Volume 37, number 1
CHE:511CAL PHYSICS LETTERS
BINDING OF AN ELECTRQN BY THE FIELD OF A MOLECULAR
J.L, CARLSTEN,
I.R. PETERSON
*
Joinf Itrsti~urr/or Laborarury Arrrophyrrcs,
Boulder,
Colorado
80302,
University
0-f Coiomdo
cmd
I January 1976
DlPOLE
h’ational
Bureau
arid Na:ionol
Burcm
-
LiCI-
of S;andar&.
USA
and’
W.C. LINEBERGER
**
frlsrirlcre for !.oboraror.r
Joinr
und
Asrroph~sics.
L;ilLwsi!_p
Dr~mrrrnenr of Ci~cmisrry~Ulriversiry o~C~!omdo,
of
Colorado
Boulder.
Colorado
80303,
of
Standards
USA
WC pre.~nt direct
cxpcri.mcnlal ruidcnsc for :hc binding oi an electron by [lx dipok moment of 3 neutral molcculc. The
pho~w!ccrr~n qxc:n:m
of LiCI- IVZSnblG.cd and srulyzcd to dcrerminc that ihc binding energy OF the elccxon in rhe
UCI- ion is (0.61 2 0.07,j CL’. This rcwl\ is 3n order of magGcudc smdlcr than rhe predictions of rhc kxcd dipole theoretical
ma!cl, but i.; consislcnl
si. +h n recent crilc~~l~~ion, which ~110~s for distoriion
of the dipolor core.
1. Incnlducrioll
The binding
of nn elcctran
by 3 dipok
fie!d
of a
ncumi P~iti~iC has attracted
conside:able
attention
owr the past years. An initial thcorelical study [I]
dcali
wirh the inrcracrion
or ncgativc mesons wit:1 Ilydrogen. Tlt;c
studies indicated a critical dipoie moment
ot’0.63’, caO formed by the proton and negative meson was needed to bind the electron. bter Ivallis cl al.
12) calcr?!ated the energy levels of an electron in the
field of a rigid dipole. A number of researchers [3-S]
then rediscovered that the minimum dipoie moment
needed to bind the electron was 0.639 eaO. Subsequcntlq’ Crawford [6) showed that, for a rigid dipole
with a moment greater than 0.639 co,-, , an infinite
number of bound states exist. Recently rotational effccts have been shown to raise the minimum dipok moment necessary to bind an electron and also to reduce
the number of bound states [7-q].
Euperirnental
evidence for the binding of an elecf
l
JILA Visiting Fellow 1974-75, on kve
Research Institute.
* Alfred P. Wan Foundation Fellow.
from SLnnford
tron by the dipole field of a polar molecule has been
indirect at best [ 1O] . Electron swarm data [ 11,12] indicate drastic changes in the electron drift velocity in
polar gases when the dipole moment becamss greater
than 0.639 cog. Such behavior has been interpreted
as
due to the formation of temporary bound negative
ions 2s the electron drifts through the polar gas. Scattering experiments
[ 131 of electrons by alkali-halides
have bezn compared with calculations
using the first
Born approximation,
and a point-dipo!e
mcdei. These
experiments
indicate that the absolute total cross sections are considerably
less than those predicted using
the first Born approximation.
Recent close coup!ing
calculations
by Allison i 141 have improved the agreement with experiment fOi momentum
transfer cross
sections.
Since LiCl is 2 closed shell molecule,
binding of an
estra electron is not obviously expected. However,
LiCl has a large dipole moment which is capable of
binding the extra electron. We have produced
LiC]-. ,
determinki
its electron affinity using photoelectron
spectroscopy,
and compared the results with the fueddipole model descriloed above, md with a recem calculation by-Jordan
[IS].
-5
Volume 37,
numt3c.r
2. Experimental
CIlEhfICAL
1
PHYSICS
3. Results
arrangement
Most,of the details of thr: experimental
apparatus
other than ,those for the ior: source have been presented
elsewhere [ 161. The LiCl- was produced in a low presure, high temperature
discllarge ion source * which
contained
some LiCl (and F;Br) crystals. The filament
in the source was made from 15 mil W wire and run at
~14 A of dc current. The resulting temperature
of the
source was estimated to be =lOOO K. A discharge was
then struck between the filament and an anode. The
discharge was found to protluce IiCl- best when run at
~40 V dc and =150 mA. A 40 mil hole in the anode
allowed extraction
of t;le ions with 1500 V applied between the anode and the first focusing electrode.
After tI;e ions were decelerated to 680 eV, they
were deflected -10” to remove neutrals from the beam.
Ttle ions were then mass selected.with
a Wien filter
and directed into a field fret: interaction
region, as
shown in fig. 1. There the ions were crossed with a
focused, intracavity
beam of an argon-ion laser operating at 458 nm (/IV= 2.540 eV). The energy of the
photodetached
electrons was determined
by a hemispherical energy analyzer which had a resolution of
x.50 meV (fwhm) and the resulting spectrum was accumulated
in a multichannel
ana1yze.r.
* Colutron
Corporation.
LETTERS
Boulder. Cclorado
80302, USA.
With tfle source operated as described in section 2,
LiCI- beam currents of 0.5 nA were obtained in the intemction region. The identification
of LiCl- was confirmed by mass scans which showed a peak at mass 42
and another at mass 44 in the natural isotopic ratio of
approximately
3: 1 for Cl (mass 35) and Cl (mass 37).
The resulting photoelectron
energy spectrum is
shown in fig. 2. A scan of photodetached
electrons for
I<- was also taken: to determine
the energy scale and
eliminate any contact potentials. The peaks correspond
to (IJ’JJ”) transitions
from the u” vibrational
level of the
ground eIectronic state of LiCl- to the u’ vibrational
level of the ground electronic state of LiCI. The spa”ing of the three peaks is 8E + 10 meV, in reasonable
agreement with the 80 meV vibrational
spacing of
LiCl [17]. Other data also show a weak peak at 2.0 eV
which could be the (0,l) transition due to thermal
excitation in the source. Its location leads to a value
of 60 + 10 mcV for the oc of LiCl- _
Effects of the rotational distributions
upon the positions of the peaks for LiCl-- shown in fig. 2 have been
considered and indicate that any shifts will be much
smaller than our present measuring errors.
The peaks at Z.94 eV, 1.85 eV and 1.76 eV are
ascribed to the (O,O), (I ,O) and (2,O) transitions
respectively. If the peak at 1.94 eV were a transition with
u’ > 0, there would then be an observable peak appear-
I.3 I.4 I.5 1.6 1.7 1.:’ 1.9 2.0 2.1
ELECTRON
Fig. I. Dia&ram of
the interaction
re_rion. The
LiCl ions were
crossed .with an Ar+ laser (/IY = 1.540 eV) and the cnqy
of
.,lhqrcscltinp photodek&cd
electrons ~ralyzcd by the hrmisphcriclll energy ailalyzcr.
ENERGY
2.0 2.1
(eV1
I%. 2. Photoelectron cnerev suectrum of LiCr _The ~caks at
I .%
eV, 1.65 eV, and
1.76‘&
carre$Tond
used
and (2.0) transitions. A’scan of R-was
electron encr,~ scale.
t0
ihs
(O,Oj, (I ,O)
to cdibratc
the
CHEMICAL PHYSICS LE’lTERS
Volume 37: number 1
ing 80 mV hipher in energy, due to the fact that FranckCondon factors vary relatively slowly with u‘ for trsnsitions that originate from u” = 0. Since the only
other peak is one that is very weak and also located
at 60 mV higher instead of 80 mY, we conclude that
the peak at 1.94 cV is the (0,O) transition.
The ratios of the intensities of the (O,O), (1 ,O) and
(3,O) transitions were used to calculate [16] the bond
length of LiCI-. Using simple harmonic oscillator wavethe Franc-Condon
factors were studied as
a function of the bond length difference between the
ion and neutral. The resulting equilibrium distance for
JX-- is
functions,
where the error arises from the uncertainty in the w,
of the transmission of the
energy analyzer.
Using the value of 1.94 eV for the location of the
(0,O) transition, we can determine rhe electron affinity
of LiCl. Applying conservation of energy and momentum
one obtains the foflowiqg expression involving the electron affinity EA [f6]
of LiCI- and the nonlinearity
IW = E A + .Q + m W/M + Ecp )
(1)
where hv is the indicenr: photon energy (hv = 7,540 eV),
R is the labomtory energy of the detached photoclectron leaving the molecule in the ground vibrational
level of the neutral molecule, &l&f is the ratio of eiectron to ion mass, FVis the kinetic energy of the ion, and
E,, is the contact potential energy difference between
the interaction region and the ciectron energy analyzer.
The fz W/M correction is due to the transformation from
center-of-mass frame to the laboratory frame. The E,
is eliminated by s~ultaneo~sl~ obsming electrons
from another ion whose EA is known. In this experiment we used K- whose EA is (0.5012 2 0.0005) eV
[W
*
Using cq- (1)
EL7
4. Comparbn
I January 1976
w&h theory
A number of researchers have studied a fried dipole
model which considers two point charges (+e and -c)
separated by a distance R, thus forming a dipdle rnoment of magnitude eR; When CR > 0.639 eaO, this
dipole moment is capable of binding an extra electron.
Wallis et al. [‘I!] calculated the binding energy of such ;1
system for various values of eR _
The dipole moment of LiCl is 2.79 eeO [IQ] . IJsing this value, the theory of Wallis et al. [?-I predicts
that an electron in the ground state of LiCl- should be
bound by 4.65 eV, and that the first excited level
shouid be bound 0.085 eV. It is very unlikeiy that the
observed electrons have come from an escitcd state of
LiCl- . The reasor! that the experimental value of
0.61 eV is JO much smaller than the theoretical value
of 4.68 eV is undoubtedly due to the Fact that the
theory considers only a pure dipole field. This model
is only appropriate at large distances, since it ignores
the effects of core repulsion and the finite size of the
atoms which are important at small distances.
Using the tied dipole model, Turner et aI. [30]
have c*alcuIated the mean distance (rI) of the bound
eIectron from the po&ivc dipole charge. For a dipole
moment of 2.79 rag, they find that the mean distance
(rt) is 1.79go. Thus for the electrons bound by l_iCI,
the fivced dipole mode! predicts that the electron will
he *I .79a0 from the Li’ ion. Since this is comparable
to the radius of the Lii ion (z 1.29~~12 1,22] ), ovcrlap will occur and the efrcct of cars repulsion will certainly be important.
Recently Jordan [Is] has calculated the potential
curve for LiCi- . He r”inds that the electron affinity for
LiCl is W.5
eV, which is reasonably
close to our ex-
perjme~t~ result. This difference: is attributed
[15) to
correlation effects which have not as yet been included in the theory.
for both LiCl and K, onz eliminates
and obtains:
5. Conclusion
EA(LICI) = EA(K) + {a(K) - fi(LiCl)]
-mFV~I/lMOC_)
- I/M(LiCI)]
.
ca
Therefore using the data shown in fig. 2 and applying
eq.. (2) we can obtain
EA(UCI) = (0.61 t 0.02) eV .
(3)
We have piesented direct esperimenrsl evidence for
the binding of an electron by the dipole moment of a
neutral molecule, The binding eneqy of the electron i&
the LiCl- ion is an order of magnitude smaller than the
prediction of the f%ed dipole theoretical model. More
sophisticated theoretical models, such 2s the one being
,.‘.
‘..
..
7
Volume 37, ntimber
I
CHEhllCAL
PHYSICS LETTERS
d&loped
by Jordan [15], which take account of the
properties of a.re,a.l molecule, are needed before ac&rat& predictions
can be made about the interacton
of
)’~&trons
in polar gases or other situations in which the
‘effect of the dipole field is important.
[S] J.&l. Levy-Lebland, Phys. Rev.,153 (1967) 1.
[6] O.H. Crawford, P-rot. Phys. Sac. (London),91 (1967) 279.
[71 O.H. Crawford,_hlol. Phys. 20 (1971) j8j_
W.R. Garrett, whys. Rev. A.3 (1971) 961;Chem. Phys.
Letters 5 (1970) 393: Mol. Phys. 20 (1971) 7.5 I.
C. Bottcher, MO!. Phys. 19 (1970) 193.
O.H. Crawford, Mol. Phys. 26 (1973) 139.
H. Lehning, Phys. Letters t8A (1968) 103.
J.h.D.Srockdale,
L.G. Christophcrson, J. Turner and V.
:’
Acknolwledgcmen:
E. Anderson,
Phys. Letters
25A (1967)
519.
R.C. Slater, h1.G. Fickes, W.G. Becker and R.C. Stern,
J. Chem. Phys. 61 (1974) 2290.
A-C. Allison, 3. Phys. B 8 (1975) 3’5.
K.D. Jordan, private communication;
to be published.
M-W. Sic@, R.J. Cclotta, J.L.. ffafl, 3. Levine and i2.A.
Bennett. Phys. Rev. A 6 (1932) 607.
P. Brunq and hl. Karplus, J. Chem. Phys. 58 (1973) 3903.
T.A. Patterson, H. Hotop, A. Kssdan, D.W. Norcross
and W-C. LinebcrEer, Phys. Rev. Lcttcrs. 32 (1974) 189.
A.J. Herbert, P.J. Lovas, C.A. hlelcndres, C.D. Hoilowell,
T.L. Story Jr. and K. Street Jr., J. Chcm. Phys. 48 (iY68)
zg24.
J.E. Turner. V-E. Anderson and K. Fox, P’nys. Rev. 174
(1968) !74.
R.C. WWS~, ed., Handbook of chemistry and physics
KXemical Rubber Company, Cleveland, 1969) p. 152.
L.A. Curtiss, C.W. Kern and R.L. hlatachcr, J. Chcm. Phys.
63 (1975) 1621.
The authors wish to thank S. Ceftman and K.K.
‘Dockenfor
helpful comments and K.P. Jordan for providing some recent unpubfished
results. This work was
,supported by the National Science Foundation
under
:. contract No. GP-39309X.
Xiferences
[ 1) AS. Wi:ghtman, Phys. Rev. 77 (1950) 52 1.
IT] R.F. %%llis, R. FIerman and H.W. hliines, 3. Mol. Spxrry. 4 (1360) 51.
131 X1.H. Siittleman 2nd V.P. hfvcri-cough, Phys. Letters 23
(1966) 545.
(41 J.6. Turner ;I& K. Fox, Ph>x Ltztters 23 (1966) 547.
,.
‘.
.”
-..,
:
.
‘.
...”
;__)
..
-’
.‘...
;.
‘.
.,
:.
.’
_;
.:
:
‘.
.:._
..’
..
.:,.
.‘l
.‘.
.
.
‘.
__
..
.“.-’
‘.’