3 March 1995

CHEMICAL
PHYSICS
LETTERS
ELSEVIER Chemical Physics Letters 234 (1995) 113-118

A measure of the effective electric-dipole polarizability of argon

R.L. Asher a,b, D. Bellert a,b, T. Buthelezi a,b, V. Lewis a,b, p.j. Brucat a,b
a Department of Chemistry, University of Florida, Gainesville, FL 32611-7200, USA
b Center for Chemical Physics, University of Florida, Gainesville, FL 32611-7200, USA
Received 21 November 1994; in final form 13 December 1994

Abstract

The electric dipole polarizability has been determined in a unique way. Binary complexes of argon with transition metal
cations have been studied by optical spectroscopy. The vibrational structure of electronic transitions near the dissociation
limit of the upper state quantifies the forces between the atoms. The polarizability of argon is determined from the
charge-induced-dipole force in four different molecular ions as 1.49 ± 0.10 ~3. This is lower than the accepted value by
9%. This discrepancy may imply that the multipole expansion of the charge distortion does not have linearly independent
coefficients in the field of an ion at close range. Induced higher-order moments appear to reduce the effective dipole
polarizability.

1. I n t r o d u c t i o n the polarizability volume of the molecule. At optical

frequencies the index of refraction, n, is more easily
Normally, the determination of the polarizability measured than the impedance of the sample. In this
of an atom or molecule is made from a bulk mea- case, the dielectric constant may be obtained from
surement of the relative electric permittivity (dielec- the relation ~/~0 =n2 and the polarizability ob-
tric constant) in the radio frequency or from the tained from Eq. (1), above.
index of refraction at optical frequencies. The theory While the above treatment is not without hidden
behind this determination may be found in any un- and undiscussed approximations, let us focus on the
dergraduate physical chemistry textbook [1]. In the fact that a molecular property, a , is derived from a
absence of a permanent dipole, or at the high temper- bulk phase measurement. In the bulk the molecules
ature limit, the reorientational contribution to the are in an environment created by their neighbors.
dielectric shielding of the applied electric field may The bulk electric permittivity of the sample is merely
be ignored and the polarizability (more precisely, the apportioned equally among the number of molecules
polarizability volume) of a molecule may be deter- in the sample, without regard to how the condensed
mined from the Clausius-Mossotti equation [1] as phase affects individual molecular polarizability. This
(e/E0) - 1 has the consequence of ignoring any possible many-
= (47rp/3) a, (1) body effects in the distortion of the molecules in the
(~/~0) + 2 externally applied electric field.
where ~/~0 is the dielectric constant, p is the This approximation is pragmatic in that dielectric
number density of molecules in the sample, and a is measurements on bulk samples are easy, in contrast

0009-2614/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved

SSDI 0009-2614(95)00005-4
114 R.L. Asher et al. / Chemical Physics Letters 234 (1995) 113-118

to the measurement of the properties of isolated 3. Results

molecules. We will describe the distortion of an
individual atom of argon in the electric field of a Sharp resonant photodissociation spectra have
nearby atomic ion. The interaction of these two been observed in our laboratory for four tran-
atoms will be studied without interference from col- sitional-metal argon binary complexes: VAr + [4],
lisions and even the application of an external elec- NiAr + [5], CoAr + [6], and ZrAr + [7]. In all cases,
tric field. The distortion of the argon is the origin of more than one excited state of the ion has been
a long-ranged binding force in an ion rare-gas di- observed. The properties of these states may vary
atomic, and the vibrational structure of this complex drastically, even in a single molecule, because of
determines the effective value of the polarizability in differences in the electronic configuration of the
this unique environment. It is clear that the technique metal ion which leads to that state. However, the
described in this Letter is not capable of routinely behavior of the vibrational energy level spacing of
placing a measure on the polarizability of an arbi- these states near their diabatic dissociation limit, D,
trary molecule, but is capable of determining the shows a regular structure and is the focus of our
effective polarizability of argon to an accuracy wor- attention. For a molecular potential of the form
thy of discussion and comparison.
C,
U(r) = repulsion rn (2)

2. Experimental the vibrational energy level structure of the highest

bound vibrational states have the property [8,9]
MAr + ions are generated in a supersonic beam
A G ( v ) = - i1[ E ( v + 1) - E ( v - 1)1
ion source, mass selected in a customized TOF mass
spectrometer, and studied spectroscopically by reso- = -K[E(v) -- D] (n+2)/2n, (3)
nant photodissociation with visible laser light. The
details of the experimental procedure for the acquisi- where E(v) is the energy of the vibrational level in a
tion of molecular ion spectra have been published given electronic state labeled by quantum number v
before [2,3]. Briefly, a laser driven plasma contain- and D is the energy of the dissociation limit of that
ing metal cations is entrained in mixture of = 5% Ar electronic state. AG(v) is 'approximately' the
derivative of the total energy with respect to vibra-
in He and expanded through a supersonic nozzle.
The beam of cold ions thus formed is skimmed and tional quantum number. [ E ( v ) - D ] is the diabatic
injected into a tandem mass spectrometer. Here, the binding energy of the vibrational level labelled by v.
The constant, K, has been shown [8,9] to be
ions are accelerated to 1.5 keV, mass-selected by
TOF, photoexcited with one or two tunable laser hC;1/"nr((n + 1 ) / n )
pulses, and kinetic energy analyzed by an electric K= (2~rg)l/2r((n + 2 ) / 2 n ) (4)
sector. The fragmentation following photoexcitation
is monitored as a function of dissociation laser fre- In our study, the long-ranged force binding the atoms
quency to obtain an excitation action spectrum. The in MAr + ions is expected to be charge-induced-di-
spectrum so obtained is analogous to the absorption pole (n = 4) and thus
spectrum of the isolated ion and is unambiguous 1¢ 20~\),
with respect to the identity of the carrier of the C4 = ~l.q (5)
spectrum. The supersonic expansion generates target where q is the charge of the metal ion and a is the
ions with very low internal energy (T < 10 K) allow- dipole polarizability of the argon. Application of Eq.
ing for the routine production of unstable and reac- (3) to MAr + ions suggests that for such charge-in-
tive species. The ultimate resolution of the spectrom- duced-dipole bound species a plot of A G 4/3 v e r s u s
eter is limited by the laser-linewidth and is on the [ E ( v ) - D ] should be linear for the highest fre-
order of 0.1 cm -~. Ne optogalvanic transitions are quency transitions in any electronic band system.
acquired simultaneously with the unknown spectra This analysis has been used by our group [3-6] to
and serve as frequency calibration. accurately extrapolate bound state vibrational level
 R.L. Asher et al. / Chemical Physics Letters 234 (1995) 113-118 115

energies to the dissociation limit. In deference to the CoAt+ I . a l q o y - l ~ i n

200 m A'-X
formulators of Eq. (3) (Refs. [8,9]), we dub this plot
a LeRoy-Bernstein plot. (Note that a LB plot of
potentials with a long-ranged force characteristic of a
power law with n ~ 4 will have a different power of
AG on the vertical axis, but we would call these
LeRoy-Bernstein plots as well.) This analysis is
basically an improvement of the Birge-Sponer plot
(AG 1 versus [ E ( v ) - D ] ) which works only for
molecular potentials well described by a Morse oscil- -500 -400 4.q(~ -200 -I O0 0 1O0
lator. It should be noted, however, that a traditional Diabatic Binding [wavenurnbem]
t~(v) -D]
Birge-Sponer analysis has been successfully applied
to MgRg ÷ ions [10] which are clearly not Morse Fig. 2. A LeRoy-Bernstein plot of the vibrational structure of the
h~ state of the CoAr + ion. The axes and legend are as in Fig. 1.
oscillators near dissociation.
Fig. 1 shows a LeRoy-Bernstein plot for three
quartet electronic states of the NiAr + molecule [5].
Although the experimental data deviates from the the argon polarizability, the reduced mass of the
form of Eq. (3) far from dissociation, gratifying diatomic, and fundamental constants,
linearity is observed near E ( v ) = D . This confirms - 3.898
the origin of the long-ranged force between the --K4/3 .~ - -
~z/30t 1/3 , (6)
dissociating atoms as charge-induced-dipole. The
limiting slope of the data near the dissociation may where the numerical constants have been evaluated
be obtained from Eq. (3) for n = 4 as a function of a s s u m i n g g 4/3 in units of (cm-1) 1/3, /z in units of
o3
g / m o l , a in units of A , and a singly charged M +
ion. The data in Fig. 1 indicate an average limiting
58NiAr+ LeRoy-Bernstein slope of - 0 . 4 1 1 (cm-1)1/3 for three electronic states
200 * (c-X) of NiAr + which ,yields an 'effective' polarizability
• = q,~ = (~-x)
of argon of 1.53 A3 for this molecule.
Fig. 2 shows a similar LeRoy-Bernstein plot for
[] Ca-x)
the so-called /~ state [6] of the CoAr + molecule.
Again, excellent agreement of the data to the form of
Eq. (3) is observed near the dissociation limit D.
The limiting slope derived from this plot is - 0 . 4 2 1
( c m - )1 1/3 yielding an effective polarizability of ar-
gon in this molecule of 1.40 ,~3.
-400 -300 -200 -100 0 100
Figs. 3 and 4 show LeRoy-Bemstein plots for
Dial:~tic Binding [wmvenumbem] electronic states of two additional molecular ion
rE(v)- D] complexes containing At: VAt + [4] and ZrAr + [7].
Fig. 1. A LeRoy-Bernstein plot of the vibrational structure of In all cases, the LB plot is linear near dissociation,
three quartet states of the NiAr + ion. As this ion is expected to
allowing the limiting slope to be determined.
exhibit a long-range 1 / r 4 attractive force, a plot of AG 4/3
versus [ E ( v ) - D ] should be linear for the highest frequency To combine the results of all the molecules inves-
transitions in any electronic band system (see Eq. (3)), where tigated we examine the cube of Eq. (6),
E ( v ) is the energy of the vibrational level in a given electronic
state labeled by quantum number v, and D is the energy of the -1 ((128)(1024)N2h 3 /'(5/4) 4 )
- g4/x2 =
dissociation limit of that electronic state. [ E ( o ) - D] is therefore ot "IT2q2c /'(3/4) 4
the diabatic binding energy of the vibrational level labelled by v.
Data for three different electronic states, marked with symbols, - - 59.217
are derived from three different electronic band systems which are (7)
listed in the legend on the right-hand side of the figure.
116 R.L. Asher et al. / Chemical Physics Letters 234 (1995) 113-118

VAt + I.hRoy Bemstoin

200 = A-X Limiting L-B Slopes for MAr+

= = = = m N i A r+

D e = "~.8 + / - 2 7
~-~ = CoAr+

~o ~ . . z,A:

~- -0.~l ~ ' " " " "

-400 ~00 -200 -100
Dialytic Binding [wavenurnbero]
0 100
[E(v) - o] -0,10 0,001 0,00~
Fig. 3, A LeRoy-Berustein plot of the vibrational structure of the Ix-2 [(mol/g) 2 ]
A state Of the VAt + ion, The axes and legend are as in Fig. 1.
Fig. 5. A plot for the experimentally determined - K 4 values
Significantly perturbed levels have been removed from the plot,
versus the inverse square of the reduced mass of the molecule
from the six limiting slopes in Figs. 1-4. This plot, according to
Eq. (7), should have a slope inversely proportional to the polariz-
ability of argon and an intercept of zero. The solid line shows a
Eq. (7) gives the slope of a plot of - K 4 versus /x-2 least squares fit to such a function and the dotted lines bound the
estimated error to this fit. The slope determined from this graph is
as a function of a and fundamental constants only,
- 3 9 . 8 +2.8 which yields a value of the dipole polarizability
w h e r e K 4 is in cm -1, /z is in units of g / m o l , a is volume of argon of ot = 1.49+0.10 ,~3.
in units of .~3, q is in esu, h is in erg s, c is in c m / s ,
and N A is Avogadro's number. Fig. 5 shows such a
plot for the experimentally determined - K 4 values
versus the inverse square of the reduced mass of the determined from this graph (Eq. (7)) is - 39.8 + 2.8
molecule from the 6 limiting slopes in Figs. 1-4. which yields an effective value of a = 1.49 +__0.10
This plot, according to Eq. (7), should have a slope 03
A.
inversely proportional to the dipole polarizability of
argon and an intercept of zero. The solid line shows
a least squares fit to such a function and the dotted
4. Discussion
lines bound the estimated error to this fit. The slope
The value obtained for the polarizability volume
of argon in this study is 9% lower than the accepted
90~Ar+ LeRoy Bernstoin
literature value 1.642 + 0.001 .~3 for the static polar-
200- • o-X izability volume [11,12] of this atom. This discrep-
=
ancy is actually outside the estimated error limits of
our study ( _ 7 % ) , but just barely. Is there any
significance to this discrepancy?
100 • Although the possibility of many-body effects as
a cause for a deviation in the dielectric constant from
a pressure-independent form (Eq. (1)) exists, as al-
luded to in Section 1, the effects of pressure on the
0 ............................
-500 -400 -300 -200 -1O0
Dialytic ~inding [wavenurnl~t~]
1O0
dielectric constant of the rare gases are well studied
and presumably understood [13]. Careful extrapola-
[E(v) - D] tion to low density, which is clearly possible in high
Fig. 4. A LeRoy-Bemstein plot of the vibrational structure of the
sensitivity measurements, can eliminate the un-
c state of the ZrAr + ion. The axes and legend are as in Fig. 1. wanted confusion introduced by such phenomena.
Significantly perturbed levels have been removed from the plot. The literature value of the static polarizability from
 R.L. Asher et al. / Chemical Physics Letters 234 (1995) 113-118 117

Ref. [11] has been determined in this way and is thus diatomic complex (Fig. 5) which supports the valid-
taken as an intrinsic property of the isolated atom. ity of the LeRoy-Bernstein analysis [8,9]. The polar-
For atoms and molecules with absorptions, the izability of argon determined in this study is found to
index of refraction, dielectric constant, and polariz- be an 'effective' value, which will be of use in the
ability are all functions of frequency. Measurements modelling of other ion-argon complexes. But what
of the 'static' polarizability are not done with static makes it just an effective measure?
electric fields, but with rf [11] or microwave [12] In both the optical and rf measurements of the
fields. These frequencies are much lower than any polarizability, the applied electric field has parallel
resonance frequencies in the argon atom and are thus equipotential surfaces (plane wave) over the dimen-
considered to be varying so slowly in time as to be sions of the atom, i.e., there is no appreciable electric
'static'. The vibrational energy level structure near field gradient. In the case of the diatomic ion com-
dissociation for a diatomic ion is determined by the plex, the electric field lines of the ion radiate from its
charge-induced-dipole forces at the outer turning center of charge. Even at the highest investigated
point of the potential and on a timescale of a vibra- vibrational levels, the outer turning point of the
tional period. Thus, our spectroscopic technique nuclear motion of the ion coml~lex only has an
should determine the polarizability of argon at a internuclear separation of about 8 A. This means that
frequency comparable to that corresponding to the the electric field that the argon experiences within
diabatic dissociation energy of the ion complex, the molecule has a significant gradient over its di-
typically about 3000 cm-1 for the ions in this study. mensions. This asymmetry of the electric field in-
The frequency dependence of the polarizability duces higher moments in the charge distribution of
has several distinct origins. For polar species, the the argon than would be induced by the field within
dielectric constant drops as the frequency of the an infinite parallel-plate capacitor or in either the
applied electric field exceeds the reorientation time optical or rf studies referred to above. We think that
of the molecule. But, for argon, no permanent mo- the induction of these higher moments will effec-
ments exist to be reoriented, so no such drop is tively reduce the polarizability in the dipole distor-
expected to occur. The argon atom has resonant tion, which is consistent with observation.
electronic absorptions in the deep UV. The
Kramers-Kronig relations [14] describe how the di-
electric constant (and thus the molecular polarizabil- 5. Conclusion
ity) is affected by this absorption. Since the charac-
teristic frequency of the electric polarization in this
The vibrational structure of ion-argon complexes
experiment is well below the frequency of electronic
has been studied by application of the formalism of
absorption, the dielectric constant and thus the polar-
LeRoy and Bernstein [8,9]. The effective polarizabil-
izability are expected to increase with increasing
ity of argon obtained from these diatomic complex
frequency. This is supported by the measured value
ions (1.49 +__0.10 .~3) is comparable to, but smaller
[15] for the polarizability of Ar at an optical wave-
than, that measured by more conventional methods.
length of 6328 A (corresponding to a frequency of
This reduction in dipole polarizability within the
15800 cm - 1 ) of 1.663 A 3, an increase of 1.3%
complex may be due to the effect of the simultane-
over the 'static' value. An effect of the characteristic
ous distortion of the atom into higher multipole
frequency of the measurement then cannot be the
moments.
cause for the observed 9% discrepancy.
It seems from the above discussion that neither
many-body effects nor frequency dependence can
explain the discrepancy between the effective polar- Acknowledgement
izability of argon as measured from the spectroscopy
of ion-argon complexes and bulk determinations. This work was financially supported by the Na-
Yet, the spectroscopic data is consistent with the tional Science Foundation under Grant CHE-
predicted trend with respect to reduced mass of the 9242111.
118 R.L. Asher et al. / Chemical Physics Letters 234 (1995) 113-118

References [8] R.J. LeRoy and R.B. Bernstein, J. Chem. Phys. 52 (1970)
3869.
[9] R.J. LeRoy, Chemical Physics Research Report CP-500,
[1] P.W. Atldns, Physical chemistry, 5th Ed. (W.H. Freeman, University of Waterloo (1994).
New York, 1994). [10] J.S. Pilgrim, C.S. Yeh, K.R. Berry and M.A. Duncan, J.
[2] D. Lessen and P.J. Brucat, J. Chem. Phys. 90 (1989) 6296. Chem. Phys. 100 (1994) 7945.
[3] D.E. Lessen, R.L. Asher and P.J. Brucat, Intern. J. Mass. [11] R.H. Orcutt and R.H. Cole, J. Chem. Phys. 46 (1967) 697.
Spec. Ion. Processes 102 (1990) 331. [12] A.C. Newell and R.C. Baird, J. Appl. Phys. 36 (1965) 3751.
[4] D. Lessen and P.J. Brucat, J. Chem. Phys. 91 (1989) 4522. [13] N.E. Hill, W.E. Vaughan, A.H. Price and M. Davies, Dielec-
[5] R.L. Asher, D. Bellert, T. Buthelezi and P.J. Brucat, Chem. tric properties and molecular behavior (Van Nostrand Rein-
Phys. Letters 228 (1994) 599. hold, New York, 1969).
[6] R.L. Asher, D. Bellert, T. Buthelezi and P.J. Brucat, Chem. [14] H.A. Kramers, Atti del congr, intern, dei. fisici, Como 2
Phys. Letters 227 (1994) 277. (1927) 545.
[7] R.L. Asher, D. Bellert, T. Buthelezi, D. Lessen and P.J. [15] A. Dalgrano and A.E. Kingston, Proc. Roy. Soc. A 259
Brucat, Chem. Phys. Letters 234 (1995) 119. (1960) 424.