You are on page 1of 13

The Chemical Bond

John Murrell

Early Ideas
Atoms come together to form molecules, and most of these have well defined structures which
conform to well established chemical rules. Molecules have varying stabilities measured either by
the energies required to break them into their constituent atoms, or by their reactivities with other
atoms or molecules; again there are well established rules that explain most observations.
Theories of the chemical bond aim to explain the structures and stabilities of molecules, and
different theories aim for different levels of explanation. The most important theories are those that
give simple explanations for the structures and stabilities of the most stable molecules ; broadly
speaking, those that we find in bottles on the shelf of a laboratory. However, as experimental
chemists have moved forward their ability to make and study very reactive, maybe only transient,
molecular species, so our theories had also to advance. Moreover, computational chemistry has now
reached an advanced level, so that in some cases it is possible to predict their properties before the
molecules are made.
Let me take a simple example: a carbon atom can combine with four hydrogen atoms to make a
stable molecule called methane, whose structure and stability conforms to the simple rules of
valence; electron pairing and inert gas structure. But, there is also a well known dihydride of carbon
called methylene, CH2,which is an important transient in many chemical reactions. Moreover, this
molecule can exist in two different low energy forms (so called singlet and triplet spin states)
whose structures and reaction characteristics are quite different. These two forms of methylene do
not conform to the simple rules, and we do not explain their properties by simple theories. We need
more elaborate theories and computation to understand them. For this species computation was
ahead of experiment in predicting the correct structures which were wrongly assigned by the first
experiments.
At the beginning of the 19th century chemistry was developing as a subject from the discovery of
new elements and the realisation that in forming a compound the weights of the elements it
contained were in definite proportions; recognising the difference between mixtures and compounds
was a key step in chemical progress. Experiments based on the gas laws, which had been known
since the middle of the eighteenth century, were crucial for these advances. We generally give
Dalton the main credit for proposing that different atoms combine together in characteristic relative
proportions. These proportions were attributed to a property which Frankland called the valency of
the elements. Thus the valencies of hydrogen (1), oxygen (2), and carbon (4) were satisfied in the
stable compounds H2O,CO2,and CH4. Frankland also noted that the elements seemed to form into
groups having the same valencies: nitrogen, phosphorus, arsenic, and antimony, showed valencies
of three, for example, and this later led to the periodic table of the elements.
The important thing about valency is that it is such a good rule for stable molecules, and that its
exceptions are often molecules of great interest. Its application to organic chemistry led to major
advances in understanding structure;conversely the proliferation of organic molecules in the
nineteenth century greatly extended our concepts of the chemical bond; take Kekules benzene, for
example. Unarguably valency is the most important idea in chemistry.
Drawing chemical structures with bonds, as we now do, was well established by the second half of
the nineteenth century, and recognising that these bonds made well defined angles with one another
was another important advance. This idea came from several sources, particularly from the work of
vant Hoff and Le Bell who proposed the tetrahedrally oriented four bonds emanating from carbon.
Further advances in understanding valence had to await the theory of the electronic structure of
atoms, following Thomsons discovery of the electron in 1897. This came with Rutherfords model
of the atom in 1911, and Bohrs explanation of atomic spectra in 1913.
Although electrical forces had been invoked to explain the chemical bond in the 19th century by
Berzelius and others, it was G .N.Lewis who took the theory forward from the Bohr model of the
atom. It is noteworthy that Lewis made no use of the quantum numbers which arise in Bohrs theory
( later extended by Sommerfeld to include elliptical orbits),as we would today. Lewis concentrated
his attention only on the closed electron shell numbers of the atom: 2,8,18, etc.
Lewis theory contains two essential ingredients, electron pairing, and the attainment of closed shell
(inert gas) structures. Both of these ideas find support from the later, quantum mechanical theory of
the chemical bond, which is why it was such an important advance in chemical thinking.
Lewis structures for hydrogen and hydrogen chloride are written as follows:

H:H H:C l:

The two dots between the atoms represent electron-pair bonds. The other dots around the Cl
represent the other outer shell electrons of the atom. At the time that this proposal was made the
origin of the pairing was unknown because electron spin had not been discovered. There is some
implication in Lewiss work that magnetism might be relevant; most importantly, he noted that the
tendency of electrons to form pairs was not a property of electrons in general but a property of
electrons in atoms.
We can write the Lewis structures given above in a short notation as follows:

H - H H - C l:

The lines indicate the electron pairs. We could also miss out the dots altogether if they are not
relevant to further understanding. In this way we simply identify an electron pair with the bonds
that are formed to satisfy the rules of valence. Extending this to multiple bonds gives structures like
O=O NN H2C = CH2
Again, valence is satisfied and electron pairing is implied .
Going back to the original structures, we see that by sharing electrons we have inert gas structures
for each atom; He -like for H, and Ar -like for Cl. By analogy we would have Ne-like structures for
O,N,and C in the second group of molecules we have looked at. But we know that the structure for
O2 does not reveal the paramagnetism of this molecule, as compared with the diamagnetism of N2,
nor do we see that different bond strengths can arise from different types of electron pairs.
A structure for BF3 as follows
F
B
F F
satisfies the valency of three for boron and one for fluorine, and can also be taken to imply
electron-pair bonds, but the boron in this molecule does not have an inert gas structure; its electron
shell has only six electrons. BF3 is a perfectly stable molecule when isolated or in the presence of
other BF3 molecules, but it is very reactive with many other molecules, and we can attribute this
reactivity to the tendency of the boron to obtain its inert gas structure. Reaction with F- is typical,
the (BF4)- now having an inert shell structure.
In summary, the failure of each atom to reach an inert gas structure does not prevent stability but it
does imply reactivity. BF3 is called a Lewis acid (a very strong one at that) arising from its strong

2
affinity for extra electrons. In contrast, a molecule like NH3 in which not all of the eight electrons
around the N are forming electron pair bonds (the one left is what we call a lone pair), is a strong
electron donor or Lewis base.
A Lewis acid and a Lewis base in combination can give strong complexes such as NH3BF3. Writing
its structure as

H F
H N B F
H F

implies a total of seven electron -pair bonds ,which is correct, but it incorrectly attributes valencies
of four to both N and B. We satisfy both valency and electron pairing if we write the alternative
structure

H F
+ -
H N B F
H F
because N+ and B- have the same number of electrons as (are isoelectronic with) carbon, and hence
would be four -valent. An alternative representation

H F
H N B F
H F
is also used in chemistry, the arrow indicating an electron-pair bond in which both electrons are
contributed by the atom at the tail of the arrow.

Resonance
Although we have given two representations of the NH3BH3 complex they are not what chemists
call different structure, rather, different representations of the same structure. Different structures
are different bond arrangements which satisfy the normal rules of valence. The most important
example of this is benzene, which we can draw as two distinct Kekul structures, both satisfying the
tetra-valence of carbon.

Another example would be the three structures that can be given for the carbonate anion

O- O O-
- -
O C O C O C
- -
O O O

The correct valencies are satisfied because O- is iso-electronic with F, and is mono-valent.
Two or more structures which can be drawn for a molecule satisfying the rules of valence are called
resonance structures. It is not correct to suppose that the molecule oscillates between these bond
structures; benzene in the bulk does not consist of an equal mixture of two molecular forms. Rather,

3
the actual form is an average of the two structural representations and we say that the bonds, or
rather the electrons contributing to the bonds, are delocalized. All the CC bonds in benzene are
intermediate between single and double. The normal length of a CC single bond is 1.54, and a
normal double bond is 1.34; the bond length in benzene is 1.40.
Chemists recognize that when resonance structures can be written the molecule has extra stability
compared with a molecule for which there is only one structure. This is particularly true if the
structures are equivalent, as they are for the examples used here. Thermochemical studies, and other
chemical properties, show that benzene is appreciably more stable than a molecule which has three
localized CC double bonds. The correct interpretation of resonance by quantum mechanics was
given by Pauling, and has been widely used in organic chemistry. We now know that for
quantitative work resonance is a model that has to be handled with caution.
Electron Pair Repulsion
We can use Lewis theory to understand some features of molecular structure if we analyse the
consequences of electrostatic repulsion. We start by adopting the principle that electron pairing is an
important feature of the chemical bond, and count the number of electron pairs around an atom.
Consider the simple example of NH3. Should this have a planar or a pyramidal arrangement of the
three N- H bonds? Nitrogen has five electrons in its outer shell (inner shell electrons can be ignored
for bonding purposes ), and with the three hydrogen electrons this makes a total of eight which we
treat as four electron pairs. If we wish to minimise the repulsion between these four electron pairs
we would direct them from the nitrogen towards the corners of a tetrahedron. The angles between
the tetrahedral directions are 109.5o , and this should be the H N H bond angle in NH3 if three of the
electron pairs are taken to be the N H bond pairs. Experiments show that this angle is 107.8o.
The simple theory therefore works quite well, but note that we would get the same result for PH3,
yet the experimental bond angle is 93.3o ; at least we are correct in concluding that PH3 ,like NH3,
is not a planar molecule with bond angles of 120o.
The fourth pair of electrons in NH3 or PH3 we call lone pair electrons. They are directed in space
towards the other corner of the tetrahedron ,and there is good evidence for this; when NH3 behaves
as a Lewis base the extra bond it makes comes in from that direction. In the simplest example of
this, NH4+ has tetrahedrally directed N H bonds.
If we accept that four electron pairs are tetrahedrally directed then the bond angle in water (two
bond pairs and two lone pairs) should also be 109.5o,but in fact it is 104.5o(in H2S it is 92.2o).
Water, also acts as a base, forming bonds along tetrahedral directions, as for example in ice. Also in
HF we could have one bond pair and three lone pairs making tetrahedral angles and this is supported
by structures in which H F behaves as a Lewis base in one component of the HF dimer.
We have seen that NH3 is not planar, and neither is NF3, and to include this molecule in the theory
we say that each F contributes just one of its valence electrons to forming the N-F bonds. But
suppose we turn to BF3; boron has two electrons less than nitrogen so we are now counting three
electron pairs ,all forming bonds, and if we want to minimize the repulsion between three electron
pairs we direct them in coplanar directions that make angles of 120o with one another. This is the
bond angle in BF3 ; unlike NF3, it is a planar molecule.
There are no good examples of molecules with two electron-pair bonds and one lone pair in which
the bond angle is close to 120o. SnCl2 is often quoted, and its bond angle is approximately 100o.
Other likely candidates such as CF2, CH2, etc, have several low energy states ,some with larger and
some with smaller bond angles. More success is found for electron repulsion theory in molecules
where the central atom has five or six electron pairs. The minimum energy structure for five pairs is
a trigonal bipyramid, and for six pairs is an octahedron.

4
The structures of PF5, SF6 and ClF3, are all based on a trigonal bipyramid. PF5 has such a structure
and note that it possesses two kinds of fluorine atoms, one in what is called the polar position and
one in a position called equatorial. However, there is exchange of atoms between these two
positions, and it is necessary to go to quite low temperatures before this is frozen out. The fluorine
atoms appear equivalent in room temperature nmr studies, for example.
SF4 has one lone pair and four bond pairs. ClF3 has two lone pairs and three bond pairs. In both
molecules the lone pairs preferentially occupy equatorial positions. SF4 has been called a sea - saw
molecule, and ClF3 is T- shaped.
Examples of structures based on an octahedron are SF6, IF5,and XeF4. IF5 is a square pyramid, and
XeF4, which has four bond pairs and two lone pairs, is square planar; the lone pairs preferentially
occupy trans positions in the octahedron.
This general model for molecular structure based on electron pair repulsion was first put forward by
Sidgwick and Powell in 1940, and later refined by Gillespie and Nyholm to explain the relative
orientation of lone pairs, and the distortions from ideal structures based on lone-pair bond -pair
repulsion. For example, in BrF5 the F B F bond angles are all less than 90o, showing that the lone
pair repels the bond pairs, or, alternatively, that the lone pair requires more space than a bond pair.

Quantum Theory
Until the beginning of the 20th century the manner in which bodies interacted with one another was
all thought to be described by what we call classical mechanics which originated in the work of
Newton. Also, visible light, which was by 1900 recognized as part of the general family of
electromagnetic radiation (including radio waves, microwave, infra-red, ultra-violet, X-ray and
gamma rays),was widely accepted to have a wave form due to its properties of diffraction and
interference. There were three fundamental developments of the 20th century which changed these
views. Firstly, in 1905,Einstein changed our view of light in the theory of special relativity (in 1916
going further with general relativity). This was a vital step for physics but ,except in special
situations, has no relevance to the chemical bond so we will not comment further. Secondly, and
also in 1905,Einstein showed that experiments in which light impinging on metals emitted
electrons, could best be interpreted by assuming that light was a stream of particles (which we call
photons),and that the energies carried by these photons are equal to h or hc/ where c is the
velocity of light, is its frequency, and is its wavelength (c=). The constant h is a fundamental
constant of nature called Plancks constant which has the value 6.626x10-34 J s (note that its units
are energy times time).
The third important development was quantum theory. This started in 1900 by the work of Planck
who showed that the distribution of electromagnetic radiation from heated bodies,( how much
radiation comes out at each frequency, the so-called black-body radiation distribution), could only
be explained by assuming that matter takes in or emits radiation in discrete amounts, which he
called quanta. To put this in a chemical context we take the following example: molecules vibrate,
and we can picture this by taking two atoms connected by a spring. But whereas classical balls and
springs can take up energy in any amount, however small, molecules can take up energy only in
units of h (multiples of h),where h is Plancks constant, mentioned above, and is the natural
frequency of vibration of the spring. Thus the energies of atoms and molecules are discrete
(quantized) and not continuous (and not describable by classical mechanics).

The quantum ideas of Planck and Einstein, together with a vast amount of data on the spectroscopy
of atoms, was developed by Bohr in 1913 into a quantum theory of the electronic structure of
atoms; called semi-classical mechanics because it was still based on Newtons equations. This was
immensely successful for the hydrogen atom, and gave a good qualitative picture of many other
5
aspects of atomic spectroscopy. However,it was clearly seen to fail for other features of atomic
spectroscopy, and it never gave quantitative results except for one-electron atoms and ions. Most
importantly for chemists, it never gave a quantitative explanation for the strength of the chemical
bond for even the simplest example,H2.

The full quantum theory as we now know it originated from an important postulate by de Broglie in
1923 that matter could exhibit wave like properties, and particularly that the wavelength of a
particle having a momentum p(p=mv where m is its mass and v its velocity) has an associated
wavelength given by =h/p. Long wavelengths are therefore
possessed by particles having small momenta, and in practice this is only physically important for
particles with very small masses; particularly for electrons and nuclei. The manifestation of wave-
like properties is shown most clearly by diffraction and this was confirmed in 1927 by experiments
on electrons by Davisson and Germer, and in 1928 by Thompson in experiments on atoms. One of
the most important matter-wave applications today for chemistry is the use of neutron diffraction to
determine the structures of molecules and solids.
If matter has wave-like properties then there must be a wave equation to describe it (just as there is
for any other wave such as light, or sound, or water waves).The appropriate equation was derived by
Schrodinger in 1926. He started with the general equation for harmonic (sine-like )waves and
introduced the de Broglie relationship. We will not give the details of the result,which is known as
the Schrodinger equation, but emphasize the fact that solutions of this equation for atoms and
molecules agree very well with experimental results. The Schrodinger equation is the correct
equation for describing the role of quantum theory in chemistry.
Atomic Orbitals and Electron Spin
(See also Chemical Spectroscopy document)
The next step in valence theory is the recognition of the relevance of electron spin, and of atomic
orbital wave functions. Electron spin was proposed in 1924 by Uhlenbeck and Goudsmit from the
observation that an electron had a magnetic moment, and that there are only two energy states for
the interaction of this magnetic moment with an external magnetic field; we say colloquially that an
electron can have two spins, and that in a typical electron-pair bond the two electrons have opposite
spins.
Atomic orbital wave functions come from the solution of the Schrodinger equation for the hydrogen
atom, and with certain approximations this model can be extended to all atoms. The symbols used
to identify atomic orbitals originate from the old form of quantum theory due to Bohr, and this in
turn used much of the symbols coming from the early experiments on atomic spectroscopy.
Atomic orbitals are associated with the discrete energy states of the electrons in atoms which are
labelled 1s,2s,2p,3s,3p,3d, etc. All s orbitals are spherically symmetrical about the nucleus (strictly
speaking it is the wave functions of the orbitals that are spherically symmetrical), and all others
have characteristic non-spherical shapes which are important when we consider how one atom
bonds to another; this will be discussed later. The p orbitals come in sets of three, all having the
same energy; d orbitals come in sets of five, etc.
The important role of electron spin in bonding, and in the Periodic Table of the elements, comes via
the Pauli Exclusion Principle. We can state this simply as no two electrons in an atom can have the
same space wave function and the same spin. It follows that an s orbital can take (be occupied by)
only two electrons, and if there are two they must have different spin states; a set of p orbitals can
take six electrons, two in each of the three orbitals; d orbitals can take ten electrons, etc.

6
The lowest energy orbital of an atom is an s orbital and is labelled 1s. There is then a large energy
gap before the next orbital, which is also an s orbital, called 2s. This is followed by a small energy
gap before we come to a set of p orbitals called 2p. There is then another large gap before we come
to the third shell of orbitals,3s,3p and 3d. The next shell is the 4 shell, although in some atoms the
first orbital in this shell,4s, has a slightly lower energy than 3d. The integers which appear in these
labels are called the principal quantum numbers (n) of the orbitals. The symbols s,p,d,etc, come
from spectroscopy and are associated with another quantum number (an integer l) which can have
the values 0,1,2, for s,p,d, respectively (and further integers and symbols which we will not deal
with here).
The full description of the electron structures of atoms is obtained by allocating electrons to the
lowest energy orbital subject to the Pauli Exclusion Principle; that is, we assign electrons to orbitals
starting with the lowest ,1s,through 2s,2p,3s.3p,3d or 4s ,4p,etc..Thus the inert gas electron shells
which come into Lewis theory of stable systems are those of He(1s2), Ne(1s2 2s22p6), Ar (1s2
2s22p63s2 3p63d10) etc. The superscripts that appear in these descriptors show the number of
electrons in each orbital set.
The periodic classification of the elements arises from the fact that chemical properties depend only
on, and are characteristic of, the outer, or valence electrons. Thus Li (1s22s1), Na (1s22s22p63s1), and
K(1s22s22p63s23p64s1), are in the same group of the Periodic Table called the alkali metal group,
and all three atoms form compounds having similar chemical properties.
The Pauli Exclusion Principle, as stated above, can be extended to molecules by noting that
electrons in molecules have wave functions ,which are called molecular orbitals, and,as for atomic
orbitals, these can hold two electrons providing they have different spins. This principle actually
comes from a more fundamental rule of quantum mechanics that the wave functions of electrons in
atoms or molecules must change sign on the exchange of two electrons;this is the so-called
Antisymmetry Principle.
The Chemical Bond
Heitler and London were the first to give an interpretation of Lewiss electron pair bond by
quantum mechanics in their 1927 paper on the hydrogen molecule. They showed that the overlap of
two 1s atomic orbitals, one on each hydrogen atom, with electron spin pairing (the two electrons
must have opposite spins,or,to be more precise the two electron spins are coupled together to give a
so - called singlet spin state) would explain why H2 had a lower energy than two hydrogen atoms.
Overlap of two atomic orbitals across a bond, and the coupling of electron spins in the way just
described, is the essence of the electron pair bond.
Let us extend this discussion to the water molecule, H2O. The oxygen atom has the electron
configuration (1s22s22p4), and as there are three 2p orbitals , two of these must contain unpaired
electrons. These orbitals can overlap with hydrogen 1s atomic orbitals. If the electron spins are
coupled this would give rise to two electron-pair O-H bonds. Maximum overlap of the oxygen 2p
and hydrogen 1s orbitals occurs if the H O H bond angle is 90o, and the observed bond angle of
104.5o is not far from this. In the periodically equivalent molecules H2S, H2Se,and H2Te, the angles
are 92o,91o, and 90o, respectively, so this simple model ,which we can call the maximum overlap
model, becomes progressively better for the heavier atoms.
The model for H2O which we have described, based upon pure 2p - 1s bonding, also works quite
well for NH3 and similar molecules, and extensions to O-halogen and N - halogen bonds is
straightforward if we invoke overlap of halogen p orbitals with oxygen or nitrogen p orbitals
directed along the bonds. However, the model clearly fails for carbon compounds because carbon is
rarely divalent, so that overlap with the carbon 2p orbitals (two of these contain unpaired

7
electrons), is not sufficient to describe chemically stable species. To explain the tetra-valency of
carbon one needs to introduce a new concept, which is hybridisation.
Hybridisation
The ground state of carbon has the configuration (1s22s22p2) and, as mentioned, might be expected
to form stable molecules like CH2 or CF2. However, such species have only transient stability. It is
energetically very favourable to extend the carbon valency to four to make stable compounds like
CH4 and CF4. To achieve this in the context of electron pair bonds one must first envisage an input
of energy to raise the atom to a higher state having the configuration (1s22s12p3). A minimum
energy of 400kJ/mol is required to do this but this allows the carbon atom to form four electron
pair-bonds and the extra bond energy is more than enough to compensate for the energy input.
However, we are not yet in the position of explaining how the four bonds in a molecule like CH4
are equivalent to one another, nor that in this molecule the four bonds are directed towards the
corners of a tetrahedron. A further step is needed to explain this, and that is to suppose that the atom
is prepared in an artificial state, called a valence state, in which four hybrid atomic orbitals are
formed, all equivalent in shape and energy, but differently directed in space.
The four hybrid orbitals used to describe the bonds in CH4 are called sp3 hybrids. A hybrid is a
mixture, in this case a mixture of one 2s and three 2p wave functions. If we label the hybrid orbitals
by the general symbol t, we can define the electron configuration of the valence state of the atom by
(t1t2t3t4),ie one electron in each of four hybrids. It can be shown that these hybrids are directed to the
corners of a tetrahedron. In CH4, each t orbital overlaps with one hydrogen 1s orbital to form an
electron pair bond - four such bonds in all, leading to a tetrahedral molecule.
A characteristic of carbon is that it can form multiple bonds as well as single bonds. In molecules
such as ethene or benzene, the carbon bonds make angles of 120o with one another. A hybrid
construction can be found to represent this by mixing the 2s orbital with two of the 2p orbitals
(these are called sp2 hybrids), and the remaining 2p orbital forms a second bond by overlapping with
a similar orbital on an adjacent atom, (to give a double bond); this is called a bond. Lastly, in
molecules like ethine, the bond angles are 180o and these can be formed by constructing two sp
hybrids, leaving the remaining two 2p orbitals to form two bonds, giving a triple bond overall.
Hybridisation is an important concept in chemistry, particularly for carbon molecules. For the
second and higher period main group elements, whose ground states have incomplete s or p shells,
it is also usual to suppose that bonding occurs via valence states in which the s electrons are
promoted, but additionally, valencies above their minimum, such as five for phosphorus or six for
sulfur, can be explained by considering valence states in which there are electrons in d orbitals, and
in which the resulting hybrids are (s, p,d) mixtures. For example, six equivalent octahedrally
oriented hybrids can be formed from sp3d2 hybrids; that is, each hybrid has (1/6) s, (3/6) p, and
(2/6)d character. Trigonal bipyramid structures such PF5 can be formed by sp3d hybrids. At one time
these hybrid models also provided a popular picture for transition metal complexes, but this
approach is now out of fashion, and has been replaced by other models.
The great merit of the electron-pair bond and hybridisation model is that it explains why bond
properties are often approximately constant from one molecule to another for a given pair of atoms;
this is what chemists call bond transferability. For example, the length of a C- H bond is always
about 1.08A, its bond stretching force constant is about 5mdyneA-1, and its bond dissociation energy
is about 400kJmol -1. This transferability can be understood by a model in which all C- H bonds are
formed by overlap of a carbon hybrid orbital and a hydrogen 1s orbital; if we put atoms into
different classes for differently hybridised carbon atoms, then transferability is even better than for
the average C- H bond.
The weakness of the approach we have described is that it gives a poor account of electron
spectroscopy , particularly of the type known as photoelectron (PE) spectroscopy. The technique of
8
this is rather simple; light having a discrete wavelength () incident on a molecule, can eject
electrons. Knowing the energy of the absorbed photon (equal to hc/) and measuring the kinetic
energy of the ejected electrons, gives ,by an energy balance, the energy required to eject an electron
from the molecule with zero kinetic energy; this is called the ionization potential. A molecule has a
spectrum of ionization potentials corresponding to the different electronic states in which the
positive ion can be left after ionization; this is called the PE spectrum.
From the electron-pair bond model we would predict that removing an electron from any of the four
bonds in methane would require the same amount of energy. There would be a single peak in the PE
spectrum from such an ionization. However, the PE spectrum of methane shows two peaks, one at
14eV (1eV=96kJmol -1),and the other at 24eV. How these two ionization potentials arise is easily
explained by another approach to the chemical bond called molecular orbital (MO) theory.
Molecular Orbitals
Molecular orbitals play the same role for molecules as atomic orbitals do for atoms; they are the
wave functions of the electrons. With each MO we associate an energy, and the electronic states of
molecules are described by assigning electrons to MOs with the restraint that two electrons in the
same orbital must have opposite spins. The lowest energy state of the molecule is obtained by
assigning electrons to the lowest energy MOs available.
There are many models for calculating MO wave functions and their energies. Some of these are
very simple, achieving not much more than an overall pattern or order of energies. Others can be
considered as exercises in numerical mathematics, the aim being to get accurate solutions of the
Schrodinger equation with the aid of powerful computers. The first MO calculations were made by
Hund in 1927, and today we find the model is the basis of most accurate molecular wave functions.
Molecular orbitals are, in general, very complicated functions in three dimensional space , and it is
not feasible to calculate them and store them by direct tabulation. The normal procedure which is
followed is to expand the wave functions in terms of a series of known functions, and it is then only
necessary to tabulate the coefficients of these functions in the expansion.
By far the most common expansion functions in this procedure are the atomic orbitals of the
constituent atoms of the molecule. This is called the linear combination of atomic orbitals (LCAO)
approximation, and this can be concisely expressed as follows
k = r crk r (1)
where k is a particular MO wave function ,and the r a set of atomic orbital wave functions. The
coefficients crk are characteristic of the particular MO but the functions r, which we call the basis,
are the same for all MOs of the molecule being considered.
In qualitative theories, which aim to produce patterns of energies and wave functions rather than
accurate numerical representations of these, very few terms in the expansion (1) are used. A simple
MO description of H2 ,for example, would have just two atomic orbitals, the 1s wave functions of
the two atoms. Likewise, a simple MO theory of unsaturated hydrocarbons (ethene, benzene, etc)
uses an expansion with one 2p-type atomic orbital for each carbon atom; this is the basis of what is
called Huckel MO theory.
An important relationship exists between molecular orbital energies and the bands in a PE spectrum.
A theorem ,due to Koopmans, says that within certain limitations the ionization potential associated
with the removal of an electron from an MO is the negative of the molecular orbital energy.
Although this might appear to be self- evident it is not a trivial theorem because electrons repel one
another, due to their charge, and therefore the total electronic energy of a molecule depends on all
the electrons present. The theorem has quite a complicated proof, and we know its limitations;

9
notably, after ionization there should be little reorganisation of the remaining electrons in the
molecule.
We will not concern ourselves in this article with the details of the computational procedure to be
followed to obtain accurate molecular orbital wave functions and energies; these are important for
research as a guide to experiments. However, by and large, such calculations have not led to great
advances in the conceptual models of the chemical bond. Of much greater importance in this
respect, are the very simple models which were developed in the 1930s.
Let us start with the simplest electron pair bond which is in H2. Forming MO wave functions by
combining the two atomic orbitals, as in (1), we find only two possibilities, namely
g = Ng( 1 + 2 ) (2)
u = Nu( 1 - 2 ) (3)
where Ng and Nu are called normalisation constants, whose significance will be explained shortly.
There are sound physical arguments for the above forms. In quantum mechanics a wave function
has the physical interpretation that its square is identified with a probability distribution, in this case
the probability distribution or electron density for the electron. As we know that the two atoms in
H2 are identical they must have the same electron densities in their surroundings. It follows that the
coefficients multiplying 1 and 2 in the wave functions must be equal ,or equal in magnitude but
opposite in sign; expressions (2) and (3) conform to this rule and there are no other possibilities.
The numbers Ng and Nu which appear in (2) and (3) also have important physical significance. If
2 is a probability density for an electron, then integrating this probability over all space must give
unity (the electron must be somewhere with probability one), hence
g2 dv = Ng2 ( 1 + 2 )2 dv = 1 (4)
dv being a volume element in three dimensional space, and the integration being over all space.
There is a similar integral which defines the value of Nu. If the individual atomic orbitals are
normalised, that is,
12dv = 22dv =1 (5)
then expression (4) becomes
Ng2( 2 + 2 S12 ) = 1 (6)
where
S12 = 12 dv (7)
is called the overlap integral between the two atomic orbitals; in qualitative terms we can say that it
represents the amount of space shared by the two atomic orbitals.
S12 is quite easily calculated if we know explicitly the atomic wave functions. For example,for two
hydrogen atom wave functions with the atoms separated by a distance R, the integral (7) can be
shown to be given by the expression
S12 = ( 1+ R + R2/3) exp(-R) (8)
If we know the value of S12 we can determine the value of Ng from (6),or of Nu from a similar
expression.
At large values of R the wave function just looks like the sum of two separate hydrogen atom wave
functions, and when R=0,that is the two nuclei have been combined to give effectively a He
nucleus, then the wave function must look like the lowest orbital of He, which is also s - type.
Between these two situations the wave function takes on a form that overlaps both nuclei, and at a

10
value of R close to the equilibrium distance in H2, the wave function has roughly the shape of an
ellipse.
Note that the function u is positive in the region of one nucleus (a peak) and negative over the
other (a trough). The wave function must be zero in a plane passing through the centre (and
perpendicular to the bond), and we call this a nodal plane.
In expressions (2) and (3) the wave functions of the MOs have been distinguished by the suffixes g
and u; the normalisation constants have likewise been distinguished. These suffixes are standard
symbols used in symmetry group theory; g means that the wave function is symmetric to inversion
through the centre of the bond, and u that it is anti-symmetric (changes sign) on inversion.

Let us now look at the electron density for the H2 wave functions in more detail. By squaring (2)
and (3) we get the following expressions
g2 = Ng2 (12 + 22 + 212) (9)
u2 = Nu2 (12 +22 - 212) (10)
We see that g leads to a build-up of electron density in the middle of the molecule arising from the
positive contribution of the overlap density 12; u , in contrast, has this term coming in with a
negative sign so that the electron density is depleted in the centre of the molecule.
The energies associated with the MOs can be calculated from the Schrodinger equation, but we can
see qualitatively how these energies are related to the energy of the H atomic orbital by the
following argument. The region in the molecule where electrons experience the greatest attraction
will be between the two nuclei. We therefore conclude that the energy associated with g is lower
than that associated with,u. It can be shown more explicitly that the energy of g is lower than that
of a separate hydrogen atom atomic orbital ,,and the energy of u is higher than this energy. We
call g a bonding MO ( having an energy lower than its atomic orbital components) and u is called
an antibonding MO (having an energy higher than that of its component orbitals). In simple models
the stabilization energy of the bonding MO is equal in magnitude to the destabilization energy of the
antibonding MO.
The MO description of diatomic molecules
We can use the pattern of MO energies just discussed to explain the stabilities of the diatomic
species ,H2+, H2, He2+,and He2. For He as well as H the wave functions have the forms shown in (2),
and (3), being the 1s atomic orbital of He in this case. This series of molecules has 1,2,3, and 4,
electrons respectively. The first two electrons can go into the bonding MO,g, providing that they
have opposite spins, and the last two must go into the anti- bonding MO u . The net number of
bonding over antibonding electrons in the series is therefore 1,2,1,and 0,respectively, and we would
therefore expect bond strengths (dissociation enthalpies) to increase from H2+ to H2, then decrease to
He2+, and for He2 to be essentially zero. This is in accord with the experimental facts. The
dissociation enthalpies for the first three members of the series are 256, 432, and approximately
300kJmol-1, and He2 is not a stable molecule.
This treatment of H2, He2, and their positive ions might be considered rather trivial, but the model
produces very important and non- trivial results when applied to the diatomic molecules of the first
row of the periodic table, Li2 to F2. Lennard-Jones in 1929 was the first person to note the success of
this theory, and he used it to explain why O2 is a paramagnetic molecule.
The valence atomic orbitals of the atoms Li - F are 2s, and 2p, and in their diatomic species we
expect, following the above discussion, that these would interact to give bonding and antibonding
MOs. Firstly, the 2s atomic orbitals would give bonding and antibonding orbitals very similar in
form to those we have examined for H2. We will now label these 2g and 2u respectively, the g

11
and u referring to the inversion symmetry as before, and the symmetry label showing that an
electron in this orbital is cylindrically symmetric about the internuclear axis; more properly that an
electron in this MO has zero angular momentum about the axis. The label 2 is to indicate that these
are the second orbitals of this symmetry type ,the first being those that arise from the interaction of
the inner shell 1s orbitals. The general form of the MO wave functions is precisely the same as in
(2), and (3), but now the atomic wave functions, , are 2s functions.
The manner in which p orbitals interact to give bonding and antibonding combinations is more
complicated; there are three 2p orbitals on each atom, and these produce three bonding and three
antibonding MOs. We have to distinguish the 2p orbitals that are directed along the internuclear axis
(z say), which are of - type, from those that are directed perpendicular to this axis (called -type).
We note that there will be bonding and anti- bonding MOs (3g and 3u respectively) and
bonding and antibonding MOs (1u and 1g respectively) the last two occurring in pairs because the
two perpendicular axes (x and y) are equivalent.
Because the overlap of p- type orbitals on the two atoms is generally smaller than between p-
type for typical equilibrium distances, the bonding - antibonding energy split is less for p than for
p. Experiment confirms this as the pattern of energies for O2 and F2. However, for the earlier
members of the period, Li2 to N2, the difference in energy between s and p atomic orbitals is
sufficiently small that one cannot ignore s - p interaction (which in the context of electron pair
bonds would be called hybridisation). The effect of this is to push the 3g bonding MO above the
1u bonding MO, giving rise to a different energy pattern.We know that this is the correct pattern
from the following observations.
1.The lowest energy configuration of B2 is (Be2)(1u2), not (Be2)(3g2),and this gives a
paramagnetic ground state (a triplet spin state).((Be2) is short-hand for the configuration of all the
lower energy electrons,namely (1g2 1u2 2g2 2u2)).
2.The lowest energy configuration of C2 is (Be2)(1u4), and gives a closed shell (singlet spin)
ground state, a diamagnetic state.
3.The lowest energy configuration of N2+ is (Be2)(1u4 3g1) , and this gives a doublet spin state
without orbital angular momentum. If it were the pattern in figure 9 the state would have orbital
angular momentum, and this would be detected by spectroscopy.
The most important results that emerge from figures 9 and 10 are that O2 has an open shell
configuration, (Be2)(3g2 1u4 1g2) giving a paramagnetic ground state ( like B2); that removing an
electron from N2 weakens the bond, because this electron comes from a bonding MO, but removing
an electron from O2 strengthens the bond, because it is coming from an antibonding MO.
Although we will not deal in detail with the MOs of CH4,we note that the theory gives rise to two
bonding MOs for this molecule. The one of lowest energy has equal weight on each of the four H
atoms (we say that it is totally symmetric),and the next in energy is a set of three equivalent MOs
(degeneracy three),all of these being occupied by electrons. Thus the PE spectrum, referred to
earlier, consists of two bands.
The pattern of M O energies for the first row of diatomic molecules can roughly be determined from
the atomic orbital energies and the symmetry of the molecule (but needs the confirmation of
experiment). For other molecules of high symmetry one can reach similar conclusions, and this
approach has been particularly fruitful for the high symmetry octahedral and tetrahedral complexes
of transition metals, and for cyclic conjugated hydrocarbons. We deal here with the latter because
theory has an important role in understanding the stability and chemical reactivity of this important
class of molecules.

12
Molecular Orbitals of Conjugated Hydrocarbons
For planar conjugated hydrocarbons, such as ethene and benzene, the MOs can be classified into
two sets, those that are symmetric to reflection in the plane containing the nuclei, and those that are
antisymmetric. There is good evidence from experiment and calculation that the latter are
responsible for most of the interesting chemistry of these molecules; we call them orbitals, by
virtue of their shapes, particularly their nodal planes.
If we use the LCAO expansion to derive the wave functions of the orbitals then the only atomic
orbitals which have symmetry are the 2p orbitals of the carbon atoms which are pointing in a
direction perpendicular to the plane of the molecule (have a nodal plane in the plane of the nuclei).
In this model, therefore, the molecular orbitals are built up by a linear combination of the set of 2p
atomic orbitals, one for each carbon atom in the conjugated system.
The simplest model for calculating wave functions and energies of these molecular orbitals was
developed by Huckel in the 1930s, and this theory has become an important part of chemical
language. The mathematical procedure for deriving Huckel orbitals is not difficult, but is
inappropriate to give at this point. The important point to stress is that the orbitals are delocalized
over all the conjugated carbon atoms, and that they have smaller energies than the other molecular
orbitals in the molecule, which we call the orbitals. The use of the word smaller is not very
explicit; the point is that the bonding orbitals are less strongly bonding, and the antibonding
orbitals are less strongly antibonding than the orbitals.
The orbitals of the cyclic conjugate hydrocarbons have a particularly simple pattern. Their
energies are such that the lowest orbital is one which has equal LCAO coefficients for each carbon
atom in the ring, and the higher orbitals appear in pairs of equal energy (we say they are doubly
degenerate) until one gets to the highest energy orbital. For rings with an even number of carbon
atoms the last orbital is non- degenerate. For rings with an odd number of carbon atoms the last
orbital is one of a degenerate pair.
Huckel noted that to obtain closed shells of electrons with this pattern (all orbitals or degenerate
pairs of orbitals are either completely filled or completely empty of electrons) it is necessary to
have 4n+2 electrons where n is an integer. This is known as Huckels rule for aromaticity. Thus, for
neutral molecules in this series the six member ring (benzene) is stable, and the next stable planar
conjugated cyclic hydrocarbon would be one with 10 carbon atoms. However, we can achieve other
six electronic systems by adding an electron to the five membered ring, or subtracting one from the
seven member ring; both of these ions are well known to be stable.
The systems having an odd number of electrons, are called free radicals; they are chemically
reactive because they can form more stable systems by forming new electron pair bonds; but
molecules like this can be seen as transient species. Planar molecules having 4n rather than 4n+2
electrons are unstable but they can change their shape to form more stable closed-shell systems.For
example, the 8-membered ring distorts to a non-planar boat-like structure called cyclo-octatetraene
which effectively has four localized double bonds, and the 4-membered ring (cyclobutadiene)
distorts to a rectangular structure, although this is still chemically very reactive.
Figures associated with points made in this article will help understanding and they can be found in
most books dealing with the chemical bond. The book The chemical bond by Murrell , Tedder and
Kettle has them all.

John Murrell

13

You might also like