The Chemical Bond Revisited

N. Colin Baird
University of Western Ontario, London, Ontarlo, Canada NBA 587

A chemical bond is a concept for which ab initio quantum sion. We shall concentrate our attention upon the hydrogen
mechanics provides a necessary and sufficient theory in that atom, which, since i t contains only one electron, has no
(for small molecules a t least) it predicts properties such as potential energy contributions from electron-electron repul-
bond lengths, dipole moments, vibration frequencies, and sions.
intensities with hieh accuracv and dissociation enereies with
a lower (but usefur) accurac{. I t would be very helpkd if one
. &

of attraction between two charged

-
Recall from elementarv ~ h v s i e that
s the ~ o t e n t i aenerw
l
is proportioc&
could "interpret". "exdain". "understand". or "analvze" to the product of their charees and is inverselv-~.r o ~ o r t i o n a l
A

the subtle, cknplicated a b initio theory in simpler t&ms, t o the; separation r. In atomic units of charge, distance, and
i.e., in terms of forces, energies, simple wavefunctions, elec- energy, the potential energy P of attraction at any instant
tron densities, etc. This is a difficult undertaking. between an electron and a nucleus of atomic number Z is
Some textbook discussions of the nature of the chemical given by
bond are oversimplified to such an extent that many of the
statements they make in fact are incorrect.' This deficiency
is understandable, given that correct explanations of cova-
lent bond formation are rare even in quantum chemistry
texts! There are two fundamental difficulties associated where r is the electron-nucleus distance.
The potential energy Pa, hetween the electron and nucleus
~~~~~~ ~ ~~~

with the rectification of this state of affairs. First, there is no

one unique, "correct" way of discussing the chemical bond- is therefore given by -Z times the average value of l l r , i.e.,
for example, one can analyze either forces or energies, wave- the average of the inverse separation between them
functions or electron densities. etc. with eaual validitv but a
difference in point of view as to which are the more funda-
mental quantities. Second, the actual situation is rather
more subtle and complicated than one might have hoped For convenience, we redefine the average of the inverse dis-
and certainly more so than is apparent from the simple but tance as X ,
often misleading explanations in most texts. One complete
and interesting analysis of the chemical bond has been pro-
vided by Ruedenberg and co- worker^,^ particularly for the
fundamental systems Hz+ and Hz. The pivotal role of and thus P, = -Zx. Clearly the potential energy of attrac-
changes in the kinetic energy of the electron(s) is empha- tion improves (i.e., becomes more negative) as the average
sized in this analysis. In the discussion that follows, I have electron-to-nucleus distance decreases.
attempted t o provide a simplified and accessible, but still We now turnour attention t o the kineticenergy associated
quantum-mechanically correct, version of Ruedenberg's with the electron's motion about the nucleus. Although we
theory. know nothing about the trajectory of the electron's move-
As we shall see, the factors whose interplay determine the ments, we do possess information concerning the probability
energy and size of free atoms are also those which operate that the electron is to be found a t anv . ~. o i nin
t soace. From
(though in modified form) in molecules and which are re- such information, we can infer theeffective size of the sphere
sponsible for covalent bond formation. Thus, t o understand to which the electron is restricted most of the time (for a 1s
Ruedenberg's theory of the chemical bond, we must first orbital). I t is well known for aparticle in a (three-dimension-
learn something about the energetics of simple atoms. al) box that the kinetic energy of motion is inversely propor-
tional to the square of the box's dimensions; this result
Free Atoms applies also to a sphere. Thus it is not unexpected to find
Accordine to classical as well as wave mechanics.. there are ~ ~
that the kinetic energy of the electron in a hydrogen atom is
two significant contril~utionsto the internal energy of a free proportional to the square of the average inverse of the
atom: thr kinetic energy of the electrons due to their motim electron-nucleus separation. In atomic units of energy i t is
about the nuclrus and the potential enrrgy of the electrons found to be:
dueto their attraction to the nucieusand their mutual repul-

Clearly, the kinetic average is always positive and thus "op-

' An example is the implication that the potential energy decreases poses" electron-nucleus binding; the kinetic energy becomes
substantially in the molecule because the attractive potential is more worse (i.e., more positive) as the average electron-to-nucleus
negative at the midpoint between the nuclei than it is close to either distance decreases, since the electron is thereby confined t o
nucleus, and the transfer of density to this region is favorable for that a smaller and smaller volume.
reason. In fact,as a simple summation of -l/r, and -1la indicates, The notential enerw and kinetic enerw -,of the electron in
the potential close to either nucleus is superior to that at midbond. a hydrogen atom are plotted against the average electron-
Ruedenberg, K. Rev. Mod. Phys. 1962,34, 326. Feinberg, M. J.; nucleu~seoaration.~ lartuallv Il/r.,)-lI . in Fieure 1. The total
Ruedenberg, K. J. Chern. Phys. 1971, 54, 1495. Feinberg, M. J.;
Ruedenberg, K.; Mehier, E. L. Adv. Ouanturn Chem. 1970, 5, 27. energy of the hydrogen atom is the sum of these two terms,
and is also shown for the region of interest (Fig. 1). The best
- seoaration
The inverse of ll/rl.. is rouahlv related to the averaae .
althougn not ioent~c.sl10%. For tieburposes of this qua ilative discus- (lowest) energy occurs for a value of one atomic unit.
sion. I sna I referto(lIr).,-' as the 'average raol~s",a though it is not The same result can be obtained by differentiating the
In fact equal to <r>. expression for the total energy E

660 Journal of Chemical Education

 Figure 2. Probability of finding electron in atomic hydrogen at varlous db-
tames rfrom the nucleus.

i.e., it is not restricted to any particular finite volume. How-

Figure 1. Kinetic (0,potential (0.and tdal energy versus average electron- ever, the total probability of finding the electron far from its
nucleus separation in atomic hydrogen. \ nucleus is very small, and for our purposes i t can be ignored.
The familiar plot for the probability of finding the electron
a t any particular distance r from the nucleus is shown in
Figure 2. The function peaks a t r = 1atomic unit, which also
E,=K,+P, is the "average radius" obtained above from energy consid-
erations. Indeed the actual potentialenergy is identical to an
with respect to x and setting the result equal to zero, atom in which the electron is restricted to travel on the
surface of a sphere of radius 1.0 atomic unit. The actual
kineticenergyis identical t o that of an electronin a spherical
box of constant ~otential,and with a wall a t a radius of 1.57
Thus, a t the best separation, atomic units. ~ h u s both, the kinetic and potential energy
functions are consistent with a hydrogen atom in which the
dK,
-=- -@,
electron travels a t a distance of about 1atomic unit from the
dx dx nucleus, i.e., on the surface of a sphere of radius F. 1 a.u.,
Now K., = 0.5x2 and Pa, = - Z X Thus, venturine
- - ~ ~ outward and inward.
~

This model for electron binding can now be extended to

bond formation. Clearlv. - . the kinetic enerm of the electron in
a hydrogen atom could be improved, i.e.,reduced in positive
value. bv increasine the effective radius of the sphere in
and
which i t i s enclosed,or alternatively by extending the sphere
in some one direction. Such an extention does not occur in
dx atomic hydrogen because it would place, for a substantial
fraction of the time, the electron in a region in which the
By setting the derivatives equal and opposite, we obtain potential is poor. T o he more precise, let us for convenience
X =z consider a cross section of the above spherical surface for the
hydrogen atom; we choose a plane passing through the nu-
at the optimum. For Z = 1(i.e., hydrogen), x = 1,and thus cleus. This cross section is a circle (of radius 1.0 a.u.); in a
sense we have returned to the Bohr model for the atom, but
this causes no real difficulty for the 1s state if we are con-
cerned solely with potential energy.
The "avera e radius" of one atomic unit (i.e., one Bohr
1
radius, 0.529 )represents the optimum compromise for the
total enerev. If the size of the atom were to be further re-
duced, theidestabi~izin~) increase in kinetic energy would
more than offset anv advantage to the potential energy. T o
restate thisconclusibn, the hyhrogen atom is prevented from
contracting to a smaller size (even though this would im-
prove the potential energy component of the total energy)
due to the large increase in kinetic energy that would result
from the increase in the electron's velocity that accompanies Now consider the potential energy of an electron located a t
any reduction in the volume of space accessible to it. some point along the hcrirontal line that touches tangential-
I t is well known that the electron in a hydrogen atom has a ly the circle (see diagram abo5.e). The variation in instants-
finite probability that i t will be found a t any point in space, nrous potential energy P with position d along this line (d =

Volume 63 Number 8 August 1986 661

 Since the potential experienced by the electron is relative-
ly constant between (and ahove or below) the nuclei in Hz+,
the probability distribution G2 for the electron also is rela-
tively smooth in this region in the groundstate of thesystem.
J.2is larger in the midbond region than is the average of the
probabilities for the two atomic I s orbitals, 41and b,, one
centered on each atom. Indeed, $ is found to he well repre-
sented hv the normalized sum of d.. l ...
d. +
. with the conm-~~~.
~ ~~ ~

quence that d 'runrains the crosnterm Zwlw,, which is large in

thr midhond rrgion. i.e.. the volume in which thr sum o? +
4r2is smaller than near either nucleus
* = N(4, + 4,)

Thus,

Figure 3. Potential energy (0versus distance d along a line 1 a,", above

nucleus in Hand H Z . Internuclear separation is 2 a,". The kinetic energy function in Hzf is related to the value
of (Ilr)., associated with the atomic orbital function +I (or
0 represents the point of intersection with the circle) is 4,) by the same inverse square dependence as in atomic
shown as the lower curve in Figure 3. Clearly P peaks (in a hydrogen, hut with a proportionality consrant which is only
negative sense) a t d = 0. I t becomes substantially smaller in a t m t 516 as large. The reduction of 116 orcurs hecause J
magnitude whenever the electron is located away from this extends well bevond the t.ffcctive limit of m, alone" the hnnd
~~-~~~ ~

point in either direction to any appreciable extent. direction; as a result, the component of kinetic energy in this
direction is only about half that in the isolated atom. (A
Molecule Formatlon further increase in the length of the "box" in the direction
In order to formulate a model for the chemical bond in along the internuclear axis makes little difference to the
Hz+, we bring a second nucleus close to the hydrogen atom; kinetic energy as the box is long in that direction anyway.)
in particular we place it 2.0 atomic units from the first Thus, for Hz+,
nucleus (since this corresponds to the internuclear separa-
tion in HZ+).Let us now reconsider the potential energy of
the electron along the horizontal line which lies 1.0 a.u.
above the line joining the nuclei.

The total potential energy for Hz+ includes two terms: the
attraction of the electron to the two nuclei and the mutual
repulsion of the nuclei. The attraction term is approximately
equal to that of -(l/r), of the hydrogen atom plus about
-1lR for the attraction of the electron to the second nucleus.
The repulsion between nuclei is exactly 11R. Thus, for Hz+,
Due to the presence of the second nucleus, the potential felt
by the electron is almost 0.5 units more negative a t points
along this line than in atomic hydrogen; however, this ad-
vantage to the energy is cancelled by an approximately equal
but opposite term which is due to the nucleus-nucleus repul-
sion.
Murh more important is the fact that the potential en erg^.
function I S r s s r n r i o l l ~flat, ie.. P is almost constant. n,hen In other words, the potential energy function is essentially
the electron travels aiokg the line segment between the nu- identical to that in the isolated atom!
clei but 1a.u. ahove them-see the upper curve in Fiaure 3. If no change were to occur in the average electron-nucleus
Thus the space accessible to the eleciron has been greatly distance, then for HZ+we would have
increased4 in one direction in space-i.e., toward the second
nucleus, where "accessible space" is defined as the region in
which the potential is essentially identical with that found
for the 1 a.u. radius around the nucleus. Thus in 3-D. ~.~the
space accessible to the electron in Hz+, and in other diatomic
since
molecules, is not a sphere but a cylinder with rounded ends:

i.e., E = -0.583 a.u., an improvement of -0.083 a.u., or about

52 kcal mol-' compared to that of a hydrogen atom (plus
bare proton).
However, there is amechanism by which the energy in Hz+
can he improved even further. Let us discover whether 1.0
(Extension of the sphere in other directions is achieved in
polyatomic molecules.) This is the essence of chemical bond
formation-the extension in space of the electron t o vol- 'Note that there is no advantage to the kinetic (or potential)energy
in allowing the electron to jump back and forth between two infinitely
umes having advantageous potential energy beyond those separated hydrogen atoms; the accessible volume at any instant is
which are available to free atoms and which results in im- just that of the one hydrogen atom at which the electron is located, not
provement in the kinetic energy value. double this value.

662 Journal of Chemical Education

a.u. is the optimum radius for the electron in Hz+ as well as in
atomic H. In functional form, the energy in HZ+is given by
where
Em = K,, + ,
p
AP = P d - patom,
and
AK = K,, - '%to,,
+
Given that E = P K, i t also follows that the change AE in
total energy upon molecule formation
Differentiating and setting dE,ldx = 0, we find that to
optimize E, x = 615 = 1.20. With (llr), = 1.20, AE=AK+e
K , = +0.60 a.u. = -m

P
, = -1.20 a.u. =+el2

E, = -0.60 am. -
Since AE must be neeative if amolecule forms. it follows that
AK is positive and A P is twice as large and is negative. Thus
i.e., an improvement to E., over the hydrogen atom of -0.10 the phenomenon of contraction apparently occurs in all mol-
a.u., about 60 kcal mol-I. (The experimental value is 64 kcal ecules (possibly along with other changes which produce an
mol-' for Hz+.) improved potential enerw -.a t the cost of onlv one-half the
As a result of this re-optimization, the average electron- corresponding increase in kinetic energy).
nucleus distance in Hzf becomes substantially smaller than Finally, it is interesting to explore the changes to the
i t was in atomic hydrogen (0.83 versus 1.00 atomic units). energy components as a function of variations in internucle-
This contraction in size of the "atoms" in Hz+ allows a 20% ar separation R. The potential energy function never departs
improvement (i.e., -0.20 a.u.) to the potential energy since much from the form
the electron comes much closer on average to a nucleus. The
"penalty" in terms of the increase in kinetic energy amounts
to only 0.183 a.u. In contrast, a decrease in radius to this
value in the original H atom would have resulted in an
energy increase of 0.22 a x . I t is the relief of some of the
kinetic energy "pressure" to expand the radius which allows In actual fact. the electron-to-electron nucleus attraction is
the atom to contract, and the potential energy to improve, in not quite HS largeas 11R;the revidual positive 6 % repulsive)
HZ+.Similar conclusions appiy also to Hz and to other cova- contribution from thesecond and third terms taken together
lent bonds, although the analysis is more complicated in becomes more serious as R becomes small, i.e., when the
these multielectron~molecules. internuclear sevaratiou is reduced much below the 2.0 a.u.
In terms of the redistribution of electron density (proha-
value discussedahove.
hilitv) which accomvanies formation of the chemical hond.
thereis therefore a iharacteristic contraction of peak radius P = -(l/r),, + small number1R
toward the nuclei as well as the well-known increase in den-
sity in the midhond region. Well before this effect becomes serious as R decreases. how-
Parenthetically i t should be added that the confusion ap-
parent in some texts concerning explanations of the chemi-
cal bond are understandable, given
- the analysis ahove. For
limit R -
ever, the kinetic energy begins to rise rather quickly. 1n the
0. the region of constant ootential contours be-
tween the nuclei must disappear and the kinetic energy
example, if no allowance for atomic contraction is made, it function returns to the form
appears as if the kinetic energy decreases and that the po-
tential energy becomes less favorable upon hond formation,
and the bond is attributable to the decrease in kinetic ener- which is identical to that for the separated atoms! The real
gy. As we have seen, however, the bonding energy can be characteristic of a bond in a molecule is the existence of a
improved by about 20%by contraction. However, the accom- region between the nuclei of constant potential contours
panyingenergetic changes are so large that the net change in that allows the potential energy to improue substantially
kinetic energy is actually positive rather than negative, and by atomic contraction a t the expense of only a small in-
the potential energy is actually lowered. The latter conclu- crease in kinetic energy.
sions are completely general for molecule formation accord- Apparently the equilibrium distance Re (i.e., the hond
ing to the uirial theorem of quantum mechanics. According length) in a molecule is that separation of nuclei which is
to this principle, the potential energy of any molecule a t its small enough such that the potential in the internuclear
equilibrium geometry5 is equal to minus twice its kinetic region is relatively constant, but not so small that this con-
energy, stant-potential region is negligihly small.
The nature of the (very common) two-electron chemical
P
,
d = -ur,, hond differs in no substantial way from the one-electron
The theorem applies also to the isolated atoms before bond- linkage discussed above. For example, in Hz the (llr) coutri-
ing butions to the energyterms are doubled. The potential func-
tion also contains an electron-electron repulsion term; this
positive (destabilizing) contribution to the energy also varies
approximately as lIR, but is slightly smaller in magnitude
Thus the change in potential energy due to molecular forma- than the extra electron-nucleus attraction which in turn is
tion is minus twice the change in kinetic energy, smaller than the ever-present nuclear-nuclear repulsion.
Overall, we obtain
"he theorem applies only to optimized wavefunctions (at the
theoretical optimum geometries) and to the real systems; thus, the
results calculated for (llr), = 1 do not obey the theorem.

Volume 63 Number 6 August 1986 683

Thus there is no special significance to two-electron bonds in travel in the region of both atoms rather than in just one and
this theory; they simply are multiples of the basic one-elec- to correlate their motions so as to avoid being very close
tron bond and correspond to the maximum number of elec- together a t any instant-i.e., their "jumps" between atoms
trons which according to quantum mechanics can travel are synchronized!)
according to the same optimum motion "path" as described
above. (In particular, there is no special significance to the Acknowledgments
phenomenon of electron interchange between the two atom- The author is grateful to K. Ruedenberg and to the re-
ic orbitals, except as amechanism which allows them both to ferees for useful comments on the manuscript.

664 Journal of Chemical Education