View Online / Journal Homepage / Table of Contents for this issue

The 16 and 18 Electron Rule in Organometallic

Chemistry and Homogeneous Catalysis

By C. A. Tolman
c E N T RA L R E S E A R C H DEPARTMENT* E X P E R I M E N T A L STATION,
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

E. I . D U P O N T D E N E M O U R S A N D C O M P A N Y , W I L M I N G T O N ,
D E L A W A R E 1 9 8 9 8 , U.S.A.

1 Introduction
Downloaded by University of Sussex on 10 October 2012

There has been a great increase in interest and activity in recent years in organo-
metallic chemistry, especially in the area of homogeneous catalysis by transition-
metal complexes. A number of publications have appeared dealing with par-
ticular organometallic reactions which occur in catalysis, such as oxidative
addition,lJ insertion,, ligand e~change,~ and reactions of complexes with Lewis
acids.6A few authors have attempted to outline the principles of catalysis more
There appears, however, to be no satisfactory scheme which relates
the various reaction types or which permits mechanistic predictions.
In this review the various types of organometallic reactions involved in
catalysis are organized in a systematic way, and some empirical rules are pre-
sented which may be used predictively to restrict the kinds of organometallic
compounds which may exist under mild conditions and the types of reactions
they may undergo.
The tendency of transition metals to form complexes in which the metal has
an effective atomic number corresponding to the next higher inert gas has long
been recogni~ed.~ There are, however, many exceptions, as illustrated by the
following examples:
Compound NVE
TiClc 8
MeAuPPh, 14
Co(CN),S- 17
Ni(HzO)5z+ 20
The number of valence electrons (NVE) consists of the valence electrons of the
* Contribution No. 1821.
J. P. Collman, Accounts Chem. Res., 1968, 1, 136; J. P. Collman and W. R. Roper, Adv,
Organometallic Chem., 1968, 7 , 53.
J. Halpern, Accounts Chem. Res., 1970, 3, 386.
a R. F. Heck, Adv. Chem. Ser., 1965, 49, 181.
C. H. Langford and H. B. Gray, ‘Ligand Substitution Processes’, Benjamin, New York.
1965.
D. F. Shriver, Accounts Chem. Res., 1970, 3, 231.
J. Halpern, Discuss. Faraday SOC.,1968, 46,7.
J. Halpern, Adv. Chem. Ser., 1968, 70, 1.
* R. Ugo, Chimica e Industria, 1969, 51, 1319.
N. V. Sidgwick, ‘The Electronic Theory of Valency’, O.U.P.,London, 1929, p. 163.

337
 16 and 18 Electron Rule in Organometallic Chemistry and Homogeneous Catalysis
View Online
metal and those electrons donated by or shared with the ligands, and would be
18 for an inert-gas configuration. If, however, one restricts attention to the
diamagnetic organometallic complexes of Groups IVB-VIII, essentially all of
the well-characterized compounds have 16 or 18 metal valence electrons.lOsll
This fact has not generally been adequately appreciated by organometallic
chemists. In fact, the literature contains numerous examples of organometallic
compounds which, as formulated, appeared to be exceptions to what can be
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

called the 16 and 18 Electron Rule. Careful subsequent study has almostinvari-
ably shown that the original formulation was incorrect. The accessibility of 16
and 18 electron configurations also has important consequences for mechanisms
of organometallic reactions, as outlined below.
Organometallic compounds, for the purposes of this discussion, are transition-
Downloaded by University of Sussex on 10 October 2012

metal complexes containing one or more ligands such as CO, N2, CN-, RNC,
PR3, P(OR)3, olefin, acetylene, n-allyl, 7r-cyclopentadienyl, 7r-aryl, acyl, -SiR3,
-R, or -H. These ‘soft’lBor ‘class b’lSligands are characterized by a high ligand-
field strength and covalent bonding character and are the types typically found
in homogeneous catalytic reactions such as hydrogenation, hydroformylation,
hydrosilylation, and olefin isomerization and oligomerization.

The 16 and 18 Electron Rule.-Two postulates or rules for organometallic com-

plexes and their reactions are proposed.
1. Diamagnetic organometallic complexes of transition metals may exist in a
significant concentration at moderate temperatures only if the metal’s
valence shell contains 16 or 18 electrons. A significant concentration is one
that may be detected spectroscopically or kinetically and may be in the
gaseous, liquid, or solid state.
2. Organometallic reactions, including catalytic ones, proceed by elementary
steps involving only intermediates with 16 or 18 metal valence electrons.
It is apparent that application of these rules requires that care be exercised in
reaching conclusions on the NVE of a metal complex. Association or dissociation
of the compound may occur. For example, Ni[PPh3I4is substantially dissociated
in solution into Ni[PPh3I3and PPh3.14r-Ally1 palladium chloride is a 16 electron
complex, with chloride ions bridging in a dimeric Polyfunctional
ligands may co-ordinate at more than one site, and additional M-C and M-H
bonds may form by reaction of a metal with C-H bonds of the ligands.ls
When the complex is in solution, there is a potential ambiguity as to whether
solvent is co-ordinated. The mere presence of solvent in the analysis of a crystal-
line solid is not good evidencefor solvent co-ordination. One convenient criterion
lo F. Basolo and R. G. Pearson, ‘Mechanisms of Inorganic Reactions’, Wiley, New York,
1967, p. 526.
I1 G. E. Coates, M. L. H. Green, and K. Wade, ‘Organometallic Compounds’, 1’01. 2, ‘The
Transition Elements’, Methuen and Co., Ltd., London, 1968, p. 6.
l* R. G. Pearson, Science, 1966, 151, 172.
S. Ahrland, J. Chatt, and N. R. Davies, Quart. Rev., 1958, 12, 265.
l4 C. A. Tolman, W. C. Seidel, and D. H. Gerlach, J. Amer. Chem. SOC.,1972, 94, 2669.
l6 W. E. Oberhansli and L. F. Dahl, J. Organometallic Chem., 1965, 3, 43.
G. W. Parshall, Accounts Chem. Res., 1970, 3, 139.

338
 Tolman
View Online
for solution studies is whether the electronic spectrum of the complex is sub-
stantially different in different solvents. Frequently a coloured complex will give
dramatic colour changes when dissolved in a co-ordinating solvent. For example,
N i [ P ( O ~ - t o l y l ) ~a] ~red-orange
,~~ solid, gives red-orange solutions in benzene,
tetrahydrofuran, or methylene chloride, in which it remains three-co-ordinate,
but gives a colourless solution in acetonitrile, from which the unstable adduct
MeCNNi[P(Oo-tolyl),I3 has been isolated.’
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

2 Types of Organometallic Reactions

Reactions of ligands which do not directly involve the transition metal are ex-
cluded from this classification.Some examples of such reactions are methanolysis
of Ni(PF& to give Ni(PF,) ,[PF,-n(OMe)n], Friedel-Crafts acylation of ferro-
Downloaded by University of Sussex on 10 October 2012

cene, and nucleophilic attack on a co-ordinated carbonyl by methoxide ion.

Reactions which involve the transition metal directly can be broken down into
five elementary reactions, each with a microscopic reverse. The classification is
based upon the changes in number of metal valence electrons, formal oxidation
state, and co-ordination number which accompany each reaction. The five
elementary reactions are shown with examples in the Table. It will be seen that
reactions 1 and 2 involve dissociation and association of ligands whereas 3-5
involve reactions between co-ordinated ligands. The latter can be classified
according to the parity of the groups coupled, considering them to be uncharged.
Thus coupling of two hydrogen atoms or two m-ally1 groups (odd-odd coupling)
is a reductive elimination. Coupling of an odd-electron fragment, such as a
n-allyl, with an even one, such as ethylene, is an insertion. Even-even coupling,
as between two m-bonded olefin molecules, is oxidative coupling.
Some of the terms in the Table are used in a way which may be unfamiliar to
the reader and warrant further discussion, The term ‘Lewis base ligand’ is used
to include ligands such as CO which are not normally considered as bases, in
addition to more usual ones such as phosphines, on the basis that they contribute
two electrons to the metal’s valence shell. Lewis acid ligands, such as H+, BF,,
and SO,, contribute no valence electrons.
In common parlance ‘oxidative addition’ is used to describe an over-all reac-
tion in which the formal oxidation state of the metal and the co-ordination
number increase by either one or two.2 In this discussion only a concerted one-
step reaction which increases both co-ordination number and oxidation state
by two is termed an oxidative addition. The narrower definition is used in order
to break reactions down into elementary steps and to distinguish between differ-

-
ent types of mechanisms. An increase in co-ordination number by one as in

H++CO(CO)[ HCo(CO),

L. W. Gosser and C.A. Tolman, fnorg. Chem., 1970, 10,2350,

C. A. Tolman, Inorg. Chem., 1971, 7, 1540.

339
 Downloaded by University of Sussex on 10 October 2012
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

Table Elementary organometallic reactions

Reverse ri.
Reaction d W E a dOSb dNc Example Reaction dNVEdOSdN $
1. Lewis acid Lewis acid
ligand dissociation 0 0 -1 CpRh(C2H4),S02f SO, CpRh(C,H,),+ association 0 0 +l
zj$
or 0 -2 -1 HCo(CO), + H+ CoC0,- + or 0 +2 + 1 0
4
2. Lewis base Lewis base
ligand dissociation -2 0 -1 NiL4 +NiL, +L association +2 0 +1
43
3. Reductive Oxidative
elimination -2 -2 -2 H,IrCI(CO)L,
+ H, + IrCl(CO)L, addition +2 +2 +2
3
0
4. Insertion -2 0 -1 MeMn(CO), + MeCOMn(CO), Deinsertion +2 0 +1 $
CFZ-CF,
5
I \ .;I
\ Reductive 5
5. Oxidative coupling -2 +2 0 a4
Fe(c0)3 decoupling +2 -2 0
/

a Change in the number of metal valence electrons; Change in the formal oxidation state of the metal. The usual convention which regards hydrides,
3
alkyls, r-allyls, and 7r-cyclopentadienyls as uninegative ions is used ; C Change in co-ordination number. 2
E
View Online
 View Online
is regarded as a Lewis acid ligand association reaction. A two-step reaction, as
in
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

is regarded as a Lewis acid ligand association followed by a Lewis base ligand

association.
Reaction of molecules such as 0, or tetracyanoethylene with a transition
metal can be regarded as an oxidative addition in which only part of a multiple
bond is broken to form two new metal-ligand bonds, leaving the two Iigands
Downloaded by University of Sussex on 10 October 2012

joined. With less electronegative olefins, such as C,H4 or C2F4, it may be more
convenient to regard their co-ordination as a Lewis base addition rather than as
an oxidative addition. The real extent of electron transfer must be determined by
ESCA or some other means. The count of metal valence electrons in either case
is the same.
The terms 'oxidative coupling' and 'reductive decoupling' are new. The
former is used to denote reactions such as that shown in reaction (4), in which the
formal oxidation state of the metal increases by two but the co-ordination
number does not change.

3 Applications of the Rule to General Organometallic Reactions

The 16 and 18 Electron Rule severely restricts the types of reaction which a
particular complex may undergo. Dissociation or association of Lewis acid
ligands may occur with either 16- or 18-electron complexes. Lewis base ligand
dissociation, reductive elimination, insertion, and oxidative coupling are re-
stricted to 1%electron complexes. Lewis base ligand association, oxidative
addition, deinsertion, and reductive decoupling reactions can occur only with
16-electron complexes. A number of examples from the literature will serve to
illustrate these points.

A. Lewis Acid Dissociation-Association.-That Lewis acids may react with

either 16- or 18-electroncomplexes is shown by the formation of adducts of BF,
with 16-electron IrC1(CO)(PPh,),ls or with 18-electron HR~(T-C~HJ,.~O Other

l9 R.N.Scott, D. F. Shriver, and L. Vaska, J. Amer. Chem. SOC.,1968,90, 1079.

*O M.P. Johnson and D. F. Shriver, J. Amer. Chem. Soc., 1966, 88, 301.

341
 16 and 18 Electron Rule in Organometallic Chemistry and Homogeneous Catalysis
View Online
examples of reactions of 18-electron complexes with Lewis acids include forma-
tion of [(n-C5H5)Rh(CO)PMePh2Cl]+from ( ~ T - C ~ H , ) R ~ ( C O ) Pand
M ~C1221
P~~
and of HNi[P(OEt),14+from Ni[P(OEt),], and H+.22

B. Lewis Base Dissociation-Association.-Ligand-exchange reactions are the

most thoroughly studied type of organometallicreaction, They generally proceed
via Lewis base ligand dissociation as a first step for 18-electron complexes and
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

via ligand association for 16-electron complexes. Examples include exchange

of CO from Ni(CO)4by L = CO or PPh3in reaction (5),23 displacement of H 2 0
fr0m[Co(CN),(H,0)]~- by X- = N3-or SCN- in reaction (6),24and the reaction
Downloaded by University of Sussex on 10 October 2012

of pyridine with trans-RPtCI(PEt3)2c~mplexe~in reaction (71a51nthese reactions

the number beneath each complex indicates the NVE.

The initially puzzling observations by Cramer2s that (acac)Rh(C,H,), ex-

changes ethylene rapidly on an n.m.r. time scale, whereas (n-C,H,)Rh(C,H,),
is very inert to exchange are readily explained. The acetylacetonate complex can
exchange via the associative mechanism shown in reaction (8). For the 18-electron
(n-C,H,)Rh(C,H,), the associative pathway is blocked. The 16-electron (C,H,)-

*l A. J. Oliver arid W. A. G. Graham, Inorg. Chem., 1970, 9, 243.

la W. C. Drinkard, D. R. Eaton, J. P. Jesson, and R. V. Lindsey, jun., Znorg. Chem., 1970,
9, 392.
*3 J. P. Day, F. Basolo, and R. G. Pearson, J. Arner. Chem. SOC., 1968, 90, 6927.
a4 A. Haim and W. K. Wilmarth, Znorg. Chem., 1962, 1, 573.
a5 F. Basolo, J. Chatt, H. B. Gray, R. G . Pearson, and B. L. Shaw, J. Chem. Soc., 1961,2207.
** R. Cramer, J . Arner. Chem. Soc., 1967, 89,4621.

342
 Tolman
View Online
Ni[P(Oo-tolyl),], exchanges ethylene very rapidly but can be recovered un-
changed after prolonged exposure of a solution to vacuum.27Again, the 16-
electron complex can associate a two-electron ligand but does not dissociate.

C. Reductive Elimination-Oxidative Addition.-The kinetics of reaction (9),

Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

involving reductive elimination of H,, have been studied.28The rate of the reac-
tion is independent of the concentration and nature of L1[L1= CO, P(OMe),,
Downloaded by University of Sussex on 10 October 2012

P(OPh),, or PMePhJ as required for rate-determining loss of H, in the mechan-

ism shown in reaction (10).The 16-electronintermediatewas actually isolated for
Ls = P(CaH11)3.

Oxidative addition of H, to the 16-electron complex IrCl(CO)(PPh,), is first-

order in both H, and Ir complex,2Das required for a concerted reaction. Reac-
tion of the 18-electron complex HIr(CO)(PPh,), with HSi(OEt),, as expected,
does not occur in a single ~tep.~OThemechanism found is shown in reaction (11)
(L = PPh3, R = OEt).

The need to dissociate a ligand from an 18-electron complex before oxidative

addition can occur is also shown in the reactions of H, or HCI with EtCOM-
(CO),(PPh,),. High concentrations of CO inhibit the formation of aldehyde.s1
Pearson has recently shown that optically active MeCHBrC0,Et reacts with
Mn(CO),- with inversion of configurati~n.~~ The 18-electron complex cannot
undergo an oxidative addition without first losing a CO ligand. Since CO dis-
sociation from Mn(CO),- is extremely slow,,, the reaction is forced to proceed
in a way which does not increase the NVE of Mn. The S NWalden ~ inversion
which occurs at the asymmetric carbon can be regarded as a special case of Lewis

W. C. Seidel and C. A. Tolman, Inorg. Chem., 1970, 9, 2354.

*8 M. J. Mays, R. N. F. Simpson, and F. P. Stefanini, J . Chem. SOC.(A), 1970, 3000.
19 P. B. Chock and J. Halpern, J. Amer. Chem. SOC.,1966, 88, 3511.
3 0 J. F. Harrod and C. A. Smith, Cunad, J . Chem., 1970,48, 870.
31 G. Yagupsky, C. K. Brown, and G. Wilkinson, J. Chem. Soc. (A), 1970, 1392.
32 R. W. Johnson and R. G. Pearson, Chem. Comm., 1970,986.
33 W. Hieber and K. Wollmann, Chem. Ber., 1962, 95, 1552, found no exchange with "CO
in 20 h at 40 "C.

343
 16 and 18 Electron Rule in Organometallic Chemistry and Homogeneous Catalysis
View Online
acid ligand association, the bromine leaving with the electron pair it had shared
with carbon (reaction 12).

H Me

MeCHBrCOzEt + Mn(CO)~’+[(COlj M n ----\C/

Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

&+I - - - 0
Brl’
&-
(12)

C02 Et
Downloaded by University of Sussex on 10 October 2012

On the other hand, reaction of IrCl(CO)(PMePh,), with MeCHBrC0,Et

is reported to proceed with retention.34A concerted oxidative addition with
retention of configuration is possible for the Ir complex because it has 16 valence
electrons. Lewis acid association with inversion is also possible.

D. Insertion-Deinsertion.-The insertion of CO in the reaction of 18-electron

MeMn(CO), with L = PPh3 or P(OPh)3 involves CO insertion as a first step,
followed by reaction of the 16-electron intermediate with L, as shown in reac-

-
tion (13).35 Insertion followed by ligand association is also consistent with the

MeMn(CO), MeCOMnlCO),
7-
MeCoMn(Co)4L (1 3)
18 16 L 18

result that treatment of MeMn(CO), with 14C0 gives an acyl with a non-
labelled CO;36some elegant experiments with 13C0have demonstrated that the
methyl group moves3’
Another type of insertion reaction is illustrated by reaction (14).38The existence
of a hydrido-olefin intermediate has been supported by deuterium-labelling
experiments3

34 R. G . Pearson and W. R. Muir, J. Amer. Chem. SOC.,1970,92, 5519.

3b R. J. Mawby, F. Basolo, and R. G. Pearson, J. Amer. Chem. SOC.,1964, 86, 3994.
ssT. H. CoffieId, J. Kozikowski, and R. D. Closson, J. Org. Chem., 1957, 22, 598; ‘Internat.
Conf. on Co-ord. Chem., London, 1959’, The Chemical Society, London (Special Publication
No. 13), 1959, p. 126.
K. Noack and F. Calderazzo, J. Organometallic Chem., 1967, 10, 101.
3 8 J. Chatt and B. L. Shaw, J. Chem. SOC.,1962, 5075.
J. Chatt, R. S. Coffey, A. Gough, and D. T. Thompson, J. Chem. SOC.(A), 1968,190.

344
 Tolman
View Online
E. Oxidative Coupling-Reductive Decoup1ing.-An example of oxidative coup-
ling is given by reaction (15),4O where R = C0,Me and L = PPh3. The reaction

R
\ /R

I
C
/"
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

L\
(N,) IrClL, + 2RC"-CR CL - I t +N2
/\
L C C

R
/ \R
Downloaded by University of Sussex on 10 October 2012

probably proceeds via oxidative coupling of the acetylenes in an (RC,R)IrClL,

intermediate.
Another example is given in reaction (16).41The details of the reaction are not

known; however, it probably proceeds by oxidative coupling of the olefins in

(C2F3H),Ni(PPh3)2. There is a precedent for (olefin),NiL, complexes in
(CH ,=CHCN) %Ni(PPh3) 2.42

4 Applications of the Rule to Homogeneous Catalytic Reactions

Homogeneous catalytic reactions proceed by the same elementary steps (see the
Table) as general organornetallic reactions. The unique feature of a catalytic
system is that a series of steps is connected in a cyclic way, to form a loop.
Mechanisms of some well-studied catalytic processes will now be described in
detail.

A. Hydroformylation of Olefins by H C O ( C O ) , . ~ ~ - Thydroformylation

~~ of a
terminal olefin is shown in Figure 1. The electron configurations and co-ordina-
tion numbers in the d n N n ~ t a t i o nare
~ ~shown by each complex. The mechanism
is basically that of Heck and B r e ~ l o w but , ~ ~ modified to show explicitly the

40 J. P. Collman, J. W. Kang, W. F. Little, and M. F. Sullivan, Inorg. Chem., 1968, 7 , 1298.

41 J. Ashley-Smith, M. Green, and F. G. A. Stone, J. Chem. Soc. (A), 1969, 3019.
4a G. N. Schrauzer, J. Amer. Chem. SOC., 1960, 82, 1008.
49 For a review see A. J. Chalk and J. F. Harrod, Adv. Organometallic Chem., 1968, 6, 119.
44 dn gives the number of d electrons for a particular oxidation state while N is the co-ordina-
tion number.
I 6 R. F. Heck and D. S. Breslow, J. Amer. Chem. SOC., 1961,83,4023.

345
 16 and 18 Electron Rule in Organornetallic Chemistry and Homogeneous Catalysis
View Online
oxidative addition of hydrogen and reductive elimination of aldehyde. The first
step of the reaction, starting with HCo(CO),, is Lewis base ligand dissociation
from the 18-electron complex to give HCo(CO),. The reactions proceeding
clockwise around the loop are then: (2) olefin (Lewis base) association to give an
18-electron hydrido-olefin complex; (3) insertion of olefin to give a 16-electron
alkyl; (4) association of CO to give an 18-electron alkyl; (5) insertion of CO to
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

give a 16-electron acyl; (6) oxidative addition by H, to give an 18-electron acyl

cobalt dihydride; and (7) reductive elimination of aldehyde product with re-
generation of HCo(CO),. Step (8) connects the loop reversibly to a species not
on the loop. Inhibition of hydroformylation by high CO pressure is due to
reduction in the concentrations of loop species by step (8) and the reverse of (1).
Downloaded by University of Sussex on 10 October 2012

d '4.

Figure 1 Hydroformylation of a terminal olefin by HCo(CO),

Figure 1 has been simplified in that it does not show formation of internal
aldehydes, hydrogenation of aldehydes to alcohols and of olefins to alkanes, or
olefin isomerization, which also occur in the system.

B. Hydrogenation of Olefins with RhCI(PPh,),.-Application of the 16 and 18

Electron Rule to olefin hydrogenation with Wilkinson's catalyst45is shown in

346
 Tolmati
View Online
Figure 2. Two loops arise from the possibility of co-ordinating hydrogen and
olefin in either order. Step (1) represents oxidative addition of hydrogen to the
square-planar 16-electron complex to give an 1%electron dihydride. Step (2),
Lewis base dissociation, is followed by olefin co-ordination in step (3). Insertion

RCHsCH;!
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

d84
I
Downloaded by University of Sussex on 10 October 2012

RhClL3
d85

d04

Figure 2 Hydrogenation of a terminal olefin by RhCI(PPhJ3

in step (4) gives an unstable hydrido-alkyl with NVE = 16. Phosphine associa-
tion in step ( 5 ) gives an 18-electron hydrido-alkyl which produces alkane by
reductive elimination, regenerating RhCI(PPh3),. Wilkinson's isolation of
(C2H4)RhClL2and observation that the ethylene complex is not readily hydro-
genatedq6suggest that step (9) is very slow, so that the bulk of the reaction goes
via the loop which contains steps (1)-(6). The inhibition of hydrogenation on
addition of triphenylphosphine to the system can be understood in terms of its
suppression of ligand dissociation in step (2). The mechanism in Figure 2 is
somewhat different from the one originally proposed, which included phosphine
ligand dissociation as the first step in the reaction and simultaneous addition
of both hydrogens to the double bond. The phosphine-dissociation step was
based on a low value for the molecular weight of RhCI(PPh3)3.Molecular weight
determinations4' under scrupulously oxygen-free conditions indicated that the
46 J. A. Osborn, F. H. Jardine, J. F. Young, and G . Wilkinson, J . Chem. SOC.(A), 1966,1711.
47 D. D. Lehman, D. F. Shriver, and I. Wharf, Chem. Comm., 1970, 1486.

347
 16 and 18 Electron Rule in Organometallic Chemistry and Homogeneous Catalysis
View Online
complex does not dissociate to any great extent. This conclusion was also sup-
ported by n.m.r. studies.48 Studies49employing a combination of spectro-
photometry and 31P and lH n.m.r., indicate that RhCl(PPh3)3 dissociates
only to a very small extent to give a dimer and that its reaction with H, gives
H2RhCl(PPh,),,60 rather than HZRhCl(PPh3),S (S = solvent) as originally
Dissociation to H2RhC1(PPh3),is slight, but does provide a mecha-
nism for ligand exchange.
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

Stepwise rather than simultaneous addition of hydrogens during olefin hydro-

genation is indicated by deuteriatiodl and olefin isomerization studies.s2
It should be noted that both formation of RhCl(PPh3)2 and simultaneous
addition of hydrogens to the olefin are prohibited by the 16 and 18 Electron
Rule. RhCl(PPh,), is already a 16-electron complex and would drop to 14 if
Downloaded by University of Sussex on 10 October 2012

a PPh3 were lost. Simultaneous addition of both hydrogens to an olefin would

decrease the NVE by 4, and is not allowed since dNVE can only be 0 or f 2 .

C. Olefin Lsomerization by Ni[P(OEt),l4 and H2S04.-An example of a catalytic

system involving a Lewis acid is the isomerization of butenes catalysed by
Ni[P(OEt),], and HzS04,63 shown in Figure 3. In this case, two steps are required
before the metal enters the loop: (1) Lewis acid association with 18-electron
NiL, and (2) Lewis base ligand dissociation to give 16-electron HNiL3+. The
reader can readily follow through the steps in the loop. Figure 3 is over-
simplified in that side-reactions of catalyst decomposition and P(OEt), de-
alkylation which occur in the system are not shown.

D. Orthodeuteriation of Triary1phosphines.-A catalytic reaction which involves

oxidative addition of a transition metal by a C-H bond is the ortho-deuteriation
of triaryl phosphine. Triphenylphosphine is converted catalytically under Dz
into tri-(2,6-dideuteriophenyl)phosphine by H R U C I ( P P ~ ~ ) , , C ~ H ~The
CH~.~~
complex HRuC1[P(OPh)J4 also exchanges H for D on the ortho-positions of
the ligand phenyl rings, but is a very poor catalyst for the catalytic formation of
tri-(2,6-dideuteriophenyl) phosphite; addition of P(OPh), to a solution of
HRuCl[P(OPh),], under Dzprevents deuteriation even of the original phosphite
ligands. The result can be readily understood in terms of the 16 and 18 Electron
Rule. HRuCl(PPh,), is a 16-electron complex and oxidative addition by an
ortho-C-H bond can readily proceed. The 18-electron complex HRuCl-
[P(OPh),], must dissociate a ligand, in an equilibrium suppressed by added
ligand, before oxidative addition can occur.

48 D. R. Eaton and S. R. Suart, J. Amer. Chem. SOC.,1968,90,4170.

J. P. Jesson, P. Meakin, and C. A. Tolman, J. Amer. Chem. SOC.,1972, 94, 3240.
5 0 A. Sacco, R. Ugo, and A. Moles, J. Chem. SOC. (A), 1966; 1670.
51 A. S. Hussey and Y. Takeuchi, J. Amer. Chem. SOC.,1969, 91, 672.
J. P. Candlin and A. R. Oldham, Discuss. Faraday SOC.,1968, 46, 60.
53 C. A. Tolman, J . Amer. Chem. SOC.,1972, 94, 2994.
54 G. W. Parshall, W. H. Knoth, and R. A. Schunn, J . Amer. Chem. SOC.,1969, 91,4990.

348
 Tolman
View Online
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337
Downloaded by University of Sussex on 10 October 2012

Y Y
CbCH2CH-NiH L32' d6s CH3CHzCHzCH2Ni HL32+
I 4
CH3 $
I;, // I ./
'<\xLCH3CH2CH2CH3~/ /,
'\@- L Ni2++3' A-0
Figure 3 Isomerization of butenes by Ni[P(OEt),], and H,SO,. I-B, c-2-B, and t-2-B represent
but-1 -ene, cis-but-2-ene, and trans-but-2-ene. From Ref. 53

5 Apparent Exceptions to the Rule

A. Possible Alternative Explanations.-Some compounds appear at first sight to

react in an exceptional way. The 1%electron complexes CO(NO)(CO),,~~ and
) ~ ~ ~ S N reactions
( T - C , H ~ ) R ~ ( C Oundergo ~ with phosphines to give Co(N0)-
(CO)tL, and (?r-C,H,)Rh(CO)L. Their anomalous behaviour can be explained
in terms of an electron rearrangement to make available an empty bonding
orbital.,' This rearrangement can be thought of as an intramolecular conversion
55 E. M. Thorsteinson and F. Basolo, J. Amer. Chem. SOC.,1966, 88, 3929.
56 H. G . Schuster-Woldan and F. Basolo, J. Amer. Chem. SOC.,1966, 88, 1657.
57 Ref. 10, p. 571.

349
 16 and 18 Electron Rule in Organometallic Chemistry and Homogeneous Catalysis
View Online
of the 18-electron complex into a 16-electron complex. C01lman~~ has recently
reported i.r. evidence which suggests that several CoCl,(NO)L, complexes
actually exist in solution as an equilibrium mixture of 16-electron [NO- ligand]
and 18-electron [NO+ ligand] forms.
Angelici and Graham69have reported a two-term rate law for the reaction of
18-electron Mo(CO), with phosphines such as PBu, :
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

The sN2 path may involve a 20-electron intermediate or, as the authors mention,
could involve attack by the phosphine on co-ordinated CO.
An X-ray crystal structure determination on PtI,(diars), (diars = o-phenyl-
enebisdimethylarsine) shows that the PtII is six-co-ordinate in the crystal, with
Downloaded by University of Sussex on 10 October 2012

equal bond distances from Pt to each of the I atoms.60The bond lengths are,
however, exceptionally long, 3.50 A. Though the compound appears to be a
20-electron complex, there is an alternative bonding description, first proposed
by R u d e s 1 for the analogous 13-, which preserves the noble-gas configuration
for the central atom. In nitromethane the platinum complex dissociates to give
the 18-electron [PtI(diar~),]+.~~
X-Ray crystal structures of tetrabenzylzirconiums2and tetraben~yltitanium,~~
apparently %electron complexes, have been reported. Both show anomalously
small M-C-4 bond angles at the methylene carbons, suggesting participation
of ring electrons in the bonding. Definitive evidence for participation of ring
electrons in bonding in a benzyl complex has been found in the structure of
(h3-4-MeC6H4CH,)(h6-c6H~)Mo(~o)3.6*

B. Original Formulation shown to be Incorrect.-There are a number of cases in

the organometallic literature where complexes formulated as having other than
16 or 18 metal valence electrons were proposed, but where subsequent work has
shown that the original formulation was in error. The absence of dissociation
of Wilkinson's hydrogenation catalyst to RhCI(PPh,), has already been men-
tioned. The apparent 14-electron complexes Ni(PPh3)266 and Ni[P(Oo-t~lyl)~],~~
were later shown to be the 16-electroncomplexes (C,H4)Ni(PPh,),67and (C2H4)
Ni[P(Oo-tolyl)&. 27
A 20-electron complex Ni(CN),'- was proposed on the basis of equilibrium
studies on Na,Ni(CN), containing various concentrations of added NaCN.68
s* J. P. Collman, P. Farnham, and G . Dolcetti, J . Amer. Chem. SOC.,1971,93, 1788.
59 R. J. Angelici and J. K. Graham, J . Amer. Chem. SOC.,1966, 88, 3568.
O 0 N. C. Stephenson, J. Inorg. Nuclear Chem., 1962, 24, 791.
61 R. E. Rundle, Rec. Chem. Prugr., 1962, 23, 195.

6a G. R. Davies, J. A. J. Jarvis, B. T. Kilbourn, and A. J. P. Pioli, Chem. Comm., 1971, 677.

63 I. W. Bassi, G. Allegra, R. Scordamaglia, and G. Chioccola, J . Amer. Chem. Soc., 1971,93,
3787.
64 F. A. Cotton and M. D. La Prade, J. Amer. Chem. Suc., 1968, 90, 5418.
6s G. Wilke, E. W. Muller, and M. Kroner, Angew. Chem., 1961, 73, 33.
M. A. McCall and H. W. Coover, B.P. 1 146 074/1969.
c7 G. Wilke and G. Herrmann, Angew. Chem. Internat. Edn., 1962,1, 549.
6 * R. A. Penneman, R. Bain, G . Gilbert, L. H. Jones, R. S. Nyholm, and G . K. N. Reddy,
J. Chem. SOC.,1963, 2266.

350
 Tolmari
View Online
Later studies showed that the results were consistent with the presence of
Ni(CN),,- and Ni(CN),,-, 16- and 18-electron complexes, respectively, and
that there was no evidence for the formation of 20-electron Ni(CN),4- even in
4M-NaCN.69 An X-ray crystal-structure determination has shown that the
compound originally formulated as M O ( C H , S ~ M ~ ,a) 10-electron
~,~~ complex, is
actually a cluster compound with a very short Mo-Mo bond di~tance.~‘
The lklectron complex Pt(PPh,), has been proposed as an intermediate in
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

oxidative addition reactions2 of Pt(PPh3), and (C,H4)Pt(PPh,),,72 and in

substitution reactions7, of (alkyne)Pt(PPh,),. The compound was first suggested
on the basis of low molecular weights for Pt(PPh& and Pt(PPh3)4.74The prep-
aration of a yellow material said to be the ‘probably monomeric species
Pt(PPh3),’ has recently been
Downloaded by University of Sussex on 10 October 2012

Recent careful physical studies have been made on the platinum(0)-triaryl

phosphine-ethylene system which include a combination of molecular weight
determinations, spectrophotometry, and lH and slP Thesc studies have
shown that Pt(PPh,), and (C,H,)Pt(PPh,), do not dissociate to a detectable
extent. Furthermore the behaviour of lineshapes in the n.m.r. spectra on adding
C2H4 and/or PPh, to solutions of (C,H,)Pt(PPh,), or Pt(PPh,), requires the
associative mechanisms for exchange of ethylene or phosphine expected for
these 16-electron complexes.

C. Predictions Based on the Rule.-The compound Pd[MeC(CH,PPh,) ,I2 has

been prepared in two crystalline forms.77The a-isomer was said to be a six-
co-ordinate 22-electron complex (l), on the basis of a zero dipole moment,
whereas the ,&isomer with a dipole moment of 2.25 D was said to be (2). On the
basis of the 16 and 18 Electron Rule it would be predicted that the a-isomer is
(3). The true co-ordination number of the complexes could readily be estab-
lished by 31Pn.m.r.

68 J. S. Coleman, H. Petersen, jun., and R. A. Penneman, Inorg. Chem., 1965,4, 135.

70 G. Yagupsky, W. Mowat, A. Shortland, and G . Wilkinson, Chem. Comm., 1970, 1369.
71 F. Huq, W. Mowat, A. Shortland, A. C. Skapski, and G. Wilkinson, Chem. Comm., 1971,
1079.
Dissociation constants for dissociation of Pt(PPh,), and (C2H4)Pt(PPh,), were reported
by J. P. Birk, J. Hafpern, and A. L. Pickard, Inorg. Chem., 1968, 7,2672.
7 s A. D. Allen and C. D. Cook, Canad. J . Chem., 1964,42, 1063.
74 L. Malatesta and C. Carielld, J . Chem. Soc., 1958, 2323.
R. Ugo, G . La Monica, F. Cariati, S. Cenini, and F. Conti, Inorg. Chim. Acta, 1970,4, 390.
78 C. A. Tolman, W. C. Seidel, and D. H. Gerlach, J. Amer. Chem. SOC.,1972, 94, 2669.
77 J. Chatt, F. A. Hart, and H. R. Watson, J. Chem. SOC.,1962, 2537.

351
 I6 and 18 Electron Rule in Organometallic Chemistry and Homogeneous Catalysis
View Online
Bi~(trityl)nickel,~~an apparent 12-electron complex, probably has a bis(r-
allyl) structure in which each trityl group donates three
Nickelocene is a paramagnetic complex in which the metal appears to have
20 valence electrons. Molecular orbital calculations, assuming a symmetric
ferrocene-like structure (h), show that the last two electrons should occupy
a strongly anti-bonding doubly degenerate elg orbital.80This is consistent both
with the instability of nickelocene compared with ferrocene and with the observed
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

paramagnetism of the nickel complex. However, it can be predicted that the

stable configuration of the compound is (5h-C5H5)(3h-C5H5)Ni, in which one
of the rings is displaced off-centre. Easy accessibility of a number of equivalent
configurations with a small activation energy between them should give a rapid
fluxional behaviour, consistent with the results of recent electron-diffraction
Downloaded by University of Sussex on 10 October 2012

measurements on nickelocene vapour.*l

Vaskag2has reported H-D exchange of HIrCl,(PPh& and HRuCl(C0)-
(PPh3)3with D2 and suggested that the reactions go via eight-co-ordinate inter-
mediates (20-electron complexes) formed in the direct oxidative addition of D,
to the above complexes. Vaska has even attributed a special catalytic activity
to this type of high co-ordination number (and high NVE) complex.83A pre-
diction that added CO and PPh3 will inhibit the rate of the exchange reaction,
indicating that a Lewis base ligand is lost prior to oxidative addition by Dz,
is nicely supported by the recent work of Burnett and Morrisons4which shows
that added PPh, inhibits the rate of reaction of H2 with HIr(CO)(PPh,), to give
H ,Ir(CO)(PPh ,) 2.
In homogeneous catalytic reactions it is to be expected that no reaction steps
will be found in which the NVE of the metal in the complex changes by more
than two. Thus concerted addition of two hydrogens to an olefinic double bond,
+
coupling of two olefins to form a cyclobutane, and [2 21 valence isomerism of
polycyclic hydrocarbons to cyclic dienes are excluded on the bases that ANVE =
+ 4 in these reactions. The requirement that ANVE = 0, + 2 adds severe re-
straints to those proposed by Mango and Schachtschneider based on orbital
symmetry considerations.85There is considerable experimental evidence that
catalysis of these reactions by Group VIII metal complexes is stepwise. For
example, evidence has been presented to show that valence isomerization of
cubane by [Rh(diene)Cl], catalysts proceeds via a non-concerted oxidative
addition mechanism.se
The products formed by cyclodimerization of penta-l,3-diene are consistent

78 G . Wilke and H. Schott, Angew. Chem. Internat. Edn., 1966, 5, 583.

This suggestion has also been made by R. B. King in ‘Annual Surveys of Organometallic
Chemistry’, Vol. 3, ed. D. Seyferth and R. B. King, Elsevier, Amsterdam, 1967, p. 403.
J. H. Schachtschneider, R. Prins, and P. Ros, Inorg. Chim. A d a , 1967, 1, 462.
L. Hedberg and K. Hedberg, J . Chem. Phys., 1970,53, 1228.
8a L. Vaska, Proc. 8th Internat. Con6 Co-ord. Chem. Vienna, 1964, 99.
8 3 L. Vaska, Znorg. Nuclear Chem. Letters, 1965, 1, 89.
M. G . Burnett and R. J. Morrison, J . Chem. SOC.(A), 1971, 2325.
F. D. Mango and J. Schachtschneider, J. Amer. Chem. SOC.,1971, 93, 1123.
86 L. Cassar, P. E. Eaton, and J. Halpern, J. Amer. Chem. Soc., 1970, 92, 3515.

352
 Tolman
View Online
with a stepwise rather than a concerted rea~tion.~’
Heimbachaahas also stressed
the stepwise nature of reactions of this type, whose concerted reactions are
forbidden in the absence of catalysts by the Woodward-Hoffmann rules.8n

6 Conclusion
The 16 and 18 Electron Rule in organometallicchemistry is consistent with such
Published on 01 January 1972 on http://pubs.rsc.org | doi:10.1039/CS9720100337

a large body of experimental evidence, including studies on reaction mech-

anisms, that anyone proposing an exceptional compound or reaction path must
bear the burden of proof. The Rule is, however, empirical at this point, and may
be subject to exceptions in some cases, especially near the beginning and end of
the transition s e r i e ~ NyholmgO
. ~ ~ ~ ~ ~has discussed the relationship between
transition metal, oxidation state, co-ordination number, and ligand type and
Downloaded by University of Sussex on 10 October 2012

has nicely correlated a great deal of data with atomic properties such as ionization
potentials and electron promotion energies. More quantitative theoretical studies
and further experience are now required to define the limits of applicability of
the Rule.
The basic premise of this review is that 16- and 18-electronconfigurations are
readily accessible to diamagnetic organometallic transition-metal complexes.
Species with other configurations or reactions by other paths will generally be so
energetically unfavourable by comparison that they are negligible.
If this review acts as a goad to the research of others, if only to challenge its
generality, it will have served its purpose well,

I am indebted to R. Cramer, J. P. Jesson, G. W. Parshall, L. J. Guggenberger,

and R. S. Nyholm for helpful discussions, and to G. Wilkinson and P. Heimbach
for communication of results prior to publication.

Note Added in Proofi The application of the 16 and 18 Electron Rule to catalysis
by transition-metal hydride complexes has recently been discussed by C . A.
Tolman, ‘Transition Metal Hydrides’, ed. E. L. Muetterties, Marcel Dekker,
Inc., New York, 1971, Ch. 6.

P. Heimbach and H. Hey, Angew. Chem. Internat. Edn., 1970, 9, 528.

P. Heimbach, Aspects of Homogeneous Catalysis, in press.
R. B. Woodward and R. Hoffmann, Angew. Chem. Internat. Edn., 1969, 8, 781.
R. S. Nyholm, Proc. Chem. Soc., 1961,273; ‘Proc. 3rd Internat. Conf. on Catalysis’, North
Holland, Amsterdam, 1965, p. 25.

353