A quasiclassical trajectory test for a potential energy surface of the Li+HF reaction

J. M. Alvario, P. Casavecchia, O. Gervasi, and A. Lagan

Citation: The Journal of Chemical Physics 77, 6341 (1982); doi: 10.1063/1.443840
View online: http://dx.doi.org/10.1063/1.443840
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/77/12?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Ab initio potential energy surface, variational transition state theory, and quasiclassical trajectory studies of the F
+ CH 4 HF + CH 3 reaction
J. Chem. Phys. 120, 5181 (2004); 10.1063/1.1637035
Potential energy surface and wave packet calculations on the Li+HF LiF+H reaction
J. Chem. Phys. 106, 1013 (1997); 10.1063/1.473185
The F+HDDF(HF)+H(D) reaction revisited: Quasiclassical trajectory study on an ab initio potential energy
surface and comparison with molecular beam experiments
J. Chem. Phys. 102, 9248 (1995); 10.1063/1.468875
An improvement of the Li+HF PES based on a 3D quasiclassical trajectory test
J. Chem. Phys. 84, 3059 (1986); 10.1063/1.450287
Potential energy surface for the Li+HFLiF+H reaction
J. Chem. Phys. 72, 4376 (1980); 10.1063/1.439728

Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 103.21.125.76 On: Mon, 29 Feb 2016
12:22:05

Letters to the Editor

from the first run, with the boundary sphere slowly
shrunk, and the temperature raised to the final values)
we conclude that the number of ions present is not an
appropriate measure of the populations required for
high conductivity! The prediction lO that the sample
should be essentially all hydroxide and hydronium can
not be verified by our calculations. Neither do we see
bare oxide anions, (02-), nor do we see H40Z+ as speculated upon by Pitzer (as quoted by Hamann ll ).
Room temperature, low density Monte CarlO simulations on 64 water molecule droplets show no dissociation. Recent experiments by Mitchell and Nellis 12
indicate that substantial dissociation does take place in
shocked water, but Ree 13 argues that the intermolecular
potential of water simplifies at temperatures of the
order of 1300 K and above, contrary to our predictions
here. Where ionic dissociation of water into hydronium
and hydroxide occurs to some tiny extent in room temperature 1 atm liquid water, we find an equivalent
hydronium ion concentration of 5.7 mol/f., about seven
orders of magnitude greater than that observed in
normal water.
These computations were performed with the help
of an NIH grant GM-26525 and were materially aided

6341

by substantial grants of computer time by the University

of Connecticut Research Computer Center.
IF. Stillinger and C. W. David, J. Chern. Phys. 69, 1473
(1978).
2F. H. Stillinger, Int. J. Quantum Chern. 14, 649 (1978).
3p J. Turner and C. W. David, J. Chern. Phys. 74, 512
(1981); J. Phys. Chern. 85, 3501 (1981).
4p. Turner, E. E. DaVid, and C. W. DaVid, Chern. Phys. 64,
537 (1982).
'C. W. David, Chern. Phys. Lett. 78, 337 (1981); J. Phys.
Chern. 85, 2433 (1981).
SF. H. Stillinger and T. A. Weber, Chern. Phys. Lett. 79,
259 (1981).
lC. W. David, J. Mol. Struct. 87, 219 (1982).
8E E. David and C. W. DaVid, J. Comput. Chern. 2, 414
(1981)
sH. G. David and S. D. Hamann, Trans. Faraday Soc. 55, 72
(1959); s. D. Hamann and M. Linton, ibid. 62, 2234 (1966).
IDS. D. Hamann (private communication).
liS. D. Hamann, Properties of Electrolyte Solutions at High
Pressures and Temperatures, Proceedings of the 1979 Nobel
Berzelius Symposium, on Chemistry and Geochemistry of
Solutions at High Temperatures (Pergamon, New York, 1981).
12A. C. Mitchell and W. J. Nellis, J. Chem. Phys. 76, 6273
(1982).
13F. H. Ree, J. Chern. Phys. 76, 6287 (1982).

A quasi classical trajectory test for a potential energy

surface of the Li + HF reaction
J. M. Alvarino,a) P. Casavecchia, O. Gervasi, and A. Lagana
Dipartimento di Chimica dell'Universita, 06100 Perugia, Italy
(Received 25 May 1982; accepted 5 August 1982)

Recently, increaSing attention has been devoted to

the reaction Li + HF - LiF + H, which may be regarded
as the prototype of the family of the alkali atom + hydrogen halide reactions. 1 For this reaction, center-ofmass angular distributions T(9) and translational energy
distributions P(EhANS) of the product molecules were
obtained from a crossed molecular beams experiment. 2
The integral reactive cross sections C1R were also evaluated. Values of C1R are 0.80 and 0.94 A2 at reactant
translational energies ETRANS of 3.0 and 8.7 kcal mol- l ,
respectively.2 At the lower energy T(6) has a forwardbackward symmetry whereas at ETRANS =8. 7 kcal mol- l
it becomes forward peaked.
For the LiFH system a semiempirical surface PES I
was reported, which has a saddle to reaction of 12.4
kcal mor l placed in the exit channel. 3 Quantum collinear 4 and three dimensional (3D) quasiclassical trajectory5 calculations were carried out on PES I. The calculated C1R value at ETRANS =15.0 kcal mol- l could not be
compared directly with experimental results obtained
at lower ETRANS' However, the saddle of PES I is too
J. Chern. Phys. 77( 12), 15 Dec. 1982

high to allow the reaction to occur at ETRANS =3. 0

kcal mol- l In addition, the value of C1R and the backward bias of the angular distribution dC1R/dw 2 , calculated at ETRANS =15. 0 kcal mol- l compare poorly with
the experimental results obtained at 8.7 kcal mOrl.
Recently, a functional form 6 PES II was fitted to the
potential energy points calculated USing an ab initio
technique. 7 Theoretical values were corrected at the
saddle region so as to mimic the possible effect of a
configuration interaction. Corrections were also made
in order to reproduce asymptotic properties of the diatomic fragments. The general shape of PES II is
roughly Similar to the one of PES I. Yet, PES II has a
lower barrier and a narrower cone of acceptance for
energies up to 20 kcal mol- l above the reactants' asymptote. 6 8 These two features are of crucial importance
in determining the low energy reaction dynamics. In
fact, as shown by quantum collinear calculations performed by us on PES ns and by other authors on PES I, 4
the height of the saddle is a critical parameter in determining reactivity for reactants in the ground vibrational

0219606/82/24634102$02.10

1982 American I nstitute of Physics

Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 103.21.125.76 On: Mon, 29 Feb 2016
12:22:05

Letters to the Editor

6342

Li + HF .... LiF + H

state. Furthermore, a narrower cone of acceptance is

quite effective in selecting collision paths to reaction.
On PES II we have carried out 3D quasiclassical trajectory calculations using a standard program 9 adapted
to run on a PRIME 550. In order to make a direct comparison with the experimental results we have used the
same translational energies as in Ref. 2 and set the HF
molecule in its ground vibrational state with a rotational
temperature of 60 K. The initial distance of the Li atom
was fixed at 12 A. Remaining parameters were generated randomly with an upper bound of 2 A for the impact
parameter. The results of our calculations indicate that
the system is reactive at ETRANS =3. 0 kcal mol-I, in
agreement with experimental findings, However, the
calculated aR is about one third of the corresponding
experimental value. Such a discrepancy can be due
either to an inadequacy of PES II or to low energy quantum mechanical tunneling or to both. Calculations of
the expected magnitude of tunneling in the Li + HF system would be of interest. We have also found that at
this energy a significant number of trajectories form a
long-lived complex either in the reactants' or in the
products' channel. This is in agreement with the backward-forward T(e) derived from the experiment at
ETRANS =3.0 kcal mol-i.

A more detailed investigation was performed at

ETRANS = 8.7 kcal mol- i by running a batch of 2000 tra-

jectories. A aR value of 1. 03 O. 08 A2 was obtained

in remarkable agreement with the experimental data.
A Significant test for PES II is given by the comparison
of the calculated daR/ dw 2 and daR/ dEhANS distributions
with the experimental T(e) and P(EhANS) curves as
shown in Fig. 1. For daR/ dw 2 final scattering angles
were boxed at every 20 and its value computed as indicated in Ref. 10. The calculated distribution shows
a forward peaking in agreement with the experimental
T(e) within the error bars [see Fig. lea)]. In Fig. l(b)
daR/dE~RANS is compared with the experimental
P(E~RANS)' Final translational energies E~RANS were
boxed at every kcal mol-i. The calculated distribution
is narrower but agrees reasonably well with the experimental P(EhANS) in the high energy side. A discrepancy
occurs in the low energy region. Again, this disagreement can be attributed to tunneling effects, which cannot be accounted for by our quasiclassical calculations,
or to some features of PES II, which need to be refined.
In concluSion, these quasiclassical calculations on
PES II reproduce correctly some experimental results
for this reaction, in particular the integral and differential reactive cross sections at a collision energy of
8.7 kcalmol- i . However, the too low aR calculated at
ETRANS =3.0 kcal mol- i and the small half-width of the
computed translational energy distribution at ETRANS
=8.7 kcal mol- i suggest that a refinement of some feature of PES II and/or of the dynamics is necessary to
achieve better agreement with experiment. Specifically,
the fact that in the crossed beam study of Ref, 2 reaction

ETRANS=

0.2j
~
0.1

(al

'~','

(\13

8.7 kcolmol-I

t<:l-i.-

1.0

-lJ .. f/

G3

0.5f-

----1------ ____ _

1.0 00
z

a::

-f-

05!:::l.
0..

o
FIG. 1. Product angular (a) and energy (b) distributions for
the Li + HF reaction at E TRANS ~ 8.7 kcal morl. Solid circles
are the classical trajectory results of this work (left ordinate)
with 68% confidence intervals. The area between dashed
curves encompasses limits of the experimentally determined
T(e) and P(E rnANs ) (right ordinate) renormalized to the maximum of the calculated distributions.

products were detected even at ETRANS =2. 2 kcal mol- 1 ,

indicates that the barrier height for this PES may be
too high, unless large tunnel effects are operative.
We thank V. Aquilanti for useful discussions. Financial support from C.N.R. (Italy) and C.S.I.C.
(Spain) is gratefully acknowledged.

a)On leave from: Departamento de Quimica Fisica, Facultad de

Ciencias, Universitad del Pais Vasco, 644 Bilbao, Spain.
IF. E. Bartoszek, B. A. Blackwell, J. C. Polanyi, and J. J.
Sloan, J. Chern. Phys. 74,3400 (1981); F. Heismannand
H. J. Loesch, Chern. Phys. 64, 43 (1982).
2C. H. Becker, P. Casavecchia, P. W. Tiedemann, J. J.
Valentini, and Y. T. Lee, J. Chern. Phys. 73,2833 (1980).
3 M. Shapiro and Y. Zeiri, J. Chern. Phys. 70, 5264 (1979).
4R . B. Walker, Y. Zeiri, and M. Shapiro, J. Chern. Phys.
74, 1763 (1981).
5y .

Zeiri, M. Shapiro, and E. pollak, Chern. Phys. 60, 239

(1981).
6S . Carter and J. N. Murrell, Mol. Phys. 41, 567 (1980).
1M. M. L. Chen and H. F. Schaefer III, J. Chern. Phys. 72,

4376 (1980).
M. Aivariiio, O. Gervasi, and A. Lagana, Chern. Phys.
Lett. 87, 254 (1982).
9S. Chapman, D. L. Bunker, and A. Gelb, QCPE No. 273.
IOD. G. Truhlar and J. T. Muckerman in Atom Molecule Collision Theory: A Guide jor the Experimentalist, edited by
R. B. Bernstein (Plenum, New York, 1979), pp. 505-566.
8J

J. Chern. Phys . Vol. 77, No. 12, 15 December 1982

Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 103.21.125.76 On: Mon, 29 Feb
2016 12:22:05