摘要:

Chemical Society Reviews Vol 13 No 4 1984 Page LIVERSIDGE LECTURE Molecular Tectonics: The Construction of Polyhedral Clusters By Norman N. Greenwood 353 Chemical Processes on Heterogeneous Catalysts By C. Kembali 375 Radical Cations in Condensed Phases By Martyn C. R. Symons 393 Aromatic Benzene Compounds from Acyclic Precursors By P. Bamfield and P. F. Gordon 441 1984 Indexes 489 The Royal Society of ChemistryLondon Chemical Society Reviews EDITORIAL BOARD Professor K. W. Bagnall, B.Sc., Ph.D., D.Sc., C.Chem., F.R.S.C. Professor B. T. Golding, B.Sc., M.Sc., Ph.D., C.Chem., F.R.S.C. Professor G. Pattenden, Ph.D., C.Chem., F.R.S.C. Professor P. A. H. Wyatt, B.Sc., Ph.D., C.Chem., F.R.S.C. (Chairman) Dr.D. A. Young, Ph.D., D.Sc., F.R.C.S., M.Inst. P. Editor: K. J. Wilkinson, B.Sc., M.Phi1. Chemical Society Reviews appears quarterly and comprises approximately 20 articles (ca. 500 pp) per annum. It is intended that each review article shall be of interest to chemists in general, and not merely to those with a specialist interest in the subject under review. The articles range over the whole of chemistry and its interfaces with other disciplines. Although the majority of articles are intended to be specially commissioned, the Society is always prepared to consider offers of articles for publication. In such cases a short synopsis, rather than the completed article, should be submitted to the Managing Editor, Books and Reviews Section, The Royal Society of Chemistry, Burlington House, Piccadilly, London, W 1V OBN.Members of the Royal Society of Chemistry may subscribe to Chemical Society Reoiews at f15.00 per annum; they should place their orders on the Annual Subscription renewal forms in the usual way. All other orders accompanied with payment should be sent directly to The Royal Society of Chemistry, The Distribution Centre, Blackhorse Road, Letch- worth, Herts. SG6 1HN England. 1984 annual subscription rate U.K. f43.50, Rest of World f45.50, U.S.A. $87.00. Air freight and mailing in the U.S.A. by Publica- tions Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. U.S.A. Postmaster: Send address changes to Chemical Society Reviews, Publications Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. Second class postage is paid at Jamaica, New York 11431. All other despatches outside the U.K. by Bulk Airmail within Europe, Accelerated Surface Post outside Europe. 0Copyright reserved by The Royal Society of Chemistry 1984 ISSN 0306-0012 Published by The Royal Society of Chemistry, Burlington House, London, W1V OBN Printed in England by Richard Clay (The Chaucer Press) Ltd, Bungay, Suffolk

ISSN:0306-0012    

DOI:10.1039/CS98413FP011

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS98413FX013

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

Chemical Society Reviews Vol 13 No 4 1984 Page LIVERSIDGE LECTURE Molecular Tectonics: The Construction of Polyhedral Clusters By Norman N. Greenwood 353 Chemical Processes on Heterogeneous Catalysts By C. Kembali 375 Radical Cations in Condensed Phases By Martyn C. R. Symons 393 Aromatic Benzene Compounds from Acyclic Precursors By P. Bamfield and P. F. Gordon 441 1984 Indexes 489 The Royal Society of ChemistryLondon

ISSN:0306-0012    

DOI:10.1039/CS98413BX015

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

LIVERSIDGE LECTURE* Molecular Tectonics: The Construction of Polyhedral Clusters By Norman N. Greenwood DEPARTMENT OF INORGANIC AND STRUCTURAL CHEMISTRY, UNIVERSITY OF LEEDS, LEEDS, LS2 9JT 1 Introduction The art of synthesizing new compounds and the determination of their structures and properties has been the central theme of chemistry for over 150 years. In more recent times the design of new atomic groupings and the purposeful synthesis of target molecules has become increasingly possible as a result of our growing knowledge and understanding of the underlying principles. I have introduced the term ‘molecular tectonics’ to describe this aspect of our subject since it encapsulates the dominant characteristics of this approach to chemistry. The word tectonics derives from the Greek wcrovitcos and refers to the design and constructim of buildings or other structures.In the form of plate tectonics the word has been used during the past few decades by geophysicists interested in continental drift and the structure of the Earth‘s crust. However, I believe the time is ripe for chemists to take over the term because, of all the sciences, chemistry is pre-eminently concerned with molecular architecture. In this lecture I shall adopt a tectonic approach, as an architect and builder, in describing the design and construction of polyhedral clusters. This presupposes the development of pragmatic bonding theories which enable us to appreciate the factors affecting the stability of polyhedral clusters and which help us to gain some understanding, at least at the phenomenological level, of the mechanistic pathways that permit their construction.‘Clusters’, of course, encompass a huge field but I shall concentrate almost exclusively on the boron hydrides since the story of clusters all began with Alfred Stock’s boranes 1*2and these compounds have traditionally been at the forefront of development of cluster chemistry. It is also probably true that the boranes and their derivatives constitute the widest range of cluster structure-types available today. This approach factorizes out electron-precise clusters whose vertices are elements from the Main Groups IV, V, and VI of the Periodic Table. For example, the CH group can act as a three-connected cluster vertex in compounds of stoicheiometry C,H, such as tetrahedrane, cubane, cuneane, and most recently the elegant dodecahedrane, C20H20.3These compounds present no conceptual difficulties in bonding, though of course they present major challenges in synthesis.Likewise, * Delivered at a Symposium of the Dalton Division of the Royal Society of Chemistry on 22 March 1984, at Queen Mary College, London El, and at the Universities of Dundee, Strathclyde, Birmingham, Loughborough, and Exeter A. Stock, ‘The Hydrides of Boron and Silicon’. Cornell University Press, Ithaca, 1933. W. N. Lipscomb, ‘Boron Hydrides’, W. A. Benjamin, New York,1963; Science, 1977. 1%. 1047.’L. A. Paquette, R. J. Ternansky, D. W. Balogh, and G. Kentgen, J. Am. Chem.Soc., 1983, 105, 5446. Molecular Tectonics: The Construction of Polyhedral Clusters clusters comprised of elements in Groups V and VI such as P,, P,S,, As,S,, P,06, and innumerable other^.^ Naked-metal clusters of the heavy main-group elements pose greater problems of bonding and electron counting: examples are4 SnS2-, PbS2-, BiS3+, Bi,”, Bi,’ +,Sn9,-, and As, Of even greater diversity are transition metal-carbonyl clusters and other metal-metal cluster compo~nds.~ Indeed, there are many connections between these clusters and those of the boron hydrides and there has been a particularly fruitful and rewarding interaction between these two broad classes of compound at both the conceptual and the experimental level.6 2 The Structure of Boranes and bane Anions We now know that the parent boranes and their anions form three main series of compound: closo-borane anions of stoicheiometry B,Hn2 -; nido-boranes B,H, + (and their anions formed by successive deprotonation, B,H, + ,-etc.); and arachno-boranes B,H, + 6 and related anions B,H, + 5-.The geometrical structure of these compounds can be summarized in what might be called our First Tectonic Principle: Closo-borane anions feature closed triangulated polyhedral clusters (sometimes called deltahedra); Nido-borane clusters are formed by removal of one vertex from a closo cluster; Arachno-borane clusters are formed by removal of two adjacent vertices from a closo cluster. Extensions are possible, hypothetically, to capped-closo boranes of stoicheiometry B,H, and to hypho-boranes B,H, + 8 and kludo-boranes B,H, + lo; examples of these series are not known for binary borane species B,H, or their anions, though they do occur amongst derivatives4 The construction of these various clusters implies that each BH vertex contributes its remaining three orbitals and two valence electrons to the skeletal bonding of the cluster.This can be generalized as a Second Tectonic Principle: The cluster geometries typified by the boranes can be generated by using asvertices atoms which contribute three orbitals in conical array to the cluster bonding. The perceptions embodied in these two tectonic principles find their most elegant codefication in a set of rules first enunciated by Wade in 1971., For our present purpose these may be paraphrased as follows: (a) closo-borane anions, BnHn2-, feature closed triangulated polyhedra (deltahedra) having n vertices and (n + 1) skeletal bonding electron pairs; (b) nido-boranes, B,H, + ,, have n vertices of an (n + 1)-vertex deltahedron occupied and involve (n + 2) skeletal bonding electron pairs-the four supernumerary H-atoms form BH,B bridge bonds around the open face; (c) arachno-boranes, B,H, + 6, have n vertices of an (n + 2)-vertex deltahedron occupied and involve (n + 3) skeletal bonding electron pairs-some of the six N.N. Greenwood and A. Earnshaw, ‘Chemistry of the Elements’, Pergamon, Oxford, 1984. ‘Transition Metal Clusters’, ed. B. F. G. Johnson, Wiley. Chichester, 1980.K. Wade, Chapter 3, p. 193, in ref. 5, and references therein. ’ R. E. Williams, Inorg. Chem., 1971,10,210; Adv. Inorg. Chem. Rudiochem., 1976, 18.67. K. Wade, Adv.Inorg. Chem. Radiochem., 1976, 18, 1. K. Wade, J. Chem. SOC.,Chem. Commun., 1971,792. Greenwood supernumerary H-atoms form BH,B bridge bonds and the rest contribute the second H-atom in BH, groups in the open face. Starting from this basis, which is now to be found in most modern textbooks of inorganic chemistry, we can ask several questions aimed at extending the range of cluster geometries so far encountered. First, what closo polyhedral clusters are possible, e.g. are they limited to a maximum of 12 vertices as in the icosahedral anion Bl2HIz2-; and second, are alternative geometric arrangements of the deltahedral vertices possible? Recent parameterized molecular-orbital calculations by Lipscomb's group lo have shown that supraicosahedral closo-borane anions up to n = 22 have apparent overall stabilities comparable with that of BloHloZ-. Indeed, B,4H142- (066) and B1,H1,,- (Dsd) have even greater stability than BloHlo2- (D4d)and aresurpassedoniy by theextremelystableanion Bl,Hl,2- (Ih) itself.The calculations further suggest that certain hypothetical neutral closo-boranes with only 2n skeletal electrons might also be stable, e.g. B,,H,, (Td), B19H,, (C3J, and B,,H,, (Td). The computational procedures that were used restricted detailed calculations to deltahedra with no more than 24 vertices, though certain even larger clusters with very high symmetry might also have considerable stability.One such structure for n = 32 is the elegant omnicapped dodecahedron of Ih symmetry shown in Figure 1-this has 20 seven-co-ordinate and 12 six- co-ordinate boron atoms if the 32 terminal H-atoms are also included. Figure 1 The 32 vertex structure of /,, symmetry envisaged for a hypothetical closo-borane (anion?)B, ,H , The question of alternative geometrical arrangements for the BH vertices of closo-borane anions has been considered by Kepert and his group. For example Figure 2 shows 17 polyhedra with 12 vertices. On the basis of an admittedly oversimplified procedure, it appears that no fewer than ten of these might have a stability intermediate between those of the observed icosahedral B, ,H ,22 -(a), and the frequently invoked cuboctahedral rearrangement intermediate (b).Note, however, that very few of the structures in Figure 2 are fully triangulated polyhedra, many having faces with four, five, or even six edges. The synthetic implications of these geometrical considerations are intriguing and we shall return to them shortly. lo J. Bicerano, D. S. Marynick, and W. N. Lipscomb, Inorg. Chem., 1978, 17, 3443. 'I D. J. Fuller and D. L. Kepert, Inorg. Chem., 1982, 21, 163; Polyhedron, 1983, 2, 749. Molecular Tectonics: The Construction of Polyhedral Clusters a (a) (C) 0 F D F H F J J K DC 0 HaF FH LK I L I Figure 2 Polyhedra with twelve vertices:' '(a) regular icosahedron, (b) cuboctahedron, (c)truncated tetrahedron, (d) hexagonal prism, (e)hexagonal antiprism.(0 bicapped pentagonal prism, (g)square cupola, (h) sphenomegacorona, (i) anticuboctahedron, G) D3,,icosahedron, (k)C,, icosahedron, (1) tetrarapped kite prism, (m) hebesphenocorona, (n) compressedcuboctahedron, (0)elongated cuboctahedron.(p) orthorhombically distorted cuboctahedron, (9)disphenohedron Greenwood A third type of structural question arises when we think of the formation of nido and aruchno species by the notional removal of vertices from the parent closo species. As can be seen in Figure 3, the skeletal connectivities of BH vertices in the anion B,Hn2-range from three to six and, with the exception of B,H,*-and B,,H 22-, each anion features two or more different types of vertex.Usually a nido cluster is formed by the notional removal of the most highly connected vertex. Clearly, however, a different nido cluster geometry would arise if a vertex of lower connectivity were removed. Whilst this particular type of geometrical isomerism is not yet known among the binary borane species themselves, it is becoming increasingly common among heteroatom boranes. An example is in Figure 4, which Figure 3 Cluster geometries for the closo-borane anions B,HnZ-showitzg the various skeletal connectivities of the BH vertices (a = 3; 0 = 4; 0= 5; 8 = 6). Note: the anion B,H,’-is not known but the isoelectronic carbaborane C2B3H5 is well characterized shows two nido-type ten-vertex structures formed by the notional removal of the six-connected and four-connected vertices respectively from the eleven-vertex closo polyhedron shown in Figure 3.Considerations of this sort lead to a refinement of the first tectonic principle enunciated above. This refinement can be expressed as the First Tectonic Corollary: Ifa closo cluster has more than one type of vertex (e.g. Molecular Tectonics: The Construction of Polyhedral Clusters vertices with dffering connectivities to the cluster) then thegeometricalshape of related nido clusters will depend on which vertex is removed. Likewise for arachno clusters. A fourth type of structural question concerns the possibility of clusters of clusters. So far we have considered only individual clusters but these in turn may be b (a) (b) Figure4 The decaborane-like structure of nid~-[H(PPh,)~lrB,,H l2 (a) compared with the structure of rhe deep purple iso-nido-[(PPh,),Ir(B,H,,-PPh,)] l3 (b) In this latter compound the B-B distances are in the range 177 f 10 pm except for B(8)-B(lO) which is 222 pm (i.e.essentiallv non-bonding) incorporated as the subunits of larger conjuncto-boranes. Several possibilities can be envisaged, all of which have been observed in pra~tise.~ Thus, conjuncto-boranes or conjuncto-heteroatom-boranesmay feature: (a) two (or more) clusters linked by direct €3-B bonds; (6) clusters fused at a comrno-B or commo-M atom; (c) edge-fusion via a common pair of B or M atoms; (6)confacial-fusion via a common triangle of B or M atoms; (e) more complex conjunctions.An excellent example of the structural subtlety that can arise amongst even the simplest-type (a)-conjuncto-boranes is afforded by our recent work on the isomers of BZoH,, i.e. conjuncto-(BIoHl 3)2 or bi(nido-decaboranyl).1616 There are four geometrically distinguishable types of B atom in nido-B,,H14 and this IZ S. K. Boocock, J. Bould, N. N. Greenwood, J. D. Kennedy, and W. S. McDonald,J. Chem. Soc., Dalron Trans., 1982, 713. I3 J. Bould, N. N. Greenwood, J. D. Kennedy, and W. S. McDonald, unpublished work, 1982. l4 N. N. Greenwood, J. D. Kennedy, T. R. Spalding, and D. Tay1orson.J. Chem.SOC.,Dalton Tram., 1979, 840. Is S. K. Boocock, Y. M. Cheek, N. N. Greenwood, and J.D. Kennedy, J. Chem.Sor., Dalton Trans., 1981, 1430. S. K. Boocock, N. N. Greenwood, J. D. Kennedy, and W. S. McDonald. J. Chem. SOC.,Daltorr Trans., 1981.2573. Greenwood leads to the possibility of eleven geometrical isomers of (BloH13),,see Figure 5. Furthermore, since B(5) and B(7) are not contained within either of the molecular mirror planes of B10H14,bonding to either of them generates enantiomeric pairs; there are four such pairs in addition to the unique 5,7' meso-diastereoisomer, making 15 distinct isomers in all. As expected, many of these show markedly different chemical reactions, particularly when comparison is made between isomers where the intercluster bonding involves atoms in the six-membered open face of decaborane rather than the 'hinge' (1,3) or 'apical' (2,4) boron atoms.Numerous other examples of B-B bonded conjuncto-boranes are known, e.g. [(CIOSO-B,~H~)~]~-,(do-B,H,),, and (ar~chno-B,H,),.~ An example of conjunction via a boron atom common to two subclusters is conjuncto-B, 5HZ3: as shown in Figure 6 this comprises a B7 plus a B, unit fused 1J ' 1,2' 1,6' L1,5' 17' I 52' 2,6' 6,s' 2F 2.7' I 5.6' 7,6' Figure 5 Schematic representation of the eleven geometrical isomers of (nido-B,,H ,3)2 using the conventional numbering system of decaborane(l4). The positions 5,7 (and 8, 10) are chiral and intercluster bonding involving one of' these positions leads to enantiomeric pairs as shown. The 5,7' isomer is the unique meso-diastereoisomer of the 53 and 7,7' enantiomeric pair "J.C. Huffman and R. Schaeffer. personal communication of unpublished work, 1981. Molecular Tectonics: The Construction of Polyhedral Clusters at a common boron atom (shaded). The structure may alternatively be regarded as a derivative of the ligand adduct LBgH13 (cf:arachno-BgH14-) in which the ligand is replaced by nido-B6Hlo which donates the electrons in its basal B-B bond to form a three-centre BBB bond. We shall return to this structure-type in a later section. Examples of edge-shared cluster fusion are afforded by B14H2O and the two isomers of Bl’H22 whereas confacial cluster fusion is found in (MeCN)2B20H 16 (triangulo) and c1oso-B20H16 (4B conjunction). Other more complex conjunctions will emerge in later sections.R B” Figure 6 Srrucrure ofconjuncto-BI 5H23 showing the commo-B atom shaded (see text) So far we have been dealing with polyhedral boranes, in which the cluster vertices are occupied exclusively by boron atoms. In seeking to extend the treatment we next consider the possibility of incorporating other elements as deltahedral vertices. 3 Heteroatom Boranes It has been known for over 20 years that carbon can be incorporated into borane clusters and that the resulting carbaboranes are often extremely stable compounds. ’ More recently it has become apparent that numerous other elements, both non-metals and metals, can act in this way and to date some 40 elements have been incorporated as heteroatoms in borane cluster^.'^ As implied by Wade’s rules, each boron atom contributes three frontier orbitals in conical array and two electrons to the cluster bonding.’ With carbon the bonding can still be described in terms of the same conical array of three frontier orbitals but the heteroatom now contributes three electrons rather than two to the cluster bonding.Other possibilities can easily be envisaged and examples are known in which R. N. Grimes, Carboranes’, Academic Press, New York,1970. l9 N. N. Greenwood, Pure Appl. Chem., 1983,55, 77 and 1415. Green wood heteroatoms contribute 0, 1, 2, 3, or 4 electrons to the cluster bonding. Such examples fall within the original concept of Wade's rules. More drastic extensions can, however, be imagined. For example, the heteroatom might contribute 1,2,3, or 4 frontier orbitals and, even with three frontier orbitals, the arrangement might be for example T-shaped rather than conical.Indeed, one can envisage the possibility that some transition metals could individually contribute a variable number of electrons, a variable number of orbitals, and a variable geometrical arrangement of orbitals to the cluster bonding. This diversifies the tectonic possibilities, and extends enormously the range of structures that can be synthesized. The position can be summarized in a ThirdTectonic Principle: Cluster geometry depends on both the number and the mutual arrangement of the frontier orbitals contributed by each atomic vertex. One of the first structurally characterized examples of a 1-orbital contributor was the nido-pentaborane derivative [(PPh3)2CuB~Hsl; "as shown in Figure 7 the copper(1) centre {L~CU-} replaces a bridge-hydrogen with which it is isolobal.21 Figure 7 Molecular structure of [(PPh3)2Cu(q2-BsHs)] with the six phenyl groups omitted for clarity.Note the similarity to the structure of nido-B,H,, one H, having been replaced by the ((PPh3)2Cu-) moiety Numerous other 1-orbital contributors can act in this way e.g., 22 { L2Ad-1, { L2Au1-),{L2XNi"-}, {L2XPdu-},{L2XPt"}, { LXCd"-), {(q5-CsHs)Be},{ Me2B-}, { RJSi}, { R3Ge-}, (MeSSn-}, (Me3Pb-1, and { Me2P1"-}. 2o V. T. Brice and S. G. Shore, J. Chem. Soc., Chem. Commun., 1970, 1312; J. Chem. Soc., Dalton Trans., 1975, 334; G. G. Outerson, V.T. Brice, and S. G. Shore, Inorg. Chem., 1976, 15, 1456. 21 N. N. Greenwood, J. A. Howard, and W. S. McDonald, J. Chem. Soc., Dalton Trans., 1976, 37. 22 N. N. Greenwood and J. D. Kennedy in 'Metal Interactions with Boron Clusters', ed.R. N. Grimes, Plenum Press. New York. 1982, Chapter 2. p. 43. 36 1 Molecular Tectonics: The Construction of Polyhedral Clusters Two-orbital contributors are exemplified by planar (L2Pt1'} (1) and octahedral {L3HIr"') (2): L Ir' L I"\ Metallaboranes that are thought to incorporate at least some contribution of this sort to the metal-boron bonding include [(PMe2Ph)2PtB8H I 2],23 [{(P-M~~P~)~P~)~B~HIo],~~ [(S~CNE~~)AUB~HI[(CO)(PMe&HIrBeH1 IC~],~~ ~1,~' and [{(S2CNEt2)Au)2B8H Indeed, very recently a heterobimetalladecabor- ane cluster has been synthesized which contains both these structurelbonding features within the one cluster viz.[{(PMe3)2Pt)B,Hl,(IrH(PMe,),(CO)}], (see Figure 8).26 Figure 8 Molecular structure of [{(PMe3)2Pt )B~HIo{I~H(PM~~)~(CO))] as deducedfrom detailed multi-element single- and multiple-resonance n.m.r. spectroscopy. (Remember that lines within the deltahedral cluster delineate the polyhedral geometry and do not represent two- electron two-centre bonds.) Three-orbital contributors include the classic conical borane and carbaborane vertices {BH) (3) and {CH} (4) as well as isolobal groups such as {L3Fe} (5) and {(qs-CsH~)Co)both of which, like { BH}, contribute two electrons to the cluster. T-Shaped two electron contributors such as {LPt} (6) can also be envisaged: 23 S.K. Boocock, N. N. Greenwood, M. J. Hails, J. D. Kennedy, and W. S. McDonald, J. Chem. SOC., Dalton Trans., 1981, 1415. 24 J. Bould, J. E. Crook, N. N. Greenwood, and J. D. Kennedy, J. Chem. Soc., Dalton Trans., 1984, 1903. "M.A.Beckett.J. E. Cro0k.N. N.Greenwood,and J. D.Kennedy,J. Chem.Soc.. Dalton Trans., 1984,1427. 26 J. Bould, N. N. Greenwood, and J. D. Kennedy, J. Chem. SOC.,Dalton Trans., 1984. 2477. 362 Greenwood Conical: I T-sh aped L-T (6) (7) A T-shaped disposition would also result if three meridionally grouped octahedral orbitals (7) were used for cluster bonding, but examples of this have not yet been synthesized. As an example of a compound containing a conical three-orbital metal contributor we can cite the nido-metallaborane [(CO),FeB,H,] (Figure 9) which was first obtained in a hot/cold tube reaction (at 220/20 "C) between nido-B,H, and [Fe(CO),]?' The structural analogy with do-B,H, is obvious, the apical {BH} unit having been replaced by the isolobal {Fe(CO),} group.When the reaction is carried out at 60-70 "C in dimethoxyethane in the presence of LiAlH,, the nido-diferraborane [{ (CO),Fe},B,H,] is obtained in which a second (basal) {BH} has also been subrogated.28 If all five {BH,} vertices in B,H, are notionally replaced by {Fe(CO),} vertices and the four H, are replaced by the four-electron equivalent C, then the well-known carbidopentairon carbonyl [Fe,(CO), ,C] is obtained (Figure Figure 9 The structure of nido-[(CO),FeB,H,], an analogue of nido-B,H, with the apical {BH) group subrogated by the isolobal {Fe(CO),) group ''N.N. Greenwood, C. G. Savory, R. N. Grimes, L. G. Sneddon, A. Davison, and S. S. Wreford, J. Chem. SOC.,Chem. Commun., 1974, 718. '* E. L. Andersen, K. J. Haller, and T. P. Fehlner. J. Am. Chem. SOC.,1979, 101,4390. 29 E. H. Braye, L. F. Dahl, W. Hubel, and D. L. Wampler, J. Am. Chem. SOL-.,1962,84.4633. Molecular Tectonics: The Construction of Polyhedral Clusters Two further aspects of the structure of nido-[(CO)~FeB4Hs] of great significance emerge if we rewrite the formula more explicitly as [(C0)3Fe(q4- B4H4(HJ4}]; replacement of the four (BH,} units by isoelectronic C atoms then generates the formula [(CO)3Fe(q4-C4H4)], i.e.tetrahapto-cyclobutadiene iron tricarbonyl. This introduces in explicit form not only the concept of boranes as ligands but also the possibility of stabilizing fugitive species such as cyclo-C4H4 or cyclo-B4Hs by co-ordination to the metal centre. These examples and many other^,^^*^^ show that there is an almost continuous gradation of isostructural compounds between the polyhedral boranes, the metallaboranes, organometallic compounds, and metal carbonyl clusters. The power of Wade's rules in systematizing such apparently diverse families of compound is clear. When conical 0 W 0 Figure 10 The structure of [Fe5(CO),,C], see text three-orbital contributors are involved these perceptions are also extremely helpful in assigning preliminary structures to compounds, perhaps obtained initially in small yields, before more effective syntheses have been devised.A good illustration of this approach concerns the iridaborane [(CO)(PMe,),IrB4H,] which was first unexpectedly obtained in milligram amounts as a by-product of the reaction of B4H4(HJ4}]; replacement of the four {BH,} units by isoelectronic C atoms then conventional elemental analysis but the detailed n.m.r. properties indicated that the compound could reasonably be formulated as the apex-subrogated derivative of arachno-pentaborane, B5H 1. This conviction was strengthened by the realization that the complex has two more electrons available for cluster bonding than does nZdo-[(CO),FeB,H,], since it contains one more H atom and Ir is in the group following iron in the Periodic Table.The structure was confirmed by a crystal 'O C. E. Housecroft and T. P. Fehlner, Adv. Organomet. Chem., 1982,21, 57. 31 J. Bould, N. N. Greenwood, and J. D. Kennedy. J. Chem. SOC.,Dalton Trans., 1982, 481 Green wood Figure 11 The structure of the metallahexaborane cluster in nido-[(CO)(PPh,),IrB,H, J. The Ir atom is almost coplanar with the four basal boron atoms B(3-6) in contrast to the position of the Cu atom in [(PPh,),CuB,H,] (Figure 7) which superficially has similar stoicheiometry structure determination on the closely related compound arachno-[(CO)(PMe,- Ph),IrB,H9].32 Another (much higher yield) route to metallaborane clusters involving a conical three-orbital metal contributor is the oxidative insertion of an iridium(1) complex into nido-B5H8- to give an expanded six-vertex nido-cluster: 33 nido-B,H8-+ tran~-[Ir(CO)Cl(PPh,)~]-nido-[(CO)(PPh,),IrB5H8] + CI-As can be seen in Figure 11, the compound is a direct structural analogue of nido- B6H 10’ The development of various high-yield routes to metallaboranes invites speculation that successive reactions might produce polymetallaboranes, and several examples of interesting and potentially useful compounds are now known.Thus, cluster expansion by oxidative insertion of [Os(CO)CIH(PPh,),] into nido- B5H8- affords the six-vertex nid0-[(CO)(PPh,)~0sB,H,1 in 80% yield and this can then be deprotonated and reacted with cis-[PtCl,(PMe,Ph),] to give the seven- vertex nido-[{ (CO)(PPh3),0s}H,(PtCl(PMe,Ph)}B5H,] (Figure 12).34 The com- pound is notable for the presence of an OsH,Pt grouping which exerts a stabilizing influence, since seven-vertex clusters are conspicuous by their complete absence among binary borane species themselves. The compound also illustrates the first tectonic corollary mentioned above since the nido-OsPtB, cluster (8) can be notionally generated by removal of a five-connected vertex from the parent dodecahedra1 closo-€3, structure (9) and is thus a cluster isomer of the known nido- ” S. K. Boocock, M. J. Toft, and S. G. Shore, 182nd Chem. Soc.National Meeting, New York. 23-28 August 1981, Abstract INOR 149.’’ N.N.Greenwood,J.D. Kennedy, W. S. McDonald, D. Reed, and J. Staves,J.Chem.Soc., Dalton Trans., 1979, 117. ’* J. Bould, J. E. Crook, N. N. Greenwood, and J. D. Kennedy, J. Chern. Soc., Chem. Commun., 1983,951. Molecular Tectonics: The Construction of Polyhedral Clusters [{(qS-C,Me,)Co},B,H,] (1 1) 35 which can be notionally generated by removal of a four-connected vertex (10): Figure 12 The structure ofthe seuen-vertes nido-heterohirnetallaborane [{ (CO)(PPh,),Os}-H,{ PtCl(PMe,Ph))B,H,] As an alternative to the nido+nido+nido tectonic sequence just considered it is also possible to generate nido-metallaborane clusters in high yield by adding a vertex to an arachno-cluster (i.e. by reversing the sequence implied by the first tectonic principle mentioned above).An example is afforded by the reaction 36 of arachno-B,H,,-with [IrCl(PPh3)3] which gives an 85% yield of nido-[H(PPh,),IrB,H 3] plus small amounts of the isomeric cluster iso-nido-[(PPh,),Ir(B9H ,,.PPh,)]; these two compounds have already been illustrated in Figure 4. 4 Novel Closo Deltahedral Geometries Within the context of the Williams-Wade f~rmalisrn,’~~closo structures are formally derived by capping the open face of a nido cluster with a conical three- orbital contributor such as (BH} or one of its many isolobal analogues. If a four- orbital contributor could be devised, then it would be possible to imagine the generation of a closo-type structure by the direct capping of an arachno-cluster with ’’T. L.Venable and R. N.Grimes, Inorg. Chem.. 1982, 21, 887. 36 J. Bould, N. N. Greenwood, J. D. Kennedy, and W. S. McDonald, J.Chem. SOC..Chem. Commun., 1982, 465. Greenwood a single vertex. The process is illustrated schematically below for arachno-B,, -B9, and -Blo clusters leading to the closed deltahedral structures (MB,} (12), (MB,} (13), and (MB,,)(14).37 (12 1 (14 1 It is notable that structure (12) differs from the tricapped trigonal pyramidal structure of B,H,’-(Figure 3) and that the structure of (13) differs from the familiar bicapped Archimedian antiprismatic structure of B,,H -(Figure 3). However, because the skeletal structures of arachno-B,,H 142 and nido-BloH,, are very similar, it follows that structure (14) for (MB,,} is essentially the same as that for B, ,H , -(Figure 3), with the unique six-connected (BH) vertex subrogated by the metal centre.These observations can be expressed as a Second Tectonic Corollary:The capping of an arachno cluster with a single additional vertex may result in a closo structure differing in geometrical shape from that formed by the cupping qfa nido-strucrure. Examples of all three iso-closo structure types (1 2), (1 3), n n Cl Figure 13 The structure of iso-closo-[H(PMe,),IrB,H,CI J shon.rng the 2,2,4,1 stack of the central deltahedron. The Ir-B(2,4,5,7) distances are 218.5 (& 1.0) pm and Ir-B(3,6) are 231 pm;there is no signiJicant bonding between B(2)-B(4) or B(5)-B(7) (307 and 304 pm respectioelj.) ’’J. E. Crook, M. Elrington, N.N. Greenwood. J. D. Kennedy, and J. D. Woollins, Polvhedron. 1984, 3, 901. Molecular Tectonics: The Construction of Polyhedral Clusters and (14)are now known and will be described in the ensuing paragraphs. The first example of a nine-vertex iso-closo deltahedron was obtained by the mild 1CI]: 38 at ca. 80 "C there is thermolysis of urachno-[H(CO)(PMe~)~1rB~Hl quantitative elimination of HZ to give nido-[(CO)(PMe3)~1rB~H and this, at IoC~] 135 "C, eliminates a further mole of HZ plus CO to give a 45% yield of the poppy red iso-~loso-[H(PMe3)~1rBeH7C1).As can be seen in Figure 13, the central deltahedron comprises a 2,2,4,1stack (B~BzB~IT} rather than the usual 3,3,3 stack of a tricapped trigonal prism (Figure 3). The iridium atom is thought to contribute four orbitals and four electrons to the cluster bonding, i.e.it is formally in the +5 oxidation state. The facile, thermally-induced arachno-+nido-+closo cluster closure is also particularly noteworthy as the first experimental demonstration of a sequence which is implicit in the familiar Williams-Wade cluster formalism (albeit to give an iso-closo deltahedron as end product). A good example of a ten-vertex iso-closo deltahedron is afforded by the mild thermolysis of nido-[H(PPh3)21rBgH 3] (Figure 4): heating this compound at 85 "Cresults in the elimination of 3H2 and the quantitative formation of the novel, -bright orange, orthocycloboronated iso-c~os~-[(HPPh,)(Ph~~~~~~)~rB~~~] shown in Figure 14.36Several other examples are also known.The 3,3,3,1 stack of Figure 14 The structure of isocloso-[H(PPh,)(Ph,PC6H4)IrB9H8] showing the 3331 stack of the central deltahedron. Distances from Ir( 1) to B(2,4,6) are in the range 214.8-218.8 pm and to B(3,5,7) are in the range 237.8-246.0 pm the iso-cioso deltahedron [B(8,9,10),B(3,5,7),B(2,4,6)Ir(l)] is immediately apparent and the idealized local CJVsymmetry of the cluster contrasts with the D4, symmetry of the more familiar 1,4,4,1stack of BloHIo2- (Figure 3), or the closely 38 J. Bould, J. E. Crook,N. N. Greenwood, J. D. Kennedy, and W. S. McDonald, J. Chem. Soc., Chem. Commun., 1982,346. Green wood related C4"local cluster symmetry of the 1,4,4,1 stack in [(PMe2Ph),NiB,H,C12].39 Again the iridium atom can be considered to be an IrV four-orbital, four-electron contributor to the cluster bonding.The eleven-vertex iso-closo structure (14) is exemplified by [(PPh,),RuB,,- H *(OEt),].3 5 Macropolyhedral Metallaboranes The preceding two sections have explored the tectonic implications of replacing the conical three-orbital contributor { BH} with a range of other non-isolobal metal vertices. The perceptions emerging can be summarized as a Third Tectonic Corollary: The range of cluster geometries observed for parent boranes and their anions can be considerably extended by incorporating as vertices metal atoms which differ from boron in the number or mutual arrangement of their frontier orbitals. In the present section we explore the tectonic consequences of one specific modification of vertex-bonding geometry, namely the replacement of a conical by a T-shaped three-orbital contributor.We can anticipate the detailed results by stating a Fourth Tectonic Corollary: The T-shaped distribution of three frontier orbitals facilitates the construction of complex macropolyhedral clusters. Perhaps the first indication of this structural effect was the formation of the extremely stable 14-atom cluster [(PMe,Ph),Pt,B, 2H,8] [Figure 15(a)]."' The compound is formed in several quite diverse reactions, though often in low Ph MeJP +Pt -P/Me-Me / \Me Ph\Brii'6' (b) Figure 15 (a) The structure of the macropolyhedral dimetallaborane [(PMe,Ph),Pt2B, ,H comprising two (q'-B,H,) subclusters joined in transoid conJguration to a central P-Pt-Pt-P group (Pt-Pt distance 264.4 pm).(b) Simplijied bonding scheme showing the T-shaped distribution of the three orbitals contributed to the cluster by each of the two {(PMe,Ph)Pt) vertices; the bonding in each of the borane subclusters is typical of nido-B,H,-39 N. N. Greenwood, M. J. Hails, J. D. Kennedy, and W. S. McDonald, J. Chem. SOC.,Dalfon Trans., in the press. 40 N. N. Greenwood,M. J. Hails. J. D. Kennedy, and W. S. McDonald, J. Chem. SOL-..Chem. Commun.. 1980, 37. Molecular Tectonics: The Construction of Polyhedral Ciusters yield;23.40.4 1 Th e more open transoid centrosymmetrical geometry is notably different from the cisoid configuration of the known ‘isoelectronic’ binary borane, BI4Hz0, i.e.[(H2B2)BlzH,a].42 A variant on this theme is the related compound [(PMe,Ph),Pt,B,H,,] which is shown in Figure 16(a).41 The central linear P-Pt-Pt-P array is still present as is one of the (q3-B,H9) subclusters, but the second (q3-B,H9) has been replaced by the vestigial (q3-B,H,) group [Figure 16(b)]. A fascinating feature of this latter compound [(PMe,Ph),Pt,B,H ,,J is that it is isoelectronic with the known compound [{(PMe2Ph),Pt),BaHlo], the two additional phosphine ligands providing the four electrons formerly supplied by the extra four H-atoms; despite this the structures of the two compounds are entirely different as can be seen by comparing the arachno-B,oH,42--like structure of [{(PMe2Ph),Pt)2B,H,o] (Figure 17) 23 with the macropolyhedral structure of I B-H’yy-yPh Ph \Me-P--Pt -Pt -P-/ Me /Me ‘Me (a) (b) Fi e 16 (a) The structure of the diplatinadecaborane [(PMe,Ph),PtzB8H,,] comprising an (qc2Hs) and an (q3-B,H,) subcluster joined to a central linear P-Pt-Pt-P group (Pt-Ptdistance 262.1 pm); note the cisoid location of the two phenyl rings in contrast to their transoid location in Figure lS(a).(b) Simplifiedbonding scheme showing similarity to that in Figure 15(b) Figure 17 The arachno-B, ,42 -like structure of the diplatinadecaborane cluster C{(PMezPh)zPt)2B8H101 41 R. Ahmad, J. E. Crook,N. N. Greenwood, J. D. Kennedy, and W. S. McDonald, J. Chem. Soc.. Chem. Commun., 1982, 1019. 42 J. C. Huffman, D. C. Moody, and R. Schaeffer. J.Am. Chem. Soc.. 1975,97, 1621. 370 PMezPh Figure 18 (a) The structure of the 17-vertex macropolyhedralplatinaborane [(PMe2Ph)(PtBt6H 18(PMe2Ph)}] shouing the q6-bonding of the B16 cluster to the Pt atom via B(6)B(2)B(g’)B(2’)B(6’)B(ll’)and with Pt-B distances in the range 223-229 pm. (b) A stick diagram illustrating the B6 and Blo subclusters fused at commo-B(8‘) and a simplified version of the bonding envisaged to the Pt atom Molecular Tectonics: The Construction of Polyhedral Clusters Greenwood [(PMe2Ph),Pt2B8H14] [Figure 16(a)]. An even more complex series of 17-vertex macropolyhedral metallaboranes has been obtained by mild thermolysis of the nine-vertex arachno-platina-nonaborane [(PMe2Ph),PtB8H12] in refluxing toluene (ca.110 0C).43-45The first such compound to be characterized was the air-stable, flame-red macropolyhedral cluster [(PMe2Ph)(PtB,,H,,(PMe2Ph)}] [Figure 18(a)].”, The structure can be viewed as a complex between a tetradentate q6-B16 ligand and the platinum centre. The B16 conjuncro-borane (which is at present unknown as a neutral binary borane) features a B, and a B,, subcluster fused at a commo-B(8’) atom [Figure 18(b)] and is clearly related to the (B, + B,) structure of conjuncto-B, 5H2, shown in Figure 6.As such the complex affords a further example of a hypothetical borane cluster stabilized by co-ordination to the metal centre. An alternative description of the macropolyhedral cluster is as a nido-{PtB,) subcluster edge fused along the Pt-B(8’) vector to a nido-{PtB,,} subcluster.A second 17-vertex macropolyhedral cluster isolated from the same thermolysis reaction is the unprecedented dark-green trimetallaborane [(PMe,Ph),Pt,B,,H 16] shown in Figure 19(a).44 The cluster geometry is shown in a slightly different orientation in Figure 19(b); it comprises two nido-type B, subclusters conjoined via a complex five-atom belt, PtPtB,Pt. The two B, subclusters are oriented with their bases parallel and facing. When the three Pt atoms are removed, the remaining hypothetical B,, ligand is seen to have a quite different structural motif than that in the known B,,H2, or the 14-atom skeleton in Figure 15. It is possible to devise a simple set of localized bonding orbitals to describe the bonding in both the 17-vertex clusters discussed in the preceding two paragraphs but this enterprise is unnecessarily detailed for our present purpose.Suffice it to say that the introduction of metal centres having vacant bonding orbitals or readily removable ligands often facilitates the synthesis or geometric rearrangement of clusters or the construction of macropolyhedral metallaboranes. 6 Summary Some 40 different elements can be incorporated as cluster vertices in polyhedral heteroboranes.” Over 75 different cluster geometries have been identified so far amongst the products of cluster syntheses 19,22 and well over half of these have no counterpart among the familiar structures of the binary boranes and borane anions. This diversity arises in part from the fact that metal centres can contribute varying numbers of orbitals and valence electrons to the cluster bonding.In fact, a selection of metal vertices is available which can be thought to contribute 1,2,3, or 4 orbitals and 0, 1, 2, 3, or 4 bonding electrons to the cluster. This extends considerably the Williams-Wade formalism 6-9 which has proved so helpful for ”M. A. Beckett, J. E. Crook, N. N. Greenwood, J. D. Kennedy. and W. S. McDonald, J. Chem. SOC., Chem. Commun., 1982, 552. M. A. Beckett,J. E. Crook, N. N. Greenwood, and J. D. Kennedy,J. Chem.SOC.,Chem. Commun., 1983, 1228. ”M.A.Beckett, J. E.Crook,N. N. Greenwood,and J. D. Kennedy, J. Chem.Soc.,DufronTruns.,submitted. Molecular Tectronics: The Construction of Polyhedral Clusters clusters constructed from vertices which are formal conical three-orbital contributors.A series of tectonic principles and corollaries has been formulated which incorporate these earlier perceptions and which extend them systematically to a wide variety of other circumstances. In particular, examples are presented in which: (a) iso-nido and iso-arachno clusters are obtained from closo deltahedra by notional removal of vertices with differing connectivities; (6)iso-closo deltahedra are obtained by the capping of arachno rather than nido clusters with a single vertex which can be thought to contribute four orbitals rather than three to the cluster bonding; (c)macropolyhedral conjuncto-boranes are generated by incorporation of metals which contribute a T-shaped or related suite of frontier orbitals.This tectonic approach not only helps to systematize the known types of polyhedral cluster but also places on a firm and more systematic basis the design and construction of novel polyhedral clusters, thereby assisting the planned synthesis of potentially useful or interesting target molecules.

ISSN:0306-0012    

DOI:10.1039/CS9841300353

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

Chemical Processeson Heterogeneous Catalysts By C. Kemball DEPARTMENT OF CHEMISTRY, UNIVERSITY OF EDINBURGH, WEST MAINS ROAD, EDINBURGH EH9 3JJ 1 Introduction This review is based on an address given at the University of St. Andrews as part of the Irvine Review Lectures on ‘Catalysis’ in November 1982. The title was chosen deliberately to cover two rather different aspects of catalysis. The first of these is concerned with major industrial processes and selected examples will be briefly reviewed. The approach adopted will embrace the development of processes in the present century and not concentrate solely on current practice. The second, and more substantial part of the article is concerned with the scientific investigation of catalysts to establish the chemical mechanisms which occur thereon.The subject of heterogeneous catalysis is so enormous that it is necessary to be more than usually selective in deciding on the scope of this review and the choice of what to include or exclude presents a problem. 2 Major Industrial Processes Catalysis is of the very greatest importance in the chemical industry. Most of the processes involve heterogeneous catalysis, usually of the gas/solid type, but some depend on homogeneous catalysis in a single phase. The history of the subject of industrial catalysis has been reviewed recently by Heinemann’ in a valuable account which includes a list of major innovations in the period 1935-78. For the purposes of this article a decision was made to illustrate the role that catalysts play in practice by reference to a limited number of topics or themes.If experts in the field are asked for their own list of a dozen topics which they believe to be representative examples of industrial catalysis, there is generally good agreement about some that should be included but considerable variation in the others that appear. Consequently, there is an element of subjectivity about the dozen topics which will now be briefly reviewed. A. Synthesis of Ammonia.-This reaction involving a decrease in the number of molecules and an exothermic process must be carried out under high pressures and at as low a temperature as possible to give a reasonable yield of ammonia. A catalyst is essential and the development of the first successful process was achieved by F.Haber and his associates, early in the present century. An account of this work has been given by Mittasch.’ Most of the successful catalysts seem to involve metallic iron stabilized by the presence of some alumina to act as a structural ’ H. Heinemann, in ‘Catalysis. Science, and Technology’, ed. J. R. Anderson and M. Boudart. Springer-Verlag, (Berlin. Heidelberg. New York), 1981, Vol. 1. p. 1. A. Mittasch, Adv. Caral., 1950, 3, 81. Chemical Processes on Heterogeneous Catalysts promoter and with some potash too. The latter is believed to enhance the rate of the difficult step, the dissociative chemisorption of the nitrogen gas, by some electron transfer through the iron to the anti-bonding orbitals of the diatomic molecules.A recent review by Ertl discusses not only the role of potassium on the surface of the iron catalysts but also the substantial variation in rates associated with different crystal faces of which the (1 11) is normally the most active. The catalytic synthesis of ammonia is a vital process for the chemical industry and the main route by which atmospheric nitrogen can be converted into a great range of substances. B. Catalytic Oxidation of Ammonia or Sulphur Dioxide.-These processes pre-date the synthesis of ammonia and, of course, provide a convenient source of nitric acid or sulphuric acid. A detailed account of them was given by Dixon and Longfield4 in the latest volume of a useful series of books edited by P.H. Emmett. The oxidation of ammonia is carried out over gauze catalysts (10% rhodium-90% platinum) typically with about 9% ammonia in air at temperatures around 850 “C. Two main types of catalysts have been used for the oxidation of sulphur dioxide: platinum supported on one of a number of bases or vanadium pentoxide promoted with potassium sulphate or related compounds. The manufacture of sulphuric acid is often quoted as an indicator of the economic climate of the chemical industry and the level of consumption of sulphuric acid can be used as a measure of a country’s technical development. C. Catalytic Hydrogenation.-The field of catalytic hydrogenation embraces a very wide range of reactions.6 One of the older processes is the hydrogenation of double bonds in oils and fats-the so-called ‘hardening of fats’.More recent applications include the selective hydrogenation of small quantities of acetylene in the presence of excess ethylene as part of the purification of the alkene for polymerization. Then there are numerous examples of hydrogenation of molecules containing atoms other than carbon or hydrogen such as the hydrogenation of adiponitrile to hexamethylenediamine. Raney nickel, prepared by the caustic leaching of aluminium from a Ni-A1 alloy, is a well-known catalyst in this field. But a number of transition metals are used because they exhibit the desirable features of activity or selectivity for particular reactions-advantages which are sufficient to balance the higher cost of some of the metals.The reverse process of dehydrogenation, e.g. of paraffins to olefins, is a related type of catalytic reaction. Because of thermodynamic limitations, higher temperatures are required and oxides such as chromia supported on alumina are more useful than metal catalysts which would give undesirable side reactions. D. Catalytic Cracking of Hydrocarbons, including Fluidization and Useof Zeolites-In terms of tonnage, the catalytic cracking of petroleum feed stocks must be near G. Ertl, Catal. Rev.-Sci. Eng., 1980, 21, 201. J. K. Dixon and J. E. Longfield, in ‘Catalysis’, ed. P. H. Emmett, Reinhold, (New York),1960,Vol. 7, p. 281.’B. G. Reuben and M. L. Burstall, ‘The Chemical Economy’, Longman. (London), 1973. See ‘Catalysis’, ed.P. H. Emmett, Reinhold, (New York),1955, Vol. 3. 376 Kemball the top of the league. Both the nature of the catalysts and the technology used have developed significantly over the last 40 to 50 years. Nearly all the catalysts are some variation of ‘acidic’ oxides. In the early processes natural montmorillonite clays were used but soon replaced with synthetic silica-aluminas containing 1&13% alumina. The formation of ‘coke’ on the catalyst leading to deactivation and requiring an oxidizing treatment for its removal and regeneration of the catalyst was countered by the introduction of the new technology of fluidization in the early 1940s. The catalyst was prepared with a particle size such that it could be transported by the gas stream and cycled between a reactor in which the cracking reaction occurred and a second chamber in which the coke was burnt off before the catalyst was returned to the reactor.A further development in this field was the introduction of zeolite catalysts in 1964. In the early stages, inclusion of some zeolites in the catalysts gave improved performance with less undesirable side reactions and longer life. But the whole subject of zeolite catalysis has expanded enormously in recent years because of their shape-selective proper tie^.^ There are two main aspects of shape-selectivity. First, because of the sizes of the channels in the zeolites some classes of molecules have relatively easy access to the internal surfaces upon which the majority of the acidic sites are located-other larger molecules do not.Secondly, there can be steric control of the various possible reaction rates in that the volume of the cavities in the zeolites may enable reactions with a small transition state to occur more readily than others with a less compact activated complex. E. Partial Oxidation.-In contrast to the more complete oxidations discussed in Section B above there are many important applications of catalysts in selective o~idation.~,~The formation of ethylene oxide from ethylene is a well-known reaction using silver catalysts. The manufacture of phthalic anhydride originally from naphthalene but also from o-xylene is frequently accomplished by a supported vanadium pentoxide catalyst. Likewise, the oxidation of p-xylene to terephthalic acid is a step on the route to the polyester ‘Terylene’.Another chemical of value to the polymer industry, acrylonitrile, is produced by the ammoxidation of propene with ammonia and oxygen using tin-antimony oxides or other mixed- oxide catalysts such as bismuthphosphomolybdate. lo*’’ F. Ziegler-Natta Polymerization.-A great industrial development grew from the pioneering work of K. Ziegler and G. Natta who devised a remarkable class of catalysts. The combination of a transition metal (usually titanium) compound with a reactive organometallic compound, such as triethylaluminium, was shown to ’P. B. Weisz. ‘Proceedings of the 7th Int. Congr. on Catalysis’, ed. T. Seiyama and K. Tanabe, Elsevier, (Amsterdam, Oxford, New York), 1981, Vol.A, p. 3.’J. K. Dixon and J. E. Longfield, in ref 4. p. 183. R. Higgins and P. Hayden, in .Catalysis’, ed. C. Kemball (Specialist Periodical Reports), The Chemical Society, (London), 1977, Vol. 1. p. 168. lo G. W. Keulks, L. D. Krenzhe, and T. M. Notermann, Adv. Cutal., 1978, 27. 183. B. C. Gates, J. R. Katzer, and G. C. A. Schuit, ‘Chemistry of Catalytic Processes’, McGraw-Hill. (New York), 1979. p. 325. Chemical Processes on Heterogeneous Catalysts have the ability to polymerize alkenes and dienes with high activity and selectivity. Important commercial processes have been devised for the production of high-density polyethylene, isotactic polypropylene and other polymers based on dienes, or mixed alkenes.The great advance associated with the Ziegler-Natta catalysts was the high degree of stereospecificity associated with their G. Catalytic Reforming of Hydrocarbons.-While catalytic cracking of petroleum feed stocks provides products with the required range of molecular weights, further processing is necessary to satisfy the demand for (a) gasoline and (b) starting materials for the petrochemical industry. The types of reaction which must be catalysed are isomerization of alkanes to convert straight-chain to branched-chain molecules and dehydrocyclization to form aromatics-both of these reactions raise the octane rating of the products. Reactions which have to be avoided are the formation of coke on the catalysts and also the breakdown of the feed stock to smaller hydrocarbon molecules.The reforming reactions are normally accomplished on so-called 'dual function' catalyst^,'^ which have both a hydrogenation/dehydrogenation capability associated with a supported metal like platinum or rhodium and an isomerizing function associated with acidic oxides such as halided alumina. An excellent account of the evolution of reforming catalysts was given by Sterba and Haensel lS and the book l6 by Gates, Katzer, and Schuit contains a chapter which relates the practice of catalytic reforming to the scientific principles upon which it is based. In recent years, improved performance in reforming has been achieved by the use of bimetallic supported catalysts such as the platinum-iridium catalysts developed by Exxon '' or platinum-rhenium by Chevron.' H.Hydrodesu1phurization.-Many catalysts, particularly metals, are poisoned by relatively small percentages of sulphur compounds. The removal of such impurities from feed stocks is frequently accomplished" by the conversion of sulphur- containing organic compounds into hydrogen sulphide by use of catalysts like cobalt molybdate (which under working conditions becomes a mixed sulphide). The hydrogen sulphide may be removed by reaction with a zinc oxide catalyst, although this second stage is not really a true catalytic process because the zinc oxide is converted into sulphide which is then discarded. In a wider sense hydrodesulphurization is typical of a number of processes including purification of feed stocks or removal of substances which would otherwise lead to environmental problems.M. N. Berger, G. Boocock, and R. N. Hayward, Adv. Catal., 1969, 19, 21 1. 13 A. D. Caunt, in ref. 9, p. 234. '* P. B. Weisz, Adv. Catal., 1962, 13, 137. l5 M. J. Sterba and V. Haensel, Ind. Eng. Chem., Prod. Res. Dev., 1976, 15, 2.'' Ref. 11, p. 184. J. H. Sinfelt, 'Bimetallic Catalysts,' John Wiley and Sons, (New York), 1983. R. L. Jacobson, H. E. Kluksdahl, C. S. McLoy, and R. W. Davis, Proc. Am. Petroleum Inst., Div. of Refining, 1969, 49, 504. l9 J. J. Phillipson, 'Catalyst Handbook', Wolfe Scientific Books, (London), 1970. p. 46. 378 Kernball I. Steam Reforming of Hydrocarbons-The high-temperature reaction of hydrocarbons with an excess of steam, typically over a nickel-supported catalyst, has a number of uses.'' These include the production of hydrogen for ammonia synthesis, the formation of carbon monoxide and hydrogen for methanol synthesis and, under circumstances where natural gas is not readily available, the formation of methane-rich mixtures from higher hydrocarbons for use as 'town gas'.The temperature operation and the steam ratio (moles of H,O:moles of C) have to be chosen with regard to the desired products and the relevant thermodynamic equilibria. The water gas shift equilibrium CO + H,OeCO, + H, is also an important factor and carbon dioxide is a by-product of many steam reforming processes. Because of the rather vigorous conditions involving steam at the high pressures and temperatures under which the catalysts have to operate, various solid-solid and solid-gas reactions have to be taken into account in the design of catalysts with sufficient activity, strength, and life.J. The Wacker Process.-The single-step oxidation of ethylene to acetaldehyde is an important example of the applications of homogeneous catalysis on a commercial scale. The process was developed by Smidt and Hafner and their associates.21 The hydrocarbon is oxidized in aqueous solution by dissolved PdCl, as follows: CH,CH, + H,O + PdCI, -CH,CHO + Pd + 2HC1 Cupric chloride is also present and it brings about the reoxidation of the palladium by the reaction Pd + 2CuC1, -PdCI, + 2CuCI and the cycle is completed by the reoxidation of the cuprous chloride by dissolved oxygen 2CuCI + 0,+ 2HC1-2CuC1, + H20 The amounts of palladium and copper salts required are relatively small and, of course, the overall reaction amounts merely to the desired oxidation of ethylene to acetaldehyde, i.e.CH2CH2+ 0,-CH,CHO A full account of the process has been given by Stern.,, K. Synthesis of Methanol.-There are a variety 23 of catalytic processes which are operated with carbon monoxide and hydrogen as reactants, such as the Fischer- Tropsch synthesis,24 which can be used to produce long-chain hydrocarbons or 2o G. W. Bridger and G. C. Chinchen, re$ 19, p. 64. 21 J. Smidt, W. Hafner. R. Jira, J. Sedlmeier. R. Sieber. and H. Kojer. Angew. Chem.. 1959, 71, 176.22 E. W. Stem, Catal. Rev., 1967, 1, 73. 23 P. J. Denny and D. A. Whan, 'Catalysis', ed. C. Kemball and D. A. Dowden (Specialist Periodical Reports), The Chemical Society, (London), 1978, Vol. 2. p. 46. 24 H. H. Storch, N. Golumbic, and R. B. Anderson, 'The Fischer-Tropsch and Related Syntheses', Wiley. (New York), 1951. Chemical Processes on Heterogeneous Catalysts oxygenated compounds. But the synthesis of methanol is undoubtedly one of the most important of present-day catalytic processes. The reaction CO + 2H2 CH,OH is exothermic with AH = -91 kJ mol ' and so there is advantage in carrying out the conversion at relatively low temperatures in order to achieve good yield. At one stage the synthesis used to be carried out at temperatures above 300 "C using a zinc oxide-chromia catalyst but this process has been superseded by the use of copper- containing catalysts operated at 220 and 300 "C under pressures of 50 to 100 atm. The selectivity is typically in excess of 99%.L. Automotive Emission ControL-One of the greatest challenges in the catalytic field has been the requirement to control the amounts of carbon monoxide and oxide of nitrogen emitted in the exhaust gases from cars. The problem has been not only to achieve the specified low limits for these gases but also to design a catalysis system which could continue to operate successfully for long periods under the changing conditions associated with the normal use of a car. Accounts of work in this field have been given by Wei,25 and by Shelef et aLZ6The main groups of catalysts which have been found to be effective can be classified as base metal, platinum, or promoted platinum catalysts.Prominent amongst the last group is a rhodium-platinum catalyst developed by G. J. K. Acres and others at Johnson Matthey. 3 Metals Having completed the survey of selected major industrial applications of catalysis, we now consider some of the evidence about the nature of the chemical processes that actually take place on the surface of the catalysts. In this section some aspects of catalysis by metals are reviewed and in the following some recent work on mechanistic studies on oxides is discussed. A. Catalytic Activity.-The essential property of any catalyst is that it provides a new and less energetic reaction path for passage from reactants to products.For effective catalysis, the strength of adsorption of the molecules must be in the right range. This aspect is illustrated by the diagram in Figure 1 for the decomposition of formic acid on metals to yield hydrogen and carbon dioxide. It is assumed that the reaction proceeds through an adsorbed intermediate which is essentially a surface metal f~rmate.~' Curve A in Figure 1 shows what happens when adsorption is too weak; the reaction involves passage through some activated complexes of rather high energy. Curve B is the case where adsorption is too strong; the intermediate is formed without difficulty but a substantial energy is then required to decompose it ''J.Wei, Adv. Catal., 1975. 24. 57.*' H. Shelef, K. Otto, and N. C. Otto, Adv. Catal., 1978, 27, 31 I.*' J. Fahrenfort, L. L. van Reijen, and W. M. H. Sachtler. 'The Mechanism of Heterogeneous Catalysis .ed. J. H. de Boer. Elsevier, (Amsterdam, London, New York), 1960, p. 23. Kernball into products. Curve C represents the ideal situation; an intermediate strength of adsorption such that both the formation of the intermediate from reactants and its decomposition can occur without involving substantial energies of activation. A n HCt ENERGY + co, /JAdsorbed\ intermediate REACTION PATH M Figure 1 Reaction paths for the &composition offormic acid: .formate species adsorbed+ A)too weakly, (B) too strongly, (C) with intermediate strength These ideas about the likely variation of rate with strength of adsorption can be tested by examining the relative activity of a series of metal catalysts for the decomposition of formic acid.The typical volcano-shaped curve shown in Figure 2 (based on Figure 8 of ref: 27) illustrates the variation observed. The vertical scale gives the temperature required to achieve a fixed catalytic activity for the reaction, the lower this temperature the more active is the metal. The heat of formation of bulk metal formate is plotted along the horizontal axis and it is assumed that there will be a correlation between the strength of adsorption of formic acid as a surface formate and the heat of formation of the bulk compound.Gold and silver are poor catalysts because the strength of adsorption is too small; iron, cobalt, and nickel are poor catalysts because the adsorption is too strong. The greatest activity is found for the various metals near the top of the volcano because they exhibit the appropriate strength of adsorption. This interpretation of the volcano-shape is substantiated by kinetic measurements which suggest that formation of the adsorbed intermediate is rate-determining on gold and desorption of product rate- determining on nickel, i.e. cases A and B respectively of Figure 1. Another example illustrating the importance of strength of adsorption is the variation of activity of the first row transition metals for the synthesis of ammonia shown in Figure 3 (based on Figure 2.9 of the book by G.C. Bond 28). Metals to the left react too strongly with nitrogen, those to the right too weakly and the appropriate strength of adsorption is found with iron which is widely used as an ammonia synthesis catalyst. G. C. Bond. 'Heterogeneous Catalysis-Principles and Applications.' Oxford University Press. 1974. 38 1 Chemical Processes on Heterogeneous Catalysts 300 4 00 500 HEAT OF FORMATION /kJ equiv-’ Figure 2 Activity of metals for the decomposition of formic acid correlated with the heat of formation of bulk metal formates: T represents the temperature required to achieve the same jixed rate of reaction on each metal Activity for Ammonia Synthesis L 1 I 1 I 1 I 1 V Ti Cr Mn Fe Co Ni Figure 3 The variation in the rate of ammonia synthesis across the transition metals in thefirst row of the periodic table Kemball B.Exchange Reactions of Hydrocarbons with Deuterium.4ne of the most powerful methods of learning about the nature and reactivity of adsorbed intermediates on catalyst surfaces has been the study of exchange reactions, e.g. of hydrocarbons with deuterium. This technique has been a fruitful source of information about the types of dissociated species formed from saturated hydrocarbons on metals and the way in which these species are involved in the mechanism of hydrocarbon reaction^.^^.^' In general, the main features associated with hydrocarbon exchange reactions with deuterium are the following: (a) H atoms in the molecule are replaced with D atoms.(6) This may involve dissociation of C-H followed by the formation of C-D. (c) Alternatively, particularly with unsaturated molecules, formation of C-D is followed by dissociation of C-H. (d) The mechanism may be stepwise, involving a single replacement of H by D during adsorption on the catalyst, or multiple, with two or more atoms being exchanged before products desorb. (e) Adsorbed intermediates are formed reversibly. cf) Adsorption of reactant and desorption of product occur at the same rate. Frequently, on metal catalysts multiple exchange occurs and the study of suitably chosen reactants can be valuable for obtaining an indication about the types of adsorbed intermediates which are important. Some of these reactants with the reason for their study are listed in Table 1.The use of cyclopentane as an interesting test reactant has long been recognized-early work ” using evaporated films of palladium as catalysts demonstrated two main types of multiple exchange for this molecule. At lower temperatures, the distribution of initial products showed a pronounced maximum at the compound with five deuterium atoms, CsHsDs. The formation of this was attributed to the interconversion on the surface of adsorbed cyclopentyl radicals and adsorbed cyclopentene molecules leading to the replacement of all the H atoms on one ‘side’ of the ring. At high temperatures the initial distribution products gave a maximum for the perdeutero-compound, CsDlo, and the formation of this is now believed to involve the ‘turn-over’ on the surface of an adsorbed cyclopentene species 30-a process involving a rather higher activation energy than the mechanism responsible for the multiple exchange of one side of the ring.Recent work 32 using deuterium n.m.r. spectroscopy has confirmed that the Ds product is indeed the compound with five equivalent deuterium atoms, one on each carbon atom. Figure 4 shows deuterium n.m.r. spectrum of the mixture of deuterocyclopentanes formed at an early stage in the exchange of cyclopentane over a supported platinum catalyst. The positions of the resonances are dependent not only on the nature of the deuterium atom being observed but also on isotopic shifts due to the neighbouring deuterium atoms.In favourable circumstances 29 C. Kemball, Adv. Cufaf.,1959. 11, 223. ’O R. L. Burwell, Jr., Acc. Chem. Res.. 1969. 2, 289. 31 J. R. Anderson and C. Kernball. Proc. R. SOC. London, Ser. A, 1954,226,472.’* A. C. Faro, Jr., C. Kemball, R. Brown, and I. H.Sadler, J. Chem. Res.. 1982, (S) 342. (M) 3735. Chemical Processes on Heterogeneous Catalysts various groupings of deuterium atoms can be recognized and estimated. The percentages of the various products in this sample, determined by mass spectrometry, were 2.3% D,, 1.2% D2,0.6% of D, and D,, 1.7% of D,, and a total of 0.6% for D, to DIo.The products C,H,D, cis-l,2-D,-cyclopentane,cis-1,2,3,4,5-D,-cyclopentane and the perdeutero-compound, C,D, o, are easily recognized in the spectrum as the peaks labelled A-D respectively.TaMe 1 Model reactants for multiple exchange Reactant Information gained CH4 Role of (CH2) ads. C2H6 Extent of alkyl-alkene interconversion on the surface CYCIO-C~H Extent of the propagation of10 the exchange beyond C~HSDS CH3ICH3-C-CH3 Extent of the propagation ofI the exchange beyond one methyl CH3 group 1 I I I 1 I 1150 1145 6/p.p.m Figure 4 Deuterium n.m.r. spectrum (line-narrowed) for cyclopentanes formed after 7% exchange with deuterium over a 0.51% Pt-alumina catalyst: peaks A to D correspond to CSH,D, cis-1,2-C,H,D2, cis-1 ,2,3,4,5-C5H,D5,and C,D, respectively The main intermediates which are important for exchange reactions of hydrocarbons on metals are shown in Figure 5 which also illustrates the system of labelling commonly used to describe the intermediates: this is based on the position Kemball of the hydrogen atoms that have to be removed from the corresponding alkane to form each species.CR3 CR2 I II* * U * ff aa aaa 1 * * * aP7 Figure 5 The main adsorbed intermediates involved in the exchange of hydrocarbons on metals C. Hydrogenolysis Combined with Exchange.-Hydrogenolysis of hydrocarbons involves the breaking of C-C bonds with the formation of small saturated molecules. Reactions such as C2H6 + H2-2CH4 C,H, + H, C2H6 + CH, are thermodynamically very favourable at moderate temperatures and so it is perhaps surprising that there are so many hydrocarbon/metal systems for which it is possible to observe the exchange of the hydrocarbon with deuterium without some accompanying hydrogenolysis or isomerization of the reactant.The key factor is that C-H bonds are ‘activated’ (broken and made) more easily on many transition metal catalysts than are C-C bonds, despite the higher bond dissociation energy of the C-H bonds. But on most transition metals increase of temperature will lead to both isomerization and hydrogenolysis, the onset of these reactions implies that the species shown in Figure 5 are no longer formed reversibly as the temperature is raised 33 and either give rise to alterations in the carbon skeleton of the molecules or dissociate to other species which do so.Naturally, the degree of reversibility varies from one intermediate to another and the last species shown, the my-diadsorbed entity, is often found to lead on to isomerization or hydrogenolysis. Evidence from the initial products of reaction of 2,2-dimethylpropane with deuterium on evaporated metal films of iron34 at 213 “C has shown clearly the differing degrees of reversibility of the intermediates on that metal. The main 33 C. Kemball, Cutul. Rev., 1971, 5, 33. 34 R. S.Dowie, C. Kemball, J. C. Kempling, and D. A. Whan. Proc. R. Soc. London. Ser. A. 1972.327.491. Chemical Processes on Heterogeneous Catalysts intermediates expected to be formed from 2,2-dimethylpropane are shown in Figure 6. The initial products and the consequential deductions are as follows: Stepwise exchange gives (CH,),C(CH,D). This implies that intermediate (I) is formed reversibly.The only multiply-exchanged initial product is (CH,),C(CD,). There must be efficient interconversion between intermediates (I) and (11). It is, of course, possible that the formation of an aaa-triadsorbed species plays a part in the exchange reaction in addition to (I) and (11). No initial products have deuterium in two methyl groups. It follows that the intermediate (111) is not formed reversibly at 213 “C on iron. Hydrogenolysis gives the perdeuteromethane, CD4, as the only significant product. This confirms that the route to the breaking of C-C bonds lies through intermediate (111). Figure 6 Intermediates involved in reactions of 2,2-dimethylpropane On other metals, particularly platinum, species closely related to (111) are considered to be important for the isomerization of 2,2-dimethylpropane and it has been suggested that the formation of the intermediate may involve only one and not two or more surface atoms of the metal.,’ 4 Oxides As mentioned above, many of the most important non-oxidative reactions of hydrocarbons on oxides involve the use of ‘acidic’ oxides and these lead to the formation of carbocations (carbenium ions) as surface intermediates which play a part in a wide range of processes. But the application of exchange studies over the last 20 years or so has revealed that a great variety of species can be formed by the interaction of hydrocarbons on different oxide catalysts.Some of the earlier work in this field has been reviewed 36 but interesting results continue to appear. Alkenes are frequently used as test reactants in studies of the catalytic properties of oxides and they can undergo exchange, isomerization, or hydrogenation depending on the nature of the catalyst. A. Reactions of Ethylene.-With transition metal catalysts, the reaction of 35 J. K. A. Clarke and J. J. Rooney, Adv. Curd., 1976, 25. 125. 36 C. Kemball, Ann. N.Y. Acud. Sci., 1973, 213,90. Kemball ethylene with hydrogen on deuterium almost always gives rise to hydrogenation with the production of ethane, possibly accompanied by some alkene exchange as well. The situation is entirely different with oxides and the relative ratio of ethylene exchange:ethylene deuteration can vary by a factor of lo6 depending on the oxide ~elected.~'Magnesia gives exchange of the alkene without any trace of addition; conversely, with zinc oxide or chromia, ethane formation takes place without any observable exchange of the alkene and the sole product of the reaction is the ethane molecule, C2H4D2,formed by the addition of a molecule of deuterium to the alkene (although not necessarily in a single step).There is an interesting paradox to be explained when one considers a catalyst like 7-Al,O, and the reactions of C, hydrocarbons. This oxide is an efficient catalyst for the exchange of H, and D, and the rates of various reactions of the hydrocarbons are given in Table 2.The problem is to explain why the alumina which can exchange relatively well both ethylene and ethane should be so comparatively inactive as a catalyst for the thermodynamically favoured reaction-the deuteration of ethylene; this reaction occurs lo3 more slowly than the exchange of the alkene at 16 "C.A possible explanation is given in Scheme 1. There is considerable evidence in support of the idea that hydrocarbon exchange on alumina involves the formation of carbanionic intermediates with the hydrocarbon molecule essentially dissociating like an acid RH-R-+ Ht and the products of the dissociation being suitably bonded to the alumina surface. The relative reactivity for deuterium exchange of different hydrocarbons can be correlated with hydrocarbon acidity by means of typical Bronsted linear free energy relationship^.^^ So the point illustrated in Scheme 1 is that the reversible GAS '2 H6 (H+) H-1 I H*) Scheme 1 Possible mechanisms for the exchange of ethylene and ethane on alumina and for the slow hydrogenation Table 2 Rates of reaction on alumina at 16 "C Reaction Ratelmolecules s-l m-2 C2H4 exchange 5 x 1014 C2H6exchange 6 x 1OI2 Deuteration of CzH4 3 x 10" 37 C.Kemball, J. D. Nisbet, P. J. Robertson, and M. S.Scurrell, Pror. R. SOC.London. Srr. A. 1974.338.299. P. J. Robertson, M. S. Scurrell. and C. Kemball, J. Chem. SOC.,Faratfay Trans. 1. 1975. 71, 903. Chemical Processes on Heterogeneous Catalysts dissociation of both hydrocarbons occurs in a similar manner and these reactions, involving a loss of H+ or gain of D+ ,give rise to exchange. The two exchange mechanisms are separate and the link between them is slow and difficult, requiring the acquisition of a H -species by adsorbed ethylene. The suggestion is that oxides which tend to form charged hydrocarbon intermediates may be less efficient for alkene hydrogenation, whereas oxides upon which the hydrocarbon intermediates are less highly charged (or can be neutralized by electron transfer to or from the solid) behave more like metals and act as hydrogenation catalysts.Most of the hydrocarbon intermediates formed on metals are thought to be essentially uncharged-many of those on oxides, particularly polar oxides which are not semi- conducting, are likely to be charged.B. Propene Intermediates-Propene has proved to be a useful molecule for exploring the catalytic properties of oxide catalysts, partly because microwave spectroscopy3’ can be used to determine the position of deuterium atoms in the alkene molecule. The early products from the exchange of propene with deuterium, or with heavy water, can be analysed both by mass spectrometry and microwave spectroscopy to find out which hydrogen atom or atoms have been replaced and this in turn gives a clear indication of the nature of the adsorbed intermediates involved in the mechanism. Some of the results so obtained4’ are shown in Table 3 and a surprising variety of intermediates is involved.Additional information has come from the examination of the reactions of the labelled propene, CD2=CHCH,, on various oxides using both mass spectrometry and microwave spectro~copy.~~ With this molecule it is possible to follow double bond movement and at the same time to determine the extent of self-exchange (forming initially C,H,D and C,H,D,). In this way, the relative importance of Table 3 Intermediates for propene exchange on oxide catalysts Intermediate carbenium ion CHJ-CH-CH~ + Catalysts zeolites, Zr02 propen- I-yl CHKH-CHjI A1203 propen-2- yl CH2X-CH3I x-allyl CHz=CJH:CH_2 MgO, Ti02 f o-ally1 CHz-CH=CH2 I 39 K. Hirota. ‘Proc. 5th Int. Congr. Catalysis, ed. J. W. Hightower, North-Holland/Elsevier. 1973, p. c-37.40 B. T. Hughes, C. Kemball, and J. K. Tyler, J. Chem. Soc., Furuduy Trans.I, 1975, 71, 1285. 41 C. S. John, C. Kemball, R. Dickinson, and J. K. Tyler, J. Chem. Sor., Furuduy Trans.I, 1976,72. 1782. Kemball intramolecular and intermolecular movement of hydrogen and deuterium atoms can be measured and analysis of the products will also reveal42 whether the double bond movement occurs by a dissociative process (through ally1 species) or by an associative mechanism (through propyl speciesba question which is not easily answered by other methods. C. Reactions of Cyc1opentene.-Like propene, this alkene has proved to be a useful model reactant for mechanistic studies on oxide catalysts. By appropriate techniques, which now include deuterium n.m.r.spectroscopy, it is possible to determine which of the hydrogen atoms are most readily exchanged, whether double bond movement occurs, and the extent of hydrogenation. One of the early exciting results was the discovery by Hightower and Hall43 that exchange with deuterium over alumina at temperatures below 100°C was limited to the replacement of the two olefinic hydrogen atoms without any double bond movement. The mechanism proposed was the reversible dissociation of the alkene to form cyclopentenyl species on the oxide surface and the existence of a kinetic isotope effect (dissociation of C-H being 2-3 times faster than dissociation of C-D) supported this conclusion. Recent results with cyclopentene as a reactant over zinc oxide as a catalyst44 illustrate the type of information that can be obtained in favourable circumstances.As with alumina, the exchange reaction was limited almost entirely to replacement of the two olefinic hydrogen atoms in a stepwise process. The peaks in the deuterium n.m.r. spectrum of exchanged cyclopentene for olefinic D (6 = 5.78 p.p.m.) and allylic D (6 = 2.31 p.p.m.) had a ratio of intensity of 50: 1. But in contrast to alumina, a slow addition reaction occurred over zinc oxide at about l/lOth of the rate of the exchange reaction. The initial product of this reaction was C5H,D2 but subsequently products containing 1, 3, or 4 deuterium atoms were formed. Scheme 2 shows the main reactions involved. The exchange process leads to the replacement successively of both the olefinic hydrogen atoms but without double bond movement. The addition reaction takes place without further exchange so the initial product is 1,2-dideuterocyclopentane.Subsequently, as addition takes place to exchanged cyclopentenes and as the D2 gas becomes diluted with some HD, the other products are formed but the presence of D atoms is limited to the two adjacent carbon atoms which were associated with the double bond in the alkene. The deuterium n.m.r. spectrum of the cyclopentanes observed after 13% of the cyclopentene had been saturated is shown in Figure 7. Table 4 indicates the main isotopic compounds which contribute to this spectrum and the expected position of the peaks based on subsequently determined isotope shifts3’ for the various groupings of deuterium atoms.The two prominent peaks in the spectrum correspond to the initial product, 1,2-dideuterocyclopentane which has two equivalent D atoms, and the most highly deuterated product, the 1,1,2,2-tetradeutero species which has four equivalent D atoms. Thus, the main reactions 42 C. S. John, C. E. Marsden, and R. Dickinson, J. Chem. SOC.,Faraday Trans. I, 1976.72, 2923. *’ J. W. Hightower and W. K. Hall, Trans. Faraday Sor., 1970, 66,477. 44 R. Brown, C. Kemball, and D. Taylor, J. Chem. Res.. 1982, (S) 223, (M) 2329. 389 Chemical Processes on Heterogeneous Catalysts \/ \/ \/ _____+CH=CH CD=CH ---+CD=CDA; \ CHD-CH2 CHD-CHD CDZ-CCH, CD2-CHD CD2-CCD2 Scheme 2 The main reactions of cyclopentene and deuterium on zinc oxide I 1 1 1 I I I 1.50 1.45 6 Ip.p.m. Figure 7 Deuterium n.m.r. spectrum (line-narrowed) for the cyclopentanes formed after 13% conversion of cyclopentene into alkane over zinc oxide. Peak A corresponds to CsHgD, Peak B to cis-1 ,Zdideuterocyclopentane and Peak E to 1,1,2,2-tetradeuterocyclopentane;the remaining peaks correspond to the other compounds listed in Table 4 Table 4 Position of n.m.r. peaks for deuterocyclopentanes formed over ZnO Calculated Observed Compound G1p.p.m. Y/P.P.m* \ / -1.501iHDT2 CHD-CHD a 1.494 1.494 I\ CD2-CHz 1.483 1 1.482 \I CD2-CHD 1.486, 1.476, 1.475 ) 1.476 \/CDz-CD2 1.468 1.468 this is the initial product KernbaIl which occur with cyclopentene on zinc oxide on 80 "Care the reversible formation of the adsorbed cyclopentenyl species and, more slowly, the formation of ad- sorbed cyclopentyl species which are then rapidly desorbed as cyclopentane.The addition reaction of cyclopentene is similar to the corresponding reaction of ethyl- ene with deuterium which also leads to a 1,Zdideuteroalkane as the sole produ~t.~' D. Isomerization of Alkenes on Alumina.-Pretreatment has a very important influence on the activity of most oxide catalysts and also on the nature of the intermediates involved in the reactions. Increasing temperature of pretreatment reduces the surface population of hydroxy-groups and can lead to the formation of active sites by removal of some of the surface oxygen as water molecules. Surface acidity may be increased by suitable treatment such as the use of halides on alumina or even addition of hydrogen sulphide.The results quoted above for the exchange of propene on alumina have indicated that different types of catalytic processes occur on this oxide. The same complexity has been revealed in relation to the isomerization of alkenes and this topic can be used as an illustration of some of the different types of mechanism which can be found. Some of the earlier work was reviewed by John and Scurrel146 in 1977 but the account here brings in results of more recent ' The catalytic properties of alumina are complex, but three main types of mechanism have been identified for alkene isomerization.(a) The easiest type of reaction occurs with alkenes which can isomerize by formation of a tertiary carbenium ion, R'R2R3C+,and examples include: (i) the conversion of 2,3-dimethylbut- 1-ene into 2,3-dimethylbut-2-ene7 (ii) double bond shift in the labelled isobutene, (CH,),C=CD,, (iii) the conversion of methylenecyclopentane into 1-methylcyclopentene. These processes often take place readily at temperatures around 0 "Cand little increase in rate occurs if the catalyst is pretreated with hydrogen sulphide. (b) The next type of reaction occurs with alkenes which can isomerize by formation of an adsorbed n-ally1 intermediate and examples are: (i) the interconversion of but-1-ene and the but-2-enes, (ii) the conversion of 3-methylcyclopentene into 1-methylcyclopentene. Treatment of the catalysts with hydrogen sulphide, which is known to be a poison 52 for reactions on alumina involving n-ally1 species, typically reduces these rates of isomerization by a factor of 10 or more.45 R. J. Kokes and A. L. Dent, Adv. Carol., 1972, 22. 1. 46 C. S. John and M. S. Scurrell. 'Catalysis,'ed. C. Kernball, (Specialist Periodical Reports). The Chemical Society, (London), 1977, Vol. 1, p. 136. 47 C. S. John and R. Dickinson, J. Chem. Res.. 1977, (S) 88. (M) 1020. 48 C. S. John, A. Tada, and L. V. F. Kennedy, J. Chem. Soc., Faradax Trans. I, 1978.74.498. 49 C. S. John, C. Kernball, and R. A. Rajadhyaksha. J. Caral., 1979, 57, 264. E. A. Irvine, C. S. John, C.Kernball, A. J. Pearman, M. A. Day, and R. J. Sarnpson. J. Caral.. 1980. 61, 326. 5' C. S. John, C. Kernball, R. C. Patterson,and R. A. Rajadhyaksha,J. Chem.SOC.,Chem. Commun. 1977. 894.'' M. P. Rosynek and F. L. Strey, J. Catal., 1976,41, 312. Chemical Processes on Heterogeneous Catalysts (c) A more difficult and slower type of isomerization is exemplified by the reaction of 3,3-dimethylbut- 1-ene to the 2,3-dimethylbutenes. For this process it is generally accepted that the mechanism must involve the formation of a secondary carbenium ion which rearranges by methyl shift to the more stable tertiary carbenium ion leading to the product alkenes, the presence of the quaternary carbon atom in the reactant eliminates the possibility of isomerization through a x-ally1 species.Treatment of the alumina with hydrogen sulphide enhances the rate of reaction by a small factor of 2-3 but the presence of a small percentage of fluoride on the alumina increases the rate by factors of 102-103. Increase of the pretreatment temperature of the alumina (up to 800 "C)gives rise to increased rates of reaction for all three kinds of alkene isomerization, although there are minor differences in the size of the effects. At first sight it is perhaps surprising that types (a) and (c) involving adsorbed carbenium ions do not show some evidence for a maximum as the surface becomes dehydroxylated because of the need to acquire a proton to convert the alkene into the corresponding carbocation. A possible explanation may be that there are two potential mechanisms for carbenium ion formation.The first involves the usual proton addition, e.g. + (CH~)~C-CH=CHZ+ H+ --+ (CH~)JC-CH-CHJ whereas the second is Lewis acid induced + (CHJ)~C-CH=CH~+ L ---+ (CHJ)~C-CH-CH~-L-These examples of isomerization on alumina illustrate clearly the variety of ways in which the catalyst may operate through a range of different intermediates. The reaction path is influenced by a number of factors, not least of which is the nature of the reactant itself. 5 Conclusions The catalytic processes which have been developed for the industrial applications in the present century are extensive, both in number and variety, and of the very greatest importance. The majority of these are examples of heterogeneous catalysis but homogeneous processes are becoming more significant as knowledge progresses.Selectivity is sometimes almost more important than activity and even marginal increases in efficiency of large scale processes can bring substantial economic advantages. The review presented here has concentrated almost entirely on the chemical aspects of catalysis with emphasis on the increasing knowledge that is now available about the character and reactivity of the adsorbed intermediates on metals and oxides. The behaviour of these intermediates and the part that they contribute to the mechanisms is becoming much more clearly understood. But the nature of the catalytic sites and the way in which they are related to the surface composition and structure of catalysts continue to present some formidable problems for fundamental and applied research.

ISSN:0306-0012    

DOI:10.1039/CS9841300375

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

Radical Cations in Condensed Phases By Martyn C. R. Symons DEPARTMENT OF CHEMISTRY, THE UNIVERSITY, LEICESTER, LEI 7RH 1 Introduction This Review was inspired by a series of recent e.s.r. studies of radical cations formed by radiolysis of rigid solutions at low temperature. Many novel cations have been prepared by this technique, and much new structural information has been forthcoming. The work is complementary to similar e.s.r. studies of radical anions in rigid solutions, also prepared by ionizing radiation. Curiously, techniques for preparing these anions were developed some years before the development of methods for preparing the cations, whereas in other branches of Chemistry, far more is known about radical cations than radical anions. Most of this information stems from radical cations generated in the gas phase, the two most significant techniques being mass spectroscopy and photoelectron spectroscopy (PES).The latter technique generates information on the ‘hot’ cations formed by ‘vertical’ electron loss, and vibrational features in the spectra give information on the modes involved in relaxation to the ground-state configuration of the cation. In contrast, the condensed-phase studies give information on the long-term ground-states of these cations or their rearranged products. Very recently there have been exciting new developments in the study of cations in the gas phase, which promise to give very detailed rotational and vibrational spectroscopic inf~rmation.~.~ Space limitations are such that I cannot cover the solid-state work adequately if I give room to these other areas.Indeed, they are well covered in a major new book,’ which does not in fact deal with the topics discussed in the present review. Gas-phase results, particularly by PES have inspired very extensive theoretical studies of radical cations. In most cases these are necessary for a proper interpretation of the PES spectra. Thus the e.s.r. spectroscopist studying radical cations is entering an already well trodden field. Nevertheless, the detailed information provided by this technique about the SOMO in particular extends and sharpens our knowledge. My major aim is to bring this new information to more general notice. The preparative procedure with which I am primarily concerned involves the use of ionizing radiation and low-temperature matrixes.To those who are interested in radical cations but unfamiliar with these procedures, let me stress that they are really quite simple tools which are utilised because they do an efficient job. ’ M. C. R. Symons, Pure Appl. Chem., 1981,53,223. ‘Molecular Ions: Spectroscopj., Structure and Chemistry, ed. T. A. Miller and V. E. Bondybey. North-Holland, Amsterdam. 1983.’M. I. Lester, B. R. Zegarski, and T. A. Miller, J. Phys. Chem., 1983,87, 5228. Radical Cations in Condensed Phases 2 Preparative Procedures Only relatively stable radical-cations have been studied in the liquid-phase by e.s.r. spectroscopy. These are usually prepared by conventional techniques including redox chemistry, electrolysis, or photolysis.For example, various aromatic cations were prepared in sulphuric acid or related acidic media many years ago.46 Norman, Gilbert, and their co-workers have used the Ti"'-H,O, system as a source of .OH radicals which will sometimes extract electrons from good electron donors, either directly or indirectly. The resulting radical-cations may not be detected directly by e.s.r. spectroscopy,' but secondary products, such as (Me,S)2 +,8 formed from Me,S-+ cations, may be sufficiently stable for e.s.r. study. Similarly, electrolysis has been used most effectively by Roberts and his co-workers to prepare (R3P)*+ cations, formed from the undetectable R,P*+ ions.9 To leave room for recent solid-state developments, I do not dwell on these procedures here.Two different techniques are used for solid-state studies. One, recently developed by Knight and his co-workers," l1 involves photoionization of the substrate in the gas-phase at the site of deposition on a rare-gas matrix at ca. 4K.The fate of the ejected electrons in these experiments is not clear, but good e.s.r. spectra for a range of radical-cations have been forthcoming. It seems that this procedure is very good for preparing small cations such as H,O+, NH,', and CH,', but probably less satisfactory for preparing larger cations. It may therefore prove to be complementary to the second technique described below. The second technique utilises X-or y-rays rather than ultraviolet light, and the solid matrix is irradiated rather than substrate molecules in the gas-phase. In some cases, it is satisfactory simply to irradiate the pure compounds as crystalline solids.Ideally, this gives electron-loss and electron-gain centres well separated from each other, examples being the formation of NO,. and -NO, ' from NO, in nitrates of [Mn(CO),hal]+ and [Mn(CO),hal] from Mn(CO),hal crystals.' However, in other cases, electron-return dominates, with consequent 'photolysis' as the main source of radicals, and in many others the electron-gain and -loss centres react with neighbouring molecules. Thus the majority of radical cations described below are not detected when the pure substrates are irradiated. Some years ago, we helped to develop a technique for preparing specific electron-gain or electron-loss centres from ionic substrates using doped salts.Thus, for example, CaCO, doped with NO,-ions gave -NO,'-ions,', or CaSO, doped with Po,, gave PO,' .14 This technique has the great advantage that single S. I. Weissman, E. de Boer, and J. Conradi, J. Chem. Ph-vs., 1956, 26, 963.'R. M. Dessau, S. Shih, and E. I. Heiba, J. Am. Chem. SOC.,1970,92,412.'R. Hulme and M. C. R. Symons. J. Chem. SOC.(A), 1966,446.' B. C. Gilbert, R. 0.C. Norman, and P. S. Williams, J. Chem. SOC.,Perkin Trans. 2, 1980, 647. * B. C. Gilbert, D. K. C. Hodgeman, and R. 0.C. Norman, J. Chem. SOC.,Perkin Trans. 2, 1973, 1748. 'W. B. Card and B. P.Roberts, J. Chem. SOC.,Chem. Commun., 1975,949.L. B. Knight, J. M. Bostick, R.W. Woodward, and J. Steadman, J. Chem. Phys., 1983, 78, 6415. " L. B. Knight and J. Steadman, J. Chem. Phys., 1983, 77, 1750. 12 0.P.Anderson, S.A. Fieldhouse, C. E. Forbes, and M. C. R. Symons,J. Chem.SOC.,Dalton Trans., 1976, 1329. l3 R.S. Eachus and M. C. R. Symons, J. Chem. SOC.(A),1968,790. 14 S. Subramanian, M. C. R. Symons, and H. W. Wardale, J. Chem. SOC.(A), 1970, 1239. 394 Symons crystals can be used. It has been extended recently by Morton and Preston and their co-workers,' 5*16 and Ammeter et al.' to the use of Cr(C0)6 single crystals doped with neutral molecules.' 5*16 Here neutral rather than charged substrates can be used, but there seems to be some uncertainty as to whether a given solute will form its cation or its anion on irradiation.To what extent this approach can be extended beyond the field of transition-metal carbonyls is not yet clear. The technique highlighted in this Review, however, is one in which certain solvents, notably fluorotrichloromethane, CFCI,, (freon) are used. The key to understanding their utility lies in equations (1 )-(4). CFCI, (CFCI,)' + e-(1) CF,CI, + e--(CFCI,)--C1-+ CFCI, (2) (CFCI,)' + CFCI, -CFCI, + (CFCI,)' (3) (CFCI,)' + S -CFCI, + S' (4) Ejected electrons (1) are rapidly scavenged by the solvent molecules (2), but solvent cations are mobile via electron transfer (3) Imtil they meet a solute molecule, S. Provided the ionization potential of S is less than that for CFCI, (ca.11.9 eV) reaction proceeds. However, as shown later, sometimes there is weak bonding between one solvent molecule and S, the 'hole' being shared between them (C1,FCCl---S)+.In a few cases, it has been possible to prepare cations from substrates with ionization potentials > 11.9 eV by using solvents such as SF6 or C2F6. However, no-one has yet prepared small protic cations such as H20+ or CH,+ by this method. Before proceeding to the heart of this Review, which is primarily concerned with the results obtained from freon and related solvents, I should mention another method developed some years' ago, namely the use of sulphuric or phosphoric acids as solvents for the radiolytic generation of radical cations. The basis of the method is summarized in equations (5)--(7).HSO,--L HSO,. + e-HSO,. + S -HSO,-+ S+ or HSO, + SH' -H,SO, + S+ Generally, reaction only occurs on annealing, the HSO, radicals being rigidly trapped at 77 K. In some cases, electron transfer occurs (6) and, in others, hydrogen atom transfer seems to be favoured (7). This technique is of less value than the 'freon' method, and the majority of cations prepared in freon and related solvents do not seem to form in sulphuric (or phosphoric) acid. However, it has the advantage that compounds soluble in these acids are sometimes either insoluble in freon, or dissolve as dimers or as aggregates therein. Since it is essential to have only monomers in solution, this makes the acid media advantageous at times. l5 J.A. Howard, J. R. Morton, and K. F. Preston, Chem. Phys. Lett.. 1981, 83, 226. T. Lionel, J. R. Morton, and K. F. Preston, J. Chem. Phys., 1982, 76, 234. J. H. Ammeter, L. Zoller, J. Bachmann, P. Baltzer, E. Gamp, R. Bucher, and E. Deiss, Helv. Chim.Am, 1981.64, 1063. Radical Cations in Condensed Phases Examples of the successful generation of radical cations in acidic glasses include R,N-+ from R,NH ions,'* R,P-+ from R,PH+ uracil cations, pro- + bably from the parent molecules,21 and (H2CO)+cations from H2C0.22Also, the x-cations (RSSR)' are formed in good yield by this procedure,23 as can be judged from Figure 1, but this is not a good general procedure for preparing radical cations. 1 3230G (9.130 GHz) 1 2 Figure I First derivative X-band e.s.r.spectrum from dimethyl disulphide in 6M-D2S04-D20 after exposure to 6oCoy-rays at 77 K, and annealing until features for SO4*-and 40,-were lost showing features assigned to MeS-SMe' radicals. [From Rex 231 This shows that good radical cation spectra have been obtained in early studies. It also illustrates the complexities of anisotropic solid-state e.s.r. spectra In the following sections, some of the results so far obtained are described. Analysis of the e.s.r. spectra is not easy, as can be judged from that in Figure 1, but since there is no room here for detailed explanations, I am afraid that published analyses have to be accepted herein. In M. C. R. Symons, J. Chem. Soc.. Perkin Trans. 2. 1973,797. l9 G. W. Eastland and M.C. R. Symons. J. Chem. Soc.. Perkin Trans. 2, 1977.833. 2o A. Begum, A. R. Lyons, and M. C. R. Symons. J. Chem. SOC.(A), 1971,2290. H. Riederer and J. Huttermann, J. fhys. Chem., 1982,86, 3454. S. P. Mishra and M. C. R. Symons. J. Chem. Soc.. Chem. Commun.. 1975,909. 23 R. L. Petersen. D. J. Nelson, and M. C. R. Symons, J. Chem. Soc., Perkin Trans. 2, 1978, 225. Symons 3 Alkane Cations The first alkane cation to be studied by e.s.r. spectroscopy was the hexamethylethane cation (Me3CCMe,)+.24 Hexamethylethane was selected for this study with some care, since there had been many previous unsuccessful attempts to study alkane cations by e.s.r. spectroscopy. It was thought that the C2H6+ cation would probably lose a proton to almost any medium to give C2H50, and that such proton loss would normally be rapid even at low temperatures.Also, because the methyl groups are equivalent, the e.s.r. spectrum for Me,CCMe3 + should be relatively simple in comparison with most other alkane cations. It was predicted that the ‘hole’ would be considerably confined to the central C-C o-bond which would therefore stretch, with concomitant partial flattening of the two -CMe, units, and that H coupling would arise via hyperconjugation. These conclusions have, after some criticism and scepticism, been largely accepted and confirmed. However, a1 though proton loss is indeed facile, especially for C2H6+,25 it does not occur readily at very low temperatures in fluorinated solvents. Furthermore, because many of the protons in alkane cations interact very weakly with the unpaired electron, the e.s.r.spectra for alkane cations turn out to be far simpler than had been supposed. Following the publication of our study of (Me,CCMe,)+ cations, Iwasaki and his co-workers initiated a very thorough and most perceptive study of a wide range of alkane cations, and it is mainly their work that I describe in the f~llowing.~~,~~ The results are in marked contrast with those for alkyl radicals, for which, to a first approximation, conventional, localized structures (SOMOs) suffice. They show that the unpaired electrons are delocalized in specific MOs, throughout the molecular frame in the same way as they are in normal n-radicals. This may come as a surprise to those of us who were reared on localized o-bonds, but has long been predicted by MO theory.However, in many cases, there are a range of possible SOMOs and the choice made in these freon systems is often a subtle one. Probably the original conformation of the molecular precursor dictates the selection, but this is also governed by the most favourable mode of relaxation (bond bending or stretching) which may well be influenced by the surrounding solvent molecules. The Methane Cation.-Attempts to prepare this by the technique of radiolysis have not succeeded, but Knight and his co-workers have been successful using their photolysis procedure described above.27 They found that the four protons were apparently equivalent even at CQ.4K [A(’H) = 54.8G1,but for CH2D2+ cations the results clearly establish that this is due to rapid averaging; the coupling to the two hydrogens being 121.2 G. Thus the static structure has C,,symmetry, as found for the isoelectronic -BH, radi~al,~” with two hydrogen atoms participating in the 24 I. G. Smith and M. C. R. Symons,J. Chem. Res. (S),1979,382; M.C. R. Symons. Chem. Phrs. Lett. 1980, 69, 198. ” K. Toriyama. K. Nunome, and M. Iwasaki. J. Chem. Php., 1982, 77, 5891. 26 K. Nunome, K. Toriyama, and M. Iwasaki, J. Chem. Phys.. 1983, 79, 2499. 27 L. B. Knight, J. Steadman, D. Feller, and E. R. Ddvidson, J. Am. Chem. SOL-.,1984, in press. 27 (a)M. C. R. Symons, T. Chen, and C. Glidewell, J. Chem. Soc.. Chem. Commun., 1983, 326.Radical Cations in Condensed Phases SOMO. This major distortion involves the movement of two hydrogen atoms towards each other, with the unpaired electron (or 'hole') shared between them and the carbon atom, and the movement of the other two hydrogen atoms away from each other within the nodal plane of the electron. It is most interesting that this is still in the fast averaging regime for CH4+ at ca. 4K, but that the isotope effect completely selects one of the distorted structures for -CH,D,+. The Ethane Cation.-This cation is isoelectronic with BzH6-9 and these radicals have the choice of two alternative structures, one which is related to the bridged B2H6 molecule, with two unique hydrogen atoms, and the other, with all hydrogen atoms equivalent, related to C2H,.It is remarkable that the structures have 'inverted', with B2H6- having the structure (l)28.29and C,H6+ the structure (2). The choice is clearly subtle, and theory has been undecided. Thus, a thorough ab initio study suggested that the 3a,, structure with a stretched C-C bond, as found for B2H6 ,should be fa~oured.~' Our own ab initio calculations on B2H6- and C2H, also favoured this di~tortion.~~ + However, the photoelectron spectrum indicated preferential loss from the leg orbital.,' This orbital is degenerate and a Jahn-Teller distortion is expected, lowering the symmetry to C,,, and giving the two orbitals 4a, and lb,. The e.s.r. spectrum, comprising a well defined triplet at 4.2 K (Figure 2a,b),leaves no doubt that the SOMO is the 40, orbital and that the Jahn-Teller distortion is fixed.On warming to 77K a seven-line spectrum is formed reversibly (Figure 2b). Since free relative rotation of the methyl groups is not expected, Toriyama et al. suggest that the distortion, which involves the inward movement of the two in-plane hydrogen atoms (towards the bridging structure), has become dynamic. The fact that the overall splitting remains unchanged shows that the coupling to the remaining four protons is very small indeed. These results have been confirmed using CH,CD, and CD,CD,. (There are interesting isotope effects, but these details do not alter the broad conclusions outlined herein.) Ignoring the Jahn-Teller requirement, it is clear that there are two alternative distortions available to the cation -one being primarily a bond-stretching with some flattening of the -CH, units, and the other being primarily the bending of two 28 P.H. Kasai and D. McLeod, J. Chern. Phys., 1969.51, 1250. 29 T. A. Claxton, R. E. Overill, and M. C. R. Symons, Mol. Phys.. 1974. 27. 701. 'O A. Richartz, R. J. Buenker, P. J. Bruna, and S. D. Peyerimkoff. Mo/. Phys., 1977. 33. 1345.'' J. W. Rabdlais and A. Katrib, Mu/.fhys., 1974, 27, 923. Symons a) t I I 1 i 1 1 +3 +2 +1 0 -1 -2 -3 Figure 2 E.s.r. spectra of C,H,+ (a) in SF, irradiated and measured at 4.2 K; (h) in SF, irradiated at 4.2 K and measured at 77 K.(From Re6 25) bonds, probably also with slight C-H stretching. Just such alternatives have been found in other systems on electron-addition.The Propane Cation.-Again, there are two predictable structures (3) and (4). The former (3) is in a sense a rr-structure, whilst the latter is the expected extension of H\/\ c/H Radical Cations in Condensed Phases the C,H,+ structure. Both have been detected, the n-structure in CFCI,CF,CI and the o-structure in SF,. This extraordinary matrix effect is not fully understood as yet. The 26, (II)structure is of great interest since it represents a sort of hyperconjugation limit. It can be compared, for example, with the (R-O-CH,O-R)+ structure considered below, or with the well known cyclohexadienyl structure, with respect to the central CH, group. This acquires a high spin-density in all cases, as indicated by the proton hyperfine coupling constants. For the propane cation this is very large for the central CH, protons (105.5G) which, as predicted by simple theory is ca.twice that for the four outer protons (52.5G). Other Normal Alkane Cations.-In general, a-delocalized structures (ethane type) are favoured, the extended structures H-(CH,),-H having e.s.r. spectra which are dominated by a triplet splitting from the two outermost in-plane protons. A typical orbital is shown in (5) for (n-butane)+. In some cases, outer features were also detected which were assigned to gauche structures and there are again marked solvent effects on their relative concentrations. The reader is referred to the original paper for more details.25 Branched-chain Alkane Cations.-A structural switch occurs for branched-chain alkane cations,,, the SOMO being localized quite strongly in a stretched C-X bond (X = H or alkyl), as in the alternative structure discussed for C,H,+.The scales may be tipped by hyperconjugative stabilization which tends to distribute the positive charge amongst several groups. For example, for the cations (Me,C*H)+, (Me,CCH,)+ and (Me,CCMe,)+ the Me,C group is thought to flatten considerably, thereby maximizing hyperconjugation and stabilizing the incipient carbocation Me,C+. The greater this stabilization, the more the spin-density will move onto the other group. The cation (Me,C*H)+ is remarkable in having the largest proton hyperfine coupling of any known organic radical (251 G).This result shows that there is a remarkably uniform distribution of the electron in this bond.Taking the extensive delocalization onto three of the methyl protons (48.8 G) into account, this shows that, within the C-H bond, the electron distribution indeed favours H. The hyperconjugative coupling falls to ca. 40G for (Me,GCH,) + which may mean that the valence bond structure (Me,C+ ---CH,) is even more favoured in this case. However, for (Me,CCMe,)+ this coupling has fallen to 29 G although the spin-density must now be 50% on each Me&- group. This can be explained in terms of a reduced flattening of the Me,C- units caused by steric repulsions between the six methyl groups. Symons Some interesting trends in the hyperfine coupling constants for trans C-HB protons for methylated propane and butane cations are shown in Figure 3,26 and results for a selection of n-and branched-chain alkane cations are given in Table 1.Thermal Reactions of Alkane Cations.-One or more bonds are greatly weakened by electron loss, and there is unit positive charge within the radical. The former effect must encourage the breaking of the weakest bonds, especially in those cases in which one bond is mainly involved, as with the branched-chain alkane cations. The latter should favour unimolecular breakdown to give the most stable a I I \ \ \ \ \ \ \ X \ a I 1 I 1V 0 2 4 ti “CH,) Figure 3 The dependence of the trans C-H, proton coupling constants on the number of CH3 groups.0 methyl-substituted propane radical cations; 0methyl-substituted butane radical cations; A neutral alkyl x radicals: and x the carbon number dependence of the in-plane end C-H proton coupling constants in n-alkane radical cations. (From Ref. 26) 401 Radical Cations in Condensed Phases Table 1 The observed hyperfine coupling constants of alkane cations Cations Matrices Temp.lK H.F. coupling constants/G 152.5 + 0.5 (2H) 50.3 f0.5 (6H) 153.1 f0.5 (2H) 140.5 k 1.0 (1H); 9.0 f 1.0 (2H) 21.9 f0.5 (1D) 23.0 f 1.0 (2D) 98.0 f0.5 (2H) 95.0 f0.5 (2H) 105.5 f 1.0 (2H); 52.5 f 1.0 (4H) 100.0 f 1.0 (2H); 52.0 f 1.0 (4H) 98.8 f0.5 (2H) 14.8 f0.2 (2D) CFCI2CF2CI 77 61.3 f0.5 (2H) n-GFI o 77 63.0 f0.5 (2H); 8.0 f0.2 (4H) CFClt CFzCl 77 41.0 f0.5 (2H) CFC13 77 44.0 f0.5 (2H); 4.1 f0.2 (8H) SF.5 4 58.0 f0.5 (2H) CFC12CF2CI 4 52.5 f 1.0 (2H) CFC13 4 55.0 f 1.0 (2H); 15.0 1.0 (1H) CFCl2CF 2C1 77 250.0 f 1.0 (1H); 47.5 2 1.0(3H) CFClj 77 251.2 f 1.0; (IH); 48.8 f 1.0 (3H) SF6 4 57.8 f0.2 (2H) CFClj 4 54.2 f0.5 (2H) SF6 4 unresolved single-line (AHpp= 27.0) CFC13 4 8.5 f0.5 (2D); 18.0 f0.5 (1H) SF6 4 40.0 f 1.0 (3H) CFCl2CFrCl CFCI, 4 4 39.8 f0.5 (3H) 40.7 + 0.5 (3H) CFC12CF2Cl 77 29.0 f0.5 (6H) CFC13 77 28.8 f0.5 (6H); 3.8 f0.2 (12H) carbocation, or, in the presence of a proton acceptor (B), proton transfer should occur.For example, for (Me,C-CH,)+ this would lead to (8) or (9).The former is (Me3C-CH3)+-Me$+.+ -CH3 (Me3CCH3)++ B -Me,C-CH2 + BH' not observed. Instead, in the 120-140 K range reaction (10) is detected. I am tempted to suggest that the thermolysis (8) does occur, but that in these solids a (Me3C-CH3)+-(Me2C=CH2)++ CH4 (10) very efficient cage-back reaction leads to hydrogen abstraction giving the products in (10). One major reason why these studies of alkane cations have been successful is the very low basicity of molecules such as SF, and CFCl,. I suspect that CH,+ is so Symons strong an acid that it can protonate these solvent molecules but evidently C2H6+ cannot. Nevertheless, an efficient bimolecular proton-transfer can occur, which is presumably transfer to parent alkane molecules (1 1).Such reactions are well known in the gas-phase, but it is a surprise that alkanes, having no obvious site for RH+ + RH-R* + RH2+ (1 1) protonation, are more basic than say CFCI,, for which protonation can be readily formulated! In studies of such proton-transfers the possible r6le of basic impurities such as water must always be borne in mind. For solvents SF, and CF,CI-CFCI,, proton transfer (9) occurs before the unimolecular decompositions set in, but for the more rigid CFCI,, these decompositions occur in preference to proton transfer. Photolysis of Alkane Cations.-It seems probable that most alkane cations have absorption bands in the visible and near i.r. region.,, Photolysis within these bands for a range of n-alkanes seems to result in the specific formation of but-2-ene cations with the elimination of the corresponding alkane 33 (12).Thus, for example, the n-hexane cation gave the (2-C4H,)+ cation, readily detected by e.s.r. (CnH2n+ 2)+ -(2-C4H8)t + cn 4H2n 6) (12) spectroscopy, together, presumably, with ethane. It would be most interesting to follow the course of photolysis of the branched-chain cations for comparative purposes. Cycloalkane Cations.-Cyclopropane cations have been studied at 77 K by Shida and co-workers34 and at 4K by Iwasaki and co-~orkers.~' This is another example of a parent cation with a degenerate ground-state which must undergo a Jahn-Teller distortion. The results at 77K show that this is dynamic, but the hyperfine features were poorly resolved because of the very small average proton coupling.However, at 4K the distortion is frozen, with two strongly coupled protons (+21 G) and four weakly coupled protons (-12.5G). The opposite signs are required by the averaging results, and by theory for the 'A distortion indicated H 32 T. Shida and Y. Takemura. Rudiut. Phys. Chem., 1983. 21, 157. 33 M. Tabata and A. Lund, Z. Naturforsch., 1983, %a, 428. 34 K. Ohta, H. Nakatsuzi, H. Kubodera, and T. Shida, Chem. Php. 1983, 76, 271. 35 M. Iwasaki. K. Toriyama, and K. Nunome, J. Chem. Soc.. Chem. Commun.. 1983, 202 403 Radical Cations in Condensed Phases in insert (6).The observed distortion is an elongation of one C-C bond, the SOMO having considerable o-bond character.The four protons attached to these carbon atoms are in the p-orbital nodal surfaces and hence give a negative coupling. Possible alternative distortions, such as 'B,, involving equal stretching of two C-C bonds would give quite different proton coupling constants, according to theory, and can be rejected. Similar arguments apply to cycloburane cations. Again a distortion must occur, that predicted being the 'B3 state of D,symmetry (two stretched C-C bonds giving a rectangular cation). The e.s.r. results show, once again, that this is dynamic at 77K, the average splitting being 14G.34 For the cyclopentane cation, the SOMO is clearly confined to three CH, units, with strong coupling to two protons (22.4G) and weak coupling to the remaix~der.~~.~~At 77 K the spectrum is a broad 1 :2: 1 triplet, but the hole moves around the ring rapidly at ca.100K, all 1 1 lines being detectable (6.3 G). Thus, the structure is o rather than n in character, and the ground state is Jahn-Teller distorted. The relatively low values for A('H) suggests that the SOMO is largely confined to C-C bonds, presumably in part because 'planar' arrangements favouring delocalization into the C-H orbitals cannot be achieved. Cyclohexane cations have been widely st~died,~~.~~,~~ the most definitive work being that of Iwasaki and his co-~orkers.~~ At temperatures 2 140K the expected Jahn-Teller distortion is dynamic, and six of the twelve protons couple strongly (43G each). However, at ca. 4K this changes to a set of 13 lines.The results vary somewhat with different matrixes, but they extrapolate to a situation in which there are two strongly coupled protons (103G) and four equivalent weakly coupled protons (13G). These results compare well with the values predicted by INDO calculations for the a, orbital which is the SOMO for the 'A, ground state. This orbital is shown in Figure 4 together with the alternative 6, SOMO. There are three equivalent structures having two parallel C-C bonds stretched, and equivalence is achieved when these interconvert rapidly. Further interesting details of this study are also de~cribed.~' One very interesting feature of these spectra is the large shift in gll(z)(= 2.01 11). This shift, which is much greater than is usually observed, was explained in terms of the proximity of the filled b, level which is coupled by B,.These results beautifully illustrate the phenomenon of o-delocalization. I now turn to examples in which more conventional structures can be used, beginning with the alkene cations. 4 Alkene Cations Although the tetramethyl ethylene cation, (Me,C==CMe,)+, was studied by e.s.r. spectroscopy many years ago,38 the only systematic study of alkene cations is that of Shida and his co-~orkers.~~ However, the most interesting study is of the parent (C,H4)+ cation, since the large coupling to four equivalent protons clearly 36 M. Tabata and A. Lund, Chem. Phys., 1983.75, 379. 37 K. Toriyama, K. Nunome, and M. Iwasaki, J. Chem.Soc.. Chem. Commun., 1984. 143. R. M. Dessau, J. Am. Chem. Socs., 1970.92.6356. 3y T. Shida, Y. Egdwa. and H. Kubodera, J. Chem. Phys., 1980. 73, 5963. 404 Symons -0--, / / -o---/ / bg \ / \ \ -00-' / \ -0 0-(elongated 1 (compressed) Figure 4 The SOMOs for the 2A, and 'B, states of the cyciohexane radical cation establishes that this species is twisted.40 The planar ion is expected to have negative proton couplings in the region of 11G. However, on twisting, the protons can couple as P-protons as well as a-protons. So the splitting should fall to zero and then become positive, which must be the situation for the large values ~bserved.~' I cannot say more about the results for (C,H,)+ cations since details have not been published.It is interesting to note that in 1947 Mulliken and Roothan predicted a 30" twist for this ion arising because of the hyperconjugative interaction that ensues.,' Since then there has been considerable controversy about the degree of twisting,42 and at least one recent calculation has predicted a planar structure.43 Results for the methyl-substituted cations were all discussed in terms of planar structures, and seemed to be well accommodated by INDO calculations. They need to be reconsidered now, in terms of possible twisted structures. We have considered the structures of asymmetrically substituted alkene cations and shown that trends in proton coupling constants can be understood in terms of a separation of the charge and the electron.Thus methyl substituents stabilize the 'hole' by electron release so the electron tends to favour the least methylated carbon. For example, 40 Y. Nagata. M. Shiotani. and J. Sohma, reported at the 22nd Japanese ESR Symposium, 1983. " R. S. Mulliken and C. C. J. Roothan, Chem. Rev., 1947.41, 219. 42 D. J. Bellville and N. L. Bduld, J. Am. Chem. Soc., 1982, 104, 294. 43 S. Merry and C. Thornson, Chem. Phys. Lett., 1981.82. 373. 405 Radical Cations in Condensed Phases for the cation (H,C==CMe,)+ the valence-bond structure H,~-bMe, is favoured. Hence the (H2C) proton coupling of 14G is higher, whilst the methyl- proton coupling of 16.5G is lower, than would be predicted for an even distribution of ~pin-density.~~ Cyclic Alkene Cations.-The cation of cyclopentene is of interest since, in this case, extensive twisting about the carbon-carbon double bond is unlikely.The results obtained by Tabata and Lund at 1 10K are shown in (7).36 The %-proton coupling of 10.5G is probably negative, and is in good accord with expectation for a planar cation. Also, for the cyclohexene cation (8) a splitting of (-) 9G suggests only a small degree of twist. 9G 9G 49 -v 7G (7) Diene Cations.-In the field of diene cations, quite the most interesting development has been the discovery of some persistent cyclobutadiene cations R =CH, , MeCH, etc. (9).45*46The results for cyclohexa- 1,4-diene cations are interesting in that there is complete delocalization via the two CH2 groups as evidenced by their large 'H coupling (67.1 G).36 The cation of cyclopentadiene is of interest since, at first sight, one might again have expected a large coupling from the bridging methylene protons.However, this is a 3n-electron radical, in contrast with the corresponding radical anion and the well-known cyclohexadienyl radical, which have Sn-electrons. For the 3-electron systems there is a node through the methylene group so these protons cannot participate in the wave-function (10). The results, shown in (10) confirm this conclusion. So far, there has been no publication dealing with alkyne cations. These are of particular interest in view of the degeneracy of the n-orbitals from which electron '* M. C. R. Symons and L. Harris, J.C'hem. Res.. (S)1982,268; (M) 1982,2746. Q. B. Broxterman, H. Hogeveen, and D. M. Kok, Tetrahedron Len.,1981,22, 173; 1983,23,639. 46 J, L. Courtneidge, A. G. Davies, and J. E. Parkin, J. Chem. SOC.,Chem. Commun., 1983, 1262. Symons loss is expected to occur. We have obtained some interesting spectra from MeCSMe, but are as yet somewhat uncertain of the correct analysis. However, there is no large g-value variation, so it is unlikely that the spectra are due to the undistorted parent cations. The Review continues with various cations having major spin-density on heteroatoms. Carbon-centred radicals with aromatic derivatives are again discussed in Section 14. 5 Alkyl-Halide Cations Various cations RCl=+, RBr-’ and R1.’ have been prepared in CFC1,;’ but the results are complicated by the fact that there is weak complexing between these cations and one chlorine atom of a neighbouring CFCl, molecule (Table 2).We suggest that relatively weak o* bonds are formed, as in (1 l), for example. This R I+ /C Cl CI makes spectroscopic identification easier despite the extra splitting from j5Cl and 37Cl, since the expected degeneracy of the halogen x-orbitals is lifted, and hence the g-values are relatively close to 2, and the lines are not greatly broadened. Unfortunately, hyperfine coupling to P-protons is small (ca. lOG), whereas it is predicted to be large, at least for one or possibly two protons. This difference reflects the presence of o bonding, which has been shown in several cases to reduce delocalization within the two components (see Section 11).Although the chloro- and bromo-cations gave well-defined solvent complexes, no such complex could be detected for iodoalkanes. However, at ca. 4 K six broad components were detected in the g = 3.6 region, and two unresolved broad features were detected at ca. g = 1.2 and g = 0.85. This species is thought to be the parent cation, RI-+ with very weak interaction with solvent molecules. One reason for this 47 G. W. Eastland, C. Glidewell, A. Hasegawa, M. Hayashi. S. P. Maj, M. C. R. Symons. and T. Wakabayashi, J. Chem. Soc.. Faraday Trans. I, 1984, in press. 407 Table 2 E.s.r parameters assigned to ulkyl halide cution udducts and dimers sg-values Hyperfine coupling a 5-z Host Matrix Radical R It Rl Nucleus A I1 C2D5Br CFClj ClCFCl2l' 1.922 2.392 79Br 464 159 81Br 500 171 35c1 57 ca.17 C2H5Br CFC13 [C2H &-I2 + 1.988 ca. 2.0 79Br 42 1 ca. 120 81Br 454 ca. 130 C2H5Br cc14 [C2H5Br ClCC13]+ 1.955 2.329 79Br 480 119 "Br 517 129 3 5c1 68 ca. 17 CzHsCl CFClj [CrHsCl-ClCFC12-J' 2.002 cu. 2.0 35c1(1) 110 ca. 35 3 5c1(11) 85 ca. 35 37c1( 92 ca. 30I) 3 7c1(11) 71 cu. 30 'H 10 10 CzHsCl CFC13 [CzH scllz + 2.006 cu. 2.0 35c1 99 cu. 35 37c1 82 ca. 30 CzHsCl cc14 [CzHsCl a* ClCC131' 2.003 cu. 2.0 35c1(1) 105 cu. 35 3 ~CI(II) 88 cu. 35 37c1( 87 cu. 30I) 3 ~(11) 73 cu. 30 'H 7 7 1.877 ca. 2.1 1271 460 cu. 120 Symons difference is the greater discrepancy between ionization potentials of RI and CFCl,, and the other is the large spin-orbit coupling energy for RI-’, responsible for the very large g-shifts, which needs to be overcome in order to form localized o-and o*-bonds to chlorine or fluorine.At relatively high concentrations, or on annealing, dimer cations (Rhal-halR) + were formed. These are also CY*complexes, and are the major electron-loss centres formed on irradiating pure ha loge no alkane^.^^ The tendency to form such o* dimers is very marked for all but first-row radicals provided the SOMO of the electron-loss centre and HOMO of the parent cation are not extensively delocalized. Much more work has been directed towards oxygen- and sulphur-centred radicals, which are now discussed. 6 Alcohol and Ether Cations Unfortunately, our attempts to prepare alcohol radical-cations by this procedure have failed.This is because of extensive hydrogen bonding giving dimers or clusters of alcohol molecules in the frozen solutions. Infrared studies show that even in the most dilute solutions suitable for cation studies, there is extensive hydrogen bonding, and on radiolysis reactions of type (13) predominate, only the radicals R2e0H being detected by spectroscopy. It may well be that at least CH,6H+ (R,CHOH), -e-+ R&OH + R,CHOH,+ (13) will prove to be suitable for study by the matrix-isolation technique developed by Knight and his co-workers. In contrast, ether cations, never previously detected by e.s.r. spectroscopy, have now been widely studied in freon matrices.49 51 Some results are summarized in Table 3. The most noteworthy aspect is the magnitude of the @-proton hyperfine coupling. For Me20-+ cations, this is ca.43 G. This can be compared with values for Me,tH (ca. 26G), Me,fiH+ (ca. 37G), or Me2fi (27.4G). These increases accord well with the simple picture of hyperconjugation involving, primarily, electron donation from the C-H o-orbitals towards the central atom rather than electron donation from the central atom into the C-H o*-orbital (see Figure 5). In passing, I call attention to one of the major e.s.r. puzzles of recent years. The radical Me6, trapped in solid methanol at 4K, exhibits an average proton hyperfine splitting of 52G.” This accords with results for ether cations.However, this coupling for Me00 radicals in the gas-phase is calculated to be 23.2 G, after correcting for orbital magnetic contribution^.^^ Theory cannot be relied upon to select a correct value, nor has there yet been any explanation for the 48 S. P. Mishra and M. C. R. Syrnons, J. Chem. Soc., Perkin Trans. 2, 1975, 1492. ‘’)H. Kubodera, T. Shida, and K. Shirnokoshi. J. Phys. Chem., 1981.85, 2583. ’’J. T. Wang and F.Williams, J.Am. Chem.Soc.. 1981,103,6994;L. D. Snow, J. T.Wang, and F. Williams. J. Am. Chem. Soc.. 1982, 104. 2602.’’ M. C. R. Syrnons and B. W. Wren. J. Chem. Soc., Chem. Commun.. 1982.817.’’ M. Iwasaki and K. Toriyarna, J. Am. Chem. Soc., 1978,100. 1964. 53 H. E. Radford and D. K. Russell, J.Chem. Phys., 1977.66,2222; 1980,72,2750. Radical Cations in Condensed Phases Table 3 E..s.r. purumeters ,for ether rudicul cutions Proton hyperfine coupling (G)" Cation r A (CFCIJ soloent) P-H Y-H Me20: 43 (6 H), 43' 43.6, 42.8, 42.5' 68.7 (4 H) 45 16 (4 H) 64 (4 H) 11 (2 H) 89 (2 H), 40 (2 H) 89 (2 H), 40 (2 H)' 59 (2 H), 22 (2 H)' 97 (2 H)' 95 (1 H), 35 (I H)' 83 (1 H), 42 (2 H) 78 (1 H). 49 (2 H) (146 K)' 34.5 (2 H), 14 (2 H) 11 (2 H), 3 (2 H) -OCH20-P-protons Other P-protons MeOCHzOMe' 31.3 (2 H), 6.0 (4 H)MeOCH, CH, 0Me 136.1 (2 H) ca. 10 (10 H) 153 (2 H) 11 (4 H) 153 (2 H) 11.2 (4 H) 140 (2 Hay) 26 (2 H), 12.5 (2 H) 140.6 (2 H) 26.3 (2 H), 12.4 (2 H)' 11 (4 H), 8 (4 H) 22.4 (2 H) 5.0 (2 H, aromatic) 135 (2 H) 32 (2 H) MeO QbMe ono ca.150 (1 H) HXOMe fO> + 162 (2 Hay) 23 (1 H), 11 (1 H) 0-0 160.2 (2 H) " G = 10 T. * Ref. 49. Rqf: 50. Tentative. D2S04 solvent. Ref: 56. Symons Figure 5 Qualitative scheme shouing why electron donation from a C-H o-orbital in the optimum orientation increases on going from C to 0 difference. From a qualitative viewpoint, I wonder if the difference lies in the fact that orbital motion is quenched for the radical in methanol, but not in the gas- phase. Does this orbital motion inhibit hyperconjugative interaction? Some interesting facets of the results for ethers include the change from axial (2H, 89G) and equatorial (2H, 40G) coupling of the four p protons in the THF cation to an average value (4H, 64.56) on annealing, giving an activation energy of ca.1.65 kcal mol 1.49 Both axial and equatorial protons are strongly coupled. However, for the tetrahydropyran (oxacyclohexane) cation, the couplings are greatly reduced (34.56 and 15G). We suggest that, in this case, the conformation has changed so that the total overlap has been redu~ed.~~,~~ Results for the four-membered ring (oxetane) cation were normal, with four equivalent P-protons having A = 64G. However, the three-membered ring (oxirane) cations had four equivalent protons with greatly reduced coupling (16G).54*55This result eliminates the normal x-structure (12) (bl). We originally (12) (13) 54 M. C. R. Symons and B. W. Wren. Tetrahedron Lett., 1983.24.2315.55 M. C. R. Symons and B. W. Wren. J. Chem. SOL'.,Perkin Trans. 2, 1984. 51 I. 41 I Radical Cations in Condensed Phases suggested the alternative o-structure, (1 3) (al),with a stretched C-C bond, since this should exhibit greatly reduced proton coupling constant^.'^ However, it seems that ring opening to form the allyl-type radical (14) is more pr~bable.~~*~’ This certainly fits the spectrum satisfactorily, and accords well with current the~ry,~’ but, at present, no clear distinction can be made. All three groups discovered independently that the -0-CH,-0- protons of acetal cations such as (15) exhibit unusually large splittings (ca. 140G). These results show the importance of delocalization within the -OCH,O- 7c system, indicated in (15).Indeed, the spin-density is quite evenly distributed within this unit, and delocalization onto the other two CH, groups is small. However, if one of the CH, protons is replaced by an alkyl group, the cation readily loses an alkyl + radical (14).58 Evidently, the incentive for this reaction is the stability of the (RO),CH carbocation.+ (ROCHOR)’ --+ (ROCHOR)’ + R‘* (14)kf The cation of 1,4-dioxane is of interest since the proton hyperfine coupling was far less than those for any other ether cations. This led us to consider the possibility of weak o* bonding between the two oxygen p-orbitals for the boat form of the cation of the type observed for the sulphur analogue (see below). If such bonding does occur, then the small proton coupling is a natural consequence of such bonding.” 7Carbonyl Cations In this section I discuss results for aldehyde, ketone, and ester cations, despite the fact that the ester structures (SOMOs) turn out to be quite different from those for aldehyde and ketone cations. Aldehyde Cations.-Prior to the introduction of freon and related solvents, the only cation in this class was that of formaldehyde, which we prepared in a sulphuric acid rnatri~.~’For reasons that we do not understand, no clear features for H2CO*+ s6 L.D. Snow, J. T. Wang, and F. Williams, Chem. Phvs. Lett.. 1983. 100. 193.’’ T. Clark, J. Chem. Soc.. Chem. Commun., 1984, in press. K. Ushida and T. Shida, J. Am. Chem. Soc.. 1982, 104, 7332. s9 S. P.Mishra and M. C. R. Symons, J. Chem. Soc., Chem. Commun., 1975.909. 412 Symons cations have been obtained from solutions of formaldehyde in freon solvents, but the complementary technique of photoionization using a rare-gas matrix gave very good e.s.r. features for this cation.60 These cations, which are isoelectronic with the well-known H,CN radicals, exhibit the expected large coupling to the two protons (Table 4). A careful analysis of this coupling have led Knight and his co-workers to suggest the interactions and axes indicated in insert (16). A surprising but compelling conclusion is that g,,,, does not lie along the C-0 axis as has generally been supposed because of the expected facile coupling between the K* and n orbitals. In fact, this direction is associated with a free-spin g-value, the suggestion being that coupling between the SOMO and the filled and unfilled K and K* orbitals effectively cancel.Although we were unable to obtain clear evidence for H,CO+ cations, other aldehyde and ketone cations were readily prepared in CFC13.61 The aldehyde spectra are characterized by a doublet splitting of ca. 135G. We now reali~e~,.~~ that an extra anisotropic quartet splitting which appeared on the acetaldehyde cation features was due to a weak hyperfine interaction with chlorine which is lost reversibly on annealing. The structure (17) is thought to be similar to that discussed above for the alkyl halide cations. The chlorine interaction was lost reversibly on annealing.c =o ‘O L. B. Knight and G. Steadman. J. Chem. Phys.. 1984, in press. b1 P. J. Boon and M. C. R. Symons, Chem. Phys. Larr., 1982.89, 516. 6z A. Hasegawa, J. Rideout, G. W. Eastland. and M. C. R. Symons, J. Chem. Res. (S),1983. 258. 63 L. D. Snow and F. Williams, Chem. Ph.vs. kit., 1983, 100, 198. 413 Radical Cations in Condensed Phases Table 4 Ohseraed and calculuted e.s.r. purumeters .for vurious uldehyde and ketone radical cations Hyperfine coupling constunts (G) No. Radical TIK 0-Protons y-Protons 6-Protons cI Me, c=04 77 136(63.7)' (-3.9, -3.9, -0.4) H' MeCHz 2 \,c=o* 120 135(61.9) ( -4.0, -4.0) 12.5 (1 H)'H (4.9, 0.4, 0.4) 3 MezCH >c=o* 77 I38(68.2) (1.3) 20 (2 H) H (3.1, 1.4, 0.9) x 2 77 (-2.3, -2.3, -1.6) x 2 Me MeCH2,5 ,c=o' 77 (-2.6) x 4 II (2 H)/MeCHz (5.8, 1.1, 1.1) x 2 Me,CH ,,c=o' 77 (-0.6) x 2 15 (4 H)MezCH (3.0, 1.6, 0.9) x 4 15 (4 H) (5.7, 1.0, 1.0) x 4 (2.9, 1.5, 0.9) x 2 8 cc=o* 77 (-1.6, -1.5) x 2 13 (2 H) (2.2, 0.0) x 2 9 cc=o* 77 (-1.9, -1.4) x 2 27.5 (2 H) (5.5, 0.4) x 2 A 10 0uc=o' 77 (-1.8, -1.7) x 2 19.5 (2 H) (4.1, 0.3) x 2 "G = 10 4T.bAverageg-valuescu.2.0035 0.001.AI,(~~CI)19.0G:AI(35CI)c.a.6G(CH~CHO');t~u.4 G(C03CHO+).The numbers in parentheses are the results of INDO calculations. 'Lost reversibly on annealing. f Changes to 3 (4 H) reversibly on annealing. One particularly interesting aspect of the spectra for these cations is that the Y-proton coupling [for example, the methyl group in (MeCHO)+] is very small.Our ENDOR results for the acetone cation gave splittings of 1.5G and 0.3G for two different types of protons suggesting preferred orientations for both methyl 414 Symons groups. In contrast, &protons can exhibit couplings as great as ca. 30 G.64*65These large couplings are only found for specific protons having conformations close to the ideal W plan, as in (18). For example, two of the four &protons of cyclohexanone have near optimum locations, and exhibit ca. 29 G splitting. Our result for the 1,4-diketone cation (19) is particularly interesting, since the e.s.r. spectrum comprises-a well-resolved quintet (A = 12.5G)from four equivalent protons.66 If we rule out rapid interconversion at 77K, this can only mean that there is delocalization across the two carbonyl groups so that all four equatorial protons are involved.This surprising result contrasts markedly with our result for the benzoquinone cation, which has hyperfine coupling to only two protons, thereby establishing that the SOMO is strongly l~calized.~’ In a preliminary study of the acetone cation, in addition to our ENDOR results mentioned above, we have studied the spectra for (‘3CH,),CO+ cations in CFCl, and CCl,.68 The results suggest that the spin-density on each methyl group is ca. 22%. If this is correct, then hyperconjugative delocalization must be large, though probably less than that for C-H bonds (ca. 26%) as seems to be required by our results for (Ph-Et) cations discussed below.+ Ketene Cations.-E.s.r. results 69 show that these have the expected x-SOMO shown approximately in (20). Indeed, the x-proton coupling of ca. 20.5G suggests that the spin-density on the (CH,) group is close to unity, as required by the limiting structure, R~~-c=o+. “‘0 H\C-@ 0 C -0 (20) 64 L. D. Snow and F. Williams, J. Chem. SOC.,Chem. Commun., 1983, 1090. 65 P. J. Boon, M. C. R. Symons, K. Ushida, and T. Shida, J. Chem. Soc.. Perkin Trans. 2, 1984, 1213. 66’’Unpublished result. H. Chandra and M. C. R. Symons, J. Chem. Soc., Chem. Commun., 1983,29.’’P. J. Boon, L. Harris, M. Olm, J. L. Wyatt, and M. C. R. Symons. Chem. Phys. Leu.,1984, 106,408 69 D. Becker, K. Plante, and M. D. %villa, J. Phys. Chem., 1983, 87,1648.Radical Cations in Condensed Phases Ester Cations-These are of special interest since the carbonyl n(0)orbital (21) is similar in energy to the x orbital (22). The n orbital is comparable with the SOMO for aldehydes and ketones, whilst the x orbital is strongly delocalized, but has its major spin-density on the 'ether' oxygen. Various theoretical calculations have been presented in favour of either of these orbitals as the SOMO for the cation. + I nR' (211 (22) The first ester cation prepared in freon, that of methyl formate, proved to be exceptional since a species exhibiting strong coupling to a single chlorine nucleus was obtained at 77K.62*69This is expected to be derived from the n-cation, as in (23). Me-0 @ + Lo There is a single proton coupling of ca.17G which can be assigned to the formyl proton. This coupling is much smaller than expected for the free n-cation but, as stressed above, the formation of a obond tends to localise the SOMO,so the results are in reasonable accord for this structure (see Section 11). On annealing there is an irreversible change to a new species which we think is the x-~ation.~~.~~ This exhibits strong coupling to two of the three methyl protons, the splitting (ca. 23 G) [the average coupling being significantly less than that for the methyl protons of (Me6Me)' cations (43 G)] showing that delocalization is indeed extensive. (We originally thought that this radical might be H2COCv/OH+ , but work on other ether cations does not support such a rearrangemen t.6 2, These results suggest that the first-formed cation is the n-cation, but this may be only because of stabilization of the system via the o/o*interaction.Clearly, the thermodynamically stable structure is the x-cation. Symons Other esters seem to favour the K-SOMOdirectly. Some of our results are shown in Table 5, together with those of Sevilla and his co-workers. The most noteworthy result is the relatively large hyperfine coupling often obtained for y-hydrogen atoms. Again, a preferred conformation is required. Table 5 E.vperimental e.s.r. purameters ,for n-cations of esters Hyperfine couplinglG a*b Esters A; A:l HC02CH3 5.5 23 (2H), 4 (1H) CH 3C02CH 3 5 (3H) 22(2H) CHjCOzCDj 5 3.5 (22H) CH 3CH2C02CH 3 7 (IH) 25 (2H) HC02CH2CH3 2.5 22 (2H) 10 (2H), (160 K) 16 (IH), (77 K) CH3C02CH2CH3 T.’g., -2.0026.G = 10 22 (lH), 31 (1H) 8 (lH), 17 (1H) We have also made a combined study with Sevilla’s group of electron-loss from lac tone^.'^ In general, the results are compatible with the results for esters, and in several cases there is evidence for intramolecular proton transfer to give alkyl-type radicals as indicated, for example, in reaction (1 5). (See note added in proof, p. 439.) qoo+ Carboxylic Acids.-As with the alcohols, it is impossible to isolate carboxylic acid monomers by freezing dilute solutions in freon solvents. However, in contrast with the alcohols, well-defined cyclic dimers can be isolated and these might yield specific cations on irradiation.However, for acetic acid dimers, high yields of methyl radicals were obtained, even after irradiation at ca. 4 K (16).’* ’O M. D. Sevilla er al.. unpublished results. ” J. Rideout and M. C. R. Symons. J. Chem. SOC..Perkin Trans. 2, 1984, in press. Radical Cations in Condensed Phases 8 Sulphur Cations Several of the important classes of sulphur cations had already been studied by e.s.r. spectroscopy before the advent of freon solvents. Because of the lower ionization potentials of the precursors and especially because of their lower reactivity, they are often stabilized in the pure compounds at low temperatures, in contrast with their oxygen analogues. However, in early studies it was not realised that dimerization is a major reaction and dimer-cations were frequently misidentified as monomer cations (see ref.72 for detailed arguments and references to these early studies). Using freon solvents, the dimerization step (17) can be avoided or followed on annealing.72 It is probably diffusion controlled and seems to be irreversible in the solid state. This process has been extensively studied by Asmus and co-workers in the liquid-phase, using the intense o-m* transition in the visible region which characterizes these species. R2S*++ R2S -RZS'-SR,+ (17) E.s.r. results for the monomer cations, R2S-+ follow closely those of the isostructural ether cations, R,O*+, except that the proton hyperfine coupling constants are reduced by about a factor of [Compare Me,O*+ (43G) and Me2S*+(21 G).] This reduction is expected because of the lower electron affinity of sulphur and the poorer overlap with the more diffuse sulphur orbitals.When dimer cations form, the proton coupling falls to ca. 6.8G.75This reflects the extra confinement of the unpaired electron discussed above for the alkyl halide cation dimers. An important generalization can be drawn from these and other results, namely, that whilst radicals centred on second and subsequent row elements (e.g. R3P*+, R,S-+, and RCI-+) normally react to form dimer radical cations with o*SOMOs, the corresponding first row radical cations (e.g.R,N-+,R20*+)react in other ways. This may occur because first row dimer cations are innately less stable, as suggested by Glidewell 76 and steric factors may play a significant rde.However, our inability to detect these dimers may occur because hydrogen atom abstraction processes are more favoured for the first row cations. Our results with cyclic ether cations55 suggest that if intra- and inter-molecular hydrogen atom transfers are prevented, intramolecular formation of o* SOMOs may indeed occur (see Section 6 above). Cyclic monosulphide cations again closely resemble the corresponding ether reactions except for the three-membered ring cation which, with hyperfine coupling of 31 G to four equivalent protons, appears to have the normal n-structure. It is especially noteworthy that there seems to be no tendency in the temperature range I--1 77-160K to form the ring-opened cation (H2C-S-CH,)'.''R.L. Petersen, D. J. Nelson. and M. C. R. Symons. J. Chem. Soc., Perkin Trans. 2. 1977. 2005.''D. N.R.Rao, M. C. R.Symons, and B. W. Wren, J. Chem. SOC.,Perkin Trans. 2, 1681. 74 J. T. Wang and F. Williams, J. Chem. Sot.., Chem. Commun.. 1981, 1184. 75 B.C. Gilbert, D. K. C. Hodgeman, and R. 0.C. Norman. J. Chem. Sor., Perkin Trans. 2. 1973. 1748. "C. Glidewell. J. Chem. Soc., Perkin Trans. 2. 1983. 1285. 418 Symons Cyclic disulphides tend to form intramolecular S-S bonds, giving o* SOMOS.~~.~~This aspect of sulphur chemistry has been most effectively studied by Asmus and co-workers, since the energy of the o-m* or related transitions are markedly dependent upon the strength of the bond, whereas e.s.r.parameters are relatively insensitive to such changes. It is noteworthy that even disulphides of type (RS-CH,SR)+ seem to form weak S-S bonds at 77K despite steric difficulties, rather than adopting the Jr-delocalized structure favoured by acetal cations. (RS-SR)' Cations.-Persulphides form x* cations on irradiation. These have also been studied by e.s.r. spectroscopists prior to the introduction of freon-type solvent^.^^.^^ We have shown that the same species are readily formed in CFCI, thereby confirming the previous identifications. 9 Selenium-centred Cations These have not been extensively studied, but the available results show the expected trends. The 'parent' cation, Me,Se-+ was briefly reported by Wang and Williams,74 but they only reported the methyl proton coupling (ca.15G).Our own results show that the g-values have diverged markedly relative to those for Me,S-+ and there is an extra doublet splitting of ca. 8G for solutions in CFCl, which is absent for solutions in CCl4o8' It might be thought that these radicals are good candidates for forming weak o-bonds to chlorine but, instead, they clearly select fluorine. The enhanced g-shifts are in accord with the greater spin-orbit coupling for selenium. They facilitate spectral interpretation relative to the sulphur cations since all three components are well separated instead of overlapping. 10 Nitrogen-centred Cations Although the *NH,+ and -NMe,+ cations have been extensively studied by e.s.r.spectroscopists both in liquid and solid systems,8' 83 there had been no systematic studies of R,N-+ cations prior to the introduction of freon-type solvents. We have recently studied a range of such ions in CFCl, and can draw some conclusions about their proper tie^.^^ The most significant result stemming from e.s.r. studies has been that these cations are planar, as judged by the 14N hyperfine tensor components. This is the same criterion that has been used to argue that most alkyl radicals are planar but, in contrast with the situation for alkyl radicals, the concept of preferred planarity of R,N*+ radicals has not been questioned. Our results for Et,N-+ are of interest in that strong coupling (38 G) to only three protons is observed at 77 K.This changes reversibly to coupling to all six protons 77 D. N. R. Rao, M. C. R. Symons, and B. W. Wren, Tetruherlron Lett.. 1982, 23. 4739.''H. C. Box and H. G. Freund, J. Chem. Phys., 1964.41,2571. " F. K. Truby, J. Chem. Phj..~.,1964, 40,2768.'"A. Hasegdwa and M. C. R. Symons, unpublished results. T. Cole, J. Chem. Phj-s., 1961, 35, I169.'* A. J. Tench. J. Chem. Phj-s.. 1963, 38, 593. 83 J. A. Brivati, K. D. J. Root. M. C. R. Symons, and D. J. A. Tinling. J. Chem. Sot. (A), 1969, 1942. 84 G. W. Eastland, D. N. R. Rao, and M. C. R. Symons, J. Chem. Sot., Perkin Trans. 2, 1984. 1551. 419 Radical Cations in Condensed Phases on annealing, the coupling changing to 19G.This value is well below the expected average coupling (28G)so that rotation must remain restricted, as in (24) and (25).Thus, in this case, steric control dominates. Results for other amine cations are in good accord with expe~tation.~~ Results for 1,l-diamine cations, (R,NCH,NR,)+, show that the SOMO is confined to one nitrogen atom. This makes an interesting contrast with our results for acetai cations and 1,l-disulphide cations discussed above. There are two alternative structures, one with the SOMO equally distributed on the two nitrogen atoms, with equal flattening, the other with the SOMO confined to one, planar, nitrogen atom, the other retaining the normal amine pyramidality. Clearly, the latter asymmetric structure is favoured. This is a more extreme and perhaps more obvious example of the asymmetric structure that I propose for the pyrimidine cation (see Section 15 below).As with the ethers, and in contrast with phosphine cations (see Section 11 below), there is no tendency to form amine dimer cations, (R,N”R,)+, the preferred reaction being hydrogen atom transfer. However, possible evidence for very weak intramolecular (T*bonding was adduced for the cation of N,N’-dimethylpiperazine (26), on the basis of greatly reduced proton hyperfine coupling constants.84 However, there can be no doubt that the cation of triethylenediamine (TED)(27) is truly delocalized, and this is surely aided by the fact that it is now impossible to form a cation planar at one nitrogen only. (26) (27) Before leaving the topic of amine cations, I consider another aspect of o* bonding which is certainly exhibited very well by R,N-+ cations.Some time ago, in a paper ignored by e.s.r. chemists,85 Patten established that H3N-+ radicals form F. W. Patten, Phys. Rev., 1968, 175. 1216. Symons weak o*bonds to C1-and Br-. I have shown how significant this result is relative to the far weaker interactions found for alkyl radicals and halide ions which are, at best, only weak charge-transfer interactions.86 This work has been extended to a range of alkylamine cations, all of which form CT*bonds to C1-, Br-, and I-ions.87 In all cases, the R3N- moiety is clearly pyramidal and there is extensive transfer of spin-density onto halogen. These 6+ 6-6+ 6-R3N I-X complexes are structurally similar to the R3P'X halide complexes studied by Symons and Petersen.88 Amide Cations.-Apart from a brief study of the cations of dimethylformamide and dimethyla~etamide,~~ no attention has been paid to these species.Our results show that, in accord with theoretical predictions and photoelectron spectroscopic results, the SOMO for these cations is the x orbital shown in (28). The spin is quite strongly confined to nitrogen, the spin-density calculated from the hyperfine components being ca. 78% in both cases. The N-methyl groups are 'freely' rotating and exhibit large splittings (32G) whereas the C-H and C-Me protons couple very weakly. 0 (28) +It is intefesting to compare trends for the amides and esters, i.e. for HC(O)NMe, and HC(O)OMe, since both have similar x SOMOs.The most noteworthy differenceis that theapparentlyfessrestrictedmethylgroupfor theester is firmly fixed, giving coupling to only two protons. This coupling is less than the average value for the amide, the estimated spin-densities on nitrogen and ester oxygen being ca. 78% and ca. 50% respectively. In both, the spin-density on the C-H moiety is low, so that on the carbonyl oxygen must be considerably greater for the ester. Is this possibly a reason for the restricted rotation of the methyl group? (However, see footnote p.439.) Nitroalkane Cations.-Perhaps there should have been classified under oxygen- centred radicals since the parent cations are expected to have a SOMO which is non-bonding (n or 0)on oxygen.However, these cations are evidently very unstable, and rearrange even at 77 K to give species having spectra very similar to that of *NOz radicals. It was originally argued that these must be the rearranged H6 M. C. R. Symons, J. Chem. Research (3,1981, 161. 87 1. Rowland, J. B. Raynor, and M. C. R. Symons, unpublished results. RH M. C. R. Symons and R. L. Petersen. J. Chem. Soc.. Faraday Trans. I, 1979, 210. 89 D. N. R. Rao and M. C. R. Symons. Chem. Phys. Lert.. 1982,93,495. 42 1 Radical Cations in Condensed Phases nitrite cations ONOR+ which are expected to have SOMOs closely resembling that for -NO,.90This postulate has now been modified by the observation of a large I3C splitting in the spectrum for the cation of 13CH3N02.91The results suggest that the carbon and nitrogen hyperfine tensor components are co-linear, the estimated spin-density on carbon being ca. 46% with a p:s ratio of ca.4.7. Delocalization onto carbon for ONOR’ cations is not expected to be large, so we now propose the o-structure H3C-N02+for the first formed cations, with a stretched C-N bond. [This resembles one of the structures for N,O,’ cations discussed below in Section 17.1 On annealing, however, the 13Csplitting was lost irreversibly but the 14N splitting was retained. We suggest that this represents the formation of the nitrite structure. Thus we propose the sequence shown in equation (18) for the overall process: + +R-~6),-R.NO, -O~~OR+ (18) It is interesting to consider this rearrangement in the light of reaction (18) for the isoelectronic RCO,.radicals. These do have the expected SOMO confined to oxygen but reaction (19) is generally fast at 77 K. It is, in fact, nicely illustrated by the work on carboxylic acids in freon discussed above, and by the recent study of irradiated lithium a~etate,~~.~~ There seems to be no evidence for the formation of OtOR radicals in these reactions, even though they have been prepared by other routes. Also, we have not detected any Re radicals which might have been formed together with NO2+by a reaction analogous to (19). Probably a major factor is the low reactivity of CO, relative to NO,+. R-C --+Re+ CO, 11 Phosphorus-centred Radicals Phosphine Radical Cations.-Some time ago, we reported the formation of various R,P*+ radical cations when dilute solutions in sulphuric acid glasses were irradiated.94*95Although the results looked reasonable, identification was based solely on the large hyperfine coupiing to 31P,the lines being so broad that proton hyperfine coupling was not resolved.We therefore thought it wise to study these radicals again using CFC1, solvent so that electron-loss could be g~aranteed.~~.~’ Fortunately, the results strongly supported our original assignments. The most important result is that these radicals are pyramidal, the estimated pis ratio being in the region of 10, with the electron strongly confined to phosphorus. D. N. R. Rao and M. C. R. Symons, Tetrahedron Lefr., 1983.24. 1293. 91 D. N. R. Rao and M. C. R. Symons, J. Chem. Soc., Faraday Trans. 1. in press. ” D. P. Lin and L. Kevan, Radiar. Phys. Chem., 1981, 17, 71. 93 M. C.R. Symons, J. Phjs. Chem., 1983.87. 1833. 94 A. Begum. A. R. Lyons, and M. C. R. Symons, J. Chem. Soc. (A), 1971. 2290. 95 G. W. Eastland and M. C. R. Symons, J. Cham. Soc.. Perkin Trans. 2, 1977, 833. 96 M. C. R. Symons and G. D. ti. McConnachie, J. Chem. Soc., Chem. Cornmun.. 1982,851.’’A. Hasegawa, G. D. G. McConnachie, and M. C. R. Symons, J, Chem.SOC.,Farada-v Trans. I, 1984,80, 1005. 422 Symons This makes an interesting contrast with planar R3N*+ radicals. There are various explanations for this contrast, but it is not unexpected in view of the large decrease in bond angle on going, for example, from NH, (0 z 107") to PH, (0 z 91").Various trends are of interest. Thus, on going from -AIH,-to -SiH, and *pH,+, there is very little change in bond angle despite the marked change in electronegativity of the central atom. Nevertheless, on going from PO,' to MeOP0,2, (MeO),PO and (MeO),P-+, there is a large increase in Aiso (,lP) (ca.300G) implying a marked fall in bond angle. Also, on going from *PMe,+ to -P(OMe),+ there is a large increase in Ai,, and fall in the p:s ratio, and hence, presumably, in bond angle. The methyl proton coupling for -PMe,+ radicals is ca. 11.5 G but *P(OMe),+ gave very narrow lines with no sign of proton splitting. However, the latter gave different species in CCl, and CFCI, solvents.In CCI, a single species was obtained with hyperfine coupling constants considerably greater than the major species found in CFCI,. We attempted to explain this contrast in terms of two major conformations (29) and (30), both species being detached in mixed solvents. The triethyl derivative gave only the species with smaller coupling constants in both solvents. We tentatively suggested that the sterically unfavoured but more spherically compact conformation (29) is responsible for the large splitting, because the extent of pseudo K-delocalization is expected to be reduced. Me (29) (30) A minor species formed in CFCI, from P(OMe), probably has a small 19F doublet splitting and such a splitting was also found for the *P(OPh),+ cations.97 Phosphine Dimer Cations-Dimers (R,PzPR,)+, were detected on annealing systems containing *PR, + cations, or directly for more concentrated solutions.97 The ,'P hyperfine parameters show that the spin is less delocalized onto the ligands and that the p :s ratio has greatly decreased (ca.4) relative to that for the monomer cations.Also, for (Me3P~PMe,)+, the proton hyperfine coupling is drastically reduced. This extra confinement has been interpreted in terms of the anti-bonding nature of the o* ~rbital.~'The relative increase in 3s-character is a natural consequence of the need to equalise the bond angles for the -PR,+ and PR, moieties. These dimers have been studied in the liquid phase by Gara and Roberts.98 Some Group IV heteroatom radical-cations are now discussed.''W. B. Gara and B.P. Roberts. J. Chem. SOC.,Perkin Trans. 2. 1978. 150. Radical Cations in Condensed Phases 12 Silicon- and Germanium-centred Radical Cations These have been studied mainly by Williams and his CO-WO~~~~S.~~~~'~Results for the cation of tetramethylsilane show two sets of six equivalent protons suggesting a distortion of the type shown in (31). This makes an interesting contrast with that found for the CMe,' cation (32). Unfortunately, 29Si satellite features have not yet been detected, so nothing definitive can be said about the precise nature of the SOMO for this radical. The reasonable postulate is that it is largely 3p, on silicon, where 2 bisects the small MeSiMe angle."' The selection of a non-hybridized p-orbital accords with our conclusions for (SnMe,) cations discussed below.+ Me Me Results for the di-silicon derivative, (Me,Si), +,on the other hand, suggest that the SOMO comprises mainly the Si-Si a-b~nd.~~ This structure is then similar to that originally proposed for the (Me,C),+ cation,24 although again there are no detectable 29Si satellites so details of the SOMO cannot be derived directly. In contrast with the situation for the hydrocarbon analogue, however, all the protons were equivalent on the e.s.r. time-scale for this cation. Results for the corresponding germanium cations were very similar and these are clearly isostructural with their silicon counterparts.99~'00 This is in marked contrast with our results for SnMe,' cations, as outlined in the following Section. These are somewhat more informative because satellite features from magnetic tin isotopes were detected.13 Tin- and Lead-centred Cations Although tentative evidence for radical cations and anions was adduced in e.s.r. studies of the effects of ionizing radiation on a range of tin and lead alkyls,'0'*'02 the first definitive results were obtained using CFCI, solutions. '03 In collaboration with Hasegawa and his colleagues, we haye now studied the series of cations -SnH,+, MeSnH,+, Me2SnH,+, Me,SnH+, and Me,Sn*+.'04*105 The 99 J. T. Wang and F. Williams. J. Chem. Soc., Chem. Commun.. 1981, 176. loo B. W. Walther and F. Williams, J. Chem. Soc., Chem. Commun., 1982, 270. lo' S.A. Fieldhouse, A. R. Lyons, H. C. Starkie, and M. C. R. Symons, J.Chem.Soc., Dalton Trans.. 1974. 1966. lU2R.J. Booth, S. A. Fieldhouse, A. R. Lyons, H. C. Starkie, and M. C. R. Sym0ns.J. Chem.Soc.. Dalton Tmns.. 1976, 1506. 103 M. C. R. Symons. J. Chem. Soc.. Chem. Commun., 1982.869. A. Hasegawa. S. Kaminaka,T. Wakabayashi. M. Hayashi.and M. C. R. Symons. J.Cheni. SOC..I04 Chem. Commun., 1983, 1199. '05 A. Hasegawa, S. Kaminaka, T. Wakabayashi, M. Hayashi, M. C. R. Symons, and J. Rideout, J.Chem. SOC..Dalton Trans., 1984, 1667. Symons results reveal some remarkable trends. Although .SnH,+ is one of the few purely inorganic radical-cations studied in this way, I include it here rather than in Section 17 in order to illustrate the trends observed through this sequence of radicals.The dominating result for -SnH,+ cations is the large coupling to 'I7Sn and 'I9Sn nuclei, together with the large proton coupling (Figure 6). Two cations were formed in almost equal concentrations at 77K, one, with C,, symmetry, has a single strongly coupled proton and the other, with C,, symmetry, exhibits coupling to two equivalent strongly coupled protons. The C," cations converted slowly to the C," form on standing at 77K. The major cationic species formed at 77K from MeSnH,, Me,SnH,, and Me,SnH had apparent 'C,,' symmetry and all exhibited large tin hyperfine coupling constants indicating large s-p admixture in the SOMO. In marked contrast, the I7Sn/' 19Sn coupling for Me,Sn+ cations is relatively small, indicating effectively pure p-character.Studies using ,CH, indicated that this cation has extensive Me3Sn+---CH, character and this conclusion was supported by annealing studies (see below). The results nicely illustrate that the choice for all the cations between C,, and C,, distortion is a subtle one which may well be influenced by medium effects. Also, there seems to be no clear choice as to which of the two ligands, H or CH,, participates strongly in the SOMO. Thus the major species for MeSnH,+ and Me,SnH, cations had two strongly coupled protons, whereas that for Me,SnH+ + had no resolved proton splittings. We assume that this cation (Me,SnH+) must have pseudo C," symmetry since the results are so different from those for Me,Sn+ cations.Electron loss from the Td molecules SnH, and SnMe, is expected to occur from the triply degenerate t, orbitals and should be followed by a Jahn-Teller distortion. This theorem predicts D2d or C,, distortions, whereas we observe C,, and C,, distortions, with no clear evidence for D2d species. It is interesting to compare the C3"versions of SnH,' and SnMe,' The former has structure (33) according to the e.s.r. results, with cu. 70% spin-density on tin which retains a strongly pyramidal H,Sn-unit, as evidenced by the ca. sp3 hybridization. In marked contrast, the structure for SnMe,' seems to be that shown in (34) with a nearly planar Me,Sn- moiety and probably a nearly planar CH, unit, the major spin-density having shifted away from tin onto carbon.These results nicely illustrate the relaxations involved as the tin-ligand bond stretches. The extra driving force for the latter is presumably the extent to which hyperconjugative overlap stabilizes the Me,Sn + cation. HH IMe H (331 (34) I/ LrJ '17sn uI/ I/ 1I1 II Figure 6 First derivative X-band e.s.r. spectrumfor a solid solution of 4 rnol?; SnH, in CFCI, recorded (a) immediately afer exposure to 6oCoy-rays at 77 K,showing features assigned to C,, and CJrstructures of SnH,' radical caiions. and (6) after storage of the irradiated sample for one week at 77K Symons The extra stability of the ethyl radical causes Me,SnEt+ cations to give Me3Sn+ and .CH2CH3 radicals even at 77K.'04*'05However, only the parent cations, Me,SnCMe, +,were detected for the t-butyl derivative, although CMe, radicals are even more stabilized.We have discussed this result in terms of equilibrium [20] and have presented evidence for the reversible generation of CMe, radicals on annealing. [These are unable to migrate away from the Me&+ cations so the original fragments reunite on cooling.] We suggest that, curiously, it is the effect of methyl group repulsions that helps to stabilize the weak Sn-C bond relative to that for the ethyl derivative. Thus, as the system moves towards dissociation, the methyl group cannot swing inwards to give the planar radical and hence the activation energy cannot be lowered as much as for less restricted systems.However, the difference may be more apparent than real because of the reversible nature of the reaction for the -CMe, derivative and the irreversible break-down of the ethyl derivative. -I It is of interest to compare our results for SnH,' cations with those for the CH,+ and BH, cations discussed above. The key difference is that tin uses its s-orbital extensively for both forms of SnH,', whilst only 2p orbitals are involved for BH4.27u We suggest that this is because tin selects the 2a, SOMO (35) for which s-p mixing is expected, whereas *BH4 selects the 'b, SOMO which cannot mix with s. Alternatively, the differences may be connected to the comparable differences observed for =AR, radicals. Thus, for example, =BH3- radicals are planar,lo7 with a pure 2p SOMO, whilst *AIH,- radicals are pyramidal with considerable s-p hybridi~ation.'~'.O9 The Hexamethyldistannane Radical Cation.-Our results for this cation suggest strongly that the SOMO is confined to a (5p + 5p') combination, the two Me,% units being effectively planar with a stretched Sn-Sn bond.", This is the structure B.W. Walther, F. Williams, W. Lau, and J. K. Kochi, OrganomeraNics. 1983. 2. 688 lo' R. C. Catton and M. C. R. Symons, J. Chem. Soc. (A), 1969.2001. J. R. M. Giles and B. P. Roberts, J. Chem. Soc.. Chem. Commun., 1981, 1167. lo' M. C. R. Symons and L. Harris. J. Chem. Soc., Faraday Trans. 1. 1982.78, 3109. Radical Cations in Condensed Phases that I originally proposed for the Me,C-CMe, + cation, although it now seems that the two Me$ units are unable to attain complete planarity.Walther er a1.1°6have shown that the tin coupling is small for the (Me,Sn*GeMe,)+ cation, so the Me,Sn unit, and probably the GeMe, unit, are also planar for this cation. Lead-centred Radical Cations-These have so far been less extensively studied. Walther et al. reported the spectrum for *PbMe4+ cations which shows that the structure must be very similar to that for .SnMe,+ cations, with the expected larger shift in g because of the larger spin-orbit coupling constant.lo6 Several common elements emerge from these studies of organic radical cations having their SOMOs formally localized on one hetero-atom. (i) If that atom is from the second or subsequent rows of the Periodic Table and forms one, two, or three o-bonds only, the cation shows a marked tendency to 'dimerize' to give o*dimer cations.This tends to localize the unpaired electron to the (T*bond. (ii) Nitrogen- and oxygen-centred radicals do not normally react in this way. This is, in part, because of their far greater tendency to extract hydrogen. (iii) Delocalization onto neighbouring alkyl groups can be extensive, especially for first-row cations. This is particularly marked for acetal cations. (iv) Localized o-radicals having the SOMO confined largely to the X-X o-bond are formed when Group IV dimeric molecules, R,X-XR,, lose an electron. (v) Electron-loss for aldehydes and ketones is from the non-bonding, in-plane 2p orbital on oxygen, and hyperconjugative delocalization is extensive. However, for the esters there seems to be a fine balance between this orbital and the II-orbital centred largely on the two oxygen atoms.Also, the internal transfer of hydrogen from the ester alkyl group to the carbonyl oxygen seems to be remarkably facile. Aromatic Cations The pioneer studies of aromatic cations were carried out many years ago, the first being on anthracene cation^,^" 'O.' '' which are remarkably stable. The first indication of an important positive-charge effect on methyl-group hyperconjuga- tion came from a study of the 9,lO-dimethylanthracene cation,'" and the first study of alkyl-substituted benzene cations was our work on hexamethylbenzene cations in sulphuric acid.' l3 A recent study of methyl-substituted benzene cations extends these studies,' l4 but in pride of place must go the most interesting study of Iwasaki et al.on the benzene cation.' ' 'lo J. R. Bolton and G. K. Fraenkel, J. Chem. Phys., 1964.40. 3307. 'IL A. Carrington, F. Dravnicks. and M. C. R. Symons, J. Chem. Soc., 1959. 947. 'I2 J. A. Brivati. R. Hume, and M. C. R. Symons, Proc. Chem. Soc,., 1961, 384. R. Hume and M. C. R. Symons, Proc. Chem. Soc.. 1963, 241; J. Chem. Soc., 1965, 1220. 'I4 M. C. R. Symons and L. Harris, J. Chem. Res. (S).1982, 268; (M) 1982, 2746. 'I5 M. Iwasdki. K. Toriyama, and K. Nunome, J. Chem. Sot.., Chrm. Commun., 1983, 320. 428 Symons The Benzene Cation.-This, like the anion, is an orbitally degenerate system so distortion must occur on electron loss.Many people have studied the dynamic Jahn-Teller distortion for C6H6 -anions but liquid-phase studies are complicated by the fact that exchange is extremely fast and perturbing effects from the cations may seriously interfere with the true behaviour. However, even when C6H6- ions are generated in solids by electron addition and cooled to very low temperatures, any inequivalence of the proton coupling constants is lost within the linewidths. In marked contrast, the distortion experienced by the cations is very large, as has been recently established by Iwasaki and his co-workers.' At 4.2 K in freon, the e.s.r spectrum comprises a major triplet from two equivalent, strongly coupled protons (A = 8.2G)and four weakly coupled protons (A = 2.4G).On annealing, the distortion becomes dynamic and the six couplings are averaged. Clearly, the symmetric orbital (36) has been selected. This suggests that there has been a distortion which shortens the C(2)-C(3) and C(5)-C(6) bonds and stretches the remainder, as indicated in (36), since this stabilizes vAand destabilizes ws. The interesting question as to why this turns out to be the preferred mode of distortion is not yet answered. Alkyl Substituted Cations.-The results obtained with freon solutions 14v1 confirm, in general, those previously obtained. They show that the orbital selected for the SOMO is that for which electron-release from the C-H o orbitals is maximized (bl, vS).Furthermore, proton splitting from methyl groups is very large, showing that such electron-donation really is important.There seems to be no need to postulate admixture of yAand wS of the type invoked to explain the e.s.r. spectra of the corresponding anions. An unexpected result was obtained for ethyl benzene cations,' l6 in that a major triplet splitting of ca. 29G was observed. This surely means that the preferred conformation is close to that in (37) which is expected to be the least favourable conformation on steric arguments. Having argued that this nicely illustrates the concept that C-H hyperconjugative electron-release is more important than that for C-C bonds, we then found that for thepara-diethyl derivative, the major species had four protons with coupling constants of only ca.11G, suggesting the reverse as in (38) for both ethyl groups.' '' We do not understand why this switch occurs but it certainly establishes that the choice is a subtle one. 'I6 D. N. R. Rao, H. Chandra, and M. C. R. Symons, J. Chem. Soc., Perkin 2, 1984, 1201 11' D. N. R. Rao and M. C. R. Symons, J. Chem. Soc., Perkin Trans. 2, in press. Radical Cations in Condensed Phases HI I HI I (37) (38) Cations with -C(O)R Substituents.-In general, results were unexceptional, the -C(O)R groups being found at sites of high electron density, although delocalization thereon was relatively small. However, the spectrum for benzaldehyde cations was unexpected in indicating an extra splitting of ca.12 G in addition to that assigned to the para-proton. This is tentatively assigned to coupling to a I9F nucleus of a solvent molecule but why this particular aromatic cation should uniquely opt for such an interaction is quite unclear to us. What is certain is that the n(0) orbital is not selected as the SOMO, since a search for species exhibiting a doublet splitting in the 100-1 5OG region was unsuccessful. This is interesting since calculations suggest that the IT and n.b. orbitals should be very close together and at least one group predict the SOMO to be the n.b. orbital on oxygen. For the ester cations, [ArC(O)OR) +,we detected small splittings from certain protons in groups R which, by comparison with our results for saturated ester cations gave spin-densities on the -OR groups of ca.24%. Results for various cyano-derivatives were similar and require no further comment. Coupling to 14N nuclei was not resolved and it seems that delocalization is small. Styrene and Related Cations.-The e.s.r. results show that the SOMO for cations of styrene, methyl styrene, and ethyl cinnamate are very similar. There is an estimated spin-density on the outer (CH,) carbon of ca. 3540%. Nitrobenzene and Related Cations.-Our results for PhNO,' cations (4 H, A 5 6 G) are in good accord with expectation for the b, (vA)orbital with a node through nitrogen. However, the directing power of the -NO, group is less than that of a -CH, group since, for p-nitrotoluene cations, the SOMO has switched from the wA orbital to the ws (a,) orbital.Nevertheless, there was no detectable delocalization onto I4N, in marked contrast with the SOMO for the corresponding radical-anions which has a high spin-density on the NO, group. An unexpected result was the detection of a spectrum similar to that for *NO, radicals from aromatic nitro-cations on annealing. These spectra were very similar to those obtained from (RNO,)+ cations (R = alkyl) and, presumably, have the same structure (see Section 10). This rearrangement requires a more drastic electronic change than that for the alkyl derivatives which accords with the higher temperatures required for the reaction. Symons Aryl Nitroso-cations.-Our results for PhNO' cations [A,,('4N) = 50G, A, = 30 GI' l8 give an isotropic coupling constant (36, 7G) close to that assigned to this cation in the liquid-phase.' l9 This result nicely confirms the previous assignment and supports the in-plane, non-bonding structure for this species (39).These results give a p:s ratio of ca.6 and an estimated spin-density on nitrogen of ca. 47%. (39) Halogenobenzene Cations.-As expected, the SOMO selected for the mono-substituted cations is the 'a2'orbital, placing high spin-density on the halogen.' 2o Thus, the ability of the halogen atoms to act as n-electron donors is more significant than their high electronegativities. Nevertheless, this property controls the extent of delocalization, as can be seen in the plot shown in Figure 7. Delocalization onto fluorine (cu.8%) is a little less than that onto oxygen for PhOMe+ cations, the estimated order being F > OR 3 C1 > Br > I z NR,. ee A ee ... 111 2.2 -I" H. Chandra and M. C. R. Symons, unpublished results. G. Cauguis, M. Genies, H. Lemaire, A. Rassat, and J. P. Ravel, J. Chem. Phys., 1967, 47. 4642. M. C. R. Symons, A. HaSegdwa, and S. P. Maj. Chem. Phjx. Lett.. 1982.89.254;J. Chem.Soc., Fororlay Truns. 1. 1983, 79. 1931. 431 Radical Cations in Condensed Phases The extent of delocalization varies with the nature of the para-substituents (X) for XC6H4hal+ cations. For example,p-Me groups cause a small reduction in spin- density on halogen (ca. 2%). This is more than outweighed by the large delocalization (ca. 8%) onto the methyl group so the effect of methyl is largely to remove spin-density from the ring.It is noteworthy that para-OR substituents greatly reduce the spin-density on bromine, the reduction being larger than that caused by a second bromine substituent. This result is in contrast with the delocalizing order given above. Most interesting is the large increase in spin-density on bromine induced by a para-nitro-group (ca. 5%). The nitro-group is a strong electron-acceptor and since the bromine atom directs the SOMO so as to place the nitro-group at a position of high positive-charge, the bromine feeds n-electrons into the ring to reduce this enforced constraint as far as possible. l7 Benzyl Derivatives.-These are of interest because of the range of conformations available to the -CH,X group, which may be largely electronically or largely sterically controlled.Generally the SOMO is the 'a,' orbital placing high spin and charge density on the >C-(CH,X) ring carbon atom. Hence it is of interest to compare the e.s.r results with those for the neutral radicals R,CeH,X.'It has not, in general, been possible to make this comparison with the corresponding radical anions since the SOMO is then yA(bl), with a node passing through the -CH,X group. As with the neutral radicals,12'"22 we had expected to find that 'heavy atom' groups such as C1, Br and I would favour the extreme out-of-plane site, with 1 17.123 This was fulfilled for X =Br but not for (PhCH,Cl)+. The former give two species, both having large hyperfine coupling to bromine.However, the latter gave no detectable coupling to chlorine but large coupling to the methylene protons, suggesting b90". Before drawing any firm c+onclusions from this, however, we note that the para-methyl derivative MeC,H,CH,CI gave the expected large chlorine coupling, suggesting that, as expected, b0.*'Thus, as with the ethyl derivative discussed above, there is a subtle balance of electronic and steric factors controlling the value(s) of 8adopted by a particular cation. What is clear is that hyperconjugative delocalization onto halogen, when it occurs, is large when 8is small. 15 Five-membered Hetero-aromatic Cations Both Shiotani and co-workers and my own group have shown that radical-cations of pyrrole, furan, and thiophene derivatives are readily formed in freon solvents and are well characterized by their e.s.r.spectra.'24,'25 In all cases, the SOMO is unambiguously the la, R orbital having a node through the hetero-atom, as can be "'M. C. R. Syrnons, J. Chem. SOL...Furaduy Trans. 2, 1972, 1897. A. R. Ly0ns.G. W. Neilson, S. P. Mishra, and M. C. R. Symons, J. Chem.Soc.. Faradaj. Trans. 2,1975, 363. IL3 S. P. Maj, D. N. R. Rao, and M. C. R. Syrnons, J. Chem. SOL..,Perkin Trans. 2. in press. M. Shiotani, Y. Nagata, M. Tasaki, and J. Sohma. J. Ph.r.s. Chem., 1983,87, 1170. D. N. R. Rao and M. C. R. Syrnons, J. Chem. Sor.. Perkin Trans. 2. 1983, 135. Symons judged from the selected data in Table 6. We have pointed out that the ring-proton hyperfine coupling constants are similar to those the radical cation of cyclopentadiene for which a similar SOMO is selected.Also, the large proton hyperfine splittings observed for the 2,Sdimethyl derivatives are significantly larger than those for 1,3-dimethyl cyclopentadienyl radicals (ca. 18G uersus ca. 13.5 G). This increase for the cations is probably largely due to the positive charge effect. The value quoted for the dimethyl cyclopentadienyl radical is a limiting value for the pure la, orbital, so that partial population of the alternative x-orbital, thought to be important for substituted cyclopentadienyl radicals,' 26 is not responsible for this difference. Table 6 Hyperfine coupling constants of radical cations derived from pyrrole, furan, and thiophene derivatives and of related radicals Radical Hyperfine Coupling ConstantslG" X R Hz.5 H3.4 H(CH3) Other nuclei S H 13 2.5 S CH3 3.5 18.1 0 H 14 3.5 NH H 16 3.0 ca.3(I4N) NH CH3 3.5 17.5 ca. 3(14N) NCH3 H 15.5 3.6 ca. 3.5(I4N) CH2 CH Hb H' 11.6 13 3.5 3.45 < 2(CH2) 1.O(CH) CH CH 3'.d 3.7 13.5 l.l(CH) "G = 10 T. bT.Shida, Y. Egawa, H. Kato, and H. Kubodera, J. Chem. Phys., 1980,73, 5963. 'A. G. Davies. E. Lusztyk, and J. Lusztyk, J. Chem. Soc., Perkin Trans. 2, 1982,729. These are limiting values. Our hope that, for thiophene, there might be some cations having their unpaired electrons on sulphur in the n(o)-orbital was not fulfilled. It seems that it is not possible for this five-membered ring to undergo sufficient angle distortion to lower this orbital below the 7c-level.Pyridine and Related Cations.-The Pyridine Cation. In one of the earliest reports of e.s.r. studies of radical-cations formed by radiolysis. Shida and his co-workers 127*128 established that the ground-state SOMO for the pyridine cation is the n(al)orbital on nitrogen (40) rather than one of the aromatic x-orbitals. This was not an obvious result since the x orbitals are close in energy to the n(al)orbital and theoreticians had not reached agreement as to the best choice. The e.s.r. results are definitive but I stress that they relate to the relaxed cation rather than to that formed on vertical ionization.As I show below, I think that bond-angle relaxation is extensive for these ions. P.J. Barker, A. G. Davies, and M.-W. Tse. J. Chem. Soc., Perkin Trans. 2. 1980, 941 "'T. Kato and T. Shida, J. Am. Chem. So(,..1979, 101, 6869.''' T. Shida and T. Kato. Chem. Phys. Lett., 1979, 68, 106. 433 Radical Cations in Condensed Phases 2SYO 75 H H 11 2 20 (40) E.s.r. results shown in structure (40) make an interesting comparison with those for the isoelectronic phenyl radical (41).’ 29 The approximate p:s ratio, deduced in the usual way from the 14N and 13C hyperfine tensor components, is ca. 9 for (py)’ but ca. 5 for phenyl, the total spin-density on N and C being ca. 75%. This suggests that the desire for more linear co-ordination for nitrogen is greater than that for carbon. This accords with the greater delocalization onto the ortho C-H groups for PY+.Substituted Pyridine Cations. Our aim in studying these cations was to try to modify the energy of the SOMO resulting from x-electron loss so that this was favoured over loss from the n-orbital.130 We succeeded with the cations shown in (42), (43), and (a),which gave only x-cations. For the 3-chloro-derivative both 71-and n-cations were detected. This is a particularly interesting result, since it is very rare that species which can be interchanged by simple electron-transfer are found to co- exist. We think that they are probably formed together initially and that the major ring distortion of the n-cation then provides a large barrier to interconversion.@Me Me Br (42) (43) (441 Diaza-derivatives. The definitive work is again that of Shida and Kato.I2’ Results for the 1,2- and 1,4-derivatives are clear-cut but I am uncertain about the 1,3- derivative. The natural assumption is that the SOMO is distributed between the two in-plane n-orbitals on nitrogen and this is clearly established for the 1,2-and lP-derivatives by the e.s.r. spectra. However, we find that the spectrum for the 1,3- derivative is very similar to that of the pyridine cation with one strongly coupled nitrogen and is not well accommodated by the concept of two equivalent nitrogen nuclei.’ * P. K. Kasai, E. Hedeya. and E. B. Whipple. J. Am. Chem. Soc., 1969, 91, 4364. 130 D. N. R.Rao. G. W. Eastland. and M. C. R.Symons, J. Chem. SOC.,Furuciuy Trans. I. in press. 131 D. N. R. Rao, G. W. Eastland. and M. C. R.Symons, J. Chem. Soc., Furuduy Trans. 1. in press. Symons A possible explanation lies in the preferred method of distortion. For the 1,2-derivative delocalization is unavoidable and for the 1,4-derivative a tendency to flatten can occur at both nitrogen atoms. However, we suggest that for the 1,3-derivative a distortion in which one CNC angle increases whilst the other decreases may be preferred, thereby localizing the SOMO.Given that interaction between the two nitrogen n-orbitals is small, this could well occur. The situation is somewhat reminiscent of the very ready distortion of m-dinitrobenzene anions from the symmetrical structure by solvent perturbations 132 and of that for (R,&CH,NR,)+ cations discussed in Section 10 above.16 Vinyl Cations This section links closely with Section 4 on alkene cations and the same problems connected with the degree of twist in the cations persists. Most work has been on (H,MHX)+ cations but there has been some work on di- and poly-substituted derivatives also. In some cases, the e.s.r. results are ambiguous and, certainly, more work is needed. Thus, for methyl acrylate and, especially, methyl methacrylate, the spectra are difficult to interpret. Tabata and Lund obtained a 25G triplet for the former cation 133 but we have obtained a 12G triplet in addition at very low concentrations.' 34 Whichever is correct, such a simple spectrum is surprising since it implies very little delocalization, whereas significant spin-density on the CH, unit and on the ester moiety might be expected since the ionization potentials of the separate units are comparable.However, for the methacrylate cation both groups have established that replacing the -OCH3 group by -OCD, has no effect on the e.s.r. spectrum. Thus, unless some rearrangement has occurred (which remains obscure), it seems that there is again little delocalization into the ester group. Unfortunately, the task of unravelling the very complicated spectrum has not yet been accomplished. The results may mean that the degree of twist is large for these cations.'34 The cation of acrolein (45) makes an interesting contrast '33 since, in this case, electron loss is from the in-plane oxygen orbital, as evidenced unambiguously by the large ( 125 G) 'H hyperfine coupling.13' D. Jones and M. C. R. Symons, Trans. Furuday Soc., 1971,67,961. 133 M.Tabata and A. Lund, Chem. Ph.v., 1983,75, 379. G. w.Eastland, Y. Kurita, and M. C. R. Symons, J. Chem. Soc.. Perkin Trans. 2, in press. Radical Cations in Condensed Phases The spectrum for acrylonitrile cations suggests coupling of 24G to two protons, ca. 12G to a third and parallel coupling to nitrogen of ca. 20G.'34 Thus, in this case, the expected n-delocalization seems to be occurring although the large methylene proton coupling may imply a considerable degree of twist. For vinyl bromide cations we obtained a well-defined bromine splitting characteristic of an a-bromine structure. Also, there was a poorly resolved 15G triplet tentatively assigned to the CH, protons.The estimated spin-density on bromine was ca. 39% which is remarkably large [cf. ca. 18% for R,C-Br radicals and 30%for (PhBr)' cations]. If the radical is planar, the estimated spin-density on the CH, unit is ca. 65% but if it is twisted these numbers have no significance. We have suggested that the 19-20G triplet observed for the cations (H,c-C--OR)+ and also (H,c-CH=OCOMe)+ have structures close to the limiting valence-bond structures indicated. This must be true if the ('H) coupling is negative but if, by any chance, the cations are strongly twisted it might be positive, in which case this argument is again meaningless.Results for various -SiMe3 derivatives are of particular interest since they require that the cations be twisted.'34*'35 For the vinyl derivative we found two protons with A = 40G and one with A = 18G. The former must be positive but this is not certain for the latter. These results imply considerable twisting with a tendency to favour the limiting structure (46). This is expected to be favoured since P-Me,Si groups can thereby stabilize the cationic H,C+ unit by cr-n overlap to a maximum extent. One interesting aspect of our results is the ready breakdown of these ions to give methyl radi~a1s.l~~This implies that the cation H,C=CH-SiMe, or some rearranged or cyclized version must have a relatively + high stability.H H Results for the cations of ally1 chloride and bromide are of interest since they clearly establish that the halogen adopts the out-of-plane site favoured both for steric and electronic reasons. The chlorine and bromine splittings were large, indicating extensive delocalization (ca. 23% on chlorine and 29% on bromine). Triplet splittings of ca. 16G indicate a spin-density of ca. 70% on the CH, moiety -again if we assume no twisting for the cations. The styrene cation [a(CH2) = IOG] has been discussed above. Results for 2-vinyl pyridine cations show that these are structurally similar but weak outer '" M.Kira. H. Nakazawa, and H. Sakurai, J. Am. Chem. Soc., 1983. 105. 6983. Symons features suggested the presence of a second species.Similar outer lines were favoured for 4-vinyl pyridine cations, the spectrum being very similar to that for unsubstituted pyridine cations, having the SOMO strongly confined to the n(o)N orbital (see Section 15 above).' 34 The former results represent another example in which two alternative structures, not involving major rearrangements, have been stabilized in a single matrix. Clearly interchange between them must be very slow. 18 Inorganic Cations This is, perforce, a surprisingly short section. Apart from various small cations formed in inert-gas matrices by Knight and his co-workers and the SnH,+ prepared in CFCl,, discussed above (Section 13), there has been little progress. We have prepared N204+cations in CFCl, but several other attempts to obtain good e.s.r.spectra for various inorganic cations have failed. For example, we have been unable to detect any resonance features for BrZ+ or I,' despite the fact that their formation is indicated by optical spectroscopy. Also, attempts to prepare cations of various neutral transition-metal complexes have largely failed, although some evidence for the formation of Fe(CO),+ from Fe(CO), in CFCI, was forthcoming. '36 The Cations CO+ and N, +.-These have been isolated in inert-gas matrices at CQ. 4 K after formation by photoionization just prior to, or during, deposition. lo*' 'This work is an extension of techniques developed especially by Lester Andrews for the preparation of cations for spectroscopic studies.For 13CO+ a very clean spectrum was obtained using a neon matrix at 4K (Figure 8)." Thus the e.s.r. parameters given in Table 7 are very precise. Also given are the gas-phase results of Pilch et al.' 38 Agreement is good, so one can reasonably deduce that there is very little perturbation from the matrix in this case. The Table 7 Ex. parameters for some inorganic radical cations in rare-gas matrices (refs. 10, 11, 139) g-Values Hyperfine coupling constanrslM Hz Cation MediumlT g, g, g= A, A,, A= Ai,, 2B 13CO+" Ne/4 2.004, 1.9996, 1.9996 1665, 1 527, 1 527, 1 573, 92 13co+b gas 1 506, 96.4 '*N2 Ne/4 2.0004 (av) 104.1,+ HzO+ 2.0093 (av) 73.7 ('H), 83.5 (''O), "Ref. 11. 'Ref. 139. 136 B. M. Peake, M. C. R.Symons.and J. L. Wyatt, J. Chem. Soc.. Dalton Trans.. 1983, 1171. 137 L. Andrews, Ann. Rev. Phys. Chem.. 1979.30, 79. N. D. Pilch, P. G. Szanto. T. G. Anderson, C. S. Gudeman, T. A. Dixon, and R. C. Woods, J. Chem. Phys., 1982, 76, 3385. Radical Cations in Condensed Phases 1 1 3llOG ' 369 OG 1-12 1 Figure 8 Lou-and high-jield e.s.r. potvder pattern spectrum of '3CO+in neon matri-u at 4 K. (Takenfrom ref: 10) parameters give calculated 2s and 2p, populations of 43% and 42% respectively. Thus the simple representation of the SOMO as a non-bonding spz hybrid orbital on carbon is nicely supported. In marked contrast, the N2+ cation was found to be rotating even at 4K in neon." It is not clear why this should be when 'CO is so firmly fixed that libration is apparently unimportant.Furthermore, the 14N isotropic coupling corresponds to only ca. 6% 2s character on each nitrogen. Hence the 2p character must be ca. 44% suggesting that, in this case, the SOMO is primarily 2p in character. Such a drastic switch seems very surprising to me but the results are convincing. H20+ and NH,+.-Knight and Steadman have reported a very thorough study of the water cation in many of its isotopic forms (H20+, HDO+, D,O+, and H2'70+).139 The results present no surprises but it is good to have a firm identification of this important cation and to have some precise data. The results for both this cation and -NH,+ cations in neon at 4K l1 show that they are apparently freely rotating. The only previous reports of e.s.r.studies of the H20+ cation for which reasonable parameters were forthcoming was our own for H,O+ in aqueous acid media 140 and one of a species thought to be H20+ formed in y-13' L. B. Knight and J. Steadman, J. Chem. Phys., 1983, 78, 5940. I4O T. A. Claxton. I. S.Ginns, M. J. Godfrey, K. V. S. Rao. and M. C. R. Symons. J. Chem. Soc.. Furadq~ Trans. 2. 1973, 69,217. Symons irradiated beryl.'41 It is interesting to note that, in our case, the relatively large g-shift (2.0093)for H20+ in neon was not observed. This probably reflects the effect of strong hydrogen bonding in the aqueous matrix. Results for .NH,+ agree very well with the many previous reports for this species which is nice confirmation that these results are indeed reliable.It is often implied in the literature that large environmental interactions render results for radicals in matrices other than the inert gases unreliable but, in general, my experience is that this is not usually the case. The N204+ Cation.-We have recently reported our results for this cation in a freon matrix.'42 We had previously suggested that cationic species were formed in an irradiated single crystal of N204+,14,and Morton et a!. tentatively identified a species formed from N204 in SF, as a nitrosyl nitrate ion, ONON02*.'43 The results using CFC13are clear-cut -the species detected is not the o-radical (O2N*NO,)-tentatively identified in N204 crystals but the rearranged cation 02N+---ONO, which can be viewed as an -NO, molecule relatively weakly bound to a linear NO,+ cation.The e.s.r. results are quite unambiguous; not only are the features for the 'NO,' unit similar to, but quite distinct from, those for genuine -NO,, but also weak interaction with 14N of the NO,' unit was res01ved.I~~I think that the driving force for this change lies in the high stabilities of both NO2+ and NO2. It is important that the theoretical calculations of Yoshioka and Jordan actually predicted that this should be the most stable form of the ~ati0n.l~~ Clearly the species reported by Morton et al. was the same species. 43 19 Conclusion I conclude that in a remarkably short period a new Chapter in e.s.r. spectroscopy has largely unfolded. I hope that the techniques involved will now be treated as standard and that anyone interested in the structure and reactivity of radical- cations will consider using this approach.Finally, I wish to thank my students (Dr. D. N. R. Rao, B. W. Wren, P. J. Boon, G. D. G. McConnachie, and Dr. H. Chandra) for all their help, Professor T. Shida and Dr. M. Iwasaki for many helpful discussions and, especially, Professors Akinori Hasegawa and George Eastland for extensive and continuing collaboration. I would also like to thank Miss V. Orson-Wright and Mrs. C. A. Crane for their valuable help in preparing the typescript. * Note added in proof.It now seems probable that all the ester cations studied at 77 K (p. 417) are not the n cations. but are rearranged cations of this type. (M. Iwasaki, H. Muto. K. Toriyama, and K. Numone, Chem.Phys. Lett., 1984, 105. 586.) 141 M. I. Samoilovich and A. I. Novozhilov, Zh. Neorg. Khim.. 1970, 15, 84. 142 D. N. R. Rao and M. C. R. Symons, J. Chrm. Soc., Dalton Trans., 1983, 2533. 14' J. R. Morton. K. F. Preston, and S. J. Strach, J. Ph!ps. Chem., 1979, 83. 533. 144 Y. Yoshioka and K. D. Jordan, J. Am. Chivn. Soc.. 1980. 102, 2621.

ISSN:0306-0012    

DOI:10.1039/CS9841300393

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

Aromatic Benzene Compounds from Acyclic Precursors By P. Bamfield and P. F. Gordon IMPERIAL CHEMICAL INDUSTRIES PLC ORGAN I CS DIVISION, BLACK LEY, MANCHESTER M9 3DA 1 Introduction The synthesis of substituted aromatic hydrocarbons has been a feature of synthetic organic chemistry from almost its beginnings early in the nineteenth century and owes much to the pioneering work of A. W. Hofmann from 1845 onwards. At that time coal tar provided a cheap and plentiful supply of benzene as well as several other aromatic hydrocarbons and so the early synthetic routes to substituted benzenes almost inevitably started from these cheap feedstocks. The required substitution patterns were then achieved by subjecting the appropriate hydrocarbon to a series of stepwise reactions, e.g.nitration, sulphonation, chlorination, reduction, oxidation. Consequently, a vast body of literature is now available concerned with both nucleophilic and electrophilic substitutions at the benzene ring and some very elegant and effective syntheses of substituted benzenes are to be found. Hence, today, ring functionalization is the most important method for preparing substituted benzenoids. Ring functionalization is not the only synthetic strategy that can be envisaged since it is also possible to construct the ring with some or all of the substituents already in place starting from acyclic precursors. This strategy can best be seen in the field of heterocyclic synthesis where many heterocycles, e.g. pyrroles, pyridines, and pyrimidines, are often best prepared from acyclic precursors.* It is not surprising therefore to find that benzene compounds can also be prepared by the same strategy and this is most emphatically confirmed in biochemical synthesis where nature is a most elegant practiser of the ring synthesis method, e.g.the synthesis of anthranilic acid. Several advantages can be seen in this approach such as the preparation of highly substituted compounds in only a few steps and the avoidance of ortho-meta-para mixtures common in conventional aromatic synthesis. Generally, highly substituted benzenes require long synthetic routes by conventional methods and this tends to negate the advantages of using cheap and readily available starting materials.Ring synthesis can also provide substitution patterns not easily attained by conventional synthesis and therefore new compounds previously untested, in for example pharmaceutical and plant protection products, could be made available. In addition, the introduction of labels (13C, 14C)into aromatic compounds, not an easy task by conventional synthesis, should be easier by ring synthesis in view of the greater availability of labelled acyclic compounds. * In this review heterocycles will therefore be considered as masked acyclic compounds. Aromatic Benzene Compounds, from Acyclic Precursors 2 Scope and Organization The scope of this review has been limited so as to keep its length within reasonable bounds. Therefore, only syntheses of benzene rings are considered; however, several examples of condensed-ring systems are presented but only where the additional rings are non-aromatic.Reactions involving cycloadditions, the most important of which is the Diels-Alder reaction, have been omitted since many reviews already cover this area.’ The vast majority of the syntheses considered here, therefore, involve condensation reactions (Aldol, Claisen erc.) and addition reactions, e.g. the Michael addition. In some sections a few aromatization reactions have been highlighted, although this topic in itself is a subject worthy of a major review. The review has been organized into four sections (Sections 3-6) depending upon the nature of the substituents in the benzene ring. The first section (3A-C) contains all those benzenes with at least one hydroxy-group and necessarily contains hydroxybenzenes with other groups, e.g.cyano, ester etc. The second section (4)contains all those benzenes with an amino-group, but not an amino- and a hydroxy-group since these are to be found in section one. The third and fourth sections (5 and 6) describe benzenes containing electron acceptors and alkyl groups only, respectively. Hence the synthesis of a benzene containing cyano-, hydroxy-, and amino-groups will be found in the first section. This classification is rather arbitrary but classing the syntheses by functionality probably presents the material in a much more useful format for the chemist who is looking for an alternative to a ring functionalization approach.The use of the alternative fragment approach, e.g. 4 + 2, 3 + 3 etc., or an approach which considers reactive centres in the acyclic precursors have therefore been omitted as being of less use in meeting the desired aims of this review. 3 Hydroxybenzenes This section contains the largest number of references and has been further split into three categories; phenols, dihydroxybenzenes (catechols, resorcinols, hydroquinones), and benzenes containing more than two hydroxy-groups. A. Phenols.-In the ring synthesis method the phenolic hydroxy-group usually comes from the carbonyl oxygen of either a ketone or an ester group and with the abundance of different ketones and esters a wide range of phenols can be synthesized, as the following text illustrates.Ketones usually react as three-carbon components uia the two active and nucleophilic x-positions. The co-reactant therefore must also be a three-carbon component but with two sites of complementary activity, i.e. electrophilic. The most common co-reactant is a p-dicarbonyl compound, or synthetic equivalent thereof, and a large number of reactions of this type have been reported. ’ (a)H.Wollweber,‘Diels-Alder Reaktion’,G.T. Verlag.Stuttgart, 1972;(b) S.Danishefsky, Act.Chem.Res., 1981, 14, 400;(c)G. Brieger and J. N. Bennett, Chem. Reu.. 1980,80, 63, and references cited therein. Barnfield and Gordon Malondialdehydes are the simplest P-dicarbonyl compounds and have been studied for nearly a century.They react with various ketones to give the corresponding phenols in fair to good yields as shown in Table 1. In particular the condensation of malondialdehydes carrying electron-withdrawing substituents has received close attention from a number of workers. Thus, Hill and co-workers did much of the pioneering work on 2-nitromalondialdehydes around 1900. Unfortunately, the synthetic utility of this approach is severely limited by the chemical instability and expense of many malondialdehydes and the preparation of Table 1 Condensation of malondialdehydes R &R1 OHC-CHX-CHO t RCH~C(O)CH~R~---9 y Entrv X R R' Yield (%) Ref: Comments 1 H C02R CO2R 41 2 2 NO2 H H 64 3u-d 3d 14C labelled acetone used 3 NO2 H Et 70 4 4 NO2 H CHzCOzH 82 3c 5 NO2 H Ph 88 Sub 6 NO2 H OPh 78 6 7 NO2 Me Me 94 4 8 NO2 Ph Ph 95 5a,4 9 NO2 Me Et 74 4 10 NO2 -(CH2).-16-71 7a,h 11 NO2 -CHzkO(CH 2)s- 88 8 12 NO2 H C02H 90 3c 13 NO2 CO2H CO2H 90 3c,4 14 NO2 H CH2COMe 17 9 major product 2,3-diacetyl- 5-nitrocyclopenta- 1,3-diene 15 CN C02Et COl Et 10 16 N=NAr H H 51-82 1 lu.6 17 N=NAr Me Me 74 1 la,h 18 COPh CO2R CO2R 77 12 19 CO2Et CO2R C02R 50 12 20 Ph C02R CO2R 48 12 21 Br CO2R CO2R 53 12 'S.H. Bertz, W. 0.Adams, and J. V. Silverton, J. Org. Chem., 1981,46, 2828. (a)H. B. Hill and J. Torray, Am. Chem.J.. 1899,22,892; (b)ibid., Chem. Ber., 1895,28,2597; (c)H. B. Hill, C. A. Soch, and G. Oenslager, Am. Chem. J., 1900,24, I; (d) N. T. Hales and H. Heaney, Terrahedron Lett., 1975, 4075.E. C. S. Jones and J. J. Kenner, J. Chem. SOC.,1931. 1842. (a) H. B. Hill, Chem. Ber., 1900.33, 1241; (b)H. B. Hill and W. J. Hale. Am. Chem. J., 1905.33, 1. T. R.Govindachari, S. Prabhakar, P. S. Sdnthanam, and V. Sudersanam. Indian J. Chem., 1966,4,433.'(a)V. Prelog and K. Wiesner, Helv. Chim. Am, 1947,30,1465; (h)V. Prelog, K. Wiesner, W. Ingold, and 0.Hafliger, Helv. Chim. Acta, 1948, 31, 1325. * V. Prelog, M. F. El-Neweihy, and 0.Hafliger, Helv. Chim. Acta, 1950, 33. 1937. W. J. Hale, Chem. Ber., 1912.45, 1596. lo C. Reichardt and K. Halbritter, Angen.. Chem., Int. Ed. Engl.. 1975, 14, 86. I' (a) D. Leuchs, Chem. Ber., 1965.98, 1335; (6) H. R. Hensel, Chem. Ber., 1964,97,96. "V. Prelog, J. Wursch, and K.Konigsbacher, Helv. Chim. Acta, 1951. 34, 258. 443 Aromatic Benzene Compounch .from Acyclic Precursors nitrophenols by conventional aromatic synthesis is so well known that it makes the ring synthesis approach rather unattractive for all but the most difficult phenols, e.g. entries 2, 10, and 11. Pyrimidines such as (1) 13*14 and (2) 15*16 react with various ketones (RCH,C(O)CH,R') to give the corresponding 4-substituted phenols (3) in acceptable yields. The pyrimidines thus act as protected malondialdehydes as do the +unsaturated aldehydes (4) which react with the dianion of ethyl acetoacetate to give the corresponding 4-substituted salicylates in good yield. The highly reactive fluoromalondialdehyde equivalent (5) reacts similarly to give a 4-fluorophenol with diethyl acetonedicarboxylate.' * Me x R = H,OH X = N02,COzR, CONHR X x = NMe2,Cl BFL-R = alkyl ,aryl Phenols substituted at the 3/5positions are synthesized from a ketone and either a P-ketoaldehyde or a P-diketone.Yields can be good (Table 2) though generally they tend to be lower than with malondialdehydes. Nevertheless, ketoaldehydes and diketones are usually more stable and easier to handle than malondialdehydes. They also offer a more useful alternative to conventional aromatic synthesis because of the possibilities of preparing unsymmetrical and highly substituted phenols which may be otherwise difficult to prepare. As Table 2 shows, there is considerable flexibility in the method so that various groups can be introduced into the phenol.However, a problem arises when unsymmetrical P-diketones are used since it is now possible to generate two isomers (equation 1). J. J. Fox. T.-L. Su. L. M. Stempel, and K. A. Watanabe, J. Org. Chem., 1982, 47, 1081. 14 H. C. Van der Plas and P. Barczynski. Red. rrav. Chim. Paps-Bas, 1978. W,256. K. Hirota, Y. Kitade, and S. Senda, J. Heterocyl. Chem., 1980, 17, 413. l6 K. Hirota, Y. Kitade, and S. Senda, J. Org. Chem., 1981,46, 3949. D. H. R. Barton, G. Dressaire, B. J. Willis, A. G. M. Barrett, and M. Pfeffer,J.Chem. Soc.. Perkin Trans. 1. 1983, 665. '' C. Reichardt and K. Halbritter, Justus Liehigs Ann. Chem., 1975, 470. 444 Bamfield and Gordon OH 00 There is no problem of course if the ketone (R3CH2C(0)CH2R4)is symmetrical, e.g.R3 = R4,entries 1-10; however, if it is not then preferential reactivity must be sought at one carbon in each of the co-reactants. One method is to employ a ketone that can be enolized preferentially in one direction and to react this with a p-dicarbonyl compound in which one carbonyl group is significantly more reactive than the other. This strategy can be seen in entries 11, 28, 33, 34-36, and 42. Preferential enolization in the ketone is achieved in many cases by incorporation of an ester group at the a-position, and for the f3-diketone use can be made of the difference in reactivity between an aldehyde, an aliphatic ketone, and an aromatic ketone. Annelation of the aromatic ring to an existing ring also helps to reduce the problem of isomer formation.Equation 2 illustrates this for the preparation of a hydroxyphthalide, where the preferred formation and then reaction of an enamino- ketone is utilized for final ring closure. Interestingly the 5-ring is formed in situ from the open chain product.31 Doubly deprotonated p-ketoesters, e.g. CH2C(0)CHC02Rcan be silylated to give the disilylated enol diene, e.g. (6), which has recently been used in some acid- catalysed (regiocontrolled) syntheses of salicylates, as shown below. Once again advantage is taken of the differential reactivity at the carbon centres to achieve this I9 (a) H. Muhlemann, Pharm. Acta Helv., 1949, 24, 351; (b) ibid.. p. 356. ’O (a) V. Prelog, 0.Metzler, and 0.Jeger, Heiv.Chim. Acta, 1947, 30,675; (h) ibid.,1947, 30,1883. ’I L. Ruzicka, V. Prelog, and J. Battegdy, Helv. Chim.Acta, 1948, 31, 1296. ’’E. Y.Belyaev, M. S. Torbis, and A. V. El’tsov, Zh. Org. Khim., 1978. 14, 2375. 23 G. A. Kraus, J. Urg. Chem., 1981.46, 201. 24 N. Takeuchi, K. Ochi, M. Murase, and S. Tobinaga, J. Chem. SOC., Chem. Commun., 1980, 593. 25 (a)J. H. Clark and J. M. Miller, Tetrahedron Lett., 1977, 139; (h) ibid., J. Chem. Soc.. Perkin Trans. I, 1977, 2063. 26 S. H. Bertz. Swthesis. 1980. 708. ’’T. M. Harris, T. P. Murray, C. M. Harris, and M. Gumulka, J.Chem. Soc., Chem. Commun., 1974.362. L. Claisen, JusfusLiebigs Ann. Chem., 1897, 297,40. 29 N. K. Kochetkov, L. J. Kudryashov, and B. P. Gottich, Tetrahedron, 1961, 12, 63.30 (a)E. Belyaev, L. M. Gornostaev, A. P. Es’Kin, M. S. Torbis, G. A. Suboch, and A. V. El’tsov, Zh. Org. Khim.. 1977, 13. 2307; (6)ibid.,1978, 14, 2189. ’I S. Auricchio, A. Ricca, and 0.V. de Pava, Garz. Chim. Itat., 1980, 110, 567. 32 (a)T. H. Chan and P.Brownbridge, J.Am. Chem. Soc., 1980. 102,3534; (b) ibid., Tetrahedron, 1981.37, 3387; (c) ibid., J. Chem. SOC.,Chem. Commun., 1979, 578. Table 2 Condensation of’ ketones to give phenols o\ Y X R’ R2 R3 R4 Yield (2,)Ref: Comments -H Me CO2H C02H CO2H 76 19a -H Me CHzPh CO2H CO2H 59 19b -H Me H COzEt COzEt 49 20 -H 2-thienyl H COzEt C02Et 61 20 4 b -H 3-pyridyl H CO2Et CO2Et 76 20 5 %3 -‘A’ ring of steroid H COzEt C02Et 57 21 6 -N=O Me Me H H 73 22 7 2 8 rn -N=NAr Me H Me Me 77 I 1a,h -N=NAr Me H H H 89 1la.h 9 8 -N=NAr Ph H Me Me 50 1 lu,h 10 2 -N=NAr Me H Me H 61 1la,h 11 -H Me H CO2Me C02Me 50 23 12 NMe2 -(CH2)4-H CO2Me C02Me 52 24 13 NMe2 H Ph H C02Me C02Me 23 24 14 -H Me Me H COMe 64 25 15 -H PhCH2 Me CO2Me COMe 52 25 16 -H CsHopC1 H C02Et COzEt 53 12 17 -H C6HgOMe H COzEt CO2Et -12 18 -Ph Ph H COzEt C02Et 51 12 19 -Ph PhCH2 H CO2Et CO2Et 51 12 20 Me Me H COzEt C02Et 52 12 21 H Pr H COzEt COzEt 40 12 22 H Me Me COMe COMe 50 26 23 H Et Me COMe COMe 41 26 24 H Me CF3 COMe COMe 13 26 25 H Et Me COzMe COZMe 78 26 26 -(CHz),-H COzMe C02Me 50-83 20 27 -(CHz)d-Me CO2Me C02Me 36 20 60% ref 26 28 -(CHZ)4-C02R COzMe COzMe 60 26 29 H Me H COzH H 83 20 30 H C15H31 H C02H COzH 50 20 31 H Me Me C02H CO2H 92 20 32 H Ph H C02H COzH 59 20 33 H Ph Me C02H COzH 47 20 34 H Me H C02R H 40 27 35 C02Et Me H H C02H 63 28 36 H Me H H C02R 39 29 37 (CH2)4 H H CO2H -21 38 N=O Me Me H H 30 3,5-dimethyl-2-nitrosophenol23% 39 N=O Me Me Me H 30 3,4,5-trimethyl-2-nitrosophenol36%40 N=O Et Me H H 30 3,5,6-trimethyl-2-nitrosophenol67% 41 N=O Ar Me H H 21-100 30 42 N=O 2-thiophenyl Me H H 54 30 43 N-4 Ar Me Me H 44-64 30 44 B N=O Me Me COzMe C02Me 64 30 45 Q N=O Ar Me COZMe COZMe 19-25 30 46 5 See also -'Methoden der Organischen Chemie'. Vol.6/k pt. 2. ed. E. Muller, G. T.Verlag, Stuttgart. 1976. 5 Aromatic Benzene Compounds from Acyclic Precursors OH C02Me Me Ph Me Me The alkyl Grignard reagent CH,=C(R)CH,MgX also allows the preparation of unsymmetrical phenols as demonstrated by its reaction with ketone (7), followed by an easy elaboration to the aldehyde (8).Final ring-closure then gives the phenol (9) in high yield.33 These syntheses and many of those shown in Table 2 provide synthetically useful alternatives to conventional aromatic syntheses, especially in view of the unsymmetrical substitution patterns and high levels of substitution that can be attained. In a series of papers 340-e describing studies of xanthyrones and glaucyrones, ”M. A. Tius, A. Thurkauf, and J. W. Truesdeli, Tefrahedron Lett., 1982,23, 2823. 34 (a) L. Crombie, D. E. Games, and A. W. G. James, J. Chem. Soc., Perkin Trans. I, 1979,464;(b) L. Crombie.M. Eskins, D. E. Games, and C. Loader, J. Chem. Soc., Perkin Trans. I, 1979,478; (c) S. R. Baker and L. Crombie, J. Chem. Soc., Chem. Commun., 1980.2 13; (d)ibid., 1979,666; (e)ibid., I980,2 I I. 448 Barnfield and Gordon Crombie and his co-workers report some useful reactions leading to phenols containing acetyl, ester, alkyl, and alkene groups. Strong chelation effects are observed in many of these reactions and as an example the compound (10) cyclizes exclusively to the salicylate (11) with sodium ethoxide (see also Claisen3') or magnesium methoxide and not at all to the resorcinol(l2); however, in an excess of magnesium methoxide mixtures of both result. (10) (11) (12) Zinc chloride can also play an important role in determining the direction of cyclization.Thus, in its absence, ring-closure of the diacid (13) occurs to the salicylic acid (14) whereas with ZnCI, present the salicylic acid (15) is formed.36 &COP +C02H / Me (14)CO;!H (13) OHbC4"Me (15) The above example illustrates an alternative source of the phenolic oxygen, i.e. from a carboxylic acid, and contrasts with the syntheses already discussed. Continuing in a similar vein, cyclization onto an ester [e.g. (16)] gives rise to the cyanophenols (17)37 and (18)38 and by a similar strategy the highly substituted cyanophenols (19) 39 and (40)"O are obtained in fair yields. These latter syntheses 35 L. Claisen, Justus Liebigs Ann. Chem., 1897, 2971. "G. Agnes and G. P.Chiusoli, Chim. Ind.(Milan), 1967,49,465. "J. Sepiol and J. Mirek, Synthesis, 1979, 290. 38 0.S. Wolfbeis, G. Zacharias, and H.Junek, Monatsh. Chem., 1975, 106, 1207. "(a)T. Severin, B. Bruck, and P. Adhikary, Chem. Ber., 1966,99,3097;(b) T. Severin and B. Bruck, Angew. Chem., Int. Ed. Engl., 1964, 3, 806. 40 C. Ivanov and T. Tcholakova, Synthesis, 1982, 730. Aromatic Benzene Compoundsfrom Acyclic Precursors are particularly useful because of the ease of synthesis and availability of the precursors, and also because of the degree and nature of the substitution pattern which would not be easy to achieve by alternative processes. A similar strategy is apparent in the synthesis of cyanoanilines (see Section 4). eRdo"CN CN CN (1 6) (18) OH Nc#co2Me PhAr CH3 + MeZN Nc2YNo2 A r1 A r2 CN The last two examples illustrate the use of Michael acceptors and further examples are shown in Table 3.The variety and, in many cases, cheapness of these versatile reagents makes them very useful for aromatic-ring synthesis and greatly increases the utility of the method. In many cases an eliminatable group, e.g. SOPh, halogen, OR, must be present in one of the acyclic precursors to ensure aromatization, otherwise only a cyclohexenone is formed. The eliminatable group can also play a second role, in that it can control the regiochemistry of condensation. See entries 1 and 2 in Table 3 where the sulphoxide group fulfils the dual role of directing the reaction and then eliminates to form the phenol.If an acetylene is used, entry 6, then there is no need for an eliminatable group, although this modification is of limited use since acetylenes themselves, with a few exceptions, are neither cheap nor readily available. This route is therefore only likely to be of use for more inaccessible phenols. Barnfield and Gordon PhCHXHCOMe PhSOCHZCOCHzR & R=H62 R = Et 50 Ph Me R’ CH 2 4 Rz)COCHzR R’ = H, Me30.58 42 R2 = H. Me, Ar CHz=C( F)COMe 43 CH z=CHCOMe PhC(0H )C(O)Ph 44 OH OH EtOCH=CHCOz E t 45N C02Et OH EtOCHXHCOzEt MeCH(C0zEt )Z 47 45a “‘QCozEt C02Et OH 7 EtOCH=CHC02Et 48 45ir oHc@co;” CO2Et EtOCH=C( COz Etz) RCHzCOlEt R = Ph 24 C02Et R = 2-thienyl 12 4% R = 3-pyridyl 61 PhCXCOPh 46 4’ A.A. Jaxa-Chamiec, P. G. Sammes, and P. D. Kennewell, J. Chem. Sac.. Perkin Trans. 1. 1980. 170. 42 D. L. Boger and M. D. Mullican. J. Urg. Chem., 1980.45. 5002. 43 H. Molines and C. Wakselman, J. Chem. Soc., Chem. Commun., 1975, 232. 44 C. Egli, S. E. Helali, and E. Hardegger, Helv. Chim. Acta, 1975.58, 104. ” (u) W. J. Croxall and M. F. Fegley, J.Am. Chem. Soc., 1950,72,970;(h) C. W. Bird and C. K. Wong. Tetrahedron Lert.. 1975, 1877. 46 W. Deuschel. Helv. Chim. Acfa. 1951, 34. 168. 45 1 Aromatic Benzene Compounds from Acyclic Precursors Table 4 Phenols from six-membered ring heterocycles Entry Heterocycle Co-reuctant Product Yield % Re$ 1 50 47 0 R =H. Me@ozH 2 3 0 0 R = NOZ. Me p""NO? Q 30 - 47 48 CO2 R 4 qonCO,H 30 49 Me 0 Me 5 Ar,J&+ A1 MeNO2 Ar2 22-78 50 6 Ph-PhCHz( COz Et)z 27 51 COLE( 7 CH ,(CN)COzEt p h qPh 37 52 CN 8 46 53 9 HOMe*Me Me R = H, olkyl, y: 70-80 53 I+ Me 10 MeC(0)CH~C0zEt 65-72 54 ozN*No2 02N& , NoR=Me CO2E1 77-80 5511 MeNOz Roc*coR MeMe Me Me NO? Bamjield and Gordon Heterocyclic systems have already been shown to be a useful source of phenols by ring cleavage.Several oxygen heterocycles are particularly susceptible to hydrolysis and nucleophilic attack as shown in Table 4 and provide some useful routes to highly functionalized benzenoids, although the yields are at best only mediocre in the examples illustrated. Pyridinium salts are also susceptible to cleavage and, as Table 4 shows, they provide a route to phenols, although the scope is limited in the example shown.Nevertheless further extensions might be possible. The preparation of 2,3,4,5-tetramethylphenol is claimed to be a feasible alternative to conventional synthesis but, on the whole, syntheses based upon pyrylium salts are likely to be most useful for rather inaccessible phenols. (Entry 8.) It is less common to find five-membered heterocycles as a source of phenols, with what few examples there are being dominated by oxygen heterocycles (Table 5). Interestingly, all the examples shown are to m-carboxyphenols; this complements the syntheses of o/p carboxyphenols already described in earlier tables. Table 5 Phenols ,from jive-membered ring heterocycles Heteroc?*cle Product Yield % ReJ Entry 26 56 1 57 2 PhCOCHZ PhHovcozH MeCOCH, E2s COzM e 68 58 3 HO 86 59 4 "F.C. Cheng and S. F. Tan. J. Cheni. Soc. C, 1968. 543.''E. D. Bergman. D. Ginsburg. and R.Pappo, Org. React., 1959. 10, 179. 49 H. Guilford, A. I. Scott, D. C. Skingle. and M. Yalpani, J. Chem. Soc.. Chem. Commun.. 1968. 1127. '"K. Dimroth and H. Wache, Chem. Bey., 1966.99, 399. '' (u)K. Dimroth and G. Neubauer. Angew. Chem., 1957, 69, 720; (h) ihid., Chem. Ber., 1959.92. 2046. '' H.G. Rajoharison, H.Soltani, M. Arnaud, C. Roussel, and J. Metzger, Sjnth. Commun.. 1980. 10, 195. " R. Lukes and M. Pergal. Collect. Czech. Chem. Commun.. 1959, 24, 36. 54 E. Matsumura. M. Ariga, and Y. Tohda.Bull. Chem. Soc. Jpn.. 1980, 53. 2891.''F. Eiden. H. P. Leister. D. Mayer, Ar:neim-Forsch., 1983. 33. 101.'' N. Elming. Acta Chem. Scunrl., 1956. 10, 1664. 5' A. Bianco, M. L. Scarpati, and C. Trogolo. Ann. Chim. (Rome), 1972.62, 709. '"C. Iavarone, M. L. Scarpati, and C. Trogolo. Guz:. Chim. Irul., 1971, 101, 748. 3Y R. J. Gillespie. J. Murray-Rust. P. Murray-Rust. and A. E. A. Porter, Tefruhrdrlmn.1981. 37. 743. 453 Aromatic Benzene Compounds from Acyclic Precursors Some of the syntheses already highlighted have been used in routes to natural products. Similarly, sclerin and resistomycin have both been synthesized recently and depend upon a ring-synthesis method to build up the phenol ring. In the synthesis of sclerin a general synthesis of 3-hydroxyphthalides (21) was developed,60 and for resistomycin the biphenyl (22) was elegantly elaborated.61 In both cases the ring synthesis approach is superior to that of the conventional approach.R2 C (OMeI3 R' +e + or ( R2C0 0 + OTMS (21) OMeArCOzEt + 0 I0-=--Li A+Ar-C-=--COzR I COz R Me0 OMe "TCO R -Ro&EzR ROZC ROzC Me0 Ar COzR (22) 50% A further example of phenylacetic ester formation is illustrated in the preparation of the phenol (23, 40%) from the dialdehyde (24) and involves a rearrangement; the mechanism is proposed in the paper.62 In a quite intriguing reaction, pentane-1,3-dione condenses with maleic anhydride to give a fair yield (50%) of the internally protected phenol (2S),63 and oxalyldialdehyde condenses with nitromethane to provide the acetyl derivative of 2,4-dinitrophen01.~~ More simply the decalenone (26) undergoes a ready rearrangement to the phenol (27).65 6o T.H. Chan and P. Brownbridge, J. Chem. SOC.,Chem. Commun.. 1981, 20. 61 K. James and R. A. Raphael, Tetrahedron Lett., 1979, 3895. 62 D. S. Tarbell and B. W. Hargotz. J. Am. Chem. SOC..1954.76, 5761. ''E. Berner, J. Chem. SOC..1946, 1052. 64 F. W. Lichtenthaler, Angew. Chem., Int. Ed. Engl., 1964.73. 21 1. 65 M. Kalyanasunderam, K. Rajagopalan, and S. Swarminathan, Tetrahedron Left., 1980, 21.4391 454 Bamjeld and Gordon COZEt &OH MeP‘ ’ acHo ’C02Et CHO 0 0 (23) (24) 0 (26) OH ), 0 Me (27) It is less common to find ring syntheses of phenols containing no acceptor groups whatever; however, the versatility of the method is such that it is nearly always possible to find a ring synthesis approach to any general class of phenol.This can be illustrated by the phenols in Table 6 which contain only alkyl, aryl, or halogeno substituents. Some of these examples appear fairly general, yields are fair and access is provided to some particularly difficult phenols, e.g. polyarylated phenols. The method is only limited in some cases by the inaccessibility of the starting materials. Table 6 AIkyl-. aryl-. and halogeno-substituted phenols Precursors Phenol Yield % Re$ R=H35 ,CHi C CH Me 50 R-C Et 34 661 ‘cH,-CGCH Me Pr’ 39OH Bu’ 40 OH Me -CZC-H67 67 cH, =C( Me)CHMeCOCI M e Me OH MeCZC-Me 39 67+ CHZ=CH CH, COCl Me Aromatic Benzene Compounds .from Acyclic Precursors Precursors Phenol Yield "/, Ref.0 OH + PhCOCOPh R&R R = -YHPh 20 68 COPh F? R3 = aryl R1 RZ R'R' = alkyl aryl 55-89 69R3f4xX = C1, Br 0 OH OH 70MeCO CH =C(Me) CH CO M e Me &Me Diphenylcyclopropenone reacts with carbon ylides to give, for example, diphenyl phenols (28) 71 and (29).72 However, despite the satisfactory yields and unusual substitution pattern the high cost of diphenylcyclopropemone seriously limits the usefulness of this route. Similarly, the cycloheptenone (30) is a rather exotic precursor to m-hydroxybenzaldehyde (85%) and is of more mechanistic than synthetic interest.73 MPh R Ph R R=H,Me R = H, Me, Ar R3= H, Me R4= H, Me (311 6b D.Plouin and R. Glenat, Bull. Soc. Chim., 1975. 336. "M. Karpf, Terraheriron Lett., 1982. 47. 4923.''C. F. H. Allen and J. A. Van Allen. J. Or%.Chem., 1951, 16, 716. K. Steinbeck, T.Schenke. and J. Runsink. Chem. BN., 1981, 114, 1836. 'O A. Baeyer and J. Piccard. JUSIUSLiehigs Ann. Chcm.. 1915. 407. 332. " Y. Tamura. T. Miyamoto. H. Kiyokawa, and Y. Kita, J. Chcm. Soc., Perkin Trans. 1. 1974. 2053. l2 T. Sasaki, K. Kanematsu, A. Kahehi, and G. Ito. Tefrirheriron, 1972. 28. 4947. 73 G. Biggi. A. J. de Hoog, F. D. Cima. and F. Pietra, J. Am. Cheni. Soc., 1973, 95. 7108. 456 BamJeId and Gordon All the reactions thus far concern the synthesis of phenols containing alkyl, aryl, and electron-withdrawing groups.However, it is just as feasible to introduce donor groups by ring synthesis methods and aminophenols are an important case in point. For instance o-aminophenols (3 1) can be prepared from the corresponding 2-acylfuranone and a cyclic amine, albeit in rather variable yields.74 This method is of limited utility in view of the ready synthesis of o-aminophenols by conventional methods. In general the synthesis of m-aminophenols has received more attention in the literature and proves to be far more useful. Various m-aminophenols containing carboxylic acid, cyano, alkyl, acyl, and halogeno groups can be readily synthesized from simple and available acyclic precursors. The level of substitution can also be varied as can the substitution pattern.Furthermore, the products in many cases are not easily accessible by conventional aromatic synthesis. For example, the ethoxymethylene compounds (32) condense with active methylene compounds, e.g. P-ketoesters, malononitrile, cyanoacetic ester, to yield the corresponding m-aminophenols (33) in good yield^.^'-^^ Enamine (34) reacts with 1,2-dibromoacryloyl chloride to give the m- aminophenol (35),78 whereas enamines (36) dimerize to provide (37).79-80 However, dimerization can be prevented if a silyl group is included in the enamine such as in (38),*’ thus allowing cross-condensation between enamines. Interestingly, the direction of cyclization depends upon the nature of the co- reacting enamine so that enamines derived from acyclic ketones give aminophenols (39) by 3C + 3C addition, and enamines derived from cyclic ketones (n = 2-5) give aminophenols (40) by 4C + 2C addition.If n > 5 then 3C + 3C addition occurs once again. R’ HRhR CN H R = CNJC02Me R = H, C02R, alkyl R2= CN, C02R R’ = COMe,C02R 74 (a) L. Birkofer and G. Daum, Angew. Chem.. 1960.72, 707; (b)ibid., Chem. Ber., 1962.95, 183. ’’H. W. Schmidt and H. Junek. Justus Liebigs Ann. Chem.. 1979,2005. 76 S. R. Baker, L. Crombie, R. V. Dove, and D. A. Slack, J. Chem. Soc., Perkin Trans. I, 1979,677. ”D. Leaver and J. D. R. Vass, J. Chem. Soc., 1965, 1629. ”(a)P. W. Hickmott, B. J. Hopkins, and C. T. Yoxall, Tetrahedron Lett., 1970,29,2519; (b)N.F. Firrel. P. W. Hickmott, and B. J. Hopkins, J. Chem. Soc., 1970, 1477. 79 1. A. Zaitsev, M. M. Shestaeva, and V. A. Zagorevskii. Zh. Org. Khim.. 1966.36. 1769. H. Bohme, J. G. von Gratz. F. Martin, R. Matusch. and J. Nehne.Justus Liebigs Ann. Chem.. 1980,394. T. H. Chan and G. J. Kang, Tetrahedron Lett.. 1983, 24. 3051. 457 Aromatic Benzene Compounds. from Acyclic Precursors H HOMe* iR' R2 R= H,Me,Ph (351 (36) (37) Me OH SiMe, (39) In common with other phenols described earlier aminophenols can also be prepared from heterocycles as shown by the conversion of dioxazoles (41) into aminophenol (42).82 Potentially more useful is the facile acylpyrone-phenol rearrangement of (43) to (44) which is usually accomplished in good R N*R (13) 82 S.Auricchio, S. Morrocchi. and A. Ricca, Tetrahedron Left., 1974, 2193. 83 ((I) F. Eiden, E. G. Teupe, H. P. Leister, and D. Mayer, Patent, Ger. Offen 2 922 488 (Chant.Ahsfr.. 198 I, 94,191 938);(6) F. Eiden, E. G. Teupe,and H. P. Leister, Arch. Pharm., 1981,314.419;(~)ibid..p.347;(4 F. Eiden and H. P. Leister. Arch. Pharm., 1980,313,972;(e)F. Eiden and E. G. Teupe. Arch. Pharm., 1979,312, 591. 458 Bamjield and Gordon All the syntheses have so far involved condensations, additions, and cyclizations leading directly to the aromatic compound, i.e. there is sufficient potential unsaturation in the acyclic precursors to generate the aromatic ring directly. However, there is substantial literature precedent for aromatizing cyclohexenones, cyclohexanones etc., and various reagents and procedures are available.Since many cyclohexanone/enones are available this is a method well worth considering as a feasible alternative to conventional routes, see Table 7. Only a small cross- section of examples is included since this topic is outside the scope of this review. Table 7 Phenols by aromatization reactions Entry Precursors Reagent Product Yield Re$ HN R3 R4R:Qo HN R' R~ -CHI -85 oQo HO NR'R' Br2-NaOH 10--60 86 OH 25-95 87Me3SiCI-DDQ &--;; QH 88 CuBr-LiBr 80 89 HO 90a.bAr fie Ar Ar S Arz%cN Ph 12-62 91 COzEt Brz-AcOH &co2Et 82 92b.. Me Aromatic Benzene Compounds from Acyclic Precursors Entry Precursors Reagent Product Yield ReJ$.-10 Br2-CS2 45 93 Me Me Me MQ C02Et e COZEi 0 ?H 11 PhSSPh-NaOMe 42-79 94phsaR’R7 12 40 95 H Br-AcOH PhPhGarPhO-b NO2 h NH2P OH 13 14 HC02Et 10-28 97 B.Dihydroxybenzenes-Perhaps not surprisingly the strategy for the preparation of dihydroxybenzenes (catechols, resorcinols, and quinols) bears a strong resemblance to that just described for phenols. As before, the aromatic hydroxy- groups usually come from carbonyl groups and once again the Aldol and Michael reactions are important. However, the Claisen condensation is used far more in the preparation of dihydroxybenzenes than for phenols and is found in many syntheses either at the ring-closure stage or in the assembly of the precursors prior to final ring-closure.Furthermore, many of the ketones, esters, ketoesters, and their unsaturated derivatives used in the synthesis of phenols can also be used for the preparation of dihydroxybenzenes. 84 H. Ida. Y. Yuasa. and C. Kibdyashi. Synrhesis, 1982.471. ” W. H. Mueller, Patent, U.S. 4 212 823 (Chem. Ahstr., 1980,93, 239 028). S. R. Ramadas, D. Rau, and W. Sucrow. Chem. Ber.. 1980, 113. 2579.‘’ M. T. Reetz and W. Stephan, Jusrus Liehigs Ann. Chem.. 1980, 533. 8R Y. Pickholtz. Y. Sasson. and J. Blum. Tetruhedron Lett.. 1974. 1263. D. Boudon. Y. Pietrasanta. and B. Pucci. Tetruheifron Lett. 1977. 821. (u) M. A. Elhashash. A. A. Afify. and A. Nagy. Indiun J. Chem.. Sect. B. 1979, 17, 581: (h) M. Abdalla. M. A. 1. Salem, and A.Hataba, Rev. Roum. Chim., 1980, 25. 1335. 91 C. Ivanov and T. Tcholakova, Synthesis, I98 I, 392. ” F. M. Hauser and S. A. Pogany. Sjwthesis. 1980. 814. y3 K. K. Bhattacharya, P. Pal. K. Ghosh, and P. K. Sen, Indian J. Chcm.. SNI. B, 1980. 19. 191. 94 B. M. Trost and J. H. Rigby. Tetruhedron Lctt., 1978. 1667. 95 D. W. Theobald. Tetruheifrlron. 1983. 39. 1605. 96 G. Pattenden and D. Whybrow. J. Chem. Soc.. Perkin Trims. 1. 1981, 3147. ’’ S. Labidalle. E. Jean. H. Moskowitz, H. Miocque, and C.Thal, Tetrirhrtlron Lett.. 1981. 22, 2869. 460 Barnfield and Gordon In this category, syntheses of resorcinols are the most commonly found in the literature and many prove to be useful alternatives to conventional methods. Several reports have appeared pertaining to the synthesis of natural products containing the resorcinol ring and activity in this area is likely to continue in view of the physiological activity of resorcinol derivatives, e.g.benzochromans. Simple dimerization of P-ketoesters can lead to interesting resorcinols. For instance diethyl acetonedicarboxylate dimerizes to the highly substituted resorcinol(45)98a*b and the enol ether of ethyl acetoacetate (46) undergoes TiC1,- catalysed dimerization to the monoethyl ether of a re~orcinol.~~ Interestingly, the nature of the enol ether is critical in deciding the final product since (46, R = Et) yields the 2-substituted resorcinol (47) whereas (46, R = Ph) yields the 2-unsubstituted resorcinol (48). In both cases, the formation of the monoether is specific with no diether formation and no isomer mixtures.The disubstituted resorcinol (49), has been the subject of successful synthetic studies as shown in Scheme 1. Several routes provide good alternatives to conventional syntheses and show elegant use of dianions, and their silyl derivatives, in regiospecific condensations. loo-lo4 2-Substituted resorcinols are also easily made and this can be illustrated by the synthesis of the resorcinol (50) from the silyl derivative (51) and the ketalester (52).'05 This reaction contrasts with one shown in Scheme 1 [to give the 2-unsubstituted (49) lo5] and demonstrates the control that can be exercised in these condensations by careful manipulation of reactive groups.Continuing upon this theme, ethyl acetoacetate condenses with diethyl acetonedicarboxylate to give resorcinol (53), which again shows that cross-condensations as well as 98 (a)H. Cornelius and H. von Pechman, Chem. Ber., 1886,19,1446; (6) D. S. Jerdan, J. Chem. SOC.,1899, 75, 808. 99 G. Declerq, G. Moutardier. and P. Mastagli, C.R. Hebd. Seances Acad. Sci., Ser. C, 1975,281,279. loo (a)T.A. Haw, E. Suokas. and K. McCoy, Acta Chem. Scand., Ser. B, 1978,32,701;(6) T. Kato, M. Sato, and T. Hozumi, J. Chem. Soc., Perkin Trans. I. 1979,529; (c)T. Kato and T. Hozumi, Chem. Pharm. Bull.. 1972, 20. 1574. lo' (a)S. N. Huckin and L. Weiler, Can.J. Chem., 1974,52,1343; (6)P. E. Sum and L. Weiler,J. Chem. Soc.. Chem. Commun., 1977,9 1. lo2 J.E. Hill and T. M. Harris, Synth. Cornmun., 1982, 621. Io3 A. G. M. Barrett. T. M. Morris, and D. H. R. Barton, J. Chem. SOC.,Perkin Trans. I, 1980, 2272. lU4 T. M. Harris and C. M. Harris, Tetrahedron. 1977,33,2159. lo5 T. H. Chan and T. Chaly, Tetrahedron Lett., 1982, 23. 2935. 461 Aromatic Benzene Compounds from Acyclic Precursors R)(,C02R \1 TlOEt or NaH / 2. Diketene Ref100a.b\H30+ \ J OTMS OMe COz R’ R =Me,Ph 00 Ref103 Ref 102 UnTMS &-Me02Cyco2Me &COCl HO R C5Hll (49) 00 R ,(MeCO),O Scheme 1 dimerizations are possible, O6 and dimethylmalonate condenses with the z,P-unsaturated ketone (54) to yield the 2-methylresorcinol (55). lo’ (51) HO C5H11 Me (50) Me0 OMe ,X,,CO2Me (53) C5H1 1 (52) Me Me02Cqx::Me OH lo‘ G.Koller and E. Krakauer, Monarsh. Chem., 1929,5%54,931. lo’ K. K. Light, Patent, US.3 928 419; Chem. Ahsrr., 1976,84, 105 224. BamJield and Gordon The ketone analogues of resorcinolester (49) are also available from acyclic precursors, as can be seen in the preparation of resorcinols (56)82*108from open- chain keto-compounds. OH OH HO R = alkyl, aryl, OH heteroc yclyl (56) (57) Various substituted resorcinols can be prepared from alkoxymethylene compounds (57). For example, (57, R' = CN)yields the cyanoresorcinol(58, R' = CN, R = C02R) when reacted with acetonedicarboxylic esters," whereas the ketoresorcinol(58, R' = COR, R = H)is formed from (57, R' = COR)and ethyl O9acetoa~etate.~~'" Similarly, diethyl acetonedicarboxylate reacts with 2-dimethylaminonitroethylene to provide the nitroresorcinol (58, R' = NO2, R2 = H).39 Because of the simplicity and availability of these precursors and the highly substituted compounds that can be obtained, these routes constitute very attractive and convenient methods for synthesizing resorcinols.Alkylresorcinols can be prepared by hydrolysing and decarboxylating the corresponding resorcinol esters. However, a particularly convenient direct route involves the reaction between methyl phenylsulphinylacetate and various substituted x,P-unsaturated ketones; the orcinols (59) are obtained generally in moderate yield.41 The need for an eliminatable group when using the Michael reaction in ethylene compounds is obviated when acetylenes are used since the precursors already possess the correct level of unsaturation for the final aromatization.Thus, conjugated acetylene compounds (R-CK-COR') react with 1,l'-diphenylacetone to give resorcinols (60),'lo with diethyl acetonedicarboxylate to yield (61)," ' and with diethyl malonate to provide resorcinols (62).''' The yields are moderate but in OH OH Me I Ph OH (60) (62) lo8 H. Zak and U. Schmidt, Chem. Ber., 1973, 106, 3652. 109 L. Crombie, D. E. Games, and A. W. G. James. J. Chem. Soc., Perkin Trans. I, 1979,472. 110 I. El-Kholy, M. M. Mishrikey, F. K. Rafla, G. Soliman, J. Chem. SOC.,1962, 5153. "I Patent, D.O.S. 1973, 2 359 410. 'Iz (a)R. M.Anker and A. H. Co0k.J. Chem. Soc., 1945,31 l;(h)J.C. Bardhan.J. Chem.Sot,.. 1929,2223; (c)E. P. Kohler. J. Am. Chem. SOC.,1922,44, 379; (d)L. Bickel, J. Am. Chem. SOC..1950. 72. 1022. 463 Aromatic Benzene Compounhfrom Acyclic Precursors view of the limited availability, and in many cases the inaccessibility of useful acetylene compounds, this method is only likely to be of use for less easily obtainable resorcinols. As with phenols, certain oxygen heterocycles can be easily converted into resorcinols after ring cleavage. Scheme 2 demonstrates this for lactones. Another OH Ref L9/ HO HOTOH X Scheme 2 lactone, dehydroacetic acid, is converted into 5-methylresorcinol simply by heating in water; this reaction was discovered some 90 years ago.' Isoxazolines have already been shown to be useful in the preparation of aminophenols and phenols and the same author has utilized them for the synthesis of dihydroxy-phthalides (63) in three simple steps from (64)in an overall yield of 20X.l l6 The phthalide (63) can then be converted into mycophenolic acid.The cyclopentenedione (65) provides another example of a regiocontrolled reaction with either the resorcinol(66) or the catechol(67) formed, according to the 0 0 HO+o Me Me-oxM; "o (64) ;go0 (63) (65) R' ph$;; ph)$ \Ph Ph OH OH OH (66) (67) 'I3 R. Bentley and P. W. Zwitkowits, J. Am. Chem. Soc., 1967,89,676. Patent, Fr. Demande, 410 795; Chem. Absrr., 1966.64, 3423. '" N. Collie and W. S. Myers, J. Chem. SOC.,1893.63, 122. 'I6 S.Auricchio, A. Ricca. and 0.V. de PWd, J. Org. Chem., 1983,48,602 Barnfield and Gordon reagents and conditions used.' '' Thus, (66) is preferentially formed when diazo- compounds (R'CHN2,R' = COPh, H, or Br) are used with zinc chloride catalysis, whereas (67) is favoured when N2CH2C02Ris the co-reactant and no catalyst is present. In general, far fewer references are available to the preparation of catechols; however, a useful general method for preparing substituted catechol monoethers (68) has been reported by Tius,' '* as shown below, and proceeds in moderate to good yields. Mono-ethers of catechol have also been prepared by treating 2,2,6-trichlorohexanone with alcohols and base.' l9 Both these methods are useful because of the easy and selective formation of the mono-ethers which can be troublesome to prepare by direct alkylation of catechols.R R = HI alkyl, acetyl 3-Substituted catechols (69) are made upon treatment of the saturated furan (70);120 the yields are rather poor and the method would therefore seem to be of limited use. Hydroquinone can be made in very high yield by the action of acetic anhydride and mineral acid upon the cyclohexa- 1,4-dione.' ' However, no indication of whether this reaction is generally applicable is given, although catechol and resorcinol can also be synthesized in very high yield from the corresponding diketones. 2,5-Dimethylquinol is obtained by dimerization of butane-2,3-dione in 300//,yield, but once again no indication is given of the generality of the reaction.' 22 Just as for phenols, dihydroxybenzenes can be obtained by aromatization.The B. Eistert and E. A. Hackrnann, Jusius Liehigs Ann. Chrm., 1962. 657. 120. 'IHM. A. Tius and A. Thurkauf, J. Org. Chrm.. 1983, 48.3839. 'Iy Van Winckel. Patent, U.S. 4 267 388. I2O (N) W. R.Boehrne,J. Am. Chrni.Soc., 1960.82.498;(h)J. T. Nielson. N. Elrning. and N. Clauson-Kaas. Actu Chem. Scund.. 1955. 9. 9."' M. S. Kablaoui, J. Org. Chmi., 1974, 39. 3696. 12' Patent. Ger.. I 220 437; Clwnt. Ahstr.. 65. 10531. 465 Aromatic Benzene Compounak from Acyclic Precursors aromatization to quinone just described is such an example,'" and another is the synthesis of doubly labelled resorcylic esters (71) which are obtained by condensation of MeCH:CRCOCH,R' with CH2( ''C0,Et)2 followed by dehydrogenation.*23 Many further examples are to be found in the literature.Me OH C. Polyhydroxybenzen-In this category the 1,3,5-trihydroxybenzeneshave received most attention, probably because of the commercial importance of phloroglucinol and some of its derivatives. Table 8 lists some of the routes published and several of these would appear attractive when compared to conventional routes via T.N.T.; the preparation of the bicycle (entry 9) particularly would not be trivial by conventional routes. Table 8 Synthesis of trihydrosybenzenes Entrj. Precursors Product Yield % Ref OH ' Mg (OMeI2 40 124 MeO2C&CO2Me -125 HO OH YE1 3 -126 Me 0 no OH ?H 4 -127 cIoc\/cocI HO OH '23 A.J. Bartlett, J. S. E. Holker. E. OBrien, and T. J. Simpson, J. Chem. Soc., Perkin Trlms. 1. 1983,667. 124 T. M. Harris, M. P. Wachter. G. A. Wiseman, J. Chem. SOC.,Chem. Commun.. 1969, 177. '25 Patent, Neth. Appl.. 76 09529; Chem. Abstr., 89. 23951. G. Waudan, Patent, Ger. Offen, 270 5874. T. Komninos. Bull. SOC.Chim. Fr.. 1918.23.449. Barnfield and Gordon Entry Precursors Product Yield % Ref. OH 5 C Hz(CO2 Et )z -128 HOJ3bH EtCHZCOCI -1296 HOEt@Et Et OH Me0,C &0Me 7 Mg (OMe)t -130 Me HO OH Et 02C&C02.' 8 25 131 HO OH 9 18-70 132 2-Ethylfurancarboxylic ester provides a route oia (72) to 3-acetyl- 1,2,4- trihydroxybenzene (40"/,),involving several steps; '33 however, 1,2,4-trihydroxy- benzene (73) is obtained directly upon reaction of 1,l'-diphenylacetone with 3- phenylcyclobutendione under mild conditions in fairly good yield.' 34 1,2,3-Trihydroxybenzenesfigure in several natural products, e.g.gallic acid, pyrogallol. Gallic acid (74, R = C0,H) and pyrogallol (74, R = H) can both be synthesized from ketal (75) and diester (76, R = C02H) or (76, R = H), respectively.' 35 The yield is fairly good (65%) in both cases. Pyrogallol derivatives can also be prepared by hydrolysis of 2,2;6,6'-tetrahalogenocyclohexanones although the yield is rather poor (20%).'36 Nevertheless, this does provide a convenient synthesis of some potentially highly substituted systems. Finally the tetra-acetylbenzene (77) is prepared from the cyclohexenones (78).13' 128 A.Baeyer, Chem. Ber.. 1885, 18, 3454. A. Combes, Bull. Soc. Chim., 1894, 11, 710. 130 T. M. Harris and C. M. Harris, Tetrahedron. 1977.33, 2159. 13' H. Leuchs and A. Geserick, Chem. Ber., 1908.41. 4171. 132 F. Effenberger, K.-H. Schonwalder, and J. J. Stezowski, Angew. Chem., Inr. Ed. Engl., 1982, 21. 871. 133 W. R. Boehme. Patent. U.S. 2 907 794; Chem. Absfr., 1961, 55, 463. 13' W. Ried and W. Kunkel, Jusfus Liebigs Ann. Chem., 1968, 717, 54. 13' M. T. Shipchandler. C. A. Peters, and C. D. Hurd, J. Chem. Soc., ferkin Trans. I, 1975, 1400. 136 Patent, G.B. I 574 713. 13' T.Posternak and J. Deshusses, Helv. Chim. Am. 1961.44, 2088. Aromatic Benzene Compounds from Acyclic Precursors OH R Ph OH OH R Me02CAC02Me A Me02C C02Me (75) OAc OAc OAc (77) 4 Anilines The synthesis of some anilines has already been discussed in the previous section under the guise of aminophenols.However, the application of ring synthesis methods to anilines is not restricted solely to aminophenols since a wide variety of substituted anilines can be obtained easily from acyclic precursors. In comparison, the amino-group is usually introduced by nitration of the aromatic ring, followed by reduction in the ring functionalization approach, and because this method is such an efficient way of producing anilines, ring synthesis methods can be considered as viable alternatives in only a few cases. One such case is to be found in the preparation of 2-cyanoanilines which probably best illustrates the usefulness of the ring synthesis method. Starting from readily obtained acyclic compounds various highly functionalized cyanoanilines are accessible in quite short synthetic sequences in contrast to the lengthy and tedious routes sometimes found in conventional methods. Scheme 3 illustrates the strategy uiu the cyclization of ylidene malononitriles, already the subject of a major review.13* Type I cyclization reactions are carried out under strongly acidic conditions and several examples are shown in Table 9.E. Campaigne and S. W. Schnellar, Synthesis, 1976, 705. Barnfield and Gordon Table 9 2-Aminohenzonitriles Starting marerials Product Yield Re6 55 139 / 31 140cQNqCN + %N"'e;" N' not 141 I givenMe NH 100 142 54 143 NH, 13') J.Sepiol. J. Mirek. and R. L. Soulen. Pol. J. Ch~m..1978, 52, 1389. I4O J. Sepiol. B. Kawalek, and J. Mirek, Srnrhrsis, 1977, 701. 14' R. Gompper. W. Elser. and H. J. Muller, Angrw. ChtJm.,Inr. Ed Engl.. 1967, 6, 453. I42 H. Jager, Client. Bey., 1962, 95. 242. H. Junek. 0.Wolfbeis, and G. Zacharias. Monritsh. ('hm., 1975. 106. 1207. 14' Aromatic Benzene Compound from Acyclic Precursors Type I CN“Q: -10acid H2 R ’R R R (or bond isomer) CN“TCNRll :eNH2- + base R Z R Z = CN or R 0 =Donor R W = CN, CO,R Scheme 3 In contrast, Type I1 cyclizations are carried out under basic conditions and where nitroalkenes (Z = NO,) are used 2-cyano-6-nitroanilines result (see Table 10).If the nitroalkene carries a p-amino- or a P-thio-substituent then the group at the P-position is eliminated during the cyclization reaction, whereas an aryl group is retained, as might be expected.Type I1 and 111cyclizations offer very effective routes into 2,6-dicyanoanilines as Table 10 6-Nitro-2-aminobenzonitriles Starting materials Product Yield ”/, ReJ /” NH + Br BamJield and Gordon Slarring malerials Product Yield % Ref 45 144 + 36 145dNo2aN N&cN / / 76-90 146 + MeSkNo2R R1 SMe MeSNTN can be seen from Table 11, and as Table 11 shows, the ylidenemalonitriles required for Type 111cyclization can also be generated in situ from 1,3-diketones and their N and S analogues.These routes are particularly important in view of the laborious, multi-step syntheses which are necessary with conventional methods. Table 11 2-Amino-I ,3-phthalodinitriles Starting materials Product Yield % Ref NCTCN + NCIX N C 3 G N d H R R R1 -90 145 R2 R2 R1 = Ph. Et R3 = Ph, Me R2= H, Me X = CN, C0,Me R~,R~=-(CH~)&- 144 T. Severin, B. Bruck, and P. Adhikary. Chem. Ber., 1966.99. 3097. '45 K. Gewald and W. Schill, J. Prakr. Chem., 1971, 313, 678. K. Peseke and J. Q. Suarez, Z. Chem., 1981, 21,405. 47 1 Aromatic Benzene Compounds from Acyclic Precursors Starting materials Product Yield % Ref Nc&N 75-77 147NcyN ’~r + NcIcNAr Me Ar Ar Ar = 4-R-C6H4-; R = H, CI, F CN -148A( CH-Jn + RCHO + CH~(CN)Z R n = 1-3 R = But,indolyl, or R’ C~H&-NCns::’ 72-77 149 RR Nc&N SMe R = Ph, 4-MeOC6HL-t CH2,CN NC@CN 83-92 150 MeS ‘C02Me SH 1 2 R2 R ,R =Me, Et, -KH& CN CN NC*CN 55 151 NC Me CN Nc*cNNC CN Me 14’ Y.Abramenko, Y. A. Baskakov, Y. A. Sharanin, N. A. Kiseleva, U. N. Vlasov, Y. G. Putsykin. and V. V. Negrebetskii. Zh. Vses. Khim. Ovu. 1979, 24. 409 (Chew. Ahsrr.. 1979. 91. 193 126). 14’ Y. Abramenko. Y. A. Baskakov, Y. A. Sharanin. A. F. Vasilev. Y. G. Putsykin.and E. B. Nazarova. J. Org. Chem. USSR, 1980, 16. 1870. 14’ K. Peseke. J. Prukr. Ckem.. 1983. 323, 499. 15() K. Gewald and H. Schafer. Z. Chem.. 1981. 21, 183. Is’ H.C. Gardner and J. K. Kochi. J. Am. Chem. Soc.. 1976, 98, 558. R.Hull. J. Chem. So<,..1951. 1136. Barnfield and Gordon Starting materials Product Yield % Ref. 0 0 Me-Me -ditto- 50-55 153 85 154 0 0 31-75 155 Nc&NR CF3 70 156 PhfiPh PhNc&cN Ph 42 157 AcO &” AcO dCN Patent, U.S.S.R. 521, 260, 1974 (Chem. Abstr., 85. 177041). I54 Y. A. Sharanin. L. A. Rodionovskaya. V. K. Prornonenkov, and A. M. Shestopalov, Zh. Org. Khim., 1983. 19. 1781. E. Gudriniece, A. V. Guttsait. S. V. Belyakov. and A. N. Fornin, Lutv. PSR Zinar. Akad. Vestis, Kim. Ser., 1983, 245 (Chem.Absrr., 1983.99, 53298d). Y. A. Sharanin and K. Y. Lopatinskaya, J. Org. Chem. U.S.S.R.,1976, 12, 688.”’B. Green, 1.S. Khaidern, R. I. Crane, and S. S. Newdz, Tetrahedron, 1976, 32, 2997. 473 Aromatic Benzene Compounds. from Acyclic Precursors 6-Acyl-2-cyanoaniline (79) can also be conveniently synthesized by a Type I11 cyclization and once again the intermediate ylidenemalononitrile is generated in situ, this time from a pyrylium salt and malononitrile.'58*159 An excellent though limited method for making a 2,4-dicyanoaniline (80) has been described,' 6o and involves the condensation of P-methyleneglutaric acid dinitrile or P-methylglutaconic acid dinitrile in high yield. ,cdcN+ NaOMe Me F ~ MeON'Me CN CN (80) In all the examples so far described the amino-group is formed by condensation at a nitrile group which inevitably leads to the formation of a primary amine.However, the amino-group in the aniline can originate from an enamino-group; this then allows for the direct synthesis of secondary and tertiary anilines. Self condensation of P-enaminoketones (8 1) provides a simple though rather unattractive route to 2-aminophenones as described in Scheme 4. If trimethylsilylchloride is used as co-reactant then the yield of phenone (82) is good although pyridines are formed as by-products.'6' If the reaction is catalysed by acid 3,5-xylidenes (83) are also formed.*62 Nevertheless the POCl, induced dimerization of enaminoesters does provide good yields of the aniline esters (84). There have been several reports on the use of enamines and enediamines to make N,N-dialkylanilines, but they have been generally of limited synthetic utility.'64 However, a recent paper has described a route which appears to offer some potential as a general method for the synthesis 2,6-disubstituted anilines.Thus 1,3-dichloro- 1,3-dimethoxypropane condenses with enamines (85) to give the 2,6-disubstituted anilines (86) in moderate yields and would presumably be of wider applicati~n.'~~ 158 K. Dirnroth and G. Neubauer. Angew. Chem.. 1957.69, 720. K. Dirnroth and G. Neubauer, Chem. Ber.. 1959.92, 2046. ''O U.K. Patent, I 374 954, 1972. C. Kashirna and Y. Yarnamoto, J. Heteroc:vcf. Chem., 1980, 17, 1141. 16' S. Auricchio, R. Bernardi, A. Ricca, Tetrahedron Lett., 1976, 4831. lb3 R. L. N. Harris, J. L. Huppdtz, and J. N.Phillips, Angew.. Chem., In!. €d. Engl.. 1976, 15. 498. 164 P. N. Hickmott and C. T. Yoxall. J. Chem. SOC.C, 1971, 1829. F. Camps, C. Jaime, and J. Molas, Tetrahedron Lett., 1981, 22, 2487. 474 Barnfield and Gordon R’, ,Rz Me (82) 12R .R (85) + NO MeUR (811 I + (84) =Me, Et, -(CHZIn-n = 4 or 5 Yields 78-89% Scheme 4 EtNlPr lz MeCN X R Yield % Me 38 EN- Ph 52 n WN-0 Ph 54 475 Aromatic Benzene Compounds .from Acyclic Precursors The dichlorodimethoxypropane just described is a synthetic equivalent of malondialdehyde and it is therefore not surprising to see that nitromalondialdehyde reacts with enamines, generated in situ,to give substituted 4-nitroanilines (87)."' Similarly enamines react with oximino-diketones to give the corresponding 4- nitroso-anilines (88) in poor to excellent yield.16' These latter two reactions are simple extensions of the phenol syntheses, described in Section 3, e.g. see Hill and Prelog syntheses, and so many of the comments recorded there apply equally to these reactions. MeC0CH2R R1, ,R2 R1R2NH N ___) OHCYCHO NO2 NO2 R', ,R5 R3& \ R3 + 4 R3@ R3 2 -8O"/o R' R2 R1vR2 NON 'OH (88) The double enamines (89) have been the focus of some fairly recent attention and provide access to either l12,4-triaminobenzenes (90) or l12-diaminobiphenyls (91) depending upon conditions and co-reactants. Although the yields are not high the substitution patterns are not common.' 68*169 In the preparation of (90) and (91) highly reactive imminium salts were used and this approach is further exemplified in the preparation of 1,3,5-tri-(N,N-diethy1amino)benzene in quantitative yield from the vinamidinium salt (91).''O Unfortunately, the reactions would appear to be of rather limited scope. The similarity between several of the routes to anilines and those to phenols (Section I) has already been highlighted.Continuing on this theme, heterocyclic compounds such as pyrylium and pyridinium salts have both been shown to provide routes into phenols and by use of the appropriate co-reactants will also R. A. Sagituh, S. P.Gromov. and A. N. Kost, J. Org. Cheni. USSR, 1978. 14. 1554. 16' E. Yu. Belyaev. G. A. Suboch.and A. V. El'tsov. J. Org. Chem. USSR, 1978, 14. 1407 l6U S. Baroni. E. Rivera. R. Stradi. and M. L. Saccarello, Trrruhedron Lerr.. 1980. 21. 889. 169 G. Crispi, P. Giacconi, E. Rossi. and R. Stradi, Sjnthesis. 1982, 787. R.Gompper and U. Heinemann, AngeM,. Chum.. In[. Ed. EngI., 1981. 20. 297. Barnfield and Gordon (90) (911 X Y Yield % X' X2 X3 Y Yield% (91) 0 0 50 HHHO25 CH, 0 55 H H H CH, 25 NMe 0 70 OMe H H 0 47 0 CH, 58 H NO, H 0 15 CH2 CH, 46 FHHO25 NMe CH, 62 OMe OMe H 0 25 Scheme 5 Me Me, N+ OEt Me2Nyy-2 OEt'+ Me2N NMez -M e,N NMe, (92) yield anilines. For example the pyrylium salt (93) gives rn-aminomethyl anilines (91) and alkylpyridiniurn salts rearrange to anilines when treated with amine U.S.S.R.Patent 659 562, 1979 (Chem. Abstr.. 1979,91, 93122). Aromatic Benzene Compoundr from Acyclic Precursors sulphates at elevated temperature.‘ 72-4 A potentially useful reaction of this type is the preparation of 2-aminodiphenylamines (95) from the corresponding amine sulphate and alkylpyridinium salt (96).’73 qH,N H R (93) (94) (95) (96) Aromatization reactions have been largely ignored in this review despite extensive literature precedent. However, the Semmler-Wolf aromatization of cyclohexenone oximes is worthy of mention. The Semmler-Wolf reaction was first published last century and was effected by heating cyclohexenone oximes with strong acids, e.g. equation 3.’” Since then a variety of reagents have been found, Me Table 12 Method A = pTosOH/CH2=C=O/MeCN, 7&80 “C B = p-TosOH/CH2=C(OEt)OAc/MeCN,80 “C R’ R2 R3 Yield Yo Method H Ph H 85 H H H 50 H Me H 61 H Me Me 61 H Me Me 62 H H OMe c1 Me Ph 76 67 -CH2-CH(OMe)-O- H 40 H SCN H 50 T. V.Stupnikova, A. I. Sedyink, V. N. Kalafat. R. S. Segitullin, and V. P. Marshtup, Khim. Gererorsikl Soedin., I98 1, 508. ’73 A. N. Kost. L. G. Yudin, A. N. Rumyanstev. and R. S. Sagitullin, Khim. Geterorsiki. Soedin., 1982,270. 174 A. N. Kost, L. G. Yudin, A. N. Rumyanstev, and R. S. Sagitullin, Khim. Geterorsiki. Soedin., 1983.63. 175 W. Semmler, Berichre. 1892, 25, 3352. BamJeld and Gordon such as acetic anhydride, polyphosphoric acid, and mixed acid, which will effect the reaction and various anilines are now accessible by this r0~te.l'~ More recently a mild and synthetically useful reaction has been published (see Table 12) and involves the use of toluenesulphonic acid with ketene, or with a synthetic equivalent of ketene, at 70-80 "Cwith acetonitrile as solvent.The yields are fairly good and a number of differently substituted anilines are readily accessible.' 77 Dehydrogenations have also been used extensively to generate anilines, usually from imines, and Table 13 illustrates some of these reactions. Particularly noteworthy is the preparation of 2,6-dimethylaniline, a commercially useful compound used in the pharmaceutical industry.' 79 Table 13 Conditions Ref: Pd/A1203 178-R'6R2R'&R2 210 "C NH2 ~ Pd/C*NH3 179 290 "C 5 Benzenes containing at least one Acceptor Group The reaction of pyrylium salts with an active methylene compound is a general route into benzenes having only one electron-withdrawing group and bearing alkyl or aryl groups in the 2,4, and 6 positions.' *'The electron-withdrawing group can be chosen from nitro, cyano, ester, and ketone groups and, as can be seen from 176 Y.Tamura, Y. Yoshimoto, K. Sakai, and Y. Kita, Synthesis, 1980,483, and references therein. 177 Y. Tamura, Y.Yoshimoto. K. Sakai, J. Haruta. and Y. Kita. S.ynthesis, 1980, 887. 17' Eur. Pat. Appl. EP 51805, 1982 (Chem. Abstr., 1982,!V, 162556). Brit. Pat. I 565 849, 1975 (Chem. Absrr., 1981. 94,30328). P. A. Grieco and N. Marinovic, Tetrahedron Lett., 1978, 2545."I K. Dimroth and K. H. Wolf, 'Newer Methods of Organic Chemistry', Vol. 3, ed. W. Foerst, Academic Press. New York, 1964, p. 357. 479 Aromatic Benzene Compounds from Acyclic Precursors Table 14, the reaction is quite versatile. Those entries shown in Table 14 are il-lustrative examples and many more are to be found in the general literature.' TaMe 14 YCHZX =.. R3 R1 X R' R2 R3 X Y Yield % ReJ Me Me NO2 H 72 182a Ph Ph NO2 H 48 182a Ph Ph NO2 H 85 182a Ph Ph NO2 H 64 182a C6H40Me(4) C6HsOMe(4) NO2 H 60 182a SMe Me NO2 H 42 183 Ph Ph COzEt COMe 32 1826 Ph Ph COMe COMe 70 1826 Ph Ph CN CO2Et 81 1826 OMe OMe C02Et WOMe)2 65 184 Early work by Hill and his co-workers showed that the sodium salt of nitromalondialdehyde could be trimerized to 1,3,5-trinitrobenzene but this is of no practical significance.More recently, two routes to 1,3-dinitrobenzenes have been devised which offer greater practical utility; both used nitroacetylene synthons. The first of these involves a stepwise reaction of p-N,N-dimethylaminonitroethylene with an aldehyde (97) to give low yields of the 1,3- dinitrobenzenes (98) carrying alkyl (or aryl) substituents at the 5-position. 86 R (98) R Yield % H 20 Me 24 Ph 40 IaZ (a)K. Dimroth. G. Brauniger, and G. Neubauer, Chem. Ber.. 1957,90, 1634; (6)K. Dimroth and G. Brauniger, Chem. Ber., 1959.92.2042. M. Ohta and H. Kato. Bull. Chem. Soc.Jpn., 1959,32, 707. D. A. Griffin and J. Staunton, J. Chem. Soc., Chem. Commun., 1971, 675. H. B. Hill and J. Torrey, Berichre, 1895, 28, 2597. T. Severin, P. Adhikary, E. Dehmel, and 1. Eberhard, Chem. Ber., 1971,104,2856. Bamjield and Gordon The second method uses P-chloronitrothylene in a condensation reaction with enamines to give aryl-substituted 1,3-dinitrobenzenes (99) or annelated derivatives (loo).' 13' On the limited evidence the yields are better than in the previous example and further exploration would seem justified. A related reaction is the thermal cyclization of 6-nitrosubstituted 1-dimethylaminohexatrienes. ''I3 j3-Ketoaldehydes will also trimerize (cf: 2-nitromalondialdehyde) under acid conditions to provide a useful route to 1,3,5-triketobenzenes (101); experimental conditions have been described in Organic Synthese~,'~' although it would appear that enamines give higher yields when used.'" COCHzR I RCH$OCH2CH0 acid RCHzCO COCHzR (101) R Yield % Ref.H 46 189 Me 21 190 PhCH, 29 190 (via enamine) H 94 191 On the other hand aromatic aldehydes have been prepared by an interesting route which involves the use of chloromethyleneiminium salts. 19* Thus, with vinylketones (102) the products are 6-chloro-1,3-benzenedicarboxaldehydes(103); la' H. G. Viehe and R. Verbruggen, Chimia, 1975, 352. "'C. Jutz and R. M. Wagner, Angew. Chem., Int. Edn. Engl., 1972, 11, 315. la9 Organic Syntheses, Coll. Vol. 111, John Wiley, 1955, p.829. 190 T. M. Harris, S. Boatman, and C. R. Hauser, J. Am. Chem. SOC.,1963,85, 3273. 191 N. K. Kochetkov, lzv. Akud. Nuuk SSSR, 1953,991 (Chem. Abstr., 1955,49,2308). 19' A. Holy and 2.Arnold, Collect. Czech. Chem. Commun., 1965, 30,53. 48 1 Aromatic Benzene Compoundsfrom Acyclic Precursors however, the yields are only moderate although the reagents are cheap and the conditions relatively mild. In a similar fashion acetylacetone was found to give 2,4-dichlorobenzaldehyde in 84% yield. This interesting reaction has not been extended to substituted 1,3-diketones. (102) (1031 R = H, Me, NMe2, Yields 3348% Polymethines (104) will also react with an iminium salt to give the tricarboxaldehyde (103, and Russian workers have claimed that piperylene (106) can be converted directly into (105) under Vilsmeier-Haak condition^.'^^ The yields from these routes are 40% and 28% respectively.-3----OHC CHO Mew In a variant of this reaction it has been shown that condensation of the pentamethylene iminium salt (107) with methyl ketones will give the phenones (108) in excellent ~ie1ds.I~~ This reaction will however be of little practical importance until the effect of substituents on the chain has been studied. COR ( 107 1 (108) 193 T. I. Lonshchakova, B. I. Buzykin, A. G. Liakumovich. and V. S.Tsivunin,J. Org. Chem. USSR, 1978, 14, 550. 194 C. Jutz, R. M. Wagner, A. Kratz. and H. G. Lobering. Ann.. 1975, 874. 482 Barnfield and Gordon A related electrocyclic cyclization of a 6-hydroxyhexatrienecarboxaldehyde to benzaldehyde suffers from the same limitations.' 95 Of greater utility, for annelated products, is the condensation rearrangement reaction between (109) and enamines to give (I R2 3 R (109) (110) R' R2 R3 n Yield % Ph Ph Ph 3 64.5 Me Ph Ph 3 17 Ph Ph Ph 4 56 Ph Ph Ph 6 63.2 Ph Ph Ph 10 19.2 Condensation routes to benzene carboxylic acids or esters without any other functional groups on the ring are very few in number.They are, however, readily made by Diels-Alder cycloaddition reaction; as stated earlier this is outside the scope of this report. In general those reactions which give simple monocarboxylic acids are, apart from those made uiupyrylium salts described earlier, of little import to the synthetic chemist.Exceptions might be those involving annelated benzene rings.'" The same comments can be made about routes to simple benzo- or phthalo-nitriles. There are however several useful routes to mono-, di-, and tri- nitriles in the naphthalene series.' 98-200 An interesting recent example involves the extrusion of sulphur from the 2-thiabicyclo[3.2.0]hepta-3,6-diones (1 11) to give the CN I Me (111) (112) R' , R2,and R3 = H, Me, or CMe3 LV5 H. Althoft, B. Bornowski, and S. Dahne, J. Prakt. Cham., 1977. 319, 890 G. Mark1 and H. Baier, Tetrahedron Lett.. 1968,4379. J. B. Dickenson and W. Reusch, Synrh. Commun.. 1983, 13, 303. C. Jutz and H. G. Peuker, Synthesis, 1975, 431.I99 H. Moureu, P. Chovin, and G. Rivoal, Bull. SOC.Chim. Fr., 1946. 106. L. Heiss, E. F. Paulus. and H. Rehling, Ann., 1980, 1583. Aromatic Benzene Compounds from Acyclic Precursors substituted phthalonitriles (1 12) in 652% yield.20' The starting materials are fairly readily obtainable by the thermal rearrangement of the (2 + 2) cycloaddition products from the corresponding thiophenes and 2-butynedinitrile. 6 Alkyl-and Aryl-substituted Benzenes An important method for synthesizing polyalkyl- or polyaryl-benzenes from acyclic precursors is the cyclotrimerization of alkynes. The reaction was first described by Reffe in 1948,202 and requires the intervention of a transition-metal catalyst. The most common catalysts are metal carbonyl complexes of Ni, Co, and Cr or Ziegler-Nattar catalysts of the type R,AI-TiCI,.There have been many good reviews on this topic which are essential reading for anyone intending to carry out work in this area.203-207 Th ese reviews also cover the introduction of alkene, alkyne, ether, carboxylic acid (esters), amino, chloro, cyano, and silyl groups into benzene via the appropriately substituted alkyne, a route ignored in previous sections of this review. The balance will therefore be redressed in this section. There is little doubt that this type of reaction is of practical value in the synthesis of tri- to hexa-substituted alkyl- or aryl-benzenes. However, a disadvantage of the method when using monosubstituted alkynes is that the products are often mixtures of the two possible isomers, the 1,2,4- or 1,3,5-trisubstituted benzenes.Similarly the co-oligomerization of different alkynes often gives a gross mixture of R R R (114) (115) all the possible products with little or no selecti~ity.~~~-~~~ An elegant solution to this problem is to preform an alkyne-metal complex and then add a second different alkyne to give the substituted benzene. This is exemplified by the formation of 1,2,4,5-tetra-t-butylbenzene(1 18), where, although the overall yield is only 21%, its synthesis by a more conventional route would be difficult.208 A very interesting development on this theme is the so-called 'cobalt system'. 201 R. H. Hall, H. J. den Hertog. D. N. Reinhoudt, S. Harkema, G.J. van Hummel, and J. W. H. M. Uiterwijk, J. Org. Chem., 1982, 47.977. 'O' W. H. Reppe, 0.Schlichting, K. Klager, and T. Toepel. Ann., 1948.560. 1. '03 C. W. Bird, 'Transition Metal Intermediates in Organic Synthesis'. Logos Press, London, 1967, Chapter 1. *04 L. D. Yur'eva, Russ. Chem. Rev., 1974,43.48. 'OS K. P. C. Vollhardt, Acc. Chem. Res., 1977, 10, I. K. M. Nicholas, M. 0. Nestle, and D. Seyferth, in Transition Metal Organometallics in Organic Synthesis', Vol. 2, ed. H. Alper, Academic Press, New York, 1978, p. 26. 'O' C. Hoogzand and W. Hubel, 'Organic Synthesis viu Metal Carbonyls'. Vol. I.ed. I. Wender and P. Pino, Wiley, New York, 1968, p. 343. '08 U. Kruerke. C. Hoogzand, and W. Hubel. Chem. Ber.. 1961.94, 2817. 484 Bamjield and Gordon This involves the reaction of CpCo(PPh,),, with two or three different alkynes in a - R' C~COPP~~)~+ R'C=CR2 CpCo , R3CECR4 ~ R4 RZ R' R' (Cp=cyclopentadiene) (119) (120) stepwise manner via a cobaltocycle (1 19)to the benzene (120), as outlined below.209 ~1: R3qR2 This method offers a much higher degree of control than previous methods and is of value in producing highly substituted benzenes.A major disadvantage of this method is that it requires stoicheiometric quantities of the complex. However, a catalytic cobalt system has been described which makes use of a sterically hindered acetylene, e.g. bistrimethylsilylacetylene, that will add to other acetylenes without undergoing trimeriza tion. '05 A procedure can then be adopted where bistrimethylsilylacetylene is used both as a reactant and as a part of a recyclable solvent system, together with catalytic quantities of the commercially available CpCo(CO),.This method is particularly useful in the production of annelated benzenes (121). Recent developments in the chemistry of the trisilylmethyl groups should allow the introduction of further functionality into the derived products.* A miscellany of methods for the formation of polyalkyl- or polyaryl-benzenes by the condensation of carbonyl compounds have been described over the years but very few of these have any practical significance. A notable exception is the conversion of acetone into mesitylene, full details of which are given in Organic Syntheses." A particularly interesting and useful objective in this area is a method for producing unsymmetrical poly-aryl compounds and this is shown below.'09 T. Wakatsuki, T. Kuramitsu, and H. Yamazaki, 7etrahedron Lett., 1974,4549. 'lo T. H. Chan and I. Fleming, Synthesis, 1979, 761. '"Organic Syntheses, Coll. Vol. 1. John Wiley. May 1961, p. 341. 485 Aromatic Benzene Compounds from Acyclic Precursors R' R' Me2" (122) R' R2 TempPC Time Yield % (122) H Ph 140 1 hour 70 H Me 200 3 days 60 H Me 140 7 days 58 Ph H 140 3days 95 Me H 200 12 hours 95 Although good to excellent yields of (122) can be obtained the process tends to be slow and unsuitable for thermally labile A good route has recently been proposed using Grignard reagents and is illustrated in Scheme 6.213-215The stepwise process is especially useful for unsymmetrical m-terphenyls.2 l5 R' R' R2 Yield % (123) H H 41 H Me 58 R' 4-OEt H 90 (123) Ph H 55 Scheme 6 'I2 R.W. Jemison, T. Laird, and W. D. Ollis. J. Chem. Soc.. Chem. Commun., 1972, 556. '"M. A. Tius, Tetrahedron Lett,, 1981.22, 3335. M. A. Tius and S.Ali, J. Org. Chem., 1982,47. 3163. 'I' M. A. Tius and S. Savariar, Synthesis, 1983, 467. Barnfield and Gordon Pyrylium salts, as described elsewhere in this review, provide an entry point into polysubstituted aromatic compounds.’ 81 The routes which are applicable to alkyl- or aryl-benzenes are shown in Scheme 7. R‘ R’R2hRz2~CHZ=PPh3 R =-CH RL R 2 R RL Scheme 7 The original substitution pattern can be maintained in the derived 1,3,5-trisubstituted benzene by reaction with Wittig reagents, or a new Ssubstituent can be introduced by reacting a Grignard reagent with a 2-methylpyrylium salt.By suitable choice of substituents the pyrylium salt can be ring opened and then closed into a 1,3,4-pattern. Reaction with enamines provides a route into annelated benzenes, but in this case there is a competing reaction which gives the aryl ketone. In general the yields from the reactions of pyrylium salts are only moderate and the synthesis of unsymmetrical pyrylium compounds is often not easy. In spite of these drawbacks this method is worth considering especially as a route to 1,3,5-triarylbenzenes. The importance of phenylalkanoic acids as intermediates in the manufacture of drugs has lead to a great deal of synthetic activity in the l8 An important target of this work has been the drug Ibuprofen (124).The Upjohn Company have devised an elegant route to this compound from readily available aliphatic starting materials2 Patent, US. 4 189 596, 1980 (Chem. Abstr., 1980,93, 7862). ’”Patent, U.S. 4 154 962 (Chem. Abstr., 1976.85, 142842). ”* Patent, Ger. 2 806 424, 1978 (Chem. Abstr., 1978,89, 214917). 487 Aromatic Benzene Compounh from Acyclic Precursors 0 0 &CO, Et &C02Et CI Atop 6-CO, H 0 Acknowledgement. We would like to thank Dr. T. S. B. Sayer for his invaluable assistance in the preparation of this review.

ISSN:0306-0012    

DOI:10.1039/CS9841300441

出版商:RSC    

年代:1984

数据来源: RSC