A.D.J. Critchley et al. / Chemical Physics 303 (2004) 235–241
237
3. Results and discussion
for the proposed neutral products are also unknown,
and shown accordingly.
The rate coefficients of the reactions between the 19
cations and PSCl3 are shown in Table 1, together with
the product ions, the branching ratios, and the proposed
neutral products. All the ions studied react with PSCl3
with a rate coefficient greater than 10ꢀ10 cm3 sꢀ1 and,
with the exception of NOþ, all approach the collisional
limit; this latter quantity is shown in square brackets in
column 2 of Table 1. These coefficients, representing
unit efficiency of reaction with capture, were obtained
from modified average dipole orientation calculations
which utilise parameterised fits to results obtained from
trajectory calculations [8]. In the absence of a literature
value, the polarisability of PSCl3 was estimated to be
1.22 ꢂ 10ꢀ29 m3 using the technique of Miller of adding
atomic hybrid components [9], while the value used for
the dipole moment of PSCl3, 1.41 D, was that measured
by Smyth et al. [10]. As found in our previous study of
the ion–molecule reactions of perfluorocarbons [3], the
changes in rate coefficient with recombination energy of
the reactant ion did not always follow the variation of
photoelectron signal with photon energy. This work
suggests that long-range CT may occur provided there is
an energy resonance, but its efficiency is not necessarily
dependent on Franck–Condon factors.
To produce the branching ratios shown in Table 1,
the experiment measured the relative intensities of ions
of different masses produced from the ion–molecule re-
action under consideration. The identities of the likely
neutral products associated with a specific product ion
were deduced via chemical reasoning. Knowledge of the
constituent atoms of the reactant ion and the co-reagent
was sufficient in most cases to make an unambiguous
determination of the ionic products. However, some
ambiguities still remain, particularly in the reactions of
CFþx with PSCl3 since Pþ has the same mass as CFþ, but
thermochemical arguments helped to restrict the num-
ber of possible channels. The channels considered most
plausible for each reaction are listed in Table 1. The
majority of the enthalpies of formation at 298 K used
for the thermochemical calculations were taken from
standard sources [11,12]. The values used for the en-
thalpies of formation of CFx, CFþx and SFx, SFxþ were
taken from Ricca [13] and Bauschlicher and Ricca [14],
respectively, from coupled cluster ab initio calculations.
The enthalpies of formation of PSClþ2 and PCl2þ were
3.1. Reactions of ions with recombination energies below
the onset of ionisation of PSCl3
At low recombination energy, ions are produced that
are formed from bond cleavage and formation in ion–
molecule complexes. The reaction between H3Oþ and
PSCl3 proceeds exclusively via proton transfer to
HPSClþ3 , indicating that the proton affinity of PSCl3 is
greater than that of water. Oþ transfer occurs with NO2þ,
OPSClþ3 being the major product. The reactions of NOþ
and CFþ with PSCl3 warrant some discussion. A signif-
icant amount of NO ꢃ PSClþ3 is formed from the first re-
action, but the major product is PSClþ3 (47%). As the rate
constant observed is less than a quarter of the capture
value, one might expect this reaction to be slightly en-
dothermic, and indeed the energy available from the re-
combination of NOþ is significantly less than the
experimentally-determined ionisation energy of PSCl3. It
has been found, however, that a significant amount of
vibrationally excited NOþ is created in selected ion flow
tubes by electron impact ionisation [17], which may ac-
count for the shortfall in the energy available to the re-
action; the v ¼ 2 level of NOþ X 1Rþ has a recombination
energy of 9.83 eV, making its reaction with PSCl3 just
exothermic. Despite this fact, the majority of the ions
should be in the ground vibrational state, making it dif-
ficult to account for a 47% branching ratio for this
channel. Furthermore, 7% of the products of this reac-
tion are PSClþ2 ions, which is even more endothermic
than production of PSClþ3 . We note that the endo-
thermicities for production of PSClþ2 + ClNO given in
Table 1 are uþpper limits, and they will take lower values if
Df H2098(PSCl2 ) is significantly less than 629 kJ molꢀ1. A
possible mechanism for the formation of PSClþ3 and
PSClþ2 is via collisionally-induced dissociation of
NO ꢃ PSClþ3 as it accelerates into the detection region.
Unlike NOþ, the major difficulty in interpreting the
CFþ data is that of product assignment. The major
product of the reaction of CFþ with PSCl3, with a yield of
64%, has mass 63 u. Its identity could be CFSþ or PSþ.
The production of CFSþ requires the breaking of one
P@S bond, whilst that of PSþ needs three P–Cl bonds to
break. The reaction CFþ + PSCl3 ! CFSþ + PCl3 is
exothermic, provided Df H2098(CFSþ) is less than 1037
kJ molꢀ1. However, the PSþ channel is also exothermic,
since the reaction CFþ + PSCl3 ! PSþ + CFCl3 has an
enthalpy of reaction, DrH2098 ¼ ꢀ13 kJ molꢀ1. Evidence
from other reactions illustrates that the three P–Cl bonds
may be broken. Thus, Nþ unambiguously produces PSþ,
admittedly at a low level of 1%, via the same chemical
mechanism (Nþ + PSCl3 ! PSþ + NCl3; DrH2098 ¼ ꢀ203
kJ molꢀ1). However, the reaction also produces NSþ at
the 5% level via the same potential mechanism for pro-
determined from the appearance energy at 298 K (AE298
)
for these ions from PSCl3 and PCl3 in complementary
TPEPICO experiments [2,15], allowance being made for
the small correction needed to convert AE298 of a frag-
ment ion into DrH2098 for the unimolecular reaction [16].
Upper limit enthalpies of formation of 629 (PSClþ2 ) and
722 (PClþ2 ) kJ molꢀ1 were obtained [2]. Values for
Df H2098 of PClþ and PSClþ are unknown, and shown as
such in column 5 of Table 1. Many of the Df H2098 values
Critchley, Andrew D.J.
