Inorganic Reaction Mechanisms Volume 5

A Review of the Literature Published Between January 1975 and June 1976

By A. McAuley

The Royal Society of Chemistry

Copyright © 1977 The Chemical Society
All rights reserved.
ISBN: 978-0-85186-295-8

Contents

Part I Electron Transfer Processes,
Chapter 1 Reactions Between Two Metal Complexes By R. D. Cannon, 3,
Chapter 2 Metal Ion–Ligand Redox Reactions By A. G. Lappin, 42,
Chapter 3 Reactions of Oxygen and Hydrogen Peroxide By A. McAuley, 107,
Part II Substitution and Related Reactions,
Chapter 1 Non-metallic Elements By G. Stedman, 121,
Chapter 2 Inert Metal Complexes: Co-ordination Numbers Four and Five By J. Burgess, 142,
Chapter 3 Inert Metal Complexes: Co-ordination Numbers Six and Higher By P. Moore, 162,
Chapter 4 Labile Metal Complexes By D. N. Hague, 240,
Chapter 5 Solvent Effects By J. Burgess, 260,
Part III Reactions of Biochemical Interest By O. N. Hague,
1 General, 277,
2 Metal Ion Transport through Membranes, 277,
3 Metal Complex Formation: Non-redox Systems, 280,
4 Reactions involving Metals in Porphyrins and Related Ring Systems, 290,
5 Redox Reactions involving Metals in other Biological and Model Systems, 299,
Part I V Organometallic Compounds By J. L. Davidson,
Chapter 1 Substitution, 305,
Chapter 2 Metal–Alkyl, –Aryl, and –Allyl Bond Formation and Cleavage, 333,
Chapter 3 Homogeneous Catalysis, 346,
Chapter 4 Insertion Reactions, 374,
Chapter 5 Reactions of Co-ordinated Ligands, 383,
Chapter 6 Oxidative Addition and Reductive Elimination, 398,
Chapter 7 Isomerization: Intramolecular Processes, 408,
Author Index, 440,

CHAPTER 1

Part I

ELECTRON TRANSFER PROCESSES

By

R. D. CANNON
A. G. LAPPIN
A. McAULEY

1

Reactions Between Two Metal Complexes

BY R. D. CANNON

1 General and Theoretical

Theory of the Electron Transfer Process. — A lengthy review of electron-transfer theory has appeared in a companion publication to this series, and a further review is promised in the next volume. New studies have also appeared of outer-sphere electron transfer in solution, covering the intermediate range between adiabatic and non-adiabatic conditions, and of inner-sphere transfer in the solid state. The relationship between electron transfer and magnetic exchange interactions has been pointed out. Antiferromagnetic coupling has been demonstrated in copper(II)–copper(II) dimers with no direct inner-sphere bridging ligands, and it is suggested that further work on such systems will lead to a better understanding of outer-sphere electron transfer in solution. There is continuing discussion of the possibility of electron transfer over long distances, but the evidence for such a process in solution still seems inconclusive.

Of more direct interest to kineticists is a reconsideration by Marcus and Sutin of the basis for the well known equation (1) relating activation free energies ΔG≠ to

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

standard free-energy changes ΔG[??]. Equation (1) is commonly illustrated by diagrams of the type of Figure 1 (cf. Figures 2 and 3, pp. 9 and 10). For restricted ranges of ΔG[??], it predicts an approximate linear correlation of ΔG[not equal to] with ΔG[??], with slope α [equation (2)], such that when ΔG[??] is close to zero, α is ca. 0,5, and this has been verified in several cases:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

However, equation (2) also predicts that as ΔG[??] becomes more negative α decreases, until when ΔG<–4A further decrease in ΔG[??] leads to an increase in ΔG≠. Marcus and Sutin now point out that this is not to be expected. The region ΔG[??]<–4A corresponds to intersection of reactants’ and products’ curves on the left-hand side of the diagram (x<0, point b, Figure 1), in contrast to ‘normal’ intersections at x > 0 (point c), and this requires a non-adiabatic transition. It is suggested that for these transitions nuclear tunnelling may become important, and that as a result ΔG[not equal to] will not increase with decreasing ΔG[??] but will remain constant, with the specific rate at the diffusion-controlled limit. An analogous situation has indeed already been recognized for electron transfer between organic molecules. Over a range of ΔG[??] from -5 to +5 kcal mol-1, Rehm and Weller obtained a parabolic plot of ΔG[not equal to] against ΔG[??], but fromΔG[??] = –5 to –62 kcal mol-1 ΔG[??] remained constant. Another theory based on the non-adiabatic description leads to a linear dependence of ΔG[??] on ΔG≠ in this region, but constancy of ΔG≠ can also be explained in a different way: it has been suggested that when the ground-state products’ curve passes the minimum on the reactants’ curve (point d), other curves for higher electronic levels become available, as for example the dotted curve (iii) in Figure 1; hence the reaction may still proceed with zero thermal activation energy along the route a -> d -> e. More sophisticated quantum mechanical models involving multiple energy states have also been discussed.

The minimum value of ΔG≠ (on the old theory) or the limiting value (on the new) is expected to be zero only if the work terms are negligible, and if there are no rapid equilibrium steps preceding the electron-transfer process. Previously, Hyde et al. had found a non-linear correlation of ΔG≠ with ΔG[??], for a reaction of a series of reductants with the common oxidant [Co(H2O)6]3+. The limiting value of ΔG≠ was found to be ca. 9 kcal mol-1. Ekstrom et al. have now added an additional point for the reaction Co3+ + U3+, which fits well on to the curve, but they point out that data for a large number of cation-cation reactions involving oxidants other than [Co(H2O)6]3+ also cluster quite well round the same curve and Falcinella et al. have produced an analogous correlation for reactions of Tl2+, acting alternatively as oxidant and reductant.

The limiting value, ΔG≠ ≈ 9 kcal mol-1, had previously been felt to be rather high, and one possible explanation considered was a rapid equilibrium between the two spin states of cobalt(III) ([MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]) prior to the electron transfer. The data from other oxidants have weakened if not altogether removed the necessity for this. It has moreover been argued from ligand-field considerations that for the hexa-aquo-cobalt(III) ion the spin-change free energy may in any case be quite small.

It is accordingly now argued that the high limiting value is sufficiently accounted for by the work terms, i.e. the limiting rate constant does represent the diffusion-controlled value for the reactants in equation. The specific collision rate calculated by the Debye equation for the reaction U3+ + Co3+ is 1.7 × 104 mol-1 s- 1 (at 25 °C, zero ionic strength), to be compared with the experimental 7.1 × 103 mol- 1 s-1. In theory there should be different limiting rates depending on the ionic charges, and it is perhaps a cause for concern that reactions as diverse as Co3+ + NpO2+, UO22+ + Eu2+, and Fe3+ + U3+ all fall near the same curve.

Activation Parameters. — Negative heats of activation have previously been reported for some outer-sphere reactions,’ and it has been suggested that some special factor, not covered by the simple Marcus theory, might need to be invoked to explain them. (It will be recalled that negative heats of activation of some inner-sphere reactions have been rationalized in terms of the energetics of formation of precursor complexes.) Marcus and Sutin point out, however, that negative ΔH≠ values are theoretically predicted, under certain conditions. On differentiating equation (1) with respect to temperature, they obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3a)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3b)

where ΔHA and ΔHA are activation parameters corresponding to the intrinsic free-energy barrier A [equation (1)] and β = 2ΔG[??]/A The term β may be positive or negative, but is usually small; hence if the overall enthalpy change is sufficiently negative ΔH≠12 may be negative even when ΔHA is positive. A complete calculation requires a knowledge of ΔHA and ΔHA, from the appropriate cross-reactions, as well as of ΔH[??] and ΔS[??]. For two reactions, [Fe(bipy)3]3+ and [Ru(bipy)3]3+ with Fe2+, the calculated parameters agree with experiment at least as regards the negative sign, i.e. ΔH≠ = –3.2 and –2.9 kcal mol-1 respectively by calculation, –0.8 and –0.3 kcal mol-1 by experiment. Further confirmation is provided by new measurements on related systems such as [Ru(bipy)(NH3)4]2+ + [Fe(H2O)6]3+: a graph of (ΔH≠ – ½ΔH≠) against ½ΔH[??](1 + 2β) conforms reasonably well to the expected straight line of unit slope for seven systems over a range of ca. 25 kcal mol-1. The actual rate data have not yet been reported.

Stereoselectivity. — An earlier report that reactions between chiral oxidants and reductants could show stereoselectivity, e.g. that (+)D-[Co(phen)3]3+ reacted at different rates with the two optical isomers of [Cr(phen)3]2+, has been shown to be in error.

2 Intramolecular Electron Transfer

The Precursor Complex. — Interest in intramolecular electron-transfer processes continues to grow, and the search continues for additional long-lived dinuclear complexes preceding the electron-transfer reaction proper.

Jwo and Haim have reported reactions of [FeII(CN)5(H2O)]3- with the isomeric oxygen-bonded pyridinecarboxylatopenta-amminecobalt(III) complexes. With the 2-isomer there is no reaction, presumably for steric reasons; with the 3-isomer a dinuclear complex is produced which shows no tendency to react further (ket ≤ 3 × 10-5 s-1), but with the 4-isomer the reaction sequence is as shown in equation (4), all steps proceeding at measurable rate (kt = 1.5 × 103 mol-1 s-1, kd = 2.5 × 10-3 s-1, ket = 1.75 × 10-4 s-1).

[FORMULA NOT REPRODUCIBLE IN ASCII] (4)

The reactions of [Co(NH3)5X]2+ (X = F, Cl, or Br) with iron(II) in DMF conform to the Michaelis–Menten rate law (5) where the value of K3 is interpreted as the

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

equilibrium constant of formation of dinuclear complexes [Co(NH3)5XFe]4+. This is the first report of such complexes involving like-charged reactants. It should be noted, however, that the values of K3 are quite small (e.g. K3= 19 1 mol- 1 for X = Cl) and although the ionic strength was maintained constant with KClO4 the range of iron concentrations was such as to require complete replacement of K+ ion by Fe2+ at the highest concentrations used.

Haim and Sutin have studied the rapid reversible reaction (6). Although both the

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

precursor and the successor complexes are formed of oppositely charged ions, they could not be detected, and only upper limits could be placed on their stability constants, of 20 and 100 l mol-1 respectively. This contrasts with the earlier results on the reaction [Co(NH3)5(OH2)]3+ + [Fe(CN)6]4-, where a precursor ion pair was observed, with formation constant 1500 l mol-1. The difference between the two systems underlines the importance of specific, short-range interactions, as opposed to overall Coulombic attraction, in determining the stabilities of ion pairs. Nevertheless, the limiting values cited above are close to the actual values predicted by the Fuoss equation, and an argument based on the Marcus theory has been invoked to show that these are in fact the correct orders of magnitude. A rate constant k for reaction (6) can be predicted from the rate constants of the appropriate cross-reactions together with the measured equilibrium constant. The experimental value of k is ca. 3000 times higher than predicted, and this may be explained by the fact that the Coulombic effect in equation (6) is favourable to electron transfer, but the Coulombic effects in the cross-reactions are unfavourable. By choosing reasonable values for the stabilities of all three precursor complexes, based on electrostatic effects, it is possible to recalculate the predicted k and obtain close agreement with experiment. Thus the role of the precursor complexes in determining rates is demonstrated even though the complexes themselves cannot be detected. The argument is similar to that previously put forward by Fay and Sutin to rationalize the relative rates of electron-transfer reactions with NCS, NNN, and SCN bridging groups.

An example of reversible intramolecular electron transfer has now been reported though not for a metal–metal system. It had previously been shown that the complexes [NiII(TPP)] (TPP2 = tetraphenylporphinate) when electrochemically oxidized in benzonitrile solution yield first the brown-coloured nickel(III) complex, which then decays at a measurable rate by a process of ligand-to-metal electron transfer [reaction (7), forward step; (TPP·)- = radical-ion]:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

When the solvent is changed to methylene dichloride this process is apparently more rapid since electrolysis of [NiII(TPP)] leads directly to the green [NiII (TPP·)]+ with no detectable intermediate. On cooling to 77 K, however, the green solution changes to an orange solid, and e.p.r. data are consistent with the reverse of reaction (7). The optical spectra of the two complexes are reported.

Another example has been reported of a direct comparison of rates of intramolecular and intermolecular electron transfer between the same or very similar pairs of oxidizing and reducing centres. Previously, Hofmann and Simic had generated the complex [CoIII(NH3)5(O2CC6H4NO2- p)]+, containing the p-nitrophenylbenzoate radical-ion, and had measured the rate of reaction (8). Cohen and Meyerstein have

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)

now obtained rate constants of reaction (9) (X = NH3, p– NO2C6H4CO2-, or various

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (9)

other ligands). Interpreting their rate constant k2 as Kk1, where K is the stability constant of the outer-sphere precursor complex and k1 the specific rate of electron transfer within the precursor complex, and assuming maximum values of K consistent with the observed rate law, they conclude that the intramolecular reaction (8) is not significantly faster than any of the intermolecular outer-sphere reactions. These results confirm that the carboxylate group is a poor mediator of electrons, and this is borne out by the previous observation that intramolecular transfer in [CoIII(NH3)5- (O2CC6H4NO2-o)]+ is more than 102 times faster than in the para-isomer. It seems likely that the electron transfer is a ‘through-space’ interaction, not directly involving the carboxylate bridge, as has also been suggested in other cases.

(Continues…)Excerpted from Inorganic Reaction Mechanisms Volume 5 by A. McAuley. Copyright © 1977 The Chemical Society. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.