ELEMENTS of the QUANTUM THEORY*

X. THE HYDROGEN MOLECULE. PART I. HEITLER-LONDON THEORY

SAUL DUSHMAN
Research Laboratory, The General Electric Co., Schenectady, New York

INTRODUCTORY REMARKS where

problem of determining the solution of the S. ha (1) -==

'van3
e~ra, fao )
f
THE equation for the helium atom is a limiting case of
the problem which is presented by the hydrogen
(402)

molecule. If in the case of the latter we permit the

two nuclei to approach until they coalesce, we obviously
Mb (2) = .
-

f
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

obtain the same system as that of the helium atom. and r,\, and ra„, refer to distances indicated in Figure 49.
But corresponding to the same energy, 2Eo, it is also
Consequently, we expect to find in the solution of the
S. equation for the hydrogen molecule certain similari- possible to have the eigenfunction obtained by inter-
ties with the results obtained in the previous chapter. changing the electrons, that is,
Downloaded via PEKING UNIV on September 23, 2022 at 04:19:41 (UTC).

However, there is one feature of the hydrogen mole- 02 =

«a(2)«b(1) (403)
cule which is of extreme importance, and which does
not occur in the case of the helium atom. This con- =
i-£_(r-+rni)/“" (4°4>
cerns the interpretation of the valence bond, which, on
the basis of the Lewis-Langmuir theory, is regarded as Consequently, the system is degenerate and the
due to the sharing of two electrons between the two zero-th order eigenfunction should be represented, as
hydrogen nuclei. The first successful application of in the case of the excited helium atom, by the two
quantum mechanics to the solution of the problem of linear combinations.
the hydrogen molecule was made by W. Heitler and
0«° =
d0i + bp 3
(405)
F. London,! and while subsequent investigators have
developed methods of attacking the problem by which
<t>$°
=
c0i + dpz (406)
more accurate results have been obtained, we shall find where a, b, c, and d are constants.
it advantageous to discuss the Heitler-London (HL) The eigenfunctions representing to first-order pertur-
theory in some detail before describing some of the bation terms, the perturbed state, that is, the state in
other methods. which the two atoms are interacting, will be given by
METHOD OF HEITLER AND LONDON pa —
pa2 T pa (407)
and
We consider a system consisting of two nuclei A and
pg =
pga + pg (408)
B, and two electrons 1 and
2, as indicated in Figure 49. / while the corresponding eigenvalues will be
In the unperturbed state, Ea 2A(j -j- zja (409)
—

where the two atoms are and

quite separate, it is obvious Eg =
2E0 + rig (410)
that the energy, E 2Eo,
=

where Eo is the energy of These will represent the eigenfunctions and eigen-
the hydrogen atom .in the values corresponding to the S. equation:
normal state, as given by
Vdp + V220 + « [E -
F„ + Ff1 + F2b +
equation (235) of Chapter ingFigure 49.—Illustrat-
Notation Used in For- F,b + F2A] 0 =
0 (411a)
VI. For electron 1 attached mulating Potential En- or
to nucleus A and electron 2 ergy Function for Inter-
action of Two Hydrogen Hp =
Ep (4116)
to nucleus B, the eigenfunc-
Atoms
tion is given by where the operator, H, is defined by the relation
0i »a(1)»b(2)
=
(401) H= {- 1/« (vd + V2a) + F„ -

(+ VzB +
*
This is the thirteenth of a series of articles presenting a more FiB + FSA) J (412)
detailed and extended treatment of the subject matter covered and
in Dr. Dushman’s contribution to the symposium on Modernizing « =
Stt'V/A"
the Course in General Chemistry conducted by the Division of
Chemical Education at the eighty-eighth meeting of the American
Fo
J+
=
Chemical Society, Cleveland, Ohio, September 12, 1934. The
-

K r 12

author reserves the right to publication in book form,

f Heitler and London, Z. Physik, 44, 455 (1927). F,A =
e1/rA1, Vz* =
eV^Bs, F,E =
ea/rBlp F2A =
e2/rAt (413)
287
288 Journal of Chemical Education
The suffixes on the Laplacian operators refer to elec- so that the orthonormalized functions are as follows:
trons 1 and 2.
1
Since and 4>/ must be orthonormalized functions, (0i + <h) (422)
\/2(1 + S2)
it follows that
1
(01
—

0-l) (423)
J' (tjta'ydvidvi =

J" (<bfi°)2dvidv2 =
(414) V2(l -

S2)

and By direct substitution of these relations in equations

(414) and (415) it is readily shown that the latter con-
= 0 (415) ditions are satisfied, so that equations (422) and (423)
J' p°dv\dv2
represent orthogonal and normalized functions.
where the integration is carried out over the whole con- From symmetry considerations it is also evident that
figuration space for each electron.
S (424)
By substituting in these last two equations from
= =

^J'ma(1)mb(1)(1i'i j'uA(2)ui,(2)dv2
equations (405) and (406), it follows that
Furthermore, these four eigenfunctions satisfy the
a2 + b2 + 2 abS2 = 1 (416) four S. equations:
c2 + +
d2 2 cdS2 =
1 (417) V2«a (1) + k(E„ + F,A)»A( 1) 01 =

ac + bd + (ad + 6c)52 = 0 (418) V2UB (2) + k(E0 + nB)«B (2) = 0 V (425)

V2»b (1) + k(E0 + F,b)»b (1) = 0 j
where
and
S1 =
0102<fol dv2 V2«A (2) + k(E„ + V^)ua (2) = 0
J’
Hence,
=

J* UA(l)ua(2)u\(2)uB(^)dtidv1 (419) V20i =

uA(l) V2 mb (2) + uB (2) v2wA( 1)
= -« (2E0 + vy + FsB)0i (426)
Now it is evident that for R “, that is, for the un- =
V202 =
-<c(2Eo + vy + Fib)02 (427)
perturbed state, S 0, since either fpi or % represents
=
and
the two separate atoms. On the other hand, for R =
V2{0i + 02) + k\2Eo (0i + 0s) + (FiA + K»B)0i
0, as mentioned already, the system becomes identical + (Vy + 1^)02( = 0 (428)
with that of the normal helium atom, for which
Now if <t>« is a solution of the S. equation (411a),
S2 =
dvi dvi = 1 then, by substituting from equations (407) and (422),
J |«(1)«(2)|- we obtain the equation,

Hence, in the case of the hydrogen molecule, 1 > S > 0, V 2(0i + <h)+ <2E„ + ya V„ + + V3* + -
ViB +
and since in the case of helium, the zero-th order eigen- (01 + 02) + y/2 + 2f>2 [vVa + k(2Eo + Tfa —
Vo + Vy
functions for the excited states are + F,B + ViB + Vy)4>a\ 0 =
(429)
From (429) and (428) it follows that
0S° =
Un(l)um(2) + Un(2)u„(l) \ ,

| V2 + 25s |V + *(2E0 + va
-

V0 + K,A + 72b + F,E +-

and FjA)0a] =
*1(7. —

Va) (01 + 0a) —

(VlA + V2B)02
-
(F,B + F2a)0i j (430)
M„(2)»„(l)
'

0A0 W„(l)«m(2) -

| |
,
This is an inhomogeneous partial differential equation
of the same type as that encountered in equation (287),
we assume a b. The validity of this assumption
=
Chapter VIII, in connection with the calculation of the
will be justified by the resulting deductions. first-order perturbation energy term. In order that it
From equation (416) it follows that may have a solution, it is necessary that the right-hand
side of equation (430) should be orthogonal with re-
1
b = a (420) spect to the solution, + <f>2, of the corresponding <f>
V2(l
1

+ S2)
homogeneous differential equation.
Also, from equation (418), it follows that Hence

b(c + d) (1 + S2) 0 = (V -

*0 (0i + 02)! dvidv? —

(V0 —
va)(2 + 2.S2)

Since
f,(FiA + (0t + 02) dvidvt +
=
F2B)02
b 9^ 0 and ^ 1, c ^ —
d.

Hence, by substituting in equation (417), we deduce + V, B)0! (0! + fr^dvidvi (431)

the result f(Vy
That is,
1 -1 En + Eti „ EnS2 -

En
c d (421) Va = —
=n —

(432)
V2(l V2(l
’ 11
1 + 52 +
1
-

S2)
-

S2) 1 S2
June, 1936 289

where (v.e.)t asfunctions of the internuclear distance R/ao,

where denotes the Bohr unit radius.
ao
2En =
1F»(0is + *.*)
—

{VP + VPW It will be observed that rja reaches a minimum value

J of —3.2 v.e. for R =
1.52, no
=
0.80 A, while vp is al-
-

{VP + 7]B) <t>P}dvidn (433)

ways positive. Thus, rja represents the lower energy
2En =
VP -
VP -

VP -

ViB)<t>itt>tdvidv2 state and must correspond to molecule formation, while

^(2V„- rjp, since it is positive, must correspond to repulsion be-
(434) tween the atoms.** Consequently, the symmetric
and S'1 is defined by equations (419) and (424). eigenfunction, <j>a, represents a stable state, while the
In a similar manner it may be deduced that the first- anti-symmetric eigenfunction, <ba, represents an un-
order perturbation energy term corresponding to 4>p is stable state.
given by the relation, Figure 51 shows plots of the same quantities in terms
of calories per mole Hs. For comparison there are also
En —
Eis EuA Eu ,,orl
ve
“
~i— S2
=
£a
r, ,

“I--1 _
-

(435J plotted the energy term, £n, and the curve calculated
by P. M. Morsel †† from observations on the band spec-
where £u and £12 have the significance defined in equa- trum of H2. The considerations upon which this cal-
tions (433) and (434), respectively.* culation has been made will be discussed in a subse-
INTERPRETATION
quent section.
PHYSICAL
Calculation shows that the term £12 is negative (over
The terms and rip represent, to a first-order approxi- a large range of values of R/ao) and greater in absolute

mation, two different values for the interaction energy of value than £n while .S'2 is positive and less than 1 for
two hydrogen atoms, and an inspection of equations all values of R/ao > 1. From the curves in Figure 51
(422) and (423) shows that )?„ corresponds to the sym- it is seen that for values of R/ao f 1.5, approximately,
metrical zero-th order eigenfunction <ffa, while ijj corre- the relations between ija, vp, £n> and £12 are those indi-
sponds to the antisymmetrical function 4>p. Thus the cated diagrammatically in Figure 52.
quantum mechanics treatment of the problem leads to
the conclusion that two hydrogen atoms can interact
in two different modes. The fundamental reason for this
is the fact that it is possible for the two electrons to in-
terchange places; or, stated in more technical language,
the existence of two modes of vibrations and corre-
sponding eigenvalues is due to the degeneracy of the
system in the unperturbed state. This, in turn, occurs
because the two electrons are absolutely equivalent, so
that it is impossible to distinguish between them. In
other words, the whole argument is a logical deduction
from the principle of indeterminism in the sense that
when two hydrogen atoms approach each other very
closely, there is no experimental method by which each
electron may be “tagged” and observed.
Evidently the interaction energy terms rja and vp are
functions of R. While the actual details of evaluating
the expressions for Eu and
£12 are discussed in a subse-
quent section, Figure 50 shows Figure 51.—Plots of the Total Energy, Coulomb
the results deduced by Y. Energy, and Exchange Energy as Functions of Inter-
nuclear Distance, for the Two Modes of Interac-
Sugiura,! which are more ac- tion of Two Hydrogen Atoms
curate than those obtained by
Heitler and London in their
A consideration of equations (433) and (434) throws
investigation. The values of considerable light on the nature of the attractive en-
rja (curve S) and rjp (curve
A) are given in electron-volts ergy forces which lead to molecule formation. In
equation (433), the term Va(4>i2 + 4>/1) represents physi-
Figure 50.—Energy cally the total repulsive energy due to the nuclear
*
The notation Eu and En is that Curves for the Two
used by Heitler and London. In a Modes of Interaction
charges and the electronic charge distributions. Simi-
subsequent section these will be shown of Two Hydrogen
to correspond to matrix elements Atoms; Curve S Cor- ‡ 1 v.e, 23,054 cal. per mole.
=

which occur in a second-degree secu- responds to the Sym- **

Thus, curve S is similar to the plots shown in Figure 46,
lar equation. metric, and Curve A Chapter VII-B, for the energy of interaction of two molecules.
† Sugiura, Z. Physik, 45, 4S4 to the Antisymmetric ††† Morse, Phys. Rev., 25, 57 (1929); also Condon and
(1927). Mode Morse, "Quantum mechanics," p. 163.
290 Journal of Chemical Education

larly, the other terms onthe right- —t '

On the basis of these considerations we are led to
hand side of (433) represent the identify the quantum mechanics exchange energy, £12,
attractive energy between the nu- ’Vs: _ with the energy of the non-polar or shared electron bond.
clei and the electronic charge dis- This is most readily evident from the plots, calculated
tributions. Hence, we may desig- r from (4>i)2 and (<t>p)2, of the density of charge distribution
.S—*-
nate En as the Coulomb interaction for the symmetric and antisymmetric cases, respec-
Figure 52.—Re-
energy. On the basis of classical lation between
considerations this should consti- the Terms in the
tute the whole of the interaction Expressions for
the Energy of In-
energy between two hydrogen teraction of Two
atoms. Actual calculation, as Hydrogen Atoms
shown in Figure 51, indicates
that the largest absolute value of En (calculated for
the value of R at which it is a minimum) is approxi-
mately 10,000 cal. per mole. Since the observed
Figure 53.—Electron Distribution
energy of formation of H2 from the atoms is 109,100
for
Elastic Reflection of Two Hydrogen
cal. per mole,* it is evident that most of the energy of Atoms
formation must be accounted for by the value of £12.
What is the physical significance of the expression tively. Figures 53 and 54, taken from a paper by F.
for En as given in (434) ? A similar difficulty was ex- London, ‡ show the results obtained.
perienced in the case of the exchange-energy term, Hn, The density is constant on each curve (so that they
which was derived in the solution of the helium atom
correspond to the isobars in the case of atmospheric
problem. [See Chapter IX, equations (347) and pressure measurements). The numbers attached to
(354). ] In the present instance also we designate En each curve give the relative densities or probabilities of
as the exchange-energy term (“Austauschenergie”), and occurrence of the electrons. Figure 53 illustrates the
the occurrence of such a term in the solution for t?u case of elastic reflection and shows that the electrons
(or vp) is a purely quantum mechanics phenomenon tend to occur during most of the time in the regions
for which there is no analog in classical experience, un- removed from the center. They avoid the region be-
less it be in the interaction of two linear harmonic os- tween the nuclei. On the other hand, Figure 54, which
cillators. This was discussed in Chapter VII, and it
was pointed out in that connection that whenever a
system is capable of vibrating in two different modes,
the actual motion will be obtained as a superposition of
these two modes. In the case of the ionized hydrogen
molecule it was shown that the application of the prin-
ciple of superposition of vibrations leads to the conclu-
sion that the system may exist in either a symmetric or
antisymmetric state. †
In the case of the hydrogen molecule a similar result
is obtained because of the possibility of interchange of
electrons between the nuclei, and if we calculate a fre-
quency vn =
En/h, we can regard the exchange-energy
term as a measure of the frequency of interchange of Figure 54.—Electron Distribu-
tion for Hydrogen Molecule
electrons. Formation
There is this important difference between the en-
ergy terms Hn and £12, that while the latter is negative represents the distribution for molecule formation
in the case of the hydrogen molecule (and in most other
(.homo polar combination), shows that in this case the
cases of interaction of similar atoms), the term Hn is electrons tend to occur during most of the time in the
positive, as is readily evident from equation (347) and region between the nuclei. Thus, the methods of quan-
the subsequent calculations in Chapter IX, Part I. tum mechanics lead to an interpretation of the shared
This is due to the fact that Hn represents a repulsive electron pair or non-polar bond of the Lewis-Langmuir
energy term which arises from the interchange of co- theory of valency. When the chemists write the elec-
ordinates by two electrons of different quantum num- tronic formula for II2 as H:H, they are therefore in
bers in the same atom. On the other hand, £12 repre-
logical agreement with the conclusion deduced from the
sents an attractive energy which occurs because of the solution of the S. equation. In other words, the theory
possibility of interchange of two equivalent electrons, of Heitler and London gives a quantitative basis for a
one from each of the atoms.
representation which the chemist had derived by in-
* This is usually given as 4.73 electron-volts. tuition.
† See discussion. Chapter VII—A, J. Chem. Educ., 12, 185
<1935). t London, Leipziger Vortrage, 1928, pp. 59-84,
June, 1936 291

THE APPLICATION OF THE PAULI EXCLUSION PRINCIPLE We shall find it advantageous to express all distances
As in the case of the helium atom, we must take into in terms of an, and designate these by p, with a corre-
consideration, in the present case, the electron spins. sponding subscript. Thus pai fAi/an, and so forth. =

In terms of these units,

Here also we find that it is possible to obtain only four
functions which are completely antisymmetrical in the ~

+ 0
S =
m>dv, (436&)
electrons and which, therefore, satisfy the Pauli exclu- -Je
sion principle. Since the spins may be either parallel
or antiparallel, we obtain, as discussed already in con- Now in evaluating integrals such as those for S, En,
nection with the helium atom, the four spin functions, En and others which occur in two-center problems, it is

a, (3, y, and 5, of which the first and last correspond to found convenient to utilize spheroidal coordinates of the
2ms =
+1 and 1, respectively, while for
—
and 7, particular type known as confocal elliptic. These are
2wj =
0. From these we derive three eigenfunctions defined as follows:
which are symmetrical in the electron spins, viz.:
4- obi)
X
g (pai
—

a, (/3 -j- y)l 2, and 5,

p =

g (pai —

pbi) (437)
and one which is antisymmetrical, viz., (/3 7) j\/2. —

The only combinations of these functions with the or-

where D R/an.\ As third coordinate, we use the
=
bital eigenfunctions pa and Pp, which can occur in ac-
cordance with Pauli’s principle are the following: angle 9, which a plane passing through the two nuclei
and the instantaneous position of the electron makes
<t>a (fi —

7)/ V2 with a fixed plane through the two nuclei. It will be

and recognized that A defines an ellipse described about the
two nuclei as focal points and passing through the
pga] pg(@ + t)/V2: <t>e5 point designating the position of electron I. Similarly
p defines a hyperbola which passes through the same
Thus we find that there exist three possible states
in which the atoms repel each other for one state in point. By revolving such a system of confocal ellipses
and hyperbolas about the line joining the nuclei as
which they attract and form a molecule. That is,
when two hydrogen atoms collide there is a 25% proba- axis, there is obtained a set of confocal ellipsoids and
bility that this collision will result in the formation hyperboloids, which (see Figure 56 in Appendix IV)
of a molecule. Furthermore, in the molecule, the
intersect along circles. These circles lie in planes per-
pendicular to the axis and have their centers on this
spins of the two electrons must be antiparallel. For axis. Hence, in order to specify the position of the
this reason the normal state of the molecule is desig-
nated spectroscopically as a singlet state (l2g), electron, we require not only X and p (which define a
whereas the repulsive state is of the triplet type, particular circle) but also 6, the angle which a line
(32«).
passing through the center and the point makes with a
fixed axis.
The suffixes g and u signify respectively "gerade,”
As shown in Appendix IV, the element of volume is
that is, even, and “ungerade,” that is, odd. These
given by
designations refer to the fact that
D3
4>a '
4-pa (—X, —y, —2) dr = “

(X2
—

p2) d\dpdg (438)

while
^(*i7ia) =
~<t>p(-x, -y, s)
Hence, in terms of these coordinate variables,
Although the energy state corresponding to Pp is ds r
unstable, “it is used to explain a certain continuous S =
I c
^
(X2
—

spectrum emitted by H3, due to transitions from an

excited triplet state, not shown in Figure 50, down to = —
I £-DX (X3
-

p2)d\dp
the 32„ curve. Transitions down to this curve would
obviously lead to a dissociation of the molecule, giving
rise to a continuous spectrum.”* since the limits of9 are 0 and 2ir, and we can therefore

integrate directly with respect to this variable.

CALCULATION OF PERTURBATION ENERGY TERMS The limits for the other variables are defined by
Let us consider first the integral 1< X <°°; —1 < p <1. Expressing the integral as
the product of one integral with respect to X and of the
S =
(424)
other with respect to p,
^j(a(1)“b(1)<Ai
S = 1 \Hp -

-
—3 fe-(rAl + rBl>/«>**
(436a) ~Dxd\j J\pHp\
*
J. H. Van Aleck and A. Sherman, “The quantum theory of † The significance of X and p. in these equations is, of course,
valence,” Rev. Modern Physics, 7, 168 (1935). This is also dis- not to be confused with the interpretations used in other connec-
cussed in Part 3 of this chapter. tions.
292 Journal of Chemical Education

=
*-ox(2x2 -
2/3)dX Hence the interaction energy of the charge distribu-
tion about the point B with a similar distribution about
1 F"
I
D2 f" the point A is given by
= tt e_xx2<ix--- I c_I dx
2jD 6 JD

where x D\. The expressions for these integrals

—
Ji =

do
-

J f —

pb jl
(
-
«-*«. d + pb) f dv,.
)

have been given previously in Chapter IX. Conse-

quently, we obtain the result,
= —

do
' -

V J("
——

PB jl
(
—

(1 -J- pn) [ dv,.

)

5 = t-v { 1 + D + (1/3) D2 ) (439) where pA =

pA1.

Figure 55 shows that the value of 52 varies from 1 Introducing the confocal elliptic coordinates defined
for p = 0 to 0 for p =>. = by equation (437), and integrating with respect to 0,
the last equation assumes the form,
D2 /* €—z>(x + y)
Ji el (X2
—

p2) dXdp.
(?0 ~2j (X
-

,u)

1 -
-D(X -

p)
0+¥-S)|
?! t-oxdX e
—

Up (X 4- p)dp —

do

\D
e—2DX dx (X -)- p) 4-
J* J' ^1 2 )dp]
= ?!i L e *
c“ dx + ( Z
dx
(fD fnD
1

do \lD xdxf-v dxf-

1
/*«.
i /•« D2 C
~

A I „
€~" xdx ~

1ft I 6-3
.4 .3 1.2 t.G 20 "24 2-3 32 3.6 4 J2£> 16 J2D n
12 J2D
Figure 55.—Plot of the Function 52 as a Function
of Internuclear Distance = ?!
a„ |_»
[L _
4-2D (I
\r>
u
+ 8 + sd+j*x]
6 Jj
†
4
† (441)

We shall now consider the relation for 2£u as given In the case of the four negative terms on the right-
in equation (433). The first term on the right-hand hand side of equation (433), which involve respectively
side is the four reciprocal distances pai-1, Pb2~1, Pas"1, and
Pbi"1, it follows from the equivalence of the electrons
-
Fo(012 + <t>%i)dvidvii = that each of these terms represents the attractive energy
J'J* due to interaction of an electron charge distribution
'//(j R + ru 'j (0i2 4-
ris/
022) dt>i dv% about one of the nuclei, say A, with the positive charge
at the other nucleus, B. Denoting each of these four
and since the two electrons are equivalent it follows that terms by Jo, it follows from equation (440) and the re-
we can write lation R aoD, that
=

J-!’ + 2A 4/*“
SS1 +o)( (442)

where Consequently
A
J* |«a(1)«b(2) | (l/n»)
= «2 dnd®j. Bn -
4- J, -

2 Jo (443)

This integral represents the repulsive energy between ?! e 5D 3D2 D3 I

+ (444)
the two electronic charge distributions represented by a0' D )1 8 4 6 t
ua2 (1) and «r2 (2), that is, by the two functions From the plot in Figure 51 of En against D it will be
observed that for D > 1.4 approximately, the expression
^ J* e-2^
uA2 (1) = dvi
has a negative value. This signifies that for distances
and beyond D 1.4, the Coulomb attractive energy between
=

dvi the nuclei and the negative charge distributions ex-

»B2(2) =

;Jf_2PE! ceeds the total repulsive energy which exists between

Now according to equation (3355), Chapter IX, the the two nuclei and also between the two electronic
potential at the point p pai Pa, due to the electron— = charge distributions. The Coulomb forces of attrac-
distribution about the point B is given by tion and of repulsion are equal at the distance D 1.9, —

approximately, for which OE,: /dD 0. =

F(p») =
™jl PB (
-
€-=<>= (1 + pb)|
)
(440) We now have to evaluate the different integrals
which occur in equation (434) for the energy En. Be-
where Pb represents pbi- cause of the equivalence of the electrons, it follows that
June, 1936 293

the four negative terms must each be equal to the same lower than that obtained by HL, but still above the
integral. Let us consider the integral observed value, —4.73 v.e. ‡
'ua(1)ub(1)<1zii
= e~

^Uk(2)un(2)dvi j*
= SK« APPENDIX IV
where TRANSFORMATION FROM CARTESIAN TO CONFOCAL ELLIP-
e2 i*{-<pai + pbi) . TIC COORDINATES
I--dvi
„

a2 - —

(IqTT J
, PA 1

In Chapter V-A the relations were developed for the

Using the transformation to confocal elliptic coordi- transformation from rectangular or Cartesian to spheri-
nates defined by equations (437) and (438), it follows cal polar coordinates. Also the expression was given
that for the Laplacian operator in terms of these coordi-
nates.
K* = -

«i
‘

W
Ui
I
J -i
(X
-

p)dy. As stated in Chapter X, Part I, it is convenient in

the caseof problems involving two-center systems to
and therefore,
choose confocal elliptic coordinates. The consideration
5ifs = -

(1 + D) of Figure 56 will illustrate the physical significance of

this coordinate system.
=

?o-'““O + 2° + T"
+t) (445)

The first integral on the right-hand side of equation

(434) is

where
/e2 Vo (t>]<f>2dvidv2 =

^
•
S2 + K-,

e2 f
Ki —
—

I &l4s(l/pia) dvidi>2 (446)

do J

Consequently we obtain the result

En - f2 .

¶+ -
2SK, (447)
Figure 56.—Confocal Elliptic Coordinates
Heitler and London did not evaluate the integral in
If 2a designates the distance between two points
equation (446), but concluded that
regarded as foci of a system of ellipses and hyperbolas,
5& then the major axis of any one ellipse is given by
8
CD = 2 a/e\
However, Y. Sugiura* showed, as a result of a
where ex (^ 1) is the eccentricity of the given ellipse.
lengthy calculation, that the integral could be repre-
sented by the relation, Similarly, for any confocal hyperbola, the major
axis is given by
C'D' =
2 a/e,

where e2 1) is the eccentricity of the hyperbola.

^ j
S2 (C + In D) + 5,2 Ei (- 4D) -
2SStEi (- 2D)
j j
(448) For any point, P', at which these two curves inter-
sect,
where AP' + BP' rA + rB
=
2a/et 2«X =
(i)=

C = Euler's constant = 0.5772 and

S1 = e »
(1
—
D + Z>2/3) AP’ BP’
—
rA
—
rB 2 a/es
—

2 ay. —
(it)=

In —
natural logarithm, where X = \/e\ and n l/e2.
=

i* —
X g —W Thus, each ellipse of the system of confocal conics
Ei(x) =
Integral logarithm! =
j du
has a definite value of X, and similarly, each confocal
—

J co u

i XI
hyperbola has a definite value of /t, so that a point, P,
Ei(-x) = C + In + -
* + J -

J g +.....(x £17) may be designated by specifying the values of X and m

for the two confocal conics which intersect there.
As result of his calculation, Sugiura derived a mini-
a If now we rotate these curves about AB as axis of
mum value for the interaction energy, 7/„, of —3.2 v.e. revolution, there are obtained a series of confocal coni-
for D 1.4 (R = 0.80 A.), which is about 0.8 v.e.
=
‡ The energy of formation of H» from the atoms is, of course,
*
+4.73 v.e. This value is greater than the observed energy of
Sugiura, Z. Physik, 45, 484(1927). dissociation of H2 because the latter does not include the “zero
† Values of this function are given in “Funktionentafeln,” by point” energy of vibration of the nuclei, as will be discussed more
E. Jahnke and F. Emde, B. G. Teubner, Berlin, 1928, pp. 19-22. fully in a subsequent section.
 294 Journal of Chemical Education
coids which in this case are known as prolate spheroids.
The ellipsoids and hyperboloids of revolution will inter-
sect along circles such as that indicated in Figure 565, (xii)
and hence it is necessary to introduce a third coordinate
variable, the angle 9, which specifies the position of P
on the circumference of the circle, with respect to a bx by
fixed plane (indicated by the line P'NP') which passes be
=

be
= a ~

^ -

^ 'cos (

(xiii)
through the axis AB. bz
It is necessary now to derive the relations between bfl
= —a V(X2 -
1) (1 -

p2) !

the variables, X,/i,0 and the rectangular variables

x,y,z. Assuming that the .r-axis coincides with the Hence,
axis of revolution, AB, that the z-axis is in the plane of a2 (X2 -

p2)
a\ (xiv)
the figure and passes through the origin, O, in the mid (X2
-

1)
point of AB, while the y-axis projects at right angles to „
a2 (X2 -
p2)
the plane containing Oz and Ox, we can set down the
_

(1
(xv)
M*)
-

following relations which are derived in any textbook

on solid geometry. ae =
a2 (X2 -

1) (1 -

p2) (xvi)
x1 y2 + z2 and
+ a2 (tit) dxdydz a2 (X2 (xvii)
x2 (1 1/x2)' p2) dXdpdO
— —

X2 -

*2 y2 + z2
The same result follows from the relation in terms of
~ '

(iv)
p2 p2 (1
-
1 /p2) the Jacobian, which is of the form,
At the point P the same values of x,y,z must satisfy bx by fe
both of these equations. Hence, bX bX bx
r2 =
y2 + z2 = a2 (X2 -

1) (1 -

p2)
dxdydz = by by bz dXdpdS
This defines the radius, r( NP), of the circle along = bp bp bp
which the two surfaces intersect, and consequently, as bx dy bz
is evident from Figure 565, b0 bp bd

y = a y/ (X2 —

1) (1
—

p2)- sin 6 (v) As may be verified by direct substitution from equa-

y/ p2)- cos e
tions (xi), (xii), and (xiii), the value of the determinant
z —
a (X2
—

1) (1
—

(pi)
is found to be a3 (X2 p2).
—

By substituting in either equation (iii) or (iv), it Applying the rules stated in Chapter V-A for ex-
is readily shown that pressing the Laplacian operator in terms of the vari-
x =
aXp (vii) ables X, p, and 9, it is readily deduced that

As in Chapter V-A, we now have to determine the l

v2 =
(1 +
three constants ax, a„, and ae, which connect the ele- a2 (X2 -
p2) M)b^
ments of volume as expressed in the two systems of co- (X2 -

p2) b2
(xviii)
ordinates, according to the relation, (x2 -

i) (i -
p2)' be2

dxdydz =
y/ a\apao-dXdpde .

When the S. equation for a two-center system is ex-

Since X, p, and 9 form an orthogonal system of co- pressed in terms of these variables it is usually much
more convenient to separate the equation into three
ordinates, it follows that
ordinary differential equations, and thus obtain a solu-
»-
(£)
-

ffi+m+(?j « tion, than when other systems of variables are used.*

There are other orthogonal systems of coordinate
variables which have been found convenient in solving
many-body problems, and a more complete discussion
of such systems may be found in the work by E.
-(S),-(sr+&copy;) +&copy; »
Madelung, † and also in that by L. Pauling and E. B.
Wilson, Jr.*
where ds2 (dx)3 + (dy)3 + (dz)3.
=
*
See discussion by H. Bethe in “Handbuch der Physik,”
These differential coefficients, as derived from equa-
XXIV, Part 1, Julius Springer, Berlin, 1933, p. 530 el seq. A
tions (v), (vi), and (vii), are as follows. further discussion of the usefulness of elliptic coordinates in the
treatment of the two-center problem will be found in the paper
bx by . by W. G. Barber and H. R. Hasse, Proc. Cambridge Phil.
= — = Xa s,n 6 Soc., 31, 564 (1935).
(xi) † E. Madelung, "Die mathematischen Hilfsmittel des
Physikers,” Julius Springer, Berlin, 1925, pp. 82-91.
bz t Pauling and Wilson, loc. cit., p. 103, and Appendix IV,
Xa cos e
bX p. 443.