JOURNAL O F T H E

AMERICAN CHEMICAL SOCIETY

(Registered in U. S. Patent 0 5 c e ) (0Copyright, 1961, by the American Chemical Society)
VOLUME 83 SEPTEMBER 18,1961 NUMBER
17

PHYSICAL AND INORGANIC CHEMISTRY

[CONTRIBUTION FROM T H E DEPARTMENT
OF CHEMISTRY, UNIVERSITY OF WISCONSIN, M A D I S O N , WISCONSIN]

Electronegativity
BY RAYMOND AND JOHN L. MARGRAVE
P. ICZKOWSKI
RECEIVED
NOVEMBER
25, 1960

The energies of positive and negative ions relative to the neutral atoms are conveniently and accurately expressed for a
given atom by a power series in N = n - 2, where n = the number of electrons around the nucleus in a given ionization state
and 2 = the atomic number of the nucleus. For a neutral atom the electronegativity is defined as x = ( - dE/dN)N,
where d E is the energy change which accompanies the change in charge d N and should be expressed in the units energy per
electron. Similarly the value of ( - dE/dN)N- -1 represents the electronegativity of the singly charged positive ion.
The E ( N ) curve exhibits a discontinuity in slope a t N values where there is a transition from one type of atomic orbital t o
another. If only the first ionization potential and the first electron affinity are known for a given species, x = ( - d E /
dN)N, o is equivalent to the well-known Mulliken relationship that electronegativity is equal to the average of the ioniza-
tion potential and the electron affinity.

Introduction due to the partially screened nuclear charge a t the

The qualitative aspects of electronegativity are covalent radius. Allred and Rochow4 found a
readily comprehensible. In a diatomic molecule, good correlation of electronegativity with Z*/rcov2
the atom which has a net negative charge compared and indicated that electronegativity represented
to the other atom is said to be more electronegative. a force on the atomic electrons a t the covalent
The specification of the electronegativity of an radius. Other correlations3 have been made with
atom (in a given valence state) on this qualitative still other powers of Y . Mulliken’s5values represent
basis would amount to a specification of its position an average of the binding energy of the outer
in an ordinal listing of the elements, such that i t electrons in an atom and its corresponding negative
would be negative in a diatomic molecule formed ion. Mulliken’s view introduces the idea that
with any element listed below it. Similarly the electronegativity represents an average of a property
criterion for setting up such a list of the atoms in the over a range of ionization instead of being solely a
order of their electronegativity would be the property of the neutral atom. Other approaches
polarity, but not the amount of the partial ionic relate electronegativity to a variety of other pro-
character of diatomic molecules, of which the nu- perties .6 -*
clear quadrupole coupling constant is an imperfect After so many approaches, there is some con-
measure. fusion as to what physical picture corresponds to
The quantitative aspects of electronegativity are the term electronegativity. This confusion is
less clear. Quantitatively, electronegativity is also manifest in the question of what should be the
defined as the tendency of an atom in a molecule to units of electronegativity. I t is difficult to under-
attract electrons.‘ G ~ r d yfound~ , ~ a correlation of stand the meaning of a quantity if one does not
electronegativity with Z*/rcovr where Z* is the know in what units i t is to be expressed. As shown
effective nuclear charge acting on the outermost in Table I, the electronegativities obtained by dif-
electrons, and rcOvis the covalent radius of the ferent investigators have different units. Clearly,
atom. Gordy took this correlation as a demon- the absolute numerical values of quantities having
stra tion that the physical quantity corresponding different units are not comparable, because they
to the term electronegativity was the potential (4) A. L. Allred and E. G. Rochow, J. Inorg. Nucl. Chem., 6, 264
(1) L. Pauling, ”The Nature of the Chemical Bond,” Cornel1 Univ. (1958).
Press, Ithaca, N. Y.,1960, p. 88. (5) R. S.Mulliken, J . Chem. Phys., 2 , 782 (1934).
(2) W. Gordy, Phys. Reo., 69, 604 (1946). (6) A. D.Walsh. Proc. Roy. SOC.(London), 8307, 13 (1951).
(3) Cf. H. 0. Pritchard and H. A. Skinner, Chcm. Rcos., 6 6 , 745 (7) W. J. Gordy, J . Chcm. Phys., 19, 792 (1951).
(19,j.j). (8) R. T.Sanderson, ibid., 23, 2467 (1955).
 3548 RAYMOND AND JOHN L. MARGRAVE
P. ICZKOWSKI Vol. 83

2, taking the energy of the neutral atom to be

zero. The abscissa, N , represents the number of
electrons present around the nucleus minus the
atomic number, Le., N = n - 2, where n is the
number of electrons around the nucleus a t any
particular state of ionization. Thus, zero cor-
responds to the neutral atom, F; one corresponds
to the negative ion, F-; minus one corresponds to
the singly ionized atom, F f ; etc. For N = 1,
E is minus the valence state electron affinity;
for N = 0, E is zero; for N = -1. E is the first
valence state ionization potential; for LY = -2,
E is the sum of the first and second valence state
ionization potentials; etc. Mulliken6 has pointed
out the importance of using the ionization poten-
tials corresponding to the appropriate valence
state of the atom in a molecule,
Fig. 1.-Atomic energy change with degree of ionization.
In a molecule, the charge on an atom is increased
or decreased due to the net increase or decrease in
represent conceptually different entities. It would the time that electrons (some of which are originally
thus be desirable to have a defining equation based from the other atom) spend on it. Thus for an
clearly on a physical concept which is in harmony atom in a molecule the quantity ( - N ) represents
with the definition of electronegativity and to the net charge on the atom, expressed in terms
clarify the question of units. of fractions of an electronic charge. If a curve in
Figure 1 is interpolated for fractional as well as
I
TABLE integral values of N , then i t may be taken to rep-
UNITSOF ELECTRONEGATIVITY resent how the energy of an atom in a molecule
Pauling' (XA - XB) = 0.208(D(AB) - would change as it gains or loses fractional elec-
+
(l/d [D(Az) D(Bz)]]'/' (Energy) tronic charge to the other atom, assuming that,
+
M u l l i k d XA = (IA I A - ) / ~ Energy except for the change in valence state, the atoms
Allred and retain the same properties that they possess in their
Rochow' X A = e2Z,rr/r2mv Force isolated state.
Gordy: X A = eZefi/roaov Energy/electron An atom, for which the E vs. N plot has a given
Walshs X A = force constant of AH slope a t the origin (-dE/dN);v 0 will take elec-
molecule Fotce/distance trons away from any atom which has a smaller
Gordy' XA - X B = 2 (fraction ionic slope, because in the process of doing so, the
character) Dimensionless energy of the system as a whole will be lowered.
Sanderson* Ratio of av. electron density Since this is the behavior of a more electronegative
t o that of corresponding atom toward a less electronegative atom, one may
rare gas atom Dimensionless identify the term electronegativity with this
slope: x = (-dE/dN)N = O. Similarly, the
Relation of Extrapolation of Ionization Potentials electronegativity of an ion would be given by the
to Electronegativity.-Recent attempts to obtain slope of the curve corresponding to that ion; e.g ,
electron affinities from the extrapolation of ioniza- xc;- = ( - dE/dN) ;,- = ._1
tion potentials have been reasonably successful. In Electronegativities of C, M, 0, F, Si, P, S and C1.
particular, the form -The coefficients in equation 1 were determined
E ( N ) = U N f b N 2 + ~h-3+- dN' (1 1 from the successive ionization potentials and the
has been shown to be a good approximation to the electron aEnity. For fluorine, chlorine, oxygen and
true equation for representing the energy of sulfur, the coefficients were determined by the
atoms in various states of ionization. When this method of averages,1° but the curve was caused to
form is used in an extrapolation procedure to de- go precisely through the origin and the experimental
termine the electron affinities of fluorine, chlorine points a t n = 1 and n = -1. For fluorine and
and oxygen, from their successive ionization po- chlorine, there are enough data to define the curve
tentials, very nearly the correct values are ob- through N = -5, and for oxygen and sulfur,
tained. Typical plots are like those shown in Fig, through N = -4. For nitrogen and phosphorus,
1, except that (1) the values shown in Fig. 1 repre- there are only enough data to determine one set of
sent the energy of the valence state for each atom coefficients, and for carbon and silicon, there are
or ion, while the othersg represent the average still fewer items of data so that only terms up
energy of all the states of the ground state configura- through the third power in equation 1 were used.
tion, and (2) in setting up Fig 1, the curve fitting was The values of the coefficients are presented in
done using the electron affinity as an experimental Table 11, It can be seen that virtually the entire
point. contribution to E: comes from the first and second
The ordinate, E , represents the total energy of powers of N and that higher terms contribute only
all the electrons around a nucleus of atomic number a minor amount.
(9) H.Hcllman and M. Mamotenko, A d o Physicochim. U R S S , I , (10) H. Margenau and G. M. Murphy, "The Mathematics of
127 (1837); J. L. Margrave, J . Chem. P h y s . , 22, 636, 1937 (1954): Physics and Chemistry," D Van Nostrand Co. Sew York, N. Y.,1956,
H.0.Pritchard and H. A. Skinner, ibid., 22, 1936 (1954). p. 516.
Sept. 5, 1961 ELECTRONEGATIVITY 3549

TABLE I1
FOR EQUATION
COEFFICIENTS 1
Element a b C J
F -12.66 8.70 0.40 -0.083
0 -9.26 7.63 - .32 .002
N -7.16 6.21 - .46 - .01
C -5.00 5.59 .37 ....
C1 -9.83 5.65 .36 - .0433
S -7.40 5.08 - .035 - .022
P -6.37 4.47 .16 - .OS2
Si -4.28 3.38 - .28 ....

p---
In Table 111, the electronegativities determined
from (-dE/dN)N -
0 are compared with electro-
w ----..
H
dN
negativities for the same valence state as given by
Mulliken's method. For C, N, 0, F and the cor-
responding elements of higher rows, the E VS. N
curve can be defined for an appreciable range of
N , because it represents a region corresponding to
the successive ionization of several electrons, all
of which are of the same type, viz., the p electrons.
The E vs. N curve has a discontinuity in slope a t a
point which represents a transition from the ioniza-
tion of p electrons to the ionization of s electrons,
zliz., the point N = -1, for the alkali metal
atoms. As a result, for the alkali metal atoms,
there are only three points on the E os. N curve
for the outermost s-orbital, E ( l ) , E(0) and E ( - 1)
which are obtainable from the first ionization po-
tential and the electron affinity. With only these
three points, the determination of the constants in
the equation E ( N ) = aN +
bN2 shows that the
value of (-ddE/dN)N = o reduces to exactly
(1/2)(1,4 + I*-) which is Mulliken's formula and
emphasizes the close relationship of the present
method to that of Mulliken.
TABLE I11
PUREp-ORBITALELECTRONEGATIVITIES
This work Mulliken
Element Valence state
(dE/dN)N -0
e.v./electron
+
P/i) (IA IA- ),
F s"x2yez, v, 12.66 12.26
0 s2x'yz. v p 9.26 9.58
N szxyz, vo 7.16 7.62
C s'xy, vz 5.00 5.58
c1 sZx'y%, v1 9.83 9.47
S SZX"2, v2 7.40 7.43
P szxyz, Va 6.37 6.2
Si s"y, vt 4.28 4.56

The values of electronegativity presented here

are also like those of Mulliken in that they refer
only to the lowest valence state of the atom.
The same procedure could be applied to higher
valence states if more data on higher ionization
potentials were available. I n order to obtain
values for electronegativity of use in practical
chemistry, i t is necessary to add together contri-
butions from all the valence states, each weighted
according to its per cent. of participation in the
actual hybridized bonds.5
For the rare gases, the point N = 0 corresponds
to a transition from the ionization of p-electrons to
the ionization of s-electrons, and therefore the
curve has a discontinuity in slope a t N = 0,
having a very steep slope for negative N and (as-
suming the electron affinity to be zero) a zero slope
3550 P. ICZKOWSKI
RAYMOND .4SD JOHN L. MARGRAVE Vol. x3

\
Fig. t'b:
I
Behavior OF M g toward Cs in the molecule MgCs Fig 2c --BehaLior of 5Ig toi%ardCs in the molecule MgCs
electronegativity of Mg may be taken as the larger implies that the binding energy of an electron and
of the two limiting slopes a t N = 0. Mg will be the atomic radius tend to change in a regular and
positive when combined with elements having a related manner in going from one atom to another.
greater electronegativity. However, in contrast The periodic table shows many regularities, and it
to most atoms which are negative, when combined is not surprising that many correlations can be
with elements having a lower electronegativity, found between seemingly unrelated quantities.
one expects Mg in molecules like MgCs to show no This is probably the basis for the correlations
appreciable tendency to form a negative ion. between electronegativity and atomic radius.
Thus, such molecules should show a very small Though these correlations may be extremely useful,
ionic component in their binding and behave much i t is hard to find one among them which would
like homonuclear diatomic molecules. qualify as the fundamental measure of electronega-
tivity. This may also be the basis of Pauling's
Discussion scale, as it is difficult enough to imagine a physical
According to the previous analysis, the units of quantity corresponding to the unit, square root of
electronegativity should be energy per electron. energy, much less to see that i t truly reflects a
The quantity (-dE/dN)x = 0 fulfills the desiderata tendency to attract electrons.
for an adequate interpretation of the definition, The methods based on partial ionic character
"tendency of an atom in a molecule to attract must also be considered to be secondary measures
electrons." The result of a small charge disloca- of electronegativity. Although the amount of
tion on interaction of two atoms is that the energy partial ionic character of a bond may be expected to
of the more electronegative atom is decreased by a parallel closely the difference in electronegativity
larger amount than the energy is increased for the of the two atoms, the relation between these
other atom, and therefore the molecule can lower quantities is not exact. The electronegativity
its energy very simply by transferring charge represents an intensity factor in taking away the
from one atom to the other. It should be noted first small amount of charge from the other atom,
that the energy change under discussion is not that while partial ionic character represents both the
which accrues from the electrostatic attraction be- intensity and capacity factors for the atoms to
tween the newly formed ions. The electronegativity give away or absorb appreciable amounts of charge.
is a property which is characteristic of the internal When one says the Li is more electronegative
constitution of an atom and the ions which can be than Na, he implies that Li will take electronic
formed from it. For example, the outer s-electron charge away from Na in the molecule LiNa;
of an alkali metal atom is less tightly bound than however, we have no assurance that the molecule
the p-electron in a halogen atom. This results NaI will have a higher ionic character than the
from the fact that the s-electron of the alkali metal molecule LiI. Evidence points to the fact that
atom is forced by the Pauli principle into an orbital LiI is more ionic than NaI.12 -4lthough Na may
where i t is strongly shielded by all of the inner elec- have a lower intensity for attracting electrons than
trons, while the p-electrons in a halogen atom are Li, i t does not necessarily have a greater capacity
only weakly shielded by each other. The energy to give away a large quantity of charge This
change under discussion results from the fact that lack of correlation between difference in electro-
the s-electron can spend part of its time in the negativity and partial ionic character may be even
lower energy region around the halogen atom. more exaggerated in comparing atoms of different
In a screening constant theory of atoms, the columns of the periodic table, e.g., oxygen and bro-
binding energy of electrons tends to behave mine.
approximately like E = - R ( Z * ) 2 / n 2 ,where R is The present approach and that of Mulliken are
the Rydberg constant, while the atomic radius, similar in that the present method reduces to that
which corresponds to the outermost maximum in of Mulliken when only the electron affinity and
the wave function behaves approximately like the first ionization potential are considered. The
r 'v ao(n2/Z"),where a" is the Bohr radius." This difference in units between those given by the
(11) K. S. Pitzer, " Q u a n t u m C h e m i s t r y . " Prentice-Hall, Sew (la) R P Dailey and C H Tolrnes J Chem Phys 23, 118
York, N. Y . , 1953, p. 81. (1955)
Sept. 5, 1961 BROMINEBY DECOMPOSITION
OF CC13Br BY y - R ~ u s 3551

present method and those given by Mulliken is have an implied unit: per electron, because they
illusory, since ionization potential and electron represent the energy change which accompanies the
affinity, though usually expressed in units of energy, removal or addition of one electron.

[COXTRIBUTION FROM THE DEPARTMENT

O F CHEMISTRY O F THE UXIVERSITY O F \vISCONSIN, MADISON, WISCONSIS]

The Influence of Phase, Temperature and Bromine Concentration on Bromine

Production in the Decomposition of CC13Brby y-Rays'a'b
BY RICHARD
F. FIRESTONE
AND JOHN E. IVILLARD
RECEIVED
DECEMBER
6, 1960

The production of Br2 by the radiolysis of CC13Br with Cos0 -prays has been studied in the solid, liquid and gas phases
and as a function of temperature and bromine concentration. In the solid state a t - 121' and below, the value of G(Br,)
is 0.12 molecule produced/100 e.v., independent of temperature. From -78 to 98' G(Br2) increases with temperature to
about 3.5 with an apparent activation energy of about 2 kcal./mole. Although this activation energy suggests a diffusion
controlled process, there is no change in G(Br2) or its temperature dependence on crossing solid phase transitions a t -35.5
and - 13.5' or in going from the solid to the liquid state a t the melting point - 5.6'. Preliminary experiments indicate
that there is no marked difference in yield in the gas phase as compared to the liquid phase a t the boiling point. At bromine
concentrations above about 0.01 M, G(Br,) is independent of bromine concentration at all temperatures in both the solid
and the liquid. Below 0.01 i V and a t temperatures of 20" and above, however, G(Br2) decreases with increasing concentra-
tion. The bromine-sensitive reaction appears t o have an activation energy of about 3 kcal./mole, slightly higher than the
bromine insensitive reaction. The presence of oxygen raises G(-CC13Br) several-fold, the production of Clz (a minor radioly-
sis product) being increased by a larger factor than Br, production.

Introduction an annular vessel surrounding a 40 curie Corny-ray source.6

Attached to the annular vessel a t a distance sufficient t o
This paper reports investigations of the effects prevent troublesome darkening by irradiation was a cell
of phase, temperature and radical scavengers on made of square Pyrex tubing of 1 cross section. By
the radiolysis of CC13Br, with particular reference pouring the liquid into this cell the optical absorbance
to the yield of bromine in the liquid and solid of the liquid could be measured periodically with a spectro-
photometer without opening the purified, air-free sample to
phases. I t is part of a more extensive investigation the air.6 Thermostating media used in maintaining desired
which has included2 a comparison of the photo- temperatures during irradiations included: liquid air,
chemical and radiation induced decomposition of -190"; liquid-solid butyl chloride, -121'; Dry Ice-
CCl?Br and the exchange of CClsBr with Br2 in acetone, -78'; liquid-solid anisole, -37'; Dry Ice-
acetone-aniline, - 12'; ice-water, 0'; thermostated
the gas and liquid phases. In the latter work2 mineral oil-bath, temperatures above 20'.
five organic products, CC14, CC12Br2, CClBr3, Dosimetry.-Dosage rates were determined by measure-
C2C16 and CzC16Br, have been observed and their ment of the F e + + +yield in air-saturated water solutions of
yields have been followed with the aid of gas chro- ferrous ammonium sulfate, 0.8 N in HzSO~, under irradiation
conditions identical with those used for irradiation of CC13Br.
matography. It was assumed that the molar absorbancy index of F e + + +
These studies were designed to obtain informa- is 2130 a t 305 mp and 25' and that 15.6 F e + + + ions are
tion which would contribute to an understanding formed per 100 e.v. absorbed.? Energy absorption in CCL-
of the mechanisms of radiolysis of organic com- Br was determined from that in the dosimeter solution by
use of the calculated ratio of the mass absorption coefficients
pounds. CCtBr was chosen because of its sim- of CC13Br and water for Co60 y-rays. The value of this
plicity, because of the ease of following its decom- ratio is 1.12.
position by measurement of the bromine produced Sample Preparation.-CC13Br from the Michigan Chemi-
and because information was already available on cal Corporation, containing 0.1% glycidyl phenyl ether as
a preservative, as well as other impurities, was purified by
its thermal3 and photochemical4 reactions and its the following steps: vigorous mechanical stirring with suc-
reactions after activation by nuclear processes.5 cessive portions of concentrated &So4 until no discoloration
of the acid occurred on several hours of additional stirring;
Experimental washing with Na2C03 and water, drying with CaCl, and
General Methods .-All irradiations of liquid and solid P205; illumination with a 1000 watt projection lamp a t
CCI3Br were made on samples of about 5 ml. contained in 6 inches for 10 hr. after making 0.01 M in Brz; passage
through a 30 X 1.5 cm. column of activated alumina; frac-
(1) (a) Presented before the Division of Physical and Inorganic tional distillation through a 12" Vigreux column; passage
Chemistry at the April, 1955, meeting of the American Chemical So- of the middle third of the distillate through a fresh activated
ciety in Cincinnati, Ohio. (b) This work is reported in greater detail alumina column. A similar procedure omitting the HzS04
in the Ph.D. thesis of Richard F. Firestone, University of Wisconsin, and alumina treatments gave a product which yielded radiol-
1954. ysis results indistinguishable from those on material pre-
(2) A. H. Young, Ph.D. thesis, University of Wisconsin (1958), pared by the more extensive procedure. CCLBr from each
rvailable from University Microfilms, Ann Arbor, Michigan. method of purification had a boiling point of 103.5' a t 745
(3) (a) A. A. Miller and J. E. Willard, J . Chcm. P h y s . , 17,168 (1949); mm., a freezing point of -5.6", and a refractive index,
(b) N. Davidson and J. H. Sullivan, i b i d . , 17, 176 (1949); (c) E. nz5D of 1.5032 =I=0.0002. CCI2Br2 (Michigan Chemical
Becker, Ph.D. thesis, University of Wisconsin, 1953. Corp.), used as a carrier for small quantities of radioactive
(4) (a) E. Paterno, Jahres6. Fouls. Chem., 24, 259 (1871); (b) CC12Br2,was prepared for use by fractional distillation in
W.Noddack, 2. Elektrochem, 27, 359 (1921); (c) H. Grusse, i b i d . , 24, a Vigreux column a t 43' under 27 mm. pressure.
144 (1923); (d) H.G.Vesper and G.K. Rollefson. J . A m . Chem. SOC., The irradiation vessels were prepared for filling by allow-
16, 1455 (1934); (e) W. Franke and H. J. Schumacher, 2. physik. ing them to stand filled with alcoholic KOH for an hour,
Chcm., B42,324 (1939); (f) W.U. Day, Ph.D. Thesis, University of following which they were rinsed, allowed to stand in boiling
Wisconsin, 1942.
(5) (a) S. Goldhaber, R. S. H. Chiang and J. E. Willard, J . A m . (6) R. F. Firestone and J. E. Willard, Reo. Sci. I n s f r . ,24,904 (1953).
C hem. SOC.,73,2271 (1951); (b) S.Goldhaber and J. E. Willard, ibid., (7) C. J. Hochanadel and J. A. Ghormley, J . Chcm. Phys., 21, 880
74, 318 (1952). (1953).