摘要:

The Royal Society of Chemistry Chemical Society Reviews Editorial Board Professor H. W. Kroto FRS (Chairman) (University of Sussex) Professor M. J. Blandamer (University of Leicester) Dr. A. R. Butler (University of St. Andrews) Professor E. C. Cdnstable (University of Basel, Switzerland) ' Professor T. C. Gallagher (University of Bristol) Professor D. M. P. Mingos FRS (Imperial College London) Consulting Editors Dr. G. G. Balint-Kurti (University of Bristol) Dr. J. M. Brown (University of Oxford) Dr. J. Burgess (University of Leicester) Dr. N. Cape (Institute of Terrestrial Ecology, Lothian) Professor B. T. Golding (University of Newcastle upon Tyne) Professor M. Green (University of Bath) Professor A. Hamnett (University of Newcastle upon Tyne) Dr.T. M. Herrington (University of Reading) Professor R. Hillman (University of Leicester) Professor R. Keese (University of Bern, Switzerland) Dr. T. H. Lilley (University of Sheffield) Dr. H. Maskill (University of Newcastle upon Tyne) Professor A. de Meijere (University of Gottingen, Germany) Professor J. N. Miller (Loughborough University of TechnoIog y ) Professor S. M. Roberts (University of Liverpool) Professor B. H. Robinson (University of East Anglia) Professor M. R. Smyth (Dublin City University, Republic of Ireland) Professor A. J. Stace (University of Sussex) Chemical Society Reviews aims to foster current progress in the chemical sciences and related disciplines. The journal has the broad appeal necessary to enable scientists to benefit from recent advances made in research outside their immediate interests.In particular, students embarking on a research career should find Chemical Society Reviews a particularly Chemical Society Reviews (ISSN 0306-0012) is published bimonthly by The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge, UK CB4 4WF. All orders accompanied by payment should be sent directly to The Royal Society of Chemistry, Turpin Distribution Services Ltd., Blackhorse Road, Letchworth, Herts., UK SG6 1HN. N.b. Turpin Distribution Services Ltd., distributors, is wholly owned by The Royal Society of Chemistry. 1996 annual subscription rate: EEA f120.00;Rest of World f123.00; USA $225.00.Customers in Canada will be charged the Rest of World price plus a surcharge to cover GST. Customers should make payments by cheque in sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank. Second-class postage is paid at Jamaica, NY 1141-9998. Airfreight and mailing in the USA by Publications Editorial Staff Managing Editor Martin Sugden Editorial Production Peter Whittington Editorial Secretary Debbie Halls Editorial Office The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge UK CB4 4WF Telephone +44 (0)1223 420066 Facsimile +44 (0) 1223 420247 Electronic Mail (Internet) rscl@rsc.org or sugdenml@rsc.org http://c hem ist ry. rsc.org/rsc/ Advertisement sales Telephone +44 (0)171 287 3091 Facsimile +44 (0) 171 494 1134 Typeset by Servis Filmsetting Ltd.Printed in Great Britain by Black Bear Press Ltd. stimulating and instructive springboard to further reading. The Editorial Board encourages an international and interdisciplinary approach to science, which is reflected in the succinct, authoritative articles commissioned. The Board members welcome comments and suggestions; these should be directed to the Managing Editor Expediting Services Inc., 200 Meacham Avenue, Elmont, NY 11003, and at additional mailing offices. US Postmaster: send address changes to Chemical Society Reviews, c/o Publications Expediting Services Inc., 200 Meacham Avenue, Elmont, NY 11003. All despatches outside the UK by Bulk airmail within Europe and Accelerated Surface Post outside Europe. PRINTED IN THE UK. 0 The Royal Society of Chemistry, 1996. All rights reserved. No parts of this publication may be repro- duced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, recording, or other- wise, without the prior permission of the publishers.

ISSN:0306-0012    

DOI:10.1039/CS99625FX005

出版商:RSC    

年代:1996

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS996250X007

出版商:RSC    

年代:1996

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99625BX011

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Nitric Oxide in Biology Its Role as a Ligand R. J. P. Williams University of Oxford Inorganic Chemistry Laboratory South Parks Road Oxford OX1 3QR UK 1 Introduction Interest in nitric oxide as a chemical affecting human life has increased dramatically over the last twenty years,' -3 at first through its damaging effects. This simple molecule was recognised in the air many years ago arising as a component of the exhaust fumes of petrol-driven vehicles and is now treated as generally injurious to health. One glance at its electronic structure as a free radical raises some alarm but it is also a powerful oxidising agent and a powerful reducing agent. Its final oxidation product as nitrate which is carried in drinking water is also a known risk especially to the health of the foetus and the very young. Much of the nitrate in drink- ing water comes in fact from agricultural fertilisers and it can be reduced back to NO'. Nitrate acts as a nutrient for nitro-bacteria. However the recent more dramatic increase in interest has arisen not from pollution but from the finding that nitric oxide is a natural endogenous chemical messenger between cells in higher animals including humans. The full role of its action is not known but the simplest is as a relaxant of smooth muscles. In essence nitric oxide is released by endothelial (surface) cells to adjacent smooth or posture muscle cells where it causes their relaxation. These findings have revealed the value of nitrates and other organic nitro-com- pounds especially nitroglycerines long used as medicines for heart Professor Williams is now Emeritus Fellow at Wadham College and Emeritus Professor Oxford University. He was born in 1926 and edu- cated at Wallasey Grammar School. He studied chemistry at Merton College Oxford graduating in 1948. He took his doctor 5 degree at Oxford in I951 working with Professor H. M.N.Irving. Wth Professor A. Tiselius (Uppsala Sweden) 1951-1 952 he developed certain analyti- cal chromatographic methods. He became lecturer and tutor in Chemistry at Wadham College and at the University of Oxford 1955-1 965. Afrer a year at Harvard University 1965-1 966 he changed direction to teach biochemistry until 1974. He became a Reader in 1972 and Napier Royal Society Research Professor at Oxford 1974-1 991 and was elected Fellow of The Royal Society in 1972 and is also a foreign member of the Swedish Portuguese Czechoslovakian and Belgian Science Academies. He has given named lecture series in several European and American Universities and plenary international lectures at many Chemistry Biochemistry and Biology Conferences. He is a medallist of The Biochemical Society (twice) The Royal Society (twice) The Royal Society of Chemistry (twice) The European Biochemical Societies (twice) and the International Union of Biochemistry. He was a founder member of The Oxford Enzyme Group which has devised many new methods for the study of in vitro and in vivo biological systems. His major research contributions concern metal ions and theirjknc- tion in biological systems. The emphasis in this work is that all biological system are derived from a combination of inorganic and organic compounds. There is no life without minerals and during the course of evolution there has been considerable changes in mineral chemistry due to life. The work is summarized in two recent books -The Biological Chemistryof The Elements (I 991) and The Natural Selection of The Chemical Elements (February 1996). 77 patients suffering from angina. In essence the action of the NO' formed on reduction of the nitrate or nitro-compound is the relax- ation of the arterial wall allowing a faster through-put of blood to the suffering muscle. As Nobel (of the Nobel Prize) remarked it is remarkable that the material nitro-glycerine from the properties of which he became rich and famous as the developer of explosives should be taken by him to alleviate angina. With this short intro- duction to the biological interest of NO' I shall now go back over its chemistry and then return to its function in living organisms. 2 Physical Properties Nitric oxide has but a small dipole moment and is gaseous down to -152"C.It is slightly and almost equally soluble in many solvents. At a vapour pressure of 1 atm it has a partition coefficient between solvents and the gaseous phase in the following order methanol 0.36,benzene 0.30,pentane 0.25 and water 0.07. Thus it is expected that like 0 it can diffuse relatively easily from water to organic sol-vents such as biological membranes and the hydrophobic interior of proteins. However it cannot be generated in high concentrations in water since its solubility is quite low and almost the same as that of dioxygen and carbon monoxide. 3 Chemioal Properties 3.1 Oxidation State Diagram for Nitrogen The oxidation state diagram for nitrogen is given in Figure 1 together with that for some other non-metals. Thermodynamically I pH=7 O&0 (ref Pt,H21e,,=10-')+'I f +5+_I+4t I1 I I I I I1 1 11 -4 -3 -2 -1 O +I +2 +3 +4 +5 +6 Oxidation state n Figure 1 The oxidation state diagram for non-metals at pH = 7.0. Note that dioxygen itself cannot react with N to give NO' but can give NO,-at thermodynamic equilibrium. NO' is formed in car engines by less stable oxidation states of 0 reacting with N,. speaking not only the oxidation of NO' by 0 is readily brought about but so is reduction even to NH3 by many hydrogen contain- ing molecules. The kinetic control of its reactions is then critical. 3.2 The Chemistry of Nitric Oxide with Oxygen Every school child knows (or did know) how to make NO' from nitric acid and copper turnings and such a child also knows that NO' does not exist for long in air due to the reaction with 0 to give NO a brown gas. However this reaction (1) is termolecular 2NO' + 0,Z2N02 (1) and hence although the free energy change is strongly in favour of NO the reaction rate is only fast at high concentrations so that in fact NO' persists in the gas or in solution phases at mol dm-3 or below for at least a minute. Immediately this warns us that to understand the action of NO' we must look not just at the thermo- dynamics of its reactions but at the kinetics of its release and binding. NO with water goes on to give both nitrite and nitrate as mentioned above while reduction of nitrate via NO' is used by both man (in fertilisers) and bacteria as a source of ammonia. NO' reacts directly with superoxide to give peroxynitrite an important fast reaction in biological systems H+ +O,-+ NO'-OI-P+NO; (2) The reaction may be involved in some biological defence mecha- nisms where it is said that OH. is deliberately generated but it exposes organisms which produce it to risk since the hydroxyl radical is an extremely potent reagent. Reduction of NO' by hydrides requires catalysts usually based on metals showing again that NO' itself is not very reactive. The cata- lysts used by organisms will be described later but almost any metal surface is catalytic. From the above redox reactions we can deduce that the standing concentration of NO' in early periods of Earth's history was extremely low since 0 had a pressure of only some 10-I2 atm. Today the 0 pressure is 0.2 atm and N gas is even more available. The equilibrium for reaction of 0 and N is just in favour of nitrate Figure 1 and though reaction to NO' is unfavourable 0 + N,c"2NO' AGO= 12.8 kJ mol-I (3) we must always always remember that NO' itself will dispropor- tionate to higher and lower oxidation states. There is always a low steady state concentration of NO' in the air. Why is the present atmosphere as it is? Obviously its gases only exist in their present concentrations in the absence of catalysts. 33 Reactivity of NO with Other Non-metals Despite its considerable potential in both oxidation and reduction reactions NO' does not attack many non-metal centres rapidly in the absence of catalyst^.^ In this it resembles dioxygen. NO+ is of course a different and very powerful electrophile in concentrated nitric-sulfuric acid but is not easily obtainable from NO' at neutral pH. Just as proteins and nucleic acids are relatively safe from attack by 0 so they are relatively kinetically stable in the presence of NO'. NO' does react with thiolates and even with sulfite but these reac- tions are not known to be of great importance in biological system except possibly in defence mechanisms (see below) and possibly in the clearance of NO'. Before we consider reactions with metal ions free or in com- plexed states we note that we can conclude that the diffusion of NO' at low concentrations (10-6~)through water or organic solvents is not greatly impeded by either physical or chemical barriers due to non-metals and it is likely to come to rest on metal ions. 3.4 Coordination Chemistry of NO NO' is an extremely powerful ligand giving binding constants to CHEMICAL SOCIETY REVIEWS 1996 metal ions free or in complex form often greatly in excess of those of CO and almost always much higher than those of 0,.The range of NO' compounds known are illustrated in Table 1? One feature is Table 1 Some nitroso complexes5 Complex Comment [Fe(CN),(NO)l2-Nitroprusside used as medicine 1Fe(CN>,(No)12-Reduced nitroprusside IRu(NH,>,(N0>l3-ICr(NO>,I Compare Cr(CO) WKO)(PPh,),(NO)I Ir-N-0 bond angle 124" the variable nature of the bound NO'. Low oxidation states of metals give M.NO- while high oxidation states give M-NO+ with increase or reduction of the effective metal oxidation state on binding. The nature of NO' in a complex which can be studied by its IR spectrum has a decided effect on the reactivity of the group in different low or high oxidation state complexes. One example is nitroprusside which is a complex formed from [Fe(CN),(H,0)l2- i.e. an Fell1 complex and can be written [Fe2+(CN)5-NO+]2-. The bound NO+ reacts with electrophilic reagents such as OH- to give a nitrito complex and with amines to give a variety of products which have been used to diagnose primary and secondary amino-compounds. Thus RNH,-ROH+N (4) R,NH+R,N,O (5) In the presence of excess of amine the nitroprusside gives finally an amino-complex of [Fe(CN),(H,0)l2- the amine replacing H,O. Other examples of the binding of NO' in different electronic states are given in Table 1. 35 The trans-Effect NO' is such a powerful ligand that it can dominate the coordination sphere of a metal. Thus reduction of [Fe(CN),(NO)J2- gives (Fe(CN),(N0)]3-which releases CN- to give five-coordinate [Fe(CN),(N0)l2-. Here the paramagnetic electron appears to reside somewhat equally on the Fe and the NO'. In any event the structure shows the Fe out of the plane of the four cyanides and the Fe-NO bond distance is very short 1.5A Figure 2? The domination of NO' as a ligand here shows that it has such a strong trans-influence as to Figure 2 The structure of [Fe(CN),(NO)12-. FeNO is almost linear. Note bond lengths. remove the sixth ligand. This situation is common to iron porphyrin complexes and probably to haem-proteins but there is a curious parallel in vitamin B chemistry where strong o-donor ligands so reduce the trans-bonding strength of Coili also of 3d6 configura- tion as to give a five-coordinate complex? The over-riding of the 18-electron rule is reminiscent of the fact that this rule often fails in complexes later in transition metal series e.g. [Ni(CN),I2- which have a larger number of nd electrons. NITRIC OXIDE IN BIOLOGY ITS ROLE AS A LIGAND-R. J. P. WILLIAMS 4 Detection of NO* The difficulty of detection of NO' especially in living organisms lies in the low range of concentration of this chemical i.e. lo-9-lo-6~. One way of detecting NO' uses spin-traps together with EPR mea-surements.* One reagent for following these levels of NO' is the N-methyl-o-glucamine dithiocarbamate (mgd) Fe*+ complex. The Fe2+ forms an NO'complex [Fe(mgd),(NO)] with a simple three- line EPR spectrum. The method can be used on urine samples to estimate the in vivo level of NO'. An alternative detection method is to use the NO' scavenger 1,2-diisopropyIidene cyclohexa-3,5- diene as a spin trap. NO' can also be detected by electrochemical oxidation using an NO' selective microprobe electrode.9 0 CO and NO. Comparative Ligand Strength Uppermost in a chemist's mind when considering the value of NO in biological organisms is the relative properties of this molecule and those of 0 and CO. Obviously 0 is overwhelmingly the most concentrated in ambient conditions since its atmospheric partial pressure is so much greater. However its binding to metal ions is intrinsically weaker for the obvious reason that it is neither a really good donor nor a good acceptor. Thus while some iron complexes bind 0 quite well and bind CO and NO' more strongly others bind NO' or CO but do not seem to bind 0 at all at ambient pressure. The factors affecting binding apart from the metal ion itself and its other coordination partners are the steric demands of the neigh- bouring groups to the coordinated complexes and the effect of their electronic structure and that of bound 0 NO or CO. While dioxy- gen binds to reduced metal ions giving a state close to M+O,- CO remains neutral and NO' may be NO- but NO+ is most likely. Clearly 0 as 0,-and in some cases NO as NO- bindings which take on negative electrostatic charge are favoured by nearby (distal) positive charge and/or H-bonding capability while NO+ binding is favoured by nearby negative charge. The conformations preferred by the small molecules themselves are M-0-0 bent M-C-0 linear and M-N-0 linear as NO+ (iso-electronic with CO) or bent NO- (iso-electronic with 0,).Thus for example steric inhibition around the binding metal-ion to the linear forms can favour bent 0,-binding over linear CO or NO+ binding. This appears to be important in the enhancement of dioxygen binding to haemoglobin relative to that of CO; see below. We turn next to the thermodynamic binding strength of the three small molecules to specific free-iron porphyrin complexes and then to haem-proteins. 5 Haem Complexes An understanding of the reactions of haem with NO' is a pre-requi- site to the understanding of the function of NO' in organisms since the receptor for NO' is a heam protein. The parallel with receptors and carriers of dioxygenlo and probably carbon monoxide is very close. The usefulness of the parallel is increased by the fact that the synthase for NO' is alsoa haem protein and its activity is due to dioxy- gen and not nitric oxide interaction with the haem. All of the 0 or NO' systems are poisoned by CO to some degree. Thus we need to understand the relative binding strengths of NO',CO and 0,to haem. Fortunately there are structures and much detailed study on the equi- librium binding constants spectra and magnetism of many model haem compounds of a variety of degrees of sophistication. If we start our discussion from free iron porphyrins in water or organic solvents then in both the ferrous Fe" and ferric Fell] states they are high-spin. The Fell ion and the Fe"' to a lesser degree are bound out-of-plane. Now on binding alternative donors to the z-axis coordination sites Figure 3 in either one or two steps the iron in both oxidation states may switch to the low-spin state and move in- plane. A weak donor such as fluoride is not effective in this regard and we can write a series of binding strengths (assisted by m-bonding) which is neutral oxygen (H,O)<charged oxygen (OH-)<amines (NH imidazole or pyridines)<thiolate<CO (CN)- NO' i.e.the strength with which the iron is driven to become low-spin. The exact order is somewhat erratic where combinations Figure 3 A typical six-coordinate haem complex here showing NO' bound opposite imidazole. Fe.NO is bent. Note bond lengths. of two donor types are used to drive the metal ions to low-spin but they do so in roughly the expected way. Thus one nitrogen base and one water in the fifth and sixth coordination positions respectively give high-spin complexes but replacement of the water by many other donors (not oxygen donors) drives the system to low-spin." NO is so strong a ligand that in the Fell condition the iron goes low- spin even in the absence of a sixth ligand. The stereochemical adjustments on binding of a molecule to the sixth position are considerable. If the Fell ion moves in-plane then the fifth ligand bound at right angles to the haem moves with it and adjusts its bond length Figure 4. On the binding of simple ligands such as ammonia or even 0,,CO or CN- the movement is always to give shorter bond lengths Figure 4. For NO' binding to a haem replacing say water but with iron bound initially to a nitrogen-base with iron out-of-plane on the nitrogen-base side the change could be dramatically different in that the NO' displacement of the water could cause a complete loss of the nitrogen-base and the creation of a low-spin five-coordinate nitroso complex with the iron moved out-of-plane on the opposite side of the porphyrin from the nitrogen base which then moves away from the iron. We shall see the value of these adjustments later when we consider haem-proteins.' 6 Reactions of NO-in Metal Complexes 6.1 Reversible Addition of NO to Non Metals in Complexes The classical test for thiols is the nitroprusside reaction [Fe(CN),(NO)]'-+ RS-% Fe(CN),Ni <" 13-which is reversible. This somewhat curious addition occurs with other sulfur compounds ,for example many thiols,and with the Fe,/S centres known in many electron transfer proteins (ferredoxins) RS- +NO'ZRSNO- (6) (RS-)4Fe4S4+NO'=[( RS -),Fe,S,NOI (7) NO also reacts with sulfite to give a variety of intermediates which decompose to give sulfate and reduced NO compounds. The impor- tance of these reactions in living systems is not known at present. 7 Biological Uses of Nitric Oxide The full involvement of nitric oxide in biological systems is not yet known.'-3 As stated already we need to have in mind the involve- ment of nitric oxide in bacterial metabolism as well as its function as a cell-to-cell messenger and detoxifying agent in higher animals CHEMICAL SOCIETY REVIEWS 1996 F-helix histidine (F8) %GHaem normal -0,C rG;Propionates PNH valine (Ell) Porphyrin PorphyrinDistal histidine (E7) (a) E-helix GOxygen XY z XY2 High-spin Low-spin or Felll) or Fe(tlt) Figure 4 (a)The binding site of 0 to iron of haemoglobin; the Fell is high-spin and five-coordinate. (b) The change in the structure of haemoglobin after 0 binding. (c)The intricate rearrangement of more distant parts of haemoglobin on binding ligands such as 0 CO or NO'. The figure is taken from ref. (1 1) but it is based on the work of M. F. Perutz. Table 2 Functions of Nitric Oxide Synthase1-3 1 Relaxation of smooth muscle by endothelial cells regulation of blood flow (NO' as messenger) 2 Central nervous system responses memory? (NO' as messenger) 3 Resistance to bacteria and lower organisms (NO' with 0,-as poison). The enzyme is in macrophages N.B. Three different but very similar enzymes are involved in the different tissues. Table 2. As discussed earlier nitric oxide entered the environment as dioxygen pressure rose and it could well be that there are sig- nalling proteins or even buffering proteins against the presence of NO' even in some early prokaryote (bacterial) cells. Thus NO' was probably treated at first by cells as a poison. The uses of NO' today in higher animals relate mainly to signalling between cells. Some of the systems are listed in Table 2. However a second use may well be deployment as a protective poison to kill certain bacteria and yet a third function is in certain activities of the brain. The use of NO' as a poison or a messenger needs a generating system but of course the generator and the receptor have to be in appropriate places and reasonably close proximity since NO' has not a long life if 0 or 0;-is available in the presence of metal ions. Only certain cells are included in a particular biological message system. It is important to realise that a message- or poison-generating system requires timed activity so that it is not always switched to the on-situation. The switches will be described below. 8 NO binding to Proteins Turning to haem proteins from the functional point of view the two factors of importance within the proteins are the relative affinity for 0,,CO and NO' and the stereochemical adjustments forced upon the protein by binding the diatomic molecules. The affinity is described in Table 3. It shows that many haem sites are likely to be partially saturated with 0 under normal aerated conditions. To avoid this binding of 0 the protein must be designed to reduce logK to below 3 i.e.binding is made slight in the atmosphere [p-(0,)=0.2 Torr]. CO is not a competitor except by accident (CO-poisoning) since CO is at very low concentration in most environments but not all. To be Table 3 Binding constants for NO CO and 0 to haem and haem- proteins in watep (37 "C) units log Klmoles 1-Ib Compound NO co 0 Model haem compounds 15.0 9.O 5.O Haemoglobin 12.0 7.O 5.O Haemocyanin 4.0 Hemerythrin 4.0 0 Note the saturation concentration of all three gases in water is around M. While 0 may be close to this level CO is far from it M) while NO is produce$ locally and effective at M in living systems. Changes in the fifth ligand or the protein can change all the values by at least ?2 so that rounded values only are given. effectively bound NO' need not be present at much greater than nanomolar levels due to the great strength of NO' binding. Thus any local production of NO' almost inevitably leads to binding by any haem which is nearby and open-sided or bound by water in the sixth position. To understand the detailed functioning of NO' we need next to analyse the nature of haem protein structures. 8.1 The Nature of Helical Proteins13 A major feature of haem proteins is that they are usually helical in construction though one or two exceptions such as cytochrome f are known. The calcium-binding protein calmodulin which will interest us too in this article is also helical. In fact calmodulin is quite like cytochrome c' (see below) in that both are based on four- helical bundles Figure 5,the one binding Ca2+ and the other haem and NO'. In both cases the protein structure is adjusted by the binding of the small group. This is a common property of many helical proteins which means that they are extremely useful as mechanical signalling devices on receipt of a message from a small molecule. If the proteins are made of subunits each of which can bind Ca*+ or NO' (or for that matter 0,) then the binding of the several units can become cooperative. This is seen in haemo-globin (40,) Figure 4(c) in calmodulin (4Ca2+) and in some cytochromes c' (NO). However a helical segment of a protein can also be part of an enzyme so that the enzyme conformation and activity are adjusted by changes in its helices. This is a common NITRIC OXIDE IN BIOLOGY ITS ROLE AS A LIGAND-R. J. P. WILLIAMS c 128 HINGE 128 c E I Figure 5 (a)The four-helical bundle of cytochrome C' shown as a side-by-side dimer. (b)The four-helical bundle of calmodulin shown as a dumb- bell covalently linked (by a long central helix) dimer. feature in kinases,I4 which transfer phosphate groups and is often called a simple helical hinge but a better description is that of a rising-hinge which requires energy from the binding energy of the substrate molecules for movement and automatically relaxes when the product molecule leaves. It takes little imagination to see that a helical signalling protein can be coupled to a helical rising-hinge enzyme so that the signal switches on the enzyme Figure 6.14It has been found that it is a helical calcium-binding protein calmodulin which switches on NO-synthase while it is a helical NO-binding signal protein con- taining haem which switches on guanyl cyclase so that a message runs through the inorganic components (Scheme 1). hinge release diffuse hinge G-protein Ca2+ -Fe haem([) -NO' -Fe haem(ir) + Phosphate (NO' synthesis) (+NO') (Guanyl cyclase) Scheme 1 In this article we are not concerned with a description of the release of calcium into the cell or with the metabolism of guanyl phos- phates but in both these cases too the machinery of protein adjust- ment to generate gating of ion movement (membrane channels in proteins) and enzyme on/off switches (protein hinges) is based on re-alignment of helices. We turn to the particular cases of the NO* BUNDLE Figure 6 A schematic diagram of the way in which a protein lower part which responds to a signal NO' (or Caz-) can adjust the activity of an enzyme upper part by mechanical changes. The receptor for NO' is shown here as a helical bundle as in haemoglobin which adjust its helices so as to push together the P-sheet binding sites of an enzyme via action on a multi-helical hinge for example of a cyclase. L-Arginine Citruiline Ornithine diimide Figure 7 The reaction of arginine to give NO' in the presence of the enzyme NOS. The changes of the iron are also shown. synthesis by enzymic action and the NO-acceptor proteins and then to the reactions which control them or which they control. 9 Nitric Oxide Synthases (NOS) This enzyme carries out the reaction Figure 7 of arginine with dioxygen to give NO and citrulline. It is a multiple redox step a2 reaction and the enzyme has a complicated requirement for both dioxygen and a source of reducing equivalents as well as for the substrate arginine This multiple requirement is common to a wide range of dioxygen-using enzymes in biological systems and there is a close similanty between NOS and the cytochromes P-450 Their peculianty using reducing power to assist oxidation is due to the need to activate dioxygen by reduction Thus 0 and even O,*-(superoxide) are not very aggressive reactants while perox- ide is a fine attacking agent especially in the presence of metal ions Many biological oxidising agents use the pathway 2e O,+metal centre (M)-.M(O,) -MO+H,O) (8) Attack on the substrate here arginine is then through the higher oxene oxidation state of the metal e g FevO or FeiVO Generally MO+RH+ROH+M (9) when the enzyme has cycled back to the Initial state A large number of such enzymes including nitric oxide synthase use iron as the metal M and as often as not in haem We can now write a simple scheme of NO' production but this leaves us with some puzzles '5 The foremost is that this haem- protein is not seriously inhibited by NO' itself despite the fact that NO' is produced adjacent to the iron Yet the site reacts with 10 The NO Receptor This protein is not so well defined It rests in or on a membrane and probably through a conformation change Figure 4 it activates for- mation of cyclic guanidine phosphate utilising a G-protein This G-protein product then generates a signal in the cell which activates many synthetic pathways as well as causing muscle relax- ation in smooth-muscle cells If one looks for a simple NO' recep-tor protein which has the required properties then the obvious choice is cytochrome c' of bacteria Cytochrome c' has already been described in outline in Figure 5 The four-helical bundle re-arranges somewhat on binding NO' (or CO) to the haem which is an effective mechanical signal The mech- anism parallels that in haemoglobin (or calmodulin) but whereas haemoglobin is to some degree protected from CO (or NO') binding cytochrome c' is protected from 0 binding Using this model we can suppose that any haem receptor for NO is protected from 02,but not CO The haem-to-protein binding alters the haem itself in such a way as to change the protein conformation Figure 4(c) This conformation change alters the binding of the NO recep-tor to its neighbour guanine cyclase enzyme and switches it to the productive mode Figure 6 There is then the problem of NO' release from the receptor Some light may be thrown on this question by the enzyme action of nitrite reductase Before turning to this enzyme note in passing that there is a further way in which NO' could generate a message-release of haem -but we do not know if this has any importance 11 Fast and Smooth Muscles An Aside It is fascinating to note that the activation of the fastest muscles for animal movement depends on the fastest inorganic messenger which can bind strongly namely calcium while the smooth muscles can be activated or deactivated by a kinetically slower inorganic messenger namely NO' The two cannot be totally dependent in that all contraction depends on the standing calcium concentration Both activities are mechanically driven in very similar ways 12 Release of Nitroso-haem from Proteins Normally the release of haem from proteins is very slow In the pres- ence of nitric oxide however dissociation is much more rapid l6 As CHEMICAL SOCIETY REVIEWS 1996 Table 4 Metals in nitrogen cycle enzymes Enzymes Metal Nitrate reductase Mo and Fe(haem) or Fen/S Nitrite reductase Fe(haem)or Cu and Fe,/S Nitric oxide reductase Fe(haem) Nitrous oxide reductase cu N1trogenase Mo or V or Fe and Fen/SII mentioned above the NO' complex of haem iron weakens any trans (to NO') bonding so that the bond may break easily The haem NO complex in such proteins as myoglobin is then free-floating Model studies on the release of haem NO from myoglobin have been sum- marised by Mitra and co-workers l6 They showed that the rate of release was rapid The physiological importance of the observations IS unknown but it is known that removal of haem from proteins such as cytochrome b and myoglobin causes very significant conforma- tional changes in the protein which could be relayed to a partner enzymes such as guanyl cyclase 13 Denitrificationl7 Bacteria can utilise NO- as a primary source of oxidising equiva- lents in the absence of dioxygen The steps of reaction have to include the intermediates or NO ,+NO ,+=NO'+=N,O+N (12) In both cases NO' is produced from nitrite and in many cases it is known that NO' is a free intermediate The presence of NO' in human exhaled breath may well be due to bacterial breakdown of nitrate In the above reaction chain there are a variety of enzymes which use different metal ions Table 4 The structures of the metal sites are of great interest individually but most important for the purposes of this review are those which handle NO' The crystal structures of two nitrite reductases are known but because of its relationship to the NO' receptor we shall concentrate on the haem-containing enzymeis leaving the copper enzymelg to one side The major inter- est concerns the way in which NO' leaves the iron since its affinity for haem iron is so high The reaction is the simple one-electron reduction NO -+2H+ +ec"0 +H,O (13) In the structure in the absence of substrate the haem iron is con- sidered to be in the Fell state and open-sided (five-coordinate) Reaction with nitrite leads on to the formation of an FeiIi-NO complex and the release of water as in the above equation with the electron coming from Fefi The displacement of NO' is then by a near-neighbour phenolate side chain (of a tyrosine amino acid) The complex which remains is an Felli phenolate It requires a reductase to release the phenol of the tyrosine and initiate a new cycle of reaction The degree of conformational rearrangement of the enzyme structure during this set of reactions remains to be determined 14 Bacteriocidal Action Wherever NO' is produced in the same location as superoxide 0;- there is an extremely rapid reaction to generate peroxynitrite ON00 This reaction has been found to occur in certain human white blood cells macrophages and neutrophils Peroxynitrite is one of the most unstable peroxides and is able to attack a large number of protein side chains such as tyrosine thiols and imidazole The presence of foreign bacterial cells captured by protective macrophages is known to generate high concentrations of both nitric oxide and superoxide Thus it is highly likely that one method NITRIC OXIDE IN BIOLOGY ITS ROLE AS A LIGAND-R J P WILLIAMS Table 5 NO-generating drugs Drug Formula Nitroprusside Na,IFe(CN),(NO)l 2H,O Glyceryl trinitrate Propane 12,3 trio1 trinitrate Amy1 (pentyl) nitrate (CHJ2CHCH2CH,NO2 fast acting Sodium nitrite NaNO slow acting lsosorbide mononitrate" 1,4 3,6 Dianhydro D glucitol Isosorbide di ni trate" 1,4 3,6 Dianhydro D glucitol Fast acting drugs but can be put into tablets (Imdur) so as to be slow acting (ovei 24 hours) for the killing of bacteria is a joint action of NO' and 0,' in the form of ON00 Examination of many primitive organisms such as the horseshoe crab which is at least 500 million years old show that it uses NO' for protection against bacteria Thus maybe NO' has been associ- ated with life for as long as 0 was available in reasonable quanti- ties some 1-2 X 109 years ago 15 Nitric Oxide and the Brain'.* There are a large number of neuro-transmitters in the brain A major transmitter is the amino acid glutamate which was thought to act directly on electrical signals However it now appears that gluta- mate also acts on a receiving nerve cell by releasing NO' which in turn diffuses back to the donor nerve cell This may be a feed-back regulation of glutamate activity There is the suggestion that this joint action of glutamate and nitric oxide could play a role in poten-tiating memory 16 Drugs used for NO* Release and Control Patients with heart disease suffer from angina nervous pains in the chest arms or even in the face The pain arises when the arteries or muscles of the heart fail to deliver sufficient oxygen to it to main- tain full heart function Major relief arises from dilation of the arter- ies which is achieved by the release of NO' from orally administered drugs The main drugs used are listed in Table 5 and may be short- acting say a few hours or slower acting say over twelve hours Unfortunately there are also thought to be diseases where too much NO' is released Here the useful drugs are those which inhibit nitric oxide synthase In all the applications of these drugs there is the risk of side effects since NO' has so many different roles as a messenger One of increasing interest is the function in the brain 17 The Evolution of Biological Signals*O The presence of NO could only arise after the advent of dioxygen 1-2X lo9 years ago Essentially NO and 0 were initially poisons since life started as a reductive chemical system for CO and N Thus the progression in evolution of oxidative systems had to be Poison-,Protective-.Functional Use Reaction (Signal) Once a functional use is uncovered for a molecule or ion which has entered the environment then the organism may find a production 83 method It appears that an early use of NO was as a source of nitro- gen and energy but before that organisms may well have protected against NO' by evolving a warning signal (cytochrome c') Once multi-cellular systems arose the signal could be used differently The great advantage of NO' as a transmitter is that it is chemically selective and acts in a very low concentration range The develop- ment of NO' in its biological function may be usefully compared with that of Ca2+ which is really just a rejected poison by prokary- otes but dominates cell-cell communication 18 Summary The intention of this article is to reveal the involvement of in par ticular inorganic elements and one simple compound NO' in living systems and to alert chemists to the ways in which they can help our understanding The inorganic chemist is in danger of letting much of the excitement of this field slip through his fingers In order to pursue the on-going transformation of biological chemistry into a mixture of organic and inorganic chemistry such as in this case he or she must study in some depth biological subjects and use the bio- logical literature The revelations concerning NO' outlined here though yet far from complete indicate what is unfolding but nearly all the relevant chemistry is being done outside what are called 'chemistry' departments Much of the biochemistry of NO' is not understood and even less is known of that of CO I hope this article will show that the combination of inorganic and biological chemistries has a very productive future ahead of it NO' is now one of the most important messengers in higher living organisms 19 References 1 S Moncada M A Marletta J B Hibbs M Feelisch and R Busse ed The Biology of Nitric Oxide vols 1-4 Portland Press Colchester 1992-1994 2 S H Snyder and D S Bredt Scientrfc American 1992,226 (May) 28 3 A R Butler and L Williams Chem Soc Rev 1993,22,233 4 M J Clarke and J B Gaul Srrucrure Bonding 1993,81 147 5 D F Shriver P W Atkins and C H Langford Inorganic Chemistrv Oxford University Press Oxford 1990,p 506 6 J Schmidt H Kuhr W I Dorn and J Kopf Inorg Nucl Chetn Lett 1974,10,55 7 J M Pratt Inorganic ChemiJtrv of Miurnin B, Academic Press New York. 1972 8 A M Komarov and C H Lai Biochrm Biophvs Acra 1995,1272,29 9 K Shibuki Neurosci Res ,1990,9,69 10 G B Jameson and J A Ibers in Bio inorganic Chemistry ed I Bertini. H B Gray S Lippard and J Valentine Unibersity Science Books Mill Valley CA USA 1994 pp 167-252 11 R J P Williams Chem Rev 1956,56,299 12 J J R Frausto da Silva and R J P Williams The Biological Chemistrv of The Elements Oxford University Press Oxford 1991 pp 344-361 13 R J P Williams Eur J Biochem 1996,to be published 14 H Joio and R J P Williams Eur J Biochem 1993,206.1 15 J Wang D L Rousseau H M Abu Soud and D J Stuehr Proc Natl Acad Sci USA 1994,91,10512 16 T K Das S Mazumdar and S Mitra J Chem Soc ,Chem Commun 1993,1447 17 W G Zumpf Arch Microbiol 1993,160,253 18 V Furlop. J W B Moir S J Ferguson and J Hajdu Cell I995,81,369 19 J W Godden,S Turley,D C Teller,E T Adman,M J Liu,W J Payne and J LeGall Science 1991,253,438 20 R J P Williams and J R R Frausto da Silva The Natural Selecrron of The Chemical Efemenrs Oxford University Press Oxford 1996

ISSN:0306-0012    

DOI:10.1039/CS9962500077

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Man and the Elements of Groups 3 and 13 John Burgess Department of Chemistry University of Leicester LEI 7RH UK 1 Introduction This Review deals with interactions between compounds of the ele- ments of Groups 3 and 13 of the Periodic Table including the lan- thanides and actinides and the human body. It touches on their levels and role in healthy individuals and covers chemical aspects of their pharmacology toxicology and control. Both the introduction and the elimination of metal-containing species and their beneficial and harmful effects will be considered. There will be specific illustra- tions of the multitude of ways in which these (and indeed many other) elements interact favourably or unfavourably with the metab- olism and with malfunctioning of the human body. At appropriate points individual Groups 3/13 elements will be used to illustrate general features such as the balance between essential traces and toxic excess and the tailoring of ligands for maximum effectiveness. Thus complex stability and selectivity will appear in the sections on aluminium and on the actinides the hydro-philic/lipophilic balance (HLB) of complexes and its relevance to transport into and around the body will appear under the headings of aluminium and of gallium and indium and the transformation of very labile metal centres into exceptionally inert species in the section on yttrium. The abundances of the Group 3/13 elements in the Earth’s crust and in oceans and rivers vary over an enormous range from the ubiquitous aluminium (8.2% of the earth’s crustal rocks) to the very scarce indium (perhaps 0.24 ppm in the earth’s crust). Boron gallium scandium and yttrium occur at the 10-30 ppm level; these abundances are similar to that of e.g. cobalt or lead. Boron is the most abundant of the elements under discussion in the oceans at a concentration of about 4.4 ppm. The concentration of aluminium in the oceans is only about 2X lop3ppm; levels of the lanthanides are in the region of 10-6 ppm with scandium and indium present at even lower concentrations. An average man contains about 0.05 g of boron but vanishingly small amounts of the other Group 3/13 elements. These are present at levels less than probably very much less than lop5 moles per person. Boron is an essential element though toxic if taken in large amounts but the other elements are all irrelevant to the normal workings of the human body -indeed many are toxic some in very small amounts. But almost all these ele- ments often in the form of complexes of appropriate isotopes are useful in diagnosis or The following pages will illus- trate these applications for this set of elements -a complementary John Burgess obtained his BA. MA. and Ph.D. degrees at the University of Cambridge. Afer two years at Fisons Fertilizers’ R & D station at Levington he went to the University ofLeicester in 1967. He has remained there ever since and is now a Reader in Inorganic Chemistry. He has been inter- ested for many years in medium and pressure efsects on inorganic kinetics. More recently his inter- est in solvent efsects on kinetic parameters has extended to sol-vation of inorganic complexes. One facet of this is the study of synthesis and solvation of hydroxypyranone and hydroxy-pyridinone ligands and their complexes with particular refer- ence to possible diagnostic and therapeutic r6les. 85 discussion of the transition metals has appeared el~ewhere.~ References are mainly to reviews and to recently published articles; sources of detailed information can be tracked through the former. 2 Boron Boron was shown to be an essential element5” for at least some plants in 1923. That it could play an important r6le in mammals was shown much later in 1981 though the antibiotic boromycin had been produced from a new strain of Streptomyces antibioticus in 1967. Boron probably plays a part in calcium and phosphate metab- olism and in bone production and upkeep. It is also claimed to be involved in membrane function and in nucleic acid and lignin biosynthesis. As for so many essential elements an excess can be toxic; in the case of boron the fatal dose for an adult is as high as 15 to 20 g. Boric acid and borax (sodium borate) were used medicinally by Persian alchemists. Boric acid is a mild antiseptic and astringent being feebly bacteriostatic and fungistatic. Its aqueous solution is used as a mouthwash and as a skin lotion. It is also used in oint-ments and in dusting powders in which it is combined with talc starch or zinc oxide. The last combination is interesting in combin-ing the antiseptic properties of the boric acid with the healing prop- erties still only dimly understood of the zinc oxide. Boric acid has also been used in combination with a variety of other compounds variously including mercury(r1) chloride ‘mercuric oxycyanide,’ zinc acetate or borax in eye lotions. Perborate a rather fragile peroxo-derivative of borate has a significant usage in dental hygiene -both as an anti-gingivitis agent and in a cosmetic role. In the latter case it is acting as a bleach removing unsightly yellow or brown coloration and stains acting analogously to perborates In washing powders. The possibility of utilising neutron capture by boron- 10to release a-particles for the in situ destruction of cancerous cells seems to have been sug- gested almost 60 years ago. At that time there were neither suitable neutron sources nor suitable boron compounds so boron neutron capture therapy was not actually attempted until the late 1950s. Severe problems were encountered then with several fatalities resulting from insufficiently complete localisation of the boron- 10; stray loB emitted a-particles in the wrong places destroying essen- tial tissues. In the past few years there has been a renaissance of boron neutron capture therapy resulting from the refinement of the neutron beams available and the development of new boron com- pounds which can be delivered more efficiently to the required site. A large number of compounds of a variety of types are being assessed -pharmacological requirements compounds and methods of delivery have recently been fully documented? Promising compounds include H,N+BH,-CO,H the boron ana- logue of glycine and more effective (~)-4-boronophenylalanine. Many compounds containing polyhedral borane and carborane units have been evaluated of which the best known is sodium boro- captate (‘borolife’) containing the anion 1. Many related com- pounds are carborane derivatives of e.g. porphyrins thiouracil and polyamines.8 These contain substituents such as 23or 4; an anionic group such as nido-7,8-C,B,Hl I -,as in 2,increase hydrophilicity. The most recent compound of this type is the dendritic or cascade species 5 containing a multitude of sulfonate groups to confer water-s~lubility.~ 86 H ,SH 3 H 1 x oqn SO H w)-)c-OSO,H 5 3 Aluminium Aluminium compounds appear to have no useful role to play in human metabolism. Until recently it was generally believed that aluminium was irrelevant to the workings of the human body -a harmless perhaps even benign element. However in view of the high charge density of Al3+ and the likelihood of strong interactions between AI3+ and ‘hard’ ligands such as phosphates,one might well be suspicious that it could be very harmful if it reached the wrong place. Stability constants for aluminium complexes of ‘hard’ oxoanions are generally several orders of magnitude higher than for analogous complexes of such biologically important metals as mag- nesium(II) calcium(I1) and zinc(rI) so there may be problems result- ing from competition between A13+ and Mg2+,Ca2+,or Zn2+. Nowadays everyone is aware of the connection between alu- minium and various neurological malfunctions especially Alzheimer’s diseaselo and dialysis dementia. Aluminium is also implicated in some bone disorders such as osteoporosis and osteo- malacia. Whether aluminium causes or accelerates these problems -for example by interfering with calmodulin the multifunctional calcium receptor that plays a central r6le in cell regulation or with guanosine triphosphatase -seems still to be a matter of debate though majority opinion at the moment seems to favour accumula- tion in the brain as a consequence rather than causative. Readers wishing to consult the voluminous literature may refer to several exhaustive reviews on aluminium and its role in food medicine and biology.’’ Fortunately the human gastrointestinal tract is extremely reluc- CHEMICAL SOCIETY REVIEWS 1996 tant to allow simple A13+(aq.) and its hydroxo and polynuclear derivatives across. This is very important in view of several wide- spread uses of aluminium compounds in water treatment as food additives and in everyday medicines. The use of aluminium salts in water treatment is primarily cosmetic to give so-called ‘polish’ -to make it look clearer and more sparkling by using finely divided alu- minium hydroxide to remove very fine suspensions of organic mate- rial iron hydroxides and the like. In recent years iron(Ir1) salts have been replacing aluminium salts for this purpose though in view of possible problems with long-term iron accumulation in the body this may not be the final solution. Simple aluminium salts such as alum (potassium aluminium sulfate) were used for many years in foods such as pickles to give a sharper flavour while the cooking of acidic fruits such as rhubarb or citrus fruits in aluminium saucepans must have added to the aluminium intake of innumerable people since the widespread superseding of enamel by aluminium for cooking vessels. Tea leaves have probably the highest aluminium content of any plant up to 3% dry weight so there have been con- cerns over the aluminium intake of heavy tea drinkers. A strong cup of tea may contain fifty times the EU-recommended maximum level of aluminium. There are a number of aluminium compounds in the lists of permitted food additives published by the Ministry of Agriculture Fisheries and Food and by the European Union. These include anticaking agents such as sodium and calcium aluminosili- cates used in some table salt and sodium aluminium phosphate used as a leavening agent in cakes and biscuits. Aluminium metal is dignified with an E-number viz. E173 for its use in ‘silvering’ confectionery. Most of these permitted additives are so insoluble under physiological conditions as to preclude significant absorption in the body. Overall it does seem that the intake of aluminium from all these ‘unnatural’ sources adds only an insignificant amount to the daily intake of between 2 and 20 mg of aluminium in a normal diet. Aluminium compounds are extensively used in antacids. Aluminium acetate hydroxide phosphate and silicate and bismuth aluminate all feature in commonly used remedies for excessive acidity in the digestive regions. In most cases the aluminium-con- taining species are of fairly high relative molecular mass which coupled with their predominantly hydrophilic peripheries means limitingly low absorption under almost all conditions -except when ingested immediately before or after consuming such materials as fruit juice or marmalade. There is some evidence for the absorption of aluminium in the form of complexes with fruit acid ligands such as citrate 6,or with vegetation-derived ligands such as humates and fulvates. Oxalatoaluminates from the cooking of rhubarb in aluminium vessels (cf.previous paragraph) on the other hand should be far too hydrophilic for significant absorption from the gastrointestinal tract. Basic magnesium compounds such as the carbonate have also long been used as antacids often in conjunction with e.g.,alu-minium hydroxide. In view of magnesium’s favourable indeed essential role in human metabolism its compounds might well be deemed to be preferable alternatives to those of aluminium. The only drawback to the use of magnesium compounds in this respect is their significant laxative action -though in many cases that could be a distinct advantage! One needs to be aware that some drugs form stable complexes with Al3+ (and indeed with Mg2+)-antacids can decrease absorption of such drugs significantly sometimes to the point of making them totally ineffective. On the other hand the buffering effects of antacids may improve absorption if a drug is absorbed more effectively from neutral rather than acidic media. It should be added here that aluminium can actually play a beneficial role in the treatment of peptic ulcers -‘Sucralfate’,7,complexes through its aluminium to exposed protein thereby protecting deli- cate tissue from stomach acid and promoting healing. Aluminium compounds are applied generally in the form of ‘alum’ -potassium aluminium sulfate -to the skin to staunch bleed- ing from minor scratches and cuts. They have also been used in face creams and are very widely used in deodorants and anti-perspi- rants.’* The first real anti-perspirant was aluminium chloride solu- tion which as it was messy an irritant and destroyed fabrics found distinctly limited popularity. In the 1940s it was found that a number MAN AND THE ELEMENTS OF GROUPS 3 AND 13 -J. BURGESS of aluminium-hydroxide-chloride species of precisely defined composition but unknown structure were equally effective. Since then the formulation of this class of antiperspirant has been refined by the addition of some zirconium and also often of glycine. All these mixtures probably result in no significant addition of alu- minium to the body since it is unlikely that any of their aluminium- containing species penetrate the skin. But the simple alum mentioned at the start of this paragraph is a different matter since it is in contact with the blood and some complexation of A13+ by transferrin seems likely. There was for a time yet another possible route for the introduc- tion of aluminium through the use of aluminium-containing water supplies in dialysis treatment for kidney malfunction. The resultant dialysis dementia is now avoided by ensuring vanishingly small levels of aluminium in water used for such treatments. 6 CH,OR r; OR RO H 7 For several years there have been discussions as to the possibil- ity of slowing halting or reversing the inexorable progress of senile dementias by decreasing or at least controlling aluminium levels in the body. There are two problems here. The first is that it has proved well-nigh impossible to reproduce Alzheimer-type neu- rofibrillary tangles in the laboratory making in vitro tests of alu-minium-control drugs impossible. The second is that it is actually very difficult to diagnose Alzheimer's disease -definitive proof is only available through a post-mortem autopsy! It is therefore diffi- cult to assess the effectiveness of chelation therapy. Moreover there is always the concern that a chelator might mobilise aluminium and transfer it to an even more undesirable-site rather than transport it through the excretion system. The choice and tailoring of potential chelators for the control of aluminium levels provides an interest- ing illustration of drug design and development. It starts with the observation of the similarity between stability constants for alu- minium(1rr) and iron(Ii1) complexes. There is a marked correlation between stability constants for respective pairs of aluminium(I1r) and iron(m) complexes as shown in Fig. 1. Iron(rI1) complexes are almost always the more stable -ail the points in Fig. 1 are on the same side of the line corresponding to equal stability constants for A1 and Fe complexes. Therefore one has to be careful not to remove iron rather than aluminium. Some simple carboxylates and hy- droxycarboxylates have been tried for chelation therapy -succi-nate 8 and malate 9 were claimed to be of some value citrate 6 of rather less. The tris-hydroxamate chelator desferrioxamine 10 (R=H) so important and well-tried for control of iron levels in the body is another obvious reagent to try. It has been used in tests for control of aluminium levels since 1980 with several claims that it can produce a significant retardation of deterioration of mental fac- ulties. But its cost is high its availability is relatively limited its administration is not totally straightforward and it often has unde- sirable side effects on sight and hearing (though admittedly these are thought to be reversible). Nonetheless it was in early 1993 the only aluminium chelator approved for long term administration to humans. What is really needed is a cheap orally-administrable drug. Again it is profitable to start with iron(r1r) chemistry though for that metal there were two complementary objectives -to get iron effi- ciently into anaemic people and later to try to get iron out of people with iron overload i.e.the very large number of patients suf- fering from such diseases as haemochromatosis and thallasaemia. 87 / 30 t /' edta . dtpa -I-// 7otrlacetate transferrin0 c)j/rate4pentanedionate tartrate maltolate I I I I I I I 10 20 30 log K,(Fe3'complex) + Figure 1 Correlation between stability constants for alurninium(i~) and iron(iri) complexes. The line corresponds to equal stability constants for complexes of the two cations. BULK AQUEOUS PEASE E -80 I-b *Qp"p 0, 2p+ 0q G.YCOCALYX + t J?4 e =(?J-J< E = 30 tr L b 8p-p 0, 50 nm zPpJ;c 0qde*'',ct< t bJ! BImYERlEHBRANE tttltltt ttrfttft E-2 rlllllllrllll111 nm Ijl~lj~Il~~~~~Il~~~i~~~~j~tt~t~~~'~~~~'twy ,;;;,:.;;,,,,w,:,t;:::~~~:lY1.5~~m~~R nm CYTOPLASH E * 80 Figure 2 Simplified representation of the barrier to absorption from the gastrointestinal tract. The values of c are estimates of the relative petmit- tivity for each layer. The pyranones 11 and 12 maltol and ethylmaltol respectively are both good chelators for iron(I1r). Moreover they have the great advantage of being permitted food additives widely used in the baking industry. Reaction of these pyranones with a range of primary amines gives the series of pyridinones 13,which form even more stable complexes (log &-36) with iron(m) and whose sol-vation properties (hydrophilic/lipophilic balance or HLB) can be tailored according to the nature of the group R'. Such solvation tailoring is particularly important for chelators for metal removal. The ligand needs sufficient water solubility for successful oral administration but sufficient lipophilicity to be absorbed on passage through the gastrointestinal tract (GIT) and to be able to cross the cell wall to reach intracellular iron. The complex formed then needs the right HLB to leave the cell and make its way suc- cessfully down the excretion pathway. Fig. 2 shows a much simpli- fied version of the barrier to absorption from the GIT. The ligand has to proceed from one aqueous medium to another through the membrane bilayer whose central region is essentially hydrocarbon and an ill-defined intervening layer of intermediate properties. Unless there happens to be a specific transmembrane transport mechanism the species crossing such a barrier needs sufficient hydrophilicity to dissolve in the aqueous media and sufficient lipophilicity to transverse the fatty centre of the bilayer membrane. But the indicated requirement of both hydrophilic and lipophilic areas on the surface of the species may mean that synergic solvation in the intermediate region or at the membrane interface makes the species particularly well solvated there such that it is strongly absorbed and does not complete its passage through the membrane. A fine balance has to be achieved for the entering ligand and for the leaving complex. Despite these complications this type of ligand has been successful in the treatment of iron overload and thus has led to their testing for the control of aluminium levels. Definitive results are awaited on the two key questions -do such ligands reduce aluminium levels at key sites by excretion and not by moving the aluminium elsewhere in the body and is there a signif- icant improvement in memory and other deteriorated mental powers? co -I a2co2-CH(0H)I II 8 9 0 0 I R' 1I R = Me I2 R=Et 13 There may well be a purely inorganic means of controlling alu- minium levels as suggested by Birchall who has been particularly interested in the role of silicon in the body. There are strong indica- tions that silicon may be an essential trace element and it could be that silicon is not essential in itself but that it may through the inter- mediacy of hydroxyaluminosilicates,play a part in the control of aluminium levels .I3 4 Gallium and Indium These elements can conveniently be considered together. They have very similar chemistry; neither has a natural metabolic role. A few simple gallium compounds e.g. nitrate and citrate have been tested for anti-tumour activity but showed no significant advantages over current treatments. Indium nitrate was similarly tested but only briefly since it is a more toxic compound than its gallium analogue. Gallium nitrate under the name 'Ganite' ,has been used in Japan for the treatment of cancer-associated hypercalczmia. The great impor- tance of these elements in relation to inorganic species in medicine is that both have isotopes useful for diagnosis or therapy. The various isotopes of gallium and indium which have at one time or another been suggested investigated or used for radio- pharmaceutical purposes are listed in Table 1. Of these isotopes Table 1 Isotopes of gallium and indium of actual or potential radiopharmacological importance; the most importance are in bold type" Decay Decay Isotope mode Half-life Isotope mode Half-life MGa EC; p+ 9.5 hours EC; /3+ 1.15 hours 67Ga EC; y 78.1 hours illln EC; y 67.4 hours 68Ga EC; p+; y 1.13 hours 113mIn IT; y 1.66 hours '2Ga p-; y 14.1 hours 114mIn IT; EC; y 49.5 days (8 Abbreviations EC = electron capture; IT = isomeric transition. CHEMICAL SOCIETY REVIEWS 1996 67Ga,68Ga and "'In are the most important. 67Ga andIIIIn have been widely used for many years; 68Ga has recently become much more readily available from 68Ge/68Ga generators. This may be expected to result in much increased use of this isotope for positron emission tomography (PET). 67Ga is used in the related y-ray imaging tech- nique of single photon emission computed tomography (SPECT-PET SPECT and other terms relevant to the use of inor- ganic complexes in imaging and therapy are succinctly explained in ref. 1). 67Ga and 68Ga are both used for scanning; both have t12suit-able for medical usage -long enough to dose the patient and to carry out the appropriate scans but short enough to avoid problems which may arise from long-term irradiation. 67Ga imaging has in recent years been used for imaging soft tissue tumours and inflammatory lesions. The half-life of 67Ga is particularly appropriate for intro- duction via monoclonal antibody techniques. %mTc very widely used for imaging a variety of parts of the body has a half-life which is rather too short (6.05 h) for introduction this way. IIIIn is used for evaluation of cerebral spinal fluid pathways for biliary reflux studies for imaging intra-abdominal abscesses for monitoring thrombosis and for early warning of transplant rejection. I 14mInhas been used in the treatment of leukaemia. Of the other much less important isotopes in Table 1 I13"In has recently found use as a tracer (vide infra) while 72Ga was used in the early 1950s to look for bone tumours. It was found to be somewhat less unsuitable than 45Ca and 47Ca but was soon superseded by isotopes better suited to this purpose. OH I4 15 16 67Ga was first administered as a lactate (14) complex. The radionuclides of gallium and indium have in recent years usually been administered complexed with citrate (6) oxine (8-hydroxy- quinolinato 15) or dtpa (16; the abbreviation is derived from its long-established trivial name diethylenetriaminepentaacetate).The citrate species are of relatively low stability and readily give up Ga3+or In3+ to transferrin as does the indium chloride which has occasionally been administered. The radionuclides then localise in sites where iron tends to concentrate -e.g. the liver spleen and kidneys and in soft tissue tumours and lesions. The dtpa complexes are somewhat more stable and are transported around the body with a markedly different distribution pattern. The oxine complexes are water-insoluble and thus more difficult to administer though their lipophilicity is an advantage in crossing cell membranes and thus assists their introduction into cells. They are used to label white and red blood cells which then take the radionuclide to the required site. Although it is relatively difficult to get the radionuclide to the right place to begin with once there it tends to stay -hydrophobicity and low solubility in aqueous media help to minimise washout. At the other end of the HLB range l~~In-labelling of haemoglobin is pos- sible through the intermediacy of the water-soluble derivative of rneso-protoporphyrin IX. Another method of administration is through attachment to mon- oclonal antibodies. Dtpa has recently been used as the chelating moiety to link I1IIn3+ to monoclonal antibodies to improve locali- sation at the required site. Species of this type have been approved for myocardial imaging and are being developed for possible appli- cation in detecting colorectal tumours. The problem now is that the I * 'In-containing species has a high relative molecular mass and slowly dissociates so that the IIIIn accumulates in the liver. Efforts are therefore currently being made to see if a fragment of the com- plete monoclonal antibody or even simply a short polypeptide chain is sufficient to direct the *!'Into the required site and to bind the indium into a more powerful chelating moiety from which it will dissociate at a much slower rate. Most of the vehicles for administration mentioned in the preced- ing paragraphs have some disadvantages so other types of com- plexes are at various stages of testing to find a better method for the MAN AND THE ELEMENTS OF GROUPS 3 AND 13 -J BURGESS introduction of gallium and indium radiopharmaceuticals into the body and for their specific localisation in the tissue or organ to be imaged Several approaches are outlined in the following para- graphs Gallium and indium complexes of pyranones and pyridinones (see formulae 11-13 in the section on aluminium above) have high stabilities (log p3for the complexes of 13 with R=R'=Me are 35 8 and 3 1 7 respectively) and offer the possibility of oral administra- tion Their partition coefficients are rather smaller than ideal but they can be tailored e g by having an X-aryl-N-substituent (17) This increases their Iipophilicity which reduces washout with its concomitant over-rapid elimination Considerably increased lipophilicity can also render them suitable for myocardial monitor- ing Further there is the possibility of including a peptide or ester linkage in the N-substituent which may undergo hydrolysis at the site of action to give a derivative with a different HLB which may have more difficulty crossing membranes and may therefore stay at the required site for a longer period The main disadvantage of this class of ligand is the presence of phenolate donor groups since even in mildly acid media these donor centres may become protonated and thus separated from the metal ion b,.rs-7 b. 18 17 Ligands with an N2S2donor set for example with the framework 18,were shown several years ago to be useful vehicles for the intro- duction of 99mTc into the body They and N,S analogues are now being assessed for the introduction of Ga and In In particular a lipophilic tetraethyl derivative has been proposed as a ligand for llIln to provide a new myocardial imaging agent Interestingly 113mInproved a useful tracer in assessing this diagnostic approach Increased stability may be introduced in the form of the macrocyclic effect familiar to inorganic solution chemists in the guise of the greater stability of complexes of e g the cyclic ligand 19 than of its open chain (garland) analogue H2NCH,CH2 NHCH2CH2NHCH,CH2NH2 Obviously it is even better to add further chelate rings and have the one ligand molecule encapsulat- ing the metal ion and occupying all coordination positions An example of a ligand of this type is dota 20 whose carboxylate pendant arms provide extra chelating units Dota with its eight potential donor sites is a powerful and popular Iigand but has the slight disadvantage that its complex with a 3+ ion will have a nett negative charge which will militate somewhat against passage across biological membranes One way around this problem is to replace carboxylate -C02 ,by an alcohol function -OH another is to go from the tetraaza-ring of dota to the triaza ring of nota 21 l4 This ligand has the added advantage of being hexadentate and thus neatly occupying the six coordination sites of the Ga3+ or In3+ as well as giving an uncharged complex with 3+ cations Fine tuning and optimisation of nota have recently led to the synthesis and study of the derivative 22 in which the simple carboxylate pendant arms of the parent ligand have been replaced by derivatised phenoxides (the -0 donors having a particularly high affinity for M3+) with methyl substituents modifying the HLB of the complex More recently phosphinate groups have been incorporated into the pendant arms These ligands form considerably more stable complexes then dtpa (vide supra),though it should be mentioned that stability and inertness of polyaminocarboxylate complexes of the dtpa and edta type have been considerably increased by moving to anions of 23 Tailoring of properties is possible here for example by the use of R=CH or CH,OH for fine tuning of the HLB The anionic ligand derived from the phenolic aminocarboxylic acid 24 is claimed to be unusually selective for gallium vs indium with log K values of 37 7 and 27 9 for its Ga3+ and In3+ complexes respec tively natural carriers such as transferrin exhibit very little differ- entiation with a difference of only 10-fold between the stability constants for its Ga3+ and In3+ complexes 19 20 "0""Me Ho)+JRMe 23 &OH ""6 24 The majority of the applications described in this section have been in diagnosis Effective radioimmunotherapy usually needs a more aggressive radionuclide such as 90Y (vide infiu) Indeed 1111n/90Yconstitute a diagnosis/therapy pair as analogous corn plexes with a given ligand have similar tumour affinities An alternative mode of tumour treatment is photodynamic therapy for which sulfonated gallium phthalocyanines containing tert-butyl groups have been evaluated as photosensitisers The lipophilicity of the tert-butyl groups is balanced by the hydrophilicity of the sulfonate groups to provide an HLB suitable for cell uptake 5 Thallium Thallium salts are poisonous -TI+ is absorbed efficiently from the human gastrointestinal tract About 1 g is sufficient to cause death which usually occurs from heart failure 3 to 4 weeks after inges- tion Earlier symptoms and effects include nausea insomnia mental confusion and swelling of the feet and legs However the best-known effect of thallium poisoning is that on the hair of the victim At first the hair can be pulled out without pain after some days it falls out anyway These phenomena have been detailed in apart from the normal sources of reference Agatha Christie's 'The Pale Horse' whose plot centres on thallium poisoning Is l6 The toxicity of thallium compounds was recognised soon after the dis covery of this element but was for a long time greatly underesti- mated There was much use of thallium compounds as depilatory agents for example in preparation for treatment of ringworm while thallium(1) acetate was at one time used for the treatment of tuberculosis By 1934 the number of fatalities recorded as a result of 692 cases of medical (mis)use had reached 31 However the main cause of thallium poisoning was the use of thallium sulfate and other salts in insecticides and rodenticides The thallium salts were often incorporated into syrups which not surprisingly formed a temptation to children At least the fact that death occurs many days after ingestion allows some scope for chelation therapy though thallium does distribute itself around the body rapidly after absorption Dimercaprol (BAL 25) is an effective antidote potas slum chloride also helps as it promotes excretion of thallium through the kidneys CHZS H I CHSH I CH-OH ,OIT1 (decay mode EC,y-emitter of relatively low toxicity t =72 hours) has been much used for tumour and especially myocardial imaging However it has now been superseded in this latter role by wmTc ,OITI is generally administered simply as TlCl ,OIT1 works as good probe as a result of the fairly close similari- ties between T1+ and K+ with the distribution of 210T1 reflecting blood flow and tissue viability It accumulates in the viable myocardium via the Na-K-ATPase pump Thus the 201Tl+ localises naturally where it is wanted without the need for specific tailoring of complexes It has to be administered to the patient under stress either from exercise or drug-induced (e g by dipyri- damolelpersantin) as a resting study only shows up long-standing scars and damage not current problems Images are taken imme- diately after stressing and again 3 to 4 hours later with the dif ferences pinpointing the problem sites Sometimes comparison of scans for two different isotopes can provide valuable information An example of this is provided by the simultaneous use of 210Tl(as TlCl) and 99mTc (as TcO ) for parathyroid imaging Subtraction of the 99mTc image from the ,O1TI image reveals the site of an adenoma since uptake of the two ele- ments differs appropriately 6 Scandium Scandium is a widely distributed but extremely thinly spread element whose aqueous solution chemistry has been somewhat neglected and is still rather incompletely characterised Thus despite confident assertions that the hydration number of Sc3+ in aqueous solution is six this has yet to be established Indeed there has been no X-ray crystal structure determination of a simple hydrated Sc3+ salt containing octahedral SC(OH,),~+ or SC(OH,)~?+(which would be uniquely interesting) or Sc(OHJg3+ [cf Y(OH,),?+ below] for that matter * There is not even a stable scandium alum Cs,SO Sc,(SO,) 24H20has been reported but is only stable below 0 "C At present the nearest thing to interaction of scandium compounds with humans is their use in some forms of lighting used in public transport vehicles and department stores However scandium salts and their solutions have been reported to be bacteriostatic or bacte- riocidal ,probably through their affecting iron-enterobactin interac- tions There have been hints of their use for specific delivery of bacteriocides in specialised applications but cost (scandium salts start at around &20/$30per gram) precludes widespread use Two radioisotopes are available from generators viz '3c (EC p+,y I =3 9 hours) and 47Sc(p ,y t =3 4 days) These might one day prove to have some radiopharmocological usefulness 46Sc(also a p emitter) in the form of citrate has also been briefly assessed for tumour imaging in rats However it seems to have no particular advantages to offset the disadvantage of its long half-life of 84 days 7 Yttrium Yttrium is another rare element which has no role in the human body Information on the solution chemistry of Y3+is again limited but at least in contrast to Sc3+(aq ) there are well-characterised aqua-yttrium(I11) ions as in [Y(OH,),I(Tc,CI,] "Y(p-,t =64 hours) has been suggested for relatively minor applications such as the treatment of arthritic joints and more sub- stantially for the in situ inadiation of malignant growths in the pleural and peritoneal cavities is potentially valuable for radioimmunotherapy but is one of the very few non actinide isotopes to come into the 'very high toxicity' category It must therefore be administered in a very stable and inert complex which can be deliv ered cleanly and accurately to the required site of action If *Since this Review was written X ray diffraction studies have established Sc(OH,)," in hydrated scandium triflate CHEMICAL SOCIETY REVIEWS 1996 nt / days -Figure 3 Inertness and lability of yttnum(i1r) complexes becomes free as Y3+(aq ) it is liable to settle in bone tissue substi- tuting for Ca2+ and cause serious problems T,O incorporated in rods or in polythene sheet has been used for implantation in the pitu- itary In this form the is unlikely to escape' The main problem associated with its use in the pleural and peritoneal cavities is that of obtaining a stable colloidal preparation though yttrium silicate col- loids have been used with some success The problem with other aspects of its use is one of ligand design -both for maximum stabil- ity and for maximum inertness so that any very slight dissociation that does take place occurs extremely slowly As the characteristic timescale for substitution by simple monodentate ligands at the very labile Y3+(aq ) is of the order of a microsecond tojudge from the rate constant established for water exchange at Y3+(aq ) the inhibition required is enormous However it can be achieved by suitable ligand design and modification as illustrated in Fig 3 The familiar multi- dentate polyaminocarboxylate ligand dtpa 16 reduces dissociation rates dramatically though not quite sufficiently The half-life for release of Y3 from its complex with the n-butyl amide derivative of dtpa is more than two days but the incorporation of a 4-nitrobenzyl substituent in the dtpa skeleton leads to an increase in dissociation half-life to a matter of weeks Going from garland ligands to pendant- arm macrocycles gives a further large increase in inertness as shown by the plot for the 4-nitrobenzyl-substituteddota complex in Fig 3 By this stage the half-life for dissociation more than 200 days at 37 "C,is considerably longer than the half-life of the wY so there is neg- ligible risk of harm arising from leakage of 9oY from the administered complex into other parts of the body 8 Lanthanum and the Lanthanides None of the lanthanides is an essential element for humans La3+ has been used from time to time as a Ca2+ substitute or antagonist to probe mechanisms of calcium action in the body The fluores- cence of Eu3+complexes has similarly found application in moni-toring the functions of such ions as Ca*+ There has been some investigation of the possibility of the therapeutic use of lanthanum salts in the treatment of disorders of calcium metabolism or action There have also been experiments mainly in Japan on the replace- ment or reinforcement of calcium in teeth by lanthanum Prolonged treatment -La3+'/edta is better than La3+(aq ) -gives a lanthanum phosphate coating on the enamel which is claimed to provide sig- nificant protection against dental caries In these and other applica- tions it has to be borne in mind that several of the lanthanides are fairly poisonous so for the administration of large or repeated doses it is advisable to have the required ion firmly complexed in a chelate which is both very stable and very inert In the latter respect it should be mentioned that rate constants for substitution of simple ligands at lanthanide(3+) ions are high in the region of 107-109 dm3 mol I s I under normal experimental conditions Purely chemotherapeutic applications are thus very rare but the magnetic and radiochemical properties of several of the lanthanides have important and wide-ranging applications in the diagnosis and treatment of disease Probably the most important application IS in MAN AND THE ELEMENTS OF GROUPS 3 AND 13 -J BURGESS magnetic resonance imaging (MRI) which concerns mainly gadolinium This topic will be covered first followed by a selection of uses of lanthanide radioisotopes MRI in essence involves running the NMR spectrum of the appropriate part of the human body The aim is to detect and locate problems by picking up differences between the signals from protons in healthy and in diseased tissue This is achieved through comparing relaxation times (TI,T,) for water protons in the two environments As the differences can be very small the effects are amplified by the addition of aFpropriate paragmagnetic ions -so-called contrast enhancement agents which increase signal intensity by enhancing relaxation These ions operate by affecting the relax- ation time of water molecules in their immediate vicinity in partic-ular water molecules which enter their normally very labile primary coordination spheres The best ions are the symmetrical half-filled shell d5 or fions Mn2+ Fe3+ or Gd3+ Compounds of manganese and iron were used first gadolinium compounds were then found to be better Dtpa (16) complexes have generally been used for the introduction of gadolinium indeed [ Gd(dtpa)12 ,as its N-methyl- glucamine (26) salt (‘Gadopentetate/Magnevist’) was the first complex approved for use in MRI for tumour detection about a decade ago It has proved valuable in diagnosis of problems affect- ing parts of the central nervous system The dtpa complex is stable but is rather labile Now new ligands many very similar to those discussed above in connection with yttrium are being designed and assessed for the required kinetic inertness of the chelating ligand The complex with dota (20) has for example been found to have the requisite stability and inertness it1 vivu Complexes need to retain at least one water ligand in the coordination shell of the Gd3+ to permit rapid exchange between free and coordinated water so that all the water in the part of the system accessible to the Gd3+ feels its paramagnetism at first hand It is also desirable that the com- plexes are non-ionic as such complexes are effective in smaller amounts and cause less pain on injection Two recently-approved complexes are ‘Gadodiamide/Omniscan’ and ‘Gadoteridol/ ProHance,’ whose ligands are an amide derivative (27) of dtpa (16) and a hydroxypropyl derivative (28) of dota (20) respectively Phosphinate analogues are also promising especially as they can be tailored by variation of no fewer than four groups viz R,R’,R” and X in 29 Tailoring of ligands with incorporation of lipophilic groups such as -CH20CH2C,H5 and -CH20C,H4CH2CH is leading to agents which may localise in the liver gall bladder and bile duct Gd3+,with its half-filled shell fconfiguration and long elec- tronic relaxation time causes so much line-broadening that con- ventional IH NMR spectra are not obtainable The other lanthanides with their much shorter electronic relaxation times generally give large shifts without significant line broadening -hence their wide use especially in organic chemistry as shift reagents This complementary behaviour has also proved to have some medical value Several Dy’+ complexes including those with tripolyphosphate and with ttha (triethylenetetraminehexaacetate 30) have been assessed for use as shift reagents in in VIVONMR but a more promising development is the investigation of Dy3+ com- plexes as magnetic susceptibility contrast agents l9 Dysprosium is the element of choice here as it has the highest magnetic moment of the lanthanides Its complex with 27 (‘Sprodiamide’) is under development as a heart and brain imaging agent Thulium com- plexes are also being assessed for imaging Tm3+ engenders shifts in the opposite direction from Dy3+ Many radioisotopes of the lanthanides are available from nuclear reactors (e g 15?Gd 169Yb) from generators (eg 134La I40Pr) or from cyclotrons (eg 157Dy 167Tm I6’Yb) *O Several lanthanide radioisotopes are of actual or potential value in diagnosis or therapy A selection of these2 is given in Table 2 p -emitting radionuclides of samarium dysprosium holmium and ytterbium have proved the most useful Samarium in the guise of I5?Sm has considerable clinical potential,2‘ it concentrates in bone tissue like 89Sr and 32P2 Phosphonate and aminocarboxylate complexes of I5?Sm have been used in palliative treatment of painful bone metastases in terminally ill patients As over 50% of patients with lung breast or prostate cancer develop bone metastases it is important to be able to allevi-ate the consequent pain Complexes of the smaller aminocarboxy- 91 CHzOH HO \N N/ R R Y =-CH2CONHCHZCH,X-CH,CONRR ‘CH OH 28 -CHZP/= 0 ‘R 29 late ligands are insufficiently well localised but the complex with the anion of the tetrakis-phosphonic acid analogue of edta edtmp (31) is very promising The relatively short half-life of I5?Sm permits treatment to be carried out in several discrete stages thereby spreading the radiation burden on the patient Other radionuclides again generally administered as aminocarboxylate complexes though sometimes in colloidal form as coprecipitates with iron(r1r) hydroxide have found employment in bone imaging in the treatment of arthriticjoints in dental radiography. and in infil-tration treatment of the lymph node To connect with applications mentioned earlier in this section. I5%d has been used to track the gadolinium administered for MRI purposes To link back to the section on boron. several lanthanide isotopes have been suggested for use in neutron capture therapy including IT7Gd (which has a high nuclear cross-section) 155Gd I4”Sm and I5lEu However development of this approach has been severely hindered by the dif- ficulty of localising a large enough number of atoms at the tumour cell site 9 Actinides Their chemistry applications and interactions with man are domi- nated by their radioactivity There are few favourable features apart from radiotherapy for cancer though a century ago or so uranium nitrate was used in treating diabetes The main chemical problem is efficient detoxification as such isotopes as 227 Ac(a naturally occur- ring isotope identified in a uranium mineral as long ago as 1899) 228Th,231Pa,*33U,*79Pu,241Am,and242Cm are all a yemitters and come into the ‘very high toxicity’ category It is essential to remove them reasonably coon after exposure to prevent them being incor- porated into bone There is particular concern about plutonium ds it is particularly toxic and is also now widely distributed -there are probably several hundred tons around the world There are also in existence considerable quantities of some of the other actinides perhaps about 10 tons each of neptunium and americium The wide- spread investigation and application of a number of the actinide\ with the consequent certainty of accidents encouraged the early Table 2 Lanthanide isotopes used in diagnosis and therapy Decay Isotope mode Half life Samarium lS3Sm P 2 47 hours Thulium Terbium '6'Tb P 6 9 days Dy sprosi um I Dy EC P+ 6 3 hours Ytterbium I S7Dy EC Y 8 1 hours Lutetium I6sDy P 2 3 hours Erbium IeyEr P 9 4 days 1-Y I7'Er P 7 5 hours 3Y establishment of detoxification agents and procedures As long ago as 198 1 some thirty chelating agents had been assessed for removal of plutonium many more have been developed since then 22 Transferrin binds plutonium and several other actinides strongly Despite this detoxification simply with (bi)carbonate is claimed to be quite effective for e g removal of ingested uranium However treatment with appropriate polydentate organic ligands is much more effective Use of diethylenetriaminepentaacetate,16 in the form of Na,Zn(dtpa) is the preferred treatment for removing transuranium elements after exposure as a result of industrial inci- dents Dtpa has the advantage over edta of being potentially octadentate It is thus more selective for the actinide cations which favour coordination numbers higher than six in comparison with ions such as Mg2+ or Zn2+ which it is undesirable to remove Triethylenetetraminehexaacetate(ttha 30) next up the series from dtpa (16) and potentially decadentate has been developed as a potential oral decorporation agent for transuranic elements Polyaminocarboxylates containing long hydrocarbon chains e g C12H25or C2,H4 derivatives of triethylenetetraminepentaacetate are effective for removal of plutonium and americium even of aged deposits They can be administered orally presumably the hydro- carbon chains assist in the transport of ligands and complexes across the various membranes which have to be traversed 23 Desferrioxamine (lo),so important for the removal of excess body burdens of aluminium (vide supra) and iron is also effective for plutonium mobilisation Its set of six hard donor atoms are appro- priate for the removal of a hard ion such as Pu4+ but it lacks the possibility of selectivity versus the essential 2+ cations which is a feature of the high denticity polyaminocarboxylates New ligands are still being developed to maximise effectiveness stability inertness and selectivity 24 Thus for example catechol (1,2-dihydroxybenzene) and pyridinone (13) units are being grafted onto appropriate backbones to maximise the desirable properties A good example is provided by the modification of desferrioxamine 10 with the moieties shown as 32 to give particularly promising chelating agents 25 R R = H C02H CONHMC 32 10 Conclusions Compounds and complexes of the Group 3 and Group 13 elements can interact in many ways favourable and unfavourable with human metabolism A selection of examples has appeared in this CHEMICAL SOCIETY REVIEWS 1996 Decay Isotope mode Half life Ih7Tm EC y 9 2 days I 7(Tm EC,P y 134days IhyYb EC y 32 days InLu P 2 6 7 days short review the interested reader will find a much greater depth of treatment and a greater variety and range of examples in the general review articles cited and in several of the multi-volume treatises published in recent years 10 References 1 D Parker Chem Br ,1994,818 2 R C Hider and A D Hall Progr Med Chem 1991 28,41 3 S Junsson,D Berning Wei Jia,and Dangshe Ma,Chem Rev 1993,93 1137 4 J Burgess Transition Met Chem ,1993,18,439 5 See e g ,R J P Williams Coord Chem Rev 1987,79,175 6 P J Sadler Adv lnorg Chem 1991,36,1 7 M F Hawthorne Angew Chem Int Ed Engl 1993,32,950 8 J R Hariharan. I M Wyzlic and A H Soloway Polyhedron 1995,14. 823 9 G R Newkome,C N Moorefield,J M Ke1th.G R BakerandG H Escamilla Angew Chem Int Ed Engl 1994,33,666 10 R Doll,Age Ageing 1993,22,138 11 P 0 Ganrot Environ Health Persp 1986 65 363 H Sigel and A Sigel (ed ),Metal Ions in Biological System5 Vol 24 Aluminium and its Riile in Biology Marcel Dekker New York 1988 R C Massey (ed ) Aluminium in Food and the Environment RSC London 1989 TE Lewis (ed ) Environmental Chemutry and Toxicology of Aluminium (ACS Symposium New Orleans 1987) Lewis Publishers Chelsea Michigan D J Chadwick and J Whelan (ed ) Aluminium in Biology and Medicine (Ciba Foundation Symposium 169 19911 Wiley Chichester 1992,R A Yokel,J Toxicol Env Health 1994,41 131 12 T Schamper J Chem Educ 1993,70,242 13 J D Birchall Chem Br 1990,141 Chem Soc Rev 1995,351 14 C J Broan J P L Cox A S Craig R Kataky D Parker A Harrison A M Randall and G Ferguson J Chem Soc Perkin Trans 2 199I 87 15 A Christie The Pale Horse Collins 1961 see. e g ch 22 16 D A Labianca J Chem Educ 1990,67,1019 17 R B Lauffer Chem Rev 1987.87,901 18 K Kumar and M F Tweedle Pure Appl Chem 1993 65 515 D Parker K Pulukkody T J Norman A Harrison L Roule and C Walker,J Chem Soc Chem Commun ,1992,1441 19 C A Chang Eur J Solid State Chem ,1991,28,237 A D Watson J Alloy5 Cpd ,1994,207/8,14 20 G V S Rayudu,Sem Nucl Med 1990,20,100 21 A Singh,R A HolmesandM Farhangi,J Nucl Med 1989,30,1814 P Sorby J Lab Cpd Rudiopharm ,1989,26,492. R A Holmes Sem Nucl Med ,1992,22,41 22 K N Raymond and W L Smith Struct Bonding 198 1,43,159 R A Bulman Struct Bonding 1987,67,91 23 F W Bruenger G Kuswik Rabiega and S C Miller J Med Chem 1992,35,112 24 F L Weitl and K N Raymond. J A m Chem Soc ,1980,102,2289.K N Raymond,G E Freemanand M J Kappe1,Inorg Chim Acta. 1984 94,193 25 Zhiguo Hou D W Whisenhunt. Jide Xu and K N Raymond J Am Chem Soc 1994,116,840

ISSN:0306-0012    

DOI:10.1039/CS9962500085

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

New Developments in Making Compounds and Materials by CondensingGaseous High-temperature Species at Atmospheric or Low Pressure Peter L. Timms School of Chemistry, University of Bristol, Bristol BS8 1TS, UK 1 Introduction It is well known that the use of high temperatures permits the forma- tion of a wide range of gaseous atomic and molecular chemical species which are often unobtainable by other means. Many of these species such as the free atoms of the metals or low oxidation state oxides and halides of Group 13 and 14 elements have the potential both thermodynamically and kinetically to be chemically very reac- tive. As soon as the temperature is lowered the species become metastable and tend to recombine to form the less reactive gaseous or solid elements or compounds from which they were made.This recombination can be very rapid and often special conditions have to be established to make use of the species while they are still reactive. Deriving useful chemistry from such species began to be a lively area of research in the 1960s following work by Skell on carbon vapour chemistry.’ There was a frenzy of activity in the 1970s fol-lowing the first work on transition metal atom chemistry at Bristol?-5 Developments have continued to the present but undoubtedly the biggest surprise in the 1980s and early 1990s again involved carbon in the synthesis of fullerenes, the remarkable mol- ecular allotropic forms of that element.6 Section 2 below, the first main part of this review, surveys some experimental considerations relevant to using high-temperature species in synthesis, while the other three parts are concerned with work over the past few years at Bristol in exploiting the chemistry of neutral atoms and small molecules: Section 3 on reactions of metal atoms with hexafluorobenzene, Section 4 on microfibres formed by condensing silicon monoxide and Section 5 on some new silylene chemistry.2 Experimental Conditions High temperatures can be attained by many different methods including electrical resistance heating, flame heating, electron bom- bardment, focussed light beams, electric arcs and plasmas, and a detailed discussion of them is beyond the scope of this review. However, one point is worth making. If gaseous species are pro- duced in thermal equilibrium with a heated solid or liquid, it is usually fairly easy to measure the temperature at which the species After graduating from the University of Oxford in 1959, Peter 7imms worked for Borax Consolidated and developed an interest in high tem- perature chemistry.This theme then shaped his D.Phi1. research at Oxford (with CS.G. Phillips), his postdoctoral research at Rice University (with J. L. Margave), and his research as Assistant Professor at Berkeley. He returned from the USA as a Ramsay Fellow at Bristol, later becoming Lecturer and Reader in Inorganic Chemistry. He was awarded a Corday Morgan Medal for work on synthetic reactions of boron, silicon and transition metal high temperature species.His current research, in addition to the topics of the review, include work relevant to the electronics industry on silica films and on destructionof toxic waste gases. 93 are formed and to predict the position of chemical equilibria. When species are formed under more dynamic conditions in flames, by laser pulses, in electric arcs or in plasmas or discharges, it is more difficult to define the temperature or know the position of chemical equilibria. In addition, high-voltage electrical discharges and plasmas will contain energetic electrons and ions. So the formation of species under dynamic conditions can create problems for the synthetic chemist which should not be underestimated. That said, dynamic methods may sometimes be the only way of generating species under particular conditions and they are being used suc- cessfully.To give three examples: working in the vacuum chamber of a Fourier-transform ion-cyclotron resonance (FT-ICR) mass spectrometer allows ion-molecule reactions involving metal atoms or ions to be studied to great advantage: rapid cooling-by-expan- sion of the vapour plumes created by a pulsed laser impact on carbon and metals allows many novel gaseous clusters to be made? and arc and plasma evaporation techniques are responsible for much of the spate of fascinating new work on condensing ‘nan- otubes’ of carbon, boron and other materials.9 Whatever method of generating the species is used, pressure is of crucial importance as it affects the ease of forming, condensing or reacting species.Figure 1 summarises how different pressure regions permit different types of chemistry involving high-temper- ature species to be carried out. Considering the situation at atmos- pheric pressure (lo5 Pa) first; if a metal is heated in a system containing an inert gas at 105 Pa, then rapid evaporation of the metal to form atoms will only be observed at a temperature when its vapour pressure is close to lo5 Pa. The mean free path of the atoms formed will be very short so that as soon as the temperature of the gas falls, multi-body collisions of the gaseous atoms leading to nucleation of solid or liquid metal will occur. If an attempt is made to react the gaseous atoms with a gaseous substance, the substance and any reaction products will become heated by conduction from the surrounding hot gas and they will only survive if they are ther- mally stable.These effects impose severe limitations on the subtlety of the chemistry that can be carried out. Nevertheless, as it is easier to work at atmospheric pressure than in any other pressure region, a huge amount of work has been carried out on atmospheric-pres- sure processes involving high-temperature species generated from furnaces under thermal equilibrium conditions or under non-equi- librium conditions in flames, arcs or plasmas. In some cases inter- esting aggregation effects occur after nucleation of solids which can lead to useful materials (see Figure 1 and Section 4) and in other cases the high activation energy for nucleation permits alternative gas-phase processes to occur in preference: e.g.in a high-tempera- ture synthesis of ethyne from methane in which the gas phase process 2CH,-.C2H2+3H, occurs much faster at 2000 “C than the thermodynamically favoured CH,+C(solid) +2H,. Reducing the pressure below atmospheric has important effects on the position of chemical equilibria, on the frequency of collisions of gas molecules with each other and on the mode of condensation. For example, pressures just below atmospheric are being used in arc syntheses of the ‘nanotubes’ mentioned above.9 Slightly lower pressures have proved to be the best for forming fullerenes. When carbon is evaporated under helium at 15 000 Pa the condensation of carbon vapour which occurs gives C6,, C,, and other cage forms of carbon.These molecules were first synthesised by pulsed laser evaporation of carbon in helium but it is now recognised that they can be formed by thermal or arc evaporation of carbon or even by controlled combustion of aromatic molecule^.^ The success of this chemistry owes much to the thermal stability of the carbon-carbon CHEMICAL SOCIETY REVIEWS, 1996 plasma generationof swcies carbon black, SiO micrdibres _-_-_-_--___ __-_--_ low temperature reactions of species with compounds in the condensed phase I I I I I I I I I I I I 16+ 104 10' 102 101 100 101 10-2 10-3 10" pressure Pa 10-6 1 atm. Figure 1 The pressure used for different types of synthesis involving gaseous high-temperature species.bonds in the gaseous intermediates and in the final condensed product as molecules can only exist at high temperatures if they contain strong chemical bonds. The pressure range of 10-1000 Pa has long found uses in flash vacuum pyrolysis of organic compounds, in which a compound is passed quickly through a heated tube and yields highly reactive fragments which can be used in synthesis. In this pressure range, there is still fairly efficient heat transfer from the hot walls of the tube to the gas so that reactive species are generated by energetic intermolecular collisions and then pass out of the hot zone rapidly because of the low pressure.*O When high-temperature species are generated at pressures below 0.1 Pa, individual atoms or molecules will travel many centimetres before their progress is seriously impeded by collisions with sur- rounding gas.Nucleation of solids from gases takes place slowly and may not be observed in ordinary vacuum systems in which col- lisions of species with the walls occur long before gas-phase nucle- ation. This makes it possible, for example, to transfer metal atoms efficiently from a hot zone where they are made through the vacuum to a cooler surface where they can be condensed; this is the prin- ciple behind the industrial method of vacuum metallization. These low pressures will also further facilitate formation of high-temper- ature molecular species from solid-solid or gas-solid equilibrium reactions.If the condensing surface is made sufficiently cold, it is possible to condense on it not just the gaseous high-temperature species but also the vapour of any compound which may react with the high-temperature species. Contact between the hot source of gaseous species and the added vapour can be made negligible so that even the most delicate molecules can be reacted with high-tem-perature species under these conditions. Cocondensation of high- temperature species with compounds on surfaces cooled in the range -100 to -269 "C is a valuable synthetic method which has been applied ~idely.~ The very lowest temperatures are required if permanent gases like CO or N, are to be reacted with metal atoms and other high-temperature species and such work has usually been carried out on a milligram scale using in situ spectroscopic charac- terisation of products.ll A great deal of work has used liquid nitro- gen-cooled surfaces at -196 "C (as we have used for the work on hexafluorobenzene described in Section 3) and has been conducted on a scale ranging from milligrams to tens of grams of reactants.Species may also be condensed under vacuum into liquids of low vapour pressure. Commonly the liquid is a solution of a reactive compound in an inert solvent, cooled until it has a low vapour pres- sure (as we have used for silylene chemistry described in Section 5) but in some cases the pure liquid is the reactant.', 3 Reactions of Atoms with Highly Fluorinated Arene Ligands Since the first work on transition metal atom chemistry at pressures <O.1 Pa: one of the most successful uses of transition metal atoms in chemical synthesis has been the reaction of atoms with arenes to form bisarenemetal 'sandwich' compounds and related products with coordinated arenes. Hundreds of compounds have been made this way which are inaccessible or difficult of access by the con- ventional methods of organometallic chemistry involving reduction of metal compounds in the presence of arene ligands or ligand dis- .531placement reactions 3-l6 The sandwich compounds have been made with Group 3-8 elements with the largest number in Group 6 where the products are 18-electron complexes (six electrons from the metal and six n-electrons from each arene ligand) with greater stability than for comparable products from other Groups.However, by using bulky arene ligands like 1,35-tri(tert-buty1)benzene to provide steric protection, surprisingly stable com- plexes have been made of many early transition metals including 15-electron complexes of some of the lanthanide We have been re-exploring this area using hexafluorobenzene and other highly fluorinated aromatic compounds as ligands. Relative to benzene or alkyl substituted benzenes, the electron with- drawing fluorine atoms reduce the availability of the aromatic melectrons for u-bonding to metals but enhance r-acceptor prop- erties. Early work on condensing chromium atoms with hexafluo- robenzene at -196 "C showed that no product that was stable at room temperature was formed but if benzene or trifluorophosphine was mixed with the hexafluorobenzene the stable compounds [Cr(C,F,)(PF,),] or [Cr(C6H6)(C6F6)] could be isolated.19 More recently, we have found that condensing molybdenum or tungsten atoms with hexafluorobenzene on a liquid nitrogen-cooled surface gives low yields (ca.2% based on the metal atoms used) of [M(C,F,),] (M=Mo or W) as quite volatile, air-stable, brightly coloured crystalline solids.2O The compounds were best formed in the apparatus of Figure 2 in which the tip of an uncooled rod of Mo or W was heated by a focussed electron beam allowing rapid evap- oration from a pendant drop of molten metal held by surface tension.The bisarene complexes are the only molecular products of the atom reactions apart from a trace of decafluorobiphenyl ,so they are easily isolated on warming the condensate by first pumping away the excess hexafluorobenzene then pumping them or dissolv- ing them out of the involatile/insoluble metal-rich residue on the condensation surface. Their air stability contrasts sharply with the behaviour of bisarenemol ybdenum or tungsten compounds with less electron withdrawing ligands -some bisarenemolybdenum .micrometer ligand vapour &.*'feed outlet for solution ''---'*H #-' of products -L volatile product pump-outIc. electrostaticallyevaporating focussed electron pendant drop ---------gun with heated of metal [ W filament at ground potential liquid .__-__ nitrogen .250 mm diameter II II -I stainlesssteelcondensate .of metal vacuum vessel atoms + ligand Figure 2 Apparatus used for efficient evaporation of molybdenum or tung-L.. ,.l,.,.*..--L--L....A ---. ,.+,.- NEW DEVELOPMENTS IN MAKING COMPOUNDS AND MATERIALS-P. L. TIMMS air stable, bright yellow, easily sublimed solid of structure shown below .-Figure 3 Preparation and structure of bis(hexafluorobenzene)tungsten. compounds even catch fire spontaneously in air. The hexafluo- robenzene raises the effective oxidation state of the Mo or W to a point where air oxidation cannot easily occur. Bis(hexafluoroben- zene)tungsten decomposes at about 180 "C in either argon or in air showing its indifference to the presence of oxygen.The structure of [W(C,F,),] (see Figure 3) determined by X-ray diffraction has tung- sten-carbon bond lengths very similar to those in [W(C,H,Me),J showing that the different g-donor and ?r-acceptor properties of hexafluorobenzene and toluene must balance out. Using mixtures of C,F, and C,H, or C,F, and C,H,F,- 1,3,5 with Mo and W atoms gives the mixed products [M(C,F,)(C,H,)] and [M(C,F,)(C,H,F,)] which show beautiful NMR spectra as the nuclei in the two rings couple to each other. This coupling produces well resolved septets in both the 'H and I9F NMR of [M(C,F,)(C,H,)J. For [M(C6F6)(C,H3F3-I ,3,5)],the 19F NMR (see Figure 4) shows a septet of quartets for the C,H,F, ring and a sym- metrical 10-line multiplet for the C,F, ring formed by a superim- posed quartet of quartets in which the inter-ring F-F coupling constant is twice the inter-ring H-F coupling constant.Atoms of the Group 8 metals ruthenium and osmium condensed with C,F, have not yielded any stable complex [M(C,F,),]. With these metals, mixtures of non-halogenated arenes and hexafluoro- benzene give products of type [M( q6-arene)( q4-C,F6)] in which the +C,F, ring is non-planar, i.e. the metal atom optimises its share of electrons from full coordination of benzene or an alkylbenzene and partial coordination of the poorer donor hexafluorobenzene.21 We have made a compound [Ru(C,F,)(PF,),] from Ru atoms, C,F, and PF, in which the hexafluobenzene appears to be v-coordinated but the complex decomposes at low temperatures and complete charac- terisation has proved very difficult.Experiments with hexafluorobenzene and metal atoms have enabled us to explain better the way in which all low-temperature reactions between metal atoms and arenes Earlier work had shown that condensation of chromium or molybdenum atom with pure bromobenzene or with mixtures of benzene and bromobenzene gave no molecular product containing coordinated bromobenzene . Instead, oxidation of the metals by the bromobenzene occurred. We observed that condensation of the same metal atoms with a mixture of hexafluorobenzene and bromobenzene gave [ Cr(C,F,)(C,H,Br) and [Mo(C,F,)(C,H,Br)] respectively. How had the hexafluo- robenzene permitted bromobenzene to coordinate to the metals without oxidising them? From this and many related experiments we found that the key is the sequence of addition of arene ligands to the metal atoms and the sensitivity to oxidation of the intermedi- ate 'half-sandwich' complex.At the very low temperatures at which metal atom reactions are conducted, only reactions with very low activation energies can proceed rapidly. When a chromium atom interacts with a bromo- benzene molecule at low temperatures, two types of reactions are possible: complexation of the r-electrons of the bromobenzene to the chromium to give the half-sandwich [Cr(C,H,Br)] or oxidative addition of bromobenzene to chromium to give C,H,CrBr. The evi- dence is that half-sandwich formation is the dominant process in this and all reactions of metal atoms with arenes.However, the chromium in the half-sandwich complex is more electron rich than a free chromium atom so it is more susceptible to oxidation and it is also harder for a second arene molecule to coordinate to it. Thus, bromobenzene rapidly oxidatively adds to [Cr(C,H,Br) I forming chromium(I1) products from which the coordinated arene is lost and [Cr(C,H,Br),] is not produced. In the situation in which a mixture of C,H, and C,H,Br is condensed with chromium atoms, the C,H,Br destructively oxidises [Cr(C,H,)] as well as [Cr(C,H,Br) I, but with the C,F,+C,H,Br mixture and chromium, ICr(C,F,)) is resistant to oxidation by C,H,Br and coordination occurs instead to give the observed I Cr(C,F,)(C,H,Br)] .These ideas are summarised in Figure 5.Spectroscopic evidence for [V(C,F,)] analogous to [Cr(C,F,)] was obtained by Ozin2, from EPR studies of the vanadium atom-C,F, reaction under matrix-isolation conditions and we have recently made various complexes of type V(C,F,)(arene) on a preparative scale. However, neither [Cr(C,F,)] nor [V(C,F,)] seems capable of adding another C,F, molecule to form the bisarenemetal I I I I I I I I I I -150.25 -150.30 -177.05 -177.10 -177.15 6 6 Figure 4 Fluorine-19 NMR spectrum of IM(C,F,)(C,H,F,- 1,331. CHEMICAL SOCIETY REVIEWS, 1996 oxidation,no isolable organometpI1Jc product Figure 5 Comparison of redox properties of different half-sandwich inter- mediates in the reaction of chromium atoms with arenes. complex although Mo and W can, presumably because the bonding capabilities of 3d orbitals are more strongly affected by an increase in effective oxidation state of the metal than are the 4d or 5d elec- trons of Mo or W.Hexafluorobenzene is itself capable of destructively oxidising some half-sandwich molecules. In the reaction of molybdenum atoms with a mixture of C,H, and C,F,, [Mo(C,F,)(C,H,)] is vir-tually the only molecular product because any [Mo( q-C6H6)] formed is destructively oxidised by C,F, and the [Mo(q6-C,F6)] half-sandwich adds benzene much more easily than it adds hexa- fluorobenzene. This combination of the oxidising power and weak donor ability of C,F, limits the range of metal atoms which will form stable complexes with hexafluorobenzene. For example, we have so far failed to make hexafluorobenzene complexes contain- ing Nb, Mn, Re or Fe.However, these properties of hexafluoroben- zene also complicate alternative approaches to studying its coordination chemistry without using metal atoms. Some progress has been made in this area24 but it is likely that metal atom chem- istry will remain the best approach for many of these products. 4 Condensation of SiO at High Temperatures -a Versatile Fibre-forming System The molecule SiO is well known as a gas-phase species at high tem- peratures. In Norway and other countries rich in hydroelectric power, arc furnaces are used to produce metallurgical grade silicon and ferrosilicon from SiO,/C and SiO,/Fe,O,/C on a huge scale.In these furnaces, the formation and reactions of gaseous SiO play a major part in reduction and heat transfer. Escape of hot gaseous SiO from the furnaces into the atmosphere leads to its oxidation and con- densation as finely divided silica, SO,. This material used to be the main component of smoke effluents from such industrial arc fur- naces which caused considerable damage to the local environment. After legislation forced the polluters to collect the silica instead of releasing it, they began seeking uses for the waste product. Today that collected silica is in such great demand as an additive for strengthening concrete that some arc furnaces are being run with the primary purpose of producing SiO, from oxidised SiO rather than silicon or ferrosilicon! When hot gaseous SiO condenses in an inert gas stream, it forms a brown powder originally named nearly ninety years ago as 'Monox'.25 This solid is basically a mixture of silicon and silica because there is no stable solid phase SiO and the disproportionation: 2SiO(gas)-Si(liquid or solid)+SiO, (viscous liquid or solid) accompanies condensation.In the 1950s, the B. F. Goodrich Company in the USA undertook a study of the generation and con- densation of SiO with a view to making solids of use as fillers for rubber. They reported in patents that if condensation is carried out quenchgas 60 mm 0.d.quart-z tube f-fl::a*additive -r.45 mm 0.d. Pyrex flow tube -450 kHz induction O cruciblecoil -: 0 grapbite felt + silica 0 oil seal \I cruciblegas tlow Figure 6 Apparatus for preparing and condensing gaseous SiO. from a smoothly flowing gas stream containing SiO plus an inert gas (CO, N, or Ar), the SiO condensed mainly in the form of micro- fibres and that the proportion of fibres could be increased by adding a low percentage of ammonia to the inert gas.,, We followed up this work with the objective of learning more about the conditions under which fibres form but have also discov- ered that the morphology of the fibres can be changed by chemical additives?' Using the type of apparatus shown in Figure 6 we prepare SiO gas from a mixture of silicon and silica heated induc- tively in a graphite crucible to about 1800 "C.The SiO is carried upwards away from the crucible in a non-turbulent stream of N, or Ar and, as the gas stream cools, the SiO condenses to a brown solid which is collected in a filter. By scanning electron microscopy (SEM) the solid appears as groups of bent microfibres, of diameter around 20 nm and length of 200-300 nm as shown in Figure 7(a). Closer analysis by transmission electron microscopy (TEM) and powder XRD shows the silicon is present as microcrystallites. These are at highest concentration at the head but they also appear along the length of each fibre. The accompanying silica is amor- phous. The surface area of the fibres is quite high, around 150 m2 g-I.Conventional chemical analysis of the collected fibrous solid gives the composition as Si405, but if air exposure is rigorously avoided before analysis then we have proved that the composition is very close to SiO.28 Fibre growth occurs rapidly. We estimate that the time interval between nucleation of gaseous SiO and completed growth of the microfibres in the rising, cooling gas stream is <0.1 s. As noted in the B. F. Goodrich patents, the growth process is sensitive to a variety of physical parameters and we find that the condensed SiO consists of discrete fibres only if: (a) the gas stream leaving the cru- cible contains at least 10% SiO corresponding to the equilibrium pressure of SiO over SiO,/Si at 1600"Cor above; (b) the gas flow is non-turbulent in the condensation region; and (c) the gas stream is fast enough to carry growing fibres rapidly away from the SiO source.Based on these observations, our proposals on the stages of fibre growth which occur in the apparatus of Figure 6 are as follows. The SiO vapour, mixed with inert gas, leaves the graphite crucible where it is formed and begins to cool by radiative heat loss; NEW DEVELOPMENTS IN MAKING COMPOUNDS AND MATERIALS-P. L. TIMMS (a) (a 500 nm 500 nm 200 nm 50 nm Figure 7 Electron micrographs: (a) SEM of fibrous condensed SiO; (b) TEM of SiO condensed in the presence of AICI, vapour: (c) TEM of mixed con- densate of SiO and GeO: (d) TEM of SIOcondensed in the presence of WCI, vapour. additional cooling is caused subsequently as the gases rising from the crucible mix with the cooler quench gas flow.At a distance of about 15 mm above the crucible and at a temperature estimated to be >1400'C, the vapour becomes supersaturated and nucleation begins creating tiny, hot particles (studies have been made of SiO nucleation at sub-atmospheric pressure but not under conditions in which fibres would The concentration of molecular SiO in the gas phase then drops very rapidly to a level at which it con- tributes little to further condensate development. As SiO is not stable in the condensed phase, disproportionation to liquid Si and viscous solid SiO, takes place within the tiny particles; these phases tend to segregate to give molten silicon and viscous silica-rich domains.Aggregation of the primary particles proceeds rapidly with a slight gain in enthalpy; the most favoured mode of aggrega- tion will be for the molten silicon-rich areas on two particles to stick and coalesce followed by coalescence of the silica as it flows across the silicon surface. Repetition of this mode of aggregation gives the observed strongly asymmetric growth with a molten silicon-rich growth tip moving away from the solidifying silica-rich fibre tail. Segregation of silicon and silica is imperfect and some of the silicon is left imbedded in the silica in the tail. Growth of a fibre diminishes and eventually ceases as the supply of molten particulates is used up and as its own growing-tip head solidifies. Probably after solid- ification but while the fibres are still very hot, head-head collisions between fibres may cause sticking and the development of the cauliflower-like fibre groups visible in Figure 7(a).If the gas flow is turbulent, particulates in different stages of growth are intermixed and the progression of fibre growth is impaired. Unless the partial pressure of SiO vapour leaving the crucible is quite high, initially formed droplets do not grow fast enough to maintain themselves liquid and they eventually coalesce to give particulate material which appears to be made of loose aggregates of small spheres and sometimes poorly formed microfibres. We have added a wide range of gaseous inorganic compounds to the gas stream carrying the SiO to find the effect on SiO condensa- tion and fibre growth.Some of these are chlorides added as gases or vapours to the quench gas stream, e.g. AlCI,, GeCI, or WCl, vapour. In these cases, the collected products are heated under vacuum to remove any condensed or adsorbed metal halides and then any remaining metal is present in the product as free metal or as an oxide. Other gaseous inorganic compounds are generated as high-temperature species along with the SiO from within the cru- cible mainly by adding oxides to the Si/SiO, mixture. For example, partial replacement of SiO, in the crucible by Ga,O,, GeO, or Li,SiO, yields Ga,O, GeO or Li atoms to be condensed together with the SiO. Particularly striking changes in fibre morphology are obtained when the quench gas stream contains about 4%)AlCl vapour.The product then has much longer and less bent fibres than from SiO alone with less clearly defined head regions [Figure 7(b)]. Free silicon is still present in the fibres but very little of this is crystalline. There is ca. 5% of aluminium present as oxide in a non-crystalline CHEMICAL SOCIETY REVIEWS, 1996 mixed phase with silica Long fibres are also seen when Ga,O, or GeO, is added to the Si/SiO, mixture in the crucible to form Ga,O or GeO+SiO in the gas phase Fibres containing 20% Ge as GeO, are ribbon-like [Figure 7(c)] We have recently made very long microfibres by condensing SiO in the presence of lithium to give fibrescontaininga glassy Si02/Li2Si03phase and a Si phase which also contains some lithium 28 The common feature favouring fibre lengthening among all these systems involving main group metal additives seems to be the formation of a eutectic between SiO, and a few percent of the metal oxide We postulate that the formation of this liquid Si0,-metal oxide phase acts as a flux to aid material flow into the fibre tail and allows fibres to continue growing at a lower temperature This simple eutetic idea fails when applied to the transition metal chloride additives Transition metal oxides like Ti0, form a eutec-tic with SiO, very similar to that formed with A1,0,, but mixing of vapours of transition metal chloridesor 0x0-chloridessuch as TiCI,, VOCl,, CrO,CI, or WCI, with the SiO before condensation gives solid products which contain mainly Monox-like fibres In most cases, the fibres are covered with nodules of solids rich in transition metal oxides With WCI, the nodules are only 1-5 nm in diameter [Figure 7(d)] Analysis by XPS shows oxidation states of 4+ for Ti, 5 + for V, 3+ for Cr and a mixture of 4+ and 6+ for W From chem-ical analysisthere is as much free silicon in the fibres as with normal Monox but it is non-crystalline So why the difference between main group and transition metal additives' The most important factor is likely to be the much lower solubility of transition metals than main-group metals in molten silicon This means that while main-group chlorides are reduced to metals on the hot silicon-richgrowing tip and pass through in solu-tion to scavenge oxygen from SiO, immediately behind the tip, reduction of transition metal chlorides is blocked because the metal silicide formed does not dissolve in or pass through the molten silicon Instead the transition metal chloride vapours seem only to attack the hot Monox fibres after they have formed, undergoing metathesis with silica on the surface to give the observed nodules of metal oxides We think that the range of fibre morphologies that can be obtained from SiO condensation may be greater than for any other known substance We are still exploringpractical uses of the various types of fibres that we can produce, particularly their potential as reinforcing fillers for plastics or rubbers Many of them can be dis-persed by vigorous mechanical means in solutionsof polymers with very little damage to the microfibres,1 e they are quite strong, but bonding between polymers and the fibre surface needs to be improved to optimise their reinforcing properties In all our preparations of SiO fibres so far we have used atmos-pheric pressure because it was most convenientbut we now plan to investigate the condensation behaviour of SiO at pressures above and below atmospheric These studies should throw more light on the mechanism of fibre formation 5 New Low-temperature Solution Reactions of Silylenes Silylenes,along with other carbene-likemolecules,were among the first gaseous species formed under vacuum at high temperatures to be studied by the method of low-temperature condensation in the 1960s They were used to synthesisea large range of new organosil-icon compounds and many major features of their chemistry under these conditions had been explored by the late 1970s 30 However, recently we chanced on a new facet of the chemistry of silylenes in the following way Our aim was to make silicon oxoiodides, a previously unknown class of compound, to find If their thermal decomposition was a useful way of depositing high-quality silica layers for the elec-tronics industry We considered condensing iodine vapour with SiO under vacuum at low temperatures but, to avoid the nuisance of evaporating the iodine, we chose to condense SiO into a solu-tion of iodine in a solvent with a vapour pressure of 10 Pa at its freezing point Tests showed that iodoethane or toluene were suit-able solvents We hoped to avoid using carcinogenic iodoethane, toluene(so1vent) + 12(dissolved) + SiO(gas) condense 95'C warmto 20'C and evaporate solventJ viscous oil of composition Si2002411,{C,#14(CH3)}11H4made upof the groups shown in the partial structure below -IS\ 0 si //O\ \ \ H{/\ 7s toi //a=.si\si\ toi P;P tOi -77P I \ , r c Figure 8 The formation of a polysiloxane from reaction of SiO gas with a solution of iodine in toluene at -95 "C so we tried toluene cooled to -90 "C From earlier work?] we knew that toluene reacted inefficiently with SiO at low tempera-tures partly by addition across the aromatic multiple bonds and partly by insertion into ring or methyl C-H bonds giving intractable solids but we felt it was likely that the reaction with I, would predominate The result was quite different from our expectation The only reaction product was an oil of molecular mass around 3000 32 A combination of MS, IR, IH, I3Cand 29S1NMR, and chemical analy sis studies on the product indicated that it was a polysiloxane with C,H,(CH,), I and H substituents on silicon in an overall composi-tion close to Si200241,,{ C,H,( CH,)} H, The structure was complex apparently involving both ring and chain forms of the siloxane skeleton containing linked XY Si(0-), units [ X=I, Y =C,H,(CH,), H or I] with a few chain-terminating1,SiO-groups and chain or ring branching HSi(0-), or ISi(O-)? units (Figure 8) Carbon-silicon bonds had been formed exclusively with aromatic ring carbons, the predominant mono-substitution of H by Si on toluene gave relative proportions of the resulting ortho, para and meta isomers of 6 3 1 Subsequently, we found that the dihalogenosilylenes behave similarly 33 Condensation of SiF,, SiCI, or SiBr, into solutions quartz tube furnace rotating, evacuatedfilled with Si pieces I0 litre flask vacuum pumps -\ \cold solution ofcooling arene + iodine bath Figure 9 Apparatus used for reacting dihalosilylenes under vacuum with cold solutions NEW DEVELOPMENTS IN MAKING COMPOUNDS AND MATERIALS-P L TIMMS of iodine in toluene at -90 "C using the apparatus of Figure 9 results in efficient aromatic substitution reactions giving C,H,(CH,)SiX,I(X=F, C1, Br) with respective o p m isomer ratios of 3 1 1, 10 4 3 and 5 2 2 When the toluene solution of I, is replaced by ICl in toluene in these reactions, the same reaction prod- ucts are formed, i e there is no incorporation of chlorine from the ICI in the products How do these reactions occur7 The following is our current inter- pretation It has been known for many years that when I, is dis- solved in benzene, toluene or other alkylbenzenes, complexation occurs and the polarised I, molecules are sandwiched between the planes of aromatic rings as shown in Figure 10 When an SiO or SIX, molecule condenses in the solution at -90 "C, addition of I, ii0 Figure 10 A proposal for the mechanism of the reaction of silylenes with toluene and iodine to give SiO1, or SiX,I, definitely does not occur to any appreciable extent Instead, the SiO or Six, may act as a nucleophile, approach- ing the positively charged end of the polarised I, molecule and forming toluene SiOI+ I-toluene or toluene SIX&+ I-toluene moieties within the sandwich structure, which rearrange as shown in Fig 10 so the positive charge is on the toluene This effec- tively electrophilic attack liberates H+ which associates with the I-There is at present considerable interest in silyl cations and an indi- cation from the work of S~hleyer,~ that the interaction between an arene and a silyl cation could yield an intermediate of the type shown in Figure 10 In toluene solution, ICI is complexed in the same way as I, The fact that ICI reacts the same as I, provides strong evidence against a possible alternative mechanistic explanation involving free radi- cals For if the reaction of ICI and SIX, yielded free radicals, SiX,Cl*+I' would be more thermodynamically favoured than SiX,I*+ C1' The absence of chlorine incorporation in the products shows that SiX,CP cannot be the attacking species Overall, these iodine-induced reactions of silylenes with toluene or with other alkylbenzenes provide (a) the most direct way known of forming a silicone oil -albeit a very reactive one as it contains Si-I in addition to Si-aryl bonds but the Si-I can be quantitatively converted to Si-OR by reaction with an ether at low temperatures, and (b) a way of making chemically reactive arylSiX,I products, which are generally new compounds since they are hard to prepare conventionally What of the silicon oxoiodides? Did we prepare a liquid polymer of nominal composition (SIOI,),~ by condensing SiO into a solution of iodine in iodoethane at -100"C which was shown by 29Si NMR to contain mainly I2Si(0-)* units with some branching ISi(0-), units and terminal 1,SiO-units On strong heating under vacuum it vaporised yielding mainly cycl ic Si,O,I, but this has proved very reluctant to decompose cleanly on hot surfaces to SiO, and SiI, So the original objectives of our study proved to be worthless, but a rich new vein of silylene chemistry has been discovered in the process Acknowledgements 1 am very grateful to all my co-workers, post- doctoral, post-graduate and undergraduate named in the references, who have carried out the work described and to SERC and Elkem Chemicals who supported some of these coworkers and the research 6 References 1 P S Skell.L D Westcott, J P Golstein and R R Engel, J Am Chem Soc ,1965,87,2829 2 P L Timms, Chem Commun , 1%9,1033 3 J R Blackborow and D Young, Metal Vapour Synthesis In Organometallic Chemistry, Springer Verlag, Berlin 1979 4 P L Timms, in High Energy Processes in Organometallic Chemistry, ed K S Suslick, ACS Symposium Series 333,1987, p 1 5 K J Klabunde, Free Atoms, Clusters and Nanoscale Particles, Academic Press, London, 1994 K J Klabunde, Chemistry of Free Atoms and Particles, Academic Press, 1980 6 H W Kroto,Angew Chem ,/nt Ed Engl ,1992,31,111 7 R M Pope, S L Vanorden, B T Cooper and S W Buckner, Organometallics, 1992, 11,2001 8 M D Morse, M E Gensic, J R Heath and R E Smalley.J Chem Phys , 1985,83,2293 9 S Iijima, Nature, 1991,354,56, N S Chopra, R J Luyken, K Cherry, V H Crespi, M L Cohen, S G Louie and A Zettl, Science, 1995,269, 966 10 U E Wiersum, Reel Trav Chim Pays Bas, 1982,101,3 17,365 11 M J Almond and A J Downs, Spectroscopy of Matrln IAolatedSpecies. ed R J H Clark and R E Hester (Advances in Spectroscopy Vol 13, Wiley, Chichester. 1989 12 C G Francis and P L Timms,J Chem Soc ,Dalton Trans ,1980,1980 13 M L H Green,J Organornet Chem ,1980,200,119 14 M L H Green, D O'Hare, P Mountfield and J G Watkins, J Chem Soc ,Dalton Trans , 1991, 1705 15 U Zenneck, Angew Chem , Int Ed Engl , 1990,29,126 16 Z Yao, K J Klabunde and A S Asirvatham, fnorg Chem 1995,34, 5289 17 J G Brennan,F G N C1oke.A A SamehandA Kaikin,J Chem Soc , Chem Commun ,1987,1668 18 F G N Cloke,M F Lappert,G A Lawless andA C Swain,./ Chem Soc , Chem Commun ,1987,1167 19 R Middleton, J R Hull, S R Simpson, C H Tomlinson and P L Timms, J Chem Soc , Dalton Trans, 1973, 120, T S Tan and M J McGlinchey, J Chem Soc , Chem Commun , 1976,155 20 J J Barker,A G Orpen,A J Seeley,and P L Timms,J Chem Soc Dalton Trans , 1993,3097 21 A Martin,A G Orpen,A J Seeley and P L Timms,J Chem Soc Dalton Trans , 1994,2251 22 A J Seeley, S Hudson, P D Sykes, W B Weise and P L Timms,J Organomet Chem , 1995,487, 167 23 M Andrews, S Matlar and G A Ozin, J Chem Phys , 1991,95,2738 24 T W Bell, M Helliwell, M G Partridge and R N Perutz, Organometallics, 1992,11, 1911 25 H N Potter, Trans Am Electrochem Soc , 1907,12, 191 26 D S Sears, US Pat 2 865 881 and 2 865 882,1958 27 S R Church, K Chown, T Lippard, D Mulligan, S Sakanishi.P L Timms and G C Allen, J Muter Chem ,1995,5,757 28 A Songsasen and P L Timms, unpublished results, 1995 29 J A Nuth and B Donn, J Chem Phys , 1982,77,2639 30 C Lui and T Hwang, Adv lnorg Chem Radiochem ,1985,29,1 31 E T Schaschel, D N Gray and P L Timms, J Organomet Chem , 1972,35,69, W N Rowlands,Ph D Thesis.Universityof Bristol, 1989 32 W N Rowlands and P L Timms,J Chem Soc ,Chem Commun ,1989, 1432, P L Timms and W N Rowlands, U S Pat 5,034,554 (1991) 33 S R Church,C G Davies,R Lumen, P A Mounier, G Saint and P L Timms, J Chem Soc , Dalton Trans ,1996,257 34 P v-R Schleyer, P Buzek,T Muller, Y Apeloig and H-U Siehl,Angew Chem Int Ed Engl ,1993,32,1471

ISSN:0306-0012    

DOI:10.1039/CS9962500093

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Magnetism of Large Iron-0x0 Clusters Dante Gatteschi, Andrea Caneschi and Roberta Sessoli Department of Chemistry, University of Florence, via Maragliano 77, 1-50144, Italy Andrea Cornia Department of Chemistry, University of Modena, Modena, 1-41100, Italy 1 Introduction Large iron-oxo clusters are well known materials which have been investigated for many years now in relation to the mechanisms of hydrolysis of iron in water solution.' In fact it was observed that hydrolysis products which are found in natural systems can be related to prototypes identified in synthetic media. Numerous min- erals are formed by hydrolytic processes, such as haematite, goethite, ferrihydrite, lepidocrocite and maghaemite. The same processes are also observed in the mechanisms of biomineralization of iron.2 In fact iron oxides or 0x0 hydroxides are found in many dif- ferent living organisms, with different biological roles, as shown in Table 1.The iron oxide type materials found in biological systems are often observed in small particles, of size of a few nanometres. The best example is provided by ferritin, the iron storage protein, which is found in most living organisms. It is constructed of an array of polypeptide chains organized into a roughly spherical shell, which envelopes an inorganic core of approximate stoichiometry corre- sponding to ferrihydrite, 5Fe20,-9H20, as schematically depicted in Fig. 1. The size of the inorganic core is about 7 nm. The synthesis of large iron-oxo clusters has long been attempted Table 1 Iron oxides and 0x0 hydroxides found in living organisms Oxide Mineral Location Function 5Fe20;9H,O Ferrihydrite Ferritin Storage Molluscs radula Structural a-FeOOH Goethite Hemosiderin Storage Molluscs radula Structural y-FeOOH Lepidocrocite Molluscs radula Structural Fe3O'a Magneti te Magnetosomes Sensor Molluscs radula Structural Dante Gatteschi, born in 1945, graduated in Chemistry in 1969 in Florence in the Laboratory ofLuigi Sacconi.He has spent all his aca- demic life in the same University, where he was appointed Professor of General and Inorganic Chemistry in the Faculty ofPharmacy in 1980. His research interests are in the area ofmolecular magnetism. Andrea Caneschi, born in 1958, graduated in Chemistry at the University ofFlorence in I983 and obtained his Ph.D.in 1988 in the same institution \.-..*-Figure 1 Scheme of the structure of ferrttin. Apoferritin is a segmented protein shell with an outer diameter of 12.5 nm and an inner diameter of 7.5 nm. The inorganic core contained inside has a composition very close to that of ferrihydrite (5Fe20;9H20). with the aim to provide suitable models for the understanding of the above outlined mechanisms, but in recent years there have been additional reasons for interest in such compounds, associated with the investigation of the magnetic properties of these In fact the general trend towards interest in nanoscale objects has shown that in principle nanometer size magnets can be important both from the fundamental point of view, where they can provide examples of manifestation of quantum effects in large objects, and for their applications, where they can be in principle used for such diverse reasons as for optimal magnetocaloric effects and for magneto-optical devices .6-The general idea behind this interest is that with small particles it may be possible to store information with a higher density than by using large particles, as is common up to now.However the process of reduction of the size of the particles cannot continue indefinitely, because when the size becomes too small the particles cannot be any longer permanently magnetized. It is of significant interest to inves- tigate the range of size, where the cross-over from permanent magnet to paramagnet occurs, because it can be expected to provide evidence of coexistence of quantum and classic phenomena.under the supervision ofProfessor D.Gatteschi. He hada temporaryposi- tion at the Universita della Calabria in the period 1988-1 990 then he became a researcher in the University of Florence. Andrea Cornia was born in 1968 in Modena, where he graduated in Chemistry in 1992. In 1995 he completed his Ph.D. studies under the supervision of Professors A. C. Fabretti (Universita di Modena) and D. Gatteschi. His research work involves the synthesis and magnetic characterization of high-nuclearity iron and man-ganese clusters. Roberta Sessoli, born in 1963, graduated in Chemistry in 1987 at the University of Florence where she received a Ph.D in 1992 with a thesis on molecular magnetism under the supervision of Professor D.Gatteschi. Her major interests are the magnet- ism and magnetic resonances of large clusters and low dimen- sional materials. Andrea Cornia Andrea Caneschi Roberta Sessoli Dante Gatteschi 101 In principle it would be desirable to learn how to synthesize large iron-oxo clusters, by adding the iron ions one after the other, or in small blocks, but this is not yet possible, and growing large clusters is still much of an art. However many beautiful examples of large clus- ters have been structurally characterized?- The advantage of these systems over the small magnetic particles grown by physical tech- niques is that the molecular clusters are well defined objects, whose structures can be determined exactly, and all the particles are of course identical in the lattice. Furthermore they can often be dissolved with retention of their structures and embedded in polymeric matrices.In this way it is possible to investigate the magnetic properties of the individual clusters, reducing the magnetic interactions between them practically to zero by keeping the clusters far apart from each other. This review aims to show some relevant examples of iron(III)-oxo clusters which have been structurally characterized and for which detailed magnetic data are available. Since 'large' has only a relative meaning we will start to consider a cluster large when it consists of at least six iron ions.We do not attempt to cover the literature exhaustively, but report only some examples which we feel are particularly relevant to illustrate the magnetic behaviour of the clusters, taking them essentially from our own data. 2 Synthetic Strategies Polyiron4xo species are present in aqueous solution, but, since iron(II1) ions are not able to stabilize terminal 0x0 ligands, the growth of the polyiron complexes cannot be controlled, and in the absence of additional ligands beyond 0x0, hydroxo and aquo ligands the final product of hydrolysis in water is ferrihydrite. Alkoxide ligands are different, because the organic part of the ligand itself can act as a block to the growth of the particles, and for instance it was possible to isolate a compound of formula N%[OFe,(OMe) ,,](MeOH), from a methanolic solution of iron(1rr) in the presence of sodium methoxide.I2 The situation is different in the presence of additional ligands, like carboxylates, which can block the growth of the particles, by occupying the empty coordination sites around the metal ions.Many large iron-oxo clusters have been synthesized by controlled hydrolysis of the basic iron carboxylates.'3 These are stable (p3-0x0)-hexakis(p-carboxy1ato)-triiron(1Ir) species which have the structure shown below: R .R I R The [Fe30I7+ core can be used as a building block to give rise to larger clusters. For instance the decanuclear complex [Fe(OMe),(O,CCH,Cl)] can be obtained14 by a reaction between [Fe,0(02CCH2CI),(H,0),](N0,),~4H,0 and Fe(N03);9H20.Similar strategies were used to synthesize clusters with eleven and sixteen iron(II1) ions.IO The carboxylates are not the only co-ligands used to block the growth of the ferrihydrite particles. For instance using the ligand H3heidi: 4COOH / HO-"\ CHEMICAL SOCIETY REVIEWS, 1996 it was foundi5 that, starting from solutions at pH 2.4,a compound Of IFei 7(p3-o)4(pU,-oH>,(E-0H) I O(heidi)8(H,0) 21 lFe 19(113-0),(1U3-0H),(1U2-0H)~(heidi) 1 O(H20) 1 21(N03)4*60H20was obtained. Other ligands which have been used extensively are the polyket- onates. In fact methanolysis of simple iron(1rr) salts in the presence of P-diketonates has proved to be an excellent route to Fe,, Fe,, Fe,, Fe,, and Fe,, clusters.16J7 These systems contain OMe and P-dike- tonate ligands and provide some insight into the aggregation processes which take place in alcoholic media. 3 Selected Structures 3.1 Fe, Clusters Several different types of Fe, clusters have been reported, which were classified as planar, twisted boat, chair, parallel triangles, octa- hedral and fused clusters.18 Recently a new structure was reported', as shown in Fig.2.The six iron ions define a ring, and they are bridged by p2-alkoxo ligands. The [Fe,(p2-OMe),,(dbm),] ring, [Fe, ring], where Hdbm is dibenzoylmethane, is a neutral species, but in the solid it crystallizes with NaCl. In fact the sodium ion is trapped in the centre of the iron ring, which acts like a crown ether complexing the alkaline ion.The six iron ions and the sodium ion are practically coplanar, with average deviations from the least squares plane of 2.4pm. Therefore the compound can be formulated as [NaFe,(p,-OMe) I,(dbm),]Cl . An isostructural compound trap- ping a Li+ ion has been isolated in our laboratory. NMR experi- ments have recently shown that the analogy between the Fe, ring and crown ethers is not only structural but also functional. In fact, when the lithium-containing compound is dissolved in chloroform in the presence of sodium ions, substitution of Li with Na takes place. A very interesting feature of the structure, and one which will be often encountered also with other iron-oxo clusters, is that the oxygen atoms of the bridges and of the P-diketonate ligands tend to form close-packed layers, similar to those found in the lattices of metal oxides and hydroxides.In fact the oxygen donors of [Fe, ring] are assembled essentially on two planes, with maximum deviations of 20 pm. The distances between the oxygen atoms compare well with those observed in the continuous lattices of the oxides and hydroxides. Therefore it can be concluded that a molecular cluster like that depicted in Fig. 2 is indeed a small piece of an iron hydrox- ide. A variation on the theme of structures with iron ions lying Figure 2 PLUTON view of [NaFe,(p2-OMe) ,2(dbm)61+. MAGNETISM OF LARGE IRON-OX0 CLUSTERS-D. GATTESCHI ET AL Figure 3 Schemes of the structures of the hexanuclear iron(l1r) clusters composed of two parallel triangles.Figure 4 PLUTON view of [Fe6(p6-O)(p,-OMe)12(OMe)6]2-. approximately on the same plane is provided by the compounds shown in Fig. 3. The two triangles in the Fe, clusters are defined by p3-oxo ligands, and they are connected by different ligands ranging from hydroxo to peroxo group~.I~-~~ Structures in which the iron ions define an octahedron have been observed in two clusters recently reported by Hegetschweiler et a1.12 The main feature of these two com-pounds, ~Fe,(~6-o){(oCH2>3(cMe)),12-and [Fe6(k6-0)(p22-OMe) ,,(OMe),I2-, [Fe, octahedra], is that they contain a unique p6-oxo bridge. The structure of [Fe6(p6-O)(p2-OMe),2(OMe),]2-is shown in Fig. 4.The close-packed structure of these compounds is very apparent; in fact there are three layers of oxygen atoms, in which the maximum deviation from planarity is 13 pm, and the iron ions are assembled in two planes in between the oxygen planes.Actually there are more oxygen layers, because the anions are capped by sodium ions. The cluster is therefore a nice example of a small part of a cubic close- packed structure. 3.2 Fe, Clusters Wieghardt reported a cluster of eight iron(II1) ions, [(tacn),Fe,(p3- 0)2(p2-OH),zBr,(H20)]+,[Fe,], whose structurez3 is given in Fig. 5 (tacn is 1,4,7-triazacyclononane).The four central iron ions are connected by two p3-oxo bridges in a well known disposition, observed in a few tetranuclear structures, which has been called ‘butterfly’.24 The two iron ions connected by bis-p.,-oxo bridges define the body of the butterfly, while the ions connected by single p3-oxo bridges define the wings.The overall symmetry of the LW Figure 5 PLUTON view of [(ta~n)~Fe,(p,-0)~(~~-0H),,Br,(H,O)I+ Figure 6 PLUTON view of [Fe(OMe),(O,CCH,Cl)] ,o. cluster is D,.In this case also the iron ions lie approximately on a plane which is sandwiched by two oxygen and nitrogen layers. 33 Felo Clusters An interesting cluster containing ten iron(r11) ions, with a ring struc- ture, [Fe(OMe),(O,CCH,Cl)] [Fe,,], named ferric wheel, was reported by Lippard et al.I4a few years ago. The structure of the cluster is shown in Fig. 6. From the magnetic point of view this cluster, like [Fe, ring], is a beautiful example of systems which can be used as models for the interpretation of the magnetic properties of linear chains.The iron ions are octahedrally coordinated and they are bridged by two methoxo and one carboxylate group. 3.4 Fe,, and Fe,, Clusters The largest clusters so far reported contain seventeen and nineteen iron(Ir1) ions respectively.I5 They crystallize together in the same unit cell and have two formulae [Fe,,(p3-O),(p3-OH)6(p2-oH),o(heidi),(H,o),,13+, [Fe,71, and [Fe,,(pL,-0)6(p3-0H),(E.Lz-OH),(heidi),o(H20)12]+, [Fe,,], respectively. The two clusters have very similar structures. That of [Fe,,] is shown in Fig. 7. [Fe,,] has two additional iron ions on the exterior of the cluster, but the central moiety of seven iron ions is identical in the two compounds.It is interesting to note that this central core of seven ions is very similar +to that observedI6 in [NaFe,(p,;OMe) 12(dbm)6],provided that the sodium ion of the latter is substituted by an iron ion. 4 Magnetic Properties of the Clusters: Theoretical Considerations Finite-size clusters can be considered to behave as complex para- magnets, in which there are in general strong to moderate interac- tions between the spins present in the cluster, and weak interactions between the clusters. In order to interpret the magnetic properties of the clusters it is customary to use a spin Hamiltonian in which the painvise interactions between the different spins, i andj, are para- metrized as in eqn. (1): In the case of pairs the levels can be expressed as a function of the total spin S: with the energies: E(S)= J/2 [S(S+ I)-S,(S, +1)-S,(S2+ I)] (3) When J<O (ferromagnetic coupling) the ground state has the maximum spin multiplicity S= S, + S,, while when J>O (antifer-romagnetic coupling) the ground state has S= S, -S,.Use of this Hamiltonian for dimers yields the well known Bleaney-Bowers equation25 for the magnetic susceptibility. Extention of eqn. (1) to large clusters is in principle easy, but in practice it rapidly becomes extremely difficult. In fact for a cluster of Nspins characterized by individual quantum numbers S,,the total number of states which form the Hamiltonian matrix is (2S,+ l)N.For instance, for S,= 5/2, as appropriate for iron(rrr), when N= 8 the CHEMICAL SOCIETY REVIEWS, 1996 Table 2 Symmetry classification of the total spin states of Fe, in the point group D, S A B, B2 B3 Total 20 1 - - - 1 19 2 1 2 2 7 18 10 6 6 6 28 17 22 18 22 22 84 16 60 50 50 50 210 15 118 108 118 118 462 14 243 225 224 224 916 13 419 401 420 420 1660 12 717 690 686 686 2779 11 1088 1061 1092 1092 4333 10 1614 1578 1568 1568 6328 9 2174 2138 2184 2184 8680 8 2841 2799 2780 2780 11200 7 3401 3359 3420 3420 13600 6 3927 3885 3854 3854 15520 5 4139 4097 4170 4170 16576 4 4155 4122 4076 4076 16429 3 3704 367 1 3750 3750 14875 2 3019 3001 2940 2940 11900 1 1899 1881 1960 1960 7700 0 703 703 630 630 2666 number of states is already 1 679 616.Therefore the calculation of the spin levels, and of the magnetic susceptibility, becomes a for- midable task. The best approach so far reported is that of using irre- ducible tensor operators, ITOs, which exploit at best the symmetry associated with the total spin.26 However even in this case it is nec- essary to calculate 21 matrices (corresponding to the total spin values Sranging from 0 to 20), the largest of which, S= 5, is of size 16 576X 16576. A further reduction in the size of the matrices can be obtained using the point group symmetry of the cluster. For D, symmetry the results are as shown in Table 2.In general the exchange interactions observed in oxo-bridged iron(ri1) ions are antiferromagnetic, with the general trend that J decreases in the order from 0x0 to hydroxo bridges. Since the cor- responding iron-oxygen distances are shorter for the 0x0 than for the hydroxo bridges a simple way of taking into account the differ- ence between the two types of groups is to use the average bond dis- tance, P, or in the presence of multiple bridges the average bond distance relative to the shortest superexchange pathway. Gorun and Lippard found27 a reasonable correlation for 34 compounds; we have now extended it to 50 compounds. The best fit equation, for P in 8, and J in cm-', is: 5=3.25X lo9 exp(-9.3P) (4) From Gorun and Lippard's work no correlation emerged between Fe-O-Fe angles and coupling constants.However, Kurtz2, noted that in strongly coupled oxo-bridged diiron(rrr) complexes the J value is angle-dependent, linear bridges leading to stronger interac- tions. The same conclusion has been reached recently in the case of alkoxo clusters, in which substantially weaker interactions are present. Attempts were also made to find a theoretical justification for this, but they were not extended to very small angles, <loo", which can be observed in p3-oxo bridged systems. In some cases we were obliged to use smaller J constants for a given P value at small angles. Further work is needed in this area. Large clusters are characterized by a large number of spin levels; therefore these will tend to merge into a continuum. When this occurs the system must cross over from complex paramagnetic to bulk magnetic behaviour.This transition must be expected to occur gradually. From a theoretical point of view we may imagine large clusters being built by adding metal ions one at a time. This path to infinity may occur in one, two, or three dimensions as suggested by Fig. 8. Linear clusters or rings can be models for one-dimensional magnetic materials, planar clusters for two-dimensional and finally three-dimensional clusters for bulk magnetic materials. MAGNETISM OF LARGE IRON-OX0 CLUSTERS-D. GATTESCHI ET AL f 0-\ Figure 8 Schematic path to infinity in one-, two- or three-dimensions. Magnetizationdirection Figure 9 Representation of the energetic barrier that a particle must undergo to reorient its magnetization in the absence of an external mag- netic field.In this ideal process at some stage the interacting spins of the cluster will become frozen in a preferred orientation, corresponding to a ferro-, ferri-, or antiferrornagnet, and the system becomes ordered. Associated with the ordering is the development of a mag- netic anisotropy, which will dictate the spatial orientation of the magnetic moment. The origin of the anisotropy can be associated with the shape of the particle, to the contribution of the anisotropy of the individual ions, or a contribution associated with the mag- netic interaction between the individual ions.For relatively large particles the anisotropy A can be assumed*9 to be proportional to the volume V of the particle: A=CV (5) A represents the barrier that a particle must overcome in order to reorient its magnetization to the energetically equivalent position, in the absence of an external field, as shown in Fig. 9. When A is large compared to thermal energy the magnetization will be blocked in one of the two minima, but when the barrier is small the magnetization will freely flip over from one minimum to the other. The ordered system will thus behave as a paramagnet, but one characterized by a large magnetic moment. A particle which behaves like this is called a superparamagnet. The relaxation of the magnetization T follows a thermally excited process: T= T(, exp(Alkr) (4) T~is typically s for ferromagnets like iron.A given particle behaves like a bulk magnet when Tis larger than the characteristic time T, of the investigation technique used to 1 05 monitor it. T~is 10-3-10-1 s for ac susceptibility measurements and lo-* s for Mossbauer spectroscopy. The temperature at which the relaxation time of the magnetization equals T, is called the blocking temperature of the superparamagnet, T,. In the case of clusters it can be thought that the first indication of bulk behaviour comes from a slow relaxation of the magnetization observed at low temperature. A further reason for interest in the investigation of the dynamics of the magnetization of the clusters at low temperature is that, since their size is intermediate between that of bulk objects, to which classical mechanics applies, and molecular objects, to which quantum mechanics applies, at very low tempera- ture the inversion of the magnetization may be thought to occur by a tunnelling process.Indeed this is one of the reasons for the great inter- est of solid state physicists in this new kind of magnetic material. 5 Magnetic Properties of the Clusters: Experimental Results 5.1 Fe, Clusters Here we will focus on the types which can be represented by the spin topologies depicted in Fig. 10. By spin topology we mean the set of exchange pathways which can be expected to be present between pairs of iron ions, on the basis of the presence of chemical bridges, like 0x0 groups connecting pairs of metal ions.In Fig. 10 we have not made any attempt to differentiate between different types of bridges, but we have considered them all as equivalent. Of course the differences in the intensities in the exchange interactions transmitted by the different bridges are of paramount importance in determining the actual magnetic properties of the individual clus- ters. Figure 10 Possible spin topologies for hexanuclear clusters. Up to now there is one cluster which has the spin topology (a). The magnetic properties of [Fe, ring] are rather obvious,16 domi- nated by the antiferromagnetic interaction between the nearest- neighbour iron ions. The magnetic susceptibility of [Fe, ring] goes through a broad maximum at ca. 150 K, and then decreases rapidly.The comparison with the behaviour expected for infinite chains of iron(Ir1) ions suggests a coupling constant J=20 cm-I. The quanti- tative fit of the magnetic properties of the cluster is less obvious, because the number of states to be calculated is very high, but the procedures outlined in Section 4 make it possible. A satisfactory fit to xvs.temperature yielded a coupling constant, J= 20 cm-I, which agrees with the values observed in dinuclear complexes with similar bridges. A further proof of the goodness of the fit comes from the field dependence of the magnetization at very low temperature, which provided the energies of a few excited states. Indeed magnetization measurements in large clusters can provide a wealth of information as shown below.The magnetization of [Fe, ring] was measured at 1.5 K in fields up to 52 T. At low field the magnetization is close to zero, but on increasing the field several steps are observed where the magneti- zation rapidly increases. The steps are equally spaced in field, as shown in Fig. 11. These data can be explained considering that the saturation magnetization of a paramagnet characterized by a ground spin state S is given by M= g,+S. In a system like [ Fe, ring1 for CHEMICAL SOCIETY REVIEWS, 1996 3 s=3 s=2 S=l M=Os=o M=-1 6 M=-2 M/P, M=-34 2 0 I'"""""J 0 20 40 60 Hfr Figure 11 Spin crossover diagram for the first spin levels of [Fe, ring]; (0) experimental points.very low fields the magnetization must be equal to zero, because the ground state has S= 0. However, on increasing the external field, the energies of the M= -S components of all the excited Sspin mul- tiplets rapidly decrease, with slopes proportional to -S, as shown in Fig. 11. It is clear that up to 10T the magnetization is close to 0. Above this limit it rapidly increases to reach the saturation value of S= 1.A further increase of the field determines a second step which reaches the magnetization of S=2. Finally a third step gives the magnetization of S=3. The inflection points of the magnetization curve are regularly separated in a field by about 16 T as clearly observable by the plot of the differential magnetization, dM/dH, vs.magnetic field, shown in Fig. 12. If the lowest lying spin multiplets follow a Lande interval rule, i.e. they are given by eqn (7): E(S)= (K/2)S(S+ 1) (7) it is easy to show14 that the crossover points will occur at fields given by eqn (8): where BSS+, is the field value at which the spin multiplets S and S+ 1 have the same energy. The consequence of eqns. (7) and (8) is that the crossover points will be observed at regular intervals in a I1000 1""""""" 800 s600 400 200 0 10 20 30 40 50 60 WF Figure 12 Differential magnetization, dMldH (arbitrary units), vs. mag-netic field at 1.5 K for (Fe, ring]. field of K/gpB.The calculated energies of the excited states corre- spond within experimental error to the observed ones.It might be asked why the lowest-lying S excited levels follow a Lande interval rule. The answer is rather simple: the lowest lying levels can be described to a good approximation by a spin configu- ration in which all the spins on the odd sites are up, and those on the even sites are down. In general the spins in the ring of 2n members can be partitioned into two sets, each containing II objects. The two sets will have identical intermediate spin S,=S,=nS,. If the two intermediate spins are coupled to give the total spin S the energies of the levels according to eqn. (3) are given by eqn. (9): showing that the first excited states must follow a Lande interval rule. K is an effective coupling constant, which is related to J according to eqn (10): K is expected to go to zero for large n, i.e.when the energy levels must merge to give a continuum. There are quite a few Fe, clusters with the spin topology depicted in Fig. 10(b), and they have been ~hown~~-,~ to give rise to several different ground spin states ranging from S= 0 to S= 5. The spin topology in this case shows the presence of two triangles, and these, in the presence of antiferromagnetic interactions, make it impos-sible to predict the nature of the ground state on the basis of up-down spins. In fact it is easy to see that if we put for instance spin (a) of Fig. 13 up and spin (b) down, there is no way that spin (c) can obey at the same time the antiferromagnetic interactions with the other two spins.This has been called spin frustration?O by analogy with the psychological term which applies to a person under the influence of two equally strong and opposing stimuli. The most general scheme for the exchange interactions in cen-trosymmetric clusters with spin topology (b) is as depicted3, in Fig. 10 (b). 3 d Figure 13 Schematic representation of the spin topology in an antiferro- magnetic coupled triangle of spins. In the following we will assume that J, ,J,, and J, are similar to each other, and much larger than J4.We will also assume that all are positive, corresponding to antiferromagnetic interactions. In the most symmetric case, J,=J,=J,, the ground state is always S= 0, so it is relatively unexciting.The same situation is achieved if it is assumed that J, =J,#J,. Matters become much more inter- esting if it is assumed J, =./,#],. In this case the possible ground states are variable depending on the J,/J, ratio, r. The calculated temperature dependence of the resultant peffis shown in Fig. 14,for selected values of r. For r= 1 the ground state is S= 0, for r= 0.7 it is S= I, for r= 0.55 it is S= 3, and for r=0.2 it is S= 5. In fact when r tends to zero the spin frustration in the triangles is com- pletely destroyed and the nature of the ground state is easily pre- dicted with elementary spin upspin down considerations. The magnetic s~sceptibility~~ of the [Fe, octahedra] clusters, see Fig. lOc, reported by Hegetschweilerl2 gives a clear indication of a ground S= 0 state, which is steadily expected on the basis of qual- itative considerations.The quantitative interpretation can be done in a simple way in this case, using the formalism first suggested by Kambe.33 The energies of the levels are given by eqn. (I 1): where J,. is the coupling constant for two ions which are cis to each MAGNETISM OF LARGE IRON-OX0 CLUSTERS-D GATTESCHI ETAL 10 :1 77K Figure 14 Calculated temperature dependence of the pewfor different r values other in the octahedron, and J, the corresponding constant for two ions trans (S, and S,, S, and S,, S4and S,) and S,j is the intermedi- ate spin resulting from the coupling of the 1 andj spins The best fit values are J,= 21 2 cm-I, J, = 9 9 cm-I The surprising feature is that despite the very long Fe-0 distances involved in the p, bridge the coupling constants are relatively large, certainly larger than would be expected on the basis of a largely accepted relation between Fe-0-Fe distance and coupling constant 27 The essential results of the investigation of the magnetic proper- ties of Fe, clusters can be summarised as follows (1) The magnetic properties of Fe, clusters are those of paramagnets (11) The quantitative interpretation of the magnetic properties can be arrived at using sophisticated extensions of the usual techniques In order to find deviations from simple paramagnetic behaviour it is necessary to extend the investigations to larger clusters 52 Fe, Clusters If the magnetic interaction transmitted by the 0x0 bridges is domi- nant in [ Fe,], the four central iron ions should be in an S= 0 ground state 74 In fact the central butterfly of four iron ions has the wing-body interaction transmitted by 0x0-bridges with exchange pathways close to 3 8 A, and 0-Fe-0 angles close to 130",while the body-body interaction corresponds to longer exchange path- ways, 3 91(2), 3 96(2) A, and much smaller angles [962( I)" and 97 3( 1)", respectively 1 Therefore it can be concluded that spin frus- tration effects are not very important, and the ground state is S= 0 The interactions with the external iron ions are determined by two different types of bridges, corresponding to one and two hydroxo groups, respectively If the coupling constants are suffi- ciently different from each other it can be concluded that the ground state for the cluster must have S= 10,as shown in the Scheme 1 This is qualitatively confirmed by the temperature dependence of xT, which at room temperature is 20 06 emu mol I K, much smaller than expected for eight uncoupled S=5/2 spins (35 emu mol I K), a clear indication of large antiferromagnetic coupling The XT Scheme 1 product increases on decreasing the temperature, showing that at low temperature the number of spins up is different from the number of spins down, i e the cluster is fenmagnetic High-field magnetization data at low temperature agree with an S= 10ground state The quantitative fit of the magnetic properties was attempted using the IT0 approach of Section 4 Areasonable agreement with experiment was found by setting J,= 20 cm I, J,= 120 cm-I, J,= 15 cm I, J4= 35 cm-1 J, corresponds to the body-body interac-tion in the central butterfly, J, to the wing-body interaction, J, to the bis-p-hydroxo and J, to the single p-hydroxo bridges This is so far the largest cluster for which quantitative interpretation of the magnetic properties was achieved The nature of the ground state was c0nfirmed3~ also by high fre- quency EPR data, recorded at 250 GHz and 42 K, which show 20 transitions The spectra can be interpreted with D= -0 191 cm I, E/D= 0 1675 The sign of the zero-field splitting indicates that the M components with the highest possible value, namely M= 10,lie lowest in energy This has the important consequence that the system becomes very anisotropic at low temperature (Ising type anisotropy, I e one easy axis), and the reorientation of the magneti- zation becomes very slow In fact Mossbauer spectroscopy showed that the reorientation of the magnetization becomes slow on the time scale of the experiment, lo-, s, at ca 30 K, while ac suscep- tibility measurements, which have a time scale of 10 s, show similar results below 3 K This behaviour is analogous to that of superparamagnets, and is entirely determined by the magnetic anisotropy of the cluster, which in its turn is mainly determined by single ion contributions These results are of extreme importance, because they confirm previous analogous findings in an Mn,, clu~ter,3~ which is the first reported case of molecular cluster with superparamagnetic-like behaviour showing magnetic hysteresis effects Furthermore they show that superparamagnetic behaviour previously observed in larger iron oxo-clusters4 must have been determined by an analo- gous mechanism, and not by the pseudo-three-dimensional struc-ture of the particles In fact [ Fe, I is essentially planar 53 Fe,, Clusters The magnetic properties14 of [Fe,,] are simple to interpret, by analogy with those of the cyclic \Fe, ring] cluster described above In fact assuming nearest-neighbour anti ferromagnetic interaction, a ground S= 0 state would be expected The magnetic susceptibility in fact goes through a maximum at ca 65 K, while XT decreases from 32 2 emu mol-I K at room temperature to 0 355 emu mol I K at 2 5 K By analogy with simple dimers the maximum in x sug-gests a coupling constant J-10 cm--' The quantitative interpreta- tion of the data is however impossible in this case, because the dimensions of the matrices are by far too large for us to be able to diagonalize them However, by calculating the susceptibility for rings of 4,6and 8 S= 5/2 spins, we could extrapolate the curve to 10 spins, with a very good agreement between observed and calcu- lated values The coupling constant, J= 9 6 cm I, appears to be in the right range More interesting are the magnetization data for [ Fe obtained at 0 65 K In fact in this case also, as for the cyclic [Fe, ring] clusterI6 described above, it can be expected that crossovers between differ- ent spin states are observed on increasing field In Fig 15 the mag- netization in pulsed fields up to 40T is shown The curve corresponds to the derivative of the magnetization as a function of the field Therefore the peaks in Fig 15 correspond to inflection points in the magnetization vs field curve, i e they correspond to crossovers between different ground states with increasing S The first peak in low field corresponds to the transi- tion S=O-.l, the second to 1+2, and the last at high field to 849' The fact that the field separations between the different crossovers are always the same confirms a Lande interval rule up to S= 9 Using eqns (8) and (10) it is possible to calculate J= 9 4cm 1, in excellent agreement with the value calculated from the suscepti- bility 108 041 I 142 1 I 2-13 :0 I I I I 0 10 20 30 40 m Figure 15 Differential magnetization, dMldH for [Fe(OMe)2(02CCH,CI)I,, measured in a pulsed field The crossover at 0 65 K between the S, (S+ 1) spin states is noted above each maximum 5.4 Fe,, and Fe,, Clusters The magnetic susceptibility of [ Fe,,+Fe,,] has a temperature dependence15 which is very similar to that of [ Fes] Actually If the susceptibilities per iron ion are reported for the two compounds they are seen to be practically identical Therefore it can be concluded that Fe,,+Fe,, are ferrimagnetic clusters The low-temperature magnetization shows that the ground states of one of the clusters cannot be smaller than 33/2 This is the largest ground spin state so far observed in magnetic clusters The EPR spectra of IFe I +Fe,9] recorded at room temperature show one very broad signal centred at g= 2 At low temperature the spectrum becomes anisotropic, with gL-3 This behaviour is anal- ogous to that observed in low-dimensional magnetic materials,”6 in which internal fields determine a shift in the g values No anomaly could be observed in the Mossbauer spectra, which show usual paramagnetic behaviour down to 3 2 K 6 Conclusions Iron-oxo clusters are a very important class of materials First of all they provide some insight into the mechanisms of formation of inor-ganic cores which are observed in iron biomineralization The struc- ture of the clusters seems to be dominated at an early stage by the tendency of the oxygen atoms to define close-packed structures, even when they are bound to the organic part of the Iigands The synthetic strategies developed so far are still rather naive and more work is needed in this area in order to learn how to grow larger clus- ters, to control their structures From the magnetic point of view the reported clusters provide large variability, within the general scheme of antiferromagnetic exchange interactions The cyclic [Fe, ring] and [Fe,,] clusters are excellent models of the magnetic behaviour of one-dimensional antiferromagnetic materials, and they have provided for the first time a detailed knowledge of the energies of the excited states More ring structures would be desirable, in order to clarify some pointh which remain unclear, such as the detailed behaviour of the suscep- tibility observed at low temperature, below the maximum The quantitative treatment of the magnetic properties requires drastic improvements, introducing explicitly zero-field splitting effects also However some encouraging successes have been obtained, which suggest that the whole matter is not hopeless All the clusters we have reported here, except IFe, octahedra], have essentially planar arrangements of the iron ions, and they can be considered as models of one- and two-dimensional magnetic materials It will be necessary also to investigate in more detail the magnetic properties of systems which have pseudo three-dimen- sional arrangements of the iron ions in order to model the proper- ties of three-dimensional magnets CHEMICAL SOCIETY REVIEWS, 1996 The reported data clearly show that the magnetic properties of the clusters slowly evolve toward bulk magnetic behaviour on increasing the number of ions The observation of superparamag-netic-like behaviour is associated with the development of the magnetic anisotropy of the clusters The requirements which so far emerge in order to give rise to such behaviour are those of having a large ground spin state with a magnetic anisotropy char acterized by the presence of one easy axis, I e an Ising type anisotropy More clusters are needed in order to confirm this view, making clear which are the conditions determining such type of anisotropy 7 References 1 W Schneider, Chimia, 1988,42,9 2 Biomineralization Chemical and Biochemical PerJpectives, ed S Mann, J Webb and R J P Williams, VCH, New York.1988 3 D Gatteschi ,Adv Muter, 1994,6,635 4 K L Taft, G C Papaefthymiou and S J Lippard,Science, 1993,259,1302 5 D D Awschalom, D P DiVincenzo and J F Smyth, Science. 1992,258, 4 14, S Gider, D D Awschalom,T Douglas, S Mann and M Chaparala, Science, 1995,268,77 6 R Sessoli, D Gatteschi, A Caneschi and M A Novak, Nature. 1993, 365,141 7 A J Leggett, Phvs Rev 6, 1984,30,120 8 P Politi,A Rett0ri.F Hartmann Boutronand J Villain. Phvs Rev Lett, 1995,75,537 9 A Muller, E Krickemeyer, 5 Dillinger, H Bogge, A Proust, W Plass and R Rohlfing, Naturwissenschuften, 1993,80,843 10 W Micklitz, V McKee, R L Rardin, L E Pence, G C Papaefthymiou, S G Bott and S J Lippard, J Am Chem Soc , 1994,116,8061, S M Gorun, G C Papaefthymiou, R B Frankel and S J Lippard, J Am Chem Soc ,1987,109,3337 11 H L Tsai, H J Eppley, N de Vries, K Folting, G Christou and D N Hendrickson, J Chem SOC Chem Commun ,1994,1745 12 K Hegetschweiler, H Schmalle, H M Streit and W Schneider, Inorg Chem ,1990,29,3625, K Hegetschweiler, H Schmalle, H M Streit, V Gramlich, H U Hund and I Erni, lnorg Chem ,1992,31, 1299 13 R D Cannon and R P White, Progr Inorg Chem , 1988.36.195 14 K L Taft. C D Delfs, G C Papaefthymiou, S Foner, D Gatteschi and S J Lippard, J Am Chem SOC ,1994,116,823 15 A K Powell, S L Heath, D Gatteschi, L Pardi, R Sessoli, G Spina, F Del Giallo and F Pieralli, J Am Chem SOC 1995,117,2491 16 A Caneschi, A Cornia and S J Lippard, Angew Chem Int Ed Engl , 1995,34,467 17 A Caneschi, A Cornia, A C Fabretti, D Gatteschi and W Malavasi, Inorg Chem 1995,34,4660,A Caneschi, A Cornia, A C Fabretti and D Gatteschi, Angew Chem ,1995,34,27 16 18 CA Christmas.H L Tsai,L Pardi,J M Kesselman,PK Gantzel,R K Chadha, D Gatteschi.D F Harvey and D N Hendrickson, J Am Chem Soc ,1993,115,12483 19 N V Gerbeleu, A S Batsanov, G A Timko, Y T Struchkov, K M Indrichan and G A Popovich, Dokl Akad Nauk SSSR, 1987,293, 122 20 W Micklitz and S J Lippard, Inorg Chem 1988, 27, 3067, W Micklitz, S G Bott, J G Bentsen and S J Lippard, J Am Chem SOC , 1989,111,372 21 J K McCusker,C A Christmas,PM Hagen,R K,Chadha,D F Harvey and D N Hendrickson, J Am Chem Soc , 1991,113,6114, VS Nair and K S Hagen, Inorg Chem 1992,31,4048 22 C J Harding, R K Henderson and A K Powell, Angew Chetn Int Ed Engl ,1993,32,570 23 K Wieghardt, K Pohl.I Jibril and G Huttner, Angew Chem Int Ed Engl ,1984,23,77 24 J K McCusker, J B Vincent, E A Schmitt, M L Mino, K Shin, D K Coggin, PM Hagen, J C Huffman. G Christou and D N Hendnckson, J Am Chem Soc 1991,113,3012 25 K Bleaney and K D Bowers, Proc Roy Soc 1952, A214,45 1 26 D Gatteschi and L Pardi, Gax Chim It ,1993,123,231 27 S M Gorun and S J Lippard, Inorg Chem ,1991,30. 1625 28 D M Jr Kurtz, Chem Rev 1990,90,585 29 A H Morrish, The Physical Principles of Magnetism, Wiley, New York, 1966 30 J Vannimenous and G Toulouse,J Phvs C SolidState Phvs .1977,10, L537 3 I C D Delfs, D Gatteschi and L Pardi, Comments Inorg Chem , 1993 IS,27 32 A Cornia, D Gatteschi and K Hegetschweiler Inorg Chem ,1994,33 1559 MAGNETISM OF LARGE IRON-OX0 CLUSTERS-D GATTESCHI ET AL 1 09 33 K Kambe.J Ph\J Soc Jpn , 1950,5.48 35 D Gatteschi. A Caneschi, L Pardi and R Sessoli, Science, 1994,265, 34 C D Delfs, D Gatteschi, L Pardi, R Sessoli, K Wieghardt and D 1054 Hanke, lnorg Chetn , 1993, 32, 3099, A L Barra. P Debrunner, D 36 D Gatteschi and R Sessoll, Magn Reson Rev, 1990,15,1 Gatteschi, C E Schulz and R Sessoli. submitted for publication

ISSN:0306-0012    

DOI:10.1039/CS9962500101

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Infrared Fourier Transform Emission Spectroscopy Peter F. Bernatht Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada N21 3G 7 1 Introduction Traditionally high-resolution infrared spectra of gas-phase mole- cules are recorded in absorption and the advantages of acquiring infrared emission spectra have largely been overlooked. As this article will illustrate, using mainly our own work, all else being equal, infrared emission spectroscopy is more sensitive than absorption spectroscopy. Since transient molecules have low con- centrations because of their great chemical reactivity, it is advanta- geous to use emission spectroscopy to study them. The infrared region of the electromagnetic spectrum provides a unique window for monitoring molecular processes.Mainly vibra- tional transitions of molecules are found in the infrared, although there are infrared electronic transitions and infrared pure rotation transitions. In general, vibrational transitions carry specific chemi- cal group information in a way that electronic and pure rotational transitions do not. It is for a good reason that the organic chemist refers to the 900-1400 cm-I interval as the ‘fingerprint region’. All molecules, except homonuclear diatomics, have at least one electric-dipole-allowed vibrational transition. Even homonuclear molecules such as H, have weak electric-quadrupole transitions which are used by astronomers to detect molecular clouds that are experiencing shock waves .I Vibrational transitions are thus a uni-versal, molecule-specific monitor of chemical composition; every major industrial laboratory and university chemistry department has at least one infrared spectrometer.Vibrational spectroscopy also has some deficiencies. Perhaps the most intrinsic problem is that vibrational transitions are relatively weak. A ‘typical’ molecule such as CH has a dipole moment of 1.43 Debye, a transition dipole moment of 0.134 Debye for the funda- mental v= 1 +0 vibrational transition and a transition dipole moment of 0.70 Debye for the A2A-X217 electronic Since the relative intensity depends on the square of the transition dipole moment, the vibrational transition is about 100 times weaker than the pure rotational or the electronic transition. For the spec- troscopy of stable molecules, this intrinsic weakness is hardly a problem -one simply uses a little more sample.For the infrared vibrational spectroscopy of transient molecules at low concentra- tions, this relative lack of sensitivity puts the technique at a serious disadvantage. Emission spectroscopy is one technique that improves the infrared performance of a spectrometer. The infrared region of the electromagnetic spectrum suffers from some other technical deficiencies. For example there is no infrared iAlso Drpurtment of Chrmittr), Univer,it) of Arizona, Tucwn, AZ 85721, USA Peter Bernath was born in I953 in Ottawa, Ontario, Canada. Following his B.Sc. degree from the University of Waterloo(1976), he received a PhD in physical chemistryffom MIT in 1980.After a post-doctoral stint at the National Research Council of Canada, he became a faculty member at the University of Arizona ffom 1982-1 990.In I991 he took up his present position as Professor of Chemistry & Physics at the Universiy of Waterloo, while maintaining a connection with the University of Arizona. equivalent of the visible dye laser or the microwave klystron that provides widely tunable radiation. On the horizon are infrared optical parametric oscillators (OPOs) which will provide tunable, single-mode, high-power radiation .4 Infrared detectors and infrared materials such as windows and lenses also tend to be inferior and more expensive that the corre- sponding visible components. The infrared analogue of the photo- multiplier tube for the measurement of low radiation levels is not available.Recently, however, a solid state Si: As avalanche infrared detector has been manufactured by Rockwell .5 For the visible and ultraviolet region, it is well-known that emis- sion spectroscopy is more sensitive than absorption spectroscopy because emission spectra have ‘zero background.’ In an ideal absorption experiment, the noise arises mainly from shot-noise from the background continuum. By comparison, in an emission experiment, the noise is reduced because the continuum is absent. Of course the total signal level is also reduced in an emission exper- iment compared to an absorption experiment but the overall signal- to-noise ratio is increased because the noise has declined much more than the signal.For example the violet colour often seen in a flame is due to the easily detected A2A+X2Uemission of the CH molecule. The detection of the corresponding absorption due to the CH molecule in a flame is possible but constitutes a much more dif- ficult experiment.6 What is surprising to many people is that the advantages of emis- sion spectroscopy persist into the infrared. Infrared spectrometers are routinely operated in absorption, not emission. Infrared emis- sion spectroscopy requires that the sample be at a different temper- ature (higher or lower) than the spectrometer. Higher temperatures are more favourable since the emitted power increases strongly with temperature.Infrared emission spectroscopy is plagued by thermal blackbody emission from the spectrometer itself as well as from the furnace or electrical discharge used to heat the sample. This black- body emission is a continuum that provides nothing but noise. The Planck function is a law of nature that can never be entirely avoided in the infrared. By limiting the field of view with cold apertures and by good optical design (e.g. by avoiding having the spectrometer look directly at a hot surface), as well as by limiting the spectral range of the detectors with cold filters, the effects of unwanted blackbody emission can be minimized. Another apparent difficulty is that the rates of emission are much smaller in the infrared than in the visible or ultraviolet. For example, the violet A2A+=X217transition of CH has a radiative lifetime? of 526 ns or an Einstein A value (reciprocal of the lifetime) of 1.9X 1 O6 SKI.This compares to a vibrational lifetime of 8.7 ms or an Einstein A value of 115 s-’ for the v= 1-0 transition.* Since the Einstein A value is the rate per molecule of photon emission, the emission rates are obviously much smaller for vibrational transitions than for elec- tronic transitions.This is due to both the v3term in the formula7 in eqn. (1): 1629 2 A=-‘P‘ (1)3eOhc3 for the Einstein A factor and the relative weakness of vibrational transitions. Nevertheless the infrared emission rates are adequate to provide an acceptable signal-to-noise ratio for CH8 (Fig.1) or even for a transient molecule such as LiH in the far-infrared region9 (Fig. 2).We discovered the power of infrared emission spectroscopy by serendipity. At the request of the astronomer J. Keady we were trying to record an absorption spectrum of SiSIO at 750 cm-I (Fig. 3) to match molecular features in an infrared spectrum of a carbon 111 star The experimental plan was simple heat Si powder and SiS, solid to 1000"C to make SiS by the reaction (2) The hot molecules would then be detected by absorption of infrared 28k7.6 vlcm ' Figure 1 The CH R(4) vibration-rotation emission line for the v= 1-0 transition of the X21Z state The quartet pattern is characteristic of a 2L7 state that follows Hund s case (b)coupling LIH 8( LID 100 0 2000 3000 406 0 v/cm ' Figure 2 The far infrared pure rotational emission spectrum of LiH and LID in the I OW00cm I region The absorption features are due to water vapour impurity HEATERS TUBE FURNACE CHEMICAL SOCIETY REVIEWS 1996 light from a glower Much to our amazement a very high quality emission spectrum was recorded in only 14 minutes of data collec tionlo (Fig 4) In 1989 when our SiS experiments were carried out with the McMath Pierce Fourier transform spectrometer of the National Solar Observatory in Tucson, Arizona, there were scarcely any pub lished examples of high resolution emission spectra at long wave lengths Apart from a relatively low resolution spectrum of GeS only the infrared emission spectrum of the F0'2 molecule was avail able in the literature The FO molecule was produced by the very exothermic reaction (3) F+ O,+FO +O2 (3) and for some time we were convinced that an energetic reaction was necessary to see emission Soon we recorded the thermal emission spectrum of BeF, (Fig 5), proving that no chemical reactions are necessary to provide additional vibrational excitation 13 Thermal emission spectra are, in fact, used to monitor atmos pheric molecules such as 0, in the stratosphere both from balloon platformsI4 and satellites l5 Thin films of solids and liquids also give excellent thermal emission spectral6 (Fig 6) Of course, spectra of materials in condensed phases are at low reso lutton 2 Infrared Electronic Transitions Most people find the idea of infrared electronic transition to be almost a contradiction, since most stable molecules typically have electronic transitions in the ultraviolet region Many transient mol ecules such as ions or free radicals, however, have visible or infrared electronic transitions because of the presence of unpaired electrons in low lying orbitals While these reactive molecules cannot be purchased from a chemical supply company they are, in fact very common In energetic environments such as in flames, explosions, upper atmosphere stellar atmospheres and the inter stellar medium, transient molecules are abundant and dominate the chemistry Transient molecules can be made in the laboratory by the appli cation of electrical discharges, heat or light to stable precursor TRANSFORM SPECTROMETER Figure 3 The experimental arrangement for high temperature emission and absorption spectroscopy For absorption spectroscopy the glower is turned on to provide a source of continuum radiation while for emission spectroscopy the glower is turned off Note that both emission and absorption experiments use the emission port of the Fourier transform spectrometer .....,,..., ..... .. Figure 4 Vibration-rotation emission spectrum of the SiS molecule near 13 pm (650-800 cm I) Notice the P branches 740-800 cm I and the R branches 650-740 cm I of the 1-0 24 1 3-2 etc vibrational bands INFRARED FOURIER TRANSFORM EMISSI( N SPECTROSCOPY -P 021 -02 0 101-100 011 -010 I. ........'."~''""~" 15Wi 1570 1575 vkm ' Figure 5 The vibration-rotation emission spectrum of the v3(uU)mode (antisymmetric stretching vibration) of the linear BeF, molecule The R branch band head of the v3fundamental is on the right edge of the figure In the middle is the head for the first bending hot band 011-010 and to the left of the figure are two heads associated with other hot bands 500 1000 1500 2000 2500 3000 vlcm ' Figure 6 Emission spectrum of a thin film (100-200 pm thick) of solid Na,CO, at a temperature of about 400 "C The two peaks marked with asterisks are due to Na,SO, impurity The effect of water absorption near 1500 cm I and CO, emission near 2300 cm I can be seen in the spec trum 3595 cm-' vfcm' 3805 Figure 7 The 0-0 band of the Bld,-rA'IJu electronic transition of the C, molecule Notice the band heads in the two R branches near 3605 cm I molecules For example the microwave discharge (2 45 GHz) of the hydrocarbons allene (CH,=C=CH,) or methane (CH,) resulted in the production of the C, molecule The new BiAg-AiUUinfrared electronic transitioni7 was observed in emission near 3600 cm I (Fig 7) Since the visible Swan system of C, (d3L7g-+a3L7u)and the near infrared Phillips system (AiL7,+Xl +Xg) are readily detected in flames and in comets,Is the new infrared BiAe-+Aiflu transition should also be present Infrared electronic transitions of polyatomic molecules are also well known, for example, for C,I9and HO, 2o A pioneering high resolution emission spectrum of the H0,free radical was recorded F BERNATH 113 with a SISAM spectrometer near 7000 cm (1 4.5 pm) 2o The HO, molecule is a key trace constituent of the earth's atmosphere l4 Infrared electronic transitions are particularly important for systems which contain transition elements Molecules with transi tion metals typically have a large number of low-lying electronic states and many unpaired electrons This means that the visible and ultraviolet electronic transitions are very complex because of the high density of states and the resulting perturbations In the infrared region the energies of the states are lower and the density of states is correspondingly reduced The infrared electronic transition of a molecule such as YN21 (Fig 8) is thus particularly simple and unperturbed The YN molecule was made by sputtering Y metal in a hollow cathode with Ne gas in the presence of a trace of N, P Branch R Branch I I 4170 4260 v/cm ' Figure 8 The 1-1 vibrational band of the AIC+-XrC+ electronic transi tion of the YN molecule Since the rotational constants are very similar in the AfX+ state and in the ground XlZ+ state, there is no bandhead 04 J 00 I 1750 1760 1770 1780 1790 1800 vlcm' Figure 9 The vibration-rotation emission spectrum of the R branch of the fundamental 1-0 vibrational band of A 1H The high temperature ( 1550 "C) makes it possible to see the R branch bandhead near 1800 cm I at R(26) The absorption features are due to the impurity H,O, both in the spectrometer (sharp absorptions) and in the short path of laboratory air between the furnace and the spectrometer (wider absorptions) 3 Inf r a red Vibration-Ro tation Transiti ons Vibration-rotation emission spectroscopy above 1800 cm I is not rare, even for transient molecules such as CHs (Fig 1) This magic number of 1800 cm I (55 pm) is the band gap of the InSb detec- tor The InSb detector has nearly unit quantum efficiency and is the best infrared detector but it does not work for wavenumbers below 1800cm The detection of emission at longer wavelengths is thus more difficult but excellent spectra can be recorded (Fig 4) In Fig 9 the emission due to A 1H molecules was recorded at a resolution of 0 005 cm I with a signal-to-noise ratio of more than 1SO for the strong lines 22 The measurement precision of 0 0002 cm I (6 MHz) is remarkable for a transient molecule Emission spectroscopy will work at even longer wavelengths as illustrated with the SiS spectrumlo (Fig 4) Even in the far infrared region it proved possible to record the vibration-rotation emission CHEMICAL SOCIETY REVIEWS, 1996 113 0 115 0 117 0 119 0 121.0 123.0 125 0 vlcm ' Figure 10 The vibration-rotation emission bands of CsI near 100cm-I. No individual rotational lines could be resolved but the 1-0 to 27+26 vibra-tional bandheads are marked in the spectrum.spectrum of near 100cm-I (Fig. 10). In this case the individ- ual rotational lines were not resolved because the rotational con- stant, B,has a value of 0.02 cm-'.Nevertheless the 1-0 to 30-29 bandheads could be measured. The observation of vibration-rota- tion bandheads is rare in room-temperature spectra but they are a common feature of high temperature spectra (Figs. 5,9 and 10). Vibration-rotation emission spectra are also readily recorded for polyatomic molecules such as BeF, (Fig. 5) and C,, (Fig. 11). The linear BeF, molecule is isoelectronic with CO, and the rotational structure was analysed to give an equilibrium Be-F bond length of 1.37297A.13For C, the rotational lines were not resolved (Fig. 11) but the band positions could be tracked as a function of temperature to determine an extrapolated 0 K band centre.24 The gas-phase posi- tions of the C,, bands at low temperature were used in a search for C,, in carbon-rich stars.25 - 600 , 700 1 800 .1 900 1000 1 1100 . 1200 7 vkm ' Figure 11 The vibration-rotation emission spectrum of gaseous C,. The C, bands are marked with asterisks. The sharp features between 500 and 900 cm-I, and above 1150cm-I are due to emission from hot water and carbon dioxide impurities. 4 Pure Rotational Emission Spectra Small gas-phase molecules with large rotational constants have pure rotational spectra in the far-infrared and infrared regions. The emission spectrum of the transient molecules LiH and LiD9 (Fig. 2) was recorded in the far-infrared region to show feasability and to explore the effects of the breakdown of the Born-Oppenheimer approximation.One application of pure rotational emission spectroscopy was the detection of water vapour on the sun.26 An infrared spectrum of a sunspot (Fig. 12) was recorded with the McMath-Pierce Fourier transform spectrometer at Kitt Peak. A high density of absorption lines was found and they could not be assigned to a known spectrum of a molecule. We suspected that the mystery molecule was hot water at the sunspot temperature of 2900 "C. The pure rotational emission spectrum of water vapour at 1550"C was recorded in the Figure 12 White light image of a sunspot (black) and the surrounding photosphere of the sun. The sunspot shows a dark, cool umbra (2900°C) and a sin rounding warmer penumbra. The photosphere of the sun shows granulation caused by the formation of convective cells on the surface of the sun.INFRARED FOURIER TRANSFORM EMISSION SPECTROSCOPY -P F BERNATH I i l l 1 IIII IIII IIII Ill1 Ill1 I v/cm" Figure 13 The upper panel is a portion of the pure rotational emission spectrum of hot water (1550"C) recorded in the laboratory The lower panel has two traces of the solar spectrum The upper trace is a portion of the absorption spectrum of the penumbra of the sunspot showing only absorption features caused by water in the earth's atmosphere The lower trace is the spectrum of a sunspot showing additional absorption features due to hot water vapour (2900"C) on the surface of the sun One pure rotation transition of the OH molecule can be seen near 539 4 cm I laboratory in the 400-1000 cm region The match between the line positions of the emission laboratory spectrum and the sunspot absorption spectrum was very good (Fig 13),identifying water on the sun 26 5 Future Prognosis The improvement in sensitivity obtained by working in emission rather than in absorptionwith high-temperature molecules, ions and free radicals is typically a factor of 10 in the mid-infrared region At the moment, the instrument of choice is the high-resolution Fourier transform spectrometer For astronomical applications for which the light levels are very low, the Fourier transform spectrometer is infe-rior in sensitivity to the cryogenic echelle spectrograph 27 A cryogenic echelle spectrograph is a compact spectrometer that uses an echelle grating in high orders to obtain high resolution spectra The entire instrument is cooled to 77 K with liquid nitrogen to eliminate most of the background thermal emission from the spec- trometer itself The infrared radiation is measured with a large format (512x512) array of InSb detectors Many of these spectrom- eters are under construction or have already been built at observato- ries around the world The anticipated performance for an instrument called Phoenix under construction at Kitt Peak National Observatory IS spectacular27 Phoenix will have a resolving power in excess of 100000 at 2000 cm I (a resolution of 0 02 cm-I) with a sensitivity 100 times higher than that of a Fourier transform spec- trometer This quantum leap in performance promises to revolution- ize both infrared astronomy and laboratory emission spectroscopy Acknowledgements The work described in this article was sup- ported by grants from the Natural Sciences and Engineering Research Council of Canada, The Petroleum Research Fund of the American Chemical Society, the NASA laboratory astrophysics program and the Combustion Research Facility of Sandia National Laboratories The National Optical Astronomy Observatories are operated by the Association of Universities for Research in Astronomy under cooperative agreement with the National Science Foundation 6 References 1 N Z Scoville, D N B Hall, S G Kleinmann and S T Ridgway, Astrophvs J , 1982,253,136 2 P F Bernath, Annu Rev Phys Chem ,1990,41,91 3 W Bauer, B Engelhardt, P Wiesen and K H Becker, Chem Phys Lett , 1989,158,321 4 For example, the entire November issue of J Opt Soc Am B ,1993.10 5 M D Petroff, M G Stapelbroek and W A Kleinhans, Appf Phys Lett , 1985,51,406 6 L Lynds and B A Woody, Appl Optics, 1988,27, 1225 7 P F Bernath, Spectra of Atoms and Molecules, Oxford Univesity Press, New York,1995, ch 1 8 P F Bernath, J Chem Phys , 1987,86,4838 9 B Guo, K Q Zhang, M Dulick and P F Bernath, in preparation 10 C I Frum, R Engleman and P F Bernath, J Chem Phys , 1990,93, 5457 11 H Uehara, K Horiai, K Sueoka and K Nakagawa, Chem Phys Lett , 1989,161,149 12 P D Hammer, A Sinha, J B Burkholder and C J Howard, J Mof Spectrosc ,1988,129,99 13 C I Frum, R Engleman and P F Bernath, J Chem PhyJ , 1991,95, 1435 14 B Carli and M Carlotti, in Spectroscopy of the Earth S Atmosphere and Interstellar Medium, ed K N Rao and A Weber, Academic Press, San Diego, 1992, 1 15 A E Roche, J B Kumer, J L Mergenthaler, G A Fly, W G Uplinger, J F Potter, T C James and L W Stemtt, J Geophys Res , 1993,98, 10763 16 P F Bernath et af ,in preparation 17 M Douay, R Nietmann and P F Bernath, J Mof Spectrosc ,1988,131, 261 18 M F AHearn and M C Festou in Physics and Chemistry of Cornets, ed W F Huebner, Springer Verlag, Berlin, 1990, p 101 19 H Sasada,T Amano,C JarmanandP F Bernath,J Chem Phys ,1991, 94,240 1 20 R P Tuckett,P A Freedman and W J Jones,Mol Phvs ,1979,37,379 21 R S Ram and P F Bernath,J Mof Spectrosc ,1994,165,97 22 J B White,M DulickandP F Bernath,J Chem Phys ,1993,99,8371 23 V Braun, B Guo, K Q Zhang, M Dulick, P F Bernath and G A McRae, Chem Phys Lett , 1994,228,633 24 L Nemes, R S Ram, P F Bernath, F A Tinker, M C Zumwalt, L D Lamb and D R Huffman, Chem Phys Lett, 1994,218,295 25 G C Clayton, D M Kelly, J H Lacy, I R Little Marenin, P A Feldman and P F Bernath.Asrronom J , 1995,109,2096 26 L Wallace, P Bernath, W Livingston, K Hinkle, J Busler, B Guo and K Q Zhang, Science, 1995,268, 1155 27 S T Ridgway and K H Hinkle, in High Resolution Spectroscopy with the VLT,ed M H Ulrich, ESO Conference and Workshop Proceedings, No 40, ES0,Garching bei Munchen, 1993,p 213

ISSN:0306-0012    

DOI:10.1039/CS9962500111

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Asymmetric Synthesis of &Amino Acids and a-Substituted P-Amino Acids Giuliana Cardillo and Claudia Tomasini Dipartimento di Chimica 'G. Ciamician,' Universita di Bologna Via Selmi 2 40126 Bologna Italy 1 Introduction In the recent years there has been increasing interest in the synthe- sis of proteinogenic and non-proteinogenic amino acids.' This is due to the wide utility of such compounds as components of pro- teins peptides and as starting materials for the synthesis of naturally occurring biologically active compounds. a-Amino acids2 are the most abundant in this class of com- pounds as they are the main components of peptides enzymes and proteins,? and they have been utilised as chiral reagents for a variety of synthetic applications. p-and y-amino acids although less abun- dant than their a-analogues are also present in peptides? and in the free form they show interesting pharmacological effects. Enantiomerically pure p-amino-a-hydroxy acids are of consider- able importance because they are the crucial components of medi- cinally useful molecules such as taxol? an anti-tumour agent bestatin an immunological response modifier and certain small peptides possessing antihypertensive activity. P,y-Diamino acids are also present in nature; for example the emericedins are inhibitors of long chain fatty acid oxidation and were isolated from the culture broth of Emericella quadrilineata IF0 5859.7 Furthermore p-amino acids are synthetic precursors of p-lactams which are potentially biologically active and of current interest.8 In addition many physiologically active compounds are present among the class of y-amino acids. y-Aminobutyric acid (GABA) is a very simple compound and a major inhibitory neurotransmitter; indeed when the concentration of GABA in the brain decreases seizures and other neurological disorders OCCU~.~ P-Hydroxy y- amino acids also display important biological activities (R)-carni- tine for example plays an important role in converting stored body fat into energy because its primary role is to transport large fat mol- ecules into cellular compartments where the fats can be metabolised. (2S,3S,4R)-4-Amino-3-hydroxy-2-methylpentanoic acid is a constituent fragment of the antitumour antibiotic bleomycin and the tripeptide mollusc toxin janolusimide. Dolaisoleuine (DIL) is a component of the polypeptide dolastatin 10,'O one of the most active antineoplastic substances presently known. The design of receptor-selective peptide and peptidomimetic ligands with highly potent and specific biological properties'' has become one of the most important areas in bioorganic chemistry medicinal chemistry molecular biology and other related research Giuliana Cardillo was born in Anconu Italy. She received her degree in Chemistrvfrom the University of Rome. Then she moved to the Politecnico of Milan working under Professor A. Quilico on naturally occurring chromenes. After two year5 at the Universityof Bari she moved to the Department of Chemistry 'G. Ciamician' of the Universiq of Bologna working with Profesmr G. Cainelli on the syntheJi5 of ter-penoids. She became ProfeJJor of Organic Chemistry in 1980. Her major re5earch interest5 lie in the area of asymmetric synthesis of polytknctionalized biologically active corripoundh ,and the stereo- Jelective 5ynthesiJ of amino acid5 is a recent focus. 117 areas. At the present time there is rapid growth in the number of endogenous and exogenous biologically active peptides under investigation. Most of these peptides are short lived molecules easily degraded by enzymes with little or no therapeutic use. To convert them into useful drugs it is often necessary to transform these compounds into more resistant molecules.'* Small cyclic pep- tides show increased resistance to enzymatic degradation and con- strained flexibility as compared to their linear analogues. Consequently they frequently exhibit higher biological selectivity and activity. Furthermore incorporation of a-alkyl a-amino acids into peptides results in conformational restrictions and increased rigidity leading to enhanced resistance towards protease enzymes and to the favouring of particular secondary structures. Indeed owing to severe restrictions of the rotational freedom around their N-C(a) and C(a)-C=O bond a-alkylated a-amino acids may be generally expected to display helix-inducing properties.' ? There is considerable interest in the synthesis of unnatural amino acids in order to introduce them into a peptide sequence.l4 They can be divided into a-substituted a-amino acids,I5 p-amino xids and y-amino acids. This review is intended to give a brief summary of some of the more recent developments in the synthesis of p-amino acids and a-substituted /?-amino acids. The first part is concerned with an overview of the naturally occurring compounds containing p-amino acids. The second part summarises results on the synthesis of natural and unnatural p-amino acids by routes involving the enzy- matic resolution of racemic mixtures the intermediary preparation of perhydropyrimidin-4-ones and the conjugate addition of ammonia equivalents to a P-unsaturated esters and imides which are all topics near to the scientific interest of the authors. For a more complete overview of the methods developed to obtain enan-tiomerically pure p-amino acids including hydrogenation of 3-aminoacrylates and nucleophilic addition of enolates to imines see two recent excellent reviews.i6 2 Natural Occurrence of Molecules containing P-Amino Acids Cyclic and non-cyclic peptides with important pharmacological properties have been isolated from marine organisms or from plants. These compounds contain in addition to unusual acid Claudia Tomasini was born in Bolognu and studied at the University of Bologna. wlzere she graduated in 1982 and gained her Ph.D. in 1987. Thereafter she obtained a grant from Italian C.N.R. (National Research Council) spending a year at Dyson Perrins hboratory in Oxford working with Professor Jack E. Baldwin. Currently she work5 as Research Associate at the University of Bologna. Her current research interests are centred on the stereoselective functionalisation of double bonds for the synthesis of poly-substituted a-and p-amino acids in optically pure form. moieties such as D-amino acids and hydroxy acids non-proteino- genic a-and p-amino acids a-substituted p-amino acids and 7-amino acids. Here we briefly describe the natural occurrence and the properties of some molecules containing p-amino acid derivatives. 2.1 Molecules containing P-Amino Acids and a-Alkyl P-Amino Acids The dolastatins 10 form an heterogeneous group of potent antineo- plastic and cytostatic peptides isolated from the Indian Ocean sea hare Dolabella auricularia. They are simply numbered from 1 to 12 and do not have a coherent set of structures. Indeed some are linear and some are cyclic and all contain unusual components such as modified amino acids residues lactone units polyketide fragments and thiazole rings. Dolastatins 11 and 12 (Fig. 1) are cyclic dep- sipeptides which have recently been isolated and structurally elu- cidated. These molecules exhibit cell growth inhibitory activity against the murine P388 lymphocytic leukaemia and contain a 2-methyl-3-aminopentanoicacid. This a-substituted p-amino acid is also contained in the structure of majusculamide CI7 (Fig. 2) a cyclic depsipeptide from Lyngbya rnajuscula which is a toxic blue-green alga and grows abundantly on the pinnacles in the lagoon of Enewetak Atoll in the Marshall Islands. Majusculamide C controls the growth of a number of fungal plant pathogens such as Phytophthora infestans the causative Figure 1 Dolastin 11 R =H; R,=CH,O Dolastin 12,R,=CH,; R,=H Figure 2 Majusculamide C 0P?.# Figure 3 Jasplakinolide organism of tomato late blight and Plasrnopora viticola the causative organism of grape downy mildew. /?-Tyrosine an unsubstituted p-amino acid is contained in jas- plakinolidel8 (Fig. 3) which was the first mixed macrocyclic polyketide-depsipeptide to be isolated from a marine organism. This sponge metabolite is of considerable interest because of its anthelminthic insecticidal ichthyotoxic and antifungal properties. P-Phenylglycine [ ( +)-(R)-3-amino-3-phenylpropanoicacid] is a CHEMICAL SOCIETY REVIEWS 1996 Figure 4 Astin A R,=H R,=OH R,=H Astin B R =OH R,=H R,=H Astin C R,=H R,=H R,=H Cyclochlorotine R =H R,=H R,=OH Asterin R,=H R,=H R,=CH p-amino acid structurally related to P-tyrosine and is contained in a group of cyclic pentapeptides astins A B and C and asterinI9 (Fig. 4),from the medicinal plants Aster tataricus L.f. (Compositae) known in Chinese medicine as containing several terpenoids and saponins. Also astins A B and C are antitumour agents. A very similar molecule is cyclochlorotine,20 which is one of the toxic metabolites of Penicilliurn islandium Sopp the mould of islandia yellow rice. Furthermore longer chain and more complicated a-substituted p-amino acids are contained in onchidin21 (Fig. 5) and motuporin22 (Fig. 6),which are both cytotoxic cyclic peptides. Indeed onchidin AM0 Figure 5 Figure 6 Motuporin is a cytotoxic depsipeptide isolated from the pulmonate mollusc Onchidium sp. Its cyclic structure is made of two identical halves and contains two moieties of a new p-amino acid 3-amino-2- methyl oct-7-ynoic acid (AMO). Motuporin is a cyclic pentapeptide from Theonella swinhoei Gray collected in Papua New Guinea and is a potent protein phos- phatase-1 inhibitor. It contains a very complicated a-methyl p-amino acid bearing several chiral centres simply called ADDA. 2.2 Molecules containing @-Amino a-Hydroxy Acids P-Amino-a-hydroxy acids are important components in a variety of biologically interesting compounds. A number of protease inhibitors for example derive their activity from the ability of the p-amino a-hydroxy acid motif to act as a transition state mimic of peptide hydrolysis. Here we show some p-amino a-hydroxy acids contained in biologically active molecules. Most of them are natu- rally occurring but some have been obtained by synthesis as promising anti-AIDS or antihypertensive agents. One of the best known molecules containing an p-amino a-hydroxy acid is taxo15 (Fig. 7) which is composed of a polyoxy- genated diterpene and (2/?,3S)-phenylisoserine.Although the role of the phenylisoserine side chain of taxol has not yet been fully ASYMMETRIC SYNTHESIS OF P-AMINO ACIDS AND P-SUBSTITUTED P-AMINO ACIDS-G CARDILLO ET AL Figure 7 Taxol determined it plays an import role in the biological function of this antitumour agent For instance one of the C-2' or C-3' polar func- tionalities can be removed without significant effect but the removal of both or the interchange of their position causes dramatic loss of activity Another example of a biologically active p-amino a-hydroxy acid is isothreonine which is the side chain of the glycopeptide 1-A'-( ~-threo-3-amino-2-hydroxybutanoyl)-2',3 '-dideoxykanamicin A*? (Fig 8) a good antibacterial agent Bestatin6 (Fig 9) is a simple dipeptide containing (2S,3R)-3- amino-2-hydroxy-4-phenylbutanoicacid [abbreviated as (2S,3R)- HO@I OH Figure 8 Dideoxykanamicin A Figure 9 Bestatin AHPA] and L-leucine and is an immunological response modifier It has been isolated from culture filtrates of Streptomyces olivoreti-cull and inhibits aminopeptidase B and leucine aminopeptidase but not aminopeptidase A carboxypeptidases or endopeptidases Studies on derivatives in which the amino group is protected show that the free amino group is essential for the activity and the car- boxyl group is also important The (2S,3S) epimer of AHPA is contained in two synthetic tripeptides kynostatins (KN1)-227 and (KNI)-27224 (Fig lo) which are highly potent HIV-1 protease inhibitors and are promis- Figure 10 (KNI) 227 R=Me (KNI) 272 R=H ing candidates as selective anti-AIDS agents The configuration of the hydroxy group has a profound effect indeed replacement of (2S,3S)-AHPAwith its (2S,3R) epimer leads to a significant loss of activity This stereochemical preference is surprising since it is opposite to that observed for bestatin Many others small peptides have been isolated from sponges or algae and contain p-amino a-hydroxy acids For instance kera- mamide F (KF)25 (Fig 11) is a cytotoxic natural product found in small quantities in a Theonellu sponge This cyclic peptide contains a remarkable array of unusual amino acids including an isoserine residue as a side chain ScytonemynA26(Fig 12) is a cyclic peptide from the blue-green alga Scytonernu sp (strain U-3-3) (Scytonemataceae) and possess Figure 11 Keramamide F AHDA Figure 12 Scytonemin A potent calcium antagonistic properties As well as some pro-teinogenic a-amino acids such as alanine glycine and leucine scytonemin A contains (2S,3R,SS)-3-amino-2,S,9-trihydroxy-l0-phenyldecanoic acid (AHDA) which is a novel p-amino acid including several chiral centres Micro~clerodermins~~(Fig 13) are cyclic hexapeptides from a deep water sponge of the genus Microscleroderma The structures of microsclerodermins A and B are very similar and were deter- mined by interpretation of spectral data They both contain a novel and complex p-amino acid (2S,3R,4S,SS,6S,ll Q-3-amino-6- methyl-12-(p-methoxypheny1)-2,4,5-trihydroxydodec-1 1 -enorc acid (AMMTD) and a known y-amino acid such as GABOB (3R)-4-amino-3-hydroxybutyric acid Finally microginin28 (Fig 14) is a linear tetrapeptide and contains a novel 3-amino-2-hydroxydecanoic acid (called AHDA like the previous one) in the (2S,3R) configuration The molecule has recently been isolated from the freshwater blue-green alga H Figure 13 Microsclerodermin A R=OH Microsclerodermin B R= H fYoH AHDA OH Figure 14 Microginin Microcystis aeruginosa and was found to possess angiotensin-con- verting enzyme (ACE) inhibitory properties which is of consider- able medical interest since ACE inhibitors have been developed as antihypertensive agents Naturally occurring ACE inhibitors often suffer from problems such as poor oral absorption short duration of action proteolytic instability and rapid biliary excretion For this reason some new synthetic compounds are currently being developed as promising antihypertensive agent Figs 1529and 1tj30 show the structures of synthetic rend I inhibitors containing the novel p-amino-a-hydroxy acid cyclohexylnorstatine (2R,3S)-3-amino-4 cyclo- hexyl-2-hydroxybutyric acid Figure 15 BocNt-I Figure 16 3 Synthetic Methods 3.1 Enzymatic Resolution of Racemates Several methods for the synthesis of racemic p-amino acids have been developed These methods are often cheap and easy to perform but an extra step is essential to resolve the racemate as the fundamental requirement in the design of these compounds is to obtain amino derivatives with high optical purity In addition to the classical resolution of racemates utilising the formation of diastereoisomeric salts by reaction with enantiomerically pure amines enantiomerically pure p-amino acids have recently been prepared by routes involving the enzymatic resolution of racemates Hydrolytic enzymes are especially well suited for the kinetic res olution of racemic a-amino acid derivatives 32 This method has therefore found numerous industrial applications Derivatives such as esters amides or hydantoins are selectively hydrolysed The dif- ferent approaches are classified according to the bond cleaved by enzymatic assistance The major processes are amide or nitrile hydrolysis by aminopeptidases or nitrilases cleavage of N-acyl groups by acylases ester hydrolysis by lipases or proteases and cleavage of hydantoins by hydantoinases If after separation only one enantiomer is required the undesired one can be racemised and the reaction mixture can be separated again Despite the prominent role of the enzymatic resolution of race- mates in the production of a-amino acids far fewer results on its application for the preparation of enantiopure p-amino acids are available Indeed a disadvantage of the enzymatic methods is often the narrow tolerance of the substrate thus the different position of the chiral centre usually a to the carbonyl group can be a reason why the more popular methods for the enzymatic separation of a- amino acid derivatives fail when applied to p-amino acid deriva- tives Until now the most promising method appears to be the selective hydrolysis of N-phenylacetyl derivatives of p-amino acids with penicillin acylase (PA) Indeed several N phenylacetyl p-amino acids have been selectively hydro1 ysed with good results by two research groups 33 Both enantiomers can be obtained in good yield with a high degree of optical purity These advantage are CHEMICAL SOCIETY REVIEWS 1996 rarely experienced when resolution based upon the crystallisation of diastereoisomeric derivatives is used Examination of the data shows that for all substrates examined by both groups penicillin acylase preferentially hydrolyses the (S)-enantiomer (Scheme 1 Table 1) 0 0 Scheme 1 Table 1 Substrates examined and results for the hydrolysis experiments R [a],of (S) 2 Yield of (S) 2 (%p la],,of 2 (lit )” CHj20uC H Zh 2 5 +34 8 +38 0 78 75 +35 +38 5 Ph20a +7 5 88 -6 9 CH 20h +3X 3 77 +34 3 ph2Ob +6 5 - -6 3 CFj20h +27 6 86 C2F5*Ob f37 0 78 C,F,20b +26 9 84 2 FC,H420h +3 0 67 4 FC,H420h +3 9 73 Yield referred to theoretical 50% of the (R) isomer IS reported See refs 33(a)and 33(h) The [a],,value Penicillin acylase is used on an industrial scale for the production of 6 aminopenicillanic acid the starting material for the synthesis of semi-synthetic penicillins Furthermore it is well established that it shows a high degree of stereoselectivity in hydrolysing pheny- lacetyl esters and amides together with a low degree of substrate specificity The protein has a single-amino-acid catalytic centre 34 Its in vivo role remains unclear however and the observation that expression of the Eschenchia coli enzyme in vivo is regulated by both temperature and phenylacetic acid has prompted speculation that the enzyme could be involved in the assimilation of aromatic compounds as carbon sources in the organism’s free living mode The crystal structure of penicillin acylase has recently been reported and shows that the protein is kidney shaped in cross-section with a deep cup-shaped depression in the centre The structures of the enzyme complexed with phenylacetic acid (a competitive inhibitor) locate the binding site for the side chain of the substrate The phenyl moiety of the inhibitor points towards the interior of the protein into a mainly hydrophobic pocket which is lined with many aromatic residues and hydrophobic side chain The complementary fit explains the specificity of the enzyme towards the phenyl moiety of a broad range of substrates The proposed catalytic mechanism shows the role of the bridging water (as a virtual base) and the a amino group to enhance the nucleophilicity of serine B 1 3.2 Synthesis of @-Amino Acids and a-Substituted @-Amino Acids through the Intermediary Preparation of Perhydropyrimidin-4-ones Perhydropyrimidin 4-ones and dihydropyrimidin-4-ones are an interesting class of heterocyclic compounds which represent pro- tected forms of @-amino acids so that the a-and the @-position can be functionalised as desired Following a route previously described for the preparation of 2-tert-butylimidazolidin-4 one,’5 a useful starting material in the synthesis of a-branched a-amino acids Juaristi and Seebach envis- aged the preparation of 2-terf-butylperhydropyrimidin-4-onederiv atives starting from p-alanine N-methylamide and pivalaldehyde36 (Scheme 2) The perhydropyrimidin-4-one was alkylated with high trans-diastereoselectivity via the corresponding enolate (Scheme 3) The high diastereoselectivity observed is attributed to the steric ASYMMETRIC SYNTHESIS OF @-AMINO ACIDS AND p-SUBSTITUTED @-AMINO ACIDS-G CARDILLO ET AL Scheme 2 Bz=PhC( 0) Bi Bi Scheme 3 Table 2 Diastereoselectivity of lithium-enolate alkylations RX Diastereoselectivity(%) Isolated yield (%) CHJ 96 7 77 PhCHzBr 95 5 75 Bu"1 >96 0 76 n C,H,,I >96 0 76 CH,=CHCH,CI 86 0 78 hindrance generated by an axial disposition of the tert-butyl group at C-2 which directs addition to the enolate face opposite to this group The acidic hydrolysis of the heterocycle affords in quantita- tive yield the corresponding racemic a-alkylated p-amino acids (Table 2) Furthermore the enantiomerically pure (S)-2-tert-butylperhy- dropyrimidin-4-one was obtained by replacing p-alanine with (S)a~paragine,~~following a procedure described by Konopelsky 3* Indeed the condensation of (S)-asparagine with pivalaldehyde followed by protection of the amino group afforded the (2S,6S)- 1-benzoyl-2-tert-butyl-6-carboxyperhydropyrimidin-4-oneThe car- boxy side chain was eliminated under oxidative decarboxylation conditions and afforded the corresponding enone which was sub- sequently reduced via catalytic hydrogenation (Scheme 4) ,,,>' 1 KOWf-BuCHO 1{' " 1 n-BuLdMeI 452 BzCVNaHC03 2 H2 Pd/C ~ 3 Pb(OAch/ H2N C02H cu(0Ac)2 Bi Bi (9-Asparagme (s) Scheme 4 In order to synthesise enantiomerically pure 2-substituted 3-aminobutanoic acids 6-methylperhydropyrimidin-4-oneswere prepared starting from enantiomerically pure 3-aminobutanoic acid 39 (R)-and (S)-3-Aminobutanoic acids were obtained pure through separation of the diastereoisomers of the corresponding ( 1 'S)-N-phenylethyl derivatives In the formation of the corre-sponding 2-tert-butylperhydropyrimidin-4-ones from each stereoisomer the corresponding cis-products predominate (cis trans ratio 95 5) The subsequent alkylation [lithium diisopropylamide (LDA) RXJ furnishes the 5-substituted derivatives as a single stereoisomer Hydrolysis of these 5,6-dialkylperhydropyrimidin-4-ones leads to the corresponding a-substituted p-amino acids in the free form and high optical purity (Scheme 5) (S)-3-Amioobutanoic n Scheme 5 Konopelski following the theme of 'self-reproduction of chiral- ity' pioneered by Seebach prepared enantiomerically pure dihy- dropyrimidinones via pivalaldehyde acetalisation of the potassium salt of asparagine to form the corresponding pyrimidinone car- boxylate as a single stereoisomer3s (Scheme 6) Methoxycarbonyl I Pd(0Ach NH NH 2''OH ~t-BuCHO HNAN'co2Me MeOCAI 0u*"C02H 3 CIC(0)OMe -Ad ~(a-asparagme 4 Pb(OAc)4/ CU(OAC)~ " -Scheme 6 functionalisation at the secondary amine followed by oxidative decarboxylation with lead(1v) acetate delivered the corresponding dihydropyrimidin-4-one as a single enantiomer Then the hetero- cycle was submitted to palladium-catalysed conjugate addition of aryl iodide which afforded the desired product as a single stereoiso- mer Treatment of the adduct with NaBH,/H,O+ ,followed by acidic hydrolysis afforded (S)-P-tyrosine-0-methyl ether hydrochloride in high yield In order to obtain various enantiomerically pure p-amino acids the chiral dihydropyrimidin-4-one synthesized from aspartic acid was deprotonated at C-6 with tert-butyllithium at -78°C 40 The corresponding carbanion reacts with alkyl halides and aldehydes to give the corresponding alkylated products in high yield (Scheme 7 Table 3) 3 4 Scheme 7 Table 3 Chemical yield of the alkylation of (S)-2-dihydro- pyrimidin-4-one 3 at C-6 Electrophile Yield of 4 (70) CH,I 95 PhCH2Br 55 CH,CH,I 55 PhI 27 Reduction of the carbon-carbon double bond with NaBH,CN in the presence of hydrochloric acid afforded the corresponding 2,6-cis-perhydropyrimidin-4-one with good yield and high diastereoisomeric selectivity (Scheme 8) Demethylation of the -1 CH2=CHOCOCI HO NH,+NaBH3CN HCI de = >95% 0 R = Me PhCH2 Ph yield= 75% (CH&CHCH Scheme 8 R=Me PhCH Ph (CH,),CHCHz amine nitrogen and hydrolysis were combined in the last step to give the desired p-amino acids The heterocycle was first treated with vinyl chloroformate in refluxing 1,2-dichIoromethane 41 The resulting carbamate was then refluxed with ethanolic HCI to achieved dealkylation and furnished enantiomerically pure P-alkyl p-amino acids In recent years we have been interested In the electrophile- CHEMICAL SOCIETY REVIEWS 1996 I.VYL 3. flash chromatography ..,resin Scheme 9 Table 4 Diastereoselectivity and chemical yield for the alkylation of 6-methylperhydropyrimidin-4-ones 5 and 6 Substrate 77°C LiHMDS" Equiv. of agent Alkylating 5 0 0.9 EtI 5 0 1.1 EtI 5 -20 1 EtI 5 -78 to room temp. 1 EtI 6 0 0.8 EtI 6 0 1.1 EtI 6 -78 to room temp. 1 EtI 6 0 1 BulI 6 0 1 BnBr LiHMDS =lithium hexamethyldisilazide promoted cyclofunctionalisation of unsaturated substrates contain- ing an internal nucleophile associated with the use of (S)-1-phenylethylamine as a chiral This strategy affords diastereoisomeric mixtures of heterocycles which can easily be sep- arated by flash chromatography. Following this approach our first synthesis of the 6-methylperhydropyrimidin-4-onesstarted from the amide obtained from the reaction of hexahydrotriazine and but-3- enoyl chloride (Scheme 9). The adduct was recovered in quantita- tive yield and converted in situ into the corresponding amide under an atmosphere of ammonia; mercury(I1) trifluoroacetate [ Hg(TFA),] promoted the cyclisation of the benzyloxycarbonyl protected deriv- ative. The reduction of the carbon-mercury bond with LiBH deliv- ered the perhydropyrimidin-4-ones in 70% overall yield as a 63:37 diastereoisomeric mixture. The isomers were separated by flash chromatography and separately submitted to acidic hydrolysis affording respectively (S)-and (R)-3-aminobutanoic acid after purification on cationic exchange resin. The particular conformation which each heterocycle assumes was established by detailed H NMR studies. Indeed the chiral group on C-1 shows a strong ten- dency to assume a rigid conformation with the hydrogen eclipsing the carbonyl group; the results of NOE experiments performed on both isomers confirm the stereochemical assignment. The 6-methyl-perhydropyrimidin-4-ones (1'S,6R)-5 and (1 'S,6S)-6were submitted to alkylation by treatment of the corre- sponding lithium enolate with various alkyl iodides 33 The alkyla- tion always proceeded with high trans selectivity delivering useful intermediates to a-substituted p-amino acids as shown in Table 4. The diastereoisomeric ratio increases with the use of a bulkier electrophile. Furthermore both the trans cis ratio and the amount of dialkyl derivative increase if the reaction temperature decreases. Moreover the diastereoisomeric ratio and the amount of dialkyl derivative increase with increase in the proportion of the base. The results suggest that the dialkylation occurs faster on the cis than on the trans-5,6-disubstituted diastereoisomer. In order to analyse how the bulkiness of the substituent at C-6 affects the diastereoselectivity of the alklyation reaction we prepared Yield Dial kylated trans cis (%) (%I ratio 84.2 -87 13 78.3 11.4 96:4 89.3 3.1 91:9 85.8 9.7 94 6 76.8 -85:15 92.1 7.9 94:6 88.5 10.0 97:3 56.1 -87:13 89.8 -93:7 several perhydropyrimidinonesU (Scheme lo) following a well- known procedure described by Steiger on cinnamic acid. Racemic p-amino acids were obtained by addition of 2 equivalents of free hydroxylamine to alk-2-enoic acids in ethanol at reflux. The hydrox- 1. NH2OHEtOHLR2.ZClNaOH HO 40-55%yield * Ho 1. SOCI* Ph 0 NHZ 2. (S)-PhMeCHNH2 * Me 80-85%yield H Scheme 10 a; R=Et b; R=Pr" c; R=Ph ylamine behaves both as a nucleophile and a reducing agent for the 3-hydroxylamino intermediate. Following this method p-amino acids were obtained and protected at the nitrogen with benzyl chloroformate. The products were converted into the corresponding (S)-phenylethylamido derivatives. The corresponding perhydro-pyrimidin-4-ones ( 1'S,6R)-7 and ( 1'S,6S)-8were easily obtained by treatment with paraformaldehyde and toluene-p-sulfonic acid in benzene at reflux and fully separated by flash chromatography. At first we considered the methylation reaction because a-methyl p-amino acids are becoming of great interest as they are components of a number of molecules which display interesting biological properties. For instance (2S,3/?)-2-methyl-3-aminopen-tanoic acid has been isolated as a hydrolysis fragment of several ASYMMETRIC SYNTHESIS OF P-AMINO ACIDS AND P-SUBSTITUTED P-AMINO ACIDS-G. CARDILLO ET AL biologically interesting natural products such as the marine peptide antifungal agents majuscolamide CI7 and the antitumour agents dolastatins 11 and 12.1° The methylation at C-5 of perhydropyrim- idin-4-ones afforded increasing diastereoselectivity on going from ethyl to propyl to phenyl groups (Table 5). Table 7 Diastereoisomeric product ratios and chemical yields for the alkylation reaction of perhydropyrimidin-4-ones 7 and 8 Starting Alkylating transxis material agent Yield (%) ratio Table 5 Starting Diastereoselectivity and chemical yield for the methylation of perhydropyrimidin-4-ones 7 and 8 Alkylating Yield truns:cis Dialkylated 7a 7b 7c 7b EtI 78 EtI 90 EtI 96 PhCH,Br 85 96:4 >99 1 >99 1 >99 1 material TI" C agent (%) ratio (%) 8a PhCH,Br 95 >99 1 7a 7a 7b -20 Me,SO 80 78:22 -20 Me,SO 81 78:22 -78 to room temp. Me1 60 84:16 - 8b 8c 8b EtI 92 EtI 80 PhCH,Br 88 >99 1 >99 1 >99 1 7c -20 Me,SO 50 92:8 - 8a -20 Me,SO 80 82:18 7 8b -20 Me,SO 86 77:23 - 8c -20 Me,SO 50 92:8 - 1I 6 mol I-' HCI 04.,,82. ion exchange resin We then studied in detail the dialkylation at C-5 of the cis and trans-5-methyl-6-ethyl perhydropyrimidin-4-ones while the alky- lation of the lithium enolate of the 5,6-trans-product 9 with ethyl iodide affords the dialkyl derivatives in 5% overall yield the alky- lation of the 5,6-cis-derivative 10 produces a mixture of 5-dialky-lated pyrimidin-4-ones in quantitative yield and in 91:9 diastereoisomeric ratio (Scheme I I Table 6). This outcome estab- lishes that the relative configuration of C-5 and C-6 of the hetero- cycle strongly influences the rate of dialkylation and thus the diastereoisomeric ratio of the monoalkylation. Moreover the same alkylated compound was obtained as the major product starting from either the trans-9 or the cis-10 compound. The absolute con- figuration of the major isomer was established by means of NOE difference experiments and shows that the preferred isomer is obtained from the attack of the ethyl group from the same side of the ethyl at C-6. This new approach allows chiral a,a-disubstituted p-amino acids to be obtained. Table 6 Diastereomeric product ratios and chemical yields for the dialkylation reaction of perhydropyrimidin-4-ones 9 and 10 Starting Diastereoisomeric material TI"C tlh ratio 1la:llb Yield (%) 9 0 1 96:4 5 9 room temp. 72 90 10 47 10 0 1 91:9 99 In order to test how the bulkiness of the alkylating agent affects the diastereoisomeric ratio the (1'S,6R)-and (1 'S 6S)-6-ethyl 6-propyl and 6-phenyl perhydropyrimidin-4-ones 7a 7b 7c 8a 8b and 8c were submitted to alkylation with ethyl iodide and/or benzyl bromide. The results are reported in Table 7 and show that the reac- tion occurs in high yield and high diastereoselection in all the analysed cases. The hydrolysis of the heterocycles was performed with 6 moll -HCI at reflux for 24 hours and delivered the corresponding p-amino acid hydrochlorides with no racemisation (Scheme 12). The (S)-1-phenylethylamine was separated during the work-up and the purifi- cation of the amino acid from sodium chloride was performed on a 78-81%yield '0 R Me 8145% yield Scheme 12 column of cation exchange resin using NH,OH (1.5 mol 1-I) as eluent. Finally in order to show another possible use of a-substi- tuted p-amino acids some selected p-amino acids were submitted to cyclisation with 2-chloro- I -methylpyridinium iodide in the pres- ence of triethylamine and afforded the corresponding 3,4-trans-disubstituted p-lactams in high yield. 3.3 Synthesisof Enantiomerically Pure &Amino Acids by Conjugate Addition to a /%Unsaturated Esters and Imides Among the strategies for the synthesis of enantiomerically pure p-amino acids diastereoselective 1,4-additions of chiral amines to a,P-unsaturated esters and additions of amines to chiral a,p-unsat-urated acyl derivatives have attracted much interest35 The first approach has been studied by several research groups. The general method is to add a 'chiral ammonia,' usually an asym- metric secondary amine to a Michael acceptor generally an unsatu- rated tert-butyl ester. The reaction produces a new chiral centre p to the ester group and the diastereoselectivity depends on the ester on the 'chiral ammonia' utilised and on the reaction conditions. All the proposed methods are quite recent. One of the first good results was achieved in 1986 by Hawkins and coworkers,&" who obtained a 55% diastereoisomeric excess upon addition of a C symmetric chiral secondary amine such as a dihydroazepine. On the other hand a much better diastereoisomeric excess (97% de) was obtained utilising the lithium salt 12 of the same C symmetric chiral secondary amine as nucleophile (Scheme 14). Removal of the dimethyl binaphthyl moiety under reductive conditions gave the deprotected (R)-P-amino acids. In a recent development of this work,4h the addition of the (1'S,SR,6S)-lla (1'S,5R,6S)-10+ (1'S,SR,6S)-11b Scheme 11 CHEMICAL SOCIETY REVIEWS 1996 Scheme 13 q-8Ni 97%de LI 88% yield *OM,12 Scheme 14 lithium salt of the dihydroxyazepine to tert-butyl crotonate in dimethoxyethane at -63 "C followed by quenching with methyl iodide afforded the derivative of tert-butyl 3-amino-2-methylbu- tanoate as a 65 2 2 1 mixture of diastereoisomers (Scheme 15) The lithium amide and the electrophile add anti across the RZ'N 0 1 t-butylhglmte 1 t-butylcrotonate 2 NH4Cl 2 Me1 12 -'IBu * 638vield 71% yield I 254iOdr 65221dr Scheme 15 carbon4arbon double bond of the ester This result was confirmed by transforming the 3-amino-2-methyl ester into the corresponding trans-p-lactam by reduction with Pd(OH),/C of the dihy-droazepine esterification of the amino group and cyclisation with 2-chloro-1 -methylpyridinium iodide In order to obtain the syn isomer the lithium amide was added to tert-butyl tigliate in DME at -46 "C The addition product was obtained in 63% yield and a 25 4 1 0 ratio of diastereoisomers with the svn-isomer in excess Recently the lithium amide derived from (R)-N-(a-methylben-zy1)benzylamine (R)-13has been used by Davies and coworkers as a synthetic equivalent of ammonia in a Michael addition to benzyl roto on ate,^^ the addition occurs with 95%de and gives after deben- zylation the (-)-(R)-3-arninobutanoic acid (Scheme 16) Treatment of methyl p-benzyloxycinnamate with the same lithium amide generated the corresponding adduct as a single diastereoiso- R = Me CHzPh 'Bu R =%m %% woMe,de = 95%-299% OMe yield = 23% -88% Scheme 16 mer The hydrogenolysis in EtOH with Pd(OH) on charcoal afforded the corresponding (S)-p-amino acids in quantitative yield The method has been further developed and the change from the benzyl ester to the tert-butyl ester increased the selectivity Furthermore the enolate formed by the 1,4-addrtion was trapped with various electrophiles under diastereoisomeric control of both chiral centres Indeed the conjugate addition of lithium (R)-(a-methy1benzyl)benzylamide (R)-13to tert-butyl esters followed by in situ hydroxylation with (+)-(camphorsulfony1)oxaziridine 0 1 (R)-13 PhA40tBu (+)-14 provides the corresponding anti-p-amino a-hydroxy amino acid derivative with excellent diastereoselectivity (Scheme 17) The synthesis of enantiomerically pure 3-phenylisoserine derivatives48u has been of great interest in recent years particularly with respect to the synthesis of taxol,I5 a complex diterpene which exhibits strong antitumoudantileukaemic activity and is currently consid- ered a major lead in cancer chemotherapy Thus the hydroxylation of the intermediate enolate of the tert-butyl cinnamate proceeds in 86% yield and 92% de (Scheme 17) Debenzylation of the adduct occurs in quantitative yield affording the anti-compound In order to obtain the syn-relative stereochemistry required by the natural product an inversion of the C 2 centre was realised through treat- ment of the anti amino alcohol under Mitsunobu conditions in 80% yield Following the same protocol the 3-amino-2-hydroxydecanoic acid (AHDA) an unusual amino acid present in the angiotensin- converting enzyme inhibitor microginin,28 was obtained as both the (2R,3R)-anti and (2S,3R)-syn diastereoisomePS6 (Scheme 18) In contrast the synthesis of the unusual amino acid allophenyl- norstatine the crucial fragment responsible for the activity of kyno- statins 227 and 272,24 required some modifications48c (Scheme 19) In fact the tandem addition-hydroxylation reaction performed on the homologous ester of the tert-butyl cinnamate proceeded with little selectivity at the a-centre if the reaction was performed with the (R)-enantiomer of the lithium amide (R)-13as nucleophile and (+)-(camphorsulphony1)oxaziridine (+)-14 as electrophile This reaction repeated with the (S)-enantiomer of the lithium amide (S)-13,gave a 22 I selectivity in favour of the anti-compound thus showing that the poor selectivity observed with the (R)-amine was a consequence of using the 'mismatched' pairing For the synthesis of anti-a-alkyl p-amino acids alkyl halides were utilised to trap the enolate obtained by the Michael addi- tion3" (Scheme 20) In this case the diastereoisomeric excesses are less satisfactory and indicate that a variety of factors are respon- sible for determining the extent of diastereofacial discrimination shown by an incoming electrophile For this reason a detailed study on better conditions to obtain the anti- or the syn-a-substituted p-amino acids was undertaken To obtain the anti-derivative the Michael adduct was quenched worked-up and subsequently depro- tonated and treated with the required alkyl halide On the other hand the addition of (R)-(a-methy1benzyl)benzylamide(R)-13to a-alkyl a$-unsaturated tert-butyl esters followed by quenching with a proton source afforded the syn-a-alkyl p-amino acid Following the second method the conjugate addition of (R)-13 to tert-butyl (E)-2-methylpent-2-enoate was performed in toluene and the reaction mixture was diluted with THF prior to quenching the reaction with the hindered acid 2,6-di-tert-butylphenol With this procedure the (2S,3R)-2 methyl-3-aminopentanoic a component of majusculamide C,I7 was generated in good yield and >91 1 syn anti selectivity after reduction of the amine moiety and acid hydrolysis of the ester (Scheme 21) Concerning the stereoselective conjugate addition of lithium (a-methylbenzy1)benzylamideto tert-butyl cinnamate a theoretical model has been proposed using molecular modelling techniques so The authors carried out calculations on a$?-unsaturated systems in the s cis conformation in accord with recent experimental findings Initial calculations were associated with modelling the approach of the achiral nucleophile dibenzylamide to the 3-Sr face of tert-butyl cinnamate assuming that the approach trajectory of the amide would be oriented in the plane perpendicular to the a,P-saturated system and that lithium chelation is operative When the amide is Ph PHCONH 0 -PhvObu 6H OH (2R,3R,&)86%veld 92%de Scheme 17 ASYMMETRIC SYNTHESIS OF P-AMINO ACIDS AND P-SUBSTITUTED P-AMINO ACIDS -G. CARDILLO ET AL 1. (R)-13 D 2.(+)-14 88% de Scheme 18 63% yield 41% yield 91% de 20% de Scheme 19 Ph 1. LDA ph 2. R"X R'Uobu -1. (R>132. proton source Ph wy0lSu sy" R' Scheme 20 near the system the conformation is similar to a butterfly placing the phenyl rings approximately parallel to each other (Scheme 22). Then the conformation of lithium (a-methy1benzyl)benzylamide was considered in the proximity of the electrophile. Although it is unreasonableto assume that the butterfly conformation is necessar- ily the global minimum for the transition states the authors com- pared the four possible butterfly conformations A B C and D. Conformation A has the lowest energy and alone maintains the but- terfly conformation; thus sterically demanding methyl group at 1. (R> 13toluene 2.TH.F 3. 2,6-Di-tert-butylphenol >95% de 65% yield position A can be tolerated without disrupting the butterfly arrange- ment. The structure associated with A suggests that in the tandem addi- tion-enolate quench methodology large electrophiles will be attacked by the enolate generated from a,P-unsaturated esters and amides anti to the amide moiety if the chelated structure remains intact. This hypothesis was experimentally verified by the authors .49a In a conceptually related system to the Davies work Yamamoto and coworkers achieved the synthesis of a p-lactam framework via a three-component coupling reaction.5Ia The key step of this syn- thetic method is the conjugate addition of lithium (R)-(a-methyl- benzy1)benzylamide to an a,P y,s-diunsaturated tert-butyl ester followed by quenching of the resulting enolate and deprotonation of the @amino ester followed by aldol condensation with acetalde- hyde. So three new chiral centre are obtained with good stereocon- trol (Scheme 23). This three-component coupling approach has been applied to the synthesis of a 1P-methylcarbapenem key Indeed the conjugate addition of N-benzyl-N-[(R)-1-phenylethylJamine to (R)-(a-tert-butyl 5-[(tert-butyldimethylsilyl)oxy 1-4-methylpent-2-enoate produces the (3S,4R)-syn-adduct with essentially 100% de in 84% yield where the addition of the (9-amine to the same sub- strate afforded the (3S,4R)-syn-adduct with essentially 100%de in 95% yield. c1-2.HCI (aq) 57% yield Scheme 21 Ph Ph A B C DEnergy (kJmol-') o +11 +12 +17 Scheme 22 CHEMICAL SOCIETY REVIEWS 1996 82%de 89%yield Scheme 23 R1 2. Si02 Li -R1 Rz 3. SiO 4. RaNi-H2 15 3. Ra/Ni-Hz M =90 -98% Scheme 24 In a similar way Enders and coworkers have recently devel- oped a diastereo- and enantio-selective synthesis of a-alkyl/aryl p-amino acids by tandem 1,4-addition/a-alkylation?2 The 'chiral ammonia' utilised by these authors is (S)-(-)-2-methoxymeth y 1- 1 -trimeth ylsilylaminopyrrolidine (TMS-SAMP) readily prepared from the well established auxiliary (S)-( -)-1 -amino-2-methoxymethylpyrrolidone(SAMP) by met- allation with n-butyllithium and N-silylation with chloromethyl- silane in 95% yield. By adding the lithium salt of TMS-SAMP to various methyl and tert-butyl a,P-unsaturated esters the corre- sponding (S,S)-3-amino ester derivatives were obtained with 93-98% de and 3247% yield and readily transformed into the corresponding (S)-p-amino acids by reduction of the SAMP moiety with Raney Ni/H (Scheme 24 Table 8).If the addition of the lithium salt of TMS-SAMP was followed by treatment of the reaction mixture with HMPA (3 equiv.) and the appropriate alkyl halide both the a-and the P-position were functionalised with diastereoisomeric excesses ranging from 63 to >96% in the favour of the anti-stereoisomer (Scheme 25). The second approach to enantiomerically pure p-amino acids is a result of the addition of an achiral amine to chiral acyl derivatives such as chiral imides or esters. Following this strategy d' Angelo and Maddaluno achieved very high stereocontrol in the conjugate addition of amines to chiral ap-ethylenic esters using high pressure as an activating condition to produce p-amino esters with good chemical yield and variable de ranging from 5 to >99% depending on the reaction conditions53 (Scheme 26). The authors attributed this stereochemical outcome to the 'T-stacking' model in which the aryl group of the inductor shields a face of the crotonate unit thereby directing the amine addi- tion to the other face. Our interest in this field is represented by the addition of O-ben- zylhydroxylamine to an a,P-unsaturated system obtained by reac- tion of various a$-unsaturated acyl chlorides with (4S3R)-l~-dimethyl-4-phenylimidazolidin-2-onein the presence of a Lewis acid5 (Scheme 27). The role of the Lewis acid is both to chelate the two carbonyls in order to produce a rigid chiral unsatu- rated system which reacts with the nucleophile on a preferred face and to enhance the reactivity of the electrophile imide. Indeed when the reaction was performed in the absence of a Lewis acid no addi- tion product was observed. Various Lewis acids were utilised as chelating agents and the best results were obtained with TiCl and AlMe,CI (Table 9). In particular the addition of O-benzylhydroxyl- amine catalysed by 2 equivalents of AIMe,CI furnished the @-amino derivative in 89 11 diastereoisomeric ratio and 83% yield. The nucleophilic attack on the aluminium-chelated crotonyl imide Table 8 Chemical yield and diastereoisomeric ratio of the tandem-addition of TMS-SAMP-Li and alkyl halides to tert-butyl ester 15 R RZ Yield (%) De (%) R R* Yield (96) De (%) CZHS 'ZH5 CA n-C,H CH3 C,HS 56 48 53 77 66 68 C2H5 n-C,H 33 65 C*HS PhCH 26 76 C2H.5 n-C I I H23 (CH,),COH 'zH5 51 41 >96 63 4 0 0 + ~PEHZ-NH~ MeOH 15 kbar 25 "C,24 h R= Scheme 26 Ph (35 34 >96 Ph n-C,H 68 >96 Ph n-C,H 64 >96 Ph ally1 52 >% Ph PhCH 67 >% Ph PhCH 48 >% p-MeOC,H CH 53 >96 Scheme 25 derivative would preferably occur from the Cp-Re face of the s-cis conformation resulting in the formation of the (4S,5R,3'R)isomer as PhCHzNH the major product while a complete inversion of selectivity was aoRobserved when TiCI was utilised as Lewis acid. In order to demon- 5096 yield 29996de strate the formation of a chelate system and to rationalise the experi- mental results we compared the IH NMR spectrum of the starting material with the spectra of its complexes with AlMe,Cl(2 equiv.) and TiCI (1 equiv.). The 'H NMR spectra of the complexes identical at -60°C and at room temperature suggest that both metals strongly chelate the two carbonyls of the imide. The inversion of selectivity ASYMMETRIC SYNTHESIS OF @AMINO ACIDS AND P-SUBSTITUTED P-AMINO ACIDS-G CARDILLO ET AL (4S,SR)-16 (4S,5R,3'S) 1 ZdCu(OAc)22. LIOWH202u-I 3. ion exchange resin HO (-)-(fl Scheme 27 0 0 0 N-Phthal a:R=H b:R=Me R 0 -HPh C R = R-Pr d:R=Ph Me Lewis acid Me 4O-W%de 3040% yieldI ~ -N iorK=r1 c 8:l:l:O dr benzenesulfonylbromide Mn' 98%yieldI.."(15 quiv ) dry THF Scheme 28 Table 9 Ratios of the diastereomeric products of the 1,4-addition of benzylhydroxylamine to imide 16 Yield Diastereoisomeric (%) M(L) (equiv 1 (%I (4S5R,3'S) (4S3R,3 'R) ZnCI (1) 92 55 45 TiCl (1) 80 80 20 TiCI (1) 90 77 23 TiCI,(OPri) (1) 38 50 50 AlCl(Me) (1) 65 26 74 AICl(Me) (1 4) 65 19 81 AICI(Me) (2) 70 20 80 AICI(Me) (I 4) 81 13 87 AICl(Me) (2) 83 11 89 could be ascribed to the difference in the metal-oxygen bond length and bond angles between the titanium and aluminium complexes Furthermore the ( +)-(3s)-butanoic acid was obtained from the 0-benzylhydroxylamino adduct by reduction of the N-0 bond with Zn/Cu in acetic acid and subsequent hydrolysis of imidazo-lidin-2-one with LiOH/H,O in THF/water In a development of this work the diastereoselective functional- isation of both the a-and the P-position was analysed with the addi- tion of chloromagnesium phthalimideS5 (Scheme 28) The reaction performed with 5 equivalents of nucleophile afforded the P-phthal- imido derivative with a 95 5 diastereoisomeric ratio and 90%yield Furthermore the resulting enolate was trapped by performing the reaction in the presence of benzenesulfonyl bromide and the syn-2-bromo-3-phthalimido derivative was obtained in good yield and high diastereoselectivity and then transformed into the correspond- ing anti-2-azido-3-phthalimidoderivative by displacement of the bromide with sodium azide 4 References 1 (a)Chemistry and Biochemistry of the Amino Acids ed G C Barrett Chapman and Hall London 1985 (6) J H Jones (Sen Reporter) Amino Acids and Peptides Specialist Periodical Report The Royal Society of Chemistry 1992,Vol 23 2 (a) R M Williams Synthesis of Optically Active a Amino Acids Pergamon Oxford 1989 (6) R 0 Duthaler Tetrahedron 1994 50 1539 (c)R M Williams Aldrrchimica Acta 1992,25(1) 11 3 Y Izumi I Chibata and T Itoh,Angew Chem ,Int Ed Engl ,1978,17 176 4 J P Michael and G Pattenden Angew Chem ,1nt Ed Engl 1993,32 1 5 K C Nicolaou,W -M DaiandR K Guy,Angew Chem Int Ed Engl 1994,33,15,and references therein 6 H Umezawa TAoyagi H Suda M Hamada and T Takeuchi J Antibiot 1976,29,97 7 S Shinagawa,T Kanamaru S Harada M Asai and H Okazaki J Med Chem 1987,30,1458and references therein 8 (a) D J Hart and D C Ha Chem Rev 1989 89 1447 (b) The Organic Chemistry of @-Lactamsed G I George VCH Publishers New York 1992 9 A V Delgado-Escueta A A Ward Jr ,D M Woodbury and R J Porter Basic Mechanisms of the Epilepsies Raven Press New York 1986 p 365 10 G R Pettit Y Kamano H Kizu C Dufresne C L Herald R J Bontems J M Schmidt F E Boettner and R A Nieman Heterocycles 1989,28,553 11 V J Hruby F A1 Obeidi and W Kazmierski Biochem J ,1990,268,249 12 M Chorev and M Goodman Acc Chem Res 1993,26,266 13 E Altmann K Nebel and M Mutter Helv Chim Acta 1991,74,800 14 J Brunner Chem Soc Rev 1993,183 15 Seeforexample (a)E C Roos,M C Lopez,M A Brook,H Hiemstra and W N Speckamp J Org Chem ,1993,58,3259,(b)P J Colson and L S Hegedous J Org Chem ,1993,58,5918 16 (a) E Jauristi D Quintana and J Escalante Aldrichim Acta 1994 27(l) 3 (6) D C Cole Tetrahedron 1994,50,9517 17 D C Carter,E R Moore J S Mynderse W D Niemczuraand J S Todd J Org Chem 1984,49,236 18 (a) P Crews L V Manes and M Boehler Tetrahedron Lett 1986,27 2797 (b)T M Zabriskie,J A Klocke,C M Ireland,A H Marcus,T F Molinski D J Faulkner C Xu and J C Clardy J Am Chem Soc 1986,108,3123 19 (a) J Jiang K K Schumacher M M Joulle F A Davis and R E Reddy Tetrahedron Lett 1994,35,2121 (b)S Kosemura TOgawa and K Totsuka Tetrahedron Lett ,1993,34,1291 20 H Yoshioka K Nakatsu M Sat0 and T Tatsuno Chem Lerr 1973 1319 21 J Rodriguez R Fernandez E Quinoa R Riguera C Debitus and P Bouchet Tetrahedron Lett 1994,35,9239 22 E Dilip de Silva D E Williams R J Andersen H Klix C F B Holmes and T M Allen Tetrahedron Lett 1992,33,1992 23 E Umemura T Tsuchiya and S Urnezawa J Anribior 1988,41,530 24 T Mimoto J Imai S Kisanuki H Enomoto N Hattori K Akaji and Y Kiso Chem Pharm Bull 1992,40,2251 25 F Itagaki H Shigemori M Ishibashi T Nakamura T Sasaki and J Kobayashi J Org Chem 1992,57,5540 26 G L Helms R E Moore W P Niemczura,G M L Patterson K B Tomer and M L Gross J Org Chem 1988,53,1298 27 C A Bewley C Debitus and D J Falkner J Am Chem Soc 1994 116,7631 28 T. Okino H. Matsuda M. Murakami and K. Yamaguchi Tetrahedron Lett. 1993,34,501. 29 K.Iizuka T. Kamijo H. Harada K. Akahane T. Kubota H. Umeyama and Y. Kiso J. Chem. Soc. Chem. Commun. 1989,1678. 30 D. S. Dhanoa W. H. Parsons W. J. Greenlee and A. A. Patchett Tetrahedron Lett. 1992,33,1725. 31 Topics in Medicinal Chemistry ed. P. R. Leeming Royal Society of Chemistry Cambridge 1987,p. 101. 32 See for example (a) K. Faber; Bio-transformations in Organic Synthesis Springer-Verlag 1992 pp. 40-48;(6) R. 0.Duthaler Tetrahedron Report No. 349 Tetrahedron 1994,50,1539. 33 D. Rossi,G. Lucente and A. Romeo Experientia 1977,33,1557;(b) V. A. Soloshonok V. K. Svedas V. P. Kukhar A. G. Kirilenko A. V. Rybakova V. A. Solodenko N. A. Fokina 0.V. Kogut I. Y. Galaev E. V. Kozlova I. P. Shishkina and S. V. Galushko Synfett 1993,339. 34 H.J. Duggleby S. P. Tolley C. P. Hill E. J. Dodson G. Dodson and P. C. E. Moody Nature 1995,373,264. 35 (a) D. Seebach R. Imwinkelried and T. Weber in Modern Synthetic Methods 1986,ed. R. Scheffold Springer-Verlag Berlin 1986;vol. 4; (b) R. Fitzi and D. Seebach Tetrahedron 1988,44,5277. 36 E.Jauristi D.Quintana B. Lamatsch and D. Seebach J. Org. Chem. 1991,56,2553. 37 E. Jauristi and D. Quintana Tetrahedron Asymmetry 1992,3,723. 38 (a)J. P. Konopelski K. S. Chu and G. R. Negrete J. Org. Chem. 1991 56,1355;(b) K. S.Chu G. R. Negrete J. P. Konopelski F. J. Lakner N. -T. Woo and M. M. Olmstead J.Am. Chem. SOC. 1992,114,1800;(c) F. J. Lakner K. S. Chu E. R. Negrete and J. R. Konopelski Org. Synth. 1995,73,201. 39 E.Jauristi J. Escalante B. Lamatsch and D. Seebach J. Org. Chem. 1992,57,2396. 40 K. S.Chu and J. P. Konopelski Tetrahedron 1993,49,9183. 41 R. A. Olofson R. C. Schnur L. Bunes and J. P. Pepe Tetrahedron Lett. 1977,1567. 42 (a)R. Amoroso G. Cardillo C. Tomasini and P.TortoretoJ. Org.Chem. 1992,57 1082; (6) R. Amoroso G. Cardillo and C. Tomasini Heterocycles 1992,34,349. CHEMICAL SOCIETY REVIEWS 1996 43 R. Amoroso G. Cardillo and C. Tomasini Tetrahedron Lett. 1992,33 2725. 44 (a) I. Braschi G. Cardillo C. Tomasini and R. Venezia J. Org. Chem. 1994 59,7292. For other publications regarding the derivatisation of chiral 6-alkylperhydropyrimidin-4-onessee (6) R. Amoroso G. Cardillo G. Mobbili and C. Tomasini Tetrahedron Asymmetry 1993 4,2241;(c) I. Braschi G. Cardillo and C. Tomasini Tetrahedron 1994 50,2955. 45 J. d’Angelo D. Desmaele F. Dumas and A. Guingant Tetrahedron Asymmetry 1992,3,459. 46 (a)J. M. Hawkins and G. C. Fu,J. Org. Chem. 198631,2820;(b) J. M. Hawkins and T. A. Lewis J. Org. Chem. 1994,59,649. 47 S. G. Davies and 0.Ichihara Tetrahedron Asymmetry 1991 ,2 183. 48 (a) M.E.Bunnage S. G. Davies and C. J. Goodwin J. Chem. SOC. Perkin Trans.1 1993 1375;(b) M.E.Bunnage A. J. Burke S. G. Davies and C. J. Goodwin Tetrahedron Asymmetry 1994,5,203;(c) M. E.Bunnage S. G. Davies and C. J. Goodwin Synfett 1993,731. 49 (a) S.G.Davies and I. A. S. Waiters J. Chem. SOC.,Perkin Trans 1 1994,1129;(b) S.G.Davies 0.Ichihara and I. A. S. Walters J. Chem. SOC.,Perkin Trans. I 1994,1141;(c)S. G.Davies 0.Ichihara and 1. A. S. Walters Synfett 1994 117. 50 J. F. Costello S. G. Davies and 0.Ichihara Tetrahedron Asymmetry 1994,5,1999. 51 (a) N.Asao N. Tsukada and Y. Yamamoto J. Chem. SOC.,Chem. Commun. 1993,1660;(b) N.Tsukada T. Shimada Y. S. Gyoung N. Asao and Y. Yamamoto J. Org. Chem. 1995,60,143. 52 (a) D. Enders W. Bettray G. Raabe and J. Runsink Synthesis 1994 1322;(b) D.Enders H. Wahl and W. Bettray Angew. Chem. Int. Ed. Engl. 1995,34,455. 53 J. d’Angelo and J. Maddaluno J.Am. Chem. SOC.,1986,108,8112. 54 R. Amoroso G. Cardillo P. Sabatino C. Tomasini and A. Trerk J. Org. Chem. 1993,58,5615. 55 G. Cardillo A. De Simone L. Gentilucci P. Sabatino and C. Tomasini Tetrahedron Lett.,1994,35,5051 .

ISSN:0306-0012    

DOI:10.1039/CS9962500117

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Varying Resonance Demand in Carbocationic Systems Yuho Tsuno and Mizue Fujio Institute for Fundamental Research of Organic Chemistry Kyushu University Hakozaki Fukuoka 812 Japan 1 Introduction In a variety of aromatic side-chain reactions exemplified by benzylic solvolyses where a positive charge which can be delocalized by the .rr-system of the ring is generated (Scheme I) the Brown equation (1) has been widely applied to correlate the substituent effects.' Scheme 1 Solvolysis of benzylic precursors. log (klk,,)=p+u+ (1) In eqn. (1) k (or K) is the rate (or equilibrium) constant for such a reaction of a ring-substituted derivative and k or KO is the corre- sponding value for the unsubstituted compound. The electrophilic substituent constants u+,were defined using the solvolysis rates of a,a-dimethylbenzyl (a-cumenyl) chlorides (l),R =R,=CH in Scheme 1 in 90% aqueous acetone (90A) at 25 "C. For this reac- tion the reaction constant p+ = -4.54. The substituent effects in such systems can be more generally described by the Yukawa-Tsuno equation (2),2 where d'is the normal substituent constant which does not involve any additional ?r-electronic interaction between the substituent and the reaction centre and A@+,is the resonance substituent constant measuring the capability for T-delocalization of the T-electron donor substituent and is defined by cr+-d'. The r value is a para- meter characteristic for the given reaction measuring the extent of resonance demand i.e. the degree of resonance interaction between the aryl group and the reaction site in the rate-determining transi- tion state. Since our proposal of this equation the original form Yuho Tsuno studied chemistry in Osaka University. He obtained his diploma under the supervision of Professor Y. Yukawa from the Institute of Scientijc and Industrial Research Osaka University in 1959. He then carried out post-doctoral work with Professor J. E. Lefler at Florida State University from 1959-61 and with Professor R. W. Taft at Pennsylvania State University from 1961-1 963 before returning to the Institute of Scientijc and Industrial Research Osaku University. In 1971 he moved as professor of physical organic chemistry to the Department of Chemistry in the Faculty of Science of Kyushu University. From 1989-1 991 he was a professor at the Institute for Molecular Science Okazaki. He was a recipient of the Japanese Chemical Socieh Award in 1990. In 1993 he was appointed as pro- fessor and director of the Institute for Fundamental Research of Organic Chemistry at Kyushu Universiv. He retired from Kyushu University in 1995 and is presently a professor of Chemistry at Kumamoto University. 129 using cr instead of a0 in eqn. (2) has been widely applied.*.? However we believe that the use of eqn. (2) which is a theoreti- cally more justifiable form is When r=O log(klk,)=p@ while when r= 1.OO eqn. (1) is obtained. This mod- ification of the parameter scale does not affect the original meaning and the applicability of the equation. With the Yukawa-Tsuno (Y-T) equation (2) we introduced the concept of varying resonance demand of reactions into the analysis of substituent effects. In the general application of this equation the r value changes widely with the reaction. Its value is not limited to lower than unity as found for the cumenyl system (O<r< I) but in many cases it is significantly higher than unity (r>l).This fact clearly indicates that the Brown u+scale [i.e.,eqn. (2) when r= 1-01 does not reflect the highest extreme of benzylic resonance exalta- tion but is merely a single point on the r scale. This r scale permits evaluation of the nature of the transition state and has been widely applied to the assignment and interpretation of reaction mech- anisms,'-S involving benzylic solv~lyses~~~ and neighbouring aryl- assisted reactions,"-'O with which we will be mainly concerned in this review. The application and generality of the Yukawa-Tsuno relationship were adequately and precisely reviewed by Shorter? and further developments of the Y-T analysis were reviewed both by Johnson4 and recently by US.^ In the extensive application of eqn. (2) the correlation results were always compared with those of the Brown equation. It was pointed out that the additional r term in eqn. (2) is either necessary or superflu~us.~-~ More critical analyses cast doubt on the real merit of the r parameter and argue against its significance as a measure of the resonance demand.4 Hence this point will be also considered in this review. The difficulties encountered in the analysis of substituent effects in solvolyses as a mechanistic probe mostly arise from the mech- anistic involvement of the sol~ent.~-~ Consequently the behaviour of benzylic carbocations in the gas phase should be the best model Mizue Fujio graduated from the Faculty of Pharmaceutical Science Osaka University and received her M. Phurm. Sci.from the same uni- versity in 1967. She studied physical organic chemistry under the supervision of Professors Yukawa and Tsuno at the Institute of Scient$c and Industrial Rebearch and received her Ph.D. from Osaku University in 1972. She joined Professor Tsuno's group at the Department of Chemistry in the Faculty of Science at Kyushu University in 1973. From 197840 she worked as a post-doctoral research associate with Professor R. W. Taft at the University of California at lrvine. Her research interests have been in substituent and solvent efsects mainly in solvolytic reactions. She is also actively researching ion-pair mechanisms in solvolytic reactions. From 1993 following the establishment of the Institute for Fundamental Research of Organic Chemistry she has been an associate profes- sor of physical organic chemistry in that institute. CHEMICAL SOCIETY REVIEWS 1996 The low since it gives an r value significantly lower than unity whereasfor the behaviour of the solvolysis intermediate in s~lution.~ intrinsic substituent effect on the benzylic cation stabilities in the gas formation of highly electron-deficient carbocation systems such as phase have also been analysed by eqn. (2) and will be compared those formed in the solvolysis of 1-aryl-1-(trifluoromethyl)ethyl here with the substituent effects on the benzylic solvolysis rea~tion.~ tosylates (R =CH R,=CF in Scheme 1) should model reactions In our opinion this will provide convincing evidence for the concept having distinctly higher resonance demand. Both systems will be of varying resonance demand in solvolysis. Finally we will analyse shown to be far beyond the correlative ability of the Brown p+u+ the mechanisms of a series of benzylic solvolysis reactions by using equation but to be well correlated in terms of the Y-T equation. the concept of a continuous spectrum of varying resonance demands. The results of the analysis of the varying resonance demand in solvolyses of a series of substrates are summarized in Table 1. The precision of the fit to eqn. (2) is generally found to have a standard 2 The Yukawa-Tsuno Analysis of Solvolyses deviation of 20.04-0.10 in log klk depending on the magnitude Most of the criticism of eqn. (2) appears to arise primarily from the of the p value. This is comparable to the standard deviations found relatively small change of the r value in most benzylic solv~lyses.~~~ for the rneta-correlation with a single u values set and also with A definite answer to such criticism will be provided by exploring similar standard deviations found in correlations when the solvent the substituent effects on systems whose r values differ significantly or the leaving group was changed. Thus the precision index of from unity. 0.015-0.020 in u units is taken as an appropriate reference level The P-aryl-assisted solvolysisp- lo e.g. of neophyl brosylates of acceptable conformity in the Y-T equation. (2) where R =R,=CH and R,=R,=H in Scheme 2 should be an appropriate model reaction for cases where the resonance demand is 2.1 Aryl-assisted Solvolyses The substituent effect on the acetolysis of neophyl brosylates 2 gives a reasonably good correlation with u+rather than with ucon-stants and this was taken as evidence for the anchimeric assistance by the P-aryl group.6 It was suggested that the Y-T equation results in a better correlation than the simple Brown p+u+ treatment and this has been corroborated recently by the application of an exten- sive data set? The behaviour of substituents in this reaction is illustrated by the so-called Y-T plot in Fig. 1. A linear relationship for meta and T- accepting para substituents covering four orders of magnitude in reactivity is wide enough to define a p correlation as a rigid refer- ence common to both the cr+ and the Y-T analyses. The u+plots Scheme 2 Solvolytic process of P-arylalkyl precursors. (open circles) of para T-donor substituents consistently deviate Table 1 Yukawa-Tsuno correlations for benzylic solvolyses No. System (Soh. Temp) P r R SD n 1 Neophyl OBs (AcOH 75 "C) -3.83 0.57 0.9997 0.038 29 2 2-Arylethyl OTs (AcOH 115 "C) -3 87 0.63 0.035 27 3 l-Aryl-2-propyl OTs (AcOH 100 "C) -3.53 0.54 0.025 21 4 threo-3-Aryl-2-butyl OBs (AcOH 75 "C) -3.32 0.56 0.042 24 5 22-Bisarylethyl OTs (AcOH 90.1 "C) -4.44 0.53 0.998 0.077 16 6 7 2-Ar-2-(m-CIC6HJethyI OTs (AcOH 90.1 "C) 2-Ar-2-(35-C1,C6H3)ethylOTs (AcOH 90.1 "C) -3.60 -3.54 0.62 0.66 0.9994 0.9996 0.037 0.028 14 13 8 (1-Arylcyclobutyl)CH OBs (AcOH 55 "C) -3.27 0.55 0.050 17 9 Benzyl OTs (SOA 25 "C) X24-MeS,3-CNa -5.23 1.29 099.5 0.040 17 9a (AcOH 25 "C) Xap-halogens" -5.23 1.29 0.9994 0.044 14 10 a-Me-benzyl C1 (SOA 45 "C) -4.97 1.15 0.9993 0.06 25 1 Oa (97T 45 "C) -6.47 1.11 0.998 0.12 21 11 a-Bu%enzyl OTs (SOA 25 "C) -5.54 1.09 09997 0.060 31 12 2,2-Me2indanyl C1 (SOA 25 "C) -5.81 1.14 0.9995 0.1 1 11 13 a,a-Me,-benzyl C1(90A 25 "C) -4.59 1.oo 14 Ar(Ph)CH-Cl (EtOH 25 "C) -4.09 1.10 0.9991 0.092 22 14a (85A 0 "C) -4.58 1.15 0.9983 0.10 20 15 Ar(m-ClC,H,)CH CI (EtOH 25 "C) -4.47 1.18 0.9993 0.083 18 15a (85A 0 "C) -4.93 1.23 0.9982 0.1 1 18 16 Ar(3,5-CI2C,H,)CH C1 (EtOH 25 "C) -4.37 1.38 0.998 0.15 6 17 18 19 Arb-NO,C,HJCH C1(85A 0 "C) Ar,COH (pK,',H,O-H,SOJ a,a-(Prl),-benzyl C1 (SOA 45 "C) -5.31 -11.3913 -4.88 1.32 0.76 1.01 0.9975 0.999 0.9995 0.14 0.4013 0.1 1 14 11 14 20 a-Et ,a-Me-benzyl C1(8OA 45 "C) -4.69 1.04 0.9993 0.079 17 21 a-But,a-Me-benzyl C1 (SOA 45 "C) -4.28 0.91 0.9986 0.088 22 22 a-Bu',a-Prl-benzyl OPNB (50E 75 "C) -3.08 0.68 0.998 0.059 12 23 a,a-But,benzyl OPNB (50E 75 "C) -2.19 0.26 0.994 0.07 12 24 (1 -Arylcyclopropyl)CH OTs (80E 25 "C) -1.55 0.1 1 0.996 0.041 8 25 26 27 Ar(Ph)CCF OTs (SOE 25 "C) BBCO OTP (80E 75 "C) Ar,CCF OTs (80E 25 "C) X==3,4-Me2" X S3 ,4-MeZu -2.17 -4.18 -6.1 -4.33 0.01 1.19 1.45 1.26 0.989 0.999 0996 0.999 0.008 0.12 0.15 0.074 7 12 6 12 28 29 30 3 1 Ar(m-ClC,H,)CCF OTs (SOE 25 "C) X2p-Me' Ar(3,5-CI,C6H,)CCF OTs (SOE 25 "C) X3HQ a-CF,benzyl OTs (50E 25 "C) a-Me,cu-CF,-benzyl OTs (80E 25 "C) -6.19 -5.95 -6.05 -6.29 1.57 1.69 1.53 1.39 0.996 0.998 0.9994 0.9998 0.19 0.18 0.082 0.070 7 12 15 28 XSY or XGY correlation for the range of substituents more electron-donating or electron-attracting than Y. 4-Methylbenzobicyclo(2.2.2 locten-1 -yl triflates. VARYING RESONANCE DEMAND IN CARBOCATIONIC SYSTEMS-Y 3 2 1 131TSUNO ETAL 20 15 3 n 20 z 10 P,0 P,-0 -1 3 CN 4 MeSmMesL 05 -2 \ 00 -3 -05I -1 0 -05 00 05 10 (3-scale Figure 1 The Y-T plot for acetolysis of neophyl brosylates (2) at 75 "C 0 plots against (T+ 0,against @,El,against substituent parameter scale 3 given by the Y-T equation with r=O57 .,for plots of substituents having invariant (T for the three scales The figure is reproduced wlth per mission from Bulf Chem Soc Jpn 1990,63,1121 downwards from the reference meta substituents correlation line The line segments between u+and 00 values for para n-donor sub- stituents reflect the resonance capabilities of these substituents 1 e the A@'-,values The p correlation line divides all the line segments at a constant internal ratio corresponding to the r value of the system Thus the Y-T correlation line can be defined as a unique line intersecting all line segments for para n-donors at such a con- stant ratio I e ,at a substituent parametera scale with a characteris- tic r value even without the need to use the meta substituents The solvolysis of 2 is considered to proceed through a rate-deter- mining aryl-assisted transition state which leads to the tertiary car- benium ion The substituent effect on the rate should reflect only that on the aryl assisted ionization step and therefore the observed r value of 0 57 should be characteristic of this step This exalted r value can be rationalized in terms of a direct n-interaction between the aryl n-system and the P-carbocation centre at the rate-deter- mining transition state Most solvol yses of P-arylalkyl arenesulfonates involve a mech- anistic complication arising from a concurrent solvent-assisted (k,) process as shown in Scheme 2 The measured rate constant k of the solvolysis should be a sum of the rate constants for aryl-assisted k and unassisted k processes In practice substituent effects on these competitive processes can be directly analysed by using a non-linear least squares method based on the assumption of inde- pendent Y-T correlations for both k and k processes i e eqn (3) kt =kH]OP~('~+'~~~'K)+kll~OP(~~ (3)Ll The treatment was simplified by applying r,=O for the k process of this reaction However it is instructive to apply a classical method of analysis The Y-T plot for the acetolysis of P-arylethyl tosylates 3,all R'sbeing H in Scheme 2 is shown in Fig 2 In this reaction only compounds carrying strongly electron-donating substituents react predominantly by the aryl assisted k process Clearly we see the same pattern of the Y-T plot for these reactive substrates which react predominantly by the k process as for the acetolysis of 2 The Y-T correlation he with an r value of0 63 passes through the points dividing the Au; line seg- ments as a constant internal ratio of 0 63 and collapses into the single -1 0 -05 00 05 10 0-scale Figure 2 The Y-T plot for acetolysis of 2 phenylethyl tosylates (3) at 115°C 0,plots against (T+ ,e,against @ 0,against (T with r=O 63,.,for the plots of substituents having invariant cfor the three scales The figure is reproduced with permission from Bull Chem Soc Jpn 1987 60,1091 smooth correlation curve for the r-independent electron-withdrawing substituents reacting by the k mechanism The same behaviour was observed for the solvolyses of threo-3-aryl-2-butyl brosylatess and (1-arylcyclobutyl)methyl brosylates Rates of these solvolyses were previously reported to correlate lin- early with u+without a significant break implying the operation of only the k mechanism without a concurrent k process The simple Brown p+a+ treatment appears to be incapable of providing a correct interpretation of the mechanistic details The Y-T correlations for the acetolysis of 2,2-diarylethyl tosy- lates 4 (entries 5-7 in Table 1) are instructive since the nature of the phenonium transition state may be modified by the presence of the non-participating aryl group keeping the mechanism essentially the same lo The acetolysis of differently disubstituted diarylethyl systems 4(X,Y) proceeds through two competitive assisted pathways either by the X-substituted phenyl (k;) or by the Y-substituted phenyl group (k;) as shown in Scheme 3 The substituent effect on the symmetrically disubstituted bis- arylethyl tosylates I e 4(X=Y) can be described accurately in terms of the Y-T relationship with p= -2 22 and r=O 53 The Y-T plot is excellently linear against an apparent substituent scalecr with r=O 53 for the whole set of substituents indicating a uniform mech- anism for all of them When Y f X the overall solvolysis rate constant k corresponds to the sum of the rate constants k;+ kl and hence k cannot be employed directly in the Y-T analysis The acetolysis of mono-sub- stituted diphenylethyl tosylates gave a nonlinear Y-T correlation which is ascribed to a competitive X-substituted aryl-assisted pathway k; and the unsubstituted phenyl-assisted kH pathway By application of an iterative nonlinear least-squares method to eqn (3),where the subscripts A and s are now replaced by the k and k! pathways respectively the substituent effect on k was dissected into a k; correlation with p,=-3 53,r,=O 60,and an unassisted (by aryl-X-substituents) correlation for the phenyl-assisted k! mecha-nism correlated with 00 having pH= -0 88 The p and rdvalues for CHEMICAL SOCIETY REVIEWS 1996 f V -"'* xa-OTs xQ-Scheme 3 Solvolysis processes of 22 diarylethyl tosylates 4 the effects of assisting aryl substituents are quite close to those for 2 whereas the low pH value with the unexalted 00 constants for the unassisting aryls is compatible with a remote P-aryl effect In the acetolysis of 2-aryl-2-(3,5-dichlorophenyl)ethyltosylates 4( X,CI,) the strongly electron-withdrawing 3,5-dichlorophenyl group does not compete in an aryl-assistance pathway with any other aryl groups carrying more electron-releasing substituents than p-CI A Y-T correlation with r=O 66 for 13 substituents in the range down to rn-chloro was obtained demonstrating that a single X-substituted phenyl-assisted pathway takes place Similarly for the acetolysis of 2-aryl-2-(rn-chlorophenyl)ethyl tosylates 4(X,C1) eqn (2) gives a linear correlation with r=O 63 for substituents more electron-releasing than H This may reflect the Y-T correlation for the X-substituted aryl-assisted pathway The r and p values are comparable with those for the aryl-assisted k pathway of the mono-substituted diphenylethyl system and also with those for other k solvolyses e g neophyl brosylates suggest- ing a close similarity of the aryl-assisted mechanisms On the other hand the acetolysis of 2-aryl-2-(p-methoxypheny1)ethyl tosylates 4( X ,p-MeO) probably proceeds uniformly through the p-methoxyphenyl-assistedpathway There is a linear Hammett correlation against a0 (or u),attributed to the effect of the unassisting-aryl X-substituents in the p-methoxyphenyl-assisted mechanism The p value for symmetrical bis-arylethyl 4(X=Y) systems appears too small compared with those of the neophyl system with the single aryl group This can be accounted for by the fact that only one of the two P-aryl groups participates in the rate-determining aryl-assisted transition state while the other one affects the unas- sisted mechanism and both routes must be additive Consequently when we apply a p value of -0 8 for the latter route we obtain a Y-T correlation for the k route [eqn (4)I log(klk,,),=-2 22X2(UO+O 53A~ri,)-O 8a"=-3 6(8+0 6lA~t,) (4) which is practically identical to those for the k processes of Y-fixed 4(X,Y) systems In the 2,2-diphenylethyl system 4 all the r values are within a narrow range of 0 62kO 04 and tend to increase only slightly as the substituent in the unassisting aryl becomes more electron-with- drawing This r value of 06 can be referred to the resonance demand characteristic of the P-aryl assisted solvolyses 2.2 Benzylic Solvolyses generating Stable Carbocations Most benzylic solvolyses generating relatively stable carbocations may belong to the category to which the Brown u+constants are effectively applicable A broad applicability of the Brown u+p+ equation has been demonstrated for solvolyses of a wide series of tertiary a,a-dialkylbenzyl p-nitrobenzoates loThe solvolysis of a phenylethyl chlorides 5 in 80% aqueous acetone (80A) gives an excellent linear correlation (R=O 999) with eqn (2) with an r value of 1 15,lo whereas the Brown u+treatment gave a bisected rather than a single linear correlation II Whereas the exalted r value as well as the concave Brown plot may be attributed to the nucleo- philic solvent participation for the region of electron-attracting sub- stituents in nucleophilic solvents the finding of a strict linear correlation with the same r value in the less nucleophilic aqueous trifluoroethanol (TFE)5 l2 argues against the importance of solvent nucleophilicity in this case Any S l-S,2 mechanistic complication should be absent in the solvolysis of a-terf-butylbenzyl tosylates 6 having a neopentyl- type structure I2 Indeed the substituent effect is accurately described by eqn (2) with an r value of I 09 differs from the r= 1 0 for the a-cumenyl chlorides solvolysis I2 From the linearity of sub- stituent effects between the solvolyses of 5 and 6 in 80A an S,l-S,2 mechanistic duality is unlikely to be the cause of the exalted r value observed in the solvolysis of 5 The slightly lower r value for the solvolysis of system 6 than for the a-methyl analogue 5 is presumably due to incomplete copla- narity of the aryl group with the cationic p-orbital in the transition state of 6 l2 In the solvolysis of 2,2-dimethylindan-l-yl chlorides (entry 12 in Table I) the vacant p-orbital developed at the benzylic position is in an appropriate stereoelectronic conformation to overlap the benzene wsystem I3 Hence this system attains a full resonance stabilization at the transition state and the r value of 1 14 is practically identical with that observed for the solvolysis of chlo-rides of 5 I2 Consequently the resonance demand for the S,l solvolysis of secondary a-alkylbenzyl must be appreciably and intrinsically higher than that for the transition state for the solvoly- sis of tertiary a,a-dialkylbenzyl Benzhydryl solvolyses (entries 14-17 in Table I) show Y-T correlations with similar resonance demand and the magnitude of r is significantly dependent on the electron-withdrawing effect of the second aryl groups l4 Substituent effects in the solvolysis of the triphenylmethyl system are reported to give r values close to unity The pK,+ values for symmetrically trisubstituted triarylmethanols give a complete linear Y-T correlation against a@ scale with an r value of 0 76 for the substituent range from p-dimethylamino to p-nitro This con- trasts sharply with the less satisfactory Y-T correlation obtained for the pK,+ values for monosubstituted triphenylmethanols or for the log (klk,) values for the solvolysis of the corresponding chlorides in which the strong r-donor p-methoxy substituent requires a higher r-value than that for the other weaker electron-donating groups The three aryl rings in the triarylmethyl cation are twisted out of coplanarity with the vacant orbital by steric interaction The propellor conformation of the symmetrical triarylmethyl cation pre- vents the aryl groups from exerting their maximum stabilizing effect on the carbocation However in the monosubstituted cation only the strong r-donor substituted aryl can be coplanar with the cationic orbital exerting its maximum r-effect Hence deviations from the correlation line are indicative of resonance loss due to reduced conjugation caused by twisting The solvolysis of benzyl tosylates 7 is a typical case where a sig-nificant mechanistic shift occurs when the substituent is changed VARYING RESONANCE DEMAND IN CARBOCATIONIC SYSTEMS-Y TSUNO ET AL 5-05-4-p-Me,",'"3 CI 4-Me0 o-p-MeS4p-MeO-m-CI \p-MeS-m-CI+ I 50E 75°C m-clm-CF3 1 I I -1 0 -05 00 05 CJ scale Figure 4 The Y-T plot for solvolysis of a,a di tert butylbenzyl OPNB (8) in 50%aq EtOH at 75 "C r=O 26 For interpretation of symbols see Fig 1 caption Reproduced with permission from TetrahedronLeu 1991,32 2929 value is comparable to that for the pK of benzoic acids on which the Hammett cr scale is based The solvolysis of 4-methylbenzobicyclo[ 2 2 2 Iocten 1 -yl triflates 9 would be an excellent model of such a system where any exalted T-delocalization interaction should be completely prohibited The carbocation orbital developed at the bridgehead of the bicyclic skeleton IS rigidly orthogonal to the benzo morbital and in addition any back-side attack by a nucleophile is prohibited Logarithmic rates are in fact correlated directly with conventional a0 parameters giving p= -2 17 Thus the substituted effect on this solvolysis can be referred to by the so-called resonance unexalted a0 reactivity with r=O 0 being the lowest limit of exalted .rr-delocalization 2.4 Highly Electron-deficient Carbocation Systems The solvolyses of benzylic substrates carrying a strongly electron- withdrawing a-substituent (entries 26-3 1 in Table I) generate highly electron-deficient carbocat~ons~ and are expected to show l9 highly exalted resonance demand Analyses of the substituent effects in this class of solvolyses have been reported in the last 10 years 19-24 The solvolysis of 1-aryl-1-(trifluoromethyl)ethyl tosylates 10 was analysed by eqn (2) *O2I In the Y-T plot of this reaction (Fig 5) meta-substituents and para-r-acceptors covering five orders of magnitude in reactivity fall on a single straight line and the u+plot (open circles) of the para-.rr-donor substituents deviates upwards from the correlation line The linear Y-T plot (squares) against thea scale with r= 1 39 I e ,the correlation line dividing all the resonance Iine-segments (o+-@) ofparu-v-donor substituents by an external ratio of 1 39 contrasts sharply with the poor linear plot against u+ Substituent effects on the solvolysis of 1-aryl-2,2,2-trifluoroethyl tosylates 11can also be described in the same way to give a high r value of 1 53 22 23 The behavi our of the a-trifl uoromethyl -a ,a-diary 1methyl system 12was also analysed by eqn (2) 24The substituent effect on the sym- metrically substituted bis-aryl derivative 12(X= Y) can be described accurately by the Y-T relationship with p= -4 18 and r= 1 19 The Y-T plot IS linear over the whole range of substituents indicating the absence of any mechanistic change with change in substituent The unsymmetrically substituted derivatives 12 (X,Y) with a series of fixed Y =H m-CI and m,m'-CI afford slightly bisected Y-T plots 3-\=* h $ w-0 ~ -1 0 -05 00 05 10 0-scale Figure 3 The Y-T plot for solvolysis of benzyl tosylates (7) in 80%acetone at 25 "C,r= 1 29 For symbols see Fig 1 caption The graph is repro-duced with permission from Bull Chem Soc Jpn 1990,63,1146 The bisected (bilinear) correlation with cr+ values gives different p+ values for the regions of electron-donating and electron-withdraw- ing substituents this was reasonably ascribed to a change from the S,1 mechanism for the former substituents to the S,2 mechanism for the latter substituents Since this mechanistic transition is clear we do not expect a single linear relationship for the whole range of substituents Fig 3 shows a u+plot for the solvolysis in 80Al5 which is neither linear nor bilinear but displays a significantly split pattern of apparently parallel curvatures with significant gaps In contrast the Y-T correlation (Fig 3) vs a scale with Y= 1 29 gives a linear plot over a range of reactivity of 104 for substituents more reactive than 4-MeS-3-CN and is connected to the concave plot covering the range of more electron-withdrawing substituents The solvolysis in the less nucleophilic 97% aqueous TFE (97T) gives a straight-line correlation with an identical r value of 1 27 for a wider range of substituents more electron-donating than the meta halo-gens l5 The enhanced r value of this solvolysis is independent of the mechanistic complexity involving nucleophilic solvent assistance in the region of electron-withdrawing substituents 23 Solvoyses with Low r-Values P-Aryl assisted solvolyses characteristically show r values of 0 5-0 6 The solvolysis of (I-arylcyc1opropyl)methyl tosylates,I6 an analogue of the neophyl system gave an excellent Y-T correlation with an extremely low r value of 0 11 suggesting a different mech- anism from that of other P-aryl-assisted solvolyses The intrinsic resonance demand of a carbenium ion may be reduced effectively by reduced coplanarity between the reaction centre and the phenyl ring (vide inffa) Thus a diminished r value may be observed in the solvolysis of highly congested benzylic pre- cursors A well known example is the solvolysis of a,a-di-rert- butylbenzyl p-nitrobenzoates 8 having two bulky tert-butyl groups at the reaction centre l7 The application of eqn (2) affords an excel- lent linear correlation (Fig 4) with p= -2 19and r=O 26 !' This r 9 CHEMICAL SOCIETY REVIEWS 1996 ~ based on eqn. (1) and it was suggested that exceptionally high p+ values in the range of -10 to -12 are a characteristic feature of a-these highly electron-deficient carbocation-forming reactions.23 However as summarized in Table 2;5z2all these solvolyses showed significantly nonlinear Brown cr+ correlations; p&= -10 for the 6-electron-donating range of substituents while p:= -5 to -6 for the region of electron-attracting substituents .22 Since the nonlinear rela- tionship rules out the operation of a single mechanism for the whole range of substituents all these solvolyses without exception 4-should involve a mechanistic change with a break in the o+plot in the vicinity of the unsubstituted derivative. It is remarkable that all but one of the 19 sets in Table 2 give good aP 3-CN-4-MeS linear relationships (D0.99) against the log (k/ko)lovalues from the solvolysis of 10 in the corresponding solvents and none shows a significantly higher slope than unity. This simple linearity sug- gests that the r values as well as p values must be very similar for these systems. For the solvolysis of A~CH(OMS)PO,E~,,~~ a good mMeS \ linear free energy relationship is obtained against log (k/ko)loin 80T without a significant break. On the other hand the solvolysis of A~CH(OMS)PS(OE~),*~is the only case which shows a clear break in the plot against log(k/k,), and it evidently indicates significant thio group participation for deactivating substrates. No mechanisticmCF3 change seems to take place in any of the other reactions. An elec- trophilic substituent parameter a scale with a high resonance I 1 I I demand (r=ca. 1.4)should be required for a proper description of -1.5 -1.o -0.5 0.0 0.5 1.o the substituent effects in these extremely electron-deficient benzylic 0-scale systems. A ceiling to the magnitude of the resonance demand in benzylic solvolyses appears to be an r value of 1.5-1.7. Figure 5 The Y-T plot for solvolysis of a-CF,-a-CH,-benzyl-OTs (10); r= 1.39. For symbols see Fig. I caption. The graph is redrawn from the data in ref. 21. 3 Stabilities of Carbocations in the Gas Phase The mechanistic involvement of the solvent is an important cause with highly exalted r values; those for the electron-donating sub- of the misunderstanding of substituent effects on benzylic solvoly- stituent range are 1.45 I .57 and I .69for these Y groups. Both p and ses. An effective approach to overcome this difficulty in the r values are appreciably reduced in the range of electron-withdraw- solvolytic studies and towards a general theory of substituent ing substituents in all the sets. The deviation from the linear Y-T rela- effects in benzylic system directly comparable with the theoretical tionship may be attributed to the deviation of the two aryl groups results is to investigate the behaviour of carbocations in the gas from coplanarity; a convincing example is afforded by the Y-T cor- phase a medium free of solvent participation and other complicat- relation of pK,+ for triarylmethanols. Since the r value obtained for ing factors. The intrinsic stabilities of benzylic carbocations have 12 (X=Y) is reduced due to the conformation with reduced copla- now become available from ion cyclotron resonance mass spectro- narity the resonance demand of the (hypothetical) planar system 12 scopic determinations in the gas pha~e.~6,,~ should be higher than the value of unity in the a-cumenyl system. The relative stabilities of carbocations can be estimated from the Whereas we have achieved remarkable success with the Y-T free energy changes of the ion-molecule proton-transfer equilibria analysis most studies of solvolyses of this class have so far been of the corresponding olefins [eqn. (5)1 or those of the gas-phase Table 2 Substituent effects in the extremely electron-deficient systems Brown eqn.b Log k-log k eqn .c System“ Solv. PD+ PA’ Slope= R Ar(3 ,5-CI,C6H,)CCF,-OTs 80E -9.6 -1.17 0.995 Ar(m-CIC,H,)CCF,-OTs 80E -9.1 -4.7 1.01 0.996 Ar( Ph)CCF,-OTs 80E -7.7 -4.6 0.79 0.991 ArC(CN)Me-OMS TFE -6.70 0.881 0.998 ArC(CF,),-OTs TFA -10.7 1.12 0.999 ArC(CN)CF,-OTs TFA -12.1 1.32 Ar-a-ketoNB-OTfd EtOH -5.69 -4.15 0.726 0.998 ArC(CF,) Me -0Ts 80E -8.8 -6.4 1.woe ArC(CF,)Me-Br -10.3 ArC( S0,Ph)Me-OMS MeOH -8.0 0.921 0.9999 ArC(S0Ph)Me-OMS TFE -7.2 0.778 09995 ArCH(OMs)-P( =O)(OEt) TFE -10.1 -6.1 0.918 0.993 97HFIP -10.3 1.123 0.9998 ArCH(OMs)-P(=S)(OEt) AcOH -7.15 -2.99 0.69 0.985 0.45s 0.9988 ArCH( CF,)OTs TFA -6.7 1.05 0.993 97HFIP -9.1 1.020 0994 TFE -9.8 1.046 0997 AcOH -10.1 I .08 aq.EtOH -I].%-9.7 0.982 0.998 Most of data are taken from ref. 22; otherwise see text. h p& and pi are the Brown p+ values for electron-donor and -accepr substituents respectively. Slope of logarithmic rates against the log (k/&,Jl0 for the solvolysis of ArCMe(CFJ0Ts (10) in the corresponding solvent. 2-Aryl-3-oxobicyclo[2.2.1Iheptan-2-y triflates. <’ Bydefinition. f More reactive than 3,4-Me2. Less reactive than 3,4-Me2. VARYING RESONANCE DEMAND IN CARBOCATIONIC SYSTEMS-Y TSUNO ETAL (5) .. (RSH3 Et H CF3) PI 10 2k m -0 -4 MeS 0 -1 0 -20 -1 0 00 10 ~+(sol"tlon) Figure6 The plots of gas phase stabilities of substituted a cumenyl cations (lC+)against Brown's cr+ in solution The figure is redrawn from the data in ref 26 halide-transfer equilibria of corresponding benzylic halides [eqn (6)j 2627 The substituent effect on the stability of the a-cumenyl cations CI 10 f-3-CI4-Me0 3-F-4460 3-CI-4-MeS hP5 Y 0)0 0 -5 -1 0 -05 00 05 0-scale Figure7 The Y-T plot of gas phase stabilities of substituted benzyl cations (7C+),r= 1 29 For symbols see Fig 1 caption Data taken from ref 27 and of the corresponding benzylic S,1 solvolyses have identical magnitude (Table 3) From this linearity the varying resonance demand in solvolysis should be an essential feature of the incipient carbocation intermediate as well as of the solvolysis transition state The r value of benzylic solvolyses should be essential in the intrin- lC+ based on the proton transfer equilibrium of a-methylstyrene~~~sic resonance demand of generated cations which should in turn can be correlated directly with the ordinary set of solution (T+ values (Fig 6) Unexpectedly the correlation covering the substituent range from p-NMe to 3,5-(CFJ2 is excellent and there is no diffi- culty in defining the gas-phase CT+ scale for r= 1 00 by using these gas-phase stabilities of substituted 1C+ cations 26 The relative stabilities of substituted benzyl cations are correlated with eqn (2) (Fig 7) with a higher resonance demand parameter r-1 29 than for the cumenyl system (r= 1 00)This linear correlation for the whole range of substituents contrasts sharply with the concave Y-T plot (Fig 3) of the solvolytic reactivities of 7 Furthermore the r value for the gas-phase stabilities of 7C+ is iden- tical with the value assigned for the S 1 solvolysis of tosylates of 7 Hence the r value of 1 29 must be an intrinsic feature of the S,l solvolysis mechanism of 7 rather than a correlational artifact of the nonlinear relation due to involvement of the solvent Consequently the Y-T equation is applicable to the gas-phase substituent effects on the intrinsic stabilities of benzylic cations in exactly the same manner as it applies to the solution phase The r value significantly increases as the stability of parent carbocation decreases while the p value remains constant for a series of ben- zylic carbocations It is remarkable that the resonance demand parameters characterizing the stabilities of a series of benzylic cations in the gas phase are linearly related to the intrinsic gas-phase stabilities of the parent carbocations (Fig 8) Furthermore,it is very surprising that the r value for the gas-phase stabilities of the cations depend on the instability of the cation The r value for the solvoly- sis of a-cumenyl chlorides which generates a higher stable tertiary benzylic carbocation should be in the lowest range of the r values for benzylic solvolyses Consequently any ordinary benzylic solvolysis should rarely show an r value significantly lower than unity More stable carbocation systems showing significantly low resonance demand from the aryl group will be those having strongly electron-donating a-substituents e g -OR -NR etc Significantly low r values as well as identity between solution and gas phase r values have also been observed in the protonation equi- libria (pK,,+ values) of benzoyl compounds 28 However,solvolytic reactions of benzoyl derivatives are often complicated by subsidiary reactions of the highly polar a-substituents The gas-phase stability of the ethylene-phenonium ion 3C+ and its ring substituted derivatives has also been determined by the bromide-transfer equilibria [analogous to eqn (6)1 of P-arylethyl bromides and gave an excellent Y-T correlation with p= -12 8 and r=O 60 (Fig 9) 29 The p value is significantly larger than those observed for benzylic carbocations but is similar to that obtained for the stabilitites of benzenium ions 29 It should be particularly noted that the r value of 0 60 is identical with the value observed for the corresponding solvolysis which proceeds via an intermediate phenonium ion 6-Ii Clearly there is a continuous spectrum of intrinsic resonance demand of the stabilities of benzylic carbocations in the gas phase 136 1.6 1 -1.5 Lv T1 c 1.4 -aE 1.3 -0 t a s:c 1.2 -2 1.1 -1.0 -Increasing stability F \ t.l.l.l.l.l.l -100 -80 -60 40 -20 0 20 Relative stability of the parent cation Figure 8 Plot of the r values for gas-phase cation stabilities against the sta- bilities of parent cations. Reproduced with permission from Chern. Lett. 1992,1085. Table 3 Yukawa-Tsuno correlations for the gas-phase stabilities of benzylic cations a-Rs in Y-T-correlation benzylic cation AG,a PP rg rS"1" 81.6 -10.6 1.53 1.53 67.8 -10.2 1.40 1.39 51.0 -10.3 1.29 1.29 31.4 -10.2 1.18 21.8 -9.9 1.14 1.15 0.0 -9.5 1 .oo 1 .oo -1.7 -9.5 1.oo 1.04 4.6 -9.1 0.86 0.91 41.8 -12.8 0.60 0.63 Relative stability of respective parent carbocations estimated from proton-transfer or chloride-transfer equilibria in kJ mol-I. a-Arylvinyl cation Ar-C+=CH,. Ethylenephenonium ion 3C'. This leads to the important conclusion that the resonance demand characterizing the S 1 solvolytic transition state should in principle be determined by the intrinsic resonance demand of the carbocation intermediate. 4 Relation between r and Molecular Structure Para meters The characteristic change of the r value in the solvolysis reaction and for the corresponding carbocation should provide important information concerning the solvolysis transition state. The value of r reflecting the r-delocalization within the cationic species appears to remain essentially the same in solution as in the gas phase and the degree of charge delocalization in the transition state of the solvolytic ionization should also resemble that in the carbo- cation intermediate. Moreover the r value is directly related to the intrinsic stability of the parent cation. Thus ab initio calculations can be used to find the underlying relationship between quantum chemical quantities and experimental r values.30 The relation between the r values and theoretical indices provides the basis for the physical meaning of the r parameter. For a series of benzylic cations several bond lengths optimized at the RHF/6-31G* level are plotted against the r values in Fig. 10. The C' -C7 bond length decreases significantly with increased reso- CHEMICAL SOCIETY REVIEWS 1996 t Ye pMes 3-CI-4-Me0 mMe P,90 -5 mF \mCF3 m-CN \ -10 I 4 I I -1.o -0.5 0.0 0.5 0-scale Figure 9 The Y-T plot of the stabilities of gas-phase phenonium ions; r=0.60. For symbols see Fig. 1 caption. The graph is redrawn from the data in ref. 29. CF3,Me Me,MeT/Pr,Pr \*Et,Et Me'H Et,Me ''pH 1.31 -. -. . . . -. . . . . 0.0 0.5 1.o 1.5 r value Figure 10 Bond lengths vs. r values for benzylic cations. Data taken from ref. 30. nance demand. As the r value increases the C1-C2 and C3-C4 bonds are lengthened but the C2-C3 bond is shortened. The Cl-C7 bond length at r=O is estimated to be 1.52 A which is close to the normal C-C single bond length whereas at r= 1.51 the C' -C7 bond length is 1.358 which is close to the normal C=€ double bond length. The C1-C2 and C2-C3 and C3-C4 bond lengths at r=O are all identical being 1.39 A similar to the benzene C-C bond length. The Wiberg bond order also changed consistently with the corresponding bond + 1 I1 111 IV V Scheme 5 VARYING RESONANCE DEMAND IN CARBOCATIONIC SYSTEMS-Y TSUNO ETAL I....,. 00 05 10 15 r value Figure 11 Mulliken population at ortho . metu and para-positions of phenyl ring (RHF/6 3 IG") I5 r values for benzylic cations Data taken from ref 30 lengths The Cl-C7 bond length or bond order at r=O clearly reflects contribution from structures I and V The contributions of resonance structures 11-IV become more important on increasing the resonance interaction between the ben- zylic p7r orbital and the benzene 7r system This tendency reveals that the r value changes linearly with the degree of overlap between the two orbitals thus reflecting the relative importance of the con- tribution of canonical structures I-V The behaviour of r is com- pletely consistent with what is envisaged for the resonance demand indicating that the varying resonance interaction between C7and the aromatic moiety changes in parallel with the r value The Mulliken charge distribution at the respective atomic posi- tions should be regarded as reflecting the demand of the a-cationic centre of benzylic carbocations for the 7r-electron delocalization from those atomic positions and should be directly related to the r value As seen in Fig 11 there are linear correlations of the charge populations as the para- and ortho-position against the r value but there is no significant change in the population at the metu-position As the r value increases the charge at the para-position increases significantly to +O 2 at r= 1 5 whereas the charge at the nzeta-posi- tion changes much less This trend of charge delocalization for ben- zylic cations is consistent with the varying degree of the resonance stabilization Deviations are significant for the two cations 9C + and 1C' (90° twisted) at r=O in which the benzylic pnorbitals and the benzene 7rorbitals are orthogonal the bonding uorbitals of the a-substituents and the benzene +orbital are in the approximate con- formation for overlap and the deviation may be attributed to the mixing of these two orbitals 5 Steric Loss of Resonance Interaction The variation of the r parameter has been discussed in terms of the varying extent of substituent-reaction site resonance interactions from one reaction series to another The carbenium-like orbital in the transition state for the solvolysis of a benzylic substrate carry- ing two bulky a-alkyl groups may be twisted out of coplanarity as exemplified by the solvolysis of 8,17resulting in a markedly reduced r value The observation of steric control of resonance demand is also important in clarifying the origin of the r value in the Yukawa-Tsuno correlation The degree of steric inhibition of resonance largely depends on the bulk of the a alkyl subctituents 31 The efficiency of resonance interaction should be proportional to cos2H. where H is the dihe- dral angle between the two overlapping p orbitals as in equation (7). Table 4 Calculated Torsional Angles (0 degree) between Alkyl Groups and Benzene Ring Ar C+Ri.R' OLA' Lpt R' R' 3 21G 6 31G" MP2(FU)/6-31G* Me Me 5 5 7 0 Bu' Me 24 24 I7(22") ( 14Y (15)'But Prl 37 33 32 But But 77 76 69 59 (63Y (60Y BBCO+/ 90 90 90 ' Given by ~/~,~~=cos'Owhere rmr,=l 00 for planar a (Y drmethylbenzyl system Based on gas phase data For the p Me0 derivative 4 methylbenzobicyclol2 2 2locten 1 yl cation 9C1 where rlnaxis the intrinsic r for the coplanar tert-benzylic system given by r= 1 00 for the a,a-dimethylbenzyl system The r values and the dihedral angles 8 based thereon are compared in Table 4 for a series of a ,a-dialkylbenzylic precursors The gas-phase stabilities of a-tert-butyl-a-methylbenzyl cations have been found to be correlated linearly by eqn (2) with r=O 86,72 a value close to r=O 91 derived from the solvolysis 32 The similar- ity between the empirical r values for the solvolysis transition state and the corrresponding gas-phase cation appears to comprise both sterically twisted and untwisted systems The optimized structure of the carbenium ion derived from theoretical calculations should be able to model that of the solvolytic transition state as well as the intermediate immediately following it Ail the dihedral angles of the twisted tert-benzylic cations in the optimized structures (Table 4) are in good agreement with the 'experimental' angles of the solvolysis transition states deduced from eqn (7) In the 6-31G* optimized structure of the carbenium ion SC' the calculated dihedral angle of 76" for the unsubstituted ion is lower than the angle estimated from the solvolysis r value whereas the calculated angle of 60" in the optimized structure of p-methoxy- 8C+ is in a better agreement with it This is as expected since the enhanced delocalization by the p-methoxyphenyl groups should increase its coplanarity with the reaction centre and hence the r value for the solvolysis of8 should be slightly higher (see Fig 4) The close similarity between the theoretically calculated and the empirically observed dihedral angle assigned for twisted benzylic systems corroborates the characterization of the empirical parame- ter r in the Yukawa-Tsuno equation as a parameter reflecting the degree of 7r-delocalization interaction between the aryl group and the reaction site 6 Varying Resonance Demand in Solvolyses The characteristic change of the r value in both the solvolysis reac- tion and the corresponding carbocation provides important infor- mation concerning the solvolysis transition state In a typical two-step S 1 mechanism with a single dominant tran- sition state the r values of closely-lying transition states for the various nucleophile-cation reactions should dlI be predominantly controlled by the intrinsic resonance demand of the intermediate cation the substituent effect should be described by a single scale of substituent constants (a)with an r value characteristic of the sygtem In fact in a recent laser flash photolysis study on the recombination of stable trityl and benzhydryl carbocations with nucleophiles and sol vents,33 McClelland et af have demonstrated that the substituent effects on the solvent-recombination processes can be correlated by eqn (2) with the same r values observed in the solvolyses and with smaller p values The effect of the coplanarity change was also observed similarly 33 The identity of r values has been recognized also in various benzylic cation-nucleophile recombination reactions 34 These facts imply that either the reverse ofthe kc ionization or the recombination step of the carbocation with various nucleophiles should be described by a Y-T a scale with the same r value as for the ionization step This is reminiscent of the Ritchie N relation-ship of constant susceptibility to the nucleophile regardless of the cation,35 which requires that the p for the recombination reaction with any nucleophile should be a constant of the cation regardless of nucleophile The energy profile in the vicinity of the transition state and the intermediate in the solvolysis process should be essentially non- crossing For solvolyses involving higher endothermic ionization processes the Hammond shift in the transition state coordinate rel- ative to the intermediate is reflected almost exclusively in the p value without affecting the r value The Bronsted-type analysis the so-called rate-equilibrium relationship in terms of a common Y-T a scale can be applied precisely for systems involving an apprecia- bly wide shift of solvolytic transition states For the uniform applicability of eqn (2) with either a non-unity or a non-zero r value Johnson suggested that the reaction con- forming to the Y-T equation with rfl 00 may be a two-step process involving a pre-equilibrium K followed by a rate-determining k step where one of these steps correlates with 1p (or cr) and the other with (T+ [eqn (8)] Whereas this scheme is logically capable of reproducing the Y-T correlations this cannot be general as the cause of observing non- unity r values since we can generally find widely varying resonance demand for the intrinsic stabilities of benzylic cations in the gas phase As the Y-T equation implies the unification of a substituent parameter scale in terms of varying r leads to a unique additivity relationship of substituent effects for the system of k=k k kJ [eqns (9a) and (9b)I where p'=Ip and r'=xp,rJcp Eqns (9a) and (9b) allow in theory the assignment of any reasonable rl value for each step Thus for the above overall pre-equilibrium mechanism the overall sub- stituent effect can be represented as a single linear Y-T correlation with an apparent r' value the Johnson pre-equilibrium scheme is a special case of rl=O and r2=1 00 which cannot be general 5 It is of great importance that this equation can be applicable also to the simultaneous contribution of different substituent effects on a single reaction step We have already discussed additivity of the sub- stituent effects on the solvolysis of bis-arylethyl tosylates 4(X= Y) The simple additivity inherent in the Y-T relationship makes eqn (2) a useful tool for analysis of reaction mechanisms but at the same time caution should be exercised in order to avoid Its inter- pretation being misleading In conclusion we want to emphasize that in spite of the wide spectrum of observed r values the intrinsic resonance demand is an inherent property of the intermediate or the transition state of respective reactions and is therefore characteristic of the structure of intermediate This is the basis of mechanistic analysis by using the Yukawa-Tsuno relationship Acknowledgement The authors are grateful to Professors M Mishima S Kobayashi and M Sawada for carrying out most of the important parts of these investigations and also to Professors Z Rappoport K Nishimoto T T Tidwell and J P Richard for coop- erative research Particularly the authors would like to express their sincere appreciation to Professor 2 Rappoport for critically reading the manuscript and for stimulating discussions and valuable sug- gestions In the preparation of this review paper CHEMICAL SOCIETY REVIEWS 1996 7 References I H C Brown and Y Okamoto J Am Chem SOC 1958,SO 4979 L M StockandH C Brown,Adv Phys Org Chem 1963,1,35 2 Y Yukawa and Y Tsuno Bull Chem Soc Jpn 1959 32 971 Y Yukawa Y Tsuno and M Sawada Bull Chem SOC Jpn 1966,39 2274 Y Yukawa and Y Tsuno Nippon Kagaku Zusshi 1965 86 873 3 J Shorter Correlation Analysis in Chemistry eds by N B Chapman and J Shorter Plenum Press New York 1978,ch 4 p 119,and refer ences quoted therein 4 C D Johnson J Org Chem 1978,43,1814,andreferences therein C D Johnson The Hammett Equation Cambridge University Press New York 1973,pp 52,86 and 108 5 Y Tsuno Proceedings of the 10th International Seminar on Physical Organic Chemistry Kyungpook National University Taegu Korea 1986,p 195 M FUJIO M Goto A Murata Y TSUJ~,M Mishima and Y Tsuno Mem Fac Sci Kyushu Univ Ser C 1988,16(2) 271 and references therein 6 M FUJIO K Funatsu M Goto M Mishima and Y Tsuno Tetrahedron 1987 43 307 M FUJIO M Goto M Mishima and Y Tsuno Bull Chem SOCJpn 1990,63,1121 7 M FUJIO K Funatsu M Goto Y Seki M Mishima and Y Tsuno Bull Chem Soc Jpn 1987,60,1091,1097 8 M FUJIO N Goto T Dairokuno M Goto Y Saeki Y Okusako and Y Tsuno Bull Chem Soc Jpn 1992,65 3072 9 D D Roberts J Org Chem 1974,39 1265 M FUJIO Y Saeki K Nakamoto S H Kim Z Rappoport and Y Tsuno Bull Chem Soc Jpn 1996,69,751 10 M FUJIO Y Maeda M Goto M Mishima and Y Tsuno Bull Chem Soc Jpn 1993,66,3015,3021 11 Y Tsuno Y Kusuyama M Sawada,T FUJI] and Y Yukawa Bull Chem Soc Jpn ,1975,48,3337,M FUJIO T Adachi Y Shibuya A Murata and Y Tsuno Tetrahedron Lett 1984,25,4557 12 Y TSUJI,M FUJIO and Y Tsuno Bull Chem SOC Jpn 1990,63,856 13 M FUJIO K Nakata Y TSUJI,T Otsu and Y Tsuno Tetrahedron Lett 1992,33,321 14 J R Fox and G Kohnstam Proc Chem SOC 1964,115 S Nishida J Org Chem 1%7,32,2692,2695,2697 15 M FUJIO M Goto T Susuki I Akasaka M Mishima and Y Tsuno Bull Chem SOCJpn 1990,63,1146 M Fuji0 M Goto T Susuki,M Mishima and Y Tsuno J Phys Org Chem ,1990,3,449 16 D D RobertsandT M Watson,J OrR Chem ,1970,35,978 M Fuiio Y Saeki and Y Tsuno unpublished results 17 M FUJIO TMiyamoto Y TSUJIand Y Tsuno Tetrahedron Lett ,1991 32,2929 18 M FUJIO,K Nakashima E Tokunaga Y TSUJ~ and Y Tsuno Tetrahedron Lett 1 992,33,345 19 A D Allen and T T Tidwell Advances in Carhocation Chemistry vol 1 ed X Creary JAI Press Inc ,Connecticut 1989,p 1 T T Tidwell Angew Chem Int Ed Engl 1984,23,20 20 K TLiu,M Y KuoandC F Shu,J Am Chem Soc 1982,104,211 K M Koshy D Roy and T T Tidwell J Am Chem SOC 1979,101. 357 21 A Murata,M Goto,R Fujiyama,M Mishima M F~ji~and Y Tsuno Bull Chem Sac Jpn 1990,63,1129 22 A Murata S Sakaguchi R Fujiyama M Mishima M Fuji0 and Y Tsuno Bull Chem SOC Jpn 1990,63,1138 23 A D Allen I C Ambidge,C Che,H Micheal R J Muirand T T Tidwell J Am Chem Soc 1983 105 2343 A D Allen V M Kanagasabapathy and TT Tidwell ,J Am Chem Soc 1986,108,3470 24 K T Liu,S M Chang,H 1 Chen,P F ChiuandT R Wu,J Org Chem 1991 56 1315 M FUJ~O Morimoto and Y Tsuno BullH Chem Soc Jpn ,to be published 25 X Creary J Am Chem SOC 1981 103 2463 X Creary and T L Underiner J Org Chem 1985 50 2165 X Creary and M E Mehrsheikh Mohammadi J Org Chem 1986,51,7 P G Gassman and T L Guggenheim J Org Chem 1982,47,3023 26 M Mishima S Usui H Inoue M Fuji0 and Y Tsuno Nippon Kagaku Kaishi 1989 1262 M Mishima S Usui M Fuji0 and Y Tsuno Nippon Kagaku Kaishi 1989,1269 27 M Mishima K Arima H Inoue S Usui M Fuji0 and Y Tsuno Bull Chem SOC Jpn 1995 68 3199 M Mishima H Nakamura K Nakata M Fuji0 and Y Tsuno Chem Lett 1994 1607 M Mishima H Inoue M Fuji0 and Y Tsuno Tetrahedron Lett 1989 30 2101 1990,31,685 28 M Mishima M FUJIOand Y Tsuno Tetrahedron Lett 1986 27 939 95 1 29 M Mishima Y TsunoandM Fuji0,Chem Lett 1990 2277 30 K Nakata M FUJ~O Y Saeki M Mishima K Nishimoto and Y Tsuno J Phys Org Chem in the press VARYING RESONANCE DEMAND IN CARBOCATIONIC SYSTEMS-Y. TSUNO ET AL 31 M. Fujio K. Nakata,T. Kuwamura Y. Saeki M. Mishima S. Kobayashi 34 J. P. Richard M. E. Rothenberg and W. P. Jencks J. Am. Chem. SOC. and Y. Tsuno Tetrahedron Lett. 1993,34,8309. 1984,106,1361;J. P. Richard and W. P. Jencks,J.Am. Chem. Soc. 1984 32 M. Fujio H. Nomura K. Nakata Y. Saeki M. Mishima S. Kobayashi 106,1373,1383,13%;J. P. Richard V. Jagannadham T. L. Amyes M. and Y. Tsuno Tetrahedron Lett. 1994 35 5005; M. Mishima K. Mishima and Y. Tsuno J.Am. Chem. Soc. 1994,116,6706. Nakata H. Nomura M. Fujio and Y. Tsuno Chem. Lett. 1992,2435. 35 C. D. Ritchie,Acc. Chem. Res. 1973,5,348; Can. J. Chem. 1986,64 33 R. A. McClelland V. M. Kanagasabapathy N. Banait and S. Steenken 2239. J. Am. Chem. Soc. 1989,111,2929;R. A. McClelland N. Banait and S. Steenken,J.Am. Chem. SOC.,1986,108,7023.

ISSN:0306-0012    

DOI:10.1039/CS9962500129

出版商:RSC    

年代:1996

数据来源: RSC