Platinum Metals Rev., 1975, 19, (3), 100
Stabilised Carbonium Ions in the Organic Chemistry of Platinum
The Evidence for Their Existence
This account of research on the organometallic chemistry of platinum summarises work which indicates that platinum may induce carbonium ion reactivity in π-bonded hydrocarbon ligands and stabilise carbonium ions in the reactions of σ-bonded unsaturated hydrocarbons. It is based on a lecture at the Chemical Society annual meeting held at York in April.
Historically, platinum has played a prominent role in organometallic chemistry. The compounds KPtCl3(C2H4)H2O (1), PtMe3 (X) (2) and Chugaev’s salt [PtC8H15N6]xClx (3, 4) present some of the first examples of stable metal-olefin, metal-alkyl and metal-carbene complexes, although ironically the nature of the bonding in these compounds was not immediately recognised. In the study of organometallic reaction mechanisms platinum has also been in the forefront, for example, in studies of the decomposition pathways of metal-alkyls (5, 6); oxidative-addition reactions (7–9); C-C and C-H reductive-elimination reactions (10–14); and M-H and M-C insertion reactions (15–22).
Certain aspects of the organometallic chemistry of platinum are well suited for the study of organic reactions occurring at metal centres (23):
(i) There exists a diverse organometallic chemistry for platinum in its well-defined valence states 0, 2+ and 4+.
(ii) For a given valence state platinum shows a well-defined coordinated chemistry. Pt(o) compounds are either three or four coordinate. Pt(2+) compounds show a remarkable preference for a four coordinate square planar geometry; five coordination is comparatively rare. Compounds of Pt(4+) are almost invariably six coordinate.
(iii) Platinum-carbon bonds are stable to oxygen, water and thermolysis. This stability results from a combination of thermodynamic and kinetic factors.
(iv) Platinum compounds are well suited to a wide variety of spectroscopic techniques. In particular the presence of 34 per cent natural abundance, provides a useful mechanistic probe in n.m.r. studies of organoplatinum compounds.
This account summarises work in our laboratories which indicates that platinum may induce carbonium ion reactivity in π-bonded hydrocarbon ligands and may stabilise carbonium ions in the reactions of σ-bonded unsaturated hydrocarbons.
By taking advantage of the high trans -influence (24) of the methyl ligand we set out to synthesise a series of cationic methyl-platinum compounds trans -[PtCH3(un)L2]+ PF6−, I, where un=an unsaturated hydrocarbon and L=a tertiary phosphine or arsine, according to reaction A below.
Compounds I were considered ideal for detailed studies of the bonding and reactivity of the unsaturated ligand un. However, the products derived from A were found to be dependent on the nature of the unsaturated hydrocarbon, the ligands on platinum, the solvent, and the reaction conditions. Reactions involving dialkyl or diaryl acetylenes, un= RC≡CR, in acetone solvent did lead to the isolation of cationic methylplatinum(2+) acetylenic compounds as white air stable crystalline salts (25, 26). From a comparison of v str(Pt-C), 2J 195Pt−H and 1J 195Pt−13C for the methylplatinum group in I, where un = RC ≡ CR, with related values observed for methylplatinum(2+) cations trans -[PtCH3 (L′)L2]+, where L′ =a neutral donor ligand, the following trans -influence series may be formulated: (27)
This series shows that the acetylenic triple bond exerts a relatively weak trans -influence, somewhere between the nitrogen donor ligands pyridine and alkylnitriles.
Further insight into the nature of the platinum-acetylene bond in I can be obtained from a consideration of the data shown in the Table. This pertains specifically to I, un= CH3C≡CCH3. v str(C≡C) which occurs at ∼ 2200 cm−1 in free 2-butyne is reduced only 100 cm−1 on coordination to the methyl-platinum(2+) cation. In contrast, coordination to zerovalent platinum as in (P(C6H5)3)2Pt(CH3C≡CCH3) causes a much greater reduction in this stretching frequency. Indeed the value of v str(C≡C) is lowered to a value comparable to that observed for cis -2-butene. Similarly the 13C shieldings of the acetylenic carbons (see the Table) reveal (28) the very different nature of the metalacetylene bond in trans -[PtCH3(CH3C≡CCH3) L2]+ and (P(C6H5)3)2Pt(CH3C≡CCH3). A single crystal X-ray determination on trans -[PtCH3(CH3C≡CCH3)(P(CH3)2C0H5)2]+PF6− revealed (29) a C≡C bond length of 1.22Å, c.f. (30), 1.21 Å for free 2-butyne, and an average C≡C-C angle of 168°.
In summary all available spectroscopic data suggests that, based on the original Dewar (31)-Chatt-Duncanson (32) pictorial descriptions of metal-acetylene bonding below, platinum or acetylene π* bonding is of minor importance in I. This is in marked contrast to the bonding in zero valent platinum acetylene compounds which may be considered to show a sizeable degree of cyclopropenoid character. The primary mode of bonding in I, un=RC≡CR, appears to be a σ-interaction involving donation from filled π-acetylenic orbitals to vacant platinum hybrid orbitals.
|Compound||v Str(C≡C)cm−1||δ (C)a|
Although dialkyl and diaryl acetylenes did yield I according to A, the reaction of terminal acetylenes, RC≡CH, in methanol solvent gave cationic methylplatinum methoxy-carbene complexes according to B (25,33). Analogous reactions in ethanol gave the appropriate ethoxy-carbene complexes. This formation of the alkoxy-carbene ligand in B was unexpected and requires specific comment.
Initially we proposed (25, 33) a mechanism in which a cationic acetylenic complex I reacts with the solvent to produce a cationic π-vinyl ether complex, which then by hydride shift yields the alkoxycarbene ligand. However, we were subsequently able to isolate cationic methylplatinum π-vinyl ether complexes (21). These were stable below +80°C and did not rearrange to give alkoxy-carbene derivates. This observation together with many other reactions of cationic acetylenic platinum complexes, summarised in Scheme 1, led us to propose a general carbonium ion model for the activation of unsaturated ligands in organoplatinum cations trans -[PtR (un)L2]+ (21, 34).
The Carbonium Ion Model
We proposed (21, 34) that activation occurs by delocalisation of the positive charge from platinum to the unsaturated moiety un; this has the effect of inducing carbonium ion reactivity. Specifically for acetylenes coordinated to platinum(2+) cations this can be pictured by the resonance shown below.
The carbonium ion reactivity of the coordinated acetylene is thus dependent on the nature of the substituents of the acetylene, R and R′; the ligands on platinum, X and L, since these determine the electrophilicity of the platinum cation; the solvent; and the nature and availability of a nucleophile.
Thus most of the reactions shown in Scheme 1 may be reduced to three fundamental types of reactions shown in Scheme 2. These are (i) nucleophilic attack on the coordinated acetylene leading to trans -vinylic products, (ii) proton elimination or abstraction by the solvent when R=H (this is favoured in polar aprotic solvents), and (iii) insertion into a group already bound to platinum leading to cis -vinylic products.
Aside from these three types of reactions (Scheme 2) many other aspects of organoplatinum chemistry illustrate the generality of this platinum induced carbonium ion model. For example, substitution of the methyl group bound to platinum in I by the more electron withdrawing trifluoro-methyl group increases the electrophilicity of the organoplatinum cation and hence increases the carbonium ion reactivity of the acetylene in trans -[PtCF3(RC≡CR′)L2]+relative to I. Although we have been unable to isolate such simple dialkylacetylene cations, they are the reactive intermediates in the polymerisation of acetylenes, a process which occurs much faster than for I in the absence of a polar protic solvent. A terminating step in the polymerisation of dimethylacetylene is the formation of the cationic tetra-methylcyclobutadiene complex [PtCF3(C4(CH3)4) (P(CH3)2 C6H5)2]+PF6− (35, 36).
Another reaction in organoplatinum chemistry which is consistent with the platinum induced carbonium ion model is the formation of an alkoxy-carbene ligand from a platinum (4+) cationic acetylenic intermediate (35).
Mechanism of Formation of the Carbene Ligand
The mechanism for the formation of the alkoxy-carbene ligand from the elements of RC≡CH and R′OH deserves specific comment. The proposed mechanism, based on the carbonium ion model, is shown in Scheme 3. The essential feature of this mechanism involves the intermediacy of a platinum vinyl carbonium ion which, once formed, reacts with the alcohol solvent to yield the alkoxycarbene ligand. A platinum vinyl carbonium ion may be formed by one of two routes; (i) an intramolecular mechanism involving an H− shift or (ii) an intermolecular mechanism involving deprotonation (abstraction of H+ by solvent) to give a platinum acetylide followed by protonation of the triple bond. Reactions carried out in CH3OD showed (21) that the acetylenic hydrogen was not lost to the solvent but rather was incorporated on the α-carbene carbon. This demonstrates that the intramolecular mechanism is operative and furthermore shows that the reaction of the vinyl carbonium ion with alcohol solvent is fast (at least relative to deprotonation) and irreversible. The exact role of platinum in facilitating this hydride shift is not clear and the intermediacy of a platinum-hydride cannot be excluded. However, in the reaction of cationic acetylenic intermediates of Pt(4+), which also yield the carbene ligand (35), formation of a platinum-hydride intermediate seems most unlikely since this would be formally a hydride of Pt(6+). Finally, it should be recognised that an alkyl shift rather than a hydride shift could be operative. Based on the relative migratory aptitudes of H− and R−, and the fact that acetylene, HC≡CH, gives the alkoxy (methyl) carbene ligand we prefer to believe in the H− shift mechanism. Of course, a 13C labelling experiment involving the specific labelling of one of the acetylenic sp-carbons could distinguish between these two possibilities. However, the salient feature of the proposed mechanism involves the intermediacy of a vinyl carbonium ion and is thus independent of which group migrates.
The concept of a platinum stabilised vinyl carbonium ion is an interesting one since simple vinyl carbonium ions are rather reactive species (37), indeed it is only recently that they have become respectable intermediates in the carbonium ion community (38). However metal vinyl carbonium ions could be stabilised by a metalloallene resonance form as shown above.
σ-Routes to Platinum Stabilised Carbonium Ions
If organoplatinum cations can induce carbonium ion reactivity in a π-bonded unsaturated hydrocarbon ligand and at the same time show evidence of stabilising the carbonium ion so formed, it is reasonable to expect that this carbonium ion stabilisation should be reflected in the reactions of σ-bonded unsaturated hydrocarbons.
We have now examined some reactions of divalent platinum acetylides of the type trans -Pt(C≡CR)2L2 and trans -PtCl(C≡CR) L2 where R=H, CH3, CF3 and C6H5, and L=P(CH3)2C6H5). These reactions are also of interest since they allow a direct comparison with the chemistry of organo-platinum compounds PtR2L2 and PtCl(R)L2 where R=alkyl or phenyl, and the presence of a four coordinate d8 metal and a coordinated alkynyl ligand provides the opportunity to compare the reactivity of inorganic and organic sites of unsaturation. Our initial findings show (39, 40) that these compounds exhibit a very intricate chemistry based on reaction at either the metal or the triple bond.
The reactions involving protic acids are particularly pertinent to the carbonium ion model outlined previously. For example trans -PtCl(C≡CR)(P(CH3)2C6H5)2 reacts:
(i) with anhydrous HCl to give almost quantitative formation of the vinylic product trans -PtCl(CCl=C(H)R)(P(CH3)2C6H5)2; (addition of HCl to the triple bond gives a mixture of cis- and trans -vinylic isomers but is specific in giving only the α-chlorovinylic isomer PtCCl=C(H)R);
(ii) with anhydrous HCl in methanol to give trans -PtCl(COCH2R)(P(CH3)2C6H5)2 and methylchloride;
(iii) with HPF6 in methanol to give quantitative formation of the methoxycarbene complex
where R=H, CH3 and C6H5. Addition of a one molar equivalent of LiCl to an acetone solution of
leads to quantitative formation of trans -PtCl (COC2H5)(P(CH3)2C6H5)2 with the formation of methyl chloride. Consequently we consider that propionyl formation in the reaction involving anhydrous HCl in methanol proceeds via initial formation of
which readily eliminates CH3Cl irreversibly.
These reactions provide further evidence for the intermediacy of platinum vinyl carbonium ions. However, the initial or preferential site of protonation is uncertain and again we must recognise the possible intermediacy of hydridoplatinum(4+) cations. Scheme 4 overleaf shows possible equilibria which may, either directly or indirectly, lead to platinum vinyl carbonium ions. Our work thus far has clearly demonstrated the existence of certain of these pathways. For example, cationic acetylenic intermediates yield alkoxy-carbene and acetylide complexes in alcohols and THF respectively. However, the intermediacy of hydrido-alkynylplatinum species remains a possibility. This potential route to metal vinyl carbonium ions leads us to predict a new synthetic route to alkoxycarbene ligands according to C.
Complexes suited to this type of reductive rearrangement could well involve Co, Rh and Ir.
Finally we note that the reaction of HPF6 in methanol with trans -PtX(C≡CH)L2, in which either the α or β acetylenic carbon is specifically labelled, could form the basis for ascertaining the intermediacy of a cationic π-acetylenic complex in this route to alkoxycarbene complexes.
As a further test of our hypothesis that low valent metals can stabilise vinyl carbonium ions we predict a direct solvolytic route to alkoxy-carbene ligands as in D where LnM provides an electron-rich metal site.
This type of solvolysis reaction is unprecedented in the reactions of simple organic vinyl halides but should be possible when the metal can assist in stabilising the vinyl carbonium ion as a metalloallene cation described previously.
where R=H, CH3 and C6H5 and L = P(CH3)2 C6H5.
(i) Alcoholysis reactions of PtX(CCl=C (H)R)L2 only occur when the Pt-X bond is labile towards Pt+X− solvolytic heterolysis e.g. X=Cl, Br and I.
(ii) Reaction of PtBr(CCl=C(H)R)L2 with AgPF6 in MeOH gives only the chloroplatinum complex
(Analysis showed <0.1 per cent Br).
(iii) Reaction in MeOD shows that the vinylic hydrogen is lost to the solvent.
(iv) trans -PtCl(CCl=C(H)Me)L2 dissolves in MeOH saturated with anhydrous NH3 to give trans -PtCl(C≡CMe)L2 and NH4Cl.
(v) trans -PtCl(CCl=C(H)Mc)L2 does not react with anhydrous NH3 in benzene; cf. (iv) above.
The observations (i)–(v) above demonstrate the importance of Pt-Cl heterolysis and that H+ is abstracted by the solvent. We cannot distinguish between a concerted or a stepwise elimination of the elements of HCl.
Subtle changes in the ligands bound to a metal often provide the basis for important modifications of catalytic processes. In this regard it is interesting to note that the reactivity of
in which both acetylenic hydrogens are incorporated in the α-carbene methyl group. (2) trans -Pt(C≡CH)2L2 reacts with DPF6 in MeOD to give
in which the acetylenic H/D ratio is ∼ 2:1 and the methyl H/D ratio is less than 1:6. (3) The reaction of DPF6 in MeOD with trans -PtCl(C≡CH)L2 and the reaction of trans -PtCl(CCl=CH2)L2 with AgPF6 in MeOD both yield
The above reactions indicate that the relative rates of H+ elimination versus carbene formation from
is sensitive to the nature of X. H+ elimination is favoured X=Cl>C≡CH>CH3.
H+ Catalysed Isomerisations of Vinylic Ligands
In a study of halovinyl complexes of platinum(II) Lewis and co-workers (43) noted that the reaction between trichloroethylene and PtL4, L =PMePh2, was stereospecific in giving
However, upon treatment with anhydrous HCl the apparent thermodynamic equilibrium distribution of vinylic isomers was obtained as indicated below.
Two plausible mechanisms were suggested (43) to account for this isomerisation—an addition-elimination sequence or a carbonium ion mechanism. Addition of HCl could occur to give either II or III.
Based on our concepts of platinum-stabilised carbonium ions we felt that the kinetically controlled mode of addition should lead to III. Consistent with this expectation was our observation (42) that the DCl catalysed isomerisation was accompanied by immediate vinyl-H/D scrambling as evidenced by the diminuation of the vinylic proton nmr signals. This observation does not rule out that addition of HCl may occur to give II but indicates that it is not the kinetically favoured mode of addition.
We have also found (42) that this vinylic isomerisation can be catalysed by the addition of each of the following acids: CF3COOH, HPF6 and HSO3F/SbF5 to CD2Cl2 solutions of
These observations lead us to believe that a simple protonation-deprotonation sequence involving the platinum-stabilised carbonium ion provides the basis for vinyl isomerisation.
In attempting to describe the organic reactions that occur within the coordination sphere of a metal it has often been implicitly assumed and sometimes explicitly stated (44) that the ability of a metal to act either as a source or a sink of electrons makes the common organic reactive intermediates carbonium ions and carbanions unfavourable. Contrary to this view there is now, in our opinion, a good deal of evidence to support the view that platinum may either induce or stabilise carbonium ion intermediates within its coordination sphere. Consequently we believe that the concept of metal-induced stabilised carbonium ions should be valuable both in interpreting and designing many organometallic reactions. Of course, there exists a certain fundamental problem in terminology. The reactivity of a metal-stabilised carbonium ion may be far removed from its simple organic analogue. In the extreme this may be compared to the dramatic difference in reactivity often observed for a carbene bonded to a metal (45–47) relative to its free organic existence (48). However, in regard to the reactions of organoplatinum complexes described above the carbonium ion terminology seems quite appropriate.
- 1W. C. Zeise, Mag. Pharm., 1830, 35, 105
- 2W. J. Pope and S. J. Peachey, J. Chem. Soc., 1909, 95, 571
- 3L. Chugaev,, M. Skanavy Grigorieva and A. Posniak, Z. Anorg. Chem., 1925, 148, 37
- 4G. Rouschias and B. L. Shaw, Chem. Commun., 1970, 183
- 5G. M. Whitesides,, J. F. Gaasch and E. R. Stedronsky, J. Am. Chem. Soc., 1972, 94, 5258
- 6G. M. Whitesides,, J. X. McDermott and J. F. White, J. Am. Chem. Soc., 1973, 95, 4451
- 7J. S. Bradley,, A. V. Kramer,, J. A. Labinger and J. A. Osborne, J. Am. Chem. Soc., 1974, 96, 7145
- 8A. V. Kramer and J. A. Osborne, J. Am. Chem. Soc., 1974, 96, 7832
- 9M. F. Lappert and P. W. Lednor, J. Chem. Soc., Chem. Commun., 1973, 948
- 10M. P. Brown,, R. J. Puddephatt and C. E. E. Upton, J. Organometal. Chem., 1973, 49, C 61
- 11M. P. Brown,, R. J. Puddephatt,, C. E. E. Upton and S. W. Lavington, J. Chem. Soc., Dalton Trans., 1974, 1613
- 12M. P. Brown,, R. J. Puddephatt and C. E. E Upton, J. Chem. Soc., Dalton Trans., 1974, 2457
- 13T. G. Appleton,, H. C. Clark and L. E. Manzer, J. Organometal. Chem., 1974, 65, 275
- 14H. C. Clark and L. E. Manzer, Inorg. Chem., 1973, 12, 362
- 15J. Chatt and B. L. Shaw, J. Chem. Soc., 1962, 5075
- 16R. Cramer and R. V. Lindsey, J. Am. Chem. Soc., 1966, 88, 3534
- 17H. C. Clark and H. Kurosawa, Inorg. Chem., 1972, 11, 1275
- 18A. J. Deeming,, B. F. G. Johnson and J. Lewis, Chem. Commun., 1970, 598; J. Chem. Soc., Dalton Trans., 1973, 1848
- 19H. C. Clark,, C. Jablonski,, J. Halpern,, A. Mantovani and T. A. Weil, Inorg. Chem., 1974, 13, 1541
- 20H. C. Clark and C. S. Wong, J. Am. Chem. Soc., 1974, 96, 7213
- 21M. H. Chisholm and H. C. Clark, J. Am. Chem. Soc., 1972, 94, 1532
- 22M. H. Chisholm and H. C. Clark, Accounts Chem. Res., 1973, 6, 202
- 23F. R. Hartley, “The Chemistry of Platinum and Palladium”, Applied Science Publishers, London, 1973
- 24T. G. Appleton,, H. C. Clark and L. E. Manzer, Coord. Chem. Rev., 1973, 10, 335
- 25M. H. Chisholm and H. C. Clark, Chem. Commun., 1970, 763
- 26M. H. Chisholm and H. C. Clark, Inorg. Chem., 1971, 10, 2557
- 27M. H. Chisholm,, H. C. Clark and L. E. Manzer, Inorg. Chem., 1972, 11, 1269
- 28M. H. Chisholm,, H. C. Clark,, L. E. Manzer and J. B. Stothers, J. Am. Chem. Soc., 1972, 94, 5087
- 29B. R. Davies and N. C. Payne, Can. J. Chem, 1973, 51, 3477
- 30E. Pignataro and B. Post, Acta Cryst., 1955, 8, 672
- 31M. J. S. Dewar, Bull. Soc. Chim. Fr., 1951, 18, C 79
- 32J. Chatt and L. A. Duncanson, J. Chem. Soc., 1953, 2939
- 33M. H. Chisholm and H. C. Clark, Inorg. Chem., 1971, 10, 1711
- 34M. H. Chisholm,, H. C. Clark and D. H. Hunter, Chem. Commun., 1971, 809
- 35M. H. Chisholm and H. C. Clark, Chem. Commun., 1971, 1484
- 36D. B. Crump and N. C. Payne, Inorg. Chem., 1973, 12, 1663
- 37M. Hannock, Accounts Chem. Res., 1970, 3, 209
- 38W. M. Jones and D. D. Maness, J. Am. Chem. Soc., 1969, 91, 4314
- 39M. H. Chisholm and D. A. Couch, J. Chem. Soc., Chem. Commun., 1974, 442
- 40R. A. Bell,, M. H. Chisholm,, D. A. Couch and L. Plauger, unpublished results
- 41R. A. Bell and M. H. Chisholm, J. Chem. Soc., Chem. Commun., 1974, 818
- 42R. A. Bell and M. H. Chisholm, unpublished results
- 43B. F. G. Johnson,, J. Lewis,, J. D. Jones and K. A. Taylor, J. Chem. Soc., Dalton Trans., 1974, 34
- 44P. M. Maitlis, “The Organic Chemistry of Palladium — Vol. 2 — Catalytic Reactions”, Academic Press, New York, 1968
- 45F. A. Cotton and C. M. Lukehart, Prog. Inorg. Chem., 1972, 16, 487
- 46D. J. Cardin,, B. Cetinkaya and M. F. Lappert, Chem. Rev., 1972, 72, 545
- 47E. O. Fischer, Pure Appl. Chem., 1972, 30, 353
- 48W. Kirmse, “Carbene Chemistry”, Academic Press, New York, 1964
The author is indebted to Professor H. C. Clark of the University of Western Ontario, Canada, with whom this work was initiated, to Dr. D. A. Couch and Mr. R. A. Bell, for their continuing work in this area; to Matthey Bishop, Inc. for their generous loan of platinum and to the Petroleum Research Fund and National Science Foundation for their current support of this work through grants PRF 7722-AC3 and NFS GP-4269X.