ChemTexts (2015) 1:10

DOI 10.1007/s40828-015-0010-4

LECTURE TEXT

The 18-electron rule and electron counting in transition metal

compounds: theory and application
Seth C. Rasmussen

Received: 29 January 2015 / Accepted: 18 February 2015 / Published online: 5 March 2015
Ó Springer International Publishing 2015

Abstract The 18-electron rule and the corresponding compounds, commonly referred to as ‘electron counting’,
methods for counting the total valence electrons of transi- are among the most useful basic tools in modern inorganic
tion metal complexes are among the most useful basic tools chemistry, particularly in its application to organometallic
in modern inorganic chemistry, particularly in its applica- species. Not only can it be a predictive tool for evaluating
tion to organometallic species. While in its simplest rep- the stability of some inorganic complexes, but can also be
resentation, the 18-electron rule is explained in that a used to predict potential formulas of stable compounds. In
closed, stable noble gas configuration of ns2(n-1)d10np6 is addition, changes in the electron count of metal complexes
achieved with 18 valence electrons, this does not during chemical processes can be a critical factor in de-
adequately explain the trends and exceptions seen in termining mechanistic details of reaction pathways.
practice. As such, this report presents a deeper discussion Historian William B. Jensen has traced the origins of the
of the 18-electron rule via molecular orbital models, modern 18-electron rule to 1921 [1], when Irving Langmuir
stressing the roles of both r- and p-bonding effects. This (1881–1957) [2] of the General Electric Company (Fig. 1)
discussion thus aims to provide a better understanding of presented new models of valence, in an attempt to extend
the relationship between electron count and stability, while the Lewis static-atom model to cases not adequately ex-
also illustrating which factors can determine adherence (or plained by ‘‘octet theory’’ (i.e., the modern ‘‘octet rule’’)
not) to this commonly utilized rule. Lastly, the two com- [3]. In this early work, Langmuir derived an equation
mon methods for electron counting (ionic and covalent which related the number of shared (i.e., bonding) electrons
models) are also presented with practical examples to or the covalence (vc) of a given atom in a compound to the
provide the complete ability to apply the 18-electron rule. difference between the number of valence electrons (e) in
the isolated neutral atom and the number of valence elec-
Keywords Transition metal complexes  18-Electron trons (s) after formation of the compound:
rule  Electron counting  Effective atomic number (EAN) vc ¼ s  e ð1Þ
rule
For any complete compound, s = 2, 8, 18, or 32. As such,
the value of s corresponds to complete closed shell electron
Introduction configurations and would correspond to the normal 8 for
compounds of p-block elements as illustrated by the octet
The 18-electron rule and the corresponding methods for rule. Extending this model beyond the octet, however, leads
determining the total number of valence electrons in metal to s = 18 for compounds of d-block elements and s = 32 for
compounds f-block elements. Thus, for the d-block elements,
a simple algebraic rearrangement leads to the relationship
S. C. Rasmussen (&) e þ vc ¼ 18 ð2Þ
Department of Chemistry and Biochemistry, North Dakota State
University, NDSU Dept. 2735, P.O. Box 6050, which states that the sum of the metal valence electrons
Fargo, ND 58108-6050, USA
e-mail: seth.rasmussen@ndsu.edu
and the bonding electrons contributed by the ligands equals

123
10 Page 2 of 9 ChemTexts (2015) 1:10

model, which they referred to as the ‘‘rare-gas rule’’ [8]. In

this treatment, they began with the full electron count
model, but quickly refocused the discussion to present the
work relative to the 18 valence electrons of the d-block
elements.
Overall, the common aspects of the Langmuir and
Sidgwick models relating to the d-block can be combined
to make the general statement that when the metal of a
complex obtains an outershell configuration of
ns2(n-1)d10np6, the valence orbitals contain 18 electrons
and a closed, stable configuration is achieved. This basic
relationship is commonly referred to as the ‘‘18-electron
rule’’, which has become a guiding principle of inorganic
chemistry, particularly in organometallic chemistry [8–10].
It should be pointed out that some use the terms
Fig. 1 Irving Langmuir (1881–1957) [Edgar Fahs Smith Collection,
‘‘18-electron rule’’ and ‘‘effective atomic number (or EAN)
Kislak Center for Special Collections, Rare Books and Manuscripts, rule’’ interchangeably [9, 11–13]. However, as the discus-
University of Pennsylvania] sion above correctly demonstrates, the 18-electron rule is
consistent with the EAN rule and may thus be considered a
component or outcome of Sidgwick’s model, but these
a total of 18 valence electrons. Langmuir then illustrated terms are not synonymous. In addition to the different
how the metal carbonyl compounds Ni(CO)4, Fe(CO)5, and emphasis on valence versus total electrons, the 18-electron
Mo(CO)6 all exhibited structures consistent with those rule is strictly limited to the d-block elements, while the
predicted by this relationship [3, 4]. EAN rule can be applied to all elements of the periodic
An alternative electron count model, known as the ef- table.
fective atomic number (EAN) rule, was introduced in 1927 Much like the more common octet rule, the 18-electron
by Nevil Sidgwick (1873–1952) [5] of Oxford University [1, rule is not always strictly obeyed and is subject to a number
4, 6, 7]. Like the Langmuir model, stability was assumed to of apparent exceptions [10]. Thus, while this tool is ex-
be dependent on obtaining a noble gas configuration for the tremely useful in predicting stability, examples of stable
central atom. In contrast, however, the EAN of a compound metal complexes with more or less than 18 valence elec-
was analogous to the atomic number of an atom and thus trons are also fairly common [1, 6, 9]. As such, the goal of
focused not just on the number of valence-shell electrons, but this report is to present a deeper discussion of the
the total electron count of the compound’s central atom. 18-electron rule to better understand the relationship be-
Thus, for transition metals, the EAN rule was said to be tween electron count and stability, while also illustrating
obeyed when d-block compounds obtained EAN values which factors can determine adherence (or not) to this
equal to the total electron configurations of noble gases Kr, commonly utilized rule.
Xe, and Rn (i.e. 36, 54, or 86) [7]. For the transition metals,
attainment of an 18-electron valence configuration would
therefore be equivalent to attaining the total electron count or 18-Electron rule in terms of molecular orbital models,
EAN of the nearest noble gas [1]. part I: simple view
As detailed by Jensen [1], there was a reversion to the
earlier approach of Langmuir by the late 1960s. This was Further insight into the connection between the stability of
most likely due to the fact that Sidgwick’s model includes metal compounds and the 18-electron rule can be gained by
both the valence and core electrons, thus resulting in a addressing this relationship in terms of the molecular or-
different electron count in relation to stability for each row bital (MO) description of bonding in metal complexes. As a
of the d-block. In contrast, Langmuir’s procedure, like the starting point, let us consider just the simple r bonding
octet rule, has the advantage of using a single valence between a metal and six identical r-donor ligands to
electron count for the three rows of the transition metals, generate an octahedral metal complex. Linear combination
thus eliminating the need to remember a different EAN of the ligand orbitals with the s, p, and d atomic orbitals of
value for each of the corresponding noble gases [1, 6]. the metal results in the formation of six bonding MOs (a1g,
Another factor contributing to this shift in approach could t1u, and eg) and six corresponding antibonding MOs, as
also have been a paper by Craig and Doggett published in illustrated in Fig. 2. As there are a total of nine metal-based
1963 that tried to present a theoretical basis for Sidgwick’s atomic orbitals and only six ligand-based donor orbitals,

123
ChemTexts (2015) 1:10 Page 3 of 9 10

The value of DO depends on both the central metal and

the specific ligands involved. In terms of the metal, DO
tends to increase down any particular periodic group (i.e.
3d \ 4d \ 5d) and for the same metal, it tends to increase
with the metal charge (i.e. M2? \ M3? \ M4?) [9, 14, 15].
As such, complexes of second and third row transition
metals are typically not found to have more than 18 elec-
trons [4].
The effect of the associated ligands is represented by the
spectrochemical series (Fig. 4) [14–17], which orders the
ligands from the smallest value of DO to the largest. In the
most general sense, the order within the spectrochemical
series is indicative of the ligand r-donor strength. How-
ever, ligands low on the spectrochemical series also tend to
exhibit p-donor characteristics, while those high on the
Fig. 2 Nevil Sidgwick (1873–1952) [Reproduced from [5]; Courtesy
series are typically strong p-accepting ligands. While these
of JSTOR] r and p effects are typically additive, the p effects can
sometimes play a more critical role than the simple r-
donor strength. The ligands within the series are generally
this results in three d orbitals remaining as non-bonding t2g
ordered according to the periodic group of the ligand donor
MOs.
atom. As such, group 17 (i.e., the halogens) are low on the
For such an octahedral complex, the most stable ar-
series, while group 14 are high on the spectrochemical
rangement will be that in which all of the bonding MOs (a1g,
series. Groups 15 and 16 then fall between these two ex-
t1u, and eg) are fully occupied and the anti-bonding MOs are
tremes, as can be seen in Fig. 4 [14].
empty. Occupation of the six bonding MOs, as well as the
Although the factors discussed above are a good start to
three nonbonding MOs, would thus require 18 electrons, as
understanding the working parameters of the 18-electron
predicted by the 18-electron rule [4, 11]. As such, com-
rule, it does not adequately explain the various limitations
plexes will therefore tend to adhere to the rule if the ener-
of this common rule. While not a complete list, the most
getic separation between the non-bonding t2g and anti-
common limitations can be given as:
bonding eg MOs (commonly referred to as DO, Fig. 3) is
large [6, 12]. Here, as DO increases, the eg MOs are desta- 1. This rule works best for low-valent (oxidation
bilized and ultimately reside at higher energies, thus making state B 2) metal complexes, but many high valent
occupation of these anti-bonding orbitals unfavorable. metal complexes also work.

Fig. 3 MO diagram for an octahedral metal complex (r bonding only; for simplicity, degenerate orbitals are displayed as stacked sets)

123
10 Page 4 of 9 ChemTexts (2015) 1:10

Fig. 4 Spectrochemical series of ligands

2. This rule works best for complexes containing p- nonbonding in terms of the previous r interactions. In the
acceptor ligands. case of p-donor contributions, the ligand p-orbitals are
3. Square planar complexes are generally stable as filled and lower in energy than the metal t2g. Mixing of
16-electron species. these orbitals with the metal t2g orbitals thus results in
4. There are always exceptions to the rule. production of p-bonding metal–ligand MOs, while the t2g
orbitals take on p* anti-bonding character, resulting in their
To understand the basis of these common limitations, it
destabilization and a decrease in DO (Fig. 5) [12, 16, 18].
will be necessary to consider a number of additional factors.
Things are a bit different in the case of p-acceptor
contributions. In this case, the ligand orbitals interacting
with the metal are empty p*-orbitals which are high in
18-Electron rule in terms of molecular orbital models,
energy (Fig. 5). As such, mixing of these high-lying or-
part II: more detailed view
bitals with the metal t2g orbitals results in stabilization of
the metal orbitals, that leads to an increase in the value of
In the initial treatment above, only r donor interactions
DO. Here, the stabilized t2g orbitals are now formally p-
were considered and any potential p interactions were not
bonding in nature, while the eg orbitals retain their r* anti-
taken into account. To deepen our insight into the factors
bonding character [12, 16, 18, 19].
that contribute to the 18-electron rule, it is now required to
Transition metal complexes can be divided into three
complicate our bonding picture by including any potential
groups [9, 18], which will be referred to here as Class I,
p-donor or p-acceptor contributions from the organic li-
Class II, and Class III. The first of these groups, Class I, are
gands. As outlined in Fig. 5, the initial splitting of the
complexes with a relatively small DO and generally consist
d orbitals from the r donors (i.e., the boxed orbitals in
of first row transition metals with ligands low on the
Figs. 3, 5) is further modified via interactions between p
spectrochemical series (i.e., weak r donors or p donors). If
ligands and the metal t2g orbitals, which are formally

Fig. 5 Effect of p ligands on the MO diagram for octahedral metal complexes

123
ChemTexts (2015) 1:10 Page 5 of 9 10

the bonding in these complexes is only weak r bonding, d-block elements belong to Class III, the 18-electron rule
then the metal t2g orbitals are formally nonbonding and can becomes a very powerful and reliable predictive tool for
thus be either empty or filled without negatively affecting organometallic chemistry.
the extent of metal–ligand bonding. Here, the relatively
weak r bonding also results in the eg orbitals being low in
energy and only weakly anti-bonding. As a result, the oc- Other compound geometries
cupation of these MOs can also occur with little negative
effect. Ligands with p-donor character, however, con- Although the argument used above was developed using
tribute to p bonding which results in the t2g orbitals taking the example of an octahedral complex, precisely similar
on anti-bonding character as shown in Fig. 5, and the oc- results are obtained for other coordination numbers. Gen-
cupation of these MOs is then not favored. It should be erally it can be said that if the coordination number (CN) is
pointed out that strong p-donor ligands not only result in greater than four, then [CN-4] d orbitals are required in
the t2g becoming anti-bonding, but also generate comple- addition to the metal s and p orbitals to form MOs with the
mentary filled p-donor bonding orbitals localized largely corresponding ligand r orbitals. As a result, this will result
on the ligands. In this case, occupation of these p-bonding in [CN-4] low-lying anti-bonding orbitals and [9-CN]
orbitals while keeping the t2g orbitals unoccupied maintains nonbonding orbitals. The energetic separation between
an 18 electron count [4]. For the most part, however, these these two general sets of orbitals will largely depend on the
complexes typically do not exhibit p-bonding of this strength of the r donors as discussed above [9].
magnitude and r-bonding dominates the characteristics of This general relationship is illustrated for CN = 6 for
these metal compounds. As such, the 18-electron rule has the octahedral geometry given in Fig. 3, but can also be
no special significance for the bulk of this class of com- shown for CN = 5 as given for the trigonal bipyramidal
plexes [12] and stable Class I species are possible with geometry shown in Fig. 6. As with the octahedral example,
electron counts of 12–22 electrons depending on the extent filling the bonding and nonbonding MOs here would cor-
that the t2g and eg orbitals are occupied [6, 9, 18]. respond to 18 electrons and be consistent with the
Class II complexes have a larger DO and generally 18-electron rule. Effects on adherence to this rule based on
contain second and third row transition metals (especially r-donor strength and any contribution of p-bonding would
in higher oxidation states) with r ligands intermediate to again mirror those previously discussed for the octahedral
high on the spectrochemical series. For this class, there is case [4, 19].
intermediate to strong r bonding, but no p bonding. As The situation is more complicated for tetrahedral com-
such, the t2g orbitals are still nonbonding and can thus be plexes (CN = 4) [6, 9, 19]. In this case, bonding with the
either empty or filled. The relatively strong r bonding, four r-donors can be proposed using either d3s or sp3 sets
however, raises the energy of the eg orbitals and strength- of metal orbitals, although in practice a combination of
ens their anti-bonding nature, therefore making the occu- both sets are used and the t2 orbitals of the complex will
pation of these MOs disfavored. As a result, stable Class II involve both d and p orbitals from the metal (Fig. 7) [9]. As
complexes are possible with electron counts of 12–18 the resulting four bonding MOs can only accommodate a
electrons, depending on the occupation of the t2g orbitals
[6, 9, 18]. Class II complexes can follow the 18-electron
rule, but a fair number do not.
The final group, Class III, is made up of complexes
exhibiting the largest DO. These complexes generally
consist of low oxidation state metals with ligands high on
the spectrochemical series, particularly p-acceptor ligands.
The metal–ligand interactions here constitute intermediate
to strong r bonding, as well as significant p back-bonding
into the empty p* orbitals of any p-acceptor ligands. As
such, the t2g orbitals now have p-bonding character and
need to be filled for complex stability. In addition, the eg
orbitals are still strongly anti-bonding and thus, the occu-
pation of these MOs is again not favored. As a result, stable
Class III complexes will always have the t2g orbitals oc-
cupied and the eg orbitals empty, resulting in 18 valence
electrons and strict adherence to the 18-electron rule [9, Fig. 6 MO diagram for a trigonal bipyrimidal metal complex (r
18]. As almost all organometallic compounds of the bonding only)

123
10 Page 6 of 9 ChemTexts (2015) 1:10

Fig. 7 MO diagram for a tetrahedral metal complex (r bonding only)

Fig. 8 MO diagram for a square planar metal complex (r bonding
only)

total of eight electrons, the remaining ten electrons must Electron counting methods
occupy the e and t2 sets of d orbitals, which are formally
nonbonding and at least partially antibonding [6]. The re- By counting the number of valence electrons surrounding a
duced number of ligands in comparison to the previous metal in a particular complex formula, it is possible not
octahedral case results in a smaller energetic separation only to predict whether the complex should be stable, but
between these two sets of orbitals (DT) and thus, there is in some cases give details concerning structural aspects of
little barrier to population of the t2 orbitals [9]. the complex (i.e., ligand binding modes, the presence of
Because of these factors, the 18-electron rule is only metal–metal bonds, etc.) [6]. Of course, as previously
obeyed in tetrahedral complexes of p-acceptor ligands. In noted, while the electron count can be extremely useful in
these cases, both the e and t2 sets of d orbitals donate predicting stability, examples of stable metal complexes
electrons to the empty p* orbitals of the ligands, resulting with more or less than 18 valence electrons are also fairly
in both sets of orbitals becoming bonding in nature. In common [1, 6, 9]. Nevertheless, the ability to correctly
addition, MO calculations have shown that these p-bonding determine the number of valence electrons from the com-
interactions contribute more to the metal–ligand bonding plex formula is a critical step in the application of the
than the corresponding r-donation from the ligand [19, 18-electron rule.
20]. There are two methods for electron counting, the ionic
As stated in the general limitations above, the one model and the covalent model, each with its own advan-
major geometry that does not obey the 18-electron rule is tages and limitations. Both models include the total valence
that of four-coordinate, square planar complexes [4, 6, 9, electrons of the metal and those donated by the ligands, but
19]. As shown in Fig. 8, the metal-based orbitals are split differ in the way the division of electrons between the metal
into four sets, with the majority being nonbonding or and the ligands are viewed. Either method can be used
having some partial bonding character. The remaining successfully and should provide the same answer, providing
orbital (corresponding to the dx2 y2 ) is anti-bonding and is that care is taken not to mix aspects of the two models.
fairly high in energy [19]. For these complexes, filling the In the ionic model, the metal is treated as a cationic
four low-lying r bonding MOs requires 8 electrons and center and the ligands carry the charge associated with their
population of the metal-based bonding and nonbonding non-coordinated state. As a consequence, this model re-
orbitals requires another 8 electrons. Any additional quires that we correctly assign the metal a formal oxidation
electrons would then populate the dx2 y2 (b1g) r* orbital, state. This in turn requires that we also know the formal
resulting in a decrease in stability. As such, a total of 16 charge on all the corresponding ligands. However, this
electrons represent the maximum population for a stable model has the advantage that it treats most metal–ligand
square planar complex [9, 19]. This is sometimes referred bonds as coordinate covalent (or dative) bonds which
to as the ‘‘16-electron rule’’ [6] or added to the 18-elec- means each bond donates two electrons to the total electron
tron rule to become the ‘‘16 and 18 electron rule’’ [10], count. The formal charge and corresponding electrons
but most often is just recognized as the most consistent donating by a significant number of common ligands under
exception to the 18-electron rule. the ionic model are given in Table 1.

123
ChemTexts (2015) 1:10 Page 7 of 9 10

In contrast to the ionic model, the covalent model treats 2,20 -bipyridine (bpy) and chloride ligands. Under the ionic
the metal as a neutral center and all ligands are likewise model, the combination of the overall 1? charge and the
treated as neutral species. While this does not affect the two anionic ligands allows us to assign for formal 3?
way that neutral ligands are counted, ligands that would be oxidation state to the rhodium center. As such, this d6 metal
typically viewed as charged are now treated as radical contributes six electrons, while each bpy and chloride
species for the electron count. This model removes the ligand contributes four and two electrons, respectively, the
need to determine the formal oxidation state of the metal, sum of which gives the expected 18 electrons of a stable
but is somewhat awkward for anionic ligands such as the species.
halides. However, this model is much more realistic for Under the covalent model, the rhodium of the same
many organometallic ligands, such as the simple alkyls. complex is treated as Rh(0) and is thus a d9 metal, con-
The corresponding electrons donated under the covalent tributing nine electrons to the electron count. As bpy is a
model are given for all common ligands in Table 1. neutral ligand, it is treated the same under both models and
The application of both models is illustrated via two ex- is thus a four-electron donor as before. Under the covalent
amples given in Fig. 9. The first example, [RhCl2(bpy)2]?, model, however, the anionic chloride ligand is treated as a
represents a typical coordination complex including neutral radical and is thus a one-electron donor. The sum

Table 1 Electron count Ligand Ionic model Covalent Model

contributions for common
-
ligand types Charge Donated e Donated e-

H (terminal hydride) -1 2 1
l-H (bridging halide) -1 2/2a 1/2a
X (terminal F, Cl, Br, I, CN, etc.) -1 2 1
a
l-X (bridging halide) -1 2/2 1/2a
ER2 (E = O, S, Se, Te) 0 2 2
OR, SR (terminal alkoxide or thiolate) -1 2 1
l-OR, l-SR (bridging alkoxide or thiolate) -1 2/2a 1/2a
MR3, PR3 (terminal amine or phosphine) 0 2 2
NR2, PR2 (terminal amide or phosphide) -1 2 1
l-NR2, l-PR2 (bridging amide or phosphide) -1 2/2a 1/2a
NCR (nitrile) 0 2 2
CH3 or other alkyl -1 2 1
l-CH3 or other alkyl (bridging alkyl) -1 2/2a 1/2a
1
g -aryl, alkenyl, alkynyl -1 2 1
3
g -allyl -1 4 3
g4-diene 0 4 4
g5-cyclopentadienyl (g5-Cp) -1 6 5
g6-arene 0 6 6
g1-acyl -1 2 1
2
g -ketone 0 2 2
g2-alkene 0 2 2
g2-alkyne 0 2 2
l-g2-alkyne 0 2/2a 2/2a
CO (carbonyl) 0 2 2
l-CO (bridging carbonyl) 0 1/1a 1/1a
CNR (isocyanide, isonitrile) 0 2 2
NO (linear nitrosyl) ?1 2 3
NO (bent nitrosyl) -1 2 1
CR2 (terminal alkylidene or Schrock carbene) -2 4 2
Adapted from [11] CR (terminal akylidyne or Schrock carbyne) -3 6 3
a
Electron values donated to CR2 (terminal Fisher carbene) 0 2 2
first and second metal, CR (terminal Fisher carbyne) -1 4 3
respectively

123
10 Page 8 of 9 ChemTexts (2015) 1:10

more simply treated using the ionic model, this second

(III) example is more straight-forward under the covalent
model. As such, both models have their benefits providing
each are used independently.
Some additional structural factors that need to be con-
sidered are cases of bridging ligands and metal–metal
bonding, both of which are somewhat common in
organometallic species. A common class of bridging li-
gands are anionic species containing multiple lone pairs on
the donor atom (halogens, RS-, RO-, R2P-, etc.). In this
case, the ligand is able to utilize two different lone pairs to
simultaneously donate to two separate metal centers. For
such ligands, the first bond is treated exactly the same as a
terminal anionic ligand and is thus a two-electron donor
under the ionic model and a one-electron donor for the
(II)
covalent model. As the second metal–ligand bond is via a
full lone-pair under either model, this is treated the same as
a neutral ligand and is thus a consistent two-electron donor
to the second metal.
Two additional types of bridging ligands are bridging
hydrides and bridging carbonyls, both of which only have a
single lone-pair to donate. In the more simplistic case of
the carbonyl, the donated electrons are split evenly be-
tween the two metal centers and this ligand is thus a one-
electron donor to each metal under both models. The
Fig. 9 Examples of electron counting via the ionic and covalent treatment of bridging hydrides, however, is a bit more
models
complicated. Here the first bond is treated in the same way
as a terminal hydride and is thus either a one- or two-
thus comes to 19 electrons, but as all components have electron donor. The second bond, however, is typically
been treated as neutral species, it is necessary to subtract described as a donation of the first r metal–ligand bond to
one electron to account for the overall positive charge of the second metal. As such, it is treated as a two-electron
the cationic complex. Thus, the total electron count comes donor to the second metal. Of course, this means that we
to 18 electrons and this is in complete agreement with the are including the same two hydride-electrons in the elec-
count from the ionic model. tron count of both metals. Illustrations of electron counting
In the second example, [Ni(Et)2(CO)2], represents a examples that include all three types of bridging ligands are
typical organometallic complex including both alkyl given in Fig. 10.
(ethyl) and carbonyl ligands. Under the ionic model, the Finally, metal–metal bonds are treated as classically
alkyl ligands are treated as anionic ligands and thus the covalent interactions under both models and thus always
nickel center is assigned a formal 2? oxidation state to contribute one electron to the total electron count. In this
give an overall neutral complex. As such, this d8 metal case, the donated electron comes from the valence elec-
contributes eight electrons, while both the alkyl and car- trons of the second metal center, which means that the
bonyl ligands contribute two electrons each, the sum of electron is technically included in the electron count for
which gives the expected 16 electrons for a square planar both metals. Examples of electron counting that include
species. metal–metal bonds are given in Fig. 10.
Under the covalent model, the nickel center is given a
formal charge of zero and is thus a d10 metal, contributing
ten electrons to the electron count. The neutral carbonyl Conclusion
ligand is treated as before and is therefore still a two-
electron donor. Under the covalent model, however, the Adherence to the 18-electron count is strongest for metal
ethyl ligands are now neutral radicals and are thus one- complexes that include strong p-bonding. As such exam-
electron donors. As a result, the total electron count again ples most commonly occur with p-acceptor ligands, strong
comes to 16 electrons and is in complete agreement with p-bonding interactions require low-valent metals to pro-
the count from the ionic model. While the first example is vide sufficient electron density on the metal for back-

123
ChemTexts (2015) 1:10 Page 9 of 9 10

covalent), providing the chosen model is utilized

ll l exclusively.

(II)
References
ll l ll l
1. Jensen WB (2005) The origin of the 18-electron rule. J Chem
Educ 82:28
2. Rideal E (1959) Langmuir memorial lecture. Proc Chem Soc
1959:80
3. Langmuir I (1921) Types of Valence. Science 54:59–67
4. Mingos DMP (2004) Complementary spherical electron density
model and its implications for the 18 electron rule. J Organomet
Chem 689:4420–4436
5. Tizard HT (1954) Nevil Vincent Sidgwick 1873–1952. Obit Not
Fellows R Soc 9:236
6. Huheey JE, Keiter EA, Keiter RL (1993) Inorganic chemistry,
principles of structure and reactivity, 4th edn. HarperCollins,
New York, pp 624–630
7. Sidgwick NV (1927) The electronic theory of valency. Clarendon
Press, Oxford, pp 163–184
8. Craig DP, Doggett G (1963) Theoretical basis of the ‘‘rare-gas
rule’’. J Chem Soc 1963:4189–4198
9. Mitchell PR, Parish RV (1969) The eighteen-electron rule.
J Chem Educ 46:811–814
10. Tolman CA (1972) The 16 and 18 electron rule in organometallic
chemistry and homogeneous catalysis. Chem Soc Rev 1:337–353
(III) 11. Collman JP, Hegedus LS, Norton JR, Finke RG (1987) Principles
and applications of organotransition metal chemistry. University
Science Books, Mill Valley, pp 22–34
12. Powell P (1988) Principles of organometallic chemistry, 2nd edn.
Chapman and Hall, New York, pp 148–152
13. Mingos DMP (1998) Essential trends in inorganic chemistry.
Oxford University Press, Oxford, pp 297–309
14. Wulfsberg G (2000) Inorganic Chemistry. University Science
Books, Sausalito, pp 367–374
Fig. 10 Examples of bridging ligands in electron counting 15. Huheey JE, Keiter EA, Keiter RL (1993) Inorganic chemistry,
principles of structure and reactivity, 4th edn. HarperCollins,
New York, pp 404–405
donation to the empty ligand p* orbitals. Metal complexes 16. Lever ABP (1968) Inorganic electronic spectroscopy. Elsevier
with strong r-donors, but no p-bonding can follow the Publishing Co., Amsterdam, pp 194–204
18-electron rule, but such complexes can exhibit stability 17. Osborn JA, Gillard RD, Wilkinson G (1964) J Chem Soc
1964:3168–3173
with fewer than 18 valence electrons. While these rela-
18. Porterfield WW (1993) Inorganic chemistry, a unified approach,
tionships are not limited to octahedral geometries and hold 2nd edn. Academic Press Inc, San Diego, pp 604–609
true for many other common metal geometries, four-co- 19. Wulfsberg G (2000) Inorganic chemistry. University Science
ordinate square planar complexes do not follow the Books, Sausalito, pp 532–534
20. Bauschlicher CW, Bagus PS (1984) The metal-carbonyl bond in
18-electron rule and are found to be stable with only 16
Ni(CO)4 and Fe(CO)5: a clear-cut analysis. J Chem Phys
electrons. Lastly, proficient application of the 18-electron 81:5889–5898
rule requires the ability to correctly count the total valence
electrons for a given metal complex. This can be accom-
plished using one of two electron counting models (ionic or

123