摘要:

~~~ QUARTERLY REVIEWS ~~~~~~ NITRONES By G. R. DELPIERRE* and M. LAMCHEN (UNIVERSITY OF CAPE TOWN) A. Introduction THE name “nitrone” a contraction for nitrogen ketone was suggested by Pfeiffer,l in 1916 for compounds possessing the group their striking similarity with ketones. (I) to emphasise t-0- x\ Y’ (4) We now recognise that the analogy rests on the mesomeric effects [ ( l ) f 3 ( 2 ) and (3)++(4)] which predominate in both classes of com- pounds making the nitrone or azomethine N-oxide group (1) behave as an extended carbonyl function. General usage and Chemical Abstracts have given a certain connotation to the term “nitrone” but the term is in need of a clearer definition. The mere presence of the azomethine N-oxide group (:C=?l-O) in a mole- cule does not necessarily endow it with nitrone characteristics any more than the presence of a carbonyl group confers ketonic properties to the wide variety of carbonyl compounds.The polarisation of the azomethine N-oxide group is controlled by the electronic effects of the substituents X Y and R. On the one extreme is the isolated system represented by the structure (1; X Y and R = alkyl) in which the double bond is fixed and the positive charge is localised between the nitrogen and carbon atoms of the azomethine system. On the other extreme is pyridine N-oxide (9 in which a high degree of delocalisation results from its aromatic character. In between these extremes all degrees of delocalisation could exist. The positive charge on the azomethine carbon atom controls most typical nitrone reactions and it must be decided whether any positive charge on the carbon atom however small entitles the compound to be included under nitrones or at what stage delocalisa- tion is too great to warrant inclusion in this class.The authors prefer to restrict the term “nitrone” to compounds in which the canonical forms (1) and (2) contribute most to the structure and to exclude all azomethine N-oxides in which considerable delocalisation *Present address Yale University New Haven Connecticut. P. Pfeiffer Annalen 1916,411,72. 329 1 330 QUARTERLY RE VIEWS of the positive charge occur. We would thus exclude compounds in which the C=N of the azomethine N-oxide forms part of an aromatic ring e.g. pyridine N-oxide ( 5 ) quinoline N-oxide etc. Compounds in which the nitrogen of group (1) is attached to an atom with unshared electron pairs e.g.isoxazoline 2-oxide (6) and furoxan (7) are also excluded since the electronic shifts shown will greatly diminish the contribution of the canonical form (2). If R = H the group (1) would represent the unstable tautomer of the oximes and if X or Y were either a hydroxyl group or an amino-group the structure (1) would be tautomeric with hydroxamic acids (8) or N-hydroxy- amidines (9).2 On the other hand the great stability of aromatic systems can be expected to resist loss of aromaticity even when conjugated to group (1) and so should not cause much delocalisation of positive charge. Com- pounds such as 3,4-dihydroisoquinoline N-oxide (10) and C-phenyl-N- methylnitrone (1 1 R = Me) are thus considered as drones. This view- point is strengthened by the fact that the compound (10) undergoes the typical nitrone reaction 1,3-cycloaddition [see Section D 4 c (iv)] 40,000 times faster than the heteroaromatic N-oxide (12).3 mp P hCO .CH = $ Ph (13) O’ Y (14) 0- It also follows from this that electron-attractin groups in either X o Y will decrease the electron density on the carbon and its electrophilic properties will be enhanced. This is exemplified by the fact that C-benzoyl- N-phenylnitrone (1 3) undergoes 1,3-~ycloaddition 1 10 times faster than CN-diphenylnitrone (1 1 R = Ph).4 Isatogens e.g. 2-phenylisatogen (14) similarly show nitrone properties and are considered cyclic C-a~ylnitrones.~ Compounds containing the group (1) have been known since the pre- paration of the first “N-alkylated oxime” to which Dittrich5 assigned the oxaziridine structure (1 5).During the latter part of the nineteenth century a G. D. Buckley and T. J. Elliot J. 1947 1508. R. Huisgen Angew. Chem. (Internat. Edn.) 1963 2 633. R. Huisgen Angew. Chem. (Internat. Edn.) 1963,2 565. M. Dittrich Bet-. 1890 28 2606. DELPIERRE AND LAMCHEN NITRONES 331 and the early years of the twentieth century these compounds played an important part in the stereochemical studies of the oximes. The three- membered ring structure (1 5) was discarded since no real evidence could be offered for its existence and also since stereochemical evidence weighed heavily against it. Thus Lindeman and Tschang6 failed to resolve the bromocamphorsulphonate of N-methyl-p-dimethylaminobenzaldoxime which on the basis of the oxaziridine structure should possess an asym- metric carbon atom.The discovery and optical resolution of the true oxaziridines7 have shown beyond doubt that the so-called “oxime N- ethers” were in fact true nitrones. Modern methods notably ultraviolet infrared and nuclear magnetic resonance spectroscopy clearly established the structure (1) in nitrones. A great number and wide variety of nitrones have been prepared and the field was reviewed by Smith8 in 1938 and again by Hamer and Macaluso in 1964.9 This review will thus not attempt complete coverage of the field but rather to focus attention on some interesting aspects. B. Nomenclature In Chemical Abstracts the carbon substituents [X and Y in formula (I)] are prefixed by a and the substituent on nitrogen [R in formula (l)] is prefixed by N . In other publications e.g.Angewandte Clzernie the carbon substituents are prefixed by C instead of a. I.U.P.A.C. considered nitrones as N-alkylated oximes and recommended naming them accordingly. The authors prefer and have used in this review the system of naming these compounds as C- and N-substituted nitrones. Thus compound (1 1 R = Me) is named C-phenyl-N-methylnitrone. Cyclic nitrones are usually named as the oxides of the parent heterocycle thus the compound (1 6) is 2,4,4-trirnethyl-l-pyrroline 1-oxide while the structure (1 7) is 3,4,5,6-tetrahydropyridine 1-oxide (or dl-piperideine N-ox i de) Occasionally the terms “aldonitrones” (one C-substituent = H) and “ketonitrones” (neither C-substituent = H) are used. C. Syntheses A number of routes to nitrones have been developed; the most im- portant of these are listed as follows.H. Lindeman and K. T. Tschang Ber. 1927,60 B 1725. W. D. Emmons J. Amer. Chem. SOC. 1956,78 6208; 1957 79 5739. J. Hamer and A. Macaluso Chem. Rev. 1964,64,473. a L. I. Smith Chem. Rev. 1938 23 222. 332 QUAR'IERLY REVIEWS 1. From N-Substituted Hydroxy1amines.-Analogous with the well known formation of oximes from aldehydes and ketones by their reaction with hydroxylamine reaction with N-substituted hydroxylamines gives nitrones. This method is of fairly general application and has been used to prepare a wide variety of nitrones generally in good yields. Many instances are however reported especially with aliphatic ketones having bulky alkyl groups where no nitrones are formed. Further in many cases dimeric products are produced as in the case of the reaction product of acetone and phenylhydroxylamine for which the structure (18) has been pro- posed.lo Generally speaking the condensation proceeds more readily with aldehydes than with ketones though some aldehydes are susceptible to polymerisation reactions under the alkaline conditions usually employed.Often the hydroxylamino-function is generated in situ usually by zinc dust-reduction of a nitro-compound;2s11 in other cases the aldehyde or ketonic function is liberated in situ from the bisulphite complex12 or from the a~eta1.l~~ When the hydroxylamino-group and carbonyl function are suitably placed in the same molecule condensation to produce cyclic nitrones occurs with great ease. Controlled reduction of y-nitro-aldehydes or y-nitro-ketones generally by the use of zinc dust and aqueous ammonium chloride solution give y-hydroxylamino-aldehydes or -ketones which cyclise to 5-membered cyclic nitrones in good F'h6=C(Me)CH;CMe 2 1 .NPh A- OH t- 1 yields.13 (1 8) lo F.H. Banfield and J. Kenyon J. 1926 1612. l1 E. Beckmann Annalen 1909 365 201 ; F. C. Brown V. M. Clark and Sir Alex- ander Todd Proc. Chem. Soc. 1957 97. l2 P. Grammaticakis Compt. rend. 1947 224 1568. lS (a) R. Bonnett R. F. C. Brown V. M. Clark I. 0. Sutherland and Sir Alexander Todd J. 1959 2094. (6) R. F. C. Brown V. M. Clark A. Giddey and Sir Alexander Todd J. 1959 2087. DELPIERRE AND LAMCHEN NITRONES 333 To avoid polymerisation the nitro-aldehyde (19) can be converted to an acetal e.g. a dioxalan (21) before the alkaline reduction of the nitro- group. Mild acid treatment will then remove the protecting group and on basification cyclisation occurs.Although the above method is probably the one of choice availability of the starting material may be limited or steric effects may prevent easy formation of the nitrones. It is thus necessary to study other synthetic methods. 2. From NN-Disubstituted Hydroxy1amines.-A wide variety of reagents can be used to oxidise both acyclic and cyclic NN-disubstituted hydroxyl- amines to nitrones. This method is only applicable when at least one of the carbon atoms attached to the nitrogen carries a hydrogen atom ~ 2 2 ) -+ (2311. In some cases it suffices merely to pass oxygen (or air) into an aqueous solution of the hydroxylamine provided that a suitable catalyst such as the copper-ammonia complex is pre~ent.'~J~ Other oxidising agents which have been used successfully include mercuric oxide,15 cuperic acetate,la hydro- gen peroxide," potassium ~ermanganate,'~ potassium ferricyanide,12 t-butyl hydroperoxidels and high-potential q~in0nes.l~ In some cases the course of the oxidation depends on subtle steric effects; thus oxidation with mercuric oxide of 1 -hydroxy-2-phenylpiperidine (24) gave the dimer of the nitrone (25) in which the nitrone group is not conjugated with the aromatic ring,20 whilst the same reagent oxidised 1-hydroxy-2-phenyl- pyrrolidine (26) to the nitrone (27) in which the nitrone group entered into conjugation with the phenyl 1 (24) (2 5) (26) (27) The unexpected production of the intermediate (25) was explained on conformational grounds;20 in the most stable form of the piperidine (24) l4 D.H. Johnson M. A. T. Rodgers and G. Trappe J. 1956,1093. I6 J. Thesing and W. Sirrenberg (a) Ber. 1959 92 1748; (6) 1958 91 1978; (c) L l7 G. E. Utzinger Annalen 1944,556 50. 19 W. D. S. Bowering V. M. Clark R. S. Thakur and Lord Todd Annalen 1963 *O J. Thesing and H. Meyer Annalen 1957 609,46. AIIessandri Chem. A h . 1915 9 1045. H. Rupe and R. Wittwer Helv. Chim. Actu 1922 5 217. H. E. de la Mare and G . M. Coppinger J. Org. Chem.. 1963 28 1068. 669 106. 334 QUARTERLY REVIEWS the phenyl and hydroxyl substituents would be equatorial and in this form the hydrogen atom at position-2 is axial and less readily attacked than the equatorial hydrogen at position-6. These distinctions are absent in the pyrrolidine (26). On the other hand no explanations have been offered for the fact that while oxidation of l-hydroxypiperidine (28) with cupric acetate yields the dimer (30)21 of the nitrone (29) oxidation with potassium ferricyanide gives a trimer formulated as (3 1).20 [Qj (29) (30) Whilst the dimer (30) is a typical 1,3-cycloaddition product of the expected nitrone (29) which is thus obviously the intermediate the trimer (31) would represent a 1,Zaddition of this nitrone and is contrary to reactions normally found for the nitrones.However the dimer (30) which forms under slightly acid conditions can be converted to the trimer by acid hydrolysis and addition of excess strong base. It thus appears that the reaction is pH dependent and that the catalyst is not the controlling factor. 3. From 0ximes.-The alkylation of oximes usually gives mixtures of the oxime-O-ethers (32) and nitrones.The particular isomer formed or its preponderance in the mixture depends on the nature of the oxime the alkylating agent and experimental conditions. Ra Ra (32) Thus methyl bromide and benzyl bromide give different ratios the reagent with the small methyl favouring nitrone formation whilst the reagent with the larger benzyl favours ether formation. Also the higher the dissocia- tion constant of the oxime the higher the proportion of O-ether pro- duced.22 As a general method of nitrone synthesis this method thus suffers from notable shortcomings. 4. From Aromatic Nitroso-compounds.-Aromatic nitroso-compounds react with a variety of reagents to give nitrones and this reaction has proved a fruitful route to many nitrones. The reagents which have been found effective can be divided into the following categories 21 J.Thesing and H. Mayer Bey. 1956 89 2159. 22 0. L. Brady and R. F. Goldstein J. 1926,2403; 0. L. Brady and N. M. Chokshi J. 1929 2271. DELPIERRE AND LAMCHEN NITRONES 335 (a) Reagents containing active methyl methylene or methylidene groups. Essentially the reaction involves a base-catalysed addition of the reagent to the nitroso-group to yield the NIV-disubstituted hydroxylamine (33). The fate of this product may fall along one or both of two paths either 1,Zelimination to give the ailil(34) or 1,3-elimination to form the nitrone (35)- Y base Y-tr' Y Ar-7-H Ar-7 + N - Ar-F-Y-A'I X x x 0- Y If Y = H this elimination is an oxidation the excess of nitroso-compound acting as oxidising agent. The most notable example of the above reaction is the so-called Krohnke ~ynthesis,~~ which makes use of pyridinium salts as active methylene reagents (X = H Y = N+C5H5 Cl-).Other reagents used include e.g. 2,4-dinitrotoluene (X = Y = H);24 p-nitrobenzyl chloride (X = H Y = C1);25 a-chloro-a-cyanotoluene (X = CN Y = C1);2s 1,3-diketones (X = H Y = C0.R).27 (b) Reagents containing an electropositive atom or group which acts as a good leaving group. Generally this reaction can be represented as follows + Reagents used include diphenyldiazomethane [X+= (N EN)],^* sul- phonium ylids [$+ = (S.Me2)],29 which gave good yields a d phosphorus ylids [X+ = (P.Ph3)],30 which gave variable amounts of anils as by- products. Other reagents like alkenes alkynes quinones etc. have been employed but are not generally used and the mechanisms are obscure.D. Properties 1. Solubilities-Nitrones may be either liquids or solids and due to the polar character of the nitrone group they tend to be readily soluble in OS F. Krohnke Angew. Chem. 1953,65 612; 1953,75 181. 24 J. Tananescu and J. Nanu Ber. 1942 75 650. *6 F. Barrow and E. D. Griffiths J. 1921 212. ee F. Barrow and F. J. Thorneycroft J. 1939 773. 27 A. Schoenberg and R. C. Azzam J. 1939 1428. 48 A. W. Johnson J . Org. Chem. 1963,28 252. 30 S. Trippett Quart. Rev. 1963 17 406. A. W. Johnson and R. B. La Count J . Amer. Chem. Suc. 1961,83,417. 336 QUARTERLY REVIEWS water unless hydrophobic substituents such as aryl groups are present. Extraction from aqueous solution by organic solvents is thus usually only possible after concentration and salting out.2. Spectral Properties.-(a) Ultraviolet Spectra. The ultraviolet spectra of a great number of nitrones have been reported. Since much of the work has been carried out on aryl-substituted nitrones confusion seems to have arisen as to the origin of the absorptions. A strong absorp- tion in the region of 230 mp which is present in all monomeric unconju- gated mono-nitrones even those with no plienyl substituents such as the 1-pyrroline 1-oxides must be due to electronic transitions (E-band) in the nitrone group and cannot be assigned to electronic transitions in the benzene nuclei as suggested by Wheeler and Gore.31 The position of the absorption maximum may be shifted when the nitrone function is con- jugated to phenyl groups and other E- and K-bands due to other chro- mophores and their combined effects may occur.31 In the conjugated dinitrones bathochromic shifts to 331 mp were observedq2 (see also Section E).(b) Infrared Spectra. A strong absorption in the 1600 cm.-l region is characteristic of all nitrones but the exact position of this band varies with different nitrones. In the 1 -pyrroline 1-oxides with position-2 unsub- stituted the absorption is usually in the range 1570-1590 cm.-l whilst in the 2-substituted 1-pyrroline 1-oxides the position of the band is some- what higher and generally lies between 1600 and 1620 cm.-l.13 In acyclic nitrones the position of this band varies over a slightly wider range de- pending on the substituents present. The exact origin of this band is not absolutely clear; some authors15a attribute it to the C=N stretching mode consider it as due to the C=N+-0- group as a whole.Another strong absorption in the region 11 70-1280 cm.-l is found in all nitrones and since a similar band is present in pyridine N-oxide and tri- methylamine N-oxide it must be considered to be due to the N+-0- stretching frequencies. (c) Nuclear Magnetic Resonance. The nuclear magnetic resonance spectra of a few nitrones are known. Two valuable contributions to the study of nitrones can be derived from them. (i) In the 1-pyrroline 1-oxides the double bond is localised in the A'-position since no signal in the ole- finic proton region was observed for the nitrone (16) and both possible tautomers would have olefinic protons. Similarly for the nitrone (20) only one proton was found to absorb at low field (proton at C-2); the A2- tautomer would be expected to show two olefinic proton signals.33 (ii) gem-Dimethyl groups at position-5 in nitrone (20) give a singlet at about T 8.7 which confirms that the adjacent O-v=C group is planar and not *l 0.H. Wheeler and P. H. Gore J. Amer. Chem. SOC. 1956,78 3363. 32 R. F. C. Brown V. M. Clark M. Lamchen and Sir Alexander Todd J . 1959 2116. R. Bonnett and D. E. McGreer Canad. J. Chem. 1962,40 177. DELPIERRE AND LAMCHEN NITRONES 337 in the oxaziridine form; the isomeric oxaziridine gives the expected two singlets for the two methyl groups cis and trans to the three-membered 3. Isomerism.-In cyclic nitrones the syn-form is dictated by the ring but in acyclic compounds the double bond in the nitrone group would introduce the possibility of geometrical isomers.However since resonance imparts a considerable amount of single-bond character to the system one might expect a fairly ready isomerisation between the syn- and anti-forms. It has been found that heat easily converted a-aryl-wcyano-N-arylnitrones from the syn- to the anti- form. A number of geometrical isomers of such a-cyano-nitrones have been prepared and their configuration assigned on the basis of dipole measurements.26 The presence of the polar cyano- group offers an easy way of determining the configuration of the nitrone for when the cyano-group is on the same side as the polar Nf-0- bond high dipole moments result while the opposite anti-configuration has a lower dipole moment. At least one attempt has been made to determine the geometry of nitrones by making use of ultraviolet spectroscopy.The nitrone (10) with the phenyl group and the oxygen atom fixed in the anti-configuration absorbs at Amax 304 228 and 211 mp while the nitrone (27) absorbing at Amax 288 221.5 and 205 mp is rigidly in the syn-configuration. On this basis the geometry of both the nitrones (11 R=Me) with Amax 288 221.5 and 206 mp and (11 R=cyclohexyl) with Amax 291 223 and 206 mp was assigned as ~ y 2 . l ~ ~ The validity of such studies should not neces- sarily be regarded as having been established. ring,33,34,35 4. Reactions.-It has been shown (page 329) that the nitrone group could be represented by the resonating system (l)t+(2). A third canonical form (36) could arise from the group (1) by electron shifts (c) which are usually referred to as "back polarisation".The nitrone group is thus a resonance hybrid of the extreme structures (l) (2) and (36). In most reactions the behaviour of nitrones indicates activation through polarisation scheme (a) but some properties e.g. geometrical isomerism indicate a degree of double-bond character (form l) while some reactions [section D 4 c (iv)] require (c) as the polarisation step in activation. (a) Dimerisation. ( i ) CycZic dimers. Dimerisation through form (2) will lead to charge neutralisation and may occur in various ways depending 34 R. Bonnett V. M. Clark and Sir Alexander Todd J. 1959,2102. 35 I,. Kaminsky and M. Lamchen Chem. Comm. 1965 130. 338 QUARTERLY REVIEWS on stability and steric effects. When the dimer is of lower energy than the monomer spontaneous dimerisation will occur. Thus from 1 -hydroxy- piperidine on oxidation with cupric acetate the cyclic dimer (30) was isolated and not the monomeric nitrone (29).21 On the other hand 1- hydroxypyrrolidine gives the monomeric 1 -pyrroline 1 -0~ide.l~" This difference was attributed to a lowering in the Pitzer strain when the six- membered nitrone (29) forms the dimer (30) whilst such a dimerisation of the five-membered nitrone would raise the Pitzer strain.15Q The nitrone (lo) although being six-membered is obtained as a monomer.The stability of this monomer has been attributed20 to de- activation of the nitrone group by conjugation with the aromatic nucleus. (ii) Acyclic dimers. The electrophilic character of the carbon atom of the nitrone group activates adjacent methyl and methylene groups and it is thus not surprising that suitably substituted nitrones should undergo aldol-type reactions.In some cases the reaction is spontaneous and the aldol dimer is obtained instead of the monomeric nitrone. Thus phenyl- hydroxylamine and acetone when heated together give the dimer (18) and since no additional basic catalyst is required either the hydroxylamine or the nitrone intermediate must act as base.1° Similarly the nitrone (16) slowly forms the dimer (37) on standing.36 CMe B- O P H-CH - BH t r CMe2 H2C- \ I + - MoC=$-O- +dL0- - MeC=$J-6 I YOH Me C= y-0- Ph Ph Ph ph Ph (1 8) Unless steric effects or the conditions mentioned above operate against dimerisation dimers can be expected to form when nitrones are prepared. Thus in cases where no reason is obvious the nitrones reported as mono- mers could well be re-examined to confirm their monomeric state.Even the stable monomers can be induced to dimerise under basic conditions the basic catalyst controlling the product. Thus whilst the nitrone (20) shows little tendency to dimerise when stored at room tem- perature it readily dimerises under basic conditions ; triphenylmethyl- sodium produces an aldol-type dimerisation to yield the dimer (38) whilst sodamide in liquid ammonia gives a benzoin-type dimerisation to produce the dimer (39); sodamide in triethylamine produces a mixture of the dimers (38) and (39).32 No explanation for this is obvious. 2-Methyl-substituted cyclic nitrones do not dimerise under strongly basic conditions most likely due to repulsion of the anions produced. However if 2-unsubstituted cyclic nitrones are also present aldol addition between the two nitrones occurs thus if the nitrone (16) is treated with triphenylmethylsodium and 43,s- trimethyl- 1 -pyrroline 1-oxide the aldol addition product (40) is formed.36 (b) Aldol Additions and Condensations.36 R. F. C. Brown V. M. Clark I. 0. Sutherland and Sir Alexander Todd J. 1959 2109. DELPIFXRE AND LAMCHEN NITRONES 339 The activation of a methyl group by an adjacent nitrone system is also demonstrated by the aldol-type condensation such nitrones give with aldehydes e.g. the nitrone (16) and benzaldehyde in the presence of potassium hydroxide give the styryl derivative (41).13a Me Me ,C02 Et PhCH=C Although aldol-type dimerisation occurs on the methylene group in position-3 of the nitrone (20) [see dimer (38)] no aldol condensations with aldehydes are found with 2-unsubstituted or 2-ethyl-substituted n i t r ~ n e s .~ ~ ~ A methylene group activated by being adjacent both to a nitrone and to another electron-attracting group may however condense with aldehydes. Thus the nitrone (42) gave the product (43) with benzalde- hyde in the presence of base.37u In the above the nitrones acted as the nucleophiles in the aldol condensa- tions. When the 2-unsubstituted nitrone (20) reacts with nitro-alkanes in the presence of sodium ethoxide the nitrone reacts as the electrophile and 1,3-additions to the nitrone to produce the 2-nitroalkyl-l-hydroxy- pyrrolidines occur. ( c ) Addition Reactions. Dimerisation and aldol reactions are specia cases of addition reactions. The polar character of the nitrone group makes it susceptible to a wide variety of 1,3-addition reactions.( i ) Addition of ccrrhonyl reagents. Nitrones react with the usual car- bony1 reagents to produce derivatives (45) of the parent carbonyl com- pound~.~’ The mechanism is not certain since no intermediate has been isolated but a likely route is through the adduct (a) which loses the hydroxylaminoresidue. [See also Section D 4 (f ). s7 (4 G. E. Utzinger and F. A. Regenass Helv. Chim. Acta 1954 37 1892; (6) M. Hamana B. Umezawa and Y. Goto Chem. A h . 1961,55,8405. 340 QUARTERLY REVIEWS [In (45) R2 = OH HN-Ph HN(2,4-dinitrophenyl) or NH-CONH in oximes phenylhydrazones 2,4-dinitrophenylhydrazones and semicar- bazones respectively.] Hydrogen cyanide adds 1,3 across the nitrone system to give a-hydroxylamino-nitrones.In the presence of base some of these adducts lose water to form cyano-imines. (ii) Addition of hydrogen cyanide. The 1-pyrroline 1-oxides react with hydrogen cyanide only when the 2- position is unsubstituted and these adducts were found to be stable to alkali.13" A carbonyl group attached to the carbon of nitrones exerts little effect on the mechanism and cyano-imines are formed. If however a carbonyl group is attached to the nitrogen atom its -I effect influences the nitrone system to lose the oxygen atom; thus potassium cyanide reacts with the nitrone (46) to give the deoxygenated product (47) and potassium ~ y a n a t e . ~ ~ (iii) Addition of Grignard reagents. Grignard reagents add 1,3 across the nitrone system of aldonitrones and after hydrolysis of the complex give hydroxylamines.Ketonitrones may resist such additions and be re- duced to i m i n e ~ . ~ ' ~ In the 1-pyrroline 1-oxide series normal addition proceeds both with aldo- and keto-nitrones. The reaction is potentially very useful as more highly substituted heterocycles may be produced by this method. Even the dimeric nitrone (30) which is resistant to a number of nitrone reactions reacts with phenylinagnesium bromide to give the hydroxylamine (24) in excellent yields.21 R' R* R' R S (i) R*MgBr 'C-N / -t \ +/ / \ C=N H'fL 'OH (ii) H,O 0- H V. BelIavita and N. Cagnoli Guzzetra 1939,69,583; (Chem. A h . 1940,34 1638). DELPIERRE AND LAMCHEN NITRONES 341 (iv) 1-3 Cycloadditions. Formation of the dimer (30) is a 1,3-cyclo- addition between two nitrone systems. The general reaction cycloaddition of nitrones with unsaturated compounds has been extensively studied in recent years.With olefins the nitrones react to form isoxazolidines according to the following scheme. R R’ R; I k=k + c=c’ Unless the addition is stereospecific a stereochemically pure isoxazolidine (48) can only be expected to be formed when R1=R2 and R4=R5=Rs=R7. In all other cases a number of stereoisomeric isoxazolidines may form due to the fact that up to three asymmetric centres could be created and that with unsymmetrically substituted olefins two orientations of addition are possible. In most of the work reported in this field excellent yields are obtained but the stereochemistry and substitution pattern of the isoxa- zolidines have not been determined. The ease with which these adducts are formed and their stability are utilised to trap unstable nitrones or those difficult to purify.In these cases the nitrone is formed in situ in the presence of the unsaturated compound and the adduct is i ~ o l a t e d . ~ ~ ~ ~ Although cycloaddition to an isolated double bond occurs conjugation of the olefinic bond with a group which enhances the polarisability of the double bond has a marked effect on the ease of cycloaddition. For example the rate of addition of C-phenyl-N-methylnitrone (1 1 R= Me) to olefins of the type R-CH=CH2 increases fourfold as R changes from alkyl to phenyl and one hundred and fifty fold when R = C0,Et.3 The mechanism of the cycloaddition is not known with certainty neither is it clear whether it always follows the same pathway. Basically two mechanisms have been considered (1) A concerted one-step process i.e.synchronous closure of both bonds as represented below In such a process stereospecific cis-addition must occur. With un- symmetrical olefins however two adducts are possible depending on the direction of addition i.e. R1 and R2 could be interchanged with R3 and R4. Huisgen3 favoured this mechanism and presented strong arguments for his choice. Unfortunately some of the results on which the argument is based are not yet published and are thus difficult to assess. 30 R. Grashey R. Huisgen and H. Leitermann Tetrahedron Letters 1960 No. 12 9. 342 QUARTERLY REVIEWS (2) An alternative mechanism proposed involves a two-step process. The structure of the intermediate will depend on the polarised state or polarisability of the reagents.'NY A H (54) The groups R and R' will determine the polarisation in the addend (51). The mobility of the T electrons in the nitrone system will enable formation of two different activated complexes (52) and (53) through the hypothetical canonical forms (49) and (SO) respcctively. Stereospecific ring closure of the activated complexes may follow or in the absence of steric hindrance free rotation of the intermediates may result in both cis- and trans- addition. Either products (54) or (55) may thus be formed and they may be single products or mixtures of isomers. The present authors in their study of cycloaddition of ethyl acrylate to 1-pyrroline 1-oxides favoured this latter mechanism.&O Thus the nitrone (20) reacts exothermally with ethyl acrylate at room temperature to give a quantitative yield of isomers with partial structure (54; R= H R'=CO,Et).This reaction being fast would require a low entropy of activation and is thus unlikely to follow a one-step process. If the recognised direction of polarisation of ethyl acrylate is accepted the reaction proceeds through the canonical form (49) the intermediate (52) in this case having the partial structure (56). The adduct could be isornerised to the thermally stable isomer with partial structure (55; R=H R'=CO,Et) which also forms when the nitrone (20) and ethyl acrylate are heated together. In this latter case the reaction was considered to proceed via the canonical form (50) and to give the intermediate (57) which ring-closed stereospecifically. 40 G . R. Delpierre and M. Lamchen J. 1963,4693. DELPIERRE AND LAMCHEN NITROPJES 343 This work showed that cycloadditions could be kinetically and thermo- dynamically controlled the room temperature reaction having been kinetic- ally controlled whilst the hot reaction submitted to thermodynamic control.Until more evidence is published especially a comparison of entropy of activation with overall entropy of the reaction no decision can be made as to the mechanism of cycloaddition. Under different conditions and with different reagents different mechanisms may be operative. 1,3-Cycloadditions of nitrones also occur with isocyanates isothio- cyanates and a1 k ynes .4 With tet raphenylc yclopentadienone 1,4-addi t ion across the conjugated diene system and 1,2-addition to the nitrone is found.41 (d) Reduction. The nitrogen atom in nitrones is at a high oxidation level and stepwise reduction to secondary amines is possible.Two reduc- tion pathways are possible each with a different intermediate; (i) addition of hydrogen to form the hydroxylamine and (ii) deoxygenation to the imine. / R' Judicious choice of reducing agent will allow preparation of any of these reduction products. Hydride reagents e.g. lithium aluminium hydride40942 or sodium borohydride,13" attack at the electrophilic carbon atom of the nitrone group and on hydrolysis of the complexes formed hydroxyl- amines are obtained in good yields. The reaction stops at the hydroxyl- amino-stage even when excess of lithium aluminium hydride and fairly vigorous conditions are used.15q Removal of the oxygen atom to form imines can be effected by treatment with a variety of reagents; zinc and acetic acid,13" sulphur dioxide,l3" and triphenylph~sphine~~ have proved very successful ; other reagents e.g.phosphorus tri- or oxy-chloride have also been used sometimes with attendant decomposition or side reactions. Generally zinc and mineral acid combinations reduce the nitrone group to the secondary amine. (e) Oxidation. The oxidation level of the nitrogen atom in nitrones is such that oxidation cannot raise it without disrupting the system and liberating the nitrogen atom as a nitroso- or nitro-group. This happens during the ozonisation of CN-diphenylnitrone which gives benzaldehyde 41 C. W. Brown K. Marsden M. A. T. Rogers C. M. B. Taylor and R. Wright 42 0. Exner Chem. M y 1954,48 1543; (Chem. Abs. 1955 49 11603). Proc. Chem. Soc. 1960,254.F. Angolini and R. Bonnet Canad. J. Chem. 1962,40,181. 344 QUARTERLY REVIEWS and nitr~benzene.~~ Since oxaziridines are stable to ozone cleavage of the C=N bond did not occur through a three-membered ring intermedi- ate; the following mechanism best explains this reaction. H Ph-C=N< I f ? 0- Ph -- Ph- 0 u r l U '1' (0- 'I' y ~ - < ~ ~ --+ Ph-$ + PhNO+02 0-o+ I 0 10 PhNO v Selenium dioxide oxidises the nitrone (20) to give the conjugated keto- nitrone system (58).45 This is analogous to the oxidation of a-methylene groups in aldehydes and ketones to produce the a-dicarbonyl When an adjacent methyl group is present as in 2,4,4-trimethyl-l-pyr- roline I-oxide (16) the reaction is more complex and in this case the ex- pected aldehyde (59) was not isolated; instead the ring-expanded product (60) was obtained4'-this is probably an artefact produced by acid treat- ment of the reaction mixture.Attack on the nitrone group of the nitrone (20) has also been observed with iron(Ir1) chloride; the product is the cyclic hydroxamic acid (61).4* (f) Hydrolysis. Acyclic nitrones are generally readily hydrolysed by acids to form aldehydes or ketones and N-substituted hydroxylamines. Considerable variation in the stability of nitrones to solvent action is found and whereas alkyl substituted nitrones are rapidly hydrolysed by aqueous acids and even decomposed by the 1-pyrroline 1-oxides are stable in hydroxylic solvents and in dilute aqueous acids. Acyclic aryl nitrones are intermediate between these two extremes in stability. In view of the lability of many nitrones to mildly acidic conditions some of their reactions with carbonyl reagents [Section D 4 (c) ( i ) ] which are usually carried out at low pH may well proceed through a preliminary hydrolysis and subsequent reaction of the carbonyl compound produced.44 A. H. RiebeI R. E. Erickson C. J. Abshire and P. S. Bailey J. Amer. Chem. Soc. 45 V. M. Clark B. Sklarz and Sir Alexander Todd J. 1959 2123. 46 N. Rabjohn Org. Reactions 1949 5 331. 4' R. F. C. Brown V. M. Clark and Sir Alexander Todd J. 1959 2105. 48 J. F. Elsworth and M. Lamchen unpublished results. 49 0. Exner Coll. Czech. Chem. Comm. 1951,16,258; (Chem. Abs. 1953,47 5884). 1960,82 1801. DELPIERRE AND LAMCHEN NITRONES 345 ( g ) Rearrangements. The nitrone group is susceptible to a number of rearrangements which reflect both on the reactivity of the nitrones as a class of compounds and the complexity of their chemistry.Such rearrange- ments may be divided into the following types :- ( i ) Photolysis. Irradiation of nitrones has been shown to produce the isomeric oxaziridines.M160 Activation of the v-electrons by light of the appropriate wavelength will be influenced by the substituents on and adjacent to the nitrone group and the isoiation of the oxaziridine will also be determined by its stability. Thus whilst 2-unsubstituted 1-pyrroline 1 -oxides form oxaziridines readily the 2-substituted isomers have been reported not to give oxaziridine~.~~ It has however been shown that the 2-substituent does not preclude oxaziridine formation and that 2,5,5- trimethyl- 1 -pyrroline 1 -oxide forms the oxaziridine on irradiati~n.~~ Oxaziridines may on heating either revert to the nitrone or be converted into the isomeric amides.Prolonged irradiation of nitrones may also pro- duce amides. A radical mechanism is most likely for the oxaziridine formation. H G - f b ) \cFN-R' hoot- RCO.NHR' R,q-j-j"y> H +,R' ~ - hv H ,yQ'd-& R/ ' ' q b ] shifts (b) (b) (a) *o. R C O heat shifts (a) (ii) Amide formation. In addition to the photolytic rearrangement mentioned above aldonitrones have been rearranged by a wide variety of chemical reagents. Phosphorus penta- tri- and oxy-chlorides acetyl chloride acetic anhydride sulphur dioxide and even bases in ethanolic solution have converted nitrones to amides often in good yields.51 In most cases the substituents do not migrate as in the case of the classical Beckmann rearrangement and this would suggest that the mechanism which operates must be different.Krohnke has suggested two mechanisms to explain the course of the rearrangement.52 For rearrangements in which no migration of substituents occur for example by the action of acetic anhydride51 [scheme (a) X = CH3CO] or with sodium e t h ~ x i d e ~ ~ [scheme (b) X = HI the following was suggested. 5a M. J. Kamlet and L. A. Kaplan J. Org. Chem. 1957,22 576 J. S. Splitter and M. 51 0. L. Brady and F. P. Dunn J. 1926 2411. 52 F. Krohnke Annalen 1957 604 203. 53 L. Chardonnes and P. Heinrich Helv. Chim. Ada 1944 27 321. Calvin ibid. 1958 23 651. 346 QUARTERLY REVIEWS When X = H the reaction is merely a hydration of the nitrone group to a nitrone hydrate (62 X = H) which dehydrates to the imidol form of the amide.The existence of such nitrone hydrates has been demonstrated by Krohnke.64 This mechanism cannot however explain the rearrangements of C- benzoyl-N-arylnitrones (63) to the formanilides (65) which occur with migration under basic conditions. For these rearrangements Krohnke62 proposed the following mechanism which assumes both nitrone hydrates and oxaziridines as intermediates. HOH PH Ph-COCH=Ft]-Ar - Ph-COCH-ifJ-Ar 5 &- OH (63) The conversion by heat of an oxaziridine such as (64) into an amide makes this mechanism plausible,55 but step [(64) to (65)] is probably better represented by the mechanism below. -. . PhC-CH-NAr - HCO.NAr.COPh c& w*/3 (iii) Ketonitrones to aldonitrones. Under certain conditions base- catalysed prototropic shift may occur between the N- and C-substituents; the net result is the conversion of a ketonitrone into an aldonitrone.The reaction can be explained by the following mechanism. Ph . . H+ Ph\ + P h \ n + C=ykyHPh - C-V=CHPh - CH-y=CHPh Ph' 0- H- Ph' 0- Ph' 0- OEt (iv) Oxime-0-ethersformation. Heat converts some nitrones into oxime 0-ethers but the reaction is not general. Thus whilst the nitrone (66) remained unchanged after prolonged heating at 2OO0 the nitrone (67) was quantitatively rearranged to the 0-ether (68) in 1/2 hr. A similar rear- rangement can be effected by acid treatment.5s The following mechanism was ~uggested.~' F. Krohnke and E. Barrier Ber. 1936,69,2006. 56 A. Padwa Tetrahedron Lefters 1964 2001. s6 M. Martynoff Ann. Chim. (France) 1937 7,424. I7 A. C. Cope and A.C. Haven jun. J. Amer. Chem. SOC. 1950,72,4897. DELPERRE AND LAMCHEN NITRONES 347 Ph \ Fh\c=&?7 A = C=N-OCHPh (67) Ph ’ “CHPh Ph’ The above rearrangement may precede acid-catalysed (68) hydrolysis of those nitrones in which the hydroxylamine salt and not the &substituted hydoxylamhe salt is produced. E. Dinitrones A number of dinitrones have been prepared usually by oxidation of NN‘-dihydroxy-compounds or of N’-hydroxynitrones. The dinitrone (7 1) was considered to be formed by spontaneous dimerisation of the unstable nitrone (69). The dimer (70) must have undergone oxidation during the isolation.37a + + - H + + 2 Ph-N=CH + PhN=CH-CH,-N-Ph -+ Ph-N=CH-CH=N-Ph A- (71) 0- I 0- I AH A- (69) (70) Catalytic oxidation of the dimers (38) and (39) gave the dinitrones (72) and (73) re~pectively.~~ The properties of these two nitrones were found to be similar to those of the mononitrones and nitrone (73) gave the normal reaction with phenylmagnesium bromide adding two moles of the Grignard reagent per mole of dinitrone.The nitrone (72) gave two strong absorption bands in the infrared one at 1570 cm.-l due to a 2-unsubstituted nitrone group and one at 1603 cm.-l due to a 2-substituted nitrone group. In nitrone (73) a pronounced batho- chromic shift due to conjugation of the nitrone systems was observed and absorptions were at 1509 and 1503 cm.-l. Similarly the ultraviolet absorptions at 237 mp (616,500) for nitrone (72) and 331 mp (~18,500) or nitrone (73) showed a bathochromic shift due to the conjugation of the two nitrone Only one preparation of a monocyclic dinitrone has been reporteds8 but later work has shown that the compound produced was not a nitrone6@.Recently the monocyclic dinitrone (74) was preparedss and also showed 6* J. K. Landquist J. 1956 1885. M. Lamchen and T. Mittag unpublished results. 348 QUARTERLY REVIEWS the bathochromic shift in both infrared and ultraviolet absorptions these being at 1545 cm.-l and 347 mp (€12,300) respectively. This dinitrone however did not give the normal nitrone reactions probably due to the conjugation. F. Importance and Uses of Nitrones The peculiar properties and reactivity of the nitrone group has enabled many chemists to make use of these compounds in syntheses. The easy hydrolysis of nitrones to carbonyl compounds has made this an important method for the synthesis of aldehydes and ketones especially the sensitive or otherwise not easily accessible compounds.This method has been used for aliphatic aromatic alicyclic and heterocyclic compounds and has been shown to be applicable to saturated as well as unsaturated mono- as well as di-carbonyl compounds with or without other functions such as the amino- or carboxyl functions etc. The nitrones required are usually obtained from the nitroso-compounds by the Krohnke reaction (p. 335). The great reactivity of the nitrone group and the easy removal of the oxygen atom make the nitrones excellent starting materials for a variety of products. Thus Lord Todd and co-workers made use of the reactive nature of the 1-pyrroline 1-oxides to link pyrrole rings together both directly and through a methylene group,32 in their attempted synthesis of corrins. A number of new pyrrolines and pyrrolidines were synthesised in this The 1,3-~ycloadditions also opened up new routes to N-bridged hetero- cyclic compounds and a number of new isoxazolidines have already been prepared by this route. Through these preparative methods the nitrones have become useful intermediates in many fields of chemistry and with the recent rapid expansion in nitrone research their usefulness will undoubtedly increase in the near future. One of us (M.L.) is grateful to his reseach students and colleagues with whom he has had valuable discussion on many aspects of nitrones. work.13,32,36,45,47

ISSN:0009-2681    

DOI:10.1039/QR9651900329

出版商:RSC    

年代:1965

数据来源: RSC

摘要:

By D. M. GRUEN (CHEMISTRY DIVISION ARGONNE NATIONAL LABORATORY ARGONNE ILLINOIS) Introduction HIGH-TEMPERATURE absorption spectrophotometry has been extensively employed in studies of the physicochemical properties of fused sa1ts.l-l’ This technique has reached a state of development in recent years which makes possible the quantitative measurement of absorption spectra of solute species in melts at temperatures up to and well above 1000”~. The methods which have been developed for the measurement of absorption spectra at high temperatures will be discussed. Spectral results will be surveyed to illustrate the variety of solute species which can be studied and to point out features which must be taken into consideration for quantitative measurements. The Scope of Fused-salt Spectrophotometry Molar absorptivities of solute species in fused salts can be determined from accurate measurements of absorbance optical path length and solute concentrations.The technical problems involved in making such measurements at high temperatures will be discussed later. We will first consider the absorption spectra of a variety of solute species and the effects of temperature and solvent composition on the spectra. The scope of fused-salt spectrophotometry in the present context is essentially defined by and limited to the wavelength range of available spectrophotometers ca. 50,000 to ca. 4000 cm.-l or 200 mp to 2500 mp. Electronic transitions occurring in this region of the electromagnetic spectrum can therefore be studied. The fused-salt solvents for the most part do not absorb light energy over most of this region.In fact many fused-salts such as molten alkali halides and nitrates are more transparent than water in the near infrared and infrared regions so that electronic transitions can be measured in D. M. Gruen “Fused Salts,” ed. B. R. Sundheim McGraw-Hill Book Co. New York 1964 ch. “Spectroscopy of Transition Metal Ions in Fused Salts.” * G. P. Smith “Molten Salt Chemistry,” ed. M. Blander Inter-Science Publishers New York 1964 ch. “Review of Electronic Absorption Spectra of Molten Salts.” G. Harrington and B. R. Sundheim Ann. New York Acad. Sci. 1960,79,950. J. R. Morrey Inorg. Chem. 1963 2 163. J. P. Young and J. C. White Anulyt. Chem. 1959 31 1892. ( a ) E. Rhodes and A. R. Ubbelohde Proc. Roy. SOC. 1959 A 251 156; (b) E. I. V. Tananeev and B. F. Dzhurinskii Dokludy Akad.Nauk S.S.S.R. 1961 140 R. Molina Bull. SOC. chim. France 1961 301. W. T. Carnally Analyt. Chem. 1962,34 786. Rhodes W. E. Snith and A. R. Ubbelohde Proc. Roy. SOC. 1965 A 285,263. 374. lo H. Lux and T. Niedermaier Z. anorg. Chem. 1956,285 246. l1 K. Sakai J. Phys. Chem. 1957 61 1131. 349 350 QUARTERLY REVIEWS fused-salt media which are masked by solvent absorption bands in aqueous solutions. An important early investigation by Retschinsky12 contains results on the temperature variation of the absorption edges of some 30 salts in the crystalline as well as in the molten state. Measurements on a number of salts in the molten state are reproduced in Fig. 1. The experimental I I 0- 40,000 35 I I I OOO 25,000 20,OOO 15,000 c m" FIG. 1. Temperature variation of the absorption edges of molten salts A USO,; B Na,SO,; C Rb2S04; D KCI; E RbCl; F NaBr; G CdCle; H PbCI1; J AgCl; K BiCI,; L AgN03; M KN03; N ZnC1,.[Adapted from T. Retschinsky Ann. fhys. 1908 27 100.1 points correspond to the highest energy of light transmitted through ca 1 cm. of the molten salt as recorded on a photographic plate. Depending on the salt the edge of the fundamental absorption band shifts 1000-2000 cm.-l to the red for each 100" rise in temperature. The red shift was found to be essentially of the same magnitude above and below the melting point. Because of the large number of salts investigated Retschinsky was able to correlate the energies of the absorption edges with the constitution of the salts 1. Heavy-metal salts absorb at lower energies than salts of the alkali and alkaline-earth metals.2. For the same metal with the exception of silver the energy of the absorption edge decreases in the order SO4- < C1- < Br- < I- < NO3-. With silver the order C1- < NO3- is reversed. 3. For the same anion the energy of the absorption cut-off decreases with increasing atomic number of the metal in a given group of the Periodic Table. Although Fig. 1 includes data only on chlorides nitrates and sulphates very similar results can be expected for fluoride melts and oxidic melts such as molten silicates and borates. The location of the edges of the fundamental absorption bands of the fused-salt solvents are of course very important insofar as observation and precise measurements on electronic transitions of solute species are la T. Retschinsky Ann.Phys. 1908,27 100. GRUEN FUSED-SALT SPECTROPHOTOMETRY 351 concerned. For example the intensities of inner electronic transitions occurring within the 3d 4d 5d 4f or 5fshells of the ions are lower than the intensities of charge-transfer bands by a factor of lo3 to lo4. For this reason the charge-transfer bands virtually obscure all inner electronic transitions occurring at energies greater than the edges of the charge- transfer bands. It should be pointed out that charge-transfer bands involv- ing the transition-metal ions are usually found at somewhat lower energies than the bands due to alkali-metal ions. The charge-transfer bands of the transition-metal ions therefore set a limit on the highest energy at which inner electronic transitions can be observed. Fortunately a large number of the inner electronic transitions occur at energies lower than ca.30,000 cm.-l where the edges of charge-transfer bands often begin to make their appearance in melts. Many electronic spectra of transition-metal ions can therefore be observed in fused-salt solutions. Furthermore ions of heavy metals such as T1+ Pb2+ and Bi2+ metals and intermetallic com- pounds dissolved in fused-salts and species such as dissolved sulphur selenium and tellurium have characteristic absorption bands in wave- length regions available for measurement. The number of different solute species which have so far been studied in fused salts is relatively small and one can confidently predict that as fused-salt solution chemistry develops many other species with electronic absorption bands will be studied.Spectra of Transition-metal Ions Transition-metal ions may be defined as all those ions having unfilled 3d 4d 5d 4J and 5fshells. This group includes ions of about half of all the elements in the Periodic Table. The absorption spectra (and the colours) of these ions in crystals and in solutions are due to transitions occurring within the unfilled inner electron shells. The Hamiltonian for the electrons of a transition-metal ion consists of two terms H = HP + V where HF is the Hamiltonian of the free ion and Vis the potential provided by its ligands. In fused-salt solutions the ligands are of course derived from anions provided by the melt e.g. Cl- F- 02- NO3- etc. In the Hamiltonian of the free ion HF there are two perturbing quanti- ties an electronic repulsion term R and a spin-orbit coupling term S whose magnitudes compared with that of ligand field potential V deter- mine the spectral characteristics of the ions in solution.Four broad classes distinguished by different relative magnitudes of the perturbation terms are as follows; each class is representative of a particular group (or groups) of transition-metal ions. (1) S < V < R 3d ions (2) V > R > S 4d and 5d ions; oxygenated 5f ions such as NpO 22+ (3) v < s < R 4f ions and tripositive 5’ ions (4) V g S g R tetrapositive 5f ions 3 52 QUARTERLY REMEWS This classification scheme is of course an approximate one only and represents an over-simplification of the actual situation. Nonetheless it is useful since it helps one to understand the magnitude of the effects pro- duced on the spectra by changing the number and kinds of ligands sur- rounding a particular transition-metal ion.The magnitude of the spectral changes which are caused by changing the solvent the temperature or both vary from one group of ions to another. Solvent and temperature effects are largest for the 3d and the tetra- positive Sfions smaller for the 4d and 5d ions and of least importance for the tripositive 4f and 5f ions. The underlying reason for this ordering is the magnitude of the ligand potential V relative to the S and R terms. The large solvent and temperature effects on 3d ion spectra necessitates a somewhat closer examination of their origin. The absorption spectra of the 3d ions are due to transitions between electronic levels whose energies and intensities are strongly influenced by the number symmetry and bonding of the nearest-neighbour anions.The ground states of the free ions of the 3dseries are aD(d19s); 6D(d4~*); SF(d29; 4F(d397); 6S(d5). These states are split under the action of the elec- tric field set up by the ligands.13 In a cubic (octahedral or tetrahedral) field a D-state is split into two levels designated by T2 and E and separated by an energy 10 Dq. In the case of dl and d9 these are the only levels arising from the d-configurations. In the free ion d4 and dG configurations there are upper triplet H P F and G multiplets spaced in the region 22,000-28,000 cm.-l. Transitions from the quintet ground states to these excited triplet multiplets can be expected to be of low intensity. Because of their relatively high energy they would occur very near the intense charge transfer bands and would in all likelihood be masked by them.For these reasons the D-state ions in chloride melts would be expected to display as the most prominent feature in their spectrum a single absorption band due to the T + E transition. An F-state in an electric field of cubic symmetry is split into three energy levels designated by T, T, and A where the T1-T2 separation is 8 Dq and the T,-A2 separation is 10 Dq. In addition to 3Fand 4Fgr~und states there are 3P and 4P excited states spaced in the region 10,OOO-14,000 cm.-l in the free-ion configurations d2 d3 d7 d8. Transitions from the F-ground states to the P(Tl) excited states are intense because no change in spin- multiplicity is involved. The P(T,) states are not split by a cubic crystal field.For d2 and d7 configurations in octahedral fields and d3 and d8 configurations in tetrahedral fields the A@) and T,(P) states cross at Dq M 1000 cm.-l so that at larger Dq values the T,(P) state lies below the A2(F) state. In any event the spectra of ions possessing F-ground states can be expected to have more than one band in the region accessible to measurement in chloride melts. One is likely to observe three bands cor- l3 C. J. Ballhausen “Introduction to Ligand Field Theory,” McGraw-Hill Book Co. Inc. New York 1962. GRUEN FUSED-SALT SPECTROPHOTOMETRY 353 responding to transitions from the lowest F-state component to the upper two components and in addition a transition to the P-multiplet. The 0S ground state of the Mn2+ ion is not split in a crystal field and transitions to upper 4G 4P and 4D states would be expected to occur in the region 20,000-35,000 cm.-l.It can be shown that the T2-E transition energy for ions with D- ground states and the energy between the lowest ligand field states and the P-states of ions with F-ground states increase with increasing values of Dq the crystal field splitting parameter. The empirical observation that the successive replacement of I- ligands by Br- C1- and H,O ligands shift the maxima of spin-allowed bands to progressively shorter wavelengths led to the construction of the spectro- chemical series.14J5J6 One can rationalise the ligand sequence in the spectrochemical series in terms of increasing values of Dq and if one con- siders only transitions between levels in an octahedral ligand field one would expect their energies to increase with increasing magnitude of Dq.On the other hand according to the point-charge model Dq tetrahedral = -4/9 Dq octahedral so that a change in the number of ligands from six to four would be expected to decrease the transition energy by several thousand wavenumbers. The intensities of spin-allowed bands are larger by a factor of 5-10 in tetrahedral as compared with octahedral fields because in a tetrahedral field the absence of a centre of inversion symmetry results in a relaxation of selection rules for d -f d electronic transitions. Solvent and Temperature Effects on 3d Transition-metal Ion Spectra in Fused-salt Solutions The absorption spectra of the 3d ions undergo profound changes which arise from changss in the nature and composition of the fused-salt solvents.In Table 1 are listed the absorption maxima of several transitions between ligand field levels for the ions Co" Nil1 and Crul in aqueous solution in molten LiN0,-KN03 eutectic in molten LiF-NaF-KF eutectic and in molten aluminium chloride. In all of these cases the 3d ions are in an octahedral field of six ligands. The absorption maxima are seen to shift to progressively lower energies going from aqueous solution to the aluminium chloride melt indicating that Dq decreases in the order H20 > NO3- > F- > C1-. The overall shift in the maxima varies from ca. 3000 to 5000 cm.-l depending on the ion and the particular transition involved. Changes of this magnitude in the energies of absorption bands can of course change the visually observed colours of solutions very markedly.l4 K. Fajans Naturwiss. 1923 11 165. l5 R. Tsuchida Bull. Chem. SOC. Japan 1938,13 388,436. l6 W. Moffitt and C. J. Ballhausen Ann. Rev. Phys. Chem. 1956,7 107. 354 QUARTERLY REVIEWS TABLE 1. Octahedral spectra of CrIII Colt and Niu in jluoride chloride and nitrate melts and in aqueous solutions. Absorption inaxima Water Ref. NQ3- Ref. F- Ref. C1- Ref. mP " P *P mP Con 510 17 560 19 580 22 634 21b Nin 395 17 425 19 434 22 487 21b 720 17 775 19 850 22 927 216 C P 425 18 448 22 545 21b 585 685 22 795 216 Transn. The change from an octahedral to a tetrahedral configuration produces both a decrease in band energy and an increase in band intensity. In going from an octahedral ligand field with a relatively high Dq say nitrate to a tetrahedral ligand field with a relatively low Dq say C1- very large spectral changes are encountered.The effects are illustrated in Figs. 2 and 3. In these illustrations Co" and Nirl spectra in aqueous HC104 and in 0 500 600 700 800 Wavelength (mp) FIG. 2. Absorption spectra of Con in A M-HCIO at 25"; B LiN03-KN03 eutectic at 160"; C pyridine hydrochloride at 160"; D Cs,ZnCl solid solution at 25" illustrating the efect of an octahedral-tetrahedral co-ordination number change. GRUEN FUSED-SALT SPECTROPHOTOMETRY w I- I 355 Wavelength (m) FIG. 3. Absorption spectra of Ni” in the same media as in Fig. 2 illustrating the efect molten LiN0,-KNO are compared with spectra of these two ions in molten pyridinium chloride and in crystals of Cs,ZnCl,. The aqueous solutions and nitrate melt spectra are “octahedral” spectra.The spectra in molten pyridinium chloride on the other hand are “tetrahedral” spectra owing to the ions COCI,~- and NiC142-. This is proved by the close simi- larity of the Cs,Zn(Co)CI and Cs,Zn(Ni)Cl, crystal spectra and pyridin- ium chloride melt spectra. Not only the band energies but the intensities of the bands are profoundly effected in going from the ionic nitrate to the ionic chloride solvent. Co-ordination-number changes from six to four with their attendant spectral effects are observed in alkali chloride melts as a function of temperature. For example V‘ in LiC1-KC1 eutectic undergoes the series of spectral changes shown in Fig. 4 as the temperature varies from 400”c to 8 0 0 ” ~ . The spectral changes can be interpreted as arising from an equilibrium of the sort of an octahedral-tetrahedral co-ordination number change.VC1,& + VC142- + 2C1- In the case of Vn increasing the temperature leads to a pronounced increase in the intensity of the spectrum and a shift of the band maximum to lower energies.17 Solutions of Co” in LiCI-KC1 eutectic on the other hand display a decreasing transition intensity with increasing temperature over the same temperature interval. In this ionic melt Con is four-co- D. M. Gruen and R. L. McBeth J. Phys. Chem. 1962,66 57. 356 QUARTERLY REVIEWS ordinate even at the lowest melt temperatures and the co-ordination- number does not change as a function of temperature.18 The intensity decrease appears to be related to subtle changes in the covalency of the Co-C1 bond. W >t P c > 4 .- .- n 0 0 L - f FIG.4. Wavelength (mu) Temperature variation of the absorption spectrum of Vu in LiCI-KCI eutectic A = 40O0 B = 600° C = 800" D = 1OOO". Spectra of Lanthanide Ions and the Analysis of Lanthanide-ion Mixtures in Fused-salt Solutions Studies of the absorption spectra of Pr3+ and Nd3+ in LiN0,-KNO eutectic at 184"c showed that the narrow bands so characteristic of aqueous-solution spectra of the rare-earth ions are largely preserved in fused-salt ~ o l u f i o n ~ . ~ ~ In addition to nitrates fused chlorides are also suitable as solvents in which to carry out spectrochemical studies of rare- earth ions. The usefulness of fused-salt solvents for spectroscopic studies is due to the wide region of optical transparency extending into the infrared. For example LiC1-KC1 eutectic at 400"c is transparent in the range ca.30,000 to ca. 1000 cm.-l. For the study of rare-earth mixtures the main problem is to find spectral regions in which characteristic bands due to ions of one element are not overlapped or masked by those of another. It is very desirable therefore to be able to make measurements over as wide a spectral range as possible. The ability to measure transitions in the region 7000 to 4000 cm.-l is particularly useful since many 4f ions possess electronic transitions in this energy range. Water absorbs strongly at wave-numbers less than 7000 cm.-l and cannot serve as a solvent in the infrared. By studying the absorption spectra of the lanthanides in molten nitrate systems for D. M. Gruen and R. L. McBeth Pure and Appl. Chem. 1963,6 23.l9 D. M. Gruen J. Inorg. Nuclear Chem. 1957 4,74. GRUEN FUSED-SALT SPECTROPHOTOMETRY 357 example one extends the region accessible to measurement from the 7000 cm.-l limit imposed by the presence there of an intense water- absorption band to ca. 4000 crn.-l where a strong nitrate ion absorption band is located. Carnallg has pointed out that in the 7000 to 4000 ern? interval there are some two dozen excited multiplets and multiplet com- ponents above the ground states of the ions Pr3+ to Ybw. Since there are only about a dozen additional excited multiplet components at energies below 4OOO cm.-l the rewards to be gained by pushing further into the infrared region are proportionately much less froin the point of view of analytical chemistry. Stewart and Kato20 and Banks et aL2lU have made detailed studies of the aqueous-solution spectra of the lanthanides.Banks et aZ.,21b Carnall,9 and Young and White22 have recently studied 4f ion spectra in chloride nitrate and fluoride melts respectively. The advantage of molten-salt solvents for these studies due to the extension of measurements to the near-infrared region has already been pointed out and is confirmed by the work of these investigators. However the problems in working with molten-salt solvents needs to be pointed out as well. The first consists of the increased difficulty in experimental technique with regard to solvent and solution preparation temperature control and container materials. A second disadvantage is due to the observation that the bands in the fused-salt solvents often tend to be less intense than their counterparts in aqueous solutions of equal concentrations.However several exceptions to this generalisation occur and give rise to potentially important analytical possibilities. Carnallg has concluded from his spectrophotometric study that “by having recourse to both aqueous and molten LiN03-KN03 solutions the quantitative analysis of even relatively complex mixtures of heavy and light lanthanides should be possible”. Spectrophotometric Determination of Mixtures of Oxidation States in Fused-salt SOlutionS Each oxidation state of a transition-metal ion possesses a different electronic configuration and is therefore characterised at least in principle by a unique absorption spectrum. Because of this fact it is possible to analyse spectrophotometrically not only mixtures of cations of different elements but also mixtures of different oxidation states of the same element.The law of additive absorbances states that in a solution of several light- absorbing species each species contributes additively to the absorbance. If this law holds and if the spectra of the different oxidation states are 2o D. C. Stewart and D. Kato Analyt. Chem. 1958,30 164. 21 (a) C. V. Banks J. L. Spooner and J. W. O’Laughlin Analyt. Chem. 1958 30 458; (6) C. V. Banks M. R. Heusinkveld and J. W. O’Laughlin Andy?. Chem. 1961 33 1235. aa J. P. Young and J. C. White Analyt. Chem. 1960,32,799. 358 QUARTERLY REVIEWS sufficiently different from each other spectrophotometric analysis is possible. Only one illustrative example will be cited here. S w a n ~ o n ~ ~ has studied the spectra of plutonium in various oxidation states in the LiCI-CsCl eutectic.He was able to characterise the III IV v and VI oxidation states of Pu spectrophotometricaIly. Only absorbance data are given in this paper but there appear to be spectral regions in which absorption is due predominantly to one oxidation state and it may be feasible therefore to determine the relative concentrations of Pu oxidation states in fused chloride melts. Spectra of Intermetallic Compounds in Fused-saIts The interesting phenomenon of solubility of intermetallic compounds in molten salts was first described by Heymann and his ~ ~ l l a b o r a t o r ~ . ~ Earlier Zintl et al. observed that a large number of intermetallic com- pounds form solutions in liquid ammonia.27 Absorption spectra of LiaBi dissolved in LiCl and in LiCl-LiF mixture were reported by Foster et a1.28 and of alkali-metal tellurides dissolved in molten alkali halides by Gruen et al.29 The spectra of the alkali-metal tellurides have absorption bands in the region 22,000 to 15,000 cm.-l.These transitions can be interpreted in first approximation as essentially atom-like excitations of the sort 5pg -+ 5p56p centered on tellurium. Trends of atomic energy levels with degree of ionisation in this region of the Periodic Table lead to the con- clusion that the absorbing species carry a negative charge and are probably to be identified with entities such as MTe- where M is an alkali metal. Molar extinction coefficients of the coloured species in these solutions have been obtained but further work is required to establish the functional relation between absorptivity and concentration.Spectra of Non-metallic Species in Fused-salts A number of spectra of non-metallic ions have been characterised in molten salts. By far the most extensively studied non-metallic ion is the nitrate ion. An exhaustive study of the effect of the cation on the energy of the lowest-lying nitrate ion transition at - 300 mp has been carried out by Smith and B o ~ t o n . ~ ~ ~ ~ Rhodes and Ubbelohdes have studied the energy J. L. Swanson J. Phys. Chern. 1964,68,438. 8' E. Heymann and H. P. Weber Nature 1938,141 1059. as E. Heymann and H. P. Weber Trans. Furuduy SOC. 1938,34 1492. B6 E. Heymann J. L. Martin and M. F. R. Mulcahy J. Phys. Chern. 1943 47,473. B7 E. Zintl J. Goubeau and W. Ropt Z. physik Chem. 1931 154 A 1-46.ee M. S. Foster C. E. Crouthamel D. M. Gruen and R. L. McBeth J. Phys. Chern. 1964,68,940. 29 D. M. Gruen R. L. McBeth M. S. Foster C. E. Crouthamel Abstracts of the 148th National American Chemical Society Meeting Chicago Illinois September 1964. so G. P. Smith and C. R. Boston Ann. New Yorlc Acad. Sci. 1960 79 930. 31 G. P. Smith and C. R. Boston Discuss. Furaduy Soc. 1961 32 14. GRUEN FUSED-SALT SPECTROPHOTOMETRY 359 change of the maximum of one of the nitrate absorption bands which occurs on melting. These workers also studied the spectrum of the thio- cyanate ion in molten KSCN.6 The energies and oscillator strengths of the nitrate ion absorption bands depend on the ionic potential of the melt cations. A detailed understanding of these spectral changes is important to a better understanding of theories of melting and of melt structure as shown by the recent work of Rhodes Smith and Ubbelohde.6b The spectrum of nitrite ion has been studied30 in molten KNO,.The band maximum occurs at 359 mp with E = 30.6. In dilute aqueous solu- tions these values are 366 mp and 21.9 mp respectively while in crystalline KNO the band maximum occurs at 360 mp. Solutions of halogen gases in molten alkali-metal halides have been measured by Greenberg and S ~ n d h e i m ~ ~ and were interpreted to show the presence of trihalide ions. Chlorine gas dissolved in the LiC1-KCl eutectic and bromine dissolved in the LiBr-KBr eutectic gave spectra that were similar to the trihalide spectra in aqueous ~ o l u t i o n s . ~ ~ ~ ~ ~ The spectrum of sulphur dissolved in molten LiCl-KCI LiBr-KBr and KSCN has been The sulphur solutions are blue and it was postulated that diatomic S molecules give rise to the colour.Spectrophotometers for High-temperature Measurements The chief instrumental problem that arises in measurements of absorp- tion spectra at high temperatures is due to the emission of radiation by the sample sample container and furnace. It will be instructive to follow the development of instrumentation to cope with this problem. The first instruments used in fused-salt spectrophotometry were manually operated devices modified by replacing standard sample compartments by heated compartments. The sequence of major components in these instruments is Light source + Monochromator -+ Sample -+ Detector (CONFIGURATION I) Measurements of transition-metal ion spectra in nitrate r n e l t ~ ~ ~ ~ ~ showed that emission of radiation did not constitute a problem up to 250"c.However with Configuration I the detector "sees" the entire spectrum of emitted radiation in addition to the transmitted monochro- matic beam. The signal (monochromatic beam) to noise (emitted radiation) ratio rapidly diminishes in the range 2 5 0 - 5 0 0 " ~ . ~ 32 J. Greenberg and B. R. Sundheim J. Chem. Phys. 1958,29,1029. 33 L. I. Katzin J. Chem. Phys. 1952 20 1165. 84 G. Zimmerman and F. C. Strong J. Amer. Chem. Soc. 1957 79 2063. 85 J. Greenberg B. R. Sundheim and D. M. Gruen J. Chem. Phys. 1958,29,461. 36 D. M. Gruen Nature 1956 178 1181. D. M. Gruen J. Inorg. Nuclear Chem. 1957 4 74. B. R. Sundheim and J. Greenberg Rev. Sci. Instr. 1956 27 703. 360 QUARTERLY REVIEWS The effect of emitted radiation on instrument performance is somewhat improved by adopting the sequence Light Source -+ Monochromator -+ Chopper -+ Sample -+ Detector -+ A.C.Amplifier -+ Recorder (CONFIGURATION 11) Most recording spectrophotometers provide this sequence. Adaptations of several automatically recording spectrophotometers for fused-salt work have been d e ~ c r i b e d . ~ ~ ~ ~ ~ ~ ~ ~ The reduction of interference from emit- ted radiation with Configuration I1 is due to the fact that the light beam is pulsed before passing through the heated sample. In principle this system should operate successfully even at very high temperatures since only the A.C. output of the phototube detector is amplified. In practice however the upper temperature of operation is limited to about 750"c because the detector is heated by the complete spectrum of thermal radiation from the furnace sample and sample container.Although the signal from this radiation is not amplified because of the nature of the circuit the heating of the detector increases its noise level and diminishes the resolution. Substantial gains in extending the range of measurements to higher temperatures were made by adopting the following configurations Light Source -+ Sample -+ Monochromator 3 Detector (CONFIGURATION 111) Light Source -+ Sample -+ Chopper -+ Monochromator -+ Detector 3 A.C. Amplifier -+ Recorder (CONFIGURATION IV) Modifications to operate in Configurations 11142 and W7 made it pos- sible to measure absorption spectra to temperatures of about 1050"~.As a consequence of operating an instrument in Configuration IV the emitted radiation is made to pass through the monochromator before striking the detector. The intensity of the emitted radiation to which the detector is exposed is therefore greatly reduced at temperatures comparable with those encountered in Configurations I and 11. However since the thermal radiation is chopped in Configuration IV at the same frequency as the transmitted beam it will be picked up by the detector and amplified as part of the signal. Planck's equation for black-body radiation allows one to make an approximate c a l ~ u l a t i o n ~ ~ ~ * ~ of the ratio I E / I O where I is the intensity of the furnace and sample radiation and Zo is the intensity of radiation from C. R. Boston and G. P. Smith J. Phys.Chem. 1958,62,409. 40 J. P. Young and J. C. White Analyt. Chem. 1959 31 1892. *l J. R. Morrey and A. W. Madsen Rev. Sci. Instr. 1961 32 799. D. M. Gruen and R. L. McBeth J . Inorg. Nuclear Chem. 1959,9,290. 43 J. R. Morrey and E. E. Voiland Spectrochim. Acta 1962,18 1175. GRUEN FUSED-SALT SPECTROPHOTOMETRY 36 1 the tungsten lamp. At a given furnace temperature T the ratio I E / J o is of course wavelength dependent. For example at a furnace temperature of 2400"c I E / I o has the values 0.02 and 0.09 at 1500 mp and 2500 mp respectively. Suppose however that a fused-salt solution of a transition metal ion at 1OOO"c has an optical density of 1 and this absorbs 90% of the incident light at a wavelength of say 1500 mp. The intensity of the thermally emitted radiation under these conditions constitutes 20 % of the signal.Since the instrument does not discriminate against the (mono- chromatised) emitted radiation a correction has to be applied to the measured spectrum at every wavelength. The virtual elimination of emitted radiation from the signal detection and amplification system can be achieved by using the sequence Light Source 3 Chopper -+ Sample -+ Monochromator -+ Detector -+ A.C. Amplifier -+ Recorder (CONFIGURATION V) This arrangement which has for example been used in atomic-absorp- tion spectro~copy,~~ and is the conventional mode of operation of some infrared spectrometers has two advantages. It eliminates emitted radia- tion except at the wavelength transmitted by the monochromator and amplifies only the radiation emanating from the light source.The performance of an instrument embodying these features,45 the Cary 14H has recently been t e ~ t e d . ~ ~ ~ ' A tungsten filament light bulb was placed in the sample light beam and the "spectrum" was measured with the filament at room temperature and then at 2200"~. The bulb filament simulated a glowilrg furnace and sample assembly at that tem- perature. The only effect on the performance was a slight increase in noise level showing that the instrument discriminates essentially 100 % against emitted radiation. With an instrumental arrangement such as given in Configuration V absorption spectra can be measured up to 2200"~ and probably well above that temperature. The special optical arrangement of this instrument results in a lowered light intensity at the detector. However by substituting a Sylvania Sun Gun for the more conventional tungsten lamps the light intensity in the near infrared and visible region of the spectrum is increased sufficiently so that normal slit widths are obtained.Furnaces for Spectrophotometers A variety of furnace designs for use in fused-salt spectrophotornetry 44 B. J. Russell J. P. Shelton and A. Walsh Spectrochim. Acta 1956,8 317. 46 Described in Bulletin No. 114-H Applied Physics Corporation Monrovia O6 C. R. Boston and G. P. Smith Rev. Sci. Instr. 1965 36 206. 47 D. M. Gruen unpublished observations. have been described.41 California. 2 362 QUARTERLY REVIEWS A furnace operating at temperatures up to 1450"~ has recently been designed and operated successfully using Pt-20% Rh wire as the heating element.46 An exploded view of a furnace with excellent temperature stability at temperature up to 1050" is shown in Fig.5. The water-cooled furnace FIG. 5 . Exploded view of filrnace assembly for absorption spectra measurements at temperatures up to 1000". 400w firerod heaters. 310 Stainless steel inner and outer walls of insulation container. Positioning screws for furnace. Thermocouple wells. Terminal box. Pt-Pt :10 % Rh thermocouple leads to temperature controller. Water-cooled brass furnace housing. To supply of inert gas. 0 ring seals. Quartz windows. Gold gasket. Keeper plate. 310 Stainless steel cell holder. P,. To light source. Q To Cary Model 14 Spectrophotometer. Dotted regions represent powdered zirconia insulation and hatched regions the 3 10 stainless steel furnace block. GRUEN FUSED-SALT SPECTROPHOTOMETRY 363 compartment (22.9 x 26.7 x 40-7 cm.) is made of 0.64 cm.brass through which the cooling water The light ports in the furnace compart- ment have quartz discs sealed into the openings. A lucite box (not shown) fits on the top of the furnace compartment and is connected to a laboratory exhaust duct thus providing a leak-proof system for obtaining spectral measurements on radioactive samples. The furnace itself is constructed of 310 stainless steel and is made in three parts. The two outside sections each contain eight 400-watt Firerod heaters while the centre section is machined to accommodate a cell holder for 1 cm. square absorption cells. The furnace is so designed that the centre section can be removed and replaced with other sections designed to hold cells of 2 5 or 10 cm.light path. These three parts of the furnace are contained in a double-walled 310 stainless steel container filled with powdered zirconia insulation. Insulation on top of the furnace is provided by three sealed 310 stainless steel cans also filled with zirconia insulation. Only the sample chamber is maintained at the high temperature. Power to the furnace is controlled through a self-saturating saturable reactor (choke) with a chopper amplifier. A Pt-Pt 10 % Rh thermocouple is used as a sensing element bucked against a reference source which is a 12 mv manganin wound helipot. An anti-oscillation control on the ampli- fier prevents hunting and large overshoots. Provisions exist for pre-heating the furnace rapidly with an automatic switchover to the temperature con- troller as the temperature approaches a predetermined point.A waterflow interlock system controlled by a solenoid valve turns off the power to the heaters reactor light and water sources should a leak develop anywhere within the system. The heater block is connected so that only 12 heaters are controlled and four heaters are available for auxiliary heat. The excellent stability of the controller coupled with the high heat capacity of the furnace resulted in temperatures which were constant to st0*lo for periods of several hours over the entire temperature range. Temperature Measurement.-Temperature of the molten salt inside a quartz cell was determined and controlled in the following way A Pt-Pt 10 % Rh thermocouple was calibrated according to the method reported by Roeser and WenseP of the National Bureau of Standards using standard samples of tin lead zinc aluminium and copper metals and a secondary standard of silver metal.This thermocouple was inserted in a thin-wall quartz tube immersed in molten salt in a cell in the operating position of the furnace. The e.m.f. of this thermocouple was measured on a Leeds and Northrup Double Range Potentiometer Model No. 8662 against a number of settings of the helipot on the temperature controller. After applying the corrections to the e.m.f. a graph was constructed show- ing the temperature of the molten salt as a function of the helipot settings. ‘~3 W. F. Roeser and H. T. Wensel in “Temperature-Its Measurement and Control in Science and Industry,” American Institute of Physics vol. 1 Reinhold Publ.Corp. New York N.Y. 1941 pp. 284-314. 364 QUARTERLY REVIEWS Although any temperature setting is maintained within 0.1" by the con- troller it is calculated that the temperature within the cell is &lo from any desired temperature. Spectrophotometric Determination of Solute Concentrations in Fused-salts For a particular fused-salt solvent medium and a given temperature the syectrophotometric determination of the concentration of a solute species involves the Lambert-Beer law of ideal behaviour. It is conventional in solution work where the interest is in the absorption by the solute apart from that by the solvent to define the absorbance A as where Psoln and Psolv is the radiant power transmitted through equal thicknesses of solution and solvent when both are irradiated with the same incident power.The Lambert-Beer law states that for a sample path- length by the absorptivity a a = A/cb (2) A = EbC (3) If 6 is in centimetres and c is in units of moles/litre one can write where e is the molar absorptivity at a particular frequency u. It is often desirable to measure the integrated absorption of a single electronic absorption band. The oscillator strength or f number of that band is given by cg f= 4.32 x 10-~ I Edy 0 The integral is taken over the range of v-values for which E is appreciable. The oscillator strength defined in this way does not take into account the effect of the medium on the incident light wave and a small correction term depending on the particular fused-salt solvent must be introduced. For allowed transition f numbers are near unity.However the inner- shell electronic transitions of the transition-metal ions for example have varying degrees of forbiddences and their f numbers range from ca. to ca. lo-*. In order to do quantitative spectrophotometry A b and c must be accurately determined and it must be shown that E is independent of c in the concentration range of interest. In conventional aqueous-solution spectrophotometry the absorbance of the solvent is automatically subtracted from that due to the solute species by having pure solvent in the reference path. Furthermore at ambient temperatures no complications arise owing to emitted radiation. In addi- tion aqueous-solution work allows one to use quartz or Pyrex cells with GRUEN FUSED-SALT SPECTROPHOTOMETRY 365 well defined path lengths. It is also usually relatively simple to make independent determinations of solute concentrations and to verify the linear relationship between A and c.The situation in the case of fused-salt spectrophotometry is not nearly so straightforward. Nonetheless the development of instrumentation and techniques for high-temperature spectrophotometry does make it possible to measure molar absorptivities to high accuracy in fused-salt media. In some cases such as the determination of molar absorptivities of transition-metal ions in chloride melts at temperatures up to 1000"~ the same accuracy has already been achieved as on aqueous solutions at room temperature. The full exploitation of the technique for quantitative spectroscopic measurements on melts however necessitates a detailed consideration of the special problems which arise in connection with fused-salt spectro- photometry.These problems will be dealt with in the subsequent sections of this Review. (1) Measurement of Absorbance at High Temperatures.-At high tem- peratures absorbance measurements are complicated by the emission of radiation from the sample sample container and furnace. There are three mechanisms by which emission occurs (1) Spontaneous discrete emission ; (2) Induced discrete emission ; and (3) Thermal emission. Spontaneous and thermal radiation are not while induced emission is dependent on the intensity of external radiation. In an earlier section it was noted that proper design of instrumentation allows one to dis- criminate 100 % against spontaneous and thermal radiation. When operating a spectrometer in Configuration V spontaneous and thermal radiation appear as D.C.signals and are not registered in the A.C. detection-amplification system. No correction for emission from these two sources need therefore be applied to absorbance measurements made under these experimental conditions. Induced emission on the other hand exhibits the same chopped fre- quency and direction as the incident light and will be detected by the instrument. This problem has been considered by Morrey and V ~ i l a n d ~ ~ who found that the correction for induced emission is less than 1 % when v(cm.-l)/T/"K) > 3.1 5. No appreciable error is therefore introduced by neglecting this correction at temperatures below 100O"c and wavelengths less than 2500mp. It is frequently experimentally convenient to measure the absorption spectra of the solution and solvent separately against an air reference.This is due to the fact that it is far easier to design a furnace in such a way as to heat only the cell containing the solution while leaving the reference path unheated and unobstructed. With such an experimental set-up the solute absorbance is obtained by subtraction. The subtraction procedure is relatively simple if all spectra are digitised on cards or tape since the subsequent calculations can then be carried out by a computer. 366 QUARTERLY REVIEWS Using a properly designed spectrophotometer and furnace and employing a carefal subtraction technique for solvent absorbance solute absorbance can be measured at wavelengths less than 2500 mp and at temperatures up to 1000”~ with the same accuracy as at ambient temperatures.(2) High-temperature Optical Cells.-A container for solutions on which quantitative spectrophotometric measurements are to be performed at high temperatures must meet four basic requirements (1) It must be transparent to electromagnetic radiation in the wavelength region in which measure- ments are to be made; (2) its softening or melting point must be consider- ably above the highest temperature at which measurements are contem- plated; (3) it must be chemically inert to attack by the fused-salt solutions ; (4) it must have either a precisely defined optical path length or one which can be determined by calibration. It is not surprising that these requirements are to some extent mutually exclusive and that they limit the choice of container materials which have been used or have been suggested for use together with their range of applicability both insofar as temperature and type of melt is concerned.Since requirements 1 and 3 optical transparency and inertness to chemical attack are found only infrequently in one and the same material con- siderable efforts have been devoted to developing “windowless” cells. In Table 2 materials of construction for “windowless” cells are grouped in a separate category. TABLE 2. Container materials for high-temperature optical cells. Material Pyrex Quartz Sapphire Magnesium oxide Diamond Material Gold Platinum Solvents and solutions Nitrates Sulphates chlorides bromides iodides solu- tions of some intermetallics Chlorides chloride-fluoride mixtures metal- molten salt solutions Same as sapphire but useful for alkali-fluoride melts Unvaluated Windowless cells.Solvents and solutions Hydroxides Fluorides oxides carbonates borates silicates upper temp. limit 500” 1200 1800 2500 Upper temp. limit 800” 1400 In order to determine molar absorptivities using the relation A = Ebc one must know the optical path through the sample. The path length is well defined when it is determined by the distance between plane parallel windows. In “windowless” cells however it is not simple to determine the path length accurately and calibration procedures must be employed. GRUEN FUSED-SALT SPECTROPHOTOMETRY 367 (a) Cells with plane parallel windows. As outlined in Table 2 a number of optically transparent oxidic materials can serve as container materials for fused salts.The Pyrex or fused silica absorption cells used with aqueous solutions are unsuitable for fused-salt work because they are not vacuum tight and the edges are usually cemented with a material which either decomposes or melts at elevated temperatures. Rectangular 1 cm. path length fused silica cells for high temperature work are now available commercially.* These cells are made of 100 % fused silica and are vacuum tight at least up to 1O00”c. The 1 cm. path length of these cells can be reduced by using readily available fused silica spacers. Procedures for filling cells of this type with fused-salt samples have been described in the 1iterature.l’ Cells for solutions which attack fused silica can be machined from single crystal sapphire or magnesium oxide. Both of these materials have been shaped into rectangular cells having rectangular or square wells with plane parallel ~ i d e ~ .~ ~ 9 ~ ~ Ultrasonic machining techniques have been employed for this purpose. Molten hydroxides oxides silicates certain fluoride melts and some metal-molten salt solutions attack sapphire as well as MgO. To study highly corrosive liquids of this sort “windowless” metal cells have been developed. The most thorough study of this subject is due to Young and his c ~ l l a b o r a t o r ~ ~ ~ ~ ~ ~ ~ ~ ~ although several other methods have also been d e s ~ r i b e d . ~ ~ ~ ~ ~ ~ ~ ~ In the “captive liquid cell” developed by light is transmitted through the side of a vertical column of liquid which is held in a metal cylinder container of special design. No window material or mirrors are necessary to the operation of the cell.The path length of liquid samples held in this cell is independent of the weight or volume of sample initially placed in the cell but is postulated to be dependent on the surface tension and density of the liquid. Based on absorbance measurements in 0.250 inch (0.635 cm.) outside diameter cells the path length of aqueous solutions is 0.68 -+ 0.02 cm.; the path length of the molten eutectic LiF-NaF-KF at 540”c is 0.77 & 0.02 cm. The cell has been used for spectrophotornetric studies of molten fluoride salts. (b) “ Windowless cells”. (3) Concentration of Solute Species.-The concentration of a solute species must be determined by an independent method before its molar absorptivity can be calculated. However determining the concentration of a solute species in a salt matrix is often not a trivial analytical problem.Indeed the difficulty associated with such determinations serves as a * Fisher and Porter Company Warminster Pennsylvania ; Pyrocell Manufacturing Co. New York. *t~ J. P. Young and J. C. White Analyt. Chem. 1960,32 1658. so J. P. Young Analyt. Chem. 1964 36 390. 6L W. Bues 2. anorg. allgem. Chem. 1955 279 104. IP J. Greenberg and L. J. Hallgren Rev. Sci. Instr. 1960 31,444. 368 QUARTERLY REVIEWS stimulus to using fused-salt spectrophotometry as a quantitative analytical method. In whatever way the amount of solute present is determined it is convenient to express its concentration in a fused salt solution in the con- ventional units of moles/litre of solution. To make this calculation the density of the melt as a function of temperature must be known and unless available must be obtained in separate experiments.Finally to apply the relation A = E ~ C for quantitative determinations of concentration Beer’s law which states that at a fixed wavelength A and E are constants independent of concentration must be verified. Only a few studies have been made to check Beer’s law in a precise way for solutions in melts. In one instance Boston and Smith39 measured the visible spec- trum of NiCl in the fused LiCl-KC1 eutectic at seven concentrations ranging from 0.01 to 0.4~. Beer’s law was accurately obeyed at all wave- lengths. (4) Preparation and Purification of Fused-salt Solutions.-The need to prepare fused-salt solutions of high purity is of concern for investigations of the basic physical chemistry of these systems as well as for purposes of analysis. It would carry us too far afield here to describe in detail the various and often ingenious methods that have been developed for the preparation and handling of fused-salt solvents and solutions. A survey of some of this work with literature references is given by Corbett and The original literature should be consulted for specific procedures. This Review is based on work done under the auspices of the U.S. Atomic Energy Commission. 63 J. D. Corbett and F. R. Duke “Technique of Inorganic Chemistry,” ed. H. B. Jonassen and A. Weissberger ch. on “Fused Salt Techniques,” vol. I Interscience Publishers New York 1963.

ISSN:0009-2681    

DOI:10.1039/QR9651900349

出版商:RSC    

年代:1965

数据来源: RSC

摘要:

MACROMOLECULAR STRUCTURE AND PROPERTIES OF DEOXYRIBONUCLEIC ACTD By P. A. EDWARDS and K. V. SHOOTER FULHAM ROAD LONDON S.W.3.) (CHESTER BEATTY RESEARCH INSTITUTE INSTITUTE OF CANCER RESEARCH DURING the last decade the role of deoxyribonucleic acid (DNA) as the carrier of genetic information has been studied in great detail. The essential simplicity of the basic metabolic process whereby DNA acts as a template for the synthesis of messenger RNA (ribonucleic acid) which in turn acts as a template for protein synthesis has been shown to be involved in a complex web of control mechanisms in which it now seems probable that proteins can control the operation of sections of the DNA template so that specific enzymes are made only when required by the cell. The key to the elucidation of the properties of DNA has b=en the structure proposed by Watson and Crick.l As is shown in Fig.1 native DNA consists of two polyphosphate chains the phosphate groups linking the 3’,5’ positions of neighbouring deoxyribose moieties. The bases are linked to the 1’ position of the sugar and the two chains are held together by hydrogen bonds formed between pairs of bases. The fundamental requirement is that adenine (A) on one chain pairs with thymine (T) on the other and guanine (G) with cytidine (C). It should be noted that the two chains run in opposite directions (antiparallel).* In the hydrated molecule the pairs of bases lie stacked on top of each other in planes at right angles to the long axis of the molecule; the two chains are coiled one full turn being made every ten pairs of bases.Replication of the molecule involves the separation of the two chains and the synthesis of a new complementary strand on each the composition of the new strands being dictated by the requirement for the pairing of A with T and G with C. In contrast to this the formation of single-stranded messenger RNA on the DNA template does not appear to require the separation of the two chains of the DNA It is possible that the messenger RNA is complementary to only one of the strands of DNA i.e. that one strand contains the genetic message the other strand being a nonsense code. The preparation3 of DNA involves disrupting the cell or organism the isolation of DNA or DNA-protein complex free from other cell con- stituents particularly RNA and polysaccharides and the removal of the protein bound to the DNA.The main methods so far applied depend upon *The abbreviations A T G and C have been used for adenine thymine guanine and cytosine and for the corresponding deoxyribose derivatives where no confusion occurs. dATP dCTP dGTP and dTTP are the deoxyribose triphosphates of the four bases. poly-dAT is a twin-stranded synthetic polymer with regularly alternating deoxy- adenosine and thymidine linked 3‘-5’ by phosphate groups in each chain. poly-dGdC is a twin-stranded synthetic polymer of deoxyguanosine and deoxycytidine one strand containing only guanine moieties and the other cytosine only. 1 J. D. Watson and F. H. C. Crick Nature. 1953 171 737. 2 J. Hurwitz and J. T. August Progr. Nucleic Acid Research 1963 1 53 K. S . Kirby Progr. Nucleic Acid Reseclrch 1965 3 1 3 69 370 QUARTERLY REVIEWS I I I I I I I H I I I I I I I 1 I 2 5’ FIG.1. A schematic diagram of purt of a DNA helix showing the pairing of adenine with thymine and guanine with cytosine. Reading from the top of the diagram the pairs of bases shown are guanine-cytosine thymine-adenine adenine-thymine and cytosine- guanine. the use of either a detergent or of phenol to inhibit enzyme action and to denature and remove protein. It has always been accepted that the pre- paration should have a high molecular weight and a low protein content. Most preparations of mammalian DNA with less than 0.1 % protein were found to have molecular weights in the range 6-8 x lo6. The presence of some residual protein in the most extensively purified preparations posed the question whether small proteins or amino-acids were part of the structure of DNA linking units of the molecule end to end within the cell and whether breaking occurred at some of these points during isola- t i ~ n .~ ~ However no evidence of a deoxynucleotide covalently linked to an A. Bendich and H. S. Rosenkranz Progr. Nucleic Acid Res. 1963 1 219. C. Sadron J. Chim. phys. 1961 58 877; G. Bernardi and C. Sadron “Acidi nucleici e loro funzione biologia,” Istituto Lombardo 1964 62. EDWARDS AND SHOOTER DEOXYRIBONUCLEIC ACID 37 1 amino-acid has yet been found. Within the last few years it has been shown that the whole of the DNA of a bacteriophage can be isolated as a single unkg A homogeneous preparation of DNA from T2 bacteriophage can be obtained with a molecular weight of 120 x los. Following this Cairns’ published radioautographs showing the isolated DNA of the bacterium Escherischia coli as a single unit of length 400 p ( M = 1000 X lo6).Extreme care has to be taken in the isolation of DNA of these high mole- cular weights to avoid degradation of the molecules by shear. It appears that the polydisperse preparations of DNA normally obtained from mammalian sources are degradation products of much larger molecules. The possibility that the DNA of a single chromosome may be a single unit cannot be ignored. Homogeneous preparations of DNA of high molecular weight have not so far been made in sufficient quantity to allow analysis for the presence of protein residues. Careful isolation techniques in which shearing is avoided have also shown that in some cases the DNA molecule exists as a closed circle.The existence of circular DNA has been demonstrated in the bacteriophage A> in the bacterium E. cdi7 and in the polyomag and papilloma viruses.lO The DNA of the bacteriophage 4x174 is single-stranded and both this form and the double-stranded molecule formed after infection of a bac- terium as a preliminary to replication are circular.ll The use of density-gradient centrifugation for characterising DNA preparations has proved to be extremely useful.12 In these experiments the DNA is dissolved in a high concentration of one of the soluble czsium salts usually the chloride or sulphate. Centrifugation at speeds of about 30,000-40,000 r.p.m. generates in time an equilibrium density gradient of the salt the DNA sedimenting to form a band within the gradient. It has been found that the buoyant density of DNA preparations is a linear function of the base composition increasing with G + C content.By including a marker DNA of known density and composition the technique can be used to determine the average base composition of a preparation. Theoretically a homogeneous preparation should exhibit a Gaussian distribution at equilibrium the variance being inversely related to the molecular weight. Attempts have been made to use this method for de- termining the molecular weight of the homogeneous preparations from bacteriophage but difficulties are encountered owing to light scattering by the large molecules. The variance was found to depend on the wave- length of the light used in the absorption optical system and measurements A. D. Hershey and E. Burgi J.Mol. Biol. 1960,2 143. J. Cairns J. Mot. Biol. 1962 4 407; 1953 6 208. C. A. Thomas “Acidi nucleici e lor0 funzione biologia,” Istituto Lombardo 1964 R. Dulbecco “Acidi nucleici e lor0 funzione biologia,” Istituto Lombardo 1964 lo J. D. Watson and J. W. Littlefield J. Mol. Biol. 1950 2 161. l1 W. Fiers and R. L. Sinsheimer J. Mol. Biol. 1962 5 424; A. K. Kleinschmidt la J. Vinograd and J. E. Hearst Fortschr. Chem. org. Naturstofle 1962 20 372. 17. p. 278; R. Dulbecco and M. Vogt Proc. Nat. Acad. Sci. U.S.A. 1963,50,236. A. Burton and R. L. Sinsheimer Science 1963 142 961. 372 QUARTERLY REVIEWS could not be mads over a sufficient range of wavelengths to allow extra- polation to the correct value.13 With degraded preparations of lower molecular weight the variance of the distribution is increased by the density heterogeneity of the molecules resulting from differences in base composition.The method has however been used successfully to demon- strate that circular DNA from papilloma or polyoma virus has the same molecular weight as the linear form of these molecules. With degraded samples of bacteriophage and bacterial DNA (M about 20 x lo6) the bands are extremely sharp and small differences in overall composition between two species are sufficient to give completely separated bands. Preparations of mammalian DNA of molecular weight 10-20 x 106 give much broader bands. Detailed analyses have shown that the equili- brium distributions of some DNA preparations contain satellite bands. DNA from herring sperm and mouse tissue for example exhibits a satellite band lighter than the average (low G + C content) while guinea pig and calf thymus DNA have satellite baiids denser than the average.The light satellite band accounts for 30% of the DNA from the crab Cancer boreah and 11 "/o in C. irroratus. In the last two cases the minor band contains 93% adenine and thymine and has been found to correspond closely with the synthetic copolymer poly-dAT (see footnote p. 369) in which adenine and thymine residues alternate in each chain. The signific- ance of these satellite bands is not yet understood. From studies of the buoyant density in a variety of czesiurn salts Hearst and Vinograd14 were able to show that the net hydration of the cmium salt of DNA is 30 % by weight in the physiological range of water activities from 0.9-1-0. The observation that the DNA of a bacteriophage can be isolated as a unit has helped to solve many problems relating to the characterisation of DNA in solution.Attempts have been made to correlate the molecular weight of DNA with easily measurable parameters such as sedimentation coefficient and viscosity. The basis for the comparisons has usually been the equation derived by Mandelkern and Flory where so is the sedimentation coefficient f is the partial specific volume of DNA p and rlo the density and viscosity of the solvent and N Avogadro's number. /I is a factor dependent only on the geometry of the molecule e.g. for random coils /3 = 2.5 x lo6 and for stiff rods 3.1-3.4 x 106. Most determinations of molecular weight have been made by use of light- scattering. The primary difficulty here has been to decide how high a molecular weight can be measured.Recently Froelich et a1.l5 have con- structed a light-scattering apparatus with which measurements can be l3 D. J. Cummings Biochim. Biophys. Acta 1963 72 475. l4 J. E. Hearst and J. Vinograd Proc. Nat. Acad. Sci. U.S.A. 1961,47 825 1005. D. Froelich C. Strazielle G. Bernardi and H. Benoit Biophys. J. 1963 3 115. EDWARDS AND SHOOTER DEOXYRIBONUCLEIC AClD 373 made down to 16" from the main beam; the limit for most instruments previously used was 30". For DNA of molecular weight up to 6 x los measurements to 30" or 16" gave identical results. At hgher molecular weights the values calculated diverged rapidly e.g. one preparation gave iki = 8-5 x lo6 if data to 30" were used and M = 17 x lo6 from data to 16".Light-scattering measurements can therefore only be used with cer- tainty to establish the correlations between M and s and M and [q] up to about 6 x lo6. The high-molecular weight end of the curves can be estab- lished from measurements on the homogeneous preparations of DNA from bacteriophage. Careful measurements of the sedimentation coefficient and viscosity of preparations of DNA from T2 T4 and T7 bacteriophage have been made in several laboratories and good agreement has been achieved demonstrating that present techniques can be used with these very large molecules.1gJ7~1s Sedimentation studies have to be made at low rotor speeds to avoid intertwining of the molecules and the formation of rapidly sedimenting aggregates.le The zonal sedimentation technique of Vinograd et aZ.20 avoids this difficulty and eliminates the problems resulting from self-sharpening of the boundary due to the Johnson-Ogston effect.The molecular weights of the DNA preparations have been determined from measurements of the lengths of the molecules in electron micrographs from determinations of the specific radioactivity of the DNA and the activity of the whole bacteriophage and from determinations of the weight of DNA per bacteriophage (see ref. 18). The molecular weights obtained show considerable scatter T2 and T4 DNA M = 97-162 x los; T7 DNA 19-27 x lo6. Burgi and Hershey found that when the DNA of bacteriophage T2 is broken by carefully controlled shearing the breaks occur most frequently near the middle of the molecule giving a reasonably homogeneous solution of half molecules which can in turn be sheared to quarter molecules.21 These preparations provide further standards of known molecular weight.Eigner and Doty18 have reviewed all the published data and have found that in the range M = 0-3-130 x lo6 the intrinsic viscosity and sedimentation coefficient at infinite dilution are monotonic- ally related. The relationship between M and s and M and [q] changes with molecular weight; at high molecular weights the molecules behave as random coils with a transition towards the characteristics of stiff rods as the molecular weight decreases although even at M = 1-4 x lo6 the limit of rod-like behaviour is not reached. It seems probable that all samples of native DNA have a characteristic chain stiffness. The bending of the molecules to form random coils could be associated with a limited degree of flexibility of the twin-strand structure or to the presence of junction l8 D.M. Crothers and B. Zimm J. Mol. Biol. 1965 12 525. l7 J. B. T. Aten and J. A. Cohen J. Mol. Biol. 1965 12 537. l8 J. Eigner and P. Doty J. Mol. Biol. 1965 12 549. lS J. Rosenbloom and V. N. Sciumaker Biochemistry 1963 2 1206. 2o J. Vinograd R. Bruner R. Kent and J. Weigle Proc. Nat. Acad. Sci. U.S.A. 21 E. Burgi and A. D. Hershey J. Mol. Biol. 1961 3,458. 1963 49 902. 374 QUARTERLY REVIEWS points along the molecule at which free rotation is possible. No distinc- tion between these possibilities can be made from the above data since both models lead to the same functional relationship between s and [ r ] ] and M. A small proportion of breaks in the single strands of the helix does not appear to increase the flexibility of the molecule.So far preparations of twin-stranded circular DNA from only two sources have been studied in detail. It has been found that in each case because of the more compact structure imposed upon it the circular DNA has a higher sedimentation coefficient than the corresponding linear form papilloma DNA circular 28S linear 21S:lO polyoma DNA circular 20S linear 16XB The identity of the circular and linear forms of these molecules has been confirmed by measurements of the molecular weight from the spread of the equilibrium distribution curve in casium chloride density gradients. The stiffness of the twin-stranded DNA molecule may be contrasted with the flexibility of the single chains. Changes in the ionic strength of the solution for example have little effect on the sedimentation rate of native DNA (Fig.2) indicating that the shape of the molecule remains essentially I I I t 001 0. I 1.0 Ionic strength Fla. 2. The efect of ionic strength on the sedimentation coeficient at infinite dilution. A native DNA; B denatured single-stranded DNA at p H 8; C denatured DNA at p H 12-13. (Reproduced by permission from F. W. Studier J. Mol. Biol. 1965 11 373.) unchanged. With the single chains there is a marked increase in sedimenta- tion rate as the masking of the charge effects of the phosphate groups increases.22 At pH 8 interaction and stacking of the bases occurs leading to a much tighter coiling of the molecule than is observed at pH 12-13 2a F. W. Studier J. Mol. Eiol. 1965 11 373. EDWARDS AND SHOOTER DEOXYRIBONUCLEIC ACID 375 where charged groups on the nucleotides prevent the stacking of the bases.No change in the absorption at 260 mp occurs at pH 12-13 when the ionic strength is increased but at pH 8 the progressive ordering of the bases in parallel arrays is accompanied by a decreasing absorption. Using first-order perturbation theory T i n o ~ o ~ ~ has related this hypochromicity to the geometry of the transition moments of the nucleotide moieties. Treating the phenomenon as a local field effect D e ~ o e ~ ~ has demonstrated that the w-r* transition at 260 mp should exhibit hypochromism in the ordered structure of layers of bases and that the n-n-* transition near 280 m p should be hyperchromic. The effects of changes of pH on the properties of native and single- stranded DNA are illustrated in Fig.3. Titration of the bases of the single strands at high pH leads to an elongation of the molecule the increase in frictional resistance resulting in a decreased sedimentation rate.22 At low pH the beginning of a similar effect is observed but aggregation then ensues and the sedimentation rate increases. In native DNA titration of the bases occurs at more extreme pHs than for the single strands.25 As is shown in Fig. 3 the shape of the native DNA molecules remains constant I 0.1 M NaCl O005M buffer I I I 1 I 4 6 8 10 12 PH 201 FIG. 3. The efect of pH on the sedimentation coeficient at infinite dilution. native DNA; - - - - - - denatured DNA. (Reproduced by permission over a wider range of pH than is observed with the single strands. The maxima observed at high and low pH with native DNA are associated with the collapse of the twin helix and indicate the existence of a metastable structure in which the hydrogen bonding between the chains has been broken but the two chains are not completely separated.from F. W. Studier J. Mol. Biol. 1965,11 373.) 23 I. Tinoco J. Amer. Chem. Soc. 1960,82,4785; 1961 83 5047. e4 H. Devoe Biopolymers Symposia No. 1 1964 251. 25 A. R. Peacocke Progr. Biophysics Biophys. Chem. 1960 10 55. 376 QUARTERLY REVIEWS Denaturation of DNA,as the separation of the two chains can be achieved by raising or lowering the pH as described above or by heating or by lowering the ionic strength of the solution. The coiling of the separ- ated strands can lead to the formation of intermolecular aggregates and thus make it difficult to relate the size of the denatured product to that of the original material.Working at comparatively low ionic strength and at low concentration Eigner and Dotyl* found that the molecular weight of DNA from bacteria is halved upon denaturation. The integrity of the single chains of the DNA molecule is not however necessary for biological activity. Davison Freifelder and Hol10way~~ found that denaturation of DNA from bacteriophages T2 T4 and T5 gave a completely polydisperse product as judged by the shape of the sedimenting boundary. De- naturation of DNA from nine other sources tested yielded products containing 15-100 % of polydisperse material. Since the amount of poly- disperse material found was characteristic of the bacteriophage rather than of the method of denaturation it must be concluded that breaks can occur naturally in the single chains.8 TiTm Tmc Temperature (“C) FIG. 4. The changes in absorption at 260 mp observed when native DNA is heated and then cooled. The total increment in absorption observed on heating is about 40 % of the absorption of the solution of native DNA in the Figure the curves have been normalised and the changes in absorption are shown as fractions of the total increment observed on heating. The dashed lines drawn from the mid-points of the curves obtained on heating give the Tm values which are Characteristic of the base composition of the DNA. The DNA preparations are from A D.pneumoniae 38% G 4- C B E. coli 52% G 4- C C Ps. aeruginosa 66% G f C (Reproduced by permission from J. Eigner and P. Doty J.Mol. Biol. 1965 12 549) 26 J. Marmur R. Rownd and C. L. Schildkraut Progr. Nucleic Acid Research 1963 27 P. F. Davison D. Freifelder and B W Holloway J. Mot. Biol. 1964 8 1. 1 232. EDWARDS AND SHOOTER DEOXYRIBONUCLEIC ACID 317 The process of denaturation can be readily followed by observing the changes in absorption at 260 mp (Fig. 4). At a given ionic strength the mid-point of the hyperchromic change Tm is a characteristic of each DNA and increases with the G + C content of the DNA. Reducing the ionic strength of the solution progressively lowers Tm until at very low ionic strengths denaturation occurs spontaneously owing to the repulsive forces between the unshielded phosphate groups. When solutions of DNA which have been heated above the denaturation temperature cool the absorption at 260 m p falls largely as a result of the random coiling and stacking of the bases of the single chains.A partial reversal of denaturation and the re-forming of some double-stranded molecules does however occur if the temperature has not been raised high enough to separate all the twin- stranded regions. Denaturation in the presence of formaldehyde which prevents the re-formation of the twin helix shows that all the double- stranded regions are not melted out until the full hyperchromic effect has been achieved.28 If the two strands of the DNA helix are cross-linked (e.g. by reaction with mitomycin actinomycin or bifunctional alkylating agents or as occurs naturally in circular DNA) the separation of the two chains is prevented and reversible denaturation is observed.With mam- malian DNA one cross-link per log molecular-weight units may be re- quired to ensure complete re-formation of the double helix. The observation that the temperature of denaturation increases with G + C content was at first attributed to the presence of three hydrogen bonds linking guanine and cytidine as opposed to the two linking adenine and thymine. Studies of denaturation in the presence of a variety of solutes and solvents showed that effects on the melting temperature could not be correlated with the hydrogen-bonding capacities of the various agents. Substances which are not in the usual sense hydrogen-bond breaking agents can alter the differential stability of DNA containing a high pro- portion of G + C relative to DNA rich in A + T as is shown in Table l.29 These results indicate that the hydrogen bonding of the base pairs contri- butes little to the stability of the helix.Theoretical treatments of the thermo- TABLE 1 Solvent 6*5~-CF&0,Na versene pH 7 7*2~-NaCI0, versene pH 7 O-~SM-N~CI citrate pH 7 51 % (v/v) MeOH 10-3~-NaC1 10-3~-tris pH 7 81 % (v/v) MeOH 1O4M-NaCl 103M-tris pH 7 dTm/d (G + C ) 0.60 0.56 0-4 1 0.35 about 0 Slope of the curve relating denaturation temperature Tm of DNA to the content of guanine and cytidine (G + C) in different solvent systems. dTm/d(G + C) is the change in Tm for a change in (G + C) content of 1 mole %. (Reproduced by permission of K. Hamaguchi a d P. Geiduschek J. Amer. Chern. Soc. 1962,84 1329.) z8 D. Freifelder and P. F. Davison Biophys. J. 1962 2 249. 29 K. Hamaguchi and P. E. Geiduschek J.Amer. Chem. SOC. 1962,84 1329. 378 QUARTERLY REVIEWS dynamics of denaturation and the difficulties encountered have been summarised by Crothers and Zimm?O These authors have related the free-energy change which occurs on stacking hydrogen-bonded base pairs in the helix to the slope of the thermal denaturation curves. For the syn- thetic polymer poly-dAT the calculated free-energy change amounts to about -7 kcal./mole of base pairs compared with the estimate of -2 to -3 kcal./mole for the initial hydrogen bonding of the bases. The problem of the strength and nature of the bonding between the superimposed bases in the helix has been approached from quantum mechanical considera- tions. Calculations made by Ladik and Hoffman31 show that overlap of the 7~ orbitals of the bases occurs but that the interaction is non-bonding in character.Assuming a van der Waals’s type of interaction Devoe and T i n o ~ o ~ ~ found that dipole-dipole dipole-induced dipole and London force interactions between the bases are large and that as is observed the stability of the helix should increase with G + C content. Although therefore the hydrogen bonding between the pairs of bases is essential in dictating the specificity of the structure these bonds contribute Iittle to the stability of the helix. If solutions of denatured bacteriophage DNA are held at a temperature 25” below Tm for several hours realignment of some of the molecules and the formation of twin-stranded DNA occurs.26 Theoretical considera- tions suggest that the process begins by a slow second-order reaction involving the formation of nucleation centres followed by the bonding of short sections of the molecules with appropriate base sequences and this is followed by an annealing process in which further sections of the chains become properly matched.This reversal of denaturation occurs only to a limited extent with bacteriophage DNA to a much smaller degree with bacterial DNA and has not yet been demonstrated with mammalian DNA. The greater the amount of DNA required by a cell or organism to express its genetic complexity the lower is the probability that complementary strands will meet during the annealing. When DNA from E. coli grown in a medium containing nitrogen-14 is mixed with DNA from the bacterium grown in medium containing 15N denatured and annealed half of the “renatured” DNA molecules formed are labelled in one strand with 14N and in the other strand with 15N (Fig.5). Formation of native DNA from complementary strands from different cells is thus possible. No hybrid formation has been detected between DNA molecules from organisms which are not genetically related even though the base composition of the DNA may be the same for both. The existence within the bacterial cell of RNA complementary in composition to the DNA has been demonstrated by the formation of RNA-DNA hybrids under the conditions for the renaturation of DNA described above. If bacteria or cells from dividing tissues are disrupted and the particu- 30 D. M. Crothers and B. Zimm J. Mul. Biol. 1964,9 1. 31 J. Ladik and T. A. Hoffman Biopolymers Symposia No. 1 1964 117. 32 H. Devoe and I. Tinoco J.Mol. Biol. 1962 4 500. EDWARDS AND SHOOTER DEOXYRIBONUCLEIC ACID 379 1.744 1.725 1.704 1 DcnsIty FIG. 5. Banding of native and renatured DNA from E. coli in cLesium chloride. The DNA of coli grown in medium containing l6N (upper figure) has a higher density than DNA from coli grown in normal medium (middle figure). Samples of these two preparations of DNA were denatured mixed and allowed tc renature together. Before banding denatured DNA left was degraded preferentially using E. coli phosphodiesterase. The lower figure shows that as would be expected for random renaturation equal amounts of the two originar native DNAs are formed and twice as much of the hybrid of inter- mediate density. (Reproduced by permission from C. L. Schildkraut J. Marmur and P. Doty J. Mol.Biol. 1961,3,595.) late matter removed by centrifugation the clear supernatant fraction con- tains all the enzymes required for the synthesis of DNA. Using this frac- tion with an added energy source one can demonstrate the successive phosphorylation of the four deoxymcleosides to the mono- di and tri- phosphates. The reactions normally stop when the triphosphates have been produced. If however DNA is added polymerisation of the tri- phosphates occurs with the formation of a polymer possessing all the properties associated with native twin-stranded DNA.S3 As a result of the presence of nucleases in most of the enzyme preparations degradation 33 A. Kornberg “Enzymatic Synthesis York 1961. of DNA,” John Wiley and Sons Tnc. New 380 QUARTERLY REVIEWS proceeds simultaneously with synthesis and the product has a lower molecular weight and is more polydisperse than the original primer.The polymerase enzyme has an absolute requirement for the deoxynucleoside triphosphates and the priming polymer has itself to be a deoxyribose derivative. Specificity for the synthesis of DNA does not reside in the enzyme since preparations from any one source can use DNA from any other source or a synthetic deoxyribose polymer as primer. The composition and base sequences of the primer dictate the nature of the product; for example when the synthetic copolymer of deoxyadenine and thymidine is used as primer in the presence of all four triphosphates less than one guanine is incorporated for every 28,000 adenine and thymine residues. The fundamental correctness of the mechanism of replication proposed by Watson and Crick has been confirmed by analysis of the frequency of nearest neighbours in the polymer formed when a normal DNA is used as primer.In these experiments one of the four 5'-triphosphates in the reac- tion was labelled with 32P. During synthesis the phosphate group is at- tached to the 3'-OH of the sugar at the end of the growing chain. By degrading the product of the polymerisation with an enzyme that cleaves the bond joining the phosphate group to the 5'-0H of the sugar the 32P label is transferred from its original base to its neighbour (Fig. 6). Repeti- o / *?'OH c \ o / P ?-OH / '*y-OH on OH DEG R ADAT 10 N F FIG. 6. Synthesis of DNA occurs by addition of a nucleoside 5'-phosphate to the 3'-OH at the end of the growing DNA chain with the elimination of pyrophosphate from the nucleoside 5-triphosphate.Degradation by micrococcal or spleen DNAase cleaves the bond linking the 5'-OH to the phosphate group thus transferring the labelled phosphorus atom from its original nucleoside to its neighbour. EDWARDS AND SHOOTER DEOXYRIBONUCLEIC ACID 38 1 tion of the experiment using a different labelled base each time provides data for calculating the relative frequencies of the sixteen possible pairs of bases. As an example of the application of the method data for the product obtained using DNA from Mycobacterium phlei as primer is given in Table 2. The sums of the four columns show that the amounts TABLE 2 Reaction Labelled Isolated 3’-Deoxyribonucleotide No. triphosphate T A C G 1 dAT32P TA AA CA GA 0.012 0.024 0-063 0.065 2 dTT32P TT AT CT GT 0.026 0.031 0.045 0.060 3 dGT32P TG AG CG GG 0.063 0.045 0.139 0.090 4 dCT32P TC AC cc GC 0.061 0.064 0.090 0.122 Totals 0.162 0.164 0.337 0.337 Nearest-neighbour frequencies in the product formed using DNA from Mycobac- terium phlei as primer in a polymerisation reaction.The totals of the four columns give the molar proportions of the bases in the product. The base composition of the primer DNA obtained from chemical analysis is thymine 0.165 adenine 0.162 cytosine 0.335 guanine 0.338. (Reproduced by permission from A. Kornberg “Enzymatic Synthesis of DNA,” p. 22 John Wiley and Sons Inc. New York 1961.) of adenine and thymine incorporated are equal as are the amounts of guanine and cytidine and that the chemical composition of the product is close to that of the primer.Comparison of the frequencies of the sixteen different pairs establishes that in the synthetic polymer base-pairing occurs as required by the Watson and Crick model and proves that the two chains of the helix are antiparallel. The frequency of AA and TT pairs are equal; CC and GG pairs are also equal as would be expected for either parallel or antiparallel chains. Matching of the other base pairs depends upon the polarity of the strands (Fig. 7). For antiparallel chains a pair CA will be FIG. 7. Representation of the two chains of a native DNA molecule to show the relation of diflerent pairs of bases on the two strands. For parallel strands the pairs on each chain are read from left to right. For antbarallel chains pairs on the top strand are read from left to right and on the lower strand from right to left.matched on the other chain by TG CT with AG GA with TC and GT with AC. Inspection of the data shows that there is good agreement for the frequencies of the two pairs in each of these four sets. A pair AT on one chain will be matched by an AT pair on the second chain; the frequency of this matching cannot therefore be determined. The same argument applies 382 QUARTERLY REVIEWS to the pairs TA CG and GC. If the two chains of the helix were parallel equal frequencies would be expected for the pairs TA = AT CA = GT GA = CT TG = AC TC = AG and CG = GC. Of these six sets only two show nearly equal frequencies for the pairs. Observations on polymerase preparations from a variety of sources have shown that while native DNA can act as primer denatured DNA is more effective.A preparation of the enzyme from calf thymus is the only one so far made which is free of nuclease activity.34 With this preparation in the presence of denatured DNA as primer the reaction stops when an amount of DNA equivalent to the added primer has been produced (Fig. 8). Richardson Inman and K ~ r n b e r g ~ ~ have studied the action of * I 4 8 12 - - - - o w 4 8 12 w 16 4a s 36 3 K 0 Z 24 P u 3 2 Time (hr.) FIG. 8. The amounts of DNAfrom the bacteriophage +X 174 usedasprimer in thepresence of DNA polymerase from calfthymus and dATP are shown as dashed lines at the right of the diagram. The curves show that polymerisation approaches a limit when an equal amount of DNA has been synthesised. The incorporation of mole dAMPSa correspondr approximately to the synthesis of 3 x (Reproduced by permission from F.J. Bollum Progr. Nucleic Acid Research 1963 1 17.) purified polymerase from E. coli (still containing some nuclease activity) using as primer DNA from bacteriophage T7 which had been degraded with exonuclease 111 an enzyme which removes nucleotides sequentially mole of DNA phosphorus. 3p F. J. Bollum Progr. Nucleic Acid Res. 1963 1 1. a6 C. C. Richardson R. B. Inman and A. Kornberg J. Mol. Biol. 1964,9,46. EDWARDS AND SHOOmR DEOXYRIBONUCLEIC ACID 383 from the 3'-OH ends of the two chains thus leaving a central piece of twin helix with single chains at the ends. They observed initially a synthesis at five times the rate for the undegraded primer until an amount of DNA had been formed equivalent to that removed by the enzymic degradation.Following this the reaction rate fell to that observed when undegraded primer was used. Electron micrographs showed the molecules present at the end of the first stage of the reaction to be linear rods with no evidence of single chains at the ends The product of the second stage of the re- action consisted mainly of molecules showing branching.36 The difference between the two types of polymerisation reaction is further emphasised by the observation that the first stage of the reaction the repair of the degraded DNA occurs when the system is incubated at 20". Polymerisa- tion based on the native DNA as primer proceeds rapidly at 37" but cannot be detected at 20". It appears that in the absence of specific nucleases the polymerase cannot induce the separation of the strands of native DNA.Activation of native DNA by nuclease action enables it to act as primer but the product is not a succession of molecules identical with the initial primer molecules. In the in vitru system synthesis proceeds in only one direction along each chain successive nucleotides being added to the 3'-OH of the growing chains. In contrast to this replication of DNA within the cell proceeds along the twin helix in one direction only both chains being synthesised t~gether.~ The replication of DNA in E. coli has been studied by cairn^.^ Cells were grown in the presence of C3H]thymidine then carefully lysed and the DNA deposited on millipore filters. Radioautographs showed the DNA molecules as continuous circles. Replication begins at a specific point on the ring and proceeds continuously round the circle.Cells which had been labelled for more than one generation time showed the parent DNA molecule as a continuous ring together with a new partially dupli- cated section. The surprising observation was that both ends of the replicat- ing section were still attached to the parent ring. Since the two chains of the molecule have to be untwisted to allow the daughter molecules to separate there must clearly be some point in the DNA ring where free rotation is possible. This swivel point cannot be detected in electron micro- graphs of circular DNA. The implications inherent in these observations have been reviewed by Sibatani and HiaL3' They have considered the mechanisms involved in the separation of the daughter molecules dis- cussed the system of enzymes which appears to be required and suggested a model for the replication of circular DNA.C r ~ t h e r s ~ ~ has developed a theory for the relaxation kinetics of the helix- coil transition assuming that the rate-limiting factor in the separation of the two chains is the viscous resistance to the rotation of the molecule. The frictional resistance to rotation calculated is about a thousand times 36 R. B. Inman C. L. Schildkraut and A. Kornberg J. Mol. Biol. 1965 11 285. 38 D. M. Crothers J. Mol. Biol. 1964 9 712. A. Sibatani and S. Hiai J. Theoret. Biol. 1964 7 393. 384 QUARTERLY REVIEWS greater than would be expected for the helix rotating in water. This large resistance may result from the high viscosity of the ordered structure of the water bound to the helix. Applying the calculated value of the frictional resistance suggests that DNA synthesis can proceed at the rate of M = 7 x lo6 per minute.From the extent of labelling of the circular E. coli DNA following incubation with [3H]thymidine for various times Cairns calculates that DNA is replicated at a rate of M = 40-60 x lo6 per minute.' The polymerase isolated from E. coli has been found to possess the further property of acting as a polymerisation i n i t i a t ~ r . ~ ~ ~ ~ ~ In the presence of dATP and TTP (for abbreviations see footnote p. 369) and in the absence of any added primer there is a lag period of up to 2 hours then a stage of rapid synthesis of high-molecular weight material followed finally by a period in which the polymer formed is degraded by the nucleases present in the solution.The polymer formed in this reaction was found to contain equal amounts of adenine and thymine. Nearest-neighbour analyses were performed and showed that the frequency of the pairs AT and TA were both 0.5 while the pairs AA and TT could not be detected. The product is thus a copolymer of A and T in regular alternating sequence. Similar experiments using mixtures of dGTP and dCTP were found to give again after an initial lag period a polymer containing G and C but not necessarily in equal proportions the molar G content ranging from 50 to 81 %. Nearest-neighbour analysis showed G always associated with G and C with C. The polymer thus consists of a polydeoxyguanylate chain hydrogen bonded to a polydeoxycytidylate chain. Both of these synthetic polymers were found to be excellent primers polymerisation in the presence of the appropriate triphosphates beginning immediately with no observable lag period.Investigations have shown that during the lag period there is a slow synthesis of a few macromolecules poly-dAT or poly-dGdC (see footnote p. 369) and this is followed by an autocatalytic replication of these molecules the rate of the reaction increasing exponentially. The mechanisms by which the first macromolecules are formed is not yet understood. Measurement of the molecular weight of the polymers after about 5 % total reaction and at later stages has shown that the size of the macromolecules does not change during the course of the autocatalytic reaction. The smaller the amount of enzyme used in the reaction the higher is the molecular weight of the product.When poly-dAT is used as primer no synthesis of polymer occurs if dATP or dTTP are present alone. With poly-dGdC as primers on the other hand polymerisation occurs if either dGTP or dCTP are present. This mode of replication where only one strand of the double helix is used as primer is similar to the synthesis of messenger RNA on a native DNA template. Since DNA carries the genetic information required for general meta- bolism its immutability seems to be a prerequisite of a normally functioning EDWARDS AND SHOOTER DEOXYRIBONUCLEIC ACID 385 cell or organism. Mutations or carcinogenesis which can be effected by various forms of radiation or by treatment with a variety of chemical compounds are believed to be produced primarily by the action of these agents on DNA.Carcinogenic hydrocarbons for example can be inter- calated between the bases;39 X-irradiation results in attack by the radicals formed by the radiolysis of water mainly on the cytosine moieties,40 ultraviolet irradiation acts primarily by inducing dimer formation between neighbouring thymidine moieties on the same chain,41 and the various derivatives developed from mustard gas and sulphur mustard aklylate the 7-position of guanine.42 All these effects could lead to the insertion in the sequence of a wrong base at the next replication step and hence to a mis- coding for some essential protein. A surprising discovery has been that many cells and organisms contain a system for the deletion of damaged parts of DNA molecules and the subsequent repair of the chain. If the DNA of a bacterium is labelled with [3H]thymidine and then irradiated with ultraviolet light partial degradation of the DNA and elimination of the label can be demonstrated.Concomitant with the degradation there is a synthesis of DNA. [14C]Bromouracil (which can replace thymidine) present in the medium is incorporated into the DNA during the degrada- tion phase. Studies of the DNA formed have shown that the bromouracil is incorporated in short sections distributed along the length of the molecule as would be required by a repair mechanism acting on damaged units randomly distributed throughout the DNA.43 More recent work has suggested that such repair mechanisms may exist in mammalian cells possibly even in brain heart and muscle cells which do not divide in the adult How DNA codes for its own repair is an interesting problem. 39 L. S. Lerman J. Mol. Biol. 1964 10 367. 40 G . Scholes Progr. Biophysics Biophys. Chem. 1963 13 59. 41 A. Wacker Progr. Nucleic Acid Research 1963 1 369. 42 P. D. Lawiey “Acidi nucleici e loro funzione biologia,” Istituto Lombardo 1964 25; P. D. Lawley and P. Brookes Nature 1961 192 1081. 43 D. Pettijohn and P. Hanawalt J . Mol. Biol. 1964 9 395. 44 S. R. Pelc J. Cell. Biol. 1964 22 21.

ISSN:0009-2681    

DOI:10.1039/QR9651900369

出版商:RSC    

年代:1965

数据来源: RSC

摘要:

METAL COMPLEXES OF LIGANDS CONTAINING SULPHUR SELENIUM OR TELLURIUM AS DONOR ATOMS By STANLEY E. LIVINGSTONE* (WILLIAM RAMSAY AND RALPH FORSTER LABORATORIES UNIVERSITY COLLEGE LONDON) 1. Introduction BY far the most studied ligands are the halide ions cyanide ion and ligands having oxygen or nitrogen as the donor atom. Ligands having sulphur selenium or tellurium as donor atoms have been less extensively studied although in recent years there has been a considerable interest shown in sulphur-ligands. There appears to have been no recent review devoted entirely to metal complexes of ligands with sulphur selenium and tel- lurium as donor atoms but certain aspects of sulphur-containing ligands have been discussed in recent books.1,2 This Review outlines the types of metal complexes formed by various ligands containing sulphur selenium and tellurium and where possible compares the properties of these complexes with those of complexes of analogous oxygen-containing ligands.Greater attention has been given to the more recent work especially that dealing with structure and the nature of the bonding. 2. General Considerations The electronegativities of atoms which can act as donors fall in the series:F>O>N>Cl>Br>I-S-Se-C>Te>P > As > Sb. However the effective electronegativity will be influenced by the other atoms or groups attached to the donor atom. From a consideration of an electrostatic model it can be said that for a unidentate ligand the co- ordinating ability will depend not only on the electronegativity but on the total dipole moment (p) of the ligand where P -3.the permanent dipole moment p l = the induced dipole moment a = the polarisability and E = the inducing electrostatic field. Thus although the permanent dipole moment of NH3 is less than that of H20 the total dipole moment of NH may be greater in the presence of a cation with a high polarising power. The larger size and smaller permanent dipole moment of H2S (pHBS - 1.1 D; pHZO 1.9 D) reduce its co-ordinating ability below that of water for ions of low field-strength. However H2S is more polarisable than water (molar refractivity of donor atoms HzO 3.7 c.c.; H2S 9.5 c.c.) and with ions of high field-strength (e.g. Hg2+ * Present address School of Chemistry University of New South Wales Kensington New South Wales Australia. l C. K. Jrargensen “Inorganic Complexes,” Academic Press New York 1963 p.131. C. M. Harris and S. E. Livingstone “Bidentate Chelates” in “Chelating Agents and Metal Chelates,” ed. F. P. Dwyer and D. P. Mellor Academic Press New York 1964 386 p = P + pl = P + ccE . . . . . . . (1) p. 95. LIVINGSTONE METAL COMPLEXES OF LIGANDS 387 Agt) H2S co-ordinates strongly and protons are forced off to give in- soluble sulphides. An analogous situation occurs with water Fe3+ forms [Fe(H,0),I3+ which has a strong tendency to lose a proton [Fe(H20)J3+ $ [Fe(H20) OHI2+ + H+ Such is the tendency for [Pt(en),(H,O),]*+ to lose protons to give the dihydroxo-complex [Pt(en)2(OH)2]2+ that it is a strong acid. Whereas both the permanent dipole moment and co-ordinating ability decrease in the series H20 > ROH > R20 the reverse order holds for sulphur both dipole morr-? and co-ordinating ability increasing in the order H2S < RSH < ;.Moreover although the polarisability is decreased by alkyl substitution the decrease is much less (5%) in going from H2S to R2S than the decrease (24%) in going from H20 to R20.3 If we consider the formation of a complex between a metal ion M+ and a negatively charged ligand L- on an electrostatic model the bond strength ( - A H ) will be given by where Ze = effective nuclear charge on M e = electronic charge and r = ionic radius. Accordingly we should expect that for any metal ion RO- would co-ordinate more strongly than RS- because of the greater size of the sulphur atom; hence dH(R0-) > AH(RS-). If on the other hand we consider the M-L bond to be covalent then where ZM = ionisation potential of M EL = electron affinity of L and f(xM x xL) = a function of the electronegativities (x) of M and L.Here again since oxygen is considerably more electronegative than sulphur we would expect RO- >> RS-. Now let us consider an uncharged ligand L; on an electrostatic model the induced dipole of L will be oriented towards M+ to form a complex M+ . . . . 6-L . . . . . . . . (4) ze x (P + p’) rI2 -AH = where rl = distance between centre of M+ and centre of dipole. The larger value of rl and the smaller value of P for sulphur than for oxygen probably outweigh the effect of the larger value of$ and we would expect R20 > R2S. On a covalent model if the lone pair is equally shared between Mf and L the complex will be represented by M f L+. The bond strength will be given by (5).R. W. Parry and R. N. Keller “Electrostatic Theory of Co-ordination Compounds” in “Chemistry of the Co-ordination Compounds,” ed. J. C. Bailar Reinhold New York 1956 p. 119, 388 QUARTERLY REVIEWS However the system M f L+ represents the extreme of covalency the positive charge on M will not be transferred completely to L and the terms IL and xL+ will not be as important as indicated by (5); the true picture will be somewhat intermediate between (3) and (5). Nevertheless for an uncharged ligand on both the electrostatic and covalent models we would expect R20 > R2S > R,Se > R2Te. However in the covalent models we have ignored any contribution made by n-bonding. Except in special instances4 (e.g. NO2- 2,2'-bipyridyl dimethylgly- oxime) oxygen and nitrogen have no orbitals available to accept electrons from suitably placed filled d orbitals on the metal atom.On the other hand sulphur and phosphorus have vacant d orbitals which can be used for d,,-d, bonding such as can occur with the later transition metals in their normal oxidation states and with the early transition metals in unusually low oxidation states. The extent to which such n-bonding occurs is difficult to assess but the available evidence suggests that in favourable circum- stances it does occur with ligands containing sulphur selenium and tel- lurium but to a lesser extent than with R,P R,As CN- and CO. Con- sequently if .rr-bonding can occur we might expect it to cause a reversal of the order to RS- > RO- and R2S > R,O. The conditions for n-bonding are most favourable with the later members of the second and third transition series [e.g.Pdf' Ptrr Hg"] and with the early transition metals in low oxidation states [e.g. Moo Wo Re']. The polarisabilities of sulphur ligands decrease in the order S2- > RS- > R2S. Therefore in any consideration of the bonding properties of sul- phur ligands the distinction between sulphide ion mercaptide ion and thioethers should be borne in mind. Not only the polarisabilities but the number of lone pairs also decrease. Williams5 has suggested that the princi- pal difference between thiols and thioethers as ligands is that the former are more highly polarisable but not as effective d, electron acceptors as the latter. In complexes containing charged sulphur Iigands such as RS- (EtO) PS2- and ROCS2- the sulphur atom is bicovalent and has a V-shaped configuration whereas in complexes of R2S it is tricovalent and trigonal pyramidal.When the thio-group acts as a bridge the sulphur atom is tricovalent and pyramidal. Coates6 observed that the strength of co-ordination of XMe (X = 0 S Se Te) to AlMe is in the order 0 > S > Se > Te but whereas dialkyl sulphides selenides and tellurides co-ordinate to Pdn Pt" and HgIr ethers do not. A more extensive survey of the relative affinities of ligand atoms for metal ions was made by Ahrland Chatt and Davies' R. S . Nyholm Rev. Pure and Appl. Chem. 1954 4 15; D. P. Craig A. Maccoll R. S. Nyholm L. E. Orgel and L. E. Sutton J. 1954 332. G. E. Coates J. 1951 2003. S. Ahrland J. Chatt and N. R. Davies Quart. Rev. 1958 12 265.ti R. J. P. Williams Ann. Reports Progr. Chem. 1959 56 87. LIVINGSTONE METAL COMPLEXES OF LIGANDS 389 who divided metals into two classes (a) those which form the most stable complexes with the first ligand atom of each group; (b) those which form the most stable complexes with the second or subsequent ligand atom. For class (a) metals 0 > S > Se > Te; for class (b) metals S & 0 but almost any sequence of S Se and Te may occur. Class (b) metals form a triangular area with a somewhat diffuse border in the centre of the Periodic Table. Metals of pronounced (b) character are in the centre of this area. Co- ordination of C2Hr CO and CNR occurs only with metals of pronounced (b) character. The oxidation state of the metal affects the degree of (b) character which is strongest for transition metals in low oxidation states i.e.metals having non-bonding d electrons and thus capable of forming d,-p and d,-d, bonds by donating a pair of electrons to the ligand. Using a simple model and taking into account the ionisation potential of the metal and the electrostatic energy involved when the ions Mn+ and X- are brought together from infinity Craig and Nyholms have given a semi-quantitative explanation of class (a) and (b) behaviour to- ward halide ions. Pearsong has classified metal ions and ligands into “hard” and “soft” Lewis acids and bases and he has suggested a general rule that hard acids bind strongly to hard bases and soft acids to soft bases. Hard acids are those that bind to bases which bind strongly to the proton i.e. basic in the usual sense while soft acids bind strongly to highly polarisable or un- saturated bases which often have negligible proton basicity e.g.R2S. Yet it is possible for a base to be soft and strongly binding to the proton- such a case is the highly polarisable S2- ion. Pearson’s hard and soft acids correspond roughly to (a) and (b) metals respectively. JlargensenlO has pointed out that class (b) contains three rather disparate categories (i) metals with unusually low oxidation numbers; (ii) metals with certain high oxidation numbers ; (iii) the s2 family Sn” Sbl” TP Pb” Bill1 showing (b) character to heavy halides and chalcogenides but (a) aversion to CN-. J~zrrgensen also suggested that some metals may have (6) character in a low and a high oxidation state yet (a) character in an inter- mediate oxidation state.The spectrochemical series of ligands is arranged according to the spectroscopic splitting parameter d as given by the frequency of the lowest ligand-field absorption band in a transition metal complex. The series is:ll I- < Br- < SCN- - C1- < (EtO),PSe2- - S2- - (EtO),PS2- < F- < Et2NCS2- < urea - OH- < C2042- - 02- - H20 < NCS- < glycine < py - NH < en - SO,2- < NO2- - phen - bipy < H- < CH,- < CN- < CO (py = pyridine; en = ethylendiamine; bipy = 2,2’-bipyridyl; D. P. Craig and R. S. Nyholm “The Nature of the Metal-Ligand Bond” in “Chelat- ing Agents and Metal Chelates,” ed. F. P. Dwyer and D. P. Mellor Academic Press New York 1964 p. 51. R. G. Pearson J. Amer. Chem. SOC. 1963 85 3533. lo C . K. Jsrgensen Znorg.Chem. 1964 3 1201. l1 C. K. Jsrgensen Ricerca xi. 1964 34 3. 390 QUARTERLY REVIEWS phen = 1,lO-phenanthroline). The position of sulphur ligands is not clear as relatively few complexes of sulphur ligands have been studied spectro- photometrically. However although some sulphur ligands are near C1- sulphur appears to have a wide range as SO,,- when bound through sulphur has a late position in the series near NO2-. R2S probably occurs between H20 and NCS-;12 the position of RS- has not been established. The series roughly follows the order of decreasing radii of the ligand atom i.e. I > Br > C1 > S > F > 0 > N > C. The parameter d is dependent on (i) electrostatic attraction; (ii) the effect of the lone pairs on the ligand; (iii) M+L d,-p, or d,,-d, bonding; (iv) L+M p,,-d, bonding; (i) (ii) and (iii) increase A whereas (iv) decreases it.’ The nephelauxetic series14J5 is a measure of the “covalent” tendency of the ligands.The repulsions between the d electrons are decreased by complex formation by from 5 to 40 % for 3d cornple~es.~~ /3 (= B’/B) is the ratio of the electronic repulsion parameters for the complex (B’) and for the free gaseous ion (B). The nephelauxetic series of ligands is arranged according to increasing values of (1 - /3) F- < H20 < urea < NH < en - C20d2- < NCS- < C1- - CN- < Br- < S2- - (EtO) PS2- - I- < (EtO) PSe2-. This order roughly follows the order of decreasing electro- negativities F > 0 > N > C1 > Br > S - I > Se.11,15 Because of insufficient data it is far from certain that all sulphur ligands have a high nephelauxetic effect.Optical electronegativity (xOp4) is derived from the observation that the electron transfer bands of a chromophore MX shift as X varies1 The optical electronegativity of a ligand is closely related to the electro- negativity of the donor atom and the values of xopt for sulphur ligands are those expected on the basis of the electronegativity of the sulphur atom,17 whereas the situation with oxygen-ligands is much more complicated.l8 The free-energy change AG is related to the corresponding equilibrium constant AG = AH - TAS = -RTlnK Thus the relative stability of a metal complex is related to the ligational free-energy change AG. For many reactions both heat and entropy changes favour complex formation but AS can vary widely being either positive or negative.Ligational enthalpy in aqueous solution is the small heat change which accompanies the replacement of water by other ligands and is the resultant of several often large heat terms related not only to the la C. K. Jerrgensen J. Inorg. Nuclear Chern. 1962 24 1571. lS T. M. Dunn “Visible and Ultraviolet Spectra of Complex Compounds,” in “Modern Co-ordination Chemistry,” ed. J. Lewis and R. G. Wilkins Interscience New York 1960 p. 229. l4 C. K. Jerrgensen and C. E. Schaffer J. Znorg. Nuclear Chem. 1958,8,143. l6 C. K. Jsrgensen “Absorption Spectra and Chemical Bonding in Complexes,” Pergamon Oxford 1962 p. 134. l6 C. K. Jerrgensen Adv. Chem. Phys. 1963,5,33 ; ref. 1 p. 5. l7 C. K. Jerrgensen Acta Chem. Scand. 1962 16 2017. IR A. Carrington and C.K. Jsrgensen. Mol. Phys. 1961.4 395. LIVINGSTONE METAL COMPLEXES OF LIGANDS 39 1 strength of the metal-ligand bond but also to the solvation energies of the ions. The stepwise ligational enthalpies AH for ionic ligands in aqueous solution are usually between -5 and +5 kcal.mole-l but may be as large as -10 kcal. for Ag1 and HgrI with CN- or I-. The values for neutral unidentate ligands are usually between 0 and -5 k~al.mole-~.~~ For similar ligands the ligational entropy changes (ASn) for the formation of metal complexes in the series Mn*I to Cut1 are approximately equal so that the free-energy sequence as determined by measurements of stability constants is the enthalpy sequence.lg For halide complexes when the entropy term predominates the free-energy sequence is F > C1 > Br > I but the sequence is reversed when the enthalpy term is more important.20 The latter occurs when the bonding is essentially covalent rather than electro- static i.e.with class (b) metals. Although the entropy changes for the formation of [Hg(SCN),I2- and [Hg(SeCN)p12- are roughly equal the enthalpy change is considerably more exothermic for the latter.21 Low-spin ds ions PdII PtII and Auur and d10 ions Cur AgI Aul and HgII have the highest formation constants with the heavy halides and with sulphur ligands. Being typically (b) they form strong (T bonds with soft ligands and also d,-d bonds by donation of a pair of electrons to the ligand. In general donor atoms which give large ligand-field splittings form their strongest complexes with metals which are particularly sensitive to ligand-field stabilisation (class b) whereas donor atoms which produce small ligand-fields tend to form more stable complexes with cations which are insensitive to ligand-field stabilisation (class a).22 Stability-constant measurements on complexes of some S-0- chelate ligands that for these ligands the “natural order” for log K Co < Ni < Cu > Zn is maintained but Cd and Pb fall out of line in the Mellor-Maley (Mg < Mn < Fe < Cd < Zn < Co < Pb < Ni < Cu < Pd)whichholds for oxygen- and nitrogen-ligands.The “natural order” for the first-row transition metals is a direct consequence of their ionisation potentials and ionic radii,24b with the added factor of Jahn-Teller stabilisation in the case of Crn and Cun; on the other hand Cd and Pb not being in the first-row transition series behave differently towards sulphur.For these two metals the possibility of some degree of L -f M r-bonding in the complexes studied by Irving and Ferneliu~~~ cannot be excluded. lS F. J. C. Rossotti “The Thermodynamics of Metal Ion Complex Formation in Solution,” in “Modern Co-ordination Chemistry,” ed. J. Lewis and R. J. Wilkins Interscience New York 1960 p. 1. A. A. Grinberg and K. B. Yatsimirskii Bull. Acad. Sci. U.R.S.S. Div. Chem. Sci. 1952 239; R. J. P. Williams J. Phys. Chem. 1954 58 121 ; E. L. King J. Chem. Educ. 1953 30 71; A. J. Poe and M. S. Vaidya Nature 1959 184 1139. 21 V. T. Toropova Zhur. neorg. Khim. 1956 1 243. 22 L. E. Orgel 10e Conseil de 1’Institut international de Chemie Solvay Stoops Brussels 1956 p. 289.23 R. J. Irving and W. C. Fernelius J. Phys. Chem. 1956 60 1427. 24 (a) D. P. Mellor and L. Maley Nature 1948 159 370; (b) H. Irving and R. J. P. Williams J. 1953 3192. 392 QUARTERLY REVIEWS The stability constants of complexes of a considerable number of sulphur-ligands have been measuredz5 but unfortunately in most instances the data are not available for the same metals with the oxygen analogues. It is hoped that in the future measurements will be made so that for borderline metals a comparison can be drawn between the stabilities of complexes of analogous oxygen and sulphur ligands. Thiols but not thioethers cause spin-pairing in complexes of Con and Ni11.2s The ability to cause spin-pairing in complexes of Corl and Nit1 is confined to “soft” bases which are relatively strong n-acceptors.However the polarisability of the ligand atom is the more decisive factor. The trans-effect which is dependent on the permanent dipole moment the induced dipole moment the polarisability the charge and the size of the ligand and on the degree of 7-r-bonding is very strong in sulphur- ligand~.~,~’ Although a relative order of the strength of the trans-effect among a number of ligands has been suggested,28 many of the data are based on relative yields obtained under different sets of conditions. Heiice the positions of sulphur-ligands compared with other ligands having a strong trans-directing influence may vary considerably according to the particular system being studied. 3. Complexes of Sulphide and Selenide Ion The simplest sulphur ligand is the sulphide ion.Many metal sulphides are extremely insoluble in water Selenides are similar but are soon oxidised by air with the separation of selenium while tellurides are usually even less stable to air. Quite a few metal sulphides are soluble in solutions of ammonium or potassium sulphide to give soluble complex sulphides and use was made of this coinplexing power of the sulphide ion in qualita- tive analysis. HgIr GerV SniV Ast1’ AsV SbV MoVI Wvr and Revrr form soluble complex sulphides. The thio-salts Na,SnS,,aq. and Na,SnS,,aq. are readily hydrolysed and are decomposed by dilute acid.29 Stannic selenide SnSe dissolves in a solution of potassium selenide from which crystals of K2SnSe,,3H,O can be Potassium thioarsenite K,AsS and thioantimonite K3SbS and sodium thioantimonate Na,SbS4,9H20 can be isolated.The esters As(SPh) and Sb(SPh)3 are known; they can be oxidised by sulphur to S.As(SPh) and S.Sb(SPh)3.31 Thiostannate esters Sn(SR) have been 26 L. G. Sillkn and A. E. Martell “Stability Constants of Metal-ion Complexes,” Chem. SOC. Special Publ. No. 17 1964. 26 (a) S. E. Livingstone J. 1956 1042; (b) R. S. Nyholm Chem. Rev. 1953 53 263. 87 (a) F. Bas010 and R. G. Pearson “Mechanism of Inorganic Reactions,” Wiley New York 1958 p. 172; (b) J. V. Quagliano and L. Schubert Chem. Rev. 1952 50 201. O8 J. Chatt and A. A. Williams J. 1951,3061 ; J. Chatt L. A. Duncanson and L. M. Venanzi J. 1955 4456. 28 G. Spacu and A. Pop Bull. Acad. Sci. Roumanie 1939 21 52. 8o A. Ditte Compt. rend. 1882 95 641. R. Klement and R. Reuber Ber. 1935 68 1761; H. J. Backer and J. Kramer Rec.Trav. chim. 1933,52,916. LJVINGSTONE METAL COMPLEXES OF LJGANDS 393 prepared from SnC1 and NaSR; they are volatile at low pressures.31 Molybdenum(v1) gives a series of thiomolybdates K2 [MoO,S] K2[Mo02S2] K2[MoOS3] and K,[MoS,]. A similar series is known with tungsten(vi) and the selenium analogues (NH4)2[W02Se,] and (NHh [WSe,] have been ~btained.~ Cul and AuI give polysulphide complexes [CuS,]- and [AuS,]- con- taining s-S links. Other polysulphides of Cu* have been The spectra of solutions of complex sulphides have been but little investigated although the spectra of solid M"S4 chromophores (M = Fe Co Ni Cu) indicate that d is larger than for M1lCl and that the nephe- lauxetic effect is also greater.l 4. Complexes of Negatively Charged Unidentate Ligands 4.1. Thiols-Mercaptide ions RS- being highly polarisable form strong bonds with typical class (b) metal ions.The avidity of thiols for Hgxl has long been known hence the name "mercaptan". The mercaptides Hg(SR)2 (where R is an alkyl radical from ethyl to n-octyl) are monomeric in benzene.34 Mercury compounds have been used for analytical reagents for thiols and disulpliides. The desirable properties for such a reagent are rather restrictive and only HgC1 and MeHgI especially the latter ade- quately fulfil all the requirement^.^^ With Ni" ethyl mercaptan forms the diamagnetic complex Ni(SEt) for which the polymeric structure (I) was ~uggested.~~ The Pd" mercaptides Pd(SR) (R = Et Prn Bun Amn) are associated in ethylene dibromide and Et Et chloroform and probably have a similar polymeric structure.34 Thio- phenol gives an intense colour with PdI1 [Pd(SPh) is bright vermilion] yet no similar colour is developed with compounds of Ag Au Rh Ir or Pt.34 The square-planar monomeric complexes [Ni(SPh) ,L2] and [Pd (SPh),L2] (L = Et3P PhEt,P; 2L = Ph,PCH,CH2.PPh,) have been recently de~cribed.~' Whereas halogen-bridged dimeric complexes of PdII and Ptl* are readily split by p-toluidine and other unidentate ligands the corresponding alkyl- thio-bridged complexes are cis and trans Isomers of the Pt" com- 32 J.W. Retgers 2. phys. Chem. 1892 10 548; V. Lenher and A. G. Fruehan J. Amer. Chem. SOC. 1927 49 3076. 33 G. Peyronel D. de Filippo and G. Marotrigiano Gazzetta 1961 91 1190 1196. 34 F. G. Mann and D. Purdie J. 1935 1549. 36 S. J. Leach Austral. J. Chem. 1960 13 520. 36 K. A. Jensen Z.anorg. Chem. 1944,252 227. 38 J. Chatt and F. G. Mann J. 1938 1949; J. Chatt F. G. Mann and A. F. Wells R. G. Hayter and F. S. Humiec J. Znorg. Nuclear Chem. 1964 26 807. J. 1938,2086; J. Chatt,J. 1950,2301. 3 394 QUARTERLY REVIEWS plexes [Pt(PPr3l1)(EtS)X] (X = C1 SEt) (11) and (111) have been prepared; the cis-Pdn analogue (X = Cl) was isolated but not the Another type of isomerism involving the bridging thiol has been reported to occur X Et Et Pr” P S Pr” P s kpt/ Lpt pprn3 3 k p t / kR/ X ’ Kc/ ‘x (Ill X ’ ‘s’ ‘PPr; (m) Et Et in alkylthio-bridged Ptn complexes containing two different alkylmer- captans;40 the isomers are shown (IV and V). Hexanethiol forms a Co1I1 complex Co(C6H,,S) for which the polymeric structure (VI) has been suggested as a po~sibility.~~ The reaction of cobalamin with thiols has been t t t The phenylthio-Snw complex Ph,Sn(SPh) was prepared under anhy- drous conditions ;43 however the complexes R,SnSR‘ R,Sn(SR’), and RSn(SR’) (R = Me Et; R’ = alkyl or Ph) can be obtained from aqueous solution.The Sn-S bond in Me,SnSR’ is cleaved by PCl to yield (R’S) P and Me,SnCl and by Br to yield R’S.SR‘ and Me,SnBr.u 4.2. Sulphite Ion.-The sulphite ion is said to form complexes with Be Mn Fe Ru Os Co Rh Ir Ni Pd Pt Cu Ag Au Zn Cd and Hg,45 although some are probably double salts rather than true complexes (see below). Most of the metals listed are class (b) metals and with these strong complexes are formed. 8D J. Chatt and F. A. Hart J. 1953 2363. 40 J. Chatt and F. A. Hart J. 1960 2807. 41 B. J. McCormick and G. Gorin hwrg. Chem. 1963 2 928.4a D. H. Dolphiii and A. W. Johnson Proc. Chem. Soc. 1963 3 11. 43 H. J. Backer and J. Kramer Rec. Trav. chim. 1934 53 1101. 44 E. W. Abel and D. B. Brady J. 1965 1192. 4s A. Werner “Neure Anschauunger auf dem Gebiete der anorganischen Chemie,” 5th edn. Vieweg Braunschweig 1923 p. 120. LIVINGSTONE METAL COMPLEXES OF LIGANDS 395 In many complexes the sulphite ion is unidentate but in others it occupies two co-ordination positions. There appears to be no easy way of converting unidentate into bidzntate cornplexe~.~~ Aquo-PdII complexes such as [Pd(SO,)(H,O),] cannot be dehydrated to give bidentate com- plexes,,' as has been done with sulphato-c~mplexes.~~ The values of the stability constants for HgIr (1ogl0/3, 22.9) and T P (log10/34 34) are remarkably high.25 The exceptional stability of the Hgn complex (mercury usually forms weak links with oxygen) and the fact that strong complexes are formed by I P Pd" and PtrI suggests that the ligand is S-bonded as in (VII) when unidentate.The SO3,- ion has CSv symmetry which would be effectively unchanged if the ligand is S-bonded whereas 0-bonding (VIII) would lower the symmetry to C,. If S-bonding occurs the S - 0 stretching frequency would be expected to be higher than in the free SO3,- ion but if 0 bonding is present both higher and lower frequen- cies would be expected because of the lowering of the symmetry from C, to C,. The infrared spectra of a number of unidentate sulphito- complexes of Co"' RhlI1 IP PdIT and Ptn are similar to the spectrum of (NH,),[Hg(SO,),] and are in accordance with S-bonding of the ligand.46~49~50 However T1 [Cu(SO,),] appears to be 0-bonded while the compounds (NH,),MII(SO,) (MI1 = Mg Mn Fe Co Ni Zn and Cd) and NH,CuSO are double salts since their spectra are similar to that of Na,S0,.50 Several sulphito-complexes of 0s" and 0s"' are known.51 As osmium has a low affinity for sulphur-ligands the SO,,- group may be 0-bonded in these complexes.When occupying two co-ordinetion positions the sulphito-group may be bidentate as in (IX) or (X) or it may be bridging as in (XI) or (XII). The sulphito-group in (IX) and (XI) has CSv symmetry but it has been 46 M. E. Baldwin J. 1961 3123. 47 G. A. Earwicker J.. 1960 2620. 48 C. G. Barraclough and M. L. Tobe J. 1961 1993. 4g A. V. Babaeva Yu. Ya. Kharitonov and Z. M. Novozhenyuk Zhur. neorg. Khim. 1961 6,2263 2281 ; A.V. Babaeva Yu. Ya. Kharitonov and I. B. Baranovskii ibid. 1962 7 1247; A. V. Babaeva Yu. Ya. Kharitonov and E. V. Shenderetskaya ibid. p. 1530. 61 N. V. Sidgwick "The Chemical Elements and Their Compounds," Clarendon Press Oxford 1950 p. 1496. G. Newman and D. B. Powell Spectrochim. Acta 1963 19,213. 396 QUARTERLY REVIEWS suggested that it is doubtful if infrared data can distinguish between (IX) (XI) and (XII).50 As (X) involves a 3-membered ring which might be easily ruptured and as [Co(en),SO,]+ is stable in water it has been in- ferred from this and the infrared spectrum that the sulphito-group is bound as in (TX),4s The infrared spectra of [CO(SO~),]~- and [Rh(S0,),I3- have been interpreted as indicating that the sulphito-groups are bridging as in (XI) or (XII).52 It is unlikely that the structures (X) and (XII) occur.The sulphito-group has a marked trans-effect which is greater than that of the nitro-group in Ptll Con' and Ir"1 complexes.53 Sulphitocobalamin has been recently in this compound S032- has replaced the CN- group attached to the cobalt atom in vitamin B,2.54 The fact that amine-sulphito-complexes of CoIll are yellow and those of RhllI IrIII Pd" and PtI1 are colourless indicates that has a high position in the spectrochemical series comparable with that of NH,.l 4.3. Thiosulphate Ion.-Alkaline thiosulphate will dissolve many in- soluble salts of Pb" Hgrr Cul and Agl. Strong complexes are formed with CuI AgI and Aul; bivalent copper is reduced by S2032- to the CUI complex. As with sulphite ion the stability constants for Hg" (10g10f12 29.9) and T P (log10/34 41) are remarkably high.25 The thiosulphate ion is usually unidentate as it undoubtedly is in [Hg(S2O3),I2- but it may well be bidentate in [Bi(S203)3].3- When uni- dentate it is almost certainly S-bonded through the central sulphur atom as is sulphite ion.The reactivity of co-ordinated S20,2- in Cot" complexes is lower than that of NO2- and it is not displaced by H 2 0 in Bivalent platinum forms the very stable complexes [Pt(S,O3)Cl2l2- and [Pt(S20,),I2- in which the ligand is almost certainly chelated; the com- plexes [Pt(S203)3]4- and [Pt(S203)4]6- are also In the chelate complexes the ligand is bound through one sulphur and one oxygen atom as cis and trans isomers of [Pt(S20,),I2- have been prepared.57 However only one isomer is known with PdlI.The group trans to S2032- is labilised in the same way as in thiourea complexes of Ptu;25b this is a further indica- tion that the ligand is bound through sulphur. 4.4. Thiocyanate and Selenocyanate Ion.-The standard free energies of formation of XCN- (X = 0 S Se) increase linearly with the ionic radius of X.58 62 A. V. Babaeva and Yu. Ya. Kharitonov Doklady Akad. Nauk S.S.S.R. 1962,144 559. s3 A. V. Ablov and G. P. Syrtsova Zhur. neorg. Klzim. 1960,5,1221; A. V. Babaeva and I. B. Baranovskii ibid. 1962 7 783; I. I. Charnyaev and Z . M. Novozhenyuk ibid. 1961 6 2462. 64 D. H. Dolphin W. A. Johnson and N. Shaw Nature 1963 199 170. 66 A. V. Babaeva I. B. Baranovskii and Yu. Ya. Kharatonov Zhur. neorg. Khim. 1963 8 604. s6 P. Shottlander Annalen 1866 140,200; D. I. Ryabchikov Compt.rend. Acad. Sci. 57 D. I. Ryabchikov Compt. rend. Acad. Sci. U.R.S.S. 1943 41 208. 68 S . Hamada Nippon Kagaku Zasshi 1961 82 1327. u.R.s.s. 1940,27,349; 28,231; 1 9 4 4 ~ 277. LIVINGSTONE METAL COMPLEXES OF LIGANDS 397 The only transition-metal cyanato-complexes which had been reported prior to 1964 were K,[CO(NCO)~],~~ K[M(NCO),] (M = Cu Cd) and [M{ (CH2)6N4}2(H20)4] [M(NCO),] (M = Co Ni CU).~O Recently the compounds [NEt,],[M(NCO),] (M = Mn Fe Co Ni Cu Zn and Cd) and [NEt,] [Fe"](NCO),] were reported; in these tetrahedral complexes the cyanato-group is considered to be bonded to the metal through the nitrogen atom thus strictly speaking they are isocyanato-complexes.61 On the other hand the thiocyanate ion forms complexes with most transition metals although those of thorium and niobium have only recently been r e p ~ r t e d .~ ~ ? ~ ~ Some of these complexes such as the high-spin iron(II1) complex [NEt4I3 [Fe(NCS)6]64 and the molybdenum(v) complex [pyHI4 [MO,O~(NCS),],~~ are deeply coloured. Except in ionic com- pounds such as KSCN the thiocyanate group can be co-ordinated in at least four ways as shown in (XII1)-(XVI). Y 7 w 0 M-S-CZN M-N=C=S M-S-CEN-M M-S-cM (XI 11) (XI v) @V> The structures represented by (XIII) (XIV) and (XV) only have been established by crystal structure determinations although in AgSCN,PPr,n the SCN group links three silver atoms two through sulphur and one through nitrogen.66 The deep blue insoluble CoHg(SCN) has a polymeric structure in which each cobalt atom is tetrahedrally surrounded by four nitrogen atoms of the thiocyanato-group while the sulphur atoms are bound to four different mercury atoms which are tetrahedrally surrounded by sulphur The electronic spectra of HgM(NCS) (M = Mn Fe) indicate an essenti- ally tetrahedral environment of the nitrogen atoms about Mn and Fe;'j8 the SCN- group is probably S-bonded to Hg.The SCN- group also acts as a bridge in the octahedrally co-ordinated Ni(NH3)2(NCS),,69 Co(py) (NCS),,70 CU(~~),(NCS),,~~ Cd(etu),(SCN) (etu = ethylenethi~urea),~~ in AgSCN,72 and in the dimeric AgSCN,PEt3;73 in Cu(NCS),(py) the 5 9 C. W. Rlomstrand J. prakt. Chem. 1871 (2) 3 221. 6 o R. Ripan Chem. Zentral. 1930 12 i 967. 61 D. Forster and D. M. L. Goodgame J. 1964,2790; 1965,262. 62 A. K. Molodkin and G. A. Skotnickova Zhur. neorg. Khim. 1964 9 1548. 64 S. E. Livingstone and T.N. Lockyer unpublished work. 65 R. G. James and W. Wardlaw J. 1928,2726; P. C. H. Mitchell and R. J. P. Williams 66 A. Turco C . Panattoni and E. Frasson Nature 1960 187 772. 67 J . W. Jeffrey Nature 1947 159 610. R8 D. Forster and D. M. L. Goodgame J. 1965 268. 6 9 M. A. Porai-Koshits E. K. Iukno A. S. Antsyshkina and E. Dikareva Soviet ' O M. A. Porai-Koshits and G . N. Tishchenko Soijiet Phys.-Cryst. 1960 4 216. 71 L. Cavalca M. Nardelli and G. Fava Acta Cryst. 1960 13 125. 72 I. Lindqvist Acta Cryst. 1957 l@ 29. 73 A. Turco C. Panattoni and E. Frasson Ricerca sci. 1960 30 1071. A. M. Golub and A. M. Sych Zhur. neorg. Khim. 1964 9 1085. f. 1962. 4570. Pfijx-CrjJst. 1957 2 366. 398 QUARTERLY REVIEWS Cu-S distances are longer than the Co-S distance in the Col* compound owing to Jahn-Teller di~tortion.~~ Bridging occurs in Pt,(PFrn,)2C1,(SCN)2 which has been isolated in two isomeric forms (XVII) and (XVIII).74 P‘:P, /NzC-S Pt PF3P\ ,S-C=h \4 Pt Pt CL’ \N=C-S / ‘PPr” Cl’ ‘S-CrN’ \PPrn3 (XVI I; (XVI I I) The CEN stretching frequency in a bridging thiocyanato-group is higher than in a terminal thio~yanato-group.~~ In general the thiocyanate ion co-ordinates to (a) class metals through nitrogen and to (b) class metals through sulphur but it is known from X-ray analysis that with thiocyanate ion the first-row transition metals Cr Mn Co Ni Cu and Zn form M-N bond^.^^,^^,^^ However the oxidation state of the metal the nature of the other ligands in the complex and steric factors may determine the way in which the thiocyanato-group is bound.In Cu(en),(SCN) the non-linear thiocyanato-groups are S-bonded in a tetragonally distorted octahedral config~ration.~~ The CEN stretching frequency has been used80 to distinguish between S- and N-bonding since among complexes of known structure those wlicli are S-bonded have generally high values for the C r N stretching frequency but some overlap occurs.It has been found that the C-S stretching frequency is more diagnostic as in S-bonded complexes it occurs at 690-720 cm.-l and in N-bonded complexes at 780-860 cm.-1.81,82 The M-SCN linkage is always angular but the M-NCS linkage can be collinear or angular with the M-N-C angle as low as 140°.81978 Various canonical structures can be written for the thiocyanato-group Free ion M-N bonding M-S bonding + N s C-S‘ M-NX-S- ,S-C=N ( X N (XX) M (xx 0 + S=C=N- - + -N=C=S M=N=C-S (XXI I) (XXI I I) M’ (XXIV) + - 2+ - M=S=C=N 2- [xxv) M/N=C=S (xxv I) (XXVI I) N-C=S 74 J.Chatt and F. A. Hart Nature 1952 169 673; J. Chatt L. A. Duncanson F. A. Hart and P. G. Owston ibid. 1958 181 43; P. G. Owston and J. M. Rowe Acia Cryst. 1960 13 253; J. Chatt and F. A. Hart J. 1961 1416. 76 J. Chatt and L. A. Duncanson Nature 1956 178 997. 76 G. S. Zhdanov and Z. V. Zvonkova Zhur. j i z . Khirn. U.S.S. R. 1950 24 1339. 77 A. Turco and C. Pecile Nature 1961 191 66 and references therein. 78 B. W. Brown and E. C. Lingafelter Acta Cryst. 1963 16 753. 7s B. W. Brown and E. C. Lingafelter kcta Cryst. 1964 17 254. g2 A. Tramer J. Chim.phys. 1962 59 232. B. C. H. Mitchell and R. J. P. Williams J. 1960 1912. J. Lewis R. S. Nyholm and P. W. Smith J. 1961 4590.LIVINGSTONE METAL COMPLEXES OF LIGANbS 399 It has been suggested that (XIX) predominates in KSCNS3 and (XXVI) or (XX) predominates in N-bonded complexes while for S-bonded com- plexes (XXI) is important.81 On the assumptions that (a) the lone pairs on the sulphur atom are more easily polarised and (b) the permanent lone- pair dipole on the nitrogen atom is larger than on the sulphur atom Lewis Nyholm and Smiths1 suggest that when thiocyanate ion is the only ligand the way in which the group is bound is decided by the relative bond energies of a covalent M-S bond and the more ionic M-N bond. Nuclear magnetic resonance (NMR) spectroscopy has been used to distinguish between N- and S-bonded thiocyanate. In known S-bonded complexes the 14N resonance is shifted slightly downfield compared with the free SCN- ion and in N-bonded complexes it is shifted upfield by a comparatively large amount.It has been suggested that N-bonding lowers the energy of the non-bonding orbital and shifts the 14N resonance up- field and that S-bonding while not affecting the non-bonding orbital lowers the energy of the delocalized anti-bonding orbital by mixing in its own stabilized d-orbitals thus causing a small low-field shift of the resonance.84 Infrared spectral data indicate S-bonding in [M(SCN),I2- and in [M(NH,),(SCN),] (M = Pd Pt) but N-bonding in [M(PR,),(SCN),]. The change from M-S to M-N bonding in these complexes of class (b) metals has been explained on the basis that strong .rr-electron acceptors such as tertiary phosphines can make the & orbitals of the metal less available for binding with the .rr-orbitals of the sulphur atom.77 The compounds cis-Mn(CO),(am) ,NCS (am = py p-toluidine ibipy) and trans-Mn(CO),( PEt,) ,NCS are N-bonded but cis-Mn(CO),( SbP h,) ,SCN is S-b~nded.~~ It appears that in the presence of strongly .rr-bonding ligands the effective charge on the metal plays an important r61e; thus complexes with bases weaker than PPh (e.g.CO SbPh,) are S-bonded whereas complexes containing stronger bases such as amines are N - b ~ n d e d . ~ ~ Most first-row transition metals are N-bonded but Co(SCN),(PPh,) and Cu(SCN),(PPh,) are S-bonded.86 Whereas Mn(CO),SCN is S-bonded the isoelectronic [Cr(CO),NCS]- is N-bonded ; the decrease in charge on the metal atom causes a change from S- to N-b~nding.~~ It has been showns8 that steric factors influence the manner of attach- ment of the thiocyanato-group in the complexes ML,(SCN) (M = Pd Pt; L = amine idiamine PR, SbR,) some of which are S-bonded and others are N-bonded.The steric requirements for M-SCN are greater than for the linear M-NCS. Since Pd(cc-picoline),(SCN) is S-bonded and L. H. Jones J. Chem. Phys. 1956 25 1069; 1958,28 1234. 84 0. W. Haworth R. E. Richards and L. M. Venanzi J. 1964 3335. 86 A. Wojcicki and M. F. Farona Inorg. Chem. 1964 3 151. F. A. Cotton D. M. L. Goodgame M. Goodgame and A. Sacco J. Amei. Chenr. SOC. 1961 83,4157; M. A. Poraj-Kosic Acta Cryst. Suppl. 1963 A 42 abstract 4.72. 87 A. Wojcicki and M. F. Farcna J . Inorg. Nuclear Chem. 1964 26 2289. 88 F. Bitsolo W. H. Baddeley and J. L. Burmeister Inorg. Chem. 1964 3 1202. 400 QUARTERLY REVIEWS Pd(py),(NCS) is N-bonded it seems that electronic effects are also important.88 Although selenocyanato-complexes of platinumsg and mercuryg0 have long been known very little has been reported on selenocyanato-complexes until the past few years.The manner of co-ordination of selenocyanate ion can be established by infrared spectroscopy the C-Se stretching frequency which occurs at 558 crn.-l in KSeCN,gl occurs in the range 520-543 cm.-l for Se-bonded complexes and at 639-672 cm.-l for N- bonded complexes.92 The complexes [Pt(SeCN),I2- and [Hg(SeCN),I2- are Se-bonded while [Co(NCSe),12- and [Cr(NCSe),I3- are N-bon- ded. 9 2 7 93994 X-Ray investigation has established the presence of Co-NCSe- Hg bridges in CoHg(SeCN) which is isomorphous with C O H ~ ( S C N ) ~ .~ ~ Selenocyanato-bridges are also present in CO(~~)~(NCS~),. 96 The tetra- hedral complexes Co(quinoline),(NCSe) and Co(Ph,PO),(NCSe) and the trans-octahedral complex Co(py),(NCSe) are N-bonded and isomor- plious with their thiocyanato-analogues. 94 The magnetic moment of Co(Ph,P),(NCSe) is 3.4 B.M.; it has been suggested that this compound contains equal numbers of octahedral low-spin cobalt@) and tetrahedral high-spin cobalt(r1) ions.94 An X-ray structural determination of NH [Co (SeCN),(DH),],3H20 (DH2 = dimethylglyoxime) shows that this com- plex has a trans configuration and that the SeCN- group is bound through Se.97 The selenocyanate coniplexes of AgI Cd” and HglI are more stable than the corresponding thiocyanates.21 As might be expected the position of the thiocyanate ion in the spectro- chemical series depends on whether it is S- or N-bonded.When S-bonded it occupies a position approximately equal to Cl- but when N-bonded it lies between HzO and NH3.98,99 The value of d (18,000 cm.-l) obtained for [Cr(NCSe),I3- shows that N-bonded selenocyanate creates a slightly stronger field than N-bonded thiocyanate (17,800 ~ r n . 3 . ~ ~ 5. Complexes of Neutral Unidentate Ligands 5.1. Thioethers Selenoethers and Tel1uroethers.-Thioethers do not co-ordinate very strongly to metals apart from Pt“ Pdl Rhxxr IrIXx and 8 9 A. Werner and A. Miolati 2. phys. Chem. 1894 14 507. g1 H. W. Morgan J. Inorg. Nuclear Chem. 1961 16 367. O2 C. Pecile A. Turco and G. Pizzolotto Ricerca xi. 1961 31 2A 247; A. TLU-CO 93 K. Michelson Actu Chem. Scand. 1963 17 1811. 94 F. A. Cotton D.M. L. Goodgame M. Goodgame and T. E. Haas Inorg. Chem. 95 E. Frasson A. Turco and C. Panattoni Gazzetta 1961,91 750. 96 S. M . Nelson Proc. Chem. Soc. 1961 372. 97 A. V. Ablov and V . N. Shafranskii Zhur. neorg. Khim. 1964 9 585. 98 C. E. Schaffer Internat. Conf. on Co-ordination Chem. London Chem. SOC. gg C. K. Jprrgensen “Absorption Spectra and Chemical Bonding in Complexes,” A. Rosenheim and M. Pritze 2 . unorg. Chem. 1909 63 275. C. Pecile and M. Niccolini J. 1962 3008. 1962 1 565. Special Publ. No. 13 1959 p. 153. Pergamon Oxford 1962 p. 109. LIVMGSTONE METAL COMPLEXES OF LIGANDS 401 Hg". Dimethyl sulphide forms three isomeric compounds of general formula PtC12(MezS)2.100 The y-isomer is [Pt(Me,S),] [PtCl,]. The corn- plexes Pt(R2S),C1 (R = Et Prn Bun) were among the first metal complexes to be investigated by dipole moment measurements ; the or-isomers have dipole moments of approximately 2.4 D while the @isomers have moments in the range 9.0-9.5 ~ .l O l These results show that the or-isomers are trans and the /%isomers are cis. The trans structure was confirmed by an X-ray investigation.lo2 Sulphur has a much lower affinity for Pt'V than for -Pt" and various attempts to prepare Ptw complexes of thioethers have been unsuccessful. The complexes of selenoethers resemble closely the analogous thioether complexes. The compounds Pt(R,Se),Cl (R = Me Et Prn Me2CH2CH2- CH, Ph) occur in a p and y-isomeric forms. The a- and p-isomers are monomeric and the values for their dipole moments are similar to those of their sulphur-anal~gues.~~~ The corresponding complexes of telluroethers are much less stable.The best characterised compound is Pt{ (PhCH,),Te ),Cl which probably has the trans configuration as it is soluble in chloroform. It is stable in the solid state but rapidly decomposes in solution to platinum tellurium and bibenzyl; no isomers are known.lo4 The PdlI complexes Pd(R2S),C12 can be readily prepared but in the a-form only;lo5 Pd(Me2S),C12 is completely isomorplious with trans- Pt(Me2S),C12.102 The Pd" complexes of RSPh (R = alkyl) have been used to identify the alkyl phenyl sulphides as unlike dialkyl sulphides these alkyl phenyl sulphides do not co-ordinate readily with mercury. With PdCl yellow monomeric complexes Pd(PhSR),Cl are usually formed but when R = But or Me,EtC the red complexes PhSR,2PdC12 are formed.loe These complexes probably have the tetrameric structure (XXVIII) ; if so they are the only known Pdll complexes of this type.Dialkyl and diary1 selenides form stable complexes Pd(R2Se),X2 (R = Me Et Bun Ph; X = C1 Br I).lo4 Chloro-bridged Pt" compounds of general formula L2Pt,C1 (L = R2S Ph loo L. A. Tschugaev and W. Subbotin Ber. 1910,43 1200. lol K. A. Jensen 2. anorg. Chem. 1935 225 97 11 5. lo2 E. G. Cox H. Saenger and W. Wardlaw J. 1934 182. lo3 E. Fritzmann 2. anorg. Chem. 1911 73 239; E. Fritzmann and V. V. Krinitzki lo4 E. Fritzmann J. Russ. Chem. SOC. 1915 47 588; 2. unorg. Chem. 1924 133 lo5 L. Tschugaev 2. unorg. Chem. 1924 134 277; L. Tschugaev and C. Ivanov lo6 V. N. Ipatiev and B. S. Friedman J. Amer. Chem. SOC. 1939 61 684. J. Appl. Chem. Rus. 1939,11,1610. 119 133.ibid. 1924 135 153; F. G. Mann and D. Purdie J. 1935 1549. 402 QUARTERLY REVIEWS R,Se R,Te; R = alkyl) have been prepared; the selenide and telluride complexes are brownish-orange and more soluble than the yellow sulphide complexes. The stability of these complexes and those of Group V ligands falls in the order R3P -J R2S > R,As > amine > RzTe > R3Sb > R2Se.lo7 The analogous Pdrr complexes L2Pd,C14 are more soluble and less stable than the Pt complexes but when L = R,Se they are more stable. The stability sequence of the Pd complexes is R3P > R,As - R,S > R,Se > R2Te > R3Sb. The reverse order of stability for the R,Te and R,Se complexes of Pd may be due to the relative sizes of the orbitals used for a-bonds Se being comparable with Pd and Te with Pt.lo8 A study of the N-H stretching frequencies of the series trans-[L,arnPtCl,] [am = RNH or R,NH; L = piperidine y-picoline R3P R,As R,Sb R2S R,Se R2Te P(OR), or C2H4] shows the increasing inductive effect transmitted across the Pt atom from the ligand donor atom to the N-H bond when the ligands are arranged in the series y-picoline < piperidine < R2S < R,Se < R,Te < R,As < R,P < R,Sb < P(OR) < C,H,.lo9 The ligand-field splittings inferred from the energies of the dzv+dxf-,,* transition in the complexes trans- [L,piperidinePtCI,] decrease when L is arranged in the sequence P(OMe) > Pr,P > piperidine > Pr,As > Et2S > Et,Se > Et,Te.The range of energies is small varying from 33,400 cm.-l when L = P(OMe) to 29,400 cm.-l when L = Et,Te.llO Rhodiurn(Ir1) gives complexes of the type Rh(SEt2),X (X = C1 Br I).111 The corresponding IrIr1 chloro-complex was reported in cis and trans isomeric forms.112 The yellow cis-isomer is soluble in polar solvents yet the “trans” isomer is soluble in chloroform and nitrobenzene but not in most non-polar solvents.A recent reinvestigation of these isomeric complexes by means of a number of physical techniques has shown that the yellow is indeed cis-[Ir(Et,S),Cl,] but the red isomer is in fact the salt trans- [Ir(Et,S),CI ,] trans- [Ir(Et,S),Cl,]. An IrIV complex [Ir(Et,S),CI,] was also reported.l13 The monomeric AuI complex Au(Et ,S)C1 is ~nstab1e.l~~ Thioethers form at least three types of Hgl* halogen complex (a) mono- meric (R2S),HgX2; (b) halogen-bridged dimeric [R,SHgX,] (XXIX) ; (c) complexes R,S,2HgX2 of uncertain structure.115 The co-ordinating ability of bischloroethyl sulphide (ClC2H4) ,S is weaker than that of diethyl sulphide; bis-complexes Pt( (CIC2H4),S ),X2 (X = C1,N02) are formed but this ligand unlike Et,S will not replace lo7 J.Chatt and L. M. Venanzi J. 1955 2787. lo* J. Chatt and L. M. Venanzi J. 1957 2351. lo* J. Chatt L. A Duncanson and L. M. Venanzi J. 1955,4461. 111 F. P. Dwyer and R. S. Nyholm J . Proc. Roy. SOC. New South Wales 1944,78,67. lla P. C. RGy N. Adhikari and R. Ghosh J. Indian Chem. SOC. 1933 10,275. llS G. B. Kauffman J. Hwa-San Tsai R. C. Fay and C. K. Jlargensen Inorg. Chern. 114 F. G. Mann A. F. Wells and D. Purdie J. 1937 1828. S. Smiles J. 1900,77 163; W. F. Faragher J. C. Morel] and S. Comay J. Amer. J. Chatt G .Gamlen and L. E. Orgel J 1959 1047. 1963. 2 1233. Chem. Soc. 1929,51,2774. LIVINGSTONE METAL COMPLEXES OF LIGANDS 403 (XXIX) (xxx) (xxx I) NH or RNH from Pt" complexes. As a result of the increased trans- effect compared with Et,S one ligand is extremely labile.lf6 Although TiIv is regarded as having class (a) behaviour TiCI and TiBr form ill-defined complexes with dialkyl ethers but the compounds TiX4(R2S)2 (R = Me,Et) are stable and crystalline. However thioether complexes of Ti111 are unstable and less well defined. Thioethers do not replace NMe completely from TiCI3(NMe,), although the amine is readily replaced by oxygen-ligands. Thus for Ti" S > 0 but for Ti"' 0 > S.l17 Thioxan (XXX) and dithian (XXXI) give 2 1 adducts with TiCI and TiBr,; thioxan is S-bonded in these complexes.llS With V1~* dialkyl sulplides form VCI,(R,S) (R = Me,Et); these complexes are monomeric with a dipole moment of 2.5 D; their spectra are interpreted as indicating a trans-trigonal bipyramidal configuration.l18 Complexes of Nbv and TaV have been isolated MX,,R,S (M = Nb Ta; X = F C1 Br; R = Et Me) and Tar, R2S are thermally more stable than the corresponding ether complexes.11g The ether can be replaced by the corresponding thioether in MCI,,Et ,O and in NbCI,,Prn,O.These results suggest that the bond energy of the M-S bond is greater than that of the M-0 bond in these complexes. This is surprising in view of the avidity of these metals for oxygen. However Me,O is not displaced by Me,S from MCI,,Me,O nor yet is Me,S displaced by Me,O from TaCI,,Me,S.120 The authors state that one cannot accept the explanation of simple back- co-ordination of unshared d electrons from a penultimate shell since there are none.They suggest three factors (i) a steric effect; (ii) the dipole moment of the ligand; (iii) the polarisabilities of sulphur and oxygen. The first factor is responsible for the greater stability of the methyl adducts. The second and third factors are responsible for the greater stability of the thioether adducts since the dipole moment of Et,S (1.58 D) is greater than that of Et,O (1-15 D) while the polarisability of sulphur in Et,S is much greater than that of oxygen in Et,O. Mov gives the complexes MoOC13,R2S (R = Me,Et,Prn).l2I 116 A. V. Babaeva V. A. Golovnya and L. A. Nazarova Zhur. neorg. Khim. 1959 117 K. Baker and G. W. A. Fowles Proc.Chern. Soc. 1964 362. 118 G. W. A. Fowles Proc. VIIIth Internat. Conf. on Co-ordination Chem. Springer 119 F. Fairbrother and J. F. Nixon J. 1962 150; F. Fairbrother K. H. Grundy lZo D. B. Copley F. Fairbrother and A. Thompson J. 1964 315. K. Feenan and G. W. A Fowles Znorg. Chem. 1965,4 310. 4 1741. Verlag Vienna 1964 p. 208. and A. Thompson J. 1965 765. 404 QUARTERLY REVIEWS A thermochemical study of the relative donor strengths of some oxygen and sulphur ligands towards SnCl and SbCl has shown that Et,S is co-ordinated more strongly than Et,O indicating that Snrv and Sbv should be regarded as borderline rather than class (a) metals.122 Diethyl sulphide and tetrahydrothiophen form weak complexes Et,S.BH, Et,S.BF, (CH,),S.BH, and (CH2),S.BF3; BH3 forms more stable adducts than BF with R2S but the reverse is true for R,0.12 Stable 1 :I and 1 :2 complexes are formed by SnCI with dialkyl sulphides and saturated heterocycles such as tetrahydrothiophen but the unstable thiophen adduct is not bonded through sulphur but is considered to be a n-electron complex.124 Dimethyl telluride forms the unstable complexes AgI(Me,Te) and Me,Te(AgI),; the selenium complex Me,Se(AgI) is very unstable but Me,S and Me20 show no sign of co-ordinating to AgLs The stability order towards HglI halides is also Te > Se > S.s The infrared spectra of the lY4-dithian (XXXI) complexes M(C4H,S2)C1 (M = Pt Cu Cd Hg) and Au(C4H,S2)Cl have been interpreted as denoting that in these compounds the ligand is in the chair configuration as in 1,4-dithian itself and that the complexes are most likely polymeric with each sulphur atom of the ligand bound to a different metal atom.On the other hand the thioxan (XXX) complexes M(C4H,0S)2 (M = Pt Cu Hg) are monomeric with the ligand bound through the sulphur atom 0dy.125 2,6-Dimethyl-4-thiopyrone (DMPT;XXXII) forms compounds M(DMTP),Cl (M = Cu,Ag) M(DMTP),C12 (M = Fe,Co,Ni,Pd,Pt,Hg) and M(DMTP)&l (M = Sb Bi). It is claimed that the infrared spectra indicate that the ligand is S-bonded.126 The ligand is a poor donor as it is readily displaced by water pyridine or halide ion. 5.2. Sulpho- Seleno- and Telluro-derivatives of Metal Carbonyls and Cyclopentadieny1s.-Many examples are known of metal carbonyl lZ2 I. Lindqvist and M. Zachrisson Actu Chem. Scand. 1960 14 453. 123 T. D. Coyle H. D. Kaesz and F. G. A. Stone J .Amer. Cheni. SOC. 1959,81,2989. 12* I. P. Gol’dshtein E. N. Gur’yanova and K. A. Kocheshkov Dokludy Akad. lZ5 P. Hendra and D. B. Powell J. 1960 5105. lZ6 H. B. Gray E. Billig R. Hall and L. C. King J. Inorg. Nuclear Chem. 1962 Nauk S.S.S.R. 1962 144,569 788. 24,1089. LIVINGSTONE METAL COMPLEXES OF LIGANDS 405 complexes which contain sulphur-ligands ; there are also a few carbonyl complexes known with selenium-ligands. The complex [Fe(CO),SEt ] is d i m e r i ~ l ~ ~ and is similar to the isoelectronic nitrosyl sulphide [Fe(NO)2SEt]2.128 An X-ray investigation of [Fe(CO),SEt] shows that the structure is that of two distorted tetragonal pyramids joined together as shown in (XXXIII).129 The Fe-S distance (2.26 A) is the same as in [Fe(NO),SEt], while the Fe-Fe distance is 2-54 A compared with 2.72 in the nitrosyl complex.13o The geometry gives an explanation of the exist- ence of two isomers of the analogous SMe complex,131 because the alkyl groups can be syn or anti.Data obtained from infrared and proton reson- ance spectra and from magnetic and dipole moment measurements indicate that the compounds [Fe(CO),L] (L = S Se SPh and SeEt) have two bridging S or Se atoms which do not form a plane with the Fe The structure of these compounds is no doubt similar to that of [Fe(CO),SEt] (XXXIII). Infrared spectral measurements on the trisubsti- tuted molybdenum carbonyls Mo(CO),L [L = Me,S Et,S (CH2)4S (NH2),C=S MeC(=S)NH,] show that all the ligands are bonded through sulphur. The thiourea and thioacetamide complexes are the most stable. A comparison of the C-0 stretching frequencies indicates that the thioethers have a substantial tendency to function as n-acceptors though not to the same extent as tertiary phosphines and arsines while thiourea and thioacetamide have only a small though definite tendency to do Some of the carbonyl groups in Ni(C0)4 and Mo(CO) can be directly substituted by dialkyl sLilphides to yield Ni(CO)4 (R,S) and Mo(CO), (R2S) (n = 1 or 2).The force constant of the C-0 stretching mode de- creases with increase of sustitution indicating some degree of rr-bonding by the thi~ether.~,~ The reaction of 2,5-dithiahexane (S-S) and 3,6,9-trithiaundecane (S-S-S) with the hexacarbonyls M(CO)6 (M = Cr Mo W) yields com- plexes M(S-S)(CO) (M = Cr,Mo,W) and M(S-S-S)(CO) (M = Cr Mo but not W). The reaction of M(S-S)(CO) with I2 yields the seven- co-ordinate complexes M(S-S>(CO),I (M = Mo W).On the other hand the compound Mn(S-S-S)(CO),Br is probably six-co-ordinate as it gives a sulphonium salt with methyl iodide suggesting that in this complex the ligand is co-ordinated through only two of the sulphur A thio-bridged rhenium complex [Re(CO),SPh] is known136 and thio- lZ7 H. Reihlen A. von Friedolsheim and W. Oswald Annalen 1928 465 72. lZ8 K. A. Hofmann and 0. F. Wiede Z. anorg. Chem. 1895,9 300. lZ9 L. F. Dahl and Chin-Hsuan Wei Inorg. Chem. 1963 2 328. 130 J. T. Thomas J. H. Robertson and E. G. Cox Actu Cryst. 1958 11 599. 131 R. B. King J. Amer. Chem. SOC. 1962 84 2460. 132 W. Hieber and W. Beck 2. anorg. Chem. 1960,305,265; S . F. A. Kettle and L. E. Orgel J. 1960,3890; Chin Hsuan Wei and L.F. Dahl Inorg. Chem. 1965,4,493. 133 F. A. Cotton and F. Zingales Chem. and I d . 1960 1219; Inorg. Chern. 1962,1 145. 13* G. Bouquet and M. Bigorgne Bull. SOC. chim. France 1962,433. 136 H. C. E. Mannerskantz and G. Wilkinson J. 1962,4454. 136 W. Hieber and L. Schuster 2. unorg. Chem. 1956,285,205. 406 QUARTERLY REVIEWS bridged iron complexes have been studied by King.l3l$l3' Infrared spectral measurements confirm that these compounds do not have bridging carbonyl groups. The structure of the diamagnetic volatile complex Fe3(C0)9S2 is probably that shown in (XXXIV). The reaction of sulphur with Co,(CO) yields a mixture of several sul- phur-containing cobalt carbonyl complexes.138 Other sulphur-containing cobalt carbonyls have also been reportedlag but their structures have not been elucidated.Cyclopentadienyl complexes [(C,H,)Fe(CO)SMe], [(C5H5)CoSMe], been prepared.140 The iron and cobalt complexes apparently contain two the chromium complex three and the molybdenum and vanadium complexes four sulphur-bridges. Similar nickelocene derivatives [(C,H,) NiSR] (R = Me,Et,Ph) are known.141 King and Bisnette14 recently reported a new type of complex with a sulphur-containing ligand co- ordinated through a C=S bond; the postulated structures of two of these complexes are shown in (XXXV) and .(XXXVI). [(CtiH.5)Cr,(MeS)3I [(C~H~MO(M~S)ZI, and [(C.5J&>V(MeS),I2 have oc - yo-co (tJ (xxxv) c,yf-co O oc s (XXXVI) Several organotellurium-containing metal carbon yls are known :la3 (Ph,Te)Fe(CO),Br, (Ph,Te),Mn(CO),X (X = Cl Br I) and the dinieric complex [(p-MeOC,H,-Te)Fe(CO)3]2; the last compound must contain two tellurium bridges and is similar to the sulphur-bridged complexes described above.5.3. Complexes of Thiourea and its Derivatives.-Thiourea acts as a unidentate ligand farming strong complexes with (b) class metal ions in particular Cul AgI Aul and Hg". It is noteworthy that Agl is complexed more strongly than CdIr and Pb11.25 Thiourea reduces Cu" to Cur Au111 to Au' Ptw to Pt" and TeN to TeI1 and forms complexes with the metal 13' R. B. King Inorg. Chem. 1963 2 641 ; J. Amer. Chem. SOC. 1963 85 1584. 138 L. Marko G. Bor E. Klumpp B. Mark4 and G. Almasy Chem. Ber. 1963,96 955. 130 L. Mark6 G. Bor and E. Klumpp Angew. Chem. 1963,75,248; E. Klumpp L. Mark6 and G. Bor Chem. Ber. 1964 97 926; S. A. Khatlob L. Marko G. Bor and B.Mark6 J. Organometallic Chem. 1964 1 373. 140 R. B. King P. M. Treichel and F. G. A. Stone J. Amer. Chem. SOC. 1961 83 3600; R. B. King ibid. 1963,85 1587; R. H. Holm R. B. King and F. G. A. Stone Inorg. Chern. 1963 2 219. 141 W. K. S-chropp J. Inorg. Nuclear Chem. 1962 24 1688. 142 R. B. King and M. B. Bisnette J. Amer. Chem. Soc. 1964 86 1267. 148 W. Hieber and J. Kruck Chem. Ber. 1962 95 2027. LIVINGSTONE METAL COMPLEXES OF LIGANDS 407 in the lower oxidation state; hence the claim that thiourea stabilises Rhw in is surprising. Infrared studies have shown that urea co-ordinates through nitrogen in its complexes with PdII and PtII and in Sn(urea),Cl but through oxygen in its complexes with CrIII FeIII Cu" and Zn" and in S n ( ~ r e a ) B r . ~ ~ ~ l ~ ~ On the other hand thiourea co-ordinates through sulphur not only with class (b) but with SnIv,la6 Pb11,14* and Te*I;149 however infrared spectral measurements indicate that the ligand is co-ordinated through nitrogen in Ti(tu),CI (tu = thiourea).150 The structures of a number of thiourea complexes have been determined by X-ray analysis.The nickel atom in [Ni(tu),Cl,] is surrounded by four sulphur atoms at 2-45 8 and the two chlorine atoms are at 2.40 A in the trans-octahedral Lopez-Castro and Truter15 also determined the structure of this high-spin compound and commented that in diamagnetic Nil1 complexes the Ni-S distance is 2.1-2.3 A whereas in six-co-ordinate complexes it is 24-26 A; this is in accordance with ligand-field theory which predicts that the stronger field produced by the ligands closer to nickel will cause the dy electrons to pair in the d, orbitals so producing a diamagnetic complex with no close neighbours in the z direction.The compound Ni(tu),(NCS) is octahedral and polymeric; the sulphur atom of each thiourea molecule is bound to two nickel atoms. Each nickel atom is bound to four sulphur atoms at a distance of 2-54 A and to two nitrogen atoms (at 1.99 A) from the thiocyanate g r 0 ~ p s . l ~ ~ The S-C dist- ance is abnormally long (1.77 A) compared with 1.64 A in thiourea itself154 and in Ni(tu)4C12.155,151 The compounds M(tu),(NCS) (M = Mn,Co,Cd) are isostructural with Ni(tu),(NCS) but Zn(tu),(NCS) is In the compound [Ni(py),(tu),Cl,] (py = pyridine) the sulphur atoms of two of the four thiourea molecules are probably bridging in a similar manner. 156 The complex Cd(tu),Cl is tetrahedral but Pb(tu),Cl has a polymeric structure in which the lead atom is seven-co-ordinate being surrounded by four bridging sulphur atoms at 2.9-3-1 8 and two bridging chlorine atoms at 3.2 A while the non-bridging chlorine is at a distance of 2-75 A 144 F.Pantani and P. G. Desideri Talanta 1960 5 69. 145 R. B. Penland S. Mizushima C. Curran and J. V. Quagliano J. Amer. Chem. Soc. 146 D. S. Bystrov T. N. Sumarokova and V. N. Filiminov Optics and Spectroscopy A. Yamaguchi R. B. Penland S. Mizushima T. J. Lane C. Curran and J. V. 1957,79 1575. 1960 9 239. Quagliano J. Amer. Chem. SOC. 1958 80 527. 148 M. Nardelli L. Cavalca and A. Braibanti Gazzetta 1956 86 867 1037. 149 0. Foss and S. Hauge Acta Chent. Scaitd. 1959 13 1252. 150 R. Rivest Canad. J. Chem.1962 40 2234. 161 M. Nardelli Gazzetta 1959 89 1616. lSP A. Lopez-Castro and M. R. Truter J. 1963 1309. lS3 M. Nardelli A. Braibanti and G. Fava Gazzetta 1957 87 1209. lS4 R. W. G. Wyckoff and R. B. Corey 2. Gist. 1931 81 386. 155 L. Cavalca M. Nardelli and A. Braiban Gazetta 1956 86 942. 156 A. V. Babaeva and Yan Vei-da Zhur. neorg. Khim. 1960 5 2735. 408 QUARTERLY REXIEWS from the central metal In [Te(tu),]Cl each tellurium atom has a square-planar arrangement of four sulphur at0ms.l5~ The complex [Mo(tu),CI,] has a magnetic moment of 3.71 B.M.,158 within the range 3.66-3.88 B.M. found for most Mo*II but the complex [Mo2(tu),C1,] has a moment of 0.59 B.M. In the latter complex the two molybdenum octahedra are bridged by either three chlorine or three sulphur atoms (from the thiourea groups) and the low value of the magnetic moment is ascribed to spin-orbit coupling due to metal-metal interaction either directly or via the bridging atom~.~~*J60 Complexes of NN'-substituted thioureas (XXXVII); R == alkyl or aryl R' = alkyl aryl or H have been extensively studied.Ethylenethiourea R\ H,C- N\H c= s I ,c=s d'" R\N/ R" (xxxvr I) H2C-NH (xxxvlrl) (etu; XXXVITI) with Cul and Agr gives complexes containing 1,2 3 or 4 molecules of ligand whereas with Aul complexes with 1 or 2 ligand molecules only are formed.lB1 Complexes of ethylenethiourea with Fe" Co" and NiI1 have also been i n v e ~ t i g a t e d . ~ ~ ~ $ ~ ~ ~ Cur and Ag* complexes with other substituted thioureas are known; their structures are often complex.164 TeI1 and TeTV complexes of several substituted thioureas have been reported.la5 The complexes [Ni(nt~)~X,] [X = C1,Br; ntu = 1-(1-naphthy1)- 2- thiourea] are tetrahedral ; the paramagnetic compounds Ni(etu),X (X = C1,Br) are octahedral and have been isolated in cis and trans forms but the iodo-complex Ni(etu),I is six-co-ordinate166 and is a rare example of a tetragonal diamagnetic NiII complex,167 while the perchlorate [Ni(etu),](ClO& is diamagnetic and square-planar.Holt and Carlinl66 have pointed out that systems in which Nil1 attains a variety of stereo- chemical environments with the one donor atom are rare yet with 1-( 1 -naphthyl)-2-thiourea Nirl gives tetrahedral and with ethylenethiourea paramagnetic octahedral diamagnetic tetragonal and diamagnetic square- 157 0. Foss and s. Hauge Acta Chern. Scand. 1961 15 1616.168 V. B. Evdokmov V. V. Zelenstov I. D. Kolli Wen-Hsia Tang and V. I. Spitsyn 159 B. N. Figgis and J. Lewis in "Modern Co-ordination Chemistry," ed. J. Lewis 160 L. F. Lindoy S. E. Livingstone and T. N. Lockyer Austral. J. C h e l ~ 1965 18 161 G. T. Morgan and F. H. Burstall J. 1928 143. 162 M. Nardelli I. Chierici and A. Braibanti Gazzetta 1958 88 37. 163 R. L. Carlin and S. L. Holt Inorg. Chern. 1963 2 849. 164 S. N. Banerjee and A. C. Suicthankar J. Indian Chem. SOC. 1963 40 387. 166 0. Foss and W. Johannessen Acta Chern. Scand. 1961 15 1939 1947 and 166 s. L. Holt and R. L. Carlin J. Amer. Chem. SOC. 1964 86 3017. 167 C. M. Harris and S. E. Livingstone Rev. Pure and Appl. Cheni. (Australia) DokIady Akad. Nauk S.S.S. R. 1962,145 1282. and R. G. Wilkins Interscience New York 1960 p.425. in the press. references therein. 1962 12 16. LIVINGSTONE METAL COMPLEXES OF LIGANDS 409 planar complexes in addition to what appear to be the first isolated geo- metrical isomers of octahedral Nilx. NN'-Diethylthiourea NN'-diphenylthiourea and NN-diplienylthiourea behave as bidentate chelating agents being bonded through both nitrogen atoms in the complexes 2TiC14,L (L = substituted thiourea).150 S-Acetyl- thiourea (satu) forms with CuI only a 1 :1 complex Cu(satu)C1;16* it would be interesting to know if the ligand is N- or S-bonded in this compound. Bis-chelated complexes of Co" Nil1 and Cull are formed by a number of NN'-diaryl-N-hydroxothioureas (XXXIX) ; the hydrogen of the OH group is replaced and the ligand is most probably co-ordinated through sulphur and oxygen as in (XL).las R-NH-F-Y-R' R-NH-C-N-R' dH (XXXIX) Data obtained from an investigation of Cox' complexes indicate that ethylenethiourea is higher in the spectrochemical series than R2S lying between N3- and Ph3P0.B' is 66 of the value for the free ion suggesting appreciable orbital overlap which is to be expected for the large polarisable sulphur atom; the position of etu is thus at the low end of the nephelauxetic series approximately equal to that of iodide ion.la3 5.4. Selenourea Complexes.-Selenourea (su) ghes precipitates which are soluble in the presence of excess of the ligand with Pd,Cu,Ag,Au,Hg Sn As Sb and Bi and a Pb complex which is insoluble in excess. BiIII forms a yellow complex with a Bi:su ratio of 1 :9 and a red 1 :12 c0mpIex:1~0 their structures are unknown but they are more stable than [Bi(t~)~]~+.l~l Whereas Os1I1 forms a 6 1 complex with thiourea [OS(~U)~],~+ with selen- ourea an 8:l complex [OS(SU),]~+ is formed; apparently the latter is a genuine eight-co-ordinate complex as the bluish-green colour is not produced in solution until the eighth mol.of ligand has been added.172 5.5. Thioacetamide and Thiobenzamide Complexes.-Thioacetamide CH3CS-NH2 (tarn) gives complexes [M(tam)4]+ (M = Cu Ag) which are S-bonded and tetrahedral.173 On the other hand AUI forms a two- covalent complex [Au(tam) 2]Br.174 With bivalent metals tetrahedral lB8 S. K. Siddhanta and S . N. Banerjee J . Indian Chem. Soc. 1961 38 747. 170 A. T. Pilipenko and I. P. Sereda Zhur. analit. Khim. 1958 13 3. 171 S. I. Gusev and L. A. Ketova Zhur neorg. Khim.1961 6 1881. 17* A. T. Pilipenko and I. P. Serada Zhur. neorg. Khim.. 1961 6,413. 173 N. S. Kurkaskov Zhur. Russ. Fiz.-Khim. Obshch. 1893,25 693; J. prakt. Chenz. 1895 [2] 51,234; E. G. Cox W. Wardlaw and K. C. Webster J. 1936,775; D. Rosen- thal and T. I. Taylor J. Amer. Chem. Soc. 1960,82,4169. 174 F. H. Brain and C. S. Gibson Brit. Assoc. Advaizce Sci. Report 1938 p. 37. B. Hirsch J . prakt. Chem. 1961 12 264. 410 QUARTERLY RFWfEWS complexes [M(tam),Cl,] (M = Fe,Co,Zn,Cd) and octahedral complexes [M(tam),Cl,] (M = Ni,Cd) are known.175 Dioximato-complexes of ColI1 containing one or two molecules of thiocetamide have been recently reported viz. [Co(DH),(tarn)X] and [Co(DH),(tam),]+; similar complexes are known with other unidentate sulphur ligands such as thiourea and thiosemicarbazide.97 Thiobenzamide behaves similarly to thioacetamide and forms complexes with class (b) metal i011s.l~~ 5.6. Triphenylphosphine Sulphide and Se1enide.-Whereas triphenyl- phosphine oxide Ph,PO forms a variety of interesting complexes with first-row transition the donor properties of Ph,PS and Ph,PSe appear to be rather weak ; however the insoluble complexes Pd(Ph,PS) CI, Pd(Ph,PSe),Cl, and Sn(Ph,PSe),Cl have been 5.7. Dimethyl Sulphoxide Complexes.-Dimethyl sulphoxide Me,SO forms complexes with transition metal ions. The ligand is bound through oxygen except in the PdlI and Pi" complexes where it is S-b~nded.l~~ 6. Complexes of Chelate Ligands 6.1. Chelate Ligands with Thioether or Selenoether Groups.-Chelate ligands with two thioether groups give similar complexes to those formed by unidentate thioethers.1,2-Dialkyl- and diaryl-dithioethanes RSCH2CH2.SR form stable complexes with most (b) class metals and some borderline metals such as Bilrl and Sniv.100~180~181,182 The diseleno- ether EtSe.CH,-CH,-CH,.SeEt forms a 1 :1 complex with PtCl similar to the dithioethers.lo3 Spectral data obtained from NiI1 complexes shows that 1,2-dialkyldithioethanes have a ligand-field stronger than ethylenethiourea and almost as strong as NH3.1a3 For organometallic derivatives of transi- tion metals to be stable it is necessary to havc present in the complex other ligands such as tertiary phosphines capable of accepting dm electrons from the metal atom ; 1,2-dirnethyldithioethane has a sufficieiitly strong ligand- field to stabilise the Pt-Me bonds in (EtSCH,CH2-SEt)PtMe,.194 Titan- ium(1v) form the complexes (RSCH2CH2*SR)TiX4 (R = Ph,Me; X = CI,Br).l17 Although aliphatic dithioethers form octahedral NilI 175 M.Nardelli and J. Chierici Gazzetta 1957 87 1478. 176 F. KaSpArek and J. Mollin COIL Czech. Chem. Conzm. 1960 25 2919. 17' E. Bannister and F. A. Cotton J. 1960 1873 1878. 178 E. Bannister and F. A. Cotton J. 1960 1959. F. A. Cotton and R. Francis J . Amer. Chem. Soc. 1960 82 2986; F. A. Cotton R. Francis and W. D. Horrocks J. Phys. Chem. 1960,64 1534; D. W. Meek D. K. Straub and R. S . Drago J. Amer. Chem. Soc. 1960 82,6013; J. Selbin W. E. Bull and L. H. Holmes J. Inorg. N i d e c r Chem. 1961 16 219. 180 L. Tschugaeff Eer. 1908 41 222. l*lZ. Tschougaeff and D. Fraenkel Compt. rend. Acad. Sci. 1912 154 33; L. Tschugaeff and A. Kobljanski Z .anorg. Chem. 1913 83 8. G. M. Bennctt A. N. Moses and F. S. Statham J. 1930 1668. R. L. Carlin and E. Weissberger h w g . Chem. 1964 3 61 1. l** J. Chatt and B. L. Shaw J. 1959 705. LIVINGSTONE METAL COMPLEXES OF LIGANDS 41 1 complexes 4-methylthioveratrole (XLI) shows little tendency to co- ordinate to Ni11.18035 H-N - CH; CH L Me/ SMe 0 SMe Ou) Chelate ligands having a thioether sulphur and another donor atom often co-ordinate more strongly than dithioethers. The co-ordinating ability of some /3-aminothioethers RS-CH2-CM ,.NH has been investi- gated.lss The free NH2 group in the Ptw complex (XLII) is capable of forming salts and the complex was optically resolved showing that the asymmetry is due to the tercovalent sulphur atom.ls7 o-Methylmercaptobenzoic acid (XLTII) forms inner complexes with Pd" Cu" Cd" and Hgl1 but not with Pt" which has a low affinity for CH,- SMe '2CH,-AsMe SMe CH 9 CHiSMe (XLVI) (x LV I I) (x LV I I I) (x Ll x) oxygen-ligands.la8~ls9 This and other o-alkylmercaytobenzoic acids form halogen-bridged dinuclear complexes (XLIV; M = Pd Cu Hg; X = C1 Dimethyl-o-methylthiophenylarsine (XLV) forms strong complexes with Co" RhIII II-111 Nil] Pd" PtII CuI AgI and A u ~ .~ ~ ~ ~ This ligand forms complexes of PdII and Pt" having co-ordination numbers lE6 R. Backhouse M. E. Foss and R. S. Nyholm J. 1957 1714. lE6 E. Gonick W. C. Fernelius and B. E. Douglas J. Arner. Chem. Soc. 1954 76 lE7 F. G. Mann J. 1930 1745. lEE S. E. Livingstone R. A. Plowman and J. Sorenson J. Proc. Roy. SOC. New South lS9 S . E. Livingstone and R. A. Plowma J.Proc. Roy. Sac. New South Wales 1951 lgo S. E. Livingstone and R. A. Plowman J. Proc. Roy. SUC. New South Wales 1950 lgl S . E. Livingstone J. 1956 1989. lga S. E. Livingstone Chem. and Ind. 1957 143; J. 1958 4222. lg3 B. Chiswell and S. E. Livingstone J. 1959 2931 ; 1960 97 3181. lg4 B. Chiswell and S. E. Livingstone J. 1960 1071. or Br),189,190,191 4671. Wales 1950 84 28. 85 116. 84,188. 412 QUARTERLY REVIEWS greater than the usual value of four.167J94J95 On the other hand dimethyl- 3-methylthiopropylarsine (XLVI) forms stable complexes with Pd" PVI and Cul only unstable products being formed with other (6) class rnetals.ls6 2-(2-Methylthioethyl)pyridine (XLVII ; R = Me) gives stable chelates with typically (6) metals.lS7 Measurements of stability constants indicate that 2-(2-methylthiomethyl)pyridine (XLVIII) also acts as a chelate group.lS8 8-Methylthioquinoline (XLIX; N-S) appears to give only 1 1 complexes of the type M(N-S)X2 with Pd" Pt" and Hg" but the 2:l complex [Cu(N-S),](ClO,) can be readily isolated.199 Schiff bases derived from 1,8-diarnino-3,6-dithiaoctane by the loss of two protons function as sexadentates; the bond angles about the sulphur atoms are such that the molecule can fit almost strainlessly around the six octahedral positions of a metal ion.The CoI1* complex of 1,8-bis(sali- cylideneamino)-3,6-dithiaoctane (L) was resolved and found to have a ( L) high value for the molecular rotation.200 The Schiff bases derived from pyridine-Zaldehyde and acetylacetone also function as sexadentates.201 Schiff bases formed by the condensation of salicyaldehyde with a,o- diamines of general formula H2N.[CH,];S. [CH,];S. [CH2IZ.NH (x y z = 2 or 3) form ColI1 complexes with very high values for 6.2 Chelate Ligands containing One Thiol Group.-The stability constants have been measured for some metal complexes of the following ligands :25 thioglycollic acid HSCH 2C02H ; p-mercaptopropionic acid HSCH,.CH 2C0 2H ; 2-aminoethanethiol HSCH ,CH ,.NH ; cysteine HSCH,CH(NH 2)C0 2H ; mercaptosuccinic acid HO ,CCH(SH).CH 2* C02H. However the available data do not usually allow comparisons to be made between complexes of analogous sulphur- and oxygen-ligands. It is noteworthy that cysteine forms extremely stable complexes with HglI and Fe'". The stabilities of UOZ2+ chelates of hydroxy- mercapto- and amino-derivatives of acetic propionic and succinic acids fall in the series NH2 > OH > SH203 and Am1" forms a stronger complex with glycollate lS5 S.E. Livingstone ahd B. Wheelahan Austral. J. Chem. 1964 17 219. B. Chiswell and S . E. Livingstone J. Inorg. Nuclear Chem. 1961 23 37. P. S. K. Chia S. E. Livingstone and T. N. Lockyer unpublished work. lo8 K. Kahmann H. Sigel and H. Erlenmeyer Helv. Chirn. Acta 1964 47 1754. lee L. F. Lindoy S. E. Livingstone and T. N. Lockyer unpublished work. F. P. Dwyer and F. Lions J . Amer. Chem. SOC. 1947 69 2917. 201 F. P. Dwyer N. S. Gill E. C. Gyarfas and F. Lions J. Amer. Chem. SOC. 1957 202 F. P. Dwyer N. S . Gill E. C . Gyarfas and F. Lions J. Amer. Chem. SOC. 1952 *03 M. Cefola R. C. Taylor P. S . Gentile and A. V. Celiano J. Phys. Chem. 1962 79 1269; F. Lions and K.V. Martin ibid. 1958 80 3858. 74,4188. 66,730. LIVINGSTONE METAL COMPLEXES OF LIGANDS 41 3 than with thioglycollate these results indicate the (a) class behaviour of UOZ2+ and Am1V. The stabilities of complexes of Co" Ni" Agl Zn" and Hg" with mercaptosuccinic acid are higher than for the nitrogen- and oxygen-analogues aspartic and malic The chelating tenden- cies of some a-mercaptoacetamides RNHCOCH2SH (R = aryl) have been examined; for these ligands U02 > Be > Ni > Mn.206 The chelates of ZnlI and PbI1 with o-aminobenzenethiol (LI) and 6-mercaptopurine (LII) are more stable than the corresponding ~xygen-chelates.~~~ N o-Aminobenzenethiol forms low-spin inner complexes with ColI and Ni11.26a~208 The yellowish-buff NiII complex can be converted in strongly alkaline medium into a deep blue compound which is said to contain 8-Mercaptoquinoline (LIII) forms strong chelates with class (b) metal^.^^^,^^^^^^^ There is a considerable bathochromic shift in the long- wave absorption maxima of the chelates of RuI" O P RhIII IrI** Pdl* and Pt"; this shift is ascribed to the high strength of the M-N bond.211 The chelates of the platinum metals with 5-chloro-8-mercaptoquinoline and 5-bromo-8-mercaptoquinoline are more stable than those of (LIII).The stability of the chelates of 5-chloro-8-mercaptoquinoline is in the order Pd > Pt > Rh > Ir > Ru > o-Methylthiobenzenethiol (LIV) 3-ethylthiopropane-1-thiol (LV) and NiIv.208 3-dimethylarsinopropane-1-thiol (LVI) form square-planar inner com- plexes with Ni" PdII and PtI*.26a9213 With PdII these ligands form thiolo- 204 I.Grenthe Acta Chem. Scand. 1962 16 1695. 205 G. R. Lenz and A. E. Martell Inorg. Chem. 1965 4 378. 206 D. F. Martin J. Amer. Chern. SOC. 1961 83 1076. 207 R. G. Charles and H. Freiser J. Amer. Chem. Soc. 1952,74 1385; G. E. Cheney H. Freiser and Q. Fernando ibid. 1959 81 2611. 208 W. Hieber and R. Bruck Naturwiss. 1949 36 312; Z. anorg. Chem. 1952,269 13. J. Bankovskis A. IevinS and Z . Lepina Zhur. analit. Khim. 1960 15 4; Yu.1. Usatenko and V. I. Suprunovich Latvijas P.S.R. Zinatnu Akad. Vestis Kim. Ser. 1963 181 ; A. Corsini Q. Fernando and H. Freiset Tulanta 1963,11,63. 210 E. V. Vasil'eva T. K. Nedopekin and V. E. Petrun'kin Ukrairz khim. Zhur. 1962 28 773. J. Bankovskis G. Mezarups and A. IevinS Zhur. analit. Khim. 1962 17 721. 219 3. Bankovskis G.Mezarups and A. IevinS Latvijas P.S. R. Zinatnu Akad. Vestis Kim. Ser. 1962 323; J. Cirule J. Bankovskis A. IevinS and J. Asaks ibid. 1964 135. llS S. E. Livingstone J. 1956 437. 414 QUARTERLY REVIEWS bridged compounds which are more stable than halogen-bridged com- plexes. The ligand (LV) gives similar PtlI bridged complexes and also binuclear complexes containing another metal atom in addition to Pt (LVII; M = Hg Pd). The cationic bridged complex (LVIII; X = C1 C10.J was also isol~ated.~~~ CH H2$' ';H2 x2 (LVII I) With Nil1 mercaptoacetic acid forms the monomeric complex [Ni(SCH2C02)2]2- and the multinuclear thiolo-bridged species [Ni (SCH2C02)6]4-.215 2-Aminoethanetl1.iol gives a trinuclear cationic complex (LIX; M = M' = Ni).21a 2-(2-Mercaptoethyl)pyridine (XLVII; R = H) gives inner complexes and binuclear bridged complexes with NP PdII and PtI* and also trinuclear complexes (LIX; M = Ni and M' = Ni Pd).217 Square-planar NilI complexes of some chelating thiols have been used in template reactions to produce macrocyclic ligands in sit^.^^* Template syntheses-a fascinating new development in co-ordination chemistry-involving sulphur ligands have been used to prepare quadri- dentate complexes of Zn" CdII and HgyI.219 Thioamidopyridine (LX) forms high-spin complexes [M(C,H 6N2S)3] (C1OJ2 (M = Fe Co Ni); complexes of ColI1 Cu' Cull Zn'l CdII and HgII are also known.220 N-2-Mercaptophenylene-2'-pyridylmethyleneimine (LXI) acts as a terdentate and forms binuclear complexes containing eight- 214 s.E. Livingstone J. 1956 1994. D. L. Leussing R. E.Laramy and G. S. Alberts J. Amer. Chem. Suc. 1960 82 216 D. C. Jicha and D. H. Busch Inorg. Chem. 1962 1 872 878. J. W. Wrathall and D. H. Busch Inorg. Chem. 1963 2 1182. *l* M. C. Thompson and D. H. Busch J . Amer. Chem. Soc. 1962 84 1762 D. H. Busch D. C. Jicha M. C. Thompson J. W. Wrathall and E. Blinn ibid. 1964 86 3651. 21g H. Jadamus Q. Fernando and H. Freiser J. Amer. Chem. Soc. 1964 86 3056. 220 G. J. Sutton Austral. J. Chem. 1963 16 371 1137; 1964 17 1360 4826. LMNGSTONE METAL COMPLEXES OF LIGANDS 41 5 (LXI I) co-ordinate Mo"' (LXII; X = C1,Br). Similar Mo*I1 complexes are formed by a number of other chelating thiols; all have anomalously low magnetic moments due to metal-metal interaction.lsO The thionitrosyl complex Pt(S ,N,H) has the cis-planar configuration (LX111)221 but the Nil1 complex of the mono-N-methyl derivative viz.Ni(S,N,Me), is tmns-planar.222 Complexes of neutral and deprotonated guanylthiourea (LXIV) are The crystal structure of dithio- oxamide (rubeanic acid ; LXV) has been recently determined ;224 this ligand forms insoluble complexes which are probably polymeric.225 Its use as an analytical reagent has been reviewed226 and the structures of its metal complexes and those of NN'-disubstituted dithio-oxamides have been discussed.227 Thiosemicarbazide exists in tautomeric forms (LXVI) and (LXVII) and can act as a neutral or a charged chelate group. The inner complexes of Ni" Pd" and PtIJ were each obtained in two forms which were assumed to be cis-trans isomers.22s However a structure determination of the red form of Ni(CH4N& shows it to be trans square-planar with nearly com- plete localisation of the double bond between the nitrogen and carbon 2a1 I.Lindqvist and J. Weiss J. lnorg. Nuclear Chem. 1958 6 184; I. Lindqvist and R. Rosenstein ibid. 1958 7 ,421. 222 J. Weiss and M. Ziegler Z. anorg Chem. 1963 322 184. 243 P. Riiy Chem. Rev. 1961 61 313. 224 P. J. Wheatley J. 1965 396. 225 K. A. Jensen Z. anorg Chenz. 1944,252,227; R. V. G. Ewens and C. S. Gibson J. 1949 3308. 22G R. R2y and J. Xavier J. Indian Chem. Soc. 1961 38 535. 227 R. N. Hurd G. de la blatx G. C. McElheny and L. V. Peiffer J. Amer. Chem. SOC. 1960 82 4454; R. N. Hurd G. de la Mater G. C. McElheny and J. P. Mc- Dermott in "Advances in the Chemistry of the Co-ordination Compounds," ed. S. Kiichner Macm'llan New York 1961 p. 350.228 K. A. Jensen and E Raneke-Madsen 2. anorg. Chem. 1934.219.243 K. A. Jensen ibid, 1934,221,6 11. 416 QUARTERLY REVIEWS (LXVO . ( D V I I) (LXVII I> atonis. Although both forms are diamagnetic it seems doubtful if the grey form is the cis-isomer but a structure determination was not pos- ~ i b l e . ~ ~ ~ Two isomeric forms of the inner complex of CoIII have been isolated ; their colours suggest that they are cis-trans i~omers.~~O Com- plexes of Schiff bases of thiosemicarbazide have been recently reported.231 3-Mercapto-l,5-diphenylformazan (dithizone = HDz ; LXVIII) forms highly coloured inner complexes with most (6) class and several borderline metals; it forms Et,SnDz Et,SnDz and Et,pbD~.,~~ A crystal structure determination of Hg Dz2,2py shows that the mercury atom is bound to the sulphur atoms of each dithizone but the nitrogen atom of the azo-group is weakly co-ordinated at a distance of 04-0-5 8 greater than expected for a covalent bond giving the molecule a distorted tetrahedral arrange- ment.233 However in CuDz the copper atom is square-planar and the ligand is bidentate.234 6.3.Alkyl Xanthates Dialkyl-dithiocarbamates -diselenocarbamates -dithiophosphates and -diselenophosphates.-Four membered chelate rings are formed by alkyl xanthates (LXIX) dialkyl-dithiocarbamates (LXX) -diselenocarbamates (LXXI) -dithiophosphates (LXXII) and -diseleno- phosphates (LXXIII). These ligands form inner complexes with class (b) bi- and ter-valent metals. The xanthate complexes are less stable than those of dithio~arbamates.~~~ The Nirl complexes of (LXIX) (LXX) and (LXXTI) are ~ p i n - p a i r e d .~ ~ ~ ~ ~ ~ ~ ~ ~ As might be expected the Hg" complex of diethyldithiocarbamate has a high stability constant.238 Reviews have 239 M. Nardelli and P. Boldrini Gazzetta 1961 91 280; L. Cavalca M. Nardelli 230 N. M. Samus Zhur. neorg. Khim. 1963 8 72. 231 A. V. Ablov and N. I. Belichuk Zhur. neorg. Khim. 1963 8 77 612; A. V. 232 H. Irving and C. F. Bell J. 1954 4253; H. Irving and J. J. Cox J. 1961 1470 233 M. M. Harding J. 1958 4136. 234 R. F. Bryan and P. M. Knopf Proc. Chem. Soc. 1961 203. 235 L. Malatesta Gazzetta 1940 70 541 553. 230 L. Cambi and L. Szego Ber. 1931 64 2591 ; 1933,66,656. Sa7 L. Malatesta and R. Pizzotti Chimica e industria 1945 27 6. 2a* W. Kemula A. Hulanicki and W. Nawrot Roczniki Clzem. 1962 36 1717. G. Fava Acta Cryst.1962 15 1139. Ablov and N. V. Gerbeleu ibid. 1964,9 85. and references therein. LIVINGSTONE METAL COMPLEXES OF LIGANDS 417 been written on metal complexes of d i t h i o c a r b a m a t e ~ ~ ~ ~ ~ ~ ~ and dithio- while the visible-ultraviolet spectra of the transition-metal complexes of these ligands have been d i s c ~ s s e d . ~ ~ ~ ~ ~ ~ ~ The infrared spectra of dithiocarbamate complexes show that there is considerable double-bond character in the C-N bond (LXXIV) but in the xanthate complexes the canonical form (LXXV) contributes little to the (LXX I v) (LXXV) structure. 243 The dithiophosphate dithiocarbamate about the nickel P-S bond has little double-bond character in the diethyl- comp1exes.l A structure analysis of Nil1 NN-diethyl- shows square-co-ordination of the four sulphur atoms and the C = S double bond is delocalised in the chelate ring while the C-N and C-S distances are shorter than expected for single bonds .244 In Cu" NN-diethyldithiocarbamate the copper atom is at the apex of a pyramid with four sulphur atoms at the base.245 In the di-n-propyl homologue the copper atom is five-co-ordinate and 0.4 A above the plane of the square containing the four sulphur atoms; there are four short (2-32 A) and one long intermolecular (2.71 A) Cu-S Although readily soluble in organic solvents NN-diethyldithiocarbamate complexes of Cul Ag' and AuI are polymeric the degree of polymerisation being 2 for Au 4 for Cu and 6 for Ag; the Au atom is two-covalent while Cu and Ag are tri-co-ordinate.The ligands are not chelated but bound through both sulphur atoms to different metal atoms.All the complexes have short metal-metal distances that for the Au complex (2.76 A) being shorter than in the metal. The dimethyl homologues are practically insoluble in organic solvents and are probably highly polymeric. It seems that the steric re- quirement of the ligand is the important factor in deciding the co-ordina- tion number of the metal and the configuration of the complex.247 The diamagnetic complex [Co(Me,NCS,),NO] has a rectangular pyramidal configuration with NO at the apex and the Co atom lying 0.54 above the basal plane.248 The MoV xanthate complex [(EtOCS,) 23Q K. Gleu and R. Schwab Angew. Chem. 1950 62 320. 240 M. Delepine Bull. Soc. chim. France 1958 5 . 241 A. I. Busev and M. L. Ivanyutin Trudy Komisii Anal. Khirn.Akad. Nauk S.S.S.R. 242 S. Kida and H. Yoneda J. Chem. SOC. Japan 1955 76 1059. 243 J. Chatt L. A. Duncanson and L. M. Venanzi Suomen Kemi 1956 29 75. 244 E. A. Shugam and V. M. Levina Kristallografiya 1960 5 32. 245 R. Bally Coinpt. rend. 1%3 257 425. 246 G. Peyronel and A. Pignedou Ricerca x i . 1959,29,1505; Gazzetta 1962,92,745. 247 R. Hesse "Advances in the Chemistry of the Co-ordination Compounds" (Proc. 6 ICCC Detroit) ed. S. Kirchner MacMillan New York 1961 p. 314; Arkiv. Kemi 1963 20 481. 248 P. R. H. Alderman P. G. Owston and J. M. Rowe J. 1962 668. Inst. Geokhim. i anala. Khim. 1960 11 172. 41 8 QUARTERLY REVIEWS MoO],O first reported by Malate~ta,,~~ has a configuration consisting of two distorted molybdenum octahedra sharing a bridging oxygen atom; the metal-metal interaction occurs via the bridging oxygen as the Mo-Mo distance is 3.72 A.2so In the monomeric diamagnetic complex ReCl (Et2NCS2) the Rerx1 atom is said to be tetrahedral with d3s hybridisa- t i ~ n .~ ~ l The infrared spectrum of the PtIr complex of NN-diethyldithiocarba- mate indicates single bond-order in the C-S bond while the Pt-S stretching mode occurs at 375 cm.-1.252 The M-S stretching frequencies of the Ni" Pd" CoxII and CrIIx chelates of diethyldithiophosphate occur in the range 308-358 Whereas Fexxr complexes of xanthates are spin-paired and those of monoalkylcarbamates RNHCS2- are spin-free the magnetic moments of the complexes Fe(RR'NCS 2)3 are temperature-dependent and lie between the values expected for one and five unpaired electrons. The ligand-field in these complexes is close to the critical value which determines the transi- tion 3 d ~ ~ - + 3 d ~ ~ 3dy2.29236 Complexes of dialkyl-diselenocarbamates and -diselenophosphates have been r e p ~ r t e d .~ ~ * ~ ~ ~ Although most of the known complexes of the ligands described in this section contain class (b) or borderline metals SeII complexes have been reported with dialkyldiselen~carbamates~~~ and diethyldithi~phosphate.~~~ Amines give 1 1 adducts with Zn and Cd NN'-diethyldithiocarbamate ; the latter is the first reported example of five-co-ordinate CdII.258 Dioximato-complexes of CoxrI are known with 1 and 2 molecules of dialkyldithiocarbamate which is displaced by halide The infrared spectra of the carbonyl complexes Mn(CO),(Et,NCS 2) Fe(CO),(Me,NCS,), and Rh(CO)2(Me,NCS,) show partial double- bond character in the C-N bond while the sulphur-ligands have no marked effect on the M-CO multiple bonding.260 The NiIX complexes of ethyl xanthate and diethyldithiophosphate form green 2 1 pyridine adducts which are octahedral and parainagnetic.12 The spectra of solutions of Nil* diethyldithiophosphate containing secondary or heterocyclic amines have been interpreted as indicating distorted five-co-ordinatioii.261 Ethyl xanthate diethyldithiophosphate and diethyldiselenophosphate 24g L.Malatesta Gazzetta 1939 69 408. 250 A. B. Blake F. A. Cotton and J. S. Wood J. Amer. Chern. SOC. 1964 86 3024. 251 R. Colton R. Levitns and G. Wilkinson J . 1960 5275. 252 K. Nakamoto J. Fujita R. A. Condrate and Y. Morinioto J. Chem. Phys. 253 S. H. H. Chaston S. E. Livingstone T.N. Lockyer V. A. Pickles and J. S. 254 D. Barnard and D. T. Woodbridge J. 1961 2922. 266 C. K. Jmgensen Mol. Phys. 1962 5 485. 256 A. Rosenbaum H. Kirchberg and E. Leibnitz J. prakt. Chem. 1963 19 1. 257 A. I. Busev and Khoang Min Tyau Zhur. neorg. Khitn. 1962 7 88. 288 G. M. C. Higgins and B. Saville J. 1963 2812. 259 A. V. Ablov and V. N. Shafranskii Zhur. neorg. Khitn. 1961 6 1781. 260 F. A. Cotton and J. A. McCleverty Inorg. Chem. 1964 3 1398. 261 C. K. Jerrgensen Acta Chem. Scand. 1963 17 533. 1963 39 423. Shannon Austral. J. Chem. 1965,18. 673. LIVINGSTONE METAL COMPLEXES OF LIGANDS 419 occupy an early position (between C1- and F-) in the spectrochemical series and a late position in the nephelauxetic series the last-named ligand coming after I-.11J55J52 The monothio-ligand dialkylthiocarbamate (LXXVI) forms polymeric complexes with Cur Ag’ and A u ~ .~ ~ ~ Diethylthiophosphate (LXXVII) Et-0 ,S- Et - d P + O (LXXVI) (LXXVII) acts as a unidentate sulphur-ligand in the AgI and Hg*I complexes which are more stable than the thiocyanate but less stable than the thiourea complexes of these Monothiophosphate SP0,- can act as a uni- or bi-dentate ligand in ColI1 complexes.265 6.4. Thio-derivatives of /%Diketones and p-Keto-esters.-Tanaka and Yokoyama have prepared a large number of #?-mercaptoketones and p-mercaptoesters and investigated their co-ordinating ability. p-Mercapto- ketones (LXXVIII; when R’ = Ph R = Ph m-tolyl p-chlorophenyl ? R-CH-CH,-C-R’ Ph- CH-CH-5- -Ph 6 AH 0 I SH (Lxxv I I I) (LXX I x) R-FH - CH,- 5- OR’ Ph- -W2- $-SR SH 0 SH 0 (LXXX) (LXXX I) 1-naphthyl 2-thienyl 2-furyl; when R’ = 2-thienyl R = Ph 2-furyl) and or-C-substituted p-mercaptoketones (LXXIX; R = Me Ph) react with Cu2+ ion to give yellow Cu‘ complexes but do not yield Nil1 complexes.2ss The authors suggest that the ligand is chelated in these complexes but this seems unlikely.On the other hand p-mercaptohydrocinnamic acid esters (LXXX; R = Ph R’ = alkyl Ph) substituted #?-mercaptohydrocinnamic acid esters (LXXX; when R = p-chlorophenyl R’ = alkyl; when R = p-methoxyphenyl 1-naphthyl 2-furyl; R’ = Et) and S-esters of 18- mercaptotliohydrocinnamic acid (LXXXI; R = alkyl Ph benzyl) form 262 C. K. Jsrgensen Acfa Chem. Scand. 1942 16 2017. 263 S. Akerstroni Acfa Chem. Scand. 1963 17 1187. 264 V. F. Toropova M. K. Saikina and N. K. Lutskaya Zhur. neorg.Khim. 1961 265 J. Hidaka J. Fujita Y . Shiniura and R. Tsuchida Bull. Chem. Soc. Japan 1959 266 H. Tanaka and A. Yoltoyama Chem. Pharrn. Bull. (Tokyo) 1960 8 275 280 6 2086. 32 1317. 1008 1012. 420 QUARTERLY REVIEWS stable brick red 2 1 Nil1 complexes in addition to yellow CuI complexes.267 cc,p-Unsaturated p-mercapto-acid esters (LXXII; X = H NO2 C1 OMe) give stable NP CO~~I and FeIII complexes; the 3:l MI11 complexes are obtained from Co2+ and Fe2+ salts.268 X (&=CH-;-OEt SH 0 f I XXXlI\ 0 ; " R 0 0 ; S R 0 ( L x X X l I I) (LXXXI v) Esters of o-mercaptobenzoic acid (LXXXIII; R = Et Prf isopentyl) and S-esters of o-mercaptobenzoic acid (LXXXIV; R = Et Pri isopentyl) give orange-red Nil1 and green Coili complexes. The C=O stretching frequency occurs at 1700 cm.-l in the 0-esters (LXXXIII) and at 1610- 1620 cm.-l in their NiI1 and Co*I' complexes; it occurs at 1680 cm.-l in the S-esters and at 1575-1580 cm.-l in the NiI1 and CO'I* complexes but at 1680 cm.-l in the Cu* complexes.26g This indicates that the ligands are bidentate being bound through sulphur and oxygen atoms in the NiII and ColI1 complexes whereas the carbonyl group is not co-ordinated in the Cur complex.The extra stability of the Nil1 and ColI1 complexes of the ligands (LXXXII) (LXXXIII) and (LXXXIV) is apparently due to the existence in the chelate ring of a conjugated double-bond system which allows some electron delocalisation. M ~ O ~ ~ - C H - - ~ - N H 0 SH 0 (LXXXV) (LXXXVI) Nickel@) complexes are formed by p-mercapto-p-hydrocinnamaniside (LXXXV) and o-mercaptobenzanilide (LXXXVI).As the frequencies of the two infrared bands associated with the amide group do not alter on co-ordination it has been inferred that the ligands are (S-0) rather than (S-N) ~ h e l a t e d . ~ ~ ~ A number of monothio-/3-diketones (LXXXVII; where R = Me R' = Me,Ph,CF,,OEt; where R = Ph R' = Ph,OEt; where R = 2-thienyl R' = CF,; R = R' = CMe,) have been prepared. Evidence obtained from infrared NMR and mass spectrometry shows that these compounds exist almost entirely in the enethiol form; the absence of a sharp SH absorp- tion at ca. 2570 cm.-l in their spectra indicates that the thiol hydrogen is strongly chelated between the sulphur and oxygen atoms as in 267 H. Tanaka and A. Yokoyama Chem. Pharm. Bull. (Tokyo) 1961 9 66 110; 268 H. Tanaka and A. Yokoyama Chem. Pharm. Bull.(Tokyo) 1962,10,19. 269 H. Tanaka and A. Yokoyama Chem. Pharm. Bull. (Tokyo) 1962,10,25. 270 H. Tanaka and A. Yokoyama Chem. Pharm. Bull. (Tokyo) 1962,10,556 1962 10 13. LIVINGSTONE METAL COMPLEXES OF LIGANDS 421 (LXXXVIII).271,a53 All these ligands form strong diamagnetic square- planar Nil1 inner complexes which are readily soluble in organic solvents. It is significant that the replacement of one oxygen atom in /I-diketones by sulphur causes a change in bond-type of the Ni" complexes from high- spin to low-spin. The infrared spectra of the ligands and of their Nil1 chelates display a strong band in the range 1260-1190 cm.-l; this band alters little on co-ordination and has been assigned to the C = S stretch possibly coupled with the C -= 0 or C - C stretching modes.253 The lowest frequency band in the visible spectra of the NiI* chelates occurs in the range 675-620 mp (14.8-16.13 kK) ; accordingly these ligands fall between diethyldithiocarbamate and diethyldithiophosphate in the spectro- chemical series.272 Mass spectrometry has shown that the Ni-S bond is stronger than the Ni-0 bond in these complexes as the latter is always ruptured The Nil* complexes of the @-mercapto-esters (LXXXVIII; R = Me,Ph R' = OEt) form paramagnetic octahedral bis-pyridine adducts but such is the preference of the Ni atom to retain a square- planar configuration in the complexes of p-mercaptoketones (LXXXVII ; R' # OEt) that pyridine adducts are not formed although a very unstable yellow adduct of y-picoline was isolated.253 The /I-mercaptoketones form stable complexes with a number of class (b) and borderline metals; the green Fell1 complexes are h i g h - ~ p i n .~ ~ ~ NiII chelates were obtained with the a-methyl-substituted ligands (LXXXIX; R = Me Ph) but no PdlI or Pt" complexes could be obtained whereas the non-a-substituted /I-mercaptoketones (LXXXVIII) readily yield complexes with PdI1 and Pt11.275 Tt is (LXXX I x) but steric hindrance by the M-Me group ability of these ligands. difficult to explain these results apparently lowers the chelating 271 S. H. H. Chaston and S. E. Livingstone Proc. Chem. SOC. 1964 11 1 ; Proc. VIII International Conference on Co-ordination Chemistry Verlag Vienna 1964 p. 370. 272 S. H. H. Chaston and S. E. Livingstone unpublished work. 273 S. E. Livingstone and J. S . Shannon unpublished work. 274 S. H. H. Chaston R.K. Y. Ho S. E. Livingstone and T. N. Lockyer unpublished 275 R. K. Y . Ho S. E. Livingstone and T. N. Lockyer Austral. J. Chem. in the press. work. 422 QUARTERLY REVIEWS 2-Picolylphenylthioketone (XC) forms stable clielates with Co"' Ni" PdlI Cu" and 21111; these complexes resemble those of 3-mercapto-1,3- diphenylprop-2-en-l-one (LXXXVIII; R = R' = Ph).27s 6.5. a-Dithiols.-The dithio-oxalate ion forms anionic metal chelates [M(C,0,S2),l2- (M = Ni Pd Pt) and [M(C202S2)3]3- (M = Cr Co Rh) which are intensely coloured and very stable;277 the former are square- planar with the structure (XCI)278 and the latter have been resolved.279 Measurements of formation constants show that for Ni" dithio-oxalate > dithiomalonate > thiocarbonate CS3,- ; the Nil1 chelates of these ligands are much more stable than those of their oxygen analogues.280 2- Me/ SH a):; 0 SH [ 04 oys,~d,s c\ s ' JJ .p (XCI I) (XCI I I) (XCO Several a-dithiols of general formula HS-CH,(SH)CH,CH,Y form strong complexes with FelII Snl Pb" As"' and B P as well as with typically (b) class metals; the stabilities of the complexes of Zn" CdlI Hgl* and PblI with a number of or-dithiols have been compared.211,281 Ethanedithiol HSCH2CH2-SH (esH,) gives the very stable complexes [Ni(es),12- and [Ni2(es)3]2- and the complex formation between 2,3- dimercapto-l-propanol and Mn" Fell' NP and ZnI1 has been investi- gated.282 Quinoxaline-2,3-dithiol (XCII) has been used for the colorimetric estimation of cobalt and Complexes of toulene-3,4-dithiol (XCIII; TDTH,) with zinc and molybdenum have been reported.284 In 1962 Schrauzer and M a y ~ e g ~ ~ ~ reported the first example of a very interesting series of metal chelates formed by a-dithiols by heating nickel sulphide and diphenylacetylene together in toluene in a sealed tube they obtained the diamagnetic complex Ni(Ph2C2S2),.Shortly afterwards Gray et aZ.286 reported complexes of maleonitriledithiolate (MNT2-) of 278 Von E. Uhlemann G. Klose and H. Muller 2. Naturforsch. 1964 19b 962. 277 H. 0. Jones and H. S. Tasker J. 1909,95,1904; C . S. Robinson and H. 0. Jones 278 E. G . Cox W. Wardlaw and K. C. Webster J. 1935 1475. 270 F. P. Dwyer and A. M. Sargeson J. Amer. Chem. SOC. 1959 81 2335. 280 W. A. Deskin J. Amer. Chem. SOC. 1958 80 5680. 281 E. V. Vasil'eva and T. K. Nedopekin Tiolovye Soedinen v. Med. Ukrain. Nauch- Issledovatel.Sanit. Khim. Inst. Trudy Nauch. Konf. Kiev 1957 p. 36. 282 D. L. Leussing J. Amer. Chem. Sac. 1959 81 4208; D. L. Leussing and G. S. Alberts ibid. 1960 82 4458; D. L. Leussing and T. N. Tischer ibid. 1961 83 65; D. L. Leussing and J. P. Mislan J. Phys. Chern. 1960,64 1908. 283 R. W. Burke and J. H. Yoe Analyt. Clzim. 1962,34,1378; G. H. Ayres and R. R. Annand ibid. 1963 35 33. 284 K. Wallenfels and H. Sund Riocfrem. Z. 1957 329 17; T. W. Gilbert and E. B. Sandell J. Amer. Chem. SOC. 1960 82 1087. 285 G. N. Schrauzer and V. Mayweg J. Amer. Chem. Soc. 1962 84 3221. 286 H. B. Gray R. Williams I. Bernal and E. Billig J. Amer. Chent. SOC. 1962. 84 3596. 1912 101 62. LIVINGSTONE METAL COMPLEXES OF LIGANDS 423 general formula [NBun,],[M(MNT),] (M = Co Ni Pd Pt Cu Zn).Davison et recognised the relationship between the Ni complex of Schrauzer and the MNT complexes and they reported three series of complexes related by electron-transfer reactions and of general formula (XCIV; when R = Ph CF, M = Ni z = 0 -1 -2; when R = CN M = Co Ni Pd Pt Cu Au z = -1). Toluene-3,4-dithiol forms complexes [Ph,AsMe] [ M(TDT),] (M = Fe Co Ni Cu); the values of pLLeff(B.M.) are Fe 4.39; Co 3.27; Ni 1.89 Cu 0. The spectra were interpreted as indicating that these complexes are square-planar and contain the metal in the + 1 oxidation state; the authors suggested that the unpaired electron which each radicle anion would possess must be paired with its partner in one of the n-orbitals of the complex. Thus [Co(TDT),]- was regarded as the first example of a high- spin square ds complex.288 However the complex [Co(MNT),I2- was shown to be l o w - ~ p i n ~ ~ ~ although it had been previously reported as high-spin;28s [NBun,] [Co(MNT),] is diamagnetic in the solid state and in cyclohexa- none solution but has a moment of 2.81 B.M.in dimethyl sulphoxide.290 A structure determination of “Me,] [Ni(MNT),] shows that the nickel atom is square-planar with nearest Ni-Ni distance of 8.05 A and a C-C distance (1.30 A) indicating an order of 2.1 for the bond between the two carbon atoms linked to sulphur and CN.291 The complex [NBun4I2- [Rh(MNT),] (,uueff 1.9 1 B.M.) apparently contains square-planar Rh**.292 The four- five- and six-co-ordinate diamagnetic complexes “Bun,]- [Co(MNT),] [NBun,] [Co(MNT),L] (L = pyridine 1,lO-phenanthroline diarsine) [NBun,13 [Co(MNT),] formally contain COIII but certain charac- teristics of these complexes are not compatible with what is expected for Co1I1.These are (i) the existence of stable four- and five-co-ordinate species ; (ii) the ready interconvertibility among four- five- and six-co- ordination; (iii) the ligand-affinity order P > N (soft-acid behaviour) is 287 A. Davison N. Edelstein R. H. Holm and A. H. Maki J. Amer. Chem. SOC. e8s H. B. Gray and E. Billig J. Amer. Chem. Soc.. 1953 85 2019. 289 A. Davison N. Edelstein R. PI. Holm and A. H. Maki J. Amer. Chem. SOC. 2so E. Billig H. B. Gray S. T. Shupack J. H. Waters and R. Williams Proc. Chem. 2n1 R. Eisenberg J. A. Ibers R. J. H. Clark and H. B. Gray J. Amer. Chem. Soc. 2s2 E. Billig S. I. Shupack J. H. Waters R. Williams and H. B. Gray J.Amer. Chem. 1963,85 2029; Inorg. Chem. 1963 2 1227. 1963,85 3049. SOC. 1964 110. 1964 86 113. SOC. 1964 $6 926. 424 QUARTERLY REVIEWS consistent with d8 systems. Consequently the bond-pair is considered as more nearly localised on the metal atom than on the ligand~.,~~ The complexes [M(MNT),]- (M = Fe Ni Pd Pt) exhibit singlet-triplet magnetic behaviour which on the assumption that these complexes are square-planar and contain formally MI" is attributed to spin interactions between pairs of metal ions through the sulphur atoms of adjacent ligands. 94 Shupack et aZ.295 observed that one way to stabilise square-planar geometry is to involve the dxy dyz and p z valence orbitals of the metal in an extensive n-orbital network spanning the entire complex thus allowing as much delocalisation as possible.Two ligands with good donor atoms and considerably delocalised v-orbital systems are MNT2- and TDT2-. The following sq uare-planar complexes have been reported (1) S = 1 [Co(TDT),]- (2) S = 4 [Rh(MNT),I2- (3) S = 0 [Cu(TDT),]- and [Cu(MNT),]- (4) S = 4 [M(TDT),]- and [M(MNT),]- (M = Ni,Pd,Pt) (5) S = 4 and 3/2 [Fe(TDT),]- and [Fe(MNT),]- The complexes [Ni(S,C,MeR),] (R = Me Ph) are also known.296 Shupack et report that electron spin resonance (e.s.r.) measurements indicate that the unpaired electron in these square-planar complexes is in an orbital which is localised over metal and ligand but is predominantly ligand. They are of the opinion that a general molecular orbital formula- tion of the ground state is preferable to an assignment of a definite oxida- tion state to the metal.Although MNT2- produces a weak ligand field comparable with B r the ds complexes are unusually stable and are inert to substitution by most ligands; MNT2- also appears to have a strong trar~s-effect.~~~ AI! including the Cu complex remain truly square-planar in strongly co-ordinating solv- ents showing a lack of axial perturbation by solvent molecules. This unusual stability of the square-planar configuration is regarded as evidence of the involvement of the 4pz orbital in r-bonding giving added stability to the electronic delocalisation throughout the The six-co-ordinate complexes [M(S,C,R,),I2 (when R = CF, M = Cr Mo W z = 0 -1 -2; when R = CN M = Cr V z = -2 M = Cr z = -3) are similar to the bis-complexes in that they contain metals stabilised in several oxidation states and are interconvertible by oxidation- reduction reactions.The most obvious assignments of the three iso- 293 C. H. Langford E. Billig S. I. Shupack and H. B. Gray J. Amer. Chem. SOC. 1964.86. 2958. 294'J. F. Weicher L. R. Melby and R. E. Benson J. Amer. Chem. Soc. 1964 86 295 S. I. Shupack E. Billig R. J. H. Clark R. Williams and H. B. Gray J. Amer. 296 G. N. Schrauzer and V. Mayweg 2. Naturforsch. 1964,19b 192. 297 E. Billig R. Williams 1. Bernal J. H. Waters and H. B. Gray Inorg. Chetn. 4329 Chem. SOC. 1964 86 4594. 1964 3 663. LIVINGSTONE METAL COMPLEXES OF LIGANDS 425 electronic series M = Cr Mo W; R = CF,; z = 0 -1 -2) as do dl and d2 are not at all rationalised with data obtained from e . ~ . r . ~ ~ ~ Similarly the assignment of an oxidation state of +6 (i.e.3p5) for vanadium in [V(S2C2Ph2)3] seems unrea~onable.~~~ The complexes [M(S2C2PhJ,] (M = Cr Mo W) are also known.300 In the series [V(S2C2Ra3]z (R = CF3 z = -1 -2; R = Ph z = 0 -1 -2) when z = -1 the compounds are diamagnetic but when z = 0 -2 and for [V(TDT),I2- peff = 1.73- 1.90 B.M. Proton resonance studies show that in the paramagnetic com- pounds the electron is strongly localised on the ligands.,O1 Carbonyl compounds containing bridging a-dithiols-viz. [(CO),Fe (TDT)] and [(C,H5)2M2{S2C2(CF,),}2] (M = V Cr Mo)-have been reported by King137 who suggested the structure (XCV) for the monomeric CoIII complex. The existence of the monomeric complex [PhC :S(S)Mn (CO),] in which the dithiobenzoate ion is bidentate suggests the involve- ment of the metal in the aromatic system through r-bonding between the metal and sulphur The CO and NO stretching frequencies of [Ph,PMe] [Mn(CO),MNT] and [Ph4PI2 [Mo(NO),(MNT),] suggest that MNT2- is functioning as a r-acceptor from the metal which is in a low oxidation The author thanks Professors R.S. Nyholm and D. P. Craig and Drs. M. L. Tobe and M. A. Bennett for helpful advice and the Royal Society and Nuf€ield Foundation for the award of a Commonwealth Bursary during the tenure of which this Review was written. 298 A. Davison N. Edelstein R. H. Holm and A. H. Maki J. Amer. Chem. SOC. 1964,86 2799. 8g9 J. H. Waters R. Williams H. B. Gray G. N. Schrauzer and H. W. Finck J. Amer. Chem. SOC. 1964,86,4198. 300 G. N. Schrauzer H. W. Finck and V. Mayweg 2. Nuturforsch. 1964,19b 1080. A. Davison N. Edelstein R. H.Holm and A. H. Maki Inorg. Chem. 1965,4,55. I. A. Cohen and F. Basolo Inorg. Chem. 1964,3 1641. 803 J. Locke and J. A. McCleverty Chem. Comm. 1965,102. 4

ISSN:0009-2681    

DOI:10.1039/QR9651900386

出版商:RSC    

年代:1965

数据来源: RSC

摘要:

THE STUDY OF RING INVERSIONS BY NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY J. EDGAR ANDERSON (INSTITUT DE CHIME UNIVERSITB DE STRASBOURG) CONFORMATIONAL analysis of cyclohexane derivatives has only become fully established since 1945,l but one of the basic tenets the rapid inter- conversion among various conformations was recognised as possible by Sachsela*b in 1890 when he made the first suggestion of a non-planar conformation for cyclohexane. Recently the thermodynamic constants for such interconversions “ring inversions” in cyclohexane and other cyclic compounds have become available using nuclear magnetic resonance (n.m.r.) spectroscopy. It is the purpose of this Review after showing the basis and the scope of the n.m.r. technique to discuss the results so far obtained and the conclusions which can be drawn about conformational mobility in ring compounds.It is well to point out first that there are recent determinations of bar- riers to inversion by other techniques. Several examples of optical iso- merism due to the fixed conformation of a ring are known.2 In such com- pounds ring inversion causes racemisation so that the rate of loss of optical activity can be related to the rate of ring inversion. This approach is suitable where the barrier is somewhat more than 20 kcal./mole that is higher than can be measured by n.m.r. By use of ultrasonic techniques the barrier to ring inversion in methylcyclohexane has been determined.3 This method may prove valuable in future for it can be applied only to un- symmetrical systems whereas the n.m.r. technique is more readily applied to symmetrically substituted molecules.The Technique The way in which n.m.r. spectroscopy can be applied to the study of fast reactions is best illustrated by an example. The n.m.r. spectrum of the N-methyl groups of NN-dimethylacetamide is a doublet of chemical shift4 l Historical details and genera1 background information on conformational analysis can be found in (a) E. L. Eliel N. L. Allinger S. J. Angyal and G. A. Morrison “Conformational Analysis,” In terscience New York 1965; (b) M. Hanack “Confor- mation Theory,” Academic Press New York 1965. Background information on nuclear magnetic resonance spectroscopy can be found in (c) J. A. Pople W. G. Schneider and H. J. Bernstein “High Resolution Nuclear Magnetic Resonance,” McGraw-Hill New York 1959. (a) K. Mislow S.Hyden and H. Schaefer J. Amer. Chem. SOC. 1962 84 1449; (b) K. Mislow and M. A. W. Glass ibid. 1961,83,2780; (c) K. Mislow and H. B. Hopps zbrd. 1962 84 3018; (d) K. Mislow and H. D. Perlmutter ibid. p. 3591; (e) W. Toch- terman and H. Kuppers Angew. Chem. 1965,77 173. J. E. Piercy and S. V. Subrahmanyam J. Chem. Phys. 1965,42,4011. Throughout this Review “chemical shift” means the relative chemical shift of the signals implied by the context measured at 60 Mc./sec. operating frequency. 426 ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 427 4.8 cycleslsec. at room temperature.6 This arises since because of the mesomerism (I)++(II) and the resultant partial-double-bond character of the carbon-nitrogen bond there are two environments for an N-mefhyl group either cis or trans to the carbonyl function.As the temperature is raised the two N-methyl signals broaden and collapse to a single line which becomes sharp on further heating. Raising the temperature increases the rate of rotation about the carbon-nitrogen bond until this rate becomes so high that only an averaged signal is observed. A diagrammatic repre- sentation of the N-methyl signal as the temperature is raised is shown in Fig. 1. increasing temperature * A a b C d e FIG. 1. Diagrammatic representation of the N-methyl proton spectrum0 f NN-dimethyk acetamide at various temperatures. The technique can be extended to cyclohexane inversion in the following manner. Cyclohexane is acceptedlQSb to exist as a rapidly exchanging mixture of equal-energy chair forms. In the exchange each proton changes from an axial to an equatorial position or vice-versa.The process is rapid at room temperature so that a single line is observed for the n.m.r. spectrum of cyclohexane. Below about -65" two separate signals can be observed these being ascribed to the axial and equatorial protons in any molecule for ring inversion is now much slower. However in any frozen conformation of cyclohexane each proton will be coupled to other protons in the molecule so that the resultant low-temperature spectrum is very complex. We will now give a qualitative picture of the theory which governs the derivation of reaction parameters from n.m.r. spectra; a full description is readily available in the literature.6 Provided that the rate of sweeping through a signal is slow the shape observed for a particular signal is determined by certain time-dependent "NMR Spectra Catalog," Varian Associates Palo Alto 1963 vol.IJ spectrum ( a ) See ref. lc ch. 10; (b) L. H. Piette and W. A. Anderson J. Chem. Phys. 1959 421. 30 899. 428 QUARTERLY REVIEWS processes. Thus for a hydrogen nucleus there is a finite lifetime in either of the possible spin states,lU and the signal shape depends on the relaxation time for an exchange between these spin states. In practice where there is no exchange process of the type we are considering here the signal shape is determined not by this relaxation time but by the homogeneity of the magnetic field. We can nonetheless derive a relaxation time To from the observed shape this To corresponding to the homogeneity but not being precisely defined. Likewise for a molecule undergoing exchange we can obtain an experimental relaxation time T, from the shape of exchange- modified signals.If then we include in a sample a compound undergoing an exchange process and a reference compound we can measure T and To. We can then calculate6 the value of T the average lifetime of a molecule in either conformation (that is in either chair form in the case of cyclohexane) by using these relaxation times and the chemical shift when exchange is slow (Fig. la) of the signals whose coalescence is being studied. Since the ring inversion is a zero-order reaction the rate constant for the process is the inverse of the average lifetime T. If values of this rate constant are measured at several temperatures thermodynamic parameters for the inversion can be derived using the Eyring or Arrhenius equations.It is also possible to obtain values of relaxation times by the spin-echo technique.' Allerhand and Gutowskys have recently developed the applica- tion of this technique to the measurement of rate processes and have used it to study the inversion of two ring compounds though there are earlier pioneering examples of such applications. 9910 For the ring inversion of cyclohexane the high-resolution method and the spin-echo method have given values of the free energy of activation AG* which agree well,sc though there are differences in the enthalpies and entropies of activation (AH* and AS*) found. It has however been pointed outsc that systematic errors affect AS* and AH* much more markedly than AG*. Allerhand and Gutowskysassc have given a critical comparison of the two techniques.They point out that the temperature range over which high- resolution technique can be applied is limited to that in which the contribu- tion to relaxation from the exchange process is larger than the n.m.r. spectrometer's error in measuring the reference relaxation time. This almost invariably limits measurements to a temperature range of about 30". In effect the spin-echo technique allows a much more accurate determina- tion of relaxation times so that there is a much wider range over which the * (a) A. Allerhand and H. S. Gutowsky J. Chem. Phys. 1964,41 2115; (b) A. Aller- hand and H. S. Gutowsky ibid. 1965,42,1587; (c) A. Allerhand F.-M. Chen and H. S. Gutowsky ibid. p. 3040; ( d ) A. Allerhand and H. S. Gutowsky ibid. p. 4203. L. W. Reeves and E.J. Wells Discuss. Faraday Soc. 1962 34 177; see also M. Bloom L. W. Reeves and E. J. Wells J. Chem. Phys. 1965,42 1615. la S. Meiboom Paper presented at symposium on High Resolution Nuclear Magnetic Resonance Boulder Colorado July 1962. See ref. Ic section 3.7. ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 429 exchange contribution to the experimental relaxation time is measurable. Thus Allerhand Chen and Gutowsky can reasonably claimBc that the spin-echo results are ‘probably more reliable than others’. At the coalescence temperature (Fig. le) the rate constant for inversion is independent of the relaxation time To provided that To is fairly large. so that knowing the chemical shift of the two signals coalescing one can obtain a value of the rate constant for inversion.From this knowing the temperature one can calculate a value of the free energy of activation dG* using the Eyring equation. Since AH* and AS* values even from the most careful analysis may be subject to errors it might seem at first sight that there is little to be gained by a full analysis. However comparisons of AG* should be made at the same temperature and since errors in AH* and AS* are ~elf-compensating,~~ dG* values obtained by extra- polation or intrapolation of results from a full analysis allow this compari- son. A large proportion of reports give barriers derived from temperatures of coalescence especially when there is only a passing interest in actual values or where values are difficult to obtain. When systematic examination of a series is being made a full analysis despite its certain limitations is obviously necessary.Throughout this Review when the “barrier to ring inversion” is men- tioned it is the free energy of activation for the ring inversion process that is meant unless otherwise indicated. Technical Details andIDifficulties There are two principal difficulties which hamper systematic investiga- tions of ring inversions. These are complex spectra and low barriers to inversion. It has been calculated1’ that for cyclohexane itself below the temperature of coalescence the spectrum should comprise many hundreds of lines. Thus it is not correct to apply the theory derived for the collapse of a doublet even though the lines observed group themselves into two broad bands.12 In the case of cyclohexane the difficulty has been overcome by examining undecadeuterocyclohexane with irradiation at the deuterium frequency to remove hydrogen-deuterium c ~ u p l i n g .~ ~ ~ ~ In the majority of cases however deuterated compounds are not readily available and proton decoupling of the protio-compound is impracticable. Another difficulty occurs when examining monosubstituted cyclic compounds. The substituent has an energy preference for the equatorial l1 J. I. Musher J. Chem. Phys. 35 1159. l2 F. R. Jensen D. S. Noyce C. H. Sederholm and A. J. Berlin J. Amer. Chem. SOC. l3 F. A. L. Anet M. Ahmad and L. D. Hall Proc. Chem. Suc. 1964,145. See correc- l4 F. A. Bovey F. P. Hood 111 E. W. Anderson and R. L. Kornegay Proc. Chem. 1960,82 1256; 1962,84,386. See also comment thereon in reference 8c. tion in a footnote in reference 16a.SOC. 1964,146; J. Chem. Phys. 1964,41,2041. 430 QUARTERLY REVIEWS position,lasb so that even with other factors favourable the more complex treatment15 for the collapse of an unsymmetrical doublet ought to be used. The second great difficulty is that of low barriers to inversion. It is theoretically possible to slow down inversion to obtain splitting of signals by cooling to a sufficiently low temperature; the lower the barrier to in- version the lower the temperature required. With reference to cyclohexyl compounds again few have barriers significantly higher than that of cyclo- hexane in which the coalescence temperature is about -65',13J4 so that temperatures much lower than this must be accessible. The lowest tem- perature at which high resolution proton spectra have been recorded is about -170°.16,17 This places a lower limit of about 5 kcal./mole on barriers which can be measured at present.In addition there are difficulties at such low temperatures. The reference signal becomes broad so that in some examples e.g. trioxan,l8 where splitting is not observed at the lowest attainable temperatures this may be due to a small chemical shift obscured by poor homogeneity rather than to a low barrier. At these low temperatures the choice of solvent may prove to be another difficulty. In general mixed solvents have not proved very satisfactory. For temperatures below - 1 20° certain tetrahalogenomethanes with two different halogen atoms e.g. CCI,F or vinyl chloride have been found most ~ ~ e f ~ l . ~ ~ ~ ~ ~ ~ ~ ~ In general a suitable solvent should have a low freez- ing point good solvent properties at low temperatures and a relatively simple proton n.m.r.spectrum. It is noteworthy that there is little information on the effects of solvents on barriers to ring inversion. In other kinetic processes studied by n.m.r. the solvent has been shown to have a large effect,20 particularly when the molecule studied contains polar functions. Analogous effects would be expected in polar ring compounds and in 3,3,6,6-tetramethyl- 1 ,Zdioxan a range of 3 kcal./mole in the barrier to inversion in various solvents has been found.21 It is unfortunate that the best solvents may owe this quality to their polar nature and that this may influence the results found. Results Cyclohexane and its Derivatives.-The inversion of six-membered rings is of particular interest to the organic chemist and the inversion of the 15 F.A. Bovey E. W. Anderson F. P. Hood and R. L. Kornegay J. Chem. Phys. l6 ( a ) F. A. L. Anet and M. Z . Haq J. Amer. Chem. SOC. 1965 87 3147; (b) F. R. l7 F. R. Jensen and L. A. Smith J. Amer. Chem. SOC. 1964,86,956. la J. E. Anderson and J. C. D. Brand Trans. Faraday SOC. in the press. lD F. A. L. Anet and J. S. Hartman J. Amer. Chem. Soc. 1963,85 1204. 2o See for example A. Loewenstein J. F. Neumer and J. D. Roberts J. Arner. 21 H. Friebolin S. Kabuss and H. Schmid personal communication. 1964,40 3099. Jensen and C. H. Bushweller ibid. p. 3285. Chem. SOC. 1960 82 3599. ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 43 1 parent cyclohexane has been much ~ t ~ d i e d . ~ ~ ~ ~ ~ ~ ~ ~ ~ In the period 1960-1962 the first five papers a p p e a ~ e d ~ ~ ~ ~ ~ ~ ~ each describing in greater or lesser detail the broadening and resolution into two parts of the single n.m.r.absorption line of cyclohexane on cooling to about +100". By virtue of the complexity of the spectrum mentioned above the derived equations were not strictly applicable and only approximate values of parameters could be determined. In 1962 MeiboomlO reported values of these parameters derived from an early application of the spin-echo technique. In the meantime Hendrick~on~~ had made calculations on the energies of the possible conformations of cyclohexane and had presented a plausible picture of the path of ring inversion through the structures shown in Fig. 2. The plausibility of the pathway is enhanced by the reason- able agreement between his calculated values for activation enthalpies FIG.2. Diagram of the path suggested by Hendrickson for the inversion of six-membered rings.2s and enthalpy differences and those found by e ~ p e r i m e n t . ~ ~ J ~ J ~ ~ * ~ ~ The difficulties arising from the complex low-temperature spectrum were removed in work by Anet,13 Bovey,14 and their co-workers published in 1964. They studied [2Hll]cyclohexane with deuterium decoupling to give a single- and double-line spectrum at high and low temperatures respectively. At the same time Allerhand and Gutowsky were developing the spin-echo and in early 1965 they reported a study of both cyclohexane and [2Hll]cyclohexane.8c From these three studies it was clear that the free energy of activation for ring inversion is close to 10.3 kcal./mole.Values of the enthalpy of activation range from 9.1 to 10.5 kcal./mole the spin-echo technique giving the lower value. The corresponding range of values for the entropy of activation is -5.8 to + 1.4 e.u. Calculations based on symmetry considerations alone suggest an entropy of activation of +4-9 e.u:12 this is outside the range of experi- mental values and thus there are additional factors involved. There are no reported values of barriers to inversion in alkylcyclo- hexanes. This may be due to a combination of low barriers and therefore low coalescence temperatures and very unequal populations of the two R. K. Harris and N. Sheppard Proc. Chem. Soc. 1961,418. *3 W. B. Moniz and J. A. Dixon J. Amer. Chem. SOC. 1961 83 1671. 24 N. Muller and W. C.Tosch J. Chem. P h y ~ . 1962,37 11 67. s6 J. B. Hendrickson J. Amer. Chem. SOC. 1961,83,4537. 26 N. L. Allinger and L. A. Freiberg J. Amer. Chem. Soc. 1960 82 2393. 27 W. S. Johnson V. J. Bauer J. L. Margrave M. A. Frisch L. H. Dreger and W. N. Hubbard J. Amer. Chem. Soc. 1961 83 606. 432 QUARTEWY WWEWS chair conformations in the monosubstituted cases. 1,l -Dimethylcyclo- hexane would seem to be very suitable for study but Harris and Shep- pard22 found that at -120" the methyl signal is still a singlet. Muller and T o s c ~ ~ ~ confirmed this but found that the complex signal of the ring protons broadens on cooling to -75" then sharpens on further cooling which suggests that the conformational exchange is slow at such low temperatures. The failure to resolve the methyl signal may be due to a very small chemical shift as Harris and Sheppard have suggested,22 and as is supported by unpublished results of the Freiburg group.21 For cis-1,2- cis- 1,4- and trans-1 ,3-dimethylcyclohexane Muller and T o ~ c h ~ ~ found that exchange is slow at -120".In all cases the maximum broadening was observed about -70" which suggests that generally in the simple alkyl cyclohexanes the barriers to inversion are similar to those in cyclohexane itself. Cyclohexyl halides have received considerable attention. Bovey Hood Anderson and Kornegay15 have fully studied cyclohexyl fluoride using the f9F-~pectrum with spin decoupling of adjacent protons. With equations modified for the unequal numbers of fluorine atoms in axial and equatorial positions they found the barriers to inversion slightly lower than in cyclo- hexane.For the other halides28up20 and for cyclohe~anol~~ barriers have been obtained from temperatures of coalescence and appear to be slightly larger than for cyclohexane but the difference is probably not significant. Methoxy- and ethoxy-cyclohexane are reported to invert rapidly even at -1 10°.28b the dimethylcyclohexanes have been discussed above. For both cis-1 ,2-28b and trans-1,3-dihalocyclohexanes30 the barrier is close to that of the parent compound. JonaS Allerhand and Gutowsky31 have examined the inversion of 1,l-difluorocyclohexane both by the high resolution and spin-echo technique^.^^,^^ They found the barrier to be less than that in cyclo- hexane by 0.7 kcal./mole. Several polysubstituted cyclohexanes have been ~ t u d i e d .~ ~ - ~ ~ Here the most interesting results are those of Tiers33 on the ring inversion of per- fluorocyclohexane followed from the 19F-spectrum. Because of the large chemical shift involved rate measurements could be made over a much wider temperature range 93". The free energy of activation is slightly smaller thar in cyclohexane and since the entropy of activation is high Of the disubstituted cases that have been 93 (a) L. W. Reeves and K. 0. Stromme Canad. J. Chem. 1960,38 1241 ; (b) L. W. l9 W. C. Neikam and B. P. Dailey J. Chem. Phyv. 1963 38 445. so H. M. van Dort and T. J. Sekuur Tetrahedron Letters 1963 1301. 31 K. Nagarajan and J. D. Roberts unpublished results quoted in J. D. Roberts 32 J. JonBS A. Allerhand and H. S. Gutowsky J. Chem. Phys. 1965,42 3396. 33 G. V. D. Tiers Proc.Chem. SOC. 1960 389. a4 S. Brownstein Canad. J. Chem. 1962 40 870. as R. K. Harris and N. Sheppard Mof. Phys. 1964 7 595. Reeves and K. 0. Stramme Tram. Faraday SOC. 1961,57,390. Angew. Chem. 1963. 75 20. ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 433 and negative -10.7 e.u. the enthalpy of activation is much smaller than in cyclohexane. Tiers has attributed this to fluorine-fluorine interactions' raising the energy of the ground-state chair conformation. BrownsteinS4 has examined the acetates of three configurational isomers of inositol (1,2,3,4,5,6-hexahydroxycyclohexane) and finds the enthalpies of activa- tion are low for these compounds also probably for the same reasons as in perfluorocyclo hexane. There are two other obvious systems which fall into this section and whose barriers to inversion are of interest cyclohexaiiones and cvclo- hexenes.There is no report of a barrier to inversion of cyclohexanone but [2H,]cyclohexanone appears to be inverting rapidly even at - 140",36 which would place an upper limit of about 7.7 kcal./mole on the barrier to inversion. Anet and Haq have recently studied the ring inversion of cyclohexene using the compound (111) with deuterium decoupling.16" The coalescence temperature is as low as -164" and the barrier is therefore about 5.3 kcal./mole. They discuss the path the inversion may take and suggest a conformation with five carbon atoms in one plane as the most likely transition state. The barrier to inversion of 4-bromocyclohexene is slightly larger than for the parent compound,16b but the derivative (IV) a tetrahydropyridazine presents a marked contrast.has found the barrier to inversion to be about 19 kcal./mole probably owing to the multiple interactions of bulky substituents in a near-planar transition state. Thus the information that an n.m.r. study of ring inversions gives about the conformations of simple cyclohexanes is limited. One useful conclusion is that in all cases ring inversion is rapid at ordinary temperatures. In cyclohexane for example the average time spent in any conformation is less than lo-* sec. at room temperature. Experimental results bear out Hendrickson's picture of the path of inversion25 quite satisfactorily. The majority of substituted cyclohexanes studied however have barriers close to that of the parent compound and perhaps not significantly different from it.Further examples will no doubt be investigated but there are factors which make it doubtful whether a systematic study will be possible. If for example one considers unsymmetrical cyclohexane derivatives then that conformation with more "large" groups equatorial will predomin- ate. A free-energy difference of only 1.5 kcal./mole between these confor- mations is sufficient for there to be more than 95% of the more stable ae J. E. Anderson D. D. MacNicol and B. D. Batts unpublished results ' 1 J. C. Breliere and J. M. Lehn Chem. Co- 1965,426. 434 QUARTERLY REVIEWS isomer at low temperates so that even if inversion is slow the signals due to the less stable isomer will not be readily observed. One could hope to obtain a molecule suitable for study by ensuring that in both possible conformations there is an approximately equal “weight” of axial and equatorial substituents.This is the position obtaining in perfluorocyclohexane and in the inositol acetates mentioned above but if the substituents involved are much “larger” than fluorine or the acetate group then 1,3-diaxial interactions will be of such magnitude as either to lower the barrier towards and perhaps below the present limit of measur- ability or to make flexible boat conformations energetically probable. The latter is the case in trans- 1,3-di-t-butylcy~lohexane~~ and hexamethyl- cy~lohexane-l,3,5-trione~~ which might otherwise be suitable for study by the n.m.r. method. Six-membered Heterocyclic Rings.-The inversion of six-membered rings containing one or more hetero-atoms has given more detailed information on the factors which affect barriers to inversion.Reeves and his co-workers have reported free energy values for the barrier to inversion in one pipera~ine~~ and one piperidineg derivative. From two studies of lY2-dioxans and 1,2-dithian~,~O-~~ it is clear that the replacement of two adjacent methylene groups by oxygen or sulphur atoms raises the barrier to inversion. There has been a full study of the 1,3-dioxans (V; A = B = oxy- gen),18v21s42s43 1,3-dithians (V; A = B = s u l p h ~ r ) ~ ~ ~ ~ ~ and lY3-thioxans (V; A = sulphur B = o ~ y g e n ) . ~ ~ ~ ~ The parent compounds in these series have barriers similar to that in cyclohexane. In the 1,3-dioxan series r e s ~ l t s ~ ~ ~ ~ show that introduction of two methyl groups into the 2- position lowers the barrier to inversion.The 1,3-dioxan ring is smaller than the cyclohexane ring since carbon-oxygen bonds are shorter than carbon-carbon bonds so that the axial methyl group has strong 1,3- interactions with two hydrogen atoms and the energy of the ground state is raised i.e. the barrier to inversion is lowered. The 5,5-dimethyl deriva- a* J. Date J. 1965 1028. 3* L. W. Reeves and K. 0. Strsmme J. Chem. Phys. 1961,34,1711. 4 O G. Claeson G. M. Androes and M. Calvin J. Amer. Chem. Soc. 1960 82 4428; 41 A. Liittringhaus S. Kabuss W. Maier and H. Friebolin 2. Naturforsch. 1961 42 H. Schmid H. Friebolin S. Kabuss and R. Mecke Spectrochim. Acta in the press. H. Friebolin S. Kabuss W. Maier and A. Liittringhaus Tetrahedron Letters 1961 83,4357. 16b 761 ; H. Friebolin and W. Maier ibid.1961,1& 1640. 1962 683. ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 435 tive in which the axial methyl group now interacts with two much smaller oxygen lone pairs has a barrier similar to that of the parent compound while the 2,2,5,5-tetramethyl compound has a barrier of intermediate value. The effects of substituents are less clearly marked in the dithian and thioxan s e ~ i e s ~ ~ ~ ~ in which steric compression is less because of long carbon-sulphur bonds. Lastly there is a considerable body of results on the inversion of annellated six-membered rings. It is difficult to draw any general con- clusions from these studies for no systematic investigations have been carried out. cis- and t r a n ~ - D e c a l i n ~ ~ ~ ~ ~ and derivative^,^^ and cis- and tr~ns-perhydroindane~~ are saturated systems that have been studied.9,lO-Dihydrophenanthrene has been shown to have a low barrier to inversion.45 Curtin Carlson and McCarty have studied the series (VI) in detailY4* determining dG* AH* AS* and finding agreement between the barriers to inversion in this series and the barriers to rotation in the series (VII) which as they point out is analogous. Rings Smaller than Six-membered.-There has been no n.m.r. study of the inversion of rings smaller than six-membered. In cyclopropane the protons are identical. In the favoured conformations of cyclobutane and cyclopentanelaSb there are different kinds of protons but the barriers to interconversion of conformations are much smaller than can be measured by n.m.r. Seven-membered Rings.-There is no report of a study of cycloheptane for which Hendri~kson~~ has calculated that there are two stable conforma- tions the twist-chair (VIII) and the slightly less-favoured twist-boat (IX).Both forms are flexib,, and in contrast with cyclohexane the inter- conversions of “axial” and “equatorial” positions in the twist-chair can be achieved by pseud~rotation,~~ which can be defined as a change in conformation of a ring compound involving only a small barrier a barrier which arises from non-bonded interactions rather than bond-angle strain i.e. pseudorotation is the flexing of a flexible system. The barrier involved in pseudorotation will normally be smaller than can be measured by n.m.r. 44 F. G. Riddell and M. J. T. Robinson Chem. Cumm. 1965,227. 46 M . Oki H. Iwamura and N. Hayakawa Bull.Chem. SUC. Japan 1963,36 1542; D. Y. Curtin C. G. Carlson and C. G. McCarty Cunad. J. Chem. 1964,42,565. 47 J. E. Kilpatrick J. S. Piker and R. Spitax J. Amer. Chem. Suc. 1947 69 2483. 1964,37 1865. 436 QUARTERLY REVIEWS In further contrast to cyclohexane the introduction of a double bond into cycloheptane produces a rigid conformation (X) from a flexible one (VIII). It has been shown however that the rigid conformation (X) and the more flexible tub-form (XI) are of approximately equal energies in some cases e.g. (XIT),48a and n.m.r. signals corresponding to the two conformations have been obtained at low temperatures. Grunwald and Price4sb have reported another example and have given a full discussion (XI I I) of the conformational interconversion processes involved. Cycloheptene itself has not been studied but a series of benzo-derivatives have barriers about 11 kcal./m01e.~~ When two of the skeletal carbon atoms are re- placed by sulphur atoms the barriers have similar values but with two oxygen atoms the barrier is reduced.48 The preferred conformation is again changed on the introduction of two double bonds into the ring,45s50ss1 usually as dibenzo-derivatives to give bridged biphenyls whose lowest-energy conformation appears to be a pseudo-boat (XIII).50 In the transition state of the inversion all the atoms except X appear to be planar.2a Likewise in the dihydropleiadenes (XIV)52 the transition state is taken to be a rather strained planar form and this explains the relatively high barriers up to 16 k ~ a l ./ m o l e ~ ~ ~ ~ ~ found in both series.Bridged biphenyls with even higher barriers have been resolved into optical isomers,2a and studies of their rates of racemisation have given their rates of inversion. Cycloheptatriene is interesting. The preferred conformation is boat- shaped,63 and interconversions of the two possible forms (XV) and (XVI) is rapid at room temperature. The spectrum of the methylene group splits only at -143O,l'~~~ and the barrier to inversion is thus about 6 (w (xv) 4 7 (XVI) 48 (a) H. Friebolin R. Mecke S. Kabuss and A. Lutteringhaus Tetrahedron Letters 1964,1929; (6) E. Grunwald and E. Price. J. Amer. Gem. Soc. 1965,87,3139. 4B S. Kabuss H. Friebolin and H. Schmid Tetrahedron Letters 1965,469. 61 W . Tochterman U. Walter and A. Mannschreck Tetrahedron Letters 1964. 63 P. T. Lansbury and J.F. Bieron J. Amer. Chem. Suc. 1964,86,2524. 63 M. Traetterberg J. Amer. Chem. Soc. 1964 86,4265. 64 F. A. L. Anet J. Amer. Chem. Soc. 1964,86,458. R. J. Kurland M. B. Rubin and W. B. Wise J. Chem. Phys. I964.40,2426. 298 1. ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 437 kcal./mole. An increase in the barrier on successive annellation of benzene rings is observed here also51 and this has been attributed to the interaction of the orrko-hydrogen atoms of these rings in a planar transition state. In this series too optically active isomers have been obtained,2e the asym- metry being due to hindered ring inversion. Eight-membered Rings.-It is generally accepted that an extended crown (XVII) is the most stable conformation of cyclo-octane.laVb This form is flexible but in achieving a full cycle of pseudorotations it can be seen from models that fairly drastic transannular interactions are en- countered so that there may be a relatively high barrier to ring inversion.(XVI I) (XXIII) Following earlier work by Harris and Sheppard22 and by Meiboom,lO this barrier was measured as 7-7 kcal./mole by Anet and Hartman,ls using quindecadeuterocyclo-octane with deuterium decoupling. [2H,5]Cyclo-octatetraene is known to exist in the tub form (XVIJI).56 There are two distinct exchange processes that one might hope to study by ~ ~ . m . r . ~ ~ bond shift (Fig. 3 path A) and ring inversion (Fig. 3 path B). In both processes the most reasonable transition state appears to be (XX planar) (XXI I) Bond shift Ring % version FIG. 3. The two exchange processes taking place in cyclo-octatetraene.planar. Direct examination of reactions A and B is not possible since the protons in (XVIII) are equivalent. AnetS7 and Whitesides Gwynn and observed the 13C-H satelliteslC of the cyclo-octatetraene spec- trum. Statistically one hydrogen atom in about 8% of the molecules is distinguished by being attached to a carbon-13 atom. When bond exchange 66 Ref. 26 gives leading references. F. A. L. Anet A. J. R. Bourn and Y. S. Lin J. Amer. Chern. SOC. 1964 86 3576. 67 F. A. L. Anet J. Amer. Chem. SOC. 1962,84,671. s8 ( a ) G. M. Whitesides D. E. Gwynn and J. D. Roberts unpublished results quoted in J. D. Roberts Angew. Chern. 1963,75,20; (b) D. E. Gwynn G. M. Whitesides and J. D. Roberts J. Amer. Chem. Soc. 1%5,87,2862. 438 QUARTERLY REVIEWS is slow this hydrogen atom has two different neighbours one separated from it by a single bond and coupled to it by about 12 c./sec.the other separated from it by a double bond and only weakly coupled to it. Thus at low temperatures the spectrum of this proton is a doublet which collapses at higher temperatures as the rate of bond exchange increases. AnetS7 found that the barrier to this bond exchange is 13.7 kcal./mole. Gwynn Whitesides and have recently reported that for the combined process of inversion and bond exchange in monofluorocyclo-octatetraene the activation energy is somewhat smaller. Ring inversion and bond shift have been studied in the compound (XXIII) the changes in the methyl proton and ring proton spectra being observed; the two processes can be differentiated56 and their rates found simultaneously.Thus it was shown that the planar conformation (XX) is more stable than (XIX) by 2-4 kcal./mole which is in the sense pre- dicted by theoretical calculation^.^^ The actual barrier to ring inversion for (XXIII) is 14-7 kcal./mole. In two cases2ds59 cyclo-oc tatetraene derivatives with two benzene rings annellated have been resolved into optical isomers so that increased barriers to ring inversion due to annellation are observed in this series also. cis-7,8-Dibromo-cis,cis,cis-cyclo-octa- 1,3,5-triene has been found to have a barrier to inversion of 13 kcal./mole,60 it being considered that the pre- ferred conformation is still a tub. Nine-membered and Larger Rings.-Hendrickson has calculateds1 that the most stable conformation of cyclononane is the symmetrical flexible form (XXIV) though X-ray evidence62 for the hydrobromide of cyclo- nonylamine indicates a slightly different conformation.Examination of 47.57 (xx IV) models shows that pseudorotation renders “axial” and “equatorial” positions equivalent so that barriers to inversion are probably too small to be measured by n.m.r. The only nine-membered ring whose inversion has been studied by n.m.r. is cis,cis,cis-cyclonona-1 ,4,7-trienesM5 for which the conformation N. L. Allinger W. Szkrybalo and M. A. DaRooge J. Org. Chem. 1963,28,3007. 6o R. Huisgen and G. Bosche Tetrahedron Letters 1965 1769. 61 J. B. Hendrickson J. Amer. Chem. Soc. 1964 86 4854. 62 R. F. Bryan and J. D. Dunitz Helv. Chim. Acfa 1960 43 1 g3 P. Radlick and S . Winstein J. Amer. Chem. Soc. 1963,85 344. 64 K. G. Untch and R.J. Kurland J. Amer. Chem. SOC. 1963 85 346; ibid. 1964 M W. R. Roth Annulen 1964,671 10. 86 5709. ANDERSON RING INVERSIONS BY N.M.R. SPECTROSCOPY 439 (XXV) appears to be the most At about -4” there are two dis- tinct peaks for the allylic protons but on heating to about 31” these coalesce. From this the barrier to ring inversion is 14.5 k~al./mole.~*~~~ Untch and Kurlande4 find a high negative entropy of activation for the inversion and attribute this to a strained planar transition state. A tri- benzo-derivative of cyclononatriene is known and as expected is found to invert slowly even at 180”.66 The only examples of rings larger than nine-membered which have been studied are the eleven-membered-ring compound (XXVI)67 and two polyenes.68a In (XXVI) the barrier to inversion is 19 kcal./m~le,~~ so this is yet another benzo-compound with a relatively high barrier.At low temperatures the A-isomereSb of [ 14lannulene (XXVII) shows (XXVI) H H (XXVII) two separate peaks corresponding to protons directed inwards and outwards.68a At higher temperatures only one line is observed which indicates that protons are exchanging positions relatively rapidly. Similar results are observed for the analogous [18]annulene. From the chemical shifts of the “outer” and “inner” protons in the low-temperature spectra there is an appreciable ring current in both molecules so that even though the rings are inverting they are concluded to be aromatic.68a The author acknowledges financial support from the Salters’ Company. 66 A. S. Lindsey J. 1965; 1685. 67 R. W. Griffin and R. A. Coburn Tetrahedron Letters 1965,2571. 68 (a) Y . Gaoni A. Melera F. Sondheimer and R. Wolovsky Proc. Chem. Soc. 1964 397; (b) Y . Gaoni and F. Sondheimer ibid. 1964,299.

ISSN:0009-2681    

DOI:10.1039/QR9651900426

出版商:RSC    

年代:1965

数据来源: RSC