Handout 2014b
Handout 2014b
2014
These 8 lectures build on material presented in Introduction to Molecular Orbitals (HT Year 1). They provide a basis for
analysing the shapes, properties, spectra and reactivity of a wide range of molecules and transition metal compounds.
The essentials of molecular orbital theory
1.
2.
3.
4.
6.
Symmetry and molecular orbital diagrams for the first row hydrides AH n
7.
8.
9.
10.
12.
14.
The bonding and shapes of H3+ and H3-: 3c-2e and 3c-4e bonds
16.
16.
17.
18.
21.
22.
23.
Bibliography
1. Jean, Volatron and Burdett; An Introduction to Molecular Orbitals (OUP)
2. Murrell, Kettle and Tedder: The Chemical Bond (Wiley)
3. McWeeny: Coulson's Valence (OUP)
4. DeKock and Gray: Chemical Structure and Bonding (Benjamin)
5. Burdett: Molecular Shapes (Wiley)
Preliminary remarks
Any theory of bonding must be consistent with quantum theory, in particular the Schrdinger wave
equation,
that relates the total energy of a (molecular) system to its wave properties .
+
For H2 , for example, the Hamiltonian takes the form
where the first term relates to the kinetic energy of the electron, the second two to the energy of
attraction between the electron and the two nuclei and the last to repulsion between the two
nuclei.
The orbital approximation
The wavefunction is in principle a highly complex multi-dimensional entity. The most common
simplification is to assume that the wave function for all the N electrons in a molecule can be
written as a product of N one-electron wavefunctions, .
(The
here refer to a spin orbital and | | indicates that the wave-function is anti-symmetrized
with respect to exchange of electrons. These are requirements of the Pauli Principle - no two
electrons in a system can have the same set of quantum numbers)
The square of the total wavefunction
The one-electron wavefunctions,
Definition: A molecular orbital is the wave function of an electron in a molecule moving under
the influence of the nuclear attraction and the average repulsion of all other electrons. It has
the same formulation as an atomic orbital does in a many electron atom.
An MO has an energy, , and the square of the wavefunction,
, or more correctly
,
gives the probability density for the electron.
MOs can be delocalized: an electron wave function extends throughout the molecule, and is not
confined to a "bond" between two atoms except in diatomic molecules.
inclusion of one atomic orbital or another influences the resultant molecular orbitals and
(consequently) the molecular properties.
In the simplest form of LCAO theory, only the valence orbitals of the atoms are used to
construct MOs (e.g. just the
for hydrogen, only the
and
for carbon, and so on). In more
accurate forms of calculation other orbitals are also included (see below in the treatment of H2+).
The atomic orbitals used are known as a basis set. In general, the larger the basis set the more
accurate are the calculations (e.g. of bond energies and distances). The down-side is that the
calculations become very computationally demanding as the size of the basis set increases, and
in general only the "simplest" systems can be calculated with a great degree of accuracy.
Linear combinations of atomic orbitals (LCAOs) are used to construct molecular orbitals . If
we take two AOs,
and , of energies
and
respectively, we form two molecular
orbitals:
The MOs must form an orthonormal set. This comprises two key features, namely:
the normalization condition:
and the orthogonality condition:
Overlap
It is obvious that atomic orbitals must overlap in order to combine to form an MO. In other
words, the overlap integral
between two atomic orbitals
and
must be non-zero.
In diatomics, it is relatively straightforward to see which orbitals overlap and which do not:
In polyatomic molecules (see Section 7 below), it is not so easy to see, and we need to use group
theory. The key result is that only orbitals carrying the same symmetry labels (when analysed in
the same point group) can overlap, otherwise
is zero.
Orbitals which overlap constructively (i.e. "in-phase") are bonding. Orbitals which overlap
destructively are antibonding. In general, the greater number of internuclear nodes the higher the
energy of an orbital (c.f. AOs in the H atom, particle in a box, etc.).
(but see Baird, J Chem Ed 1986, 63, 663 for a discussion of the origin of bonding (or JEM
lectures in QM supp!))
The energy of stabilization of a bonding MO, and the energy of destabilization of an
anti-bonding MO (with respect to the combining AOs) depends on:
The size of overlap,
. The energy of
stabilisation of an MO is at a maximum when the two combining AOs have similar energy.
Conversely, orbitals with very different energies will interact poorly, irrespective of the size
of the overlap.
To ascertain the best possible wave function within the LCAO approximation, we use the
variation principle. The variation principle states: The lower the energy calculated for a
wavefunction, the closer the wave function is to the true wave function.
For a trial wavefunction,
, the associated energy is given by:
(
is the true energy
in the molecular orbitals
which give the minimum possible energy (i.e. we vary them until the change
is less than a target criterion). There are a number of well-established computational procedures
for doing this.
5
5.1
An approximation to the wave function of the electron in the one-electron ion H2+ may be made
by summing the two
atomic orbitals,
on atom HA and
on atom HB. The two atoms
(i.e. H) in this species are obviously the same so the coefficients introduced in section 4 are
identical and in this case are represented by N, the normalisation constant.
5.2
However, such a "first-guess" wave function gives only 64% (!) of the experimental bond
dissociation energy of H2+ and a bond length that is 0.26 too long. In other words, the
wavefunction that we have chosen to start from is pretty inadequate.
5.3
It can be improved by contracting the 1s atomic orbitals. These wavefunctions have the form:
where Z is the nuclear charge acting on the electron. It is found that a minimum energy is
obtained when Z = 1.24 rather than Z = 1.
5.4
Doing this gets us 98% of the experimental bond energy, a considerable improvement on the
first-guess approximation of 64% in 5.2 above). The effect of mixing in the
orbital is to
polarize the electron density along the bonding direction and into the region between the two
nuclei. Use of such a high energy orbital leads to them being described as polarization
functions. Pictorially (and not strictly to scale) this improved position can be visualised thus:
6
The table below summarises the various improvements to the calculated dissociation energy and
bond length of H2+ and compares these to the experimental values.
Results of some calculations on H2+
Dissociation energy
(kcal mol-1)
Bond length ()
41
1.32
54
1.06
63
1.06
64
1.06
MO using 1s and
AO
Experimental values
5.5
Wave functions for the hydrogen molecule H2 may be generated by placing two
electrons (labelled "1" and "2") with opposite spins in the bonding MO [i.e.
]
+
that we have just derived for H2 . The spatial wavefunction for the molecule is then given by
the product of the two one-electron wavefunctions,
where
and
6.
We shall first consider linear heteronuclear diatomic molecules AH where A is one of the second
period atoms, Li to F. These atoms have one
and three
valence orbitals. In such systems it is
conventional to take z as the molecular axis (as is also the case for homonuclear diatomics A2).
Symmetry analysis: Pt grp Cv
From the group theory (character) table for Cv (see appendix):
For the A atom (4 AOs): (
)=
)=
)=
6.1
The valence
and
symmetry). The
atomic orbitals of the A atom overlap with the H 1s orbital (all are
() have no symmetry match on H, so will be non-bonding. The three
Perturbation theory tells us that the interaction between two orbitals with different energies
S2
depends on the overlap (squared) divided by the energy difference between them (
) (4.9).
E
Even if overlap is good, when orbitals have very different energies their interaction will be very
weak and they will be effectively non-bonding.
For HA we can distinguish two limiting cases.
i)
and H 1s.
and
and the lower energy orbital (1) now behaves like a core
orbital.
9
and
6.2
and
Li
Be
Ne
- 5.4
- 10.0
- 15.2
- 21.4
-26.0
-32.3
- 40.0
- 48.5
- 3.5
- 6.0
- 8.5
- 11.4
- 13.4
- 14.8
-18.1
- 21.6
1.9
4.0
6.7
10.0
12.6
17.5
21.9
26.9
Note that it is widespread practice (and generally convenient) to refer to "" and "" type
orbitals even though in a strict (group theoretical) sense these Mulliken symbols for irreducible
representations only formally apply to the cylindrical point groups Dh and Cv (even in the
latter instance it is also permitted to use the a1, e1 etc. symbols adopted here). For example, in
ethene we like to refer to H2C=CH2 "" and "" bonds even though the molecular orbitals
being referred to actually have the Mulliken symbols ag and b1u in the D2h symmetry of
ethene.
10
Symmetry and molecular orbital diagrams for the first row hydrides AHn
We have already seen for the AH molecule how knowledge of the symmetry (irreducible
representations) of atomic orbitals helps identify those atomic orbitals that (in principle at least) can
overlap to form MOs. Symmetry becomes an indispensible tool when analysing the bonding in
polyatomic molecules AHn with n > 1.
7.
So far we have a good idea of the MO pattern when 2 or 3 AOs interact. With polyatomic systems we
have a potentially very large number of AOs, so expect an equally large number of MOs. However if a
molecule is symmetric (or can be approximated as such) we can still deal with their interactions in a semiqualitative way.
If two or more atoms are chemically identical, they must be associated with equal electron density.
In H2 we have:
Therefore in any molecule with two equivalent H atoms (H2O, HCCH etc) the contribution to any
MO in the molecule will be either
or
Such linear combinations of
equivalent atomic orbitals are known as Symmetry Adapted Linear Combinations (SALCs).
If we can generate SALCs for equivalent orbitals for molecules just by using symmetry, we pre-fix the
relative coefficients of the AOs in the MOs and the interaction between the different atoms is then often
reduced to that between 2 or 3 orbital sets. Many examples of this will be presented throughout these
lectures and later courses.
SALCs for various symmetry arrays of equivalent orbitals are given in the appendix, along with relevant
character tables. You should now be familiar with these from the "Symmetry 1" 2nd year lectures.
11
8.
We shall take as our specific example H2O, but the general scheme developed will be applicable to
any C2v symmetric AH2 main group molecule. We will use group theory to classify the irreducible
representation of SALCs of the H atom 1s AOs in the C2v symmetry of bent AH2. Having then
identified the irreducible representations of the O atom
and
orbitals (in C2v symmetry) we shall
then construct an MO diagram for H2O using the principles of AO overlap and energy separation
developed above.
8.1
Symmetry analysis
First we choose the coordinate system. Conventionally we assign the z-axis as the C2 axis, take yz as
the molecular plane, and assign the x-axis as being perpendicular to molecular plane,
Now identify the AO basis sets. These are H atom 1s and the O atom
and
orbitals
Assign these AOs or SALCs to their irreducible representation in the C2v point group of H2O.
We find these from group theory (character) tables for the central atom (on an invariant point),
and either by inspection or using the reduction formula (see "Symmetry 1" lectures) for the
SALCs of the peripheral atomic orbitals. Therefore:
For the O (A) atom (4 AOs): ( ) = a1
(
) = a1 (
) = b1
For the H atoms (2 x 1s AO): (2 x 1s) = a1 + b2
Hence for the resulting H2O molecule (6 MOs): (MOs) = 3a1 + b1 + 2b2
8.2
) = b2
This simple symmetry analysis shows that the in-phase (a1) SALC of H atom 1s AOs can
interact with both the
and
or
AOs). So in
AO of oxygen, so
The
AO of oxygen cannot interact at all with either HA or HB and so will be strictly nonbonding in H2O.
8.3
Starting with the 6 AOs (4 from O, 2 from 2 H atoms), to construct an MO diagram we now
have reduced the problem to combining a set of three orbitals (2a1 + b2 on oxygen) and a set of 2
orbitals (a1 + b2 SALCs). Symmetry has told us the number of each type of MO. We must use
calculation or qualitative arguments to get the final ordering or the MOs. These are given
below. Note that the ordering of the 2a1 and the 1b2 orbitals is not obvious. As predicted in 8.2,
The net result is: three MOs of symmetry a1, two of symmetry b2 and one of symmetry b1.
C2v Mulliken symbols for the oxygen AOs and (HH) SALCs even though the real symmetry
of these fragments is much higher than that of the resultant H2O molecule.
8.4
A variational calculation for water gives the AO coefficients shown in the Table below. The results
show that there is little oxygen
or H 1s AO contribution in the lowest of the 1a1 symmetry
valence MOs.
Table: MO coefficients for H2O
Oxygen AO
Hydrogen AO
MO
1s1
1s2
1a1
-0.222
0.843
-0.132
0.151
0.151
2a1
-0.093
0.516
0.787
-0.264
-0.264
1b2
0.624
0.424
-0.424
1b1
8.5
The schematic cartoons given in the MO diagram above emphasise the AO contributions and are
helpful to visualise the LCAO origins of the MOs. However, a more accurate picture of an MO function is
given by a contour diagram such as those shown below.
14
9.
We shall take as our specific example NH3, but again the bonding scheme developed will be broadly
applicable to any C3v symmetric AH3 main group molecule. As above we will use group theory to
classify the irreducible representation of SALCs of the three H atom AOs, along with the
and
AOs of the central N atom in the C3v symmetry of the molecule.
9.1
Symmetry analysis The coordinate system conventionally chosen has the C3 axis lying along
the molecular z axis.
The
and
AOs of N and the SALCs of C3v (H)3 span the irreducible representations:
9.2
9.3
The resultant MO scheme for NH3 is illustrated below. Note that there is a fair degree of
mixing (like the BH molecule above). Therefore in the a1 orbital manifold N-H bonding is welldeveloped in the more stable orbital (i.e. the 1a1 MO ) with 2a1 possessing somewhat more nonbonding character. Indeed, the two electrons in the 2a1 level are the MO equivalent of the NH3
"lone pair" obtained in valence bond descriptions of the bonding.
15
The
and
(4 x ) = a1 + t2
(MOs) = 2a1 + 2t2
16
) = a1
) = t2
Note that (by symmetry at least) we anticipate no non-bonding MOs in this AHn compound
since all of the carbon atomic orbitals (a1 + t2) have a symmetry match among the (H)n SALCs.
In the AHn (n = 1, 2, 3) compounds discussed above symmetry analyses predicted three, two or
one non-bonding MOs (AH: 1 x a1 and 1 x e1 non-bonding MOs; AH2: 1 x a1 and 1 x b1; AH3:
1 x a1 non-bonding MO).
10.2 Note that the combinations derived from the
form a triply degenerate set of MOs
because the C3 operation in the Td point group mixes the three
AOs so they must have
identical energy.
10.3 The resultant MO scheme for CH4 is illustrated below. Note that although there is a favourable
- AO energy separation, there is no s-p mixing in CH4 because the
and
orbitals
have different irreducible representations in the Td point group (the overlap, S, is zero, and
S2
therefore so is
.
E
11.
It is helpful to have some experimental tests of the electronic structures proposed and this is where
photoelectron spectroscopy (PES) has important applications.
Just as atomic spectroscopy can give information on atomic orbitals and their energies, we can obtain
information on molecular orbitals by studying ionization of molecules.
KE h IE
In a photoelectron (PE) experiment, monochromatic radiation (single energy photons) h, are
used to ionize gas phase molecules, and the kinetic energy (KE) of the ejected electrons is
measured.
Einsteins equation is used to convert the KEs to ionization energies (IEs).
IE = h - KE
A PE spectrum consists of the number of electrons N(E) of a particular energy plotted against
the IEs.
11.1 The simplest molecular PE spectrum is that of H2. Photoejection of an electron leads to the
formation of H2+. The PE spectrum of H2 is very well understood and is reproduced below.
18
Figure: Potential energy curves for ionisation of a diatomic (a) with minimal change in geometry (b)
with large change in geometry.
The depths of the two curves for H2 and H2 correspond to the bond dissociation energy for
each species. The energy difference between the two minima corresponds to the ionization
energy (IE) (15.45 eV for H2). The horizontal straight lines drawn "within" the potential energy
for H2+ correspond to the vibrational energy states in which H2+ can be formed. Under the
conditions of the experiment ionization occurs from H2 in its lowest vibrational state (" = 0).
11.2 The fine structure in the ionization band arises because excitation of the electron happens so fast
that the nuclei do not move during the transition. This is known as the Franck-Condon
principle that you will meet in many branches of spectroscopy.
The molecular ion is formed with the same geometry as the neutral molecule. Therefore,
given that the photon energy is sufficient, a molecular ion may be formed in a number of
19
vibrationally excited states as well as in the ground state. Thus a PE spectrum consists of a
number of discrete bands of different IE, even though one is ionizing from a single orbital for
each envelope of bands.
In the case of H2+ the equilibrium bond length (re = 1.06 ) is longer than that of 0.74 for the
H2 molecule (because the bond order of H2+ is only 0.5) case (b) in the figure above.
Therefore it is most probable that the molecule is formed in a vibrationally excited state. We
can see from the vibrational structure of the PE band that the vibrational state in which H2+ is
most likely to be formed has ' = 2.
11.3 Koopmans approximation equates ionization energy to the negative of an orbital energy.
This is an approximation as orbitals tend to be slightly different in molecular ions than in
molecules, and as we have seen above MO calculations themselves are not normally accurate.
Nevertheless, Koopmans' theorem gives a good indication of orbital energies in molecules.
20
12.
12.1 Schematic representations of the PE spectra of the isoelectronic series of species: CH4, NH3, H2O,
HF and Ne are shown in the following Figure. The photon energy used is not sufficient to ionize the
core electrons so for each system we are looking at the ionization of the eight valence electrons.
21
Figure: PE spectra of the isoelectronic series of species: CH4, NH3, H2O, HF and Ne
12.2
The number of bands are related to the symmetry of the species under consideration:
Species
Symmetry
No. of bands
CH4
NH3
H2O
HF
Td
C3v
C2v
Cv
2
3
4
3
t2, a1
2a1, e, 1a1
b1, 2a1, b2, 1a1
, 2, 1
Ne
p, s
22
13.1 Consider the case of a general AH2 molecule. It can have either a linear (Dh) or a bent (C2v)
structure. We already have an MO scheme for a bent AH2 molecule such as water (see Section
8). A linear one is readily derived.
Symmetry analysis: Point group Dh
For the A atom (4 AOs):
( ) = g (
For the H atoms (2 x
AO): (2 x ) = g + u
) = u (
) = u
23
24
Figure: Walsh correlation diagram between the MO levels of linear and bent AH2
13.2 The principal feature that determines whether or not an AH2 molecule is bent or not is whether
the 2a1 orbital is occupied. The 2a1 orbital has a much lower energy in the bent geometry than
the linear one. Hence, when the 2a1 orbital is occupied, a lower energy is achieved if molecule
is bent. The shape adopted by an individual AH2 species depends on:
Primarily the number of electrons in the valence orbitals (as just stated).
Secondly (e.g. for 8 electron systems: H2O, H2S, etc.) the energy gap E (see Walsh
diagram above) between the 2g and 1u MOs (this governs the extent of s-p mixing and
the extent of stabilisation of the 2a1 MO on bending).
13.3 Known shapes of some AH2 molecules are summarised in the following table.
Known shape of some AH2 molecules
Molecular species
Shape
LiH2+
Bent
LiH2, BeH2+
Linear
Linear
Bent
Bent
Bent
Bent
BeH2, BH2+
BH2, AlH2, CH2
-,
For three or four electrons the molecules are linear as this gives a more stable 1b2 orbital. For
more than four electrons the molecules are bent as the 2a1 orbital is then occupied and this orbital
changes most rapidly with angle (i.e. is stabilised).
25
Bond angles in bent AH2 species are summarised in the table below.
Bond angles (o) in bent 5 - 8 electron AH2 species
2
1a1 1b2 +
2a1
BH2
2a1
131
CH2
NH2
BH2
AlH2
119
SiH2
2a1 1b1
110
NH2
115-120
OH2
102
CH2
93
2a1 1b1
103
OH2
110.5
FH2
+
-
104.5
118
99
NH2
PH2
92
SH2
92.1
AsH2
91
SeH2
90.6
TeH2
90
104
13.4 Molecules with just one electron in the 2a1 orbital have significantly larger H-A-H angles than
those with two. Remember that to bend the molecule destabilizes the b2 electrons. With two
electrons in the 2a1 orbital there is more stabilization on bending than with just one. Occupation
of the b1 orbital, which is non-bonding in both linear and bent configurations, makes less of a
difference to the angle.
13.5 The H-A-H valence angles of eight-electron molecules H2O, H2S and H2Se are 104.5, 92.1 and
90.6o, respectively. These molecules have the same MO configuration, namely
1a121b222a121b12. The difference in H-A-H valence angles comes from the extent of mixing of
the 2g and 1u(z) MOs of linear AH2 on bending. This is heavily influenced by the energy
separation E between them (see Walsh diagram above).
As the diagram below shows, the energy gap E decreases with decreasing electronegativity of
the atom A (e.g. down the group O, S, Se, Te).
Therefore as E decreases, the mixing of the 2g and 1u(z) MOs increases giving a more stable
2a1 level and so greater driving force for bending
13.6 It is also possible, from the changes in the population of the orbitals, to predict how the shape
will change when a molecule is electronically excited. The configuration for CH2 given above is
in fact an excited state. The ground state is a triplet, with configuration 1a121b222a111b11, and
an angle of 136. The angle is wider because there is only one 2a1 electron. SiH2 has a singlet
ground state with a 1a121b222a12 configuration and an angle of 93. The triplet and singlet
excited states with configurations 1a121b222a111b11 have a wider angle of 123.
13.7 Photoelectron spectroscopy (again) A great deal of information can also be obtained from the
vibrational structure of the bands in a PES.
The PE spectrum for H2O may be compared with the MO diagram (see page 13 and the Figure
below). There are four bands in the valence region, three "p" bands and one "s" band. The first
band, arising from the 1b1 orbital, has minimal vibrational structure and is characteristic of
ionization from a non-bonding orbital. The second (2a1) band shows a vibrational progression
associated with bending the water molecule.
27
The PE spectrum of NH3 may be also compared with its MO diagram (see 8.5 above and the Figure
below).
The PE spectrum of NH3 shows three main bands exactly as predicted by the LCAO MO scheme.
The lowest energy band is for ionization from the 2a1 MO (2A1 state) and exhibits extensive
vibrational fine structure. This is at first sight surprising since we have previously described the 2a1
MO of NH3 as partially non-bonding (corresponding formally to the "lone pair" in a valence bond
description). But again we have to recognise that we are forming the NH3+ ion in a vibrationally
excited state. The equilibrium geometry of NH3+ (like the CH3 radical and BH3) is trigonal
planar. The vibrational spacing of 900 cm-1 in the first 2A1 band corresponds to the energy
required to bring the NH3+ from a trigonal pyramidal (C3v) shape in ground state NH3 to one that is
trigonal planar (D3h) in NH3+.
28
The trigonal planar (D3h) shape favoured by 7- and 6-electron 1st row compounds AH3 can readily
be explained by reference to the appropriate Walsh diagrams for the process C3v AH3 D3h AH3
(see Albright, Burdett and Whangbo: "Orbital Interactions in Chemistry" in the Bibliography).
In the PE spectrum for CH4 the presence of two bands corresponds to ionization from the t2 and
a1 orbitals, and so the PE spectrum is fully consistent with the MO description. The fine
structure in each of the PE bands is consistent with significant bonding character in the two MOs
from which ionization is occurring.
29
14.
The molecules considered so far have obvious ties between the MO descriptions and traditional ideas
of electron pair bond formation (i.e. molecules with n bonds and 2n bonding electrons). But as you
know from "electron-deficient" compounds such as diborane, B2H6 (containing two 3-centre-2electron B-H-B bonds), this is certainly not always the case. In this Section we use Walsh diagrams to
explore the molecular structures of H3+ and H3 and discuss their bonding in terms of 3-centre-2electron (3c-2e) and 3-centre-4-electron (3c-4e) interactions. We return to the topic of 3c-2e and 3c4e bonds in our discussion of octahedral hypervalent compounds in later Sections.
14.1 The principal alternative shapes for H3 and H3 are triangular (D3h) and linear (Dh).
The SALCs of triangular H3+ are easily derived from the three H 1s AOs: (3 x 1s) = a1' + e'.
These combine to give three MOs of symmetry a1' + e' as shown below. In H3+, two electrons
occupy them.
The H3+ ion has been observed by mass spectrometry in electrical discharges through H2 gas.
The bonding orbital a1' accommodates the two electrons and the overall electronic structure can
be described as a 3c-2e bond. The H3+ ion is thermodynamically stable with respect to H+(g)
and H2(g). Calculations have shown that the enthalpy of dissociation for the process:
H3+
--
H2
H+
is about 400 kJ mol-1 which is comparable to that of the H-H bond in H2 itself. This example
shows how important delocalisation of electron pairs is for producing stability in molecules.
The SALCs of linear H3- can be constructed using a fragment orbital approach as usual. In this
case the two outer hydrogens [(
) = g + u] combine with a central H atom
[(
) = g]. This is analogous to the MO diagram in 11.1 for a Dh AH2 molecule. The
rather obvious difference is that the central H atom does not possess valence p orbitals. Four
electrons occupy the 1g and 1u MOs.
30
The situation in H3- is usually described as 3c-4e bonding. H3 is the simplest example of this
quite widespread model and we shall look at other examples later in our analyses of SF6 and the
hypothetical SH6. The "surplus" two electrons in H3 are accommodated in a HH nonbonding MO (1u) localised on the terminal HA,C atoms.
14.2 Of course, we want to know why H3+ has the triangular structure while H3 is predicted to have
a linear one. A Walsh diagram shows how the energies of the D3h and Dh H3 MOs correlate.
Clearly for 4 electrons (H3 ) the linear structure is favoured but for 2 electrons (H3+) the
triangular one gives the best electronic stabilisation.
Figure: Walsh correlation diagram between the MO levels of linear and triangular H3
31
Here we still maintain the desired octet and overall this suggests delocalised bonding with an
average Xe-F bond order of 0.5. A common approach to obtain a 2c-2e (single) valence bond
description is to use the high energy 5
orbital on Xe (not normally used in an Aufbau sense
until the 6s and 4f elements have been passed) to produce five sp3d hybrids. These five hybrid
orbitals can accommodate the 10 electrons (3 lone pairs and 2 bonding pairs) and each Xe-F
bond can now be described as a 2c-2e localised single bond. The intuitive difficulty with this
model is that the 5
AO of Xe must be very high in energy and so one should question
whether it is reasonable (and indeed necessary) to invoke its use in bonding.
In this section we explore the MO description of XeF2 with and without the use of the 5
AO
and see how the photoelectron spectrum helps support the MO picture developed. To do this we
first develop an MO description of CO2 which is related to that of XeF2.
32
15.2 Symmetry analysis: In Section 11 above we developed an MO scheme for linear (Dh) AH2.
The symmetry analysis of CO2 and XeF2 is similar except that now the terminal atoms bring np
(n = 2 or 5, respectively) valence AOs as well as the ns AOs to the bonding description:
;6x
) = u (
AOs): (2 x
(2 x
) = u
) = g + u
(2 x
) = u + g
) = u + g
The symmetries (i.e. 3g + 3u + 2u + 1g) of the MOs of CO2 and XeF2 are identical
although the contributions of the various AOs to the bonding (see below) differs, as does their
occupation by electrons. We consider first the bonding description of CO2 which has 16 valence
electrons (4 from C and 12 from 2 x O respectively).
15.3 The resultant MO diagram for CO2 is shown in the Figure below. In the -bonding manifold
there are three g and three u symmetry MOs. These MOs (without
- mixing shown for
the terminal atoms) are predominantly bonding, non-bonding and anti-bonding in character. In
the -bonding framework there are two bonding (1u) and two anti-bonding (2u) MOs. The
1g MO is based entirely on the terminal atoms as there is no symmetry match (i.e. no g AO)
on the central atom. Note that there are 4 occupied orbitals with significant bonding character
33
(2g, 2u and 1u) and four empty antibonding counterparts. The net bond order is therefore 4,
or 2 per bond (i.e. C=O double bonds).
Note also that s-p mixing will accumulate bonding character in the 1g and 1u orbitals and make
2g and 2u rather more non-bonding, but this does not affect the bond-order analysis.
effect by adding some bonding character. The g SALC is also stabilised somewhat by
interaction with 5
. It is not clear, however, to what extent this orbital mixing actually
occurs, if at all. The 5
AOs, although the electron withdrawing effects of the fluorine atoms will generate a partial
positive charge on Xe that would to some extent stabilise the 5dz2 orbital (see discussion of SF6
in Section 16 below).
The stability of XeF2 can be explained without d-orbital participation (i.e. delocalised 3c-4e
bonding). 5
orbital participation would help stabilise the 3g MO (with the latter then
becoming more -bonding in character), but it is not essential, and probably does not occur to
the extent implied by a formal sp3d hybridisation model.
15.5 The photoelectron spectrum of XeF2 is shown below. It can be assigned assuming Koopmans'
theorem (IE = -i) and the level ordering on the right hand side MO diagram above in which the
3g orbital lies below 2u
35
Figure: Qualitative MO energy level diagram for XeF2 without (left) and with (right) 5d participation.
36
16.
Figure: Some resonance forms for SF6 that conform to the octet rule
As for XeF2 such resonance structures imply a delocalised bonding description such as that
anticipated by an MO analysis. But as was the case for XeF2, valence bond 2c-2e models can at
first sight be postulated. Thus if S can use its high energy, vacant
and
atomic
orbitals (not normally employed in an Aufbau sense until at least Ca has been passed) then a set
of sp3d2 hybrid orbitals can be produced. Each of these six hybrids then contributes to a 2c-2e
S-F bond. Again there is a problem with proposing the extensive use of 3d orbitals on sulfur:
the 3d AOs lie 800 kJmol-1 higher in energy than the 3p and so the extent to which they can
contribute is arguable. On the other hand, sophisticated MO calculations have suggested that the
formal charge on S in SF6 is as large as +3.0e which would stabilise the 3d orbitals significantly.
Here we will analyse the general -bonding in molecules such as SF6 and show how a
satisfactory delocalised multicentre model (without 3d AOs) can account for the bonding.
We shall take as a starting point the hypothetical model hexahydride compound AH6. The
orbital analysis will be a general one and is applicable to both 12- and 8--electron systems with
appropriate electronegativity perturbations as we shall see later.
16.1 Symmetry analysis: The coordinate system chosen has the H atoms of AH6 arranged along
Cartesian axes:
The ns and np and nd AOs of atom A and the SALCs of Oh (H)6 span the irreducible
representations:
For the A atom (1 x ns and 3 x np AOs):
(s) = a1g
(
) = eg
) = t1u
) = t2g
Hence for the resulting AH6 molecule (15 MOs): (MOs) = 2a1g + 2t1u + 2eg + t2g
This simple symmetry analysis is a general result for any AH6 -only octahedral complex with
s, p and d AOs included in the A atom orbital basis set. You will return to this many times in
discussions of the bonding in transition metal complexes. In the AH6 post-transition metal
complexes we need to consider two distinct cases, namely with or without nd orbital
participation.
16.2 In the first instance we will treat the nd AOs as being too high in energy to contribute to the AH bonding. The symmetry analyses therefore predicts:
4 bonding MOs (a1g + t1u) and 4 anti-bonding MOs (a1g + t1u).
5 A atom-based non-bonding (i.e. the nd) orbitals (eg and t2g)
2 H atom-based non-bonding LCAO/SALCs (eg)
A fragment orbital interaction diagram for AH6 without d-orbital participation in shown on the
following page. The 12 valence electrons occupy the six lowest energy MOs. Four MOs are AH bonding (1a1g and 1t1u) and two (1eg) are non-bonding and located entirely on the H atoms.
Thus four highly delocalised electron pairs are involved in forming six A-H "bonds", and the
net bond order for each A-H bond is 2/3. There are clear parallels between the situation in 12
electron AH6 molecules and linear H3. The lowest energy level in AH6 arises from the in-phase
overlap of the central atom ns AO with all six ligand 1s orbitals, but the triply degenerate 1t1u
set are three-centre bonding orbitals just like the lowest energy MO in H3 (see Section 12.1).
The bonding in H3 leads to a single non-bonding MO (1u) located on the terminal H atoms,
whereas in AH6 there is a pair of doubly degenerate (1eg) non-bonding orbitals located on the
ligands. Therefore, just as in H3, there is a build up of electron density on the peripheral
atoms.
38
Figure: Qualitative fragment orbital interaction diagram for a 12 valence electron AH6
Compound without (left) and with (right) participation of the nd atomic orbitals on the central atom.
39
The important conclusion of this analysis is that he best stabilisation of a 12 electron AH6 type system will arise when the terminal atoms are electronegative ones (i.e. best able to
stabilise the build up of negative charge). This is in general true: for example, SF6 is known
whereas SH6 is not. In a qualitative sense we could have anticipated this from the AH6 MO
diagram below: if the peripheral ligands are electropositive then the 1eg set will lie at high
energy and the compound would be expected to undergo oxidation very easily and such a
compound would be inherently unstable.
16.3 We shall now consider the effect of bringing the five nd (eg + t2g) orbitals into the bonding
picture (see Figure above, right). The t2g set (
) are unable to find a match with any
of the (H)6 SALCs, but the eg (
) set have the correct nodal properties and symmetry
to interact with the previously non-bonding (H)6 eg SALCs. The bonding LCAOs are shown
below:
It is apparent that the previously non-bonding 1eg set is stabilised, all 12 electrons are now in
bonding orbitals and so the formal bond order is increased to 1.0 per A-H bond. However, the
extent to which d-orbital participation actually contributes to the bonding in post-transition metal
compounds is still a subject of ongoing investigation. All things considered it is currently
believed that the role of d orbitals in such compounds is at most minimal; the bonding is best
considered in terms of multi-centre bonding with d orbitals making only a minor contribution
(somewhat like the contribution of
polarisation functions to the -bonding in H2+ and H2.
40
17.
CH62+
The final class of AH6 type hypervalent compound to consider are eight valence electron
species such as CH62+ (observed only in the gas phase) and [C(AuPR3)6]2+. The latter class of
complex are stable at room temperature and a number of these have been crystallographically
characterised. The AuPR3 ligand is a one-electron, -type donor ligand like H and so
[C(AuPR3)6]2+ complexes are electronic relatives and models of CH62+. [aside: you will see in
Organometallics next year that H and AuPR3 are examples of isolobal species. They have
similar frontier orbitals with the same population by electrons; therefore they often form related
sets of compounds.]
Figure: Some resonance forms for CH62+ that conform to the octet rule
Just as for XeF2 and SF6 above, one can write down resonance structures that comply with the
octet rule. Again such structures suggest a delocalised description of the bonding in MO theory,
but now the terminal ligands carry a positive rather than negative charge. However, unlike the
situation for XeF2 and SF6 we cannot invoke the participation of high energy d orbitals on
carbon (the closest in energy would be the 3d). Therefore we are compelled to seek only a
hypervalent bonding description and this can be easily done using the basic MO scheme
developed above for 12 valence electron AH6 compounds (Section 16.2). A qualitative MO
diagram for CH62+ is given in the Figure on the following page. This is also a model for
[C(AuPR3)6]2+ for the reasons outline above.
The bonding in CH62+ features a set of four filled -bonding MOs (1a1g and 1t1u) together with
the corresponding vacant -anti-bonding MOs (2a1g and 2t1u) at high energy. Just as for 12
electron AH6 compounds there are two non-bonding MOs located entirely on the peripheral
atoms/ligands (i.e. the 1eg set). However, the key important difference in the 8 valence
electron CH62+ (and [C(AuPR3)6]2+) is that the non-bonding eg MO is vacant.
17.1 The fact that the eg level is vacant for an 8 valence electron AH6 / CY62+ species has several
important consequences for the types of complex that will form eight electron AH6 systems:
41
In the 12 electron species the ligand-based 1eg MOs of AH6 are occupied so this is favoured
by compounds with electronegative peripheral atoms and relatively electropositive central
atoms (hence SF6 is a good example). This will keep the 1eg orbital low in energy.
In the 8 electron species the ligand-based 1eg MOs of AH6 are vacant and so they need to be
high in energy if the molecule is to be stable. This requires the peripheral ligands to be
electropositive relative to the central atom. This is one of the likely reasons why the gold
complexes [C(AuPR3)6]2+ in particular are so stable: here the 1eg MO will be somewhat higher
in energy that the carbon
AO manifold.
42
18.
18.1
NH3.BH3
Simplified version:
43
18.2
B2H6
B2H6
44
19.
2 electrons per vertex atom are involved in a bond or lone pair directed radially out from the cluster. These
do not contribute to the bonding within the cluster skeleton, so delete 2 x no. of vertex atoms from the
TVEC
e.g.
P4:
SEC = 20 4 x 2 = 12
B5H9 SEC = 24 5 x 2 = 14
B6H62- SEC = 26 6 x 2 = 14
Sn52-
45
SEC = 22 5 x 2 = 12
19.2
Wades rules
B5H52-
B6H62-
B7H72-
B8H82-
B12H122-
TVEC
22
26
30
34
38 = 4n + 2
SEC
12
14
16
18
26 = 2n + 2
TVEC
SEC
TVEC
SEC
B6H62-
B5H9
B4H10
26
14
24
14
22
14
2n+2
2n+4
2n+6
B7H72-
B6H10
30
16
28
16
2n+2
2n+4
46
B5H11
26
16
2n+6
19.3
2n
Bonding orbitals:
1 strongly bonding symmetric combination of orbitals
+
n less strongly bonding combinations of the orbitals where n is the number of vertices
In general, for an n-vertex, closo polyhedron, there are n + 1 bonding orbitals in the cluster skeleton
SEC = 2n + 2.
TVEC = 4n + 2
47
t1u
eg
t1g
t2u
6x
12 x
t1u
t2g
a1g
14e-
remove BH2+
B6H62-
19.5
14e-
'B5H54-'
add 4 H+
14e-
B5H9
When we remove BH2+, we remove 3 orbitals from the cluster framework. We can replace the missing
BH2+ fragment with any other isoelectronic species, as long as it provides 3 orbitals with similar symmetry,
for example CH3+.
48
14e-
14e-
2+
remove BH
'B5H54-'
B6H62-
3+
add CH
C
C
14e
14e
CB5H6-
C2B4H6
A lone pair on each atom, pointing out of the cluster, plays the same role as the terminal B-H or C-H bond
in boranes or carboranes.
Sn
Bi
Sn
Pb
Bi
Bi
Ge
Sn
Sn
Sn
2-
Pb4
Ge
Ge
Bi
Pb
Pb
Ge
Ge
Pb
2-
Sn5
49
Ge
Ge
Ge
2-
Bi4
Ge
Ge92-
SEC = 22 10 = 12 = 2n + 2
SEC = 22 8 = 14 = 2n + 6
2-
Both B4H10 and Bi42- are arachno, but in one case 2 trans vertices are removed, in the other 2 cis vertices.
The Bi42- case is uncommon (sometimes called iso-arachno)
Transition metal chemistry: metallaboranes and metallacarboranes
Fe, Ru, Os are d8: 6 electrons are in the t2g orbitals, directed away from the ligands, leaving 2 to contribute
to the bonding. Therefore M(CO)3, M = Fe, Ru, Os are isolobal with BH.
50
20.
You have already seen electrostatic (Crystal Field) approaches to the metal-ligand interactions in
transition metal complexes. It is perhaps self-evident that such an electrostatic theory would not be
appropriate for compounds such as Cr(CO)6 which has Cr in its zero oxidation state and neutral CO
ligands. Moreover, small, anionic ligands like F- produce only a small splitting of the d-orbital t2g and
eg set energy levels while CO generally causes one of the largest. In addition there are electron spin
resonance data for complexes such as [IrCl6]2- [Ir(+4) which has a d5 configuration with an unpaired
electron] that reveal that the unpaired Ir 5d electron density is associated significantly with the 6
chloride ligands as well as the Ir centre. All of these observations are suggestive of a more covalent
bonding model that we shall now develop using molecular orbital theory.
d-Orbital energies. You have already seen for AH6 systems the effects of involving d-orbitals in
post-transition metal compounds such as SF6. In these systems there is little participation of the nd
orbitals in the bonding because they lie relatively high in energy above the valence ns and np AOs.
This all changes for the transition metals. Now the valence AOs are the (n+1)s, (n+1)p and nd (n =
3, 4 or 5). Within these three sets, the nd orbitals are the most stable of the valence orbitals in metal
complexes as illustrated in the Figure below which shows a magnified view of the orbital order near Z
= 20 where the 3d elements begin. The nd AOs are stabilised on crossing the transition series, and
after the end of the series are best considered as being core-like and playing no further direct role in
metal-ligand bonding.
d orbital overlaps. The overlap of the d orbitals (and for 3d in particular) with ligand orbitals is
small as they belong to the inner quantum shell and are not very radially extended. The overlap
51
improves going down a group as the d orbitals acquire radial nodes and extend more into the
interatomic region. This is the principal reason why transition metal-ligand bond strengths increase
going down a group. This contrasts with the situation for main group metal-ligand bond strengths
which tend to decrease with increasing principal quantum number.
The contracted nature of the 3d orbitals is made very apparent by the Figure below that shows
the radial distribution functions for the valence orbitals of chromium. The maximum probablilty
(rmax) in the RDFs are 3d(rmax) = 50 pm; 4s(rmax) = 140 pm which may be compared to the metallic
radius of chromium (128 pm). The semi-core like nature of the 3d has important bonding implications
(see above); in contrast the very diffuse 4s orbitals overlap more effectively with the ligand donor
orbitals.
Figure: Radial distribution functions (RDFs) for the valence orbitals of main group, transition metal
and lanthanide ions.
Figure: Diatomic orbital overlap, SAA, as a function of orbital size (at fixed distance)
52
) = t1u
M atom (5 x nd AOs):
) = eg
) = t2g
Figure: The -only LCAO MO diagram for a 1st row octahedral transition metal complex ML6
A number of important points follow from this LCAO scheme:
53
The 1a1g, 1eg and 1t1u orbitals form a set of M-L -bonding orbitals. Because M is less
electronegative than L they are more localized on the ligands than the metal.
In other
words, electrons in these MOs are ligand-based. We should expect the 1eg orbitals to be
more stable than the 1t1u because nd AOs are more stable than (n+1)p AOs. Indeed, the
(n+1)p AOs may be so high in energy, especially for early transition metals, that they
contribute very little to the bonding. The 1a1g MO is more stable than the 1eg due to better
overlap of the (n+1)s AO with the (L)6 SALCs as compared with that of the more
contracted nd AOs.
The 2a1g, 2eg and 2t1u orbitals form a set of M-L * anti-bonding orbitals. They are
localized on the metal. Obviously populating these with d-electrons will lead to a reduced
M-L bond strength.
The 1t2g and 2eg ( anti-bonding) MOs have entirely (in the case of 1t2g) or
predominantly (for 2eg) d orbital character. Any electrons in these orbitals are considered
to be "d-electrons".
It is therefore clear from the MO diagram that no metal complex MLn can ever have
electrons in the "s-orbitals" despite any corresponding gas-phase atomic configurations. All
metal-localized electrons in transition metal complexes are best considered as "delectrons". For example, the valence electron configuration of Cr(g) is 3d54s1 whereas for
Cr(0) in Cr(CO)6 the configuration is 3d6 [and in fact t2g6 eg0 for this Oh complex].
In the case of -only bonding ligands the 2eg 1t2g separation depends on the strength of the M-L
interaction. This should:
increase with charge on the metal, as this will stabilize the d orbitals and decrease the
energy gap between metal and ligand orbitals.
54
In crystal field theory, the 2eg orbitals are raised in energy with respect to the 1t2g orbitals
because of electrostatic interactions with the ligands while in ligand field theory they are raised
in energy because they are antibonding. The degeneracies remain the same as they are a
consequence of the symmetry of the complex rather than the particular bonding model.
Ligand field theory is easy to parameterize and therefore to use to model spectral and magnetic
properties. MO theory also gives a better qualitative understanding of the size of the d-orbital
separations in various complexes.
MO occupations in octahedral complexes
Recall Hund's rules tell us that an atom or molecule achieves a lower energy if electrons are
placed in separate degenerate orbitals with the same spin. This is because such a state keeps the
electrons further apart, decreases electron-electron repulsion, and maximise exchange energy.
If the lowest energy available orbitals are close in energy, but not actually isoenergetic, it may
still be advantageous to populate them separately with electrons of parallel spin rather than
pairing electrons in the same, lower energy orbital (c.f. Cr ground state, 3d54s1). This leads to
the concept of "low-spin" and "high-spin" complexes.
High-spin and low-spin complexes: For octahedral complexes with between 4 and 7 d-electrons
there are two alternative ways of filling the d orbitals. The complex may either maximize the
spin putting electrons into the upper eg orbitals, or fill the lower t2g orbitals pairing the electrons.
The two alternative configurations for a d5 complex are thus: t2g5 eg0 (low spin) or t2g3 eg2.
The size of o relative to the pairing energy is critical in determining whether a complex is
high or low spin. This is illustrated qualitatively for a d5 ML6 complex below.
For second and third row complexes the pairing energy is less due mainly to reduced electronelectron repusion for the more diffuse 4d and 5d AOs.
The number of unpaired electrons in a transition metal complex can be deduced from its
magnetic properties.
coupling is relatively small and the Russell-Saunders coupling scheme applies (see 3rd year
lectures).
Table: Electronic configurations of octahedral complexes
Electron confign
of metal ion
d1
d2
d3
d4
low spin
t2g1
t2g2
t2g3
t2g4
1
2
3
2
[Ti(H2O)6]3+
[V(H2O)6]3+
[Cr(H2O)6]3+
[Mn(CN)6]3-
high spin
low spin
high spin
low spin
t2g3 eg1
t2g5
t2g3 eg2
t2g6
4
1
5
0
[Cr(H2O)6]2+
[Fe(CN)6]3[Mn(H2O)6]2+
[Co(NH3)6]3+
high spin
low spin
high spin
d8
t2g4 eg2
t2g6 eg1
t2g5eg2
t2g6 eg2
4
1
3
2
[CoF6]3[Co(NO2)6]4[Co(H2O)6]2+
[Ni(NH3)6]2+
d9
t2g 6 eg3
[Cu(H2O)6]2+
d5
d6
d7
21
Ligands are ordered according to the relative magnitude of they produce. This order is known as
the spectrochemical series. For the more common ligands:
The effect of the donor ability of the ligands is rationalised by noting that the metal e g symmetry d
orbitals (2eg in the MO scheme on P53) are metal-ligand -anti-bonding. Ligands that are good
donors (i.e. their donor orbitals have a good overlap and/or energy match with the metal d orbitals)
56
form complexes in which the -bonding 1eg MO is better stabilised. Consequently the metal-based
2eg orbitals are more strongly anti-bonding and thus higher in energy.
For a full picture of most complexes we also need to consider the orbitals of the ligands. We will
take the basic picture derived for the -only system and add on the interaction. This is likely to be a
small perturbation on the system as overlap is smaller than overlap.
Symmetry analysis: Each of the ligands in a complex ML6 can offer up one or two -type orbitals for
M-L bonding.
The possible SALCs of ligand donor orbitals are shown below using for
convenience ligand p orbitals as the generating donor function (in reality they could be p orbitals on
e.g F- or * orbitals on e.g. CO only the symmetry matters for now). The twelve p orbitals (two on
each L) give four SALCs of t1g, t2g, t1u and t2u symmetry. Only the one with t2g symmetry can find a
match with the metal d orbitals. The other three sets of donor SALCs will be -non-bonding in the
metal complexes. Only one of each set of triply degenerate SALCs is shown:
57
Figure: SALCs of the orbitals on L (one of the triply degenerate set in each case).
For the t2g SALC, antibonding overlap with the metal orbitals is also shown.
Ligands tend either to have low lying orbitals that are occupied (e.g. F, Cl, NR2) or higher lying
orbitals that are unoccupied (e.g CO, CN). The effect on the largely metal t2g orbitals will differ in
the two cases. The nature and extent of the interaction is very important in determining the size of
.
-donor ligands destabilise the metal t2g d orbitals. These metal-localized MOs become
M-L -anti-bonding. The magnitude of decreases. A consequence of these effects is that
-donor ligands tend to stabilise high oxidation states (i.e. low d electron counts).
-acceptor ligands stabilise the metal t2g d orbitals. These metal-localized MOs become
M-L -bonding. The magnitude of increases. A consequence of these effects is that acceptor ligands tend to stabilise low oxidation states (i.e. higher d electron counts).
58
59
Summary figure: The two alternative effects of M-L -bonding as a perturbation of a ML6 -only
model
The spectrochemical series can therefore be rationalized in terms of the / donor / acceptor
properties of the ligands. ligands can stabilise or destabilise the t2g d orbitals, therefore
increasing or decreasing, respectively, .
60
-only
12 bonding + 6 non-bonding
([Co(NH3)6]3+)
-donor
12 bonding + 6 anti-bonding
([RuCl6]4-)
-acceptor
12 bonding + 6 bonding
(Cr(CO)6)
Electron counts < 18 are most common for donors, where the t2g orbital becomes slightly
antibonding.
61
Other factors in determining . The principal quantum number of the metal. As discussed on
Page 53 the overlap between the metal d orbitals and the ligand orbitals increases going down a
transition metal group. This results in an increase in as the metal eg orbitals become more strongly
anti-bonding.
complex
[Co(NH3)6]3+
[Rh(NH3)6]3+
[Ir(NH3)6]3+
o=
22,900 cm-1
34,000 cm-1
40,000 cm-1
The charge on the metal. The t2g eg d orbital splitting tends to increase with charge when ML interactions are the most important feature in the bonding. This is because charge stabilises
the metal orbitals and brings them closer in energy to the ligand orbitals. For example, in
[V(H2O)6]2+ o = 12,400 cm-1, whereas in [V(H2O)6]3+ o = 17,850 cm-1:
Figure: The increase in as charge on the metal increases (-only effects considered)
However, we need also to consider the balance between any improved interactions as the d
orbitals are stabilised and any decreased -bonding as the metal t2g orbitals move further away
in energy from any ligand acceptor SALCs. Thus for both [Fe(CN)6]4- and [Fe(CN)6]3- the
value of ~ 34,000 cm-1. This is because as the 3d orbitals gets closer in energy to the donor
orbital they get further away from the acceptor orbital of CN:
Figure: Opposing effects on the - and -interactions as the metal charge increase
62
22.
Apart from octahedral, the most common shape for a transition metal complex is tetrahedral ML 4.
Another four coordinate geometry is square planar (D4h) ML4 although this is typically found only for
complexes with a d8 (and occasionally d4 and Jahn-Teller distorted d9) configuration. We shall briefly
analyse bonded Td and D4h ML4 complexes.
Tetrahedral ML4 complexes
Symmetry analysis: The coordinate system chosen here has one of the C2 (S4) axes lying along the
molecular z axis.
The (n+1)s, (n+1)p and nd AOs of M and the SALCs of Td (L)4 donor orbitals span the
irreducible representations:
M atom [1 x (n+1)s and 3 x (n+1)p AOs]:
(s) = a1
M atom (5 x nd AOs):
(d) = e + t2
(4 x -donor) = a1 + t2
) = t2
Overall there will be a set of M-L bonding (a1 + t2) and anti-bonding (also a1 and t2) MOs. The nd e
orbital set (
63
orbitals that are only slightly * anti-bonding (because the metal has two t2 sets of valence AOs but
the (L)4 donor SALCs offer up only one t2 set).
An LCAO MO diagram for ML4 is illustrated below. The a1 ligand combinations form the
strongly bonding and anti-bonding MOs 1a1 and 2a1 respectively. The metal e symmetry
orbitals are non-bonding. Both of the metal t2 sets (from
and
) can, as
illustrated above, find a correct match with the ligand t2 SALCs. However, the greater overlap
with the metal (n+1)p orbitals results in strongly bonding 1t2 and anti-bonding 3t2 MOs derived
mainly from these. The smaller overlap with the metal nd orbitals yields weakly -anti-bonding
MOs (the 2t2 set) localised mainly on the metal and forming part of the "d orbital" manifold.
Thus, just as the t2g and eg MOs in Oh ML6 complexes are considered to be mainly d orbital in
character, so the 1e and 2t2 MOs in a Td ML4 complex are regarded in a similar way.
Figure: A -only LCAO MO diagram for a tetrahedral transition metal complex ML4 Note:
each orbital of t2 symmetry is a linear combination of the form
The separation between the 1e and 2t2 MOs in a Td ML4 complex is often referred to a t. The
value of t is less than half that of in an otherwise equivalent ML6 complex (i.e. for the same
64
ligands and metal). This can be attributed to the presence of fewer ligands in an ML4 complex
versus an analogous ML6;
In fact, based on d-orbital overlaps alone, it can be shown analytically (using the Angular
Overlap Model - see DeKock and Gray: "Chemical Structure and Bonding") that the relationship
between o and t for comparable MLn (n = 4 or 6) complexes is: t = 4/9 o
The consequence of the comparatively small t value is that virtually all Td ML4 complexes of
the 1st row transition metals are high spin (t < spin pairing energy even for strong field
ligands).
The (n+1)s, (n+1)p and nd AOs of atom M and the SALCs of D4h (L)4 span the irreducible
representations shown below.
M atom [(n+1)s and (n+1)p AOs]:
M atom (5 x nd AOs):
) = a1g
(s) = a1g
(
( ) = a2u
) = b1g (
) = b2g (
) = eu
) = eg
65
A general MO diagram for a square planar transition metal complex featuring only interactions is shown on the following page.
We can see from the MO diagram that:
There are four strongly -bonding (1a1g, 1eu and 1b1g) and strongly -anti-bonding
(3a1g, 2eu and 2b1g) levels. The former set accommodate the four pairs of electrons
required to form the four M-L bonds. Another a1g symmetry MO is found at
intermediate energy (i.e. the 2a1g orbital). This is because there are two AOs [the
(n+1)s and n
] on the metal than can interact with the (L)4 a1g SALC. As usual, it is
the (n+1)s metal orbital that overlaps best with the ligand donors. Hence the only
weakly anti-bonding "middle" level, namely 2a1, has predominantly n
character.
There are four strictly (by symmetry) -non-bonding MOs. These are the 1a2u, 1eg
and 1b2g orbitals. The 1a2u lies relatively high in energy because it is pure (n+1)
AO in character (remember that the (n+1)p orbitals are the least stable of the
transition metal valence orbitals). The 1eg and 1b2g orbitals are equivalent to the
triply degenerate t2g d orbital manifold in octahedral ML6 complexes. For a -only
complex, the 1eg and 1b2g orbitals are isoenergetic (have the same energy) but are not
all degenerate by symmetry - only the n
and n
pair (1eg). This is known as accidental degeneracy and will be removed in a more
sophisticated treatment (i.e. including effects).
66
Figure: A
-only LCAO MO diagram for a square planar transition metal complex ML4
SALC, so it is not used to form a bond, and it is empty, so is not used to hold non-bonding
electrons.
67
Figure: A -only Walsh diagram relating square planar and tetrahedral complexes ML4.
In deciding between tetrahedral and square planar coordination we should consider:
The nd electron count i.e. the population of the various nd orbitals. This should be done
in terms of the actual orbital energies (in terms of values) and the effects on M-L
bonding.
The t2 orbitals in Td-ML4 are somewhat antibonding (as is the 2a1g MO of D4h-ML4).
However in Td-ML4 there are four M-L antibonding electrons whereas in low spin D4hML4 there are only two. This weakening effect is shown in the average M-L bond lengths
listed below for d8 nickel complexes:
Bond
Square planar
Tetrahedral
NiN
1.68
1.96
NiP
2.14
2.28
NiS
2.15
2.28
NiBr
2.30
2.36
d8 ML4 complexes will favour a square planar structure in the case of a large value (i.e.
such that 6/9 exceeds the spin pairing energy). In this case the two electrons in the D4h-
68
ML4 HOMO (2a1g) destabilise the compound by 2 x 3/9 = 6/9 , whereas the four
electrons in the HOMO (t2) of Td-ML4 destabilise the compound by 4 x 4/9 = 16/9 .
However, if the value is small (i.e. with a weak field ligand) the energetic preferences in
terms of nd orbital occupation are insufficient to overcome the steric disadvantages of
forcing the ligands closer together and the Td-ML4 isomer is more stable.
Thus we find [Ni(CN)4]2- is square planar whereas [NiCl4]2- is tetrahedral. Note that
[PdCl4]2- and [PtCl4]2- are also square planar due to the larger values associated with the
4d and 5d elements.
69
23.
A miscellany of bonds
= {ReCl41-}2 so use square planar fragment orbitals as a starting point (from page 69)
2b1g on P69
n.b. same diagram for Cr2(CH3COO)4(H2O)2 (ICL lab course, Schlenk line expt)
70
71
72
Appendix
73
74