 times as large as those of the a-bands. Quantum mechanically the solvent-induced absorption red shifts are ascribed to a change in the dispersion forces between the solvent and the solute on excitation.471 The magnitudes of the dispersion forces depend upon the intervals between the energy levels in the solute and the solvent molecules. The energy separations between the first and the higher excited states of the absorbing molecule are smaller than those between the ground and the corresponding upper states so that the dispersive attractions are the smaller for the ground state.The resultant solvent shift of the absorp- tion to lower frequencies dv is given by471 AV = agz(Ea44 + M2)/6R6 . where a~ and ~ l g are the molecular polarisabilities of the solute and the solvent respectively M and E are the dipole moment and the energy of the transition and z is the number of solvent molecules at an average distance R surrounding each solute molecule. By giving a displacement dependent upon the sum of terms due to the energy and the moment of the transition equation (25) accounts for the appreciable red shifts of bands due to forbidden transitions and for the larger shifts of high intensity bands. The molecular polarisability a ~ is proportional to the square of the solvent refractive index n2 explaining (eqn.25) the linear relations which have been observed472 between the 468 Bayliss J. Chem. Phys. 1950 18 292. 469 Bayliss and McRae J. Phys. Chem. 1954 58 1002 1006. 470 Coggeshall and Pozefsky J. Chem. Phys. 1951 19 980. 471 Longuet-Higgins and Pople J. Chem. Phys. 1957 27 192. 472 Robertson Babb and Matsen J. Chem. Phys. 1957,26 367. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 369 absorption wavelength of a solute A and the value of n2 for the solvent. Some values of dh/dn2 are 24 8 for the a-band of benzene472 (emax -lo2) 220 A for the iodine-benzene charge-transfer band473 (emax N lo3) and 300 A for the long-wavelength absorption of l y ~ o p e n e ~ ~ ~ (emax -lo5). The dispersion forces are weak and where they provide the main solute-solvent interactions vibrational quantisation is preserved so that the electronic absorption bands appear generally with vibrational structure.The stronger dipole-dipole and particularly hydrogen-bonding inter- actions give frequency shifts larger by an order of magnitude than those due to dispersion forces and they break down vibrational quantisation resulting in smooth and structureless electronic absorption bands. The direction of the frequency shifts depends upon the sign of the permanent dipole moment change of the absorbing molecule on excitation for the case of the dipole-dipole interaction and upon the sign of the molecular polarisability change for the case of dipole-induced dipole inter- a ~ t i o n . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ The magnitudes of the shifts depend upon the values of these changes in the solute on excitation and of the dipole field due to the solvent molecules at the position of the solute dipole.475p476 Molecules which are highly dipolar or zwitterionic in the ground state commonly have a reduced and reorientated dipole moment in the excited state.The solvation cage of the ground state is subject to orientation strain in the excited state and large blue shifts of the solute absorption result on changing from a less to a more polar solvent. The absorption of the zwitterion (28) for example moves 3000 cm.-l to higher frequencies on change from dioxan to aqueous in contrast to the case of Phenol Blue (45) which has an absorption band in the visible region that moves by a similar increment to lower frequencies under the same condi- t i o n ~ .~ ~ ~ Phenol Blue can be regarded as a hybrid from (45a) and (45b) the former contributing largely to the ground state and the latter predominat- ing in the excited state. The upper state of Phenol Blue and similar molecules is formed in a solvation cage which is already partly orientated by the small ground-state dipole and although the cage is subject to some 473 Ham J . Amer. Chem. Soc. 1954 76 3881. 474 Rosen and Reid J. Chem. Phys. 1952,20 233. 475 McRae J . Phys. Chem. 1957 61 562. 476 Ooshika J. Phys. SOC. Japan 1954 9 594. 370 QUARTERLY REVIEWS strain the solvation energy is greater in the excited than the ground state. The differential stabilisation of the upper state by solvation increases with the polarity of the solvent.The absorption of merocyanine dyes in which the non-polar and the zwitterionic structures have comparable energies exhibit striking solvent In certain cases e.g. (46) the zwitterionic structure appears to be more stable than the neutral structure in polar solvents but less stable in non-polar solvents so that first red shifts and then blue shifts are observed as the solvent polarity is progressively increased480 (Fig. 16). At the same FIG. 16. (; tion in dioxan-water mixtures.480 The variation of the wavelength ( ) and the extinction coeficient -. -) of the absorption maximum of the merocyanine (46) with solvent composi- time the maximum extinction coefficient of the absorption rises and then falls (Fig. 16) but compensating changes in the band half-width maintain an approximately constant oscillator s trength.48 O The ground $G and the excited state wave function t,+b~ of (46) can be written as a combination of the functions of the neutral t + b ~ and the zwitterionic +z; structures (46a) and (46b) respectively .. . . . . . $G = a$N + b+z (26) $ ~ = b t + b ~ -Q$z (27) . . . . . . . . 478 Brooker Experientia Supplementum I1 Proceedings of the 14th International Congress of Pure and Applied Chemistry Zurich 1955. *79 Brooker Keyes Sprague Van Dyke Van Lare Van Zandt White Cressman and Dent J. Amer. Chem. SOC. 1951,73,5332. 480 McRae Spectrochim. Acra 1958 12 192. MASON MOLECULAR ELECTRONIC ABSORPTION SPECTRA 37 1 In non-polar solvents the mixing coefficients are such that a > b but in polar solvents b > a. In a suitable solvent of intermediate polarity a = b and both structures contribute equally to the ground and the excited states which then have their minimal energy separation.At this point in the solvent polarity range the bonds along the conjugated pathway have a uniform bond order in both the ground and the upper state so that there is no change in the relative positions of the atomic nuclei on excitation. The band origin appears strongly in the absorption to the virtual exclusion of higher members of any excited state vibrational pro- gression resulting in a high maximum extinction coefficient and a narrow band width. In circumstances where a # b a change in the bond alterna- tion along the conjugated chain occurs on excitation and the higher members of upper state progressions in the stretching vibrations are involved in the absorption giving a broadened band and a reduced maximum extinction The solvent shifts which characterise Burawoy’9 R-bands have been employed as criteria to distinguish between n -+ 7 ~ * and weak 7~ -f 7 ~ * b a n d ~ .l ~ ~ In general absorption involving lone-pair electrons whether of the n -+ 7 ~ * or the n -+ CT* type moves to higher frequencies on change from a non-polar to a hydroxylic solvent owing to the stabilisation of the lone-pair electrons in the ground state by hydrogen-bonding.ll However the weak 3500 8 absorption of the alkyl nitrites is insensitive to a change of solvent,175 though the band is probably due to a n + 7 ~ * transition since the absorption is polarised perpendicular to the molecular plane in the nitrite and in contrast red shifts are observed in the case of the shorte r-wavelength carbonyl absorption of the ~x-diket0nes.l~~ Solvent- induced absorption displacements are not a property of the ground state alone and a highly dipolar excited state resulting from a n -+ 7 ~ * transition should give such red shifts.Of the various solvent effects the disappearance of a weak band in acid solution due to formation of cation remains the sole universal indication of a transition involving lone-pair electrons. The author is indebted to Professor R. C. Cookson Professor A. D. Walsh and Dr. M. C . Whiting for the provision of data before publication and to Professor R. C. Cookson and Dr. J. Czekalla for their perahion to reproduce the spectral curves of Fig. 6 and Fig. 15 respectively. Woods and Schwartzman J. Amer. Chem.Soc. 1948,70 3394; Howton J Org. Armitage Jones and Whiting J. 1952 2014. Chem. 1949,14 1. 482 Price and Walsh Trans. Faraday Soc. 1945 41 381.

ISSN:0009-2681    

DOI:10.1039/QR9611500287

出版商:RSC    

年代:1961

数据来源: RSC

摘要:

AN OUTLINE OF RHENIUM CHEMISTRY By A. A. WOOLF ALDERMASTON COURT ALDERMASTON BERKS.) RHENIUM is the rarest of the naturally occurring elements. Its average concentration in the earth’s crust (0.001 part per million) is of the same order of magnitude as that of some of the platinum metals but it does not occur as the element or as a distinct mineral species. Ores or residues are considered rich in rhenium when they contain 10-100 p.p.m. though one really exceptional ore has been found which contained 3000 p.p.m. In spite of or perhaps even because of its rarity rhenium is by no means one of the less familiar elements and the extent of its chemistry can only be outlined here. However rhenium chemistry has been so well documented since its early days that the reader can consult the latest books for detail~,l-~ and allow the Reviewer a bias towards the more recent work not reported therein.(RESEARCH LABORATORY ASSOCIATED ELECTRICAL INDUSTRIES Discovery Production and Preparation of the Metal The discovery of rhenium seems conventional when viewed in retrospect. Its existence as dvi-manganese was predicted by Mendelkev. It may have been discovered in the last century but adequate confirmation had to await Moseley’s law which unequivocally related the X-ray spectra of elements with their atomic numbers. It was not until 1925 however that the X-ray spectrum expected for element 75 was found by the Noddacks together with Berg in concentrates from platinum minerals and by Loring and Druce in manganese compounds. Heyrovsky and DolejsEk also con- firmed the presence of rhenium in commercial manganese salts by the then novel technique of polarography.By 1928 the Noddacks had laboriously isolated the first gram of rhenium. Feit achieved technical scale prepara- tion (100-200 kg./annum) in the 1930’s from molybdenite residues recovered from copper schists. Nowadays most rhenium is recovered as a by-product in the manufacture of molybdenum and world production is probably in the region of 5 tonnes/annum. The flue dusts obtained from molybdenum.concentrates are enriched in rhenium because of the volatility of rhenium heptaoxide. The rhenium contained in this dust can be extracted with water containing an oxidising agent ; alternatively the rhenium in the concentrates can be solubilised without volatilisation by heating them with lime at 600”. Calcium per- rhenate is readily leached out with water from the insoluble calcium molybdate.Direct solution is possible by leaching at low temperatures Druce “Rhenium” Cambridge University Press Cambridge 1948. Tribalat “RhCnium et Technetium” Gauthier-Villars Paris 1957. Lebedev “Renii” Moscow 1960. 372 WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 373 under pressure (170-200" and 50 atmospheres). The rhenium in solution can then be recovered by (a) precipitation of potassium per-rhenate or rhenium heptasulphide (b) ion exchange4 or adsorption5 (this is particu- larly effective for separating molybdenum from rhenium because their highest-valency state ions carry different charges and by altering the conditions can be made anionic or cationic) (c) electrodeposition from acid solutions (d) cementation on to iron sponge in acid solutionY6 or (e) solvent extraction from acid or alkali s~lution.~ In most processes the impure product can be converted into potassium per-rhenate which is easily purified by recrystallisation.Originally the metal was produced by reducing this salt with hydrogen at 1000" but the product was always contaminated with potassium. Impure rhenium can be purified by conversion into the pentachloride with chlorine hydrolysing the pentachloride to rhenium dioxide and reducing the last.a On a small scale it is more convenient to convert potassium per-rhenate into the ammonium salt on ion-exchange resins and then to reduce this salt to a purer rheni~m.~ The reduction temperature can be lowered considerably by using a hydrogen pressure of 50 atmospheres. Thermal decomposition of rhenium chlorides and carbonyls is used mainly to coat other metals with rhenium.The relative instability of rhenium iodides precludes a continuous van Arkel-de Boer method of purification.1° Electrolysis of per-rhenates in aqueous solution does not yield pure rhenium;ll the electrodeposit always contains some oxide and it is usual to anneal in hydrogen at 1000" to achieve permanent coatings.la Rhenium is the highest-melting metal (3180") next to tungsten and has a density of 21.04. The natural element consists of only two isotopes la7Re and 185Re but nine other radioisotopes are known. The Oxygen Compounds of Rhenium Rhenium Heptaoxide.-This results when the metal is heated in air or oxygen. The oxidation proceeds through intermediary lower oxides since one can observe the red trioxide on some faces of a single metal crystal at lower oxygen pressures,13 and conversely the dioxide can be isolated by hydrogenating the heptaoxide at about 300".The peroxide (Re,Oa) once claimed seems to have been a fine spray of per-rhenic acid caused by Meloche and Preuss Analyt. Chem. 1954 26 191 1 ; Ryabchikov and Borisova Alexander J. Amer. Chem. SOC. 1949,71,3043; Galyaeva Zimakov and Rudenko Kovyrshin and Appolonov Tsvetnye Metally 1957 (S) 67. Tribalat Ann. Chim. 1953,8,642; Kertes and Beck,J. 1961 1921. * Rosenbaum Runck and Campbell J. Electrochem. SOC. 1956 103 518. Woolf J. Less-Common Metals 1959 1,420. lo Woolf J. Inorg. Nuclear Chem. 1958 7 291. l1 Lundeil and Knowles J. Res. Nut. Bur. Stand. 1937 18 629. la Levi and Espersen Ph-vs. Rev. 1950 78 231.l3 Reviewer's unpublished observation. Zhirr. anal. Khim. 1958 13 155,492. Tsvetn-ve Metally 1959 (5) 73. 374 QUARTERLY REVIEWS traces of moisture.2 The difficulty of collecting the heptaoxide from a flowing gas necessitates oxidation in a static atmosphere and in such a system the formation of the heptaoxide is a convenient way to purify and separate rhenium. O The melting and boiling points calculated from vapour-pressure measurement~l~ since the solid sublimes are 300.3" and 360.3". In this respect it closely resembles the neighbouring osmium tetraoxide rather than tungsten trioxide. Unlike osmium tetraoxide it is reported to be soluble in ethers alcohols and amines without being reduced. Per-rhenic Acid.-The acid is formed when the heptaoxide dissolves in water.Solutions of the acid can also be prepared by dissolving the metal powder in hydrogen peroxide but with the consolidated metal this oxida- tion is too slow to be practicable. Anodic oxidation in alkali hydroxide solution13 followed by removal of the alkali ions on a cation-exchange resin is satisfactory. The acid is strong as can be demonstrated qualitatively by the behaviour of its salts on ion-exchange resins the pH of its aqueous solutions and the attack on metals metal oxides hydroxides and carbon- ates. Quantitatively the dissociation constant has been determined by observing how the optical absorption of the acid varies with concentration.15 The first dissociation constants of perchloric permanganic per-rhenic and periodic acids are lo7 400 40 and 0.02 respectively (the perchloric constant is an estimate).Pauling16 has suggested an empirical correlation between the strength of an acid and its structure. For an acid XO,(OH) the constants are -lo5 and -los for rn = 1 2 or 3 respectively. The ortho-acid HJO, which can be isolated from aqueous solution has the acidity expected of an acid with rn = 1 but the rhenium acid with rn = 2 which would have the acidity found experimentally has never been isolated although its salts exist :17 ReO,- + 20H- + [KeO,(OH),] + Re0,3- + H,O Neither a hydrate nor the anhydrous acid can be obtained from per- rhenic acid solutions. The colourless dilute solution becomes yellow-green on evaporation and additional lines presumably of the undissociated acid appear in the Raman spectrum. Further evaporation which leads to volatilisation of the heptaoxide has been employed for analytical separa- tion sulphuric acid being added to raise the distillation temperature.ls The per-rhenates resemble the perchlorates in their aqueous solubilities.(There is no corresponding resemblance in organic solvents.) The sodium salts are about a hundred times more soluble than the potassium salts and l4 Smith Line and Bull J. Amer. Chem. Soc, 1953 74 4964. l5 Bailey Carrington Lott and Symons J. 1960 290. l6 Pauling "The Nature of the Chemical Bond" Cornell University Press 1960 l7 Scharnow 2. anorg. Chem. 1933 215 184; Scholder Angew. Chem. 1958 70 p. 324. 583. Geilmann and Bode 2. analyt. Chem. 1950 130 323. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 375 other metallic salts are extremely soluble. Silver per-rhenate which has been used to determine the atomic weight of rhenium is only slightly soluble compared with the perchlorate.Salts of large organic bases (nitron bruciiie strychnine) are insoluble enough for the analysis of rhenium. The tetraphenyl-arsonium or -phosphonium salts are very much less soluble in water than in chloroform and microanalysis of rhenium can be effected by solvent extraction of these ~a1ts.l~ Reduction of Per-rhenates.-The per-rhenate ion resembles the perchlor- ate rather than the permanganate ion in its action towards reducing agents. The reduction of per-rhenates is usually slow and the mean state reached at a particular time must depend on the kinetics of the intermediate states through which they pass. An integral reduced state can be reached by adding an excess of reducing agent preferably in a medium which can form a complex with the rhenium.In general the higher the acidity the higher the valency of the reduced state. Nitric and perchloric acids cannot be used as reducing media because reduced rhenium species catalyse their decomposition. The results of some investigations are summarised in Table 1. TABLE 1. Chemical reduction of the per-rhenate ion HCl(8-10) H2S0,(<3*6) HCl(0.6-3 -2) H2S0,( 1 0 - 1 8) H2S04(9) HCl(4.8-6.4) (10-18) HZ1 HCI Reducing agent Sn2+ Ti3+ V2+ Cr2+ Sn2+ Ti3+ V2+ Cr2+ Cr2+ Sn2+ Cr2+ 21- 31- Zn-Hg Cd-Hg Bi-Hg N2H4 W O 2 Reduced state 4 4 5 + 4 (6 transient) 5-+4 + 7 slowly 3 5 4 0 + 1 4 2 + 3 3 + 4 4 5 4 4 Ref. 20 Y Y ¶ 9 Y ? Y Y 21 Y Y Y Y Y Y 22 Y Y These results can be applied preparatively (see p. 380) and analytically.Thus per-rhenates after reduction with bismuth amalgam can be estimated volumetrically by re-oxidation to per-~henate.~~ Ref. 2 p. 124. 2o Tribalat Ann. Chim. 1949 4 289. 21 Lazarev Zhur. neorg. Khim. 1956 1 385. aa Rulfs and Meyer J. Amer. Chem. Soc. 1955 77 4505. es Spitzy Magee and Wilson Mikrochim. Acta 1957 354. 376 QUARTERLY REVIEWS The polarographic reduction of per-rhenate is summarised in Table 2. The valency state reached is calculated from the IlkoviE equation.24 The polarographic waves are irreversible and the half-wave potentials have no thermodynamic significance. The assignment of a three-electron reduction in perchloric acid has been questioned25 because this value has only been obtained at one temperature. Gas is generated at the mercury surface indicating decomposition of the acid as in chemical reductions.TABLE 2. Polarographic reduction of the per-rhenate ion Supporting electrolyte(M) KCl(2) HC1O4(4) HCl(4) KOH(0- 1) KCN(O.1) Reduced state - 1 4 4 - 1 5 and 1 The reduction to an apparent -1 state is discussed later. Rhenium Trioxide.-The trioxide is a red paramagnetic solid which disproportionates in vacuo above 300" to the di- and the hepta-oxide. It is formed by allowing the heptaoxide to react with rhenium or the dioxide at about 300". The trioxide is prepared more conveniently by decomposing the compound obtained from the heptaoxide and dioxan at 145" or by heating the heptaoxide in carbon monoxide. It is remarkably inert to dilute acids and bases provided that they are non-oxidising so that rhenates (Re0,2-) must be prepared indirectly by fusing mixtures of rhenium dioxide and per-rhenate.Rhenium Dioxide.-The dioxide is also paramagnetic. This black oxide can be obtained by partial reduction of ammonium per-rhenate or rhenium heptaoxide with hydrogen at 300" or with rhenium at 600". The hydrated oxide formed in hydrolysis of hexahalogenorlienates(rv) can be dehydrated in vacuo at 650". It is soluble in acids with which it forms complexes but insoluble in bases. Aqueous oxidising agents convert it into per-rhenate. The rhenites containing the anions ReOa2- Re044- and ReOS6- are formed from the dioxide and fused alkalis.26 The compositions of the other oxides reported are rather uncertain. Thus the solid isolated by alkaline hydrolysis of rhenium trichloride was believed to have been a hydrated sesquioxide Re,O,.Similar hydrates of Re0 and Re,O from per-rhenic acid reduced in hydrochloric acid with cadmium and zinc respectively were also ill-defined. The Halogen Compounds of Rhenium This section of rhenium chemistry has been drastically revised in recent years. New compounds have been isolated and old ones contradicted. 24 Colten Dalziel Griffith and Wilkinson J. 1960 71. 25 Rulfs and Elving J. Amer. Chem. Soc. 1951 73 3284. 26 Deschanvres Ann. Chim. 1959 4 1217. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 377 The fluorine compounds arersummarised in Table 3. TABLE 3. Fluorine compounds of rhenium Oxidation state Simple fluorides Oxy fluorides Complex anions 7 ReF,27 ReOF and Re0,F329 Re0,F3 - 6 5 4 ReF,28 ReOF430 ReO,F,( ?)2s ReFs2-( ?) 33 ReO,F,- 34 Simple Fluorides.-The heptafluoride has only recently been made2' by passing fluorine at 250 mm.pressure over rhenium heated to 300-400". The volatile product was purified by heating it at 400" in a static atmos- phere of fluorine. The purity of the product is best judged from vapour- pressure measurements and as little as 0.1 % of rhenium hexafluoride can be detected in this way. Earlier workers believed that the hexafluoride was the highest fluoride attainable and it is important to bear in mind that in all previous work with the hexafluoride contamination with the hepta- fluoride had been neglected. The hexafluoride was originally made from fluorine and rhenium at 125". Chlorine trifluoride has also been employed as a fluorinating agent. The hexafluoride should be heated with excess of rhenium metal to ensure the removal of any heptafluoride.According to Ruff and Kwasnik28 it attacks silica at room temperatures. Later workers maintain that it is stable and only disproportionates above 300". The hexafluoride can be reduced to the pentafluoride by tungsten carbonyl and an excess of tungsten hexafluoride. The pentafluoride disproportionates in vacuo above 180" to the tetra- and hexa-fl~orides.~~ Ruff and Kwasnik made a tetra- fluoride by reducing the hexafluoride with either hydrogen at 200" or sulphur trioxide at 400" but it is believed to have been a mixture of tetra- and penta-flu~rides.~~ Attempts to prepare lower fluorides from the trihalides and anhydrous hydrogen fluoride have been unsuc~essful.~~ 0xyfluorides.-Ruff and Kwasnik isolated ReOF and ReO,F when a fluorine-oxygen mixture was passed over rhenium at 125-300".27 Malm Selig and Fried J. Anzer. Chem. SOC. 1960 82 1510. 28 Ruff and Kwasnik. 2. anorg. Chem. 1932 209 113; 1934 219 65; Cady and Hargreaves J . 1961 1563; Nikolaev and Ippolitov Doklady Akad. Nauk S.S.S.R. 1960,134,358. 2 9 Aynsley Peacock and Robinson J. 1950 1622; Cady and Hargreaves J. 1961 1568. 30 Hargreaves and Peacock J. 1960 1099. 31 EmelCus and Gutmann J. 1950,2115; Peacock J. 1956 167. 32 (a) Engelbrecht and Grosse J . Amer. Chem. SOC. 1954,76 2042; (b) Aynsley and Hair J. 1958 3747; (c) Wiechert Z. anorg. Chem. 1950,261 310. 33 Hargreaves and Peacock J . 1957,4390. 34 Peacock J . 1955 602. 35 Peacock J. 1957,467. 35aWei~e 2. anorg. Chem. 1956,283 377; Peacock J. 1956,1291. 378 QUARTERLY REVIEWS Robinson and his c o - ~ o r k e r s ~ ~ disputed this and isolated only ReOF and Re02F If any sexivalent rhenium were present it would have given hydrated rhenium dioxide on hydrolysis by disproportionation but only per-rhenic and hydrofluoric acids were found.Another septivalent oxyfluoride Re03F has been prepared from the corresponding chloride with hydrogen fluoride.32u or from potassium per-rhenate with iodine pentafl~oride~~~ or anhydrous hydrogen fluoride.32c It is much more stable than permanganyl fluoride. The sexivalent ReOF results when rhenium hexafluoride is reduced with rhenium ~ a r b o n y l . ~ ~ It closely resembles the analogous molybdenum and tungsten compounds. Complex Fluorides.-Neither the hepta- nor the hexa-fluoride combines with alkali fluorides in the absence of a solvent.The hexafluoride dissolved in iodine pentafluoride or liquid sulphur dioxide does react and an impure salt K,ReF has been obtained.33 Pure salts with rhenium in the sexivalent state which contained the Re02F,- anion were readily prepared by dissolv- ing per-rhenates in bromine trifluoride at room ternperat~re.~~ No doubt more oxygen could be displaced under more extreme conditions. In the corresponding reaction with permanganates the manganese is reduced to the quadrivalent state (MnF,-). A series of salts of the ReF6- ion have been made by Peacock3 who ingeniously adapted Ruff's original observation that rhenium hexafluoride oxidised potassium iodide by using sulphur dioxide as a solvent e.g. 2ReF6 + 2KI = 1 + 2KReF $ . The hexafluororhenates(1v) are described later together with the other hexahalogenorhenates.The remaining halogen compounds isolated are summarised in Table 4. TABLE 4. Halogen compounds of rhenium Oxidation state 7 6 5 4 3 1 Simple halides Oxy halides Complex anions Re03C139 ReO3Br3? ReOCl R ~ O B I ~ ~ ~ ~ 1 ~ 3 6 ReOC152- ReOBr,,- 41 Re1438 Re16e- Re(OH)CI,* - ReCI ReBr ReC14- ReCb3- Re1338 ReBr437 Rec162- Re,OClIo4- 41 ~e138 The highest authenticated chloride is the pentachloride made by'ldirect chlorination of the metal at 400-500" or by the action of carboqtetra- s6 Knox J. Amer. Chem. Soc. 1957 79 3358. 38 Peacock Welch and Wilson J. 1958 2901. a9 Wulf and Clifford J. Arner. Chern. SOC. 1957 79 4257. 40 Brukl and Ziegler Monarsh. 1933 63 329. dl Trzebiatowski and Wajda Bull. Acad. Polon. Sci. Class III 1954 2 249; Jezowski Colton and Wilkinson Chem.and Ind. 1959 1314. and Iodka Rocznicki Chem. 1939 19 187. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 379 chloride on rhenium heptaoxide at 400°.36 It can be purified by sublimation in vacuo or in a current of chlorine. In water it disproportionates to hydrated rhenium dioxide and septivalent rhenium. The existence of a tetrachloride has been disputed although the recent isolation of a tetra-bromide and -iodide indicates that the tetrachloride may exist. Croft46 claims to have intercalated the tetrachloride in graphite. The trichloride resulted when the pentachloride was heated in an inert gas. It is also formed by thermal dissociation of silver hexachlororhenate(1v) or from sulphuryl chloride and rhenium. The original magnetic measure- ments showed the solid trichloride to be diamagneti~~~ and the molecular weight in acetic acid solution corresponding to a dimer accorded with this.However a remeasurement has established the paramagnetism of the tri~hloride.~~ Recently complexes with organic ligands have been in~estigated.~~ Thus 1 1 and 1 2 complexes of rhenium trichloride and triphenylphos- phine resulted when the acetone solutions of the components were mixed. More complicated products were obtained from nitrogen-containing ligands with the exception of pyridine. For example the complex with 1,lO-phenanthroline was diamagnetic and was believed to be polymeric. Complexes with organic multiply-bonded compounds are also known.37 Rhenium tribromide is probably very similar to the trichloride and has been prepared by analogous methods.Rhenium iodides have only recently been i ~ o l a t e d . ~ ~ ~ ~ ~ They cannot be made from the metal and iodine alone or in solvents but only indirectly via complex iodides. Thus per-rhenic acid can be reduced with hydriodic acid to hexaiodorhenic(rv) acid which decomposes thermally to rhenium tetraiodide. The tetraiodide can be further decomposed; at 350° in the presence of excess of iodine to the tri-iodide; at 1 lo" if the iodine vapour is removed in a stream of nitrogen to the monoiodide. All the halides can be decomposed to the metal at higher temperatures. Oxyhalides are usually present in crude halides unless air and moisture are rigorously excluded during preparation. Per-rhenyl chloride can be made from rhenium chlorides and oxygen39 or even from rhenium disul- phide chlorine and oxygen.The corresponding oxybromide prepared from the metal mAts at 39.5°.40 Coltoii et ai. who passed oxygen over rhenium tetrabromide at 100-120" obtained the oxybromide as a liquid at room temperature^.^^ On hydrolysis it yielded only bromide and per-rhenate. The original material was probably contaminated with the dioxydi- bromide. Complex Salts.-The series M [Re(halogen),] where M is an alkali alkaline earth or co-ordinated multivalent metal ion is next in importance 42 Schuth and Klemm 2. anorg. Chem. 1934 220 193. 43 Knox and Coffey J . Amer. Chem. SOC. 1959 81 5. 44 Colton Levitus and Wilkinson J. 1960 4121. 45 Meloche and Martin J. Amer. Chem. SOC. 1956 78 5985. 46 Croft Austral. J . Chem. 1956 9 184. 380 QUARTERLY REVIEWS to the per-rhenates as a starting point for other preparations.Apart from the fluoro-salt all can be made in the wet way by reducing acid solutions of per-rhenates with such agents as hypophosphorous acid,22 chromou~~~ and titanous chlorides,20 and hydrazine.20 Alkali iodides or bromides can also be employed as reductants e.g. 2KRe0 + 16HC1 + 6KI = 2K,ReCI + 4KCl + 31 + 8H20 This overall equation hides the complexity of the reaction. The reduction goes through intermediate oxidation states because complex oxychlorides of quinquevalent rhenium can also be recovered from the eactiron mixture. The reaction can also involve a displacement of a hexaiodorhenate ion by hydrochloric acid. The equilibrium between the hexahalgenorhenates has not been measured but qualitatively the constants pertaining to the displacement cannot be large because it is possible to traverse the series HCI HBr HBr HI Rec1~~- ReBr62- though with more ease to the right.In fact the most convenient prepara- tion of the iodo-compound is by displacement of the chloro- or bromo-salt with hydriodic acid.13 The salts can also be made in a dry way for example by passing chlorine over a mixture of rhenium and potassium chloride. Ruff and Kwasnik28 claimed to have made the hexafluoro-salt using hydrofluoric acid with the potassium iodide and per-rhenate but neither Peacock nor W e i ~ e ~ ~ ~ could repeat this preparation and had to resort to dry methods. The former heated ammonium hexaiodorhenate(1v) with potassium bifluoride at 250" the latter passed hydrogen fluoride over potassium hexabromorhenate at 450". Other salts were made by pre- cipitation or ion-exchange from the potassium salt.The thermal and hydrolytic stabilities of the hexahalogeno-salts decrease from the fluoride to the iodide. The fluoride itself is stable in concentrated hydrochloric acid and can only be slowly attacked by aqueous alkali. For complete decomposition alkali fusion is required. The hexachloro- rhenates(rr1) are stable in 0.01 N-hydrochloric acid the hexabromo- compound in 2~-hydrobromic acid and the hexaiodo-compound only in concentrated hydriodic acid. Potassium chlororhenate(1v) can be distilled at 1100" whereas the iodorhenate starts to dissociate at 300". The absorption spectra of the halogenorhenates have been utilised for rhenium analysis. The salts K,[Re(OH)CI,] K,[ReOC15] and K4[Re,OC1,,] were isolated from potassium per-rhenate hydrochloric and hydriodic acid mixtures by varying the iodide concentration and temperat~re.~~ Lower Valency States Stabilised by Ligands A series of stable compounds of all the lower valency states of rhenium has only just been completed largely because of the interest shown in recent years in the ligand chemistry of transition metals in general.WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 38 1 The Cyanide Ligand.-The cyanide ion has been extensively applied as a ligand. Morgan and Davies4' made Na3[ReV0,(CN),],2H,0 by reducing a mixture of sodium per-rhenate and cyanide with hydrazine. Recently the potassium salt has been studied in It was made from K,[ReVOCl,] with an alkaline solution of potassium cyanide. The presence of four ions in the complex oxy-cyanide was indicated by the conductivity of its dilute solutions that of quinquevalent rhenium by oxidation with dichromate.Polarographic reduction gave a six-electron change in potassium chloride and a four-electron change in potassium cyanide to the same -1 and + 1 states as were obtained from potassium per-rhenate (p. 376). Chemical reduction with sodium amalgam led to the + 1 state. A complex of this state K5 [Re(CN),],3H20 was prepared similarly by reducing a mixture of potassium hexachlororhenate(1v) with potassium amalgam,49 and may have resulted from the sodium amalgam reduction of rhenium trichloride in sodium cyanide solution.50 Other examples of this state with rhenium in the cation are the salts of the [Re(CH,-C,H,.NC),]+ A complex cyanide of quinquevalent rhenium K3Re(CN),,H20 was prepared from potassium hexaiodorhenate(1v) and potassium cyanide in rnethan01.~~ The similarity of its infrared spectrum to those of the octa- cyano-molybdenum and -tungsten complexes suggested the same dode- cahedra] co-ordination.A solution of the salt could be oxidised in acid solution by air and the hexammine-cobaltic salt isolated from the solution Alternatively it could be reduced to a tervalent complex ion [ReJ'1(CN)6]3- with potassium borohydride. Arsine and Phosphine Ligands.-The ligand o-phenylenebisdimethyl- arsine (L) stabilises the 2+ 3 + and 5+ states of rhenium. The compounds [ReI1'L2(halogen) ,] +C 1 04- were prepared by reducing per-rhenic acid with hypophosphorous acid in the presence of the diarsine and appropriate halogen in methanol. Further reduction with sodium stannite produced the uncharged complexes [Re" L,(halogen),lo ; oxida- tion with chlorine yielded the quinquevalent complex [ReVL,Cl,]f All these formulae were justified by analytical and magnetic evidence.Similarly the 2+ and the 3+ state can be stabilised with tri- phenyl-p h ~ s p h i n e . ~ ~ Rhenium Carbonyls and Carbonyl Halides.-These can be regarded as stable compounds of zero- and uni-valent rhenium with direct rhenium- 47 Morgan and Davies J. 1938 1858. 48 Trzebiatowski and Danowska 2. phys. Chem. 1959 212,29. 49 Clauss and Lissner Z. anorg. Chem. 1958 297 300. Meier and Treadwell Helv. Chim. Acta 1958 38 1679. s1 Malatesta Angew. Chem. 1960 72 323. 62 Colton Peacock and Wilkinson J. 1960 1374. 63 Curtis Fergusson and Nyholm Chem. and Znd. 1958 625. 64 Freni and Valenti 1.Inorg. Nuclear Chem. 1961 16 240. 382 QUARTERLY REVIEWS carbon bonding. Druce's claim to have prepared such a bond in a trialkyl from rhenium trichloride and a Grignard reagent has been twice dis- proved. Rhenium pentacarbonyl can be made by the action of carbon monoxide on a variety of compounds (the trioxide heptaoxide heptasulphide or potassium per-rhenate) at about 200-300 atmospheres at 250". It is stable to dilute acids and alkalis. Cryoscopic measurements as well as its diamagnetism shows that it is dimeric. Two series of carbonyl halides Re(CO),X and [Re(CO)4X]255 exist. The iodide of the first series was made by the reaction 2K2Re16 + lQC0 = 2Re(CO),I + 31 + 4KI and the others by the general reaction Re + CuX + 6CO = Re(CO),X + Cu(C0)X All are colourless solids which sublime without decomposition in an atmosphere of carbon monoxide.They are soluble in hydrocarbons but not in water. The second series was formed by heating members of the first series in inert solvents 2Re(CO),X + [Re(CO),X] + 2CO The tetracarbonyl halides cannot be melted without decomposition. A whole series of substitution reactions of the carbonyl halides has been ad~mbrated.,~ The carbonyls and carbonyl halides also react with hot methanolic potash to form the salt K [Re,(CO),0,H]57. Cyclopentadienyl Compounds.-When cyclopentadienylsodium inter- acted with rhenium pentachloride in tetrahydrofuran a yellow compound was isolated which as far as elementary analysis could show was the expected rhenocene Re(C5H5)2 a member of the metallocene series which ranges from titanium to nickel.However a substance of this formula should be paramagnetic whereas it was in fact diamagnetic. The problem was resolved by nuclear magnetic resonance spectrosc~py~~ of solutions in organic solvents which showed two peaks whose areas were in the ratio of 10 1. The first peak was the expected one corresponding to the ten ring protons the second which had a large chemical shift was due to a diamagnetically shielded proton directly bound to rhenium. Similar transition metal-hydride links have been found subseq~ently.~~ The compound was therefore a univalent rhenium hydride. The hydride 56 Abel Hargreaves and Wilkinson J. 1958 3149. b7 Hieber and Schuster Z . anorg. Chem. 1956 285 265. 68 Green Pratt and Wilkinson J. 1958 3916 4315. 5 8 Lewis Sci. Progress 1961 49 67. Hieber and Fuchs 2.anorg. Chem. 1941 248 269; Hieber and Schuster ibid. 1956 287 214. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 383 behaved as a proton acceptor or base with a dissociation constant in dioxan very close to that of ammonia in the same solvent H+ OH- (C5H5),ReH + (C,H5),ReH,+ Salts of the cyclopentadienyl cation can be precipitated by large anions. A similar but non-hydridic cation with benzene ligandseo Re(C6H8),+ can also be isolated in salts. Doubtless the metallo-organic chemistry of rhenium will expand in line with work on other elements. The Re-l State The existence of a rhenide ion analogous to halide ions seemed definite if somewhat unexpected until as recently as last year. However nuclear magnetic resonanceGoa and infrared spectroscopy of rhenide solutions revealed the presence of metal-hydrogen bonds.The existence of a negative valency state became questionable because the reducing properties of solutions could be accounted for by the presence of hydrides. For simplicity however the reduced state is referred to as a rhenide. Lundell and Knowlesll first prepared a dilute solution of' potassium rhenide by passing a cold solution of potassium per-rhenate in sulphuric acid through a column of amalgamated zinc. The colourless solution which emerged could be re-oxidised to per-rhenate in accordance with Re-l + 8Fe3+ = 8Fe2+ + Re7+ The per-rhenates were also reduced to rhenide at the dropping mercury electrode (p. 376) but significantly the reduction goes further at very low concentrations. 24 A solid rhenide approximating to KRe 4H20 has been made by reducing potassium per-rhenate with potassium in wet ethylenediamine.sl Repeated extraction with isopropyl alcohol removed potassium hydroxide.The solid was slightly paramagnetic. A square planar structure which utilised dsp2 orbitals for binding the water molecules to a rhenide ion was suggested and justified by Cobble on thermodynamic grounds. He considered but rejected the possibility of a rhenium hydride.6a Other workers used lithium reduction in aqueous solution and separated the lithium hydroxide by recrystallisation from water or isopropyl alcohol. The bulk of lithium hydroxide could be precipitated as phosphate. The lithium salt in solution could be converted into the ammonium salt on cation exchange resins and then into other salts with appropriate alkalis.63 6o Fischer and Wirzmuller Chern.Ber. 1957 90 1725. 6oa Colten Dalziel Griffith and Wilkinson Nature 1959 183 1755. 6p Cobble J . Phys. Chern. 1957 61 727. Bravo Griswald and Kleinberg J. Phys. Chem. 1954 58 18. Floss and Grosse J. Inurg. Nuclear Chem. 1959 9 318; 1960 16 36. 384 QUARTERLY REVIEWS G i n ~ b e r g ~ ~ argues from the first workers' results that their reduction product which analysed to a valency state between -0.8 and -1.3 was contaminated with per-rhenate and rhenium oxides and hence the pure product would have had a state below - 1. He repeated the reduction with amalgamated zinc and obtained a + 1 &- 0.1 state but with sodium amalgam a -2 to -3.5 state. Potassium rhenide was then prepared more carefully to yield a white diamagnetic solid reputedly free from alkali and per-rhenate.The oxidation state determined with calcium hypochlorite in alkali was -3-4(5). This solid was treated with excess of sulphuric acid and from a potassium analysis the acid consumed the hydrogen evolved and the rhenium formed the following equation was deduced K,ReH + 3H+ = 3K+ + Reo + 5H2 The empirical formula was doubled to avoid the paramagnetism which would arise from quadrivalent rhenium and the water present being allowed for the solid had the composition K6Re,H,,,6H20. Nuclear magnetic resonance spectra again provided definite evidence of a rhenium hydride by the presence of a small pesk to the high field side of the main water proton peak.64 The ratios H-/H+ = 2 and H-/Re = 6.5-7.3 were deduced and agreed with the formula suggested by chemical analyses.65 Subsequent work has not lessened the confusion.Ginsberg now believes that the solid isolated from the reduction of potassium per-rhenate in wet ethylenediamine is anhydrous K,ReH with the rhenium in a -5 state. His earlier material appears to have been contaminated with carbonate and per-rhenate. Floss and Grosse however having re-investigated their product assign to it the formula KReH with a variable amount of water of crystallisation. Further work is obviously needed to see whether separate hydrides exist. Nevertheless there is now little doubt that the existence of a simple rhenide ion analogous to halide ions is untenable. Rhenium Compounds with Metallic Bonding Rhenium forms compounds with metals and some non-metals in which the bonding is different from that considered previously.In particular this hexagonally close-packed metal tends to form intermetallic phases of wide ranges of composition and more complex structures with transition elements of Groups IV V and VI,66 e.g. ReZr, Re3Cr2 ReMo with the same structure as p-uranium or Re2,Zr and R,,Nb with the a-manganese structure. Sili~ides,~ borides,68 a germanide,69 and a nitride70 are known 64 Ginsberg Ph.D. Thesis Columbia {Jniversity 1959 and private communication. 65 Ginsberg Miller Cavanaugh and Dailey Nature 1960 185 528. " Ageev Doklady Akad. Naitk S.S.S.R. 1959 129 559. 67 Searcy and McNees J. Amer. Chem. SOC. 1953,75,1578; 1955,77,5290. 68 Searcy and McNees J. Amer. Chem. SOC. 1954,76 5287; Aronsson Stenberg and 69 Neshpor Paderno and Samsonov Dokludy Akad. Nauk S.S.S.R. 1958,118,515.'O Hahn and Konrad Z. anorg. Chem. 1951,264 174. Aselius Acra G e m . Scand. 1960 14 733 1001. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 385 but contrary to earlier rhenium bonding also occurs in rhenium dioxide and rhenium carbonyl. a carbide could not be detected.13 Direct Comparative Chemistry of Rhenium Stable compounds containing rhenium in the valency states from zero to seven have now been isolated. However some of these states are only accessible because of ligand stabilisation and the more normal chemistry of the oxygen and halogen compounds of rhenium shows the comparative instability of the uni- bi- and quinque-valent states. The last tends to disproportionate to the more stable quadri- and septi-valent states in aqueous media and sometimes in the solid. This disproportionation of a quinquevalent state has analogies in the preceding group of the Periodic Table (e.g.S2F1,, UCl,). Comparison with the Neighbowing Elements.-The significant difference between rhenium and manganese is the great stability of the latter's bi- valent state even in salts of oxy-acids. The usual argument forwarded for the stability of the Mn2+ ion is that it corresponds to a stable half-filled electronic level 3d5 yet the 5d5 configuration of Re2+ is unstable. The counter argument that the stability of a half-filled level decreases as the atom becomes larger does not appear cogent when the ability of elements on either side of gadolinium to attain the half-filled gadolinium level 4f ' is considered although the tendency is becoming apparent in the actinide series. A h r e n ~ ~ ~ believes that bivalent manganese is the more stable because the gap between the third and the second ionisation potential of manganese is greater than the difference between the rhenium potentials i.e.it is easier to oxidise bivalent rhenium. However his thesis even if true does not account for the stability of bivalent manganese. It leaves us to explain the differences in ionisation potentials and furthermore any conclusions would only apply to ions in the gas phase. Considering the transition metals as a whole we can see the tendency for the higher valency states to be more stable both in the earlier members of the series and the later members of the groups in the Periodic Table. Qualitatively this tendency can be associated with the bonding between the atomic nuclei and their valency electrons and quite simply the stability of the higher valency states of rhenium compared with those of manganese results from the fact that the valency electrons of rhenium are farther from the nucleus and are less tightly bound.Even in the quadrivalent state in which the two elements show the greatest resemblance the tendency for Mn*+ to behave as an oxidising agent is very marked. The differences between rhenium and technetium are of degree rather 'l Trzebiatowski 2. anorg. Chem. 1937 233 376. 72 Ahrens J. Inorg. Nuclear Chem. 1957 4 264. 386 QUARTERLY REVIEWS than kind. It should be remembered that all the isotopes of the latter are radioactive and this very instability may influence chemical kinetics especially in solid-state reactions. Quantitatively the differences can be shown on a potential diagram,73 which refers to acid solutions; the techne- tium potentials are in parentheses 0.510 (0.738) Re Oa2'0 (0'272) RLO 0.385 (0-8) ReO 0.768 (0.7) RL0,- i 0-367 (0.472) - 1 - Thus both the sexi- and septi-valent state of technetium are more easily reduced to the quadrivalent state in acid solution.The difference in behaviour of per-rhenate and pertechnetate on polarographic r e d u c t i ~ n ~ ~ in which only a quadrivalent technetium state is reached in alkaline solution could also be correlated with the potential diagram although the polarographic waves are irreversible. When rhenium is compared with the Group VIII metals a greater resemblance is seen between the chemistry of rhenium and ruthenium than that of the neighbouring osmium.Thus ruthenium forms a stable tribromide and trichloride but with osmium the tetrahalides are more stable; the dioxide of ruthenium is more stable than that of osmium and the complex halides of ruthenium are reminiscent of the rhenium com- plexes e.g. RuC~,~- RuCI,,- [RuCI,(OH)]~- (Ru,OC~,,)~-. Similarly one can present evidence for a greater resemblance between rhenium and uranium or molybdenum than between rhenium and tungsten. Thermochemistry.-The measured heats of formation and free energies are collected in Table 5. In the gaseous rhenium heptaoxide molecule which is known to be monomeric if tetrahedral rhenium co-ordination is TABLE 5. Heats of formation and free energies of rhenium compounds Compound - Re02 ReO 74 Re207 74 HReO (a 74 KReO ReS 75 ReCl 76 ReBr 76 ReSi 67 ReF6 28 - AH",, (kcal/mole) 101 146 296 189 263 46 63 39 275 dG020s 89 127 259 167 239 44 48 34 252 55 73 King and Cobble J.Amer. Chem. SOC. 1957 79 1959; Cartledge and Smith 74 Boyd Cobble and Smith J. Amer. Chem. SOC. 1953 75 5783 578. 76 Juza and Biltz 2. Efecktrochem. 1931 37,498. 76 King and Cobble J. Amer. Chem. SOC. 1960,82 2111. J . Phys. Chem. 1955,59 1111. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 387 assumed there are eight rhenium-oxygen bonds and an average rhenium- oxygen bond energy of 139 kcal./mole [(heat of formation 296 + heat of sublimation 34 + heats of atomisation 2 x 183 + 7 x 59-1) f 81. This can be compared with the rhenium-fluorine bond energy of 96 kcal./mole in rhenium hexafluoride. We can compare the oxides by calculating the energy required to produce an atom of oxygen and leave rhenium in the metallic state.The values are close for the di- and tri-oxides (110 and 108 kcal./g-atom of oxygen respectively) and are significantly greater than for the heptaoxide (101 kcal/g. atom). Structural Chemistry.-This chemistry is based mainly on tetrahedral and octahedral co-ordination around rhenium. Rhenium trioxide has a structure consisting of octahedra of oxygen with rhenium at the centre linked at every corner.77 This structure is closely related to the perovskite structures (M*ITiIVO,) which have additional metal ions at the centre of the unit cells. Recently similar phases [Ba2M11]ReV10 and Ba[M~,.,ReVIJ,.,]O, where MI and MI1 are uni- and bi-valent metals with interesting variations in their magnetic properties have been studied.78 Rhenium dioxide is similarly built from oxygen octahedra but in this struc- ture they share two opposite sides and the remaining two corners to give a very distorted rutile (TiO,) structure with alternately short and long Re-Re distances.79 No single value for a Re1" ionic radius can be given for such a structure.An orthorhombic modification is formed by annealing above 1050°.80 Rhenium heptaoxide has an orthorhombic lattice (a = 1525 b = 5.48 c = 12.5 A) with eight molecules in a unit ce1181 but the structure has still to be determined. The thermochemical data already mentioned indicate that the bonding is not the same as in the di- and tri-oxides and this is further indicated by its high solubility in water as well as organic solvents and its high vapour pressure. A molecular lattice with tetra- hedrally co-ordinated rhenium within the molecules is probable.Tetra- hedral co-ordination occurs in solid per-rhenatess2 and probably in the per-rhenate ion in solution because the Raman spectra of the isoelectronic triad OsO, ReO,- and W042- are so similar.s3 Per-rhenyl chloride and fluoride are also tetrahedral.42,84 There is an interesting shortening of the Re-0 bond length in the latter which is well outside the error of microwave spectroscopic measurements and again no single value can be assigned to a covalent tetrahedral radius of rhenium. The rhenium-oxygen system is quite different from the molybdenum- ?? Meisel 2. anorg. Chern. 1932 207 121. Ward and Longo J. Amer. Chem. Soc. 1960 82 5985; Sleight and Ward ibid. 7g MagnCli and Andersson Acta Chem.Scand. 1955,9 1378. * O MagnCli Actu Cryst. 1956 9 1038. 8a Broch Z.phys. Chem. 1929,136,22; Pitzer 2. Krist. 1935,92 131. 83 Woodward and Roberts Trans. Furaday Soc. 1956 52 615. 84 Lotspiel Diss. A h . 1958 19 340. 1961,83 1088. Wilhelmi Acta Chem. Scund. 1954 8 693. 388 QUARTERLY REVIEWS and tungsten-oxygen systems which contain a plethora of intermediate phases between their di- and tri-oxides. This is probably because of the volatility and stability of rhenium heptaoxide. Any intermediate oxide is able to disproportionate into the dioxide and heptaoxide. Deschanvres,26 however claims to have prepared a non-stoicheiometric dioxide rich in oxygen by thermal decomposition of ammonium per-rhenate at 350- 450" which can be converted into the orthorhombic dioxide by annealing at 600".The excess of oxygen is removed as heptaoxide. TABLE 6. Bond lengths and co-ordination in rhenium compounds Bond ReO Re-0 in Re03 Re04- Re-0 in Re03CI Re-0 in Re0,F Re-F in ReF Re-CI in Re0,CI Re-F Re03F Re-Br ReRr62- +(Re-Re) Re,(CO), * Estimate only. ReF62- ReC1,2- metal Ref. 79 80 77 82 39 84 86 35a 39 84 85a 85b 16 87 Rhenium co-ordina- tion 6 6 4 4 4 6 6 4 4 6 6 12 6 Bond Ionic radii length - (4 - 1.87 Rev* 0-55 1.76 1 6 6 1*93* 2.02 Re'" 0-66 2.23 1 -86 2.37 RelV 0.56 2.50 Re1" 0- 5 5 1.37 1-51 - - Octahedral co-ordination occurs in the hexahalogenorhenate~(rv),~~~~~ rhenium hexafluoride,86 and in the carbonyl. In the last molecule there is a direct Re-Re bond which is longer than in the metal itself and the octahedra are inclined at 45".87 The bond lengths found in various structures (Table 6) enable some radii for rhenium in different valency states to be deduced.By subtracting the covalent bond radii of fluorine chlorine and bromine from the bond lengths of the hexahalogenorhenates(1v) values of 1.36 1.38 and 1-38 A respectively are obtained for the covalent octahedral radius of rhenium. This is very close to the metallic radius for twelve-fold co-ordination (1.37 A) but the agreement is not so convincing if one uses the experi- mentally determined fluorine radius (0.72 A)88 which reduces the first value to 1.30 A. The ionic radius obtained from the hexafluororhenate ion also differs 85 (a) Aminoff 2. Krist. 1936 A 94 246; (b) Templeton and Dauben J. Arner. 86 Gaunt Trans. Furday SOC. 1954,50 209. 87 Dahl Ishishi and Rundle J. Chem. Phys. 1957 26 1750.Chem. SOC. 1951,73,4492. Ref. 16 p. 228. WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 389 from the values obtained from the chloride and bromide. The first value can be accepted as being more nearly an idealized RelV radius and it does conform with the other radii including that of Rev which because of the isomorphism between salts containing the SbF,- and ReF,- ions,35 must be close to that of SbV (0.62 A). Ahrens’s attempts9 to calculate ionic radii empirically from ionisation potentials which yielded 0.56 and 0.72 for Re7+ and Re4+ respectively is evidently more optimistic than successful. His value for Re7+ is also unlikely because it is so close to the value at which tetrahedral co-ordination can change to octahedral (limit- ing radius ratio 0.414 A). Magnet0chemistry.-The interpretation of magnetic data obtained from a variety of rhenium compounds affords a good test of modern theorys0 as well as distinguishing between alternative structures.In the rhenium atom the valency electrons are 5d56s2 the d electrons having parallel spins and separately occupying orbitals of equal energy. When ligands are placed around the ionised rhenium atom the d orbitals are no longer of equal energy. For octahedral co-ordination three orbitals (dJ have lower energy than the remaining pair (dJ whereas for tetrahedral co- ordination the order is reversed. The energy difference is greater for octahedral than for tetrahedral co-ordination and also for ions of the third transition series compared with the preceding series. The large energy separation tends to favour spin pairing as against electron promo- tion to the higher level in octahedral rhenium complexes with four to six d electrons.The stereochemical consequences expected from the pairing of these non-bonding d electrons are given in Table 7. The ideal moment is not achieved in all of these examples because of competing interactions such as spin-orbital coupling antiferromagnetism and temperature- independent paramagnetism. Magnetically concentrated solids are also exceptional. It should be noted that in all the rhenium(v) complexes examined there are deviations from ideal behaviour. Spin orbital coupling could reduce the moments to about 1.2-1.5 but for very low values polymeric structures need to be postulated. The tetrahedral co-ordination for du4 has not been fully substantiated but the measured diamagnetism is a strong indication of this geometry.The diagnostic value of magnetic measurements is well illustrated by the diamagnetism of cyclopentadienylrhenium hydride which provided an essential clue to its structure.58 Other examples are provided by the coupling of paramagnetic units to produce diamagnetic molecules e.g. rhenium carbonyls7 and complexes containing the ReIV-O-ReIV unit41 complexed with halides or organic acids. 89 Ahrens Geochim. Cosmochim. Actu 1952,2 155. 90 Gillespie and Nyholm Quart. Rev. 1957 11 339; Figgis and Lewis in “Modern Q1 Colton Levitus and Wilkinson J. 1960 5275. Coordination Chemistry,” ed. by Lewis and Wilkins London Interscience 1960. Ref. to magnetism 52 52 48 48 53 35 41 35a 35a 35a 52 53 44 44 91 91 53 49 51 TABLE 7.Con3guration and magnetic moment of some rhenium complexes Non-bonding electrons Complex Rew(CN) ReV(CN),3- ReVOC1,2- ReV0,(CN)43- ReVL2C12+ ReIV( OH)Cl ReVF6- Re'"Cls2- ReNBr6 2- RerVF62- ReII*(CN)6 3- Re"IL2Cl,+ ReIII(Me.CO.CH = COMe) Re"1Cl3 2PPh3 Re"C1,- Re1Wl2 NMe2CS2 Re"L2C120 ReI(CN):- Re'(Me-C,H,-NC) L = o-Me2AsC6H4.AsMe * Partial X-ray or spectral confirmation of structure available. Expected co-ordination Dodecahedral( ?) Dodeca hedral* Octahedral 9 9 9 7 9 9 9 9 9 7 * * * 9 9 9 9 9 9 9 9 9 9 Trigonal bipyramidal Tetrahedral Octahedral 9 9 * 9 9 9 9 w \o 0 Moment Ideal (spin Obs. only) 1 -73 0 2.83 2-83 2.83 2.83 3.87 3.87 3.87 3-87 2.83 2.83 2.83 0 0 0 1.73 0 0 2.0 0 0.48 0 -0.7 1.55 a 2.9 2 3.6 2 3.7 1.8-2.1 m 2.3 2 0 0 1.7-1.8 "3 0 0 3 WOOLF AN OUTLINE OF RHENIUM CHEMISTRY 39 1 Applications It might reasonably be asked why so much attention has been devoted to rhenium the rarest of elements over the last few years.First rarity need not be equated with expense and in fact rhenium is less expensive than some of the platinum and rare-earth metals because of its ready extraction from ores. Secondly interest in high-temperature materials in diverse technical applications has prompted many investigations involving rhenium and this in turn has tended to increase the supply and availability of the metal. The chemist is mainly interested in rhenium as another transition metal of variable valency whose properties and stabilities in different states have to be incorporated into present-day theory. The metal and its compounds have found specialised outlets in the same way as the platinum metals.The metal itself is applied as a protective coating or as an inert foil or filament in closed systems but as it is difficult to fabricate workable alloys have been developed. These find advantageous application in thermocouple 92 contact and heating elements. Chemically rhenium and its compounds are most efficient catalystsg3 for hydrogenation dehydrogenation dehydration isotopic exchange and even oxidation reactions. Rhenium sulphides are especially useful for the first two. For example rhenium heptasulphide can selectively hydrogenate multiple-bond systems without hydrogenolysis of carbon-sulphur links in the same The halides are potential but as yet unexploited catalysts. It is conceivable that in the future the rsducing power of rhenium hydride compounds will provide a further contribution to the inorganic arsenal of the organic chemist. I thank J. Lewis and R. Peacock for helpful criticism of the original manuscript various authors for providing more information on their work and T. E. Allibone C.B.E. F.R.S. Director of the Laboratory for permission to publish this review. Oa Hasse and Schneider Z . Phys. 1956 144 256. e3 Tropsch and Kassler Ber. 1930 63 2149; Platonov Zhur. obshchei Khim. 1941 O4 Broadbent Slaugh and Jarvis J. Arner. Chem. Sac. 1954 76 1519. 11 590.

ISSN:0009-2681    

DOI:10.1039/QR9611500372

出版商:RSC    

年代:1961

数据来源: RSC