A study of the gas phase reactions of various cations with two derivatives of SF6; SF5CF3 and SF5Cl

View Article Online / Journal Homepage / Table of Contents for this issue

A study of the gas phase reactions of various cations with two

derivatives of SF₆; SF₅CF₃and SF₅Cl

Clair Atterbury,^aAndrew D. J. Critchley,^bRichard A. Kennedy,*^aChris A. Mayhew*^band

Richard P. Tuckett^a

a

School of Chemical Sciences, University of Birmingham, Edgbaston, Birmingham, UK B15 2TT

School of Physics and Astronomy, University of Birmingham, Edgbaston, Birmingham, UK B15 2TT

b

Received 19th October 2001, Accepted 19th February 2002

First published as an Advance Article on the web 2nd May 2002

A selected ion ﬂow tube apparatus was used to investigate the positive ion chemistry of two derivatives of SF₆;

SF₅CF₃and SF₅Cl. This represents the ﬁrst investigation of the positive ion chemistry of SF₅Cl, and much of

the data on SF₅CF₃is being presented here for the ﬁrst time. Rate coeﬃcients and ion product branching ratios

have been determined at room temperature (300 K) for reactions with the following twenty-two cations; Ne⁺,

F⁺, Ar⁺, N₂⁺, N⁺, CO⁺, CO₂⁺, O⁺, N₂O⁺, H₂O⁺, O₂⁺, SF₄⁺, CF₂⁺, SF⁺, SF₂⁺, NO₂⁺, SF₅⁺, NO⁺, CF⁺,

CF₃⁺, SF₃⁺, and H₃O⁺(listed in order of decreasing recombination energy). Comparisons are made in the text

for the reactions of these ions with SF₆. SF₂⁺, NO₂⁺, NO⁺, SF₃⁺, and H₃O⁺are found to be unreactive with

both of the derivatives. The majority of the other reactions proceed with rate coeﬃcients which are close to the

+

capture value. The exceptions are the reactions of O₂⁺, SF⁺, SF₅⁺, and CF₃⁺with SF₅CF₃, and SF₄⁺and SF₅

with SF₅Cl, all of which have rate coeﬃcients signiﬁcantly less than the capture mechanism value. Several

distinct processes are observed among the large number of reactions studied, including dissociative charge

transfer and various abstraction channels. Non-dissociative charge transfer is not observed, implying that any

parent ions formed dissociate rapidly to the fragment ions and associated neutrals.

gas-phase study of the reactions of SF₅CF₃with several posi-

1. Introduction

tive ions of atmospheric interest.¹²A more detailed investiga-

tion of the reactions of a larger number of positive ions with

this molecule is presented here. This includes cations of impor-

There have been a number of investigations dealing with the

positive ion chemistry of the important molecule sulfur hexa-

ﬂuoride, SF₆.^1–8These have been mostly associated with stu-

dies of the fundamental ion-molecule chemistry. Recent

interest in this chemistry has come from the use of SF₆in reac-

tive plasmas for the etching of insulating and semiconductor

layers.⁷Studies of the reactions of ions with SF₆and related

molecules enhance our understanding of the complex ion

chemistry and kinetics involved in SF₆containing plasmas.

For example, the data obtained could be used to help model

plasmas in order to ascertain which physical and chemical

parameters will need to be varied in order to optimise the

plasma environment for a particular application. In this paper

we present a study of the reactions of a series of positive ions

with two derivatives of SF₆; triﬂuoromethyl sulfur pentaﬂuor-

ide, SF₅CF₃, and sulfur chloropentaﬂuoride, SF₅Cl, and com-

pare them with the corresponding reactions for SF₆.

Whilst the major stimulus for this investigation comes from

its application to plasma ion chemistry, the recent detection of

SF₅CF₃in the stratosphere provides additional motivation to

this study.⁹As found for SF₆, SF₅CF₃is not broken down

by UV photodissociation in the stratosphere.¹⁰This leads

naturally to the question as to whether there are any other

atmospheric loss mechanisms. A recent electron attachment

study shows that the molecule will be destroyed in the regions

of the ionosphere where free electrons exist.¹¹This provides an

upper limit to the atmospheric lifetime of SF₅CF₃of ca. 1000

years. Other competitive routes in the ionosphere leading to

the possible destruction of SF₅CF₃are reactions with ions.

These may play a major role in the atmospheric lifetime of

SF₅CF₃and hence have an impact on its global warming

potential, which is currently estimated to be 18 000 times that

of CO₂.⁹A preliminary communication by us reported the ﬁrst

+

tance to industrial plasmas, e.g. F⁺, Ar⁺, SF₅⁺, CF⁺, CF₃

,

and SF₃⁺. In addition, this paper presents the ﬁrst reported

ion-molecule chemistry of SF₅Cl.

Rate coeﬃcients and product ion branching ratios for ther-

mal (300 K) reactions of SF₅CF₃and SF₅Cl with the following

cations of fundamental, atmospheric and/or plasma proces-

sing interest are reported: Ne⁺, F⁺, Ar⁺, N₂⁺, N⁺, CO⁺,

+

,

CO₂⁺, O⁺, N₂O⁺, H₂O⁺, O₂⁺, SF₄⁺, CF₂⁺, SF⁺, SF₂

NO₂⁺, SF₅⁺, NO⁺, CF⁺, CF₃⁺, SF₃⁺, and H₃O⁺. These ions

cover a large range of recombination energies, from 21.56 eV

(Ne⁺+ e^ꢀ! Ne) to 6.27 eV (H₃O⁺+ e^ꢀ! H₂O + H). (1

eV ¼ 96.4845 kJ mol^ꢀ1.) Owing to this large range of available

energies, several distinct reaction processes are observed

amongst the large number of reactions studied, including dis-

sociative charge transfer and various abstraction processes.

For completeness, and ease of comparison, the results of the

reactions of the above ions with SF₆collected from various

published papers are also presented here. Where results for

SF₆are unavailable in the literature, additional experiments

have been performed as part of this investigation, the data

from which are presented here.

2. Experimental

The selected ion ﬂow tube (SIFT) apparatus used to obtain the

reaction rate coeﬃcients and product ion branching ratios has

been described in detail elsewhere,^13,14so only a brief overview

of the salient points of the experimental technique is included

here. The reagent ions were generated in an electron impact

high-pressure ion source containing an appropriate gas

2206

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

This journal is # The Owner Societies 2002

DOI: 10.1039/b109567f

View Article Online

(Ne for Ne⁺, CF₄for F⁺and CF_m⁺(m ¼ 1–3), Ar for Ar⁺, N₂

states. As for Ar⁺and Ne⁺, no diﬀerences in the reactivity

of the F⁺ion in its various spin–orbit states are to be expected.

For the other two atomic ions used in this investigation, elec-

tronically excited states of N⁺and O⁺have previously been

shown not to be present.¹⁶

+

for N₂and N⁺, CO for CO⁺, CO₂for CO₂⁺, N₂O for O⁺,

NO⁺and N₂O⁺, H₂O for H₂O⁺and H₃O⁺, O₂for O₂⁺, SF₆

+

for SF_n(n ¼ 1–5), and NO₂for NO₂⁺). The reagent ions

were mass selected using a quadrupole mass ﬁlter, before being

injected into a 300 K helium carrier gas at a pressure of ca. 0.5

Torr. The helium gas (high-purity grade (99.997%)) was passed

through a liquid nitrogen cooled zeolite trap before use. The

reagent ions were transported along the ﬂow tube and detected

by a downstream quadrupole mass spectrometer detection sys-

tem. Neutral reactant gas was added in controlled amounts to

the ion swarm/carrier gas.

Water contamination in the ﬂow tube and He buﬀer gas ﬂow

resulted in electron transfer from H₂O to those injected ions

whose recombination energies are greater than the ionisation

potential of H₂O, 12.62 eV. This resulted in an H₂O⁺signal,

which was usually less than 5% of the parent ion signal. Pro-

duct ion branching ratios have been corrected for the contribu-

tion to the signals resulting from the reaction with H₂O⁺.

Some of the H₂O⁺was converted to H₃O⁺in the ﬂow tube

(H₂O⁺+ H₂O ! H₃O⁺+ OH). This caused no problem in

the determination of the product ion branching ratios, because

H₃O⁺is unreactive with SF₅CF₃, SF₅Cl, and SF₆. Experi-

ments with H₂O⁺as the reagent ion presented a separate pro-

blem, for it was found that OH⁺and H₃O⁺ions were being

injected into the ﬂow tube together with H₂O⁺. Again, because

the H₃O⁺ions are unreactive with both SF₅CF₃and SF₅Cl,

their presence in the ﬂow tube could be ignored. However, it

was observed from the decrease in the OH⁺signal as a func-

tion of neutral concentration that this ion reacts eﬃciently

with SF₅CF₃and SF₅Cl. Reduction of the injection ion energy,

which insured that the OH⁺was not produced by collision-

induced dissociation, did not remove the problem. Attempts

to eliminate the OH⁺signal by increasing the resolution of

the injection mass spectrometer drastically reduced the H₂O⁺

signal. The best compromise between OH⁺rejection and

H₂O⁺signal gave a reagent ion ﬂux comprised of 5% OH⁺

and 95% H₂O⁺. No allowances have been made for the (small)

contributions due to the OH⁺reaction to the branching ratios

for the reaction of H₂O⁺with either SF₅CF₃or SF₅Cl. Reac-

tion rate coeﬃcients and product ions for the reaction with

OH⁺were not obtained. OH⁺was also produced in the ﬂow

tube when Ne⁺was used as the reagent ion. Again, the OH⁺

signal was small (approximately 5% of that of Ne⁺) and was

ignored in the determination of product ion branching ratios.

Ne⁺also resulted in the formation of HeNe⁺cations in the

ﬂow tube from termolecular association, which led to a HeNe⁺

signal of about 2%–3% of the Ne⁺signal. Given the low per-

centage of this impurity ion in the ﬂow tube, its contributions

to the yields of the product ions for Ne⁺were reasonably

neglected.

The loss of the reagent ions and the appearance of product

ions were monitored by the detection system. Relaxed

resolution on the quadrupole mass spectrometer, in order to

minimize mass discrimination eﬀects, was used in the determi-

nation of the branching ratios. The reaction rate coeﬃcients

and ion product distributions were then determined in the

usual way.^13–15Reaction rate coeﬃcients, which were extracted

from linear least-squares ﬁts of the plot of the logarithm of the

reagent ion signal versus reactant neutral concentration, are

considered to be accurate to ꢁ20%. The percentage product

ion branching ratios, determined from the product ion counts

as a function of reactant neutral concentration and extrapo-

lated to zero ﬂow conditions to allow for secondary reactions,

are only used to provide a qualitative indication of the relative

importance of diﬀerent channels, and therefore accurate values

are not important in our discussions. However, a statistical

analysis on the branching ratios indicates a fractional accuracy

of better than 20% for the reaction channels with branching

ratios greater than 10%. For ion channels with branching

ratios less than this value, the accuracy is reduced. For the

ion channels with branching ratios of only 1%, the uncertainty

is as high as ꢁ100%.

The high pressure of the gases used in the ionisation source,

ca. several Torr, should help to quench electronically and/or

vibrationally excited states of molecular ions prior to their

injection into the ﬂow tube. However, quenching of vibrational

+

states is not assured, as conﬁrmed by studies on the O₂

reagent ion. Whilst the majority of the O₂ions were found

+

to be in the v ¼ 0 level, approximately 30% were found to be

in the v ¼ 1 and 2 vibrational levels. This was established from

observed reactions with Xe and SO₂. The lack of a reaction of

+

the reagent ion O₂with H₂O showed that the v ꢂ 3 vibra-

+

tional levels of O₂were not populated. Similarly, for the

The samples of the reactant neutrals, SF₅CF₃and SF₅Cl,

were commercially purchased (Apollo Scientiﬁc Limited, Wha-

ley Bridge, Derbyshire, UK) with stated purities of 99% and

97%, respectively. They were used in the experiments without

further puriﬁcation. The SF₆was purchased from Fluorochem

Limited, Old Glossop, Derbyshire, UK with a stated purity of

>99%. This sample was also used directly without additional

puriﬁcation.

+

N₂reagent ions, reaction with Ar indicated that approxi-

mately 40% of the N₂⁺ions were in the v ¼ 1 vibrational level.

It is possible that other reagent molecular ions had internal

energies above the thermal value. However, no curvature

was observed in any of the pseudo-ﬁrst-order kinetic plots,

indicating that rate coeﬃcients are the same for reactions

involving the ground and any vibrationally excited states. This

does not rule out reagent ion vibrational excitation inﬂuencing

the ion product distributions for those reactions which pro-

duce more than one ion product.

3. Results and discussion

Atomic ions are expected to be in their ground electronic

states prior to reaction with a neutral molecule. Thus Ar⁺

and Ne⁺should emerge from the high-pressure ion source in

their lowest energy, ²P_3/2states. Even if this were not the case,

the energy separation between the ²P_3/2and ²P_1/2states is only

0.18 eV for Ar⁺and 0.10 eV for Ne⁺, and therefore diﬀerences

in reactivities between the two spin–orbit states are not

expected. No evidence was found for diﬀerences in the reaction

The experimental rate coeﬃcients, k_exp, the observed product

ions, and their branching ratios for the reactions of SF₅CF₃,

SF₅Cl and SF₆with 22 diﬀerent reagent ions are given in

Tables 1–3, respectively. The reagent ions are listed (top to

bottom) in order of decreasing recombination energy (RE).

Where no reaction was observed, this implies that the experi-

mental rate coeﬃcient, k_exp< 10^ꢀ13cm³molecule^{ꢀ1 ꢀ1}. The

s

2

rate coeﬃcients of the P_3/2and P_1/2states. Whether diﬀer-

ences in branching ratios result from reactions involving these

two spin–orbit states is more diﬃcult to assess, but given the

small diﬀerences in energy, none are expected. The ground

state of F⁺is a closely-spaced triplet with recombination ener-

gies of 17.42 eV (³P₂), 17.47 eV (³P₁) and 17.48 eV (³P₀). Thus,

for this ion, there will be signiﬁcant population in all three

majority of the SF₆data presented in Table 3 has been col-

lected from the literature,^1–8and are given here to aid compar-

ison. In the cases of the reactions with SF₆for which no

published data exist, we have made the appropriate measure-

ments to complete Table 3.

To assess the eﬃciency of the reactions, the experimental

rate coeﬃcients are compared in the tables with the predictions

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2207

View Article Online

Table 1 The measured 300 K reaction rate coeﬃcients (in units of 10^ꢀ9cm³molecule^ꢀ1s^ꢀ1), cation product ions, their branching ratios (percen-

tages), suggested neutral products and 0 K enthalpies for the reactions of Ne⁺, F⁺, Ar⁺, N₂⁺, N⁺, CO⁺, CO₂⁺, O⁺, N₂O⁺, H₂O⁺, O₂⁺, SF₄

,

+

CF₂⁺, SF⁺, SF₂⁺, NO₂⁺, SF₅⁺, NO⁺, CF⁺, CF₃⁺, SF₃and H₃O⁺with SF₅CF₃are presented. The recombination energies of the reagent

cations are listed in eV in round brackets under the cations. The value in square brackets is the ﬂuoride aﬃnity of the reagent cation. The capture

rate coeﬃcients, calculated using parametrized ﬁts to results from trajectory calculations,¹⁸are presented in square brackets under the experimental

values. The measured rate coeﬃcients are considered to be accurate to ꢁ20%. When dissociative charge transfer is exothermic, only this channel is

represented in the table, with the neutrals assembled to give the lowest reaction enthalpy. The trace given for the SF₄⁺product ion for a number of

reactions and for SF₄CF₃⁺from the reaction with CF₂⁺represents a detectable signal with our apparatus, but whose branching ratio is less than

+

24

+ 25

1%. 0 K enthalpies of formation for CF_m⁺, CF_m

,

and for SF_n, SF_n

are taken from recent high level ab initio calculations.

D_fH^o₀(SF₅CF₃) ¼ ꢀ1673 kJ mol^ꢀ1(see text). Other 0 K enthalpy data are taken from the JANAF tabulations,²⁷information available at the NIST

website,⁵¹and additional sources of data as identiﬁed in the table footnotes. The major source of uncertainty in the values is ꢁ40 kJ mol^ꢀ1in

D_fH^o₀(SF₅CF₃)

Reagent ion

Rate coeﬃcient/10^ꢀ9cm³molecule^ꢀ1s^ꢀ1

Product ion (%)

Proposed neutral products

D_rH^o₀/kJ mol^ꢀ1

Ne⁺

1.5

[1.8]

SF₅+ Ne

ꢀ836

+

CF₃(47)

+

SF₂(5)

(21.56)

[1753]

CF₄+ F₂+ Ne

CF₄+ F + Ne

CF₄+ Ne

ꢀ645

+

SF₃(37)

ꢀ898

+

SF₄(1)

ꢀ946

+

SF₅(10)

CF₃+ Ne

SF₅+ F

ꢀ786

F⁺

1.6

[1.9]

CF₃(73)

+

SF₃(24)

ꢀ436

+

(17.42)

[1508]

CF₄+ F + F

CF₄+ F

CF₃+ F

ꢀ498

+

SF₄(trace)

ꢀ546

+

SF₅(3)

ꢀ386

Ar⁺

1.4

[1.3]

CF₃(65)

+

SF₃(30)

SF₅+ Ar

ꢀ276

+

(15.76)

[1193]

CF₄+ F + Ar

CF₃+ Ar

F + Ar

ꢀ338

+

SF₅(4)

ꢀ226

D_fH^o₀(SF₄CF₃⁺) + 230

+

SF₄CF₃(1)

+

CF₃(65)

N₂

(15.58)

[1175]

1.6

[1.6]

SF₅+ N₂

ꢀ259

+

SF₃(28)

CF₄+ F + N₂

CF₄+ N₂

CF₃+ N₂

ꢀ321

+

SF₄(trace)

ꢀ368

+

SF₅(5)

ꢀ209

+

SF₄CF₃(2)

F + N₂

SF₅+ N

D_fH^o₀(SF₄CF₃⁺) + 247

N⁺

2.2

[2.1]

CF₃(80)

+

SF₃(17)

ꢀ158

ꢀ220

ꢀ267

ꢀ108

ꢀ107

ꢀ169

ꢀ217

ꢀ57

+

(14.53)

[1374]

CF₄+ F + N

CF₄+ N

CF₃+ N

+

SF₄(trace)

+

SF₅(3)

CO⁺

1.6

[1.6]

CF₃(75)

+

SF₃(22)

SF₅+ CO

CF₄+ F + CO

CF₄+ CO

CF₃+ CO

SF₅+ CO₂

CF₄+ F + CO₂

CF₄+ CO₂

CF₃+ CO₂

SF₅+ O

+

(14.01)

[1173]^a

+

SF₄(trace)

+

SF₅(3)

+

CF₃(76)

CO₂

(13.78)

[1049]^b

1.2

[1.3]

ꢀ85

+

SF₃(14)

ꢀ147

ꢀ194

ꢀ35

+

SF₄(8)

+

SF₅(2)

O⁺

1.9

[2.0]

CF₃(83)

+

SF₃(16)

ꢀ69

+

(13.62)

[1202]

CF₄+ F + O

CF₄+ O

CF₃+ O

ꢀ131

ꢀ179

ꢀ19

+

SF₄(trace)

+

SF₅(1)

N₂O⁺

(12.89)

[971]

1.1

[1.3]

CF₃(75)

+

SF₃(19)

SF₅+ N₂O

CF₄+ F + N₂O

CF₄+ N₂O

CF₃O + N₂

CF₂O + N₂+ F

SF₅+ H₂O

CF₄+ F + H₂O

CF₄+ HF + OH

CF₄+ H₂O

SF₅+ O₂

+1

+

ꢀ61

+

SF₄(5)

ꢀ109

ꢀ203^c

ꢀ98^d

+27

ꢀ35

+

SF₅(1)

H₂O⁺

(12.62)

[962]

1.6

[1.9]

+

CF₃(8)

+

SF₃(92)

ꢀ108

ꢀ83

+

SF₄(trace)

+

CF₃(63)

O₂

(12.07)

[887]

0.01

[1.5]

+80, +58, +36^e

29^f

18

+

SF₅O₂

+

SF₃(31)

CF₄+ F + O₂

CF₄+ FO₂

CF₄+ O₂

ꢀ32

+

SF₄(2)

+

SF₅(4)

ꢀ29

CF₃+ O₂

CF₃O₂

CF₄+ SF₅

130

ꢀ33^g

ꢀ121

+

SF₃(100%)

SF₄

0.17

[0.92]

(11.92)

[976]

+

CF₂

+

CF₃(38)

1.1

[1.2]

SF₄+ CF₃

SF₃+ CF₄

CF₃+ CF₄

C₂F₅

ꢀ52

(11.43)

[1122]

ꢀ193

+

SF₃(60)

ꢀ267

+

SF₅(2)

+

SF₄CF₃(trace)

ꢀ49^h

CF₃

D_fH^o₀(SF₄CF₃⁺) + 301

2208

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

Table 1 (continued)

Reagent ion

Rate coeﬃcient/10^ꢀ9cm³molecule^ꢀ1s^ꢀ1

Product ion (%)

Proposed neutral products

D_rH^o₀/kJ mol^ꢀ1

+

CF₃(44)

+

SF₃(56)

SF⁺

0.03

[1.2]

SF₂+ SF₄

CF₄+ SF₂

+36

ꢀ179

(10.25)

[1034]

+

SF₂

no reaction

(10.17)

[876]

+

NO₂

(9.59)

[607]

+

SF₅

+

CF₃(13)

0.14

[0.88]

SF₄+ SF₆

SF₆+ CF₄

+30

ꢀ185

+

SF₃(87)

(9.55)

[1040]

NO⁺

(9.26)

[797]

no reaction

CF⁺

1.2

[1.5]

CF₄+ SF₂

CF₂+ SF₄

CF₂+ CF₄

C₂F₆

ꢀ261

+4

+

CF₃(47)

(9.10)

[1066]

+

SF₃(51)

ꢀ211

ꢀ431ⁱ

ꢀ40^j

ꢀ230

+8ⁱ

+

SF₅(2)

C₂F₄

CF₄+ CF₄

C₂F₆

+

SF₃(98)

CF₃

0.48

[1.1]

+

SF₅(2)

(9.03)

[1085]

+

SF₃

no reaction

(8.26)

[870]

H₃O⁺

(6.27)

[843]

a

D_fH^o₀(FCO) ¼ ꢀ185 kJ mol^ꢀ1from Dixon_ꢀa₁nd Feller.^{52 b}D₀(F–CO₂) ¼ 48 kJ mol^ꢀ1from Arnold et al.^{39 c}D_fH₀^o(CF₃O) ¼ ꢀ630 kJ mol^ꢀ1from

Louis et al.^{53 d}D_fH^o₀(CF₂O) ¼ ꢀ603 kJ mol from Louis et al.^{53 e}The values given are for the v ¼ 0, 1 and 2 vibrational levels of O₂⁺, respec-

+

tively. Energy levels calculated using O₂ground state vibrational spectroscopic constants from Huber and Herzberg.^{54 f}D₀(SF₅–O₂) ¼ 51 kJ

295

mol^ꢀ1estimated from D_rH^o(SF₅O₂! SF₅+ O₂) ¼ 57.3 kJ mol^ꢀ1obtained by Sehested et al.⁵⁵D₀(CF₃–O₂) ¼ 163 kJ mol^ꢀ1from Louis

g

et al.⁵³D_fH^o₀(CF₃CF₂) ¼ ꢀ896 kJ mol^ꢀ1from Bausclicher and Ricca.⁴⁷D_fH^o₀(CF₃CF₃) ¼ ꢀ1341 kJ mol^ꢀ1from Bausclicher and Ricca.⁴⁷

h

i

j

D_fH₀^o(CF₂CF₂) ¼ ꢀ673 kJ mol^ꢀ1from Bausclicher and Ricca.⁴⁷

of capture theory, k_c. Essential pieces of information for mak-

ing capture theory calculations of k_care the polarisability and

the dipole moment of the neutral reactant molecule. Neither of

these quantities are available in the literature for SF₅CF₃and

SF₅Cl. The polarisabilities of SF₅CF₃and SF₅Cl have been

estimated to be 8.44 ꢃ 10^ꢀ30m³and 8.38 ꢃ 10^ꢀ30m³, respec-

tively, using Miller and Savchik’s empirical approach.¹⁷At

the Hartree–Fock level, the dipole moments for SF₅CF₃and

SF₅Cl have been calculated to be 0.95 and 0.93 D, respectively.

Given these values the capture rate coeﬃcients can be esti-

mated using the results of parameterised ﬁts to trajectory cal-

culations,¹⁸and these k_cvalues appear in square brackets in

Tables 1 and 2. A comparison of the experimental rate coeﬃ-

cients with the capture values provides a good indication of

whether a reaction is eﬃcient, with most captures leading to

reaction, or ineﬃcient. The capture rate coeﬃcients are readily

determined for the centrosymmetric molecule SF₆using the

Langevin equation,¹⁹and a value for the polarisability of

SF₆( ¼ 6.54 ꢃ 10^ꢀ30m^{3 20}). These k_cvalues are presented in

Table 3.

The determination of reaction pathways requires knowledge

of the ion and neutral products. In our experiments, only the

masses and relative yields of the ion products are determined,

so the neutral products can only be inferred from mass balance

and thermochemical arguments. The thermochemical analysis

is restricted to considerations of enthalpy at 0 K (i.e. D_rH^o₀< 0

0 for an observable reaction), and we have ignored any entro-

pic eﬀects. In the tables we have generally listed the neutrals

which give the most exothermic pathway to a particular

observed ion product, so that the maximum energy is available

for partitioning between the internal and translational degrees

of freedom. In many cases there is suﬃcient energy to break

chemical bonds in the neutral products.

3.1 SF₅CF₃reactions

An investigation of the reactions of a number of positive ions

of atmospheric interest (N₂⁺, N⁺, CO⁺, CO₂⁺, O⁺, N₂O⁺,

H₂O⁺, O₂⁺, NO₂⁺, NO⁺, and H₃O⁺) with SF₅CF₃has

recently been reported by us.¹²A more extensive study, which

includes many additional ions, i.e. Ne⁺, F⁺, Ar⁺, SF₄⁺, CF₂

,

+

SF⁺, SF₂⁺, SF₅⁺, CF⁺, CF₃⁺, and SF₃⁺, is presented here.

The experimental reaction rate coeﬃcients, the observed pro-

duct ions, and their percentage branching ratios are given in

Table 1.

Dissociative charge transfer is almost certainly the dominant

reaction mechanism for reagent ions whose recombination

energies are greater than about 12.5 eV. The overall process

can be viewed as a two step mechanism, transfer of an electron

from the reactant neutral to the reagent ion (often described as

occurring at long range), followed by fragmentation of the

unstable intermediate ion, here (SF₅CF₃⁺)*, with the newly

formed neutral acting as a spectator. Simple dissociations

such as:

ꢄ

ðSF₅CF₃^þÞ ! CF₃^þþ SF₅

ð1Þ

ð2Þ

! SF₅^þþ CF₃

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2209

View Article Online

Table 2 The measured 300 K reaction rate coeﬃcients (in units of 10^ꢀ9cm³molecule^ꢀ1s^ꢀ1), cation product ions, their branching ratios (percen-

tages), suggested neutral products and 0 K enthalpies for the reactions of Ne⁺, F⁺, Ar⁺, N₂⁺, N⁺, CO⁺, CO₂⁺, O⁺, N₂O⁺, H₂O⁺, O₂⁺, SF₄

+

,

CF₂⁺, SF⁺, SF₂⁺, NO₂⁺, SF₅⁺, NO⁺, CF⁺, CF₃⁺, SF₃⁺and H₃O⁺with SF₅Cl are presented. The recombination energies of the reagent cations

are listed in eV in brackets under the cations. The value in square brackets is the ﬂuoride aﬃnity of the reagent cation. The capture rate coeﬃcients,

calculated using parametrized ﬁts to results from trajectory calculations,¹⁸are presented in square brackets under the experimental values. The

measured rate coeﬃcients are considered to be accurate to ꢁ20%. It is assumed that ClF₂and ClF₃are not formed in dissociative charge-transfer

reactions. When dissociative charge transfer is exothermic, only this channel is represented in the table, with the neutrals assembled to give the

+

lowest reaction enthalpy. The trace given for the SF₃and SF₂Cl⁺product ions resulting from the reaction with CO₂represents a detectable

24

+25

signal with our apparatus, but whose branching ratio is less than 1%. 0 K enthalpies of formation for CF_m⁺, CF_m

,

and for SF_n, SF_n

are

taken from recent high level ab initio calculations. D_fH^o₀(SF₅CF₃) ¼ ꢀ985 kJ mol^ꢀ1(see text). Other 0 K enthalpy data are taken from the JANAF

tabulations,²⁷information available at the NIST website,⁵¹and additional sources of data as identiﬁed in the table footnotes

Reagent ion

Rate coeﬃcient/10^ꢀ9cm³molecule^ꢀ1s^ꢀ1

Product ion (%)

Cl⁺(4)

Proposed neutral products

D_fH^o₀/kJ mol^{ꢀ1 a}

Ne⁺

1.7

[1.8]

SF₅+ Ne

ClF + F₂+ Ne

ClF + F + Ne

ClF + Ne

Cl + Ne

ꢀ562

ꢀ457

ꢀ709

ꢀ757

ꢀ892

ꢀ868^a

ꢀ309

ꢀ357

ꢀ492

ꢀ623^a

ꢀ332

ꢀ308^a

+

SF₂(trace)

(21.56)

[1753]

+

SF₃(86)

+

SF₄(1)

+

SF₅(5)

SF₄Cl⁺(4)

F + Ne

F⁺

1.9

[1.9]

SF₃(19)

+

SF₄(30)

ClF + F + F

ClF + F

Cl + F

+

(17.42)

[1508]

+

SF₅(19)

SF₄Cl⁺(32)

F + F

Cl + Ar

F + Ar

Ar⁺

1.3

[1.4]

SF₅(47)

SF₄Cl⁺(53)

+

(15.76)

[1193]

+

N₂

+

SF₅(52)

SF₄Cl⁺(48)

1.6

[1.6]

Cl + N₂

F + N₂

ꢀ315

(15.58)

[1175]

N⁺

ꢀ291^a

+

SF₃(3)

1.9

[2.1]

ClF + F + N

Cl + N

F + N

ꢀ31

+

SF₅(77)

(14.53)

[1374]

CO⁺

ꢀ214

ꢀ190^a

ꢀ163

ꢀ140^a

SF₄Cl⁺(20)

+

SF₅(50)

1.6

[1.6]

Cl + CO

F + CO

(14.01)

[1173]

+

CO₂

SF₄Cl⁺(50)

1.2

[1.3]

SFCl⁺(5)

CF₄+ O₂

ClF₂+ CO₂

CF₂O + FO

CF₃O₂

D_fH^o₀(SFCl⁺)^cꢀ 878

(13.78)

[1049]^b

ꢀ90^d

+

SF₃(trace)

SF₂Cl⁺(trace)

D_fH^o₀(SF₂Cl⁺)^eꢀ 446^f

D_fH^o₀(SF₂Cl⁺)^eꢀ 591^g

+

SF₅(93)

Cl + CO₂

F + CO₂

Cl + O

ꢀ141

ꢀ117^a

ꢀ125

ꢀ102^a

SF₄Cl⁺(2)

O⁺

2.0

[2.0]

SF₅(84)

SF₄Cl⁺(16)

+

(13.62)

[1202]

N₂O⁺

(12.89)

[971]

H₂O⁺

(12.62)

[962]

F + O

+

SF₅(98)

SF₄Cl⁺(2)

1.2

[1.3]

Cl + N₂O

F + N₂O

ꢀ55

ꢀ31

1.7

[1.9]

SF₂OH⁺(4)

ClF₂+ FH

2 HF + ClO

2 HF + FO

Cl + H₂O

FH + OH

ClO₂

D_fH^o₀(SF₂OH^{+ h}

)

ꢀ 331^d

+

SF₃(2)

ꢀ78

SF₂Cl⁺(6)

D_fH^o₀(SF₂Cl⁺)^eꢀ 431

+

SF₅(35)

ꢀ29

ꢀ78^a

3

SF₄Cl⁺(53)

+

O₂

+

SF₅(97)

1.2

[1.5]

(12.07)

[887]

+

SF₄

SF₄Cl⁺(3)

FO₂

ꢀ3^a

0.3

[1.0]

SF₄Cl⁺(100)

SF₅

ꢀ91^a

(11.92)

[976]

+

CF₂

1.2

[1.2]

CF₂Cl⁺(17)

SF₅

CF₂Cl

CF₃

ꢀ206

ꢀ116

ꢀ237^a

ꢀ77ⁱ

+

SF₅(14)

(11.43)

[1122]

SF⁺

SF₄Cl⁺(69)

+

SF₅(93)

0.8

[1.2]

ClSF

SF₂

(10.25)

[1034]

+

SF₂

SF₄Cl⁺(7)

ꢀ149^a

no reaction

(10.17)

[876]

+

NO₂

(9.59)

[607]

+

SF₅

0.08

[0.9]

SF₄Cl⁺(100)

SF₆

ꢀ156^a

(9.55)

2210

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

Table 2 (continued)

Reagent ion

Rate coeﬃcient/10^ꢀ9cm³molecule^ꢀ1s^ꢀ1

Product ion (%)

Proposed neutral products

D_fH^o₀/kJ mol^{ꢀ1 a}

[1040]

NO⁺

(9.26)

[797]

no reaction

CF⁺

1.5

[1.5]

CF₂Cl⁺(3)

SF₄

ꢀ343

ꢀ482

ꢀ22

+

SF₃(3)

(9.10)

[1066]

CClF₃

ClF + CF₂

CClF

CF₂

+

SF₅(27)

ꢀ74

SF₄Cl⁺(67)

ꢀ181^a

ꢀ43

+

SF₅(39)

CF₃

0.7

[1.1]

CClF₃

CF₄

(9.03)

[1085]

+

SF₃

SF₄Cl⁺(61)

ꢀ201^a

no reaction

(8.26)

[870]

H₃O⁺

(6.27)

[843]

a

b

D_fH₀^o(SF₄Cl⁺) ¼ 150 kJ mol^ꢀ1, see text. D₀(F–CO₂) ¼ 48 kJ mol^ꢀ1from Arnold et al.^{39 c}D_fH^o₀(SFCl⁺) ꢅ 782 kJ mol^ꢀ1, estimate, see text.

d

D_fH^o₀(ClF₂) ¼ ꢀ105 kJ mol^ꢀ1, from Lias et al.^{56 e}D_fH^o₀(SF₂Cl⁺) ꢅ 453 kJ mol^ꢀ1, estimate, see text. ^fD_fH₀^o(CF₂O) ¼ ꢀ603 kJ mol^ꢀ1from Louis

et al.^{53 g}D_fH^o(CF₃O₂) ¼ ꢀ640 kJ mol^ꢀ1from Haworth et_ꢀa₁l.^{57 h}D_fH^o(SF₂OH⁺) ꢅ 321 kJ mol^ꢀ1, estimate, see text. ⁱAssuming bond additivity,

0

D_fH^o₀(ClSF) ¼ [D_fH₀^o(SF₂) + D_fH₀^o(SCl₂)]/2 ¼ ꢀ154 kJ mol

.

can occur, with smaller product ions possibly resulting from

more extensive fragmentation, for example:

to these fragment ions at their observed appearance energies,

it is necessary to postulate more complicated dissociation

mechanisms, which involve bond making as well as bond

breaking. The lowest energy pathways to SF₄⁺, SF₃⁺, and

ꢄ

ðSF₅CF₃^þÞ ! CF₂^þþ F þ SF₅

ð3Þ

ð4Þ

ð5Þ

+

! SF₄^þþ F þ CF₃

CF₂are:

ꢄ

ðSF₅CF₃^þÞ ! SF₄^þþ CF₄

! SF₃^þþ CF₄þ F

! CF₂^þþ SF₆

ð8Þ

ð9Þ

ð10Þ

! SF₃^þþ F þ F þ CF₃

Experimental observations of the fragmentation of (SF₆⁺)*²¹

favour a mechanism with sequential loss of F atoms:²²

ꢄ

ðSF₆^þÞ ! ðSF₅^þÞ þF ! ðSF₄^þÞ

In each case the breaking of the S–C bond is accompanied by

migration of F across this bond. One view of the overall pro-

cess is that reaction (8) corresponds to dissociation (1) fol-

lowed by F^ꢀabstraction from SF₅by CF₃⁺, similarly (10)

commences with (2). Reactions (8) and (9) may be connected,

ꢄ

þ 2F ! ðSF₃^þÞ þ3F ! . . .

Indeed, the observed threshold for the production of SF₃

ð6Þ

+

23

from SF₆corresponds closely to the energy needed to access

the dissociative ionisation limit:

with excited (SF₄⁺)* formed by (8) subsequently fragmenting

+

to produce SF₃⁺; the SF₃⁺–F bond is quite weak (D₀(SF₃

–

SF₆! SF₃^þþ 3F þ e^ꢀ

ð7Þ

F) ¼ 48 kJ mol^ꢀ1). The predicted threshold energies for

+

The dissociation of isolated energy-selected (SF₅CF₃⁺)* ions

has been investigated using the TPEPICO (Threshold Photo-

Electron PhotoIon COincidence) technique.¹⁰Making the

SF₄⁺, SF₃⁺, and CF₂produced through pathways (8), (9),

and (10) are given in Table 4. In each case the prediction lies

below the experimentally observed appearance energies for

these ions. At threshold it is concluded that production of

SF₄⁺, SF₃⁺, and CF₂⁺requires migration of an F atom across

the S–C bond. Fig. 1 includes these predicted threshold ener-

gies. Also shown are the lowest energy channels to SF₂⁺, SF⁺,

S⁺, CF⁺, F₂⁺, and F⁺, most of these channels involve migra-

tion of an F atom across the S–C bond. Many other dissocia-

tive ionisation pathways do exist.

When information on the energetics of the dissociative

ionisation pathways is combined with the recombination ener-

gies of the reagent ions, Fig. 1 provides a visual summary of

which product ions are accessible through exothermic

charge-transfer pathways for the various reagent ions. The

calculated 0 K enthalpy changes of the lowest energy charge-

transfer reactions to the observed product ions are given in

Table 1. For reagent ions with recombination energies of

12.5 eV or greater, charge transfer provides an exothermic or

close to thermoneutral pathway to each of the observed pro-

duct ions.

+

assumption that the pathway to CF₃is reaction (1), the

experimentally established adiabatic energy for the production

+

of CF₃(12.9 ꢁ 0.4 eV¹⁰) can be used as the starting point to

build up the adiabatic energies for the production of other

ions from SF₅CF₃by dissociative ionisation. The results are

displayed in Fig. 1, which was constructed using recently

+ 24

published 0 K enthalpies of formation for CF_mand CF_m

,

obtained from high level ab initio

+ 25

and for SF_nand SF_n

calculations.

The predicted lower bounds on the appearance energies for

,

the formation of SF₅⁺, SF₄⁺, SF₃⁺, and CF₂from the path-

+

ways (reactions (2) to (5)) listed above are collected in Table 4.

These ions have been detected following both photoionisa-

tion¹⁰and electron impact ionisation²⁶of SF₅CF₃, and the

experimental appearance energies (at 298 K) are presented

+

in Table 4. The appearance energy for SF₅is entirely consis-

tent with the suggestion that this ion is formed by simple scis-

+

sion of the S–C bond (reaction (2)). However, SF₄⁺, SF₃

,

+

and CF₂all appear at energies well below the predicted

thresholds. In order to ﬁnd energetically accessible pathways

It is frequently instructive to compare the product ion

branching ratios observed for a reagent ion of recombination

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2211

View Article Online

Table 3 The measured 300 K reaction rate coeﬃcients (in units of 10^ꢀ9cm³molecule^ꢀ1s^ꢀ1), cation product ions, their branching ratios (percen-

tages), suggested neutral products and 0 K enthalpies for the reactions of Ne⁺, F⁺, Ar⁺, N₂⁺, N⁺, CO⁺, CO₂⁺, O⁺, N₂O⁺, H₂O⁺, O₂⁺, SF₄

+

,

CF₂⁺, SF⁺, SF₂⁺, NO₂⁺, SF₅⁺, NO⁺, CF⁺, CF₃⁺, SF₃⁺and H₃O⁺with SF₆are presented. The recombination energies of the reagent cations

are listed in eV in brackets under the cations. The value in square brackets is the ﬂuoride aﬃnity of the reagent cation. The capture rate coeﬃcients,

calculated using Langevin theory,¹⁹are presented in square brackets under the experimental values. The measured rate coeﬃcients are considered to

be accurate to ꢁ20%. When dissociative charge transfer is exothermic, only this channel is represented in the table, with the neutrals assembled to

24

+25

give the lowest reaction enthalpy. 0 K enthalpies of formation for CF_m⁺, CF_m

,

and for SF_n, SF_n

are taken from recent high level ab initio

calculations. Other 0 K enthalpy data are taken from the JANAF tabulations,²⁷information available at the NIST website,⁵¹and additional

sources of data as identiﬁed in the table footnotes

Reagent ion

Rate coeﬃcient/10^ꢀ9cm³molecule^ꢀ1s^ꢀ1

Product ion (%)

Proposed neutral products

D_fH^o₀/kJ mol^ꢀ1

Ne⁺

0.69,^a0.8^b

[1.5]

SF₃(91)^b

F₂+ F + Ne

F₂+ Ne

F + Ne

ꢀ437

ꢀ485

ꢀ713

ꢀ38

+

(21.56)

[1753]

F⁺

SF₄(3)^b

+

SF₅(6)^b

1.1^a

[1.5]

SF₃(10)^a

F₂+ F + F

F + F

+

(17.42)

[1508]

Ar⁺

SF₅(90)^a

+

ꢀ313

+

SF₅(100)

0.91,^a0.93,^c1.2^d

[1.1]

F + Ar

F + N₂

ꢀ153

(15.76)

[1193]

+

N₂

1.0,^a1.2,^b1.3^d,e

[1.3]

ꢀ135

+

SF₅(100)

(15.58)

[1175]

N⁺

1.8,^b1.4^{d e}

[1.7]

SF₃(2)^b

F₂+ FN

F + N

ꢀ58

ꢀ34

+

(14.53)

[1374]

CO⁺

SF₅(98)^b

+

0.92,^a1.3,^d0.98^f

[1.3]

CO + F

FCO

+16

ꢀ133^g

+

SF₅(100)

(14.01)

[1173]

+

CO₂

0.009,^a0.014^e

[1.1]

F + CO₂

FCO₂

+39

ꢀ10^h

+

SF₅(100)

(13.77)

[1049]

O⁺

1.5,^d2.4ⁱ

[1.6]

0.003^a

[1.1]

F + O

FO

+54

ꢀ162

+124

+69

+

SF₅(100)

(13.62)

N₂O⁺

(12.89)

[971]

+

SF₅(100)

F + N₂O

FO + N₂

H₂O⁺

(12.6)

[962]

0.19^b

[1.5]

OSF₄

OHSF₅

2 HF

HF

D_fH^o₀(OSF₄⁺) ꢀ 317

D_fH^o₀(OHSF₅⁺) ꢀ 45

D_fH^o₀(H₂OSF₆⁺) + 228

+

H₂OSF₆

no reaction^d

+

O₂

(12.07)

[887]

+

SF₄

no reaction^j

(11.92)

[976]

+

CF₂

0.79^a

[1.0]

CF₃

ꢀ82

+

SF₅(100)

(11.43)

[1122]

SF⁺

no reaction^j

no reaction^e

no reaction^j

no reaction^{d e}

(10.25)

[1034]

+

SF₂

(10.17)

[876]

+

NO₂

(9.59)

[607]

+

SF₅

(9.55)

[1040]

NO⁺

(9.26)

[797]

CF⁺

0.88^a

[1.2]

CF₃(5)^a

SF₄

CF₂

ꢀ269

ꢀ26

+

(9.10)

[1066]

+

CF₃

SF₅(95)^a

+

0.22,^a0.25^e

[0.9]

CF₄

ꢀ45

+

SF₅(100)

(9.03)

[1085]

+

SF₃

no reaction^j

(8.26)

2212

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

Table 3 (continued)

Reagent ion

Rate coeﬃcient/10^ꢀ9cm³molecule^ꢀ1s^ꢀ1

Product ion (%)

Proposed neutral products

D_fH^o₀/kJ mol^ꢀ1

[870]

H₃O⁺

(6.27)

[843]

no reaction^a

a

Jarvis et al.^{8 b}Williams et al.^{7 c}Shul et al.^{4 d}Fehsenfeld.^{1 e}Babcock and Streit.^{3 f}Bowers et al.^{2 g}D_fH^o₀(FCO) ¼ ꢀ185 kJ mol^ꢀ1from Dixon

and Feller.^{52 h}D₀(F–CO₂) ¼ 48 kJ mol^ꢀ1, experimental value from Arnold et al.^{39 i}Morris et al.^{6 j}This study.

major product ions from photoionisation of SF₅CF₃are

+

CF₃and SF₃⁺, accompanied by small amounts of SF₄

,

SF₅⁺, and CF₂⁺. Over the photon energy range 15 to 20

eV, the TPEPICO branching ratios to CF₃and SF₃are

+

approximately constant, 70% CF₃and 20% SF₃⁺. The pro-

+

,

duct branching ratios for the reactions of SF₅CF₃with N₂

Ar⁺, and F⁺, presented in Table 1, mirror this behaviour.

Above about 20 eV, the TPEPICO branching ratio to CF₃

+

starts to fall, whilst that to SF₃rises. The product branching

ratios for the reaction of Ne⁺, compared to N₂⁺, Ar⁺, and

F⁺, with SF₅CF₃display the same trend. The TPEPICO data

show two interesting features within the photon energy range

12.7 to 15.5 eV. At about 14.9 eV, there is a rapid increase in

+

the SF₃branching ratio, from 2% (hn < 14.8 eV) to 20%

(hn > 15 eV). This change in dissociation behaviour occurs

at the onset of photoionisation into the ﬁrst excited state of

SF₅CF₃⁺. In contrast, the product branching ratios show

no evidence of a signiﬁcant change in the yield of SF₃⁺, which

remains at close to 20% as the recombination energy of the

reagent ion increases from 12.9 eV to more than 16 eV. Below

+

13 eV the TPEPICO branching ratio for SF₃shows a rapid

rise, with a matching fall for CF₃⁺. These observations at

around 12.9 eV, the observed thermodynamic limit for the

formation of CF₃by (1), reﬂect the competition between

+

(1) and (9). Reaction (1) is a simple dissociation, through a

loose transition state, while the more complicated mechanism

envisaged for (9), requiring intramolecular F^ꢀabstraction as

well as dissociation, will have a tight transition state. Once

Fig. 1 Adiabatic dissociative ionisation limits below 22 eV for

SF₅CF₃. The limits are divided into those attained by simple fragmen-

tation, and those involving the migration of an F atom across the S–C

bond. The recombination energies of the reagent ions with RE > 11 eV

are also presented in the ﬁgure. The electrons have been omitted from

the labelling of the various limits.

+

(1) opens, the yield of SF₃(calculated threshold 12.26 eV)

rapidly decreases, and CF₃is the dominant product ion.

+

This may account for the very high SF₃branching ratio

(92%) from the reaction of H₂O⁺(RE ¼ 12.61 eV) with

SF₅CF₃, compared to 19% from the reaction of N₂O⁺

(RE ¼ 12.89 eV). An alternative view¹²is that there is a

change in reaction mechanism between N₂O⁺and H₂O⁺. It

is postulated that N₂O⁺reacts by charge transfer, leading

energy RE with the TPEPICO ion product branching ratios

obtained with photons of energy RE. Similar values and trends

for branching ratios are regarded as an indicator that

+

to CF₃as the major product. Since the recombination

energy of H₂O⁺(12.62 eV) is below the observed onset of

photoionisation of SF₅CF₃(12.9 + 0.2 eV¹⁰), H₂O⁺is unlikely

the charge-transfer reaction proceeds via

a long-range

mechanism. TPEPICO experiments¹⁰have shown that the

Table 4 Calculated and observed appearance energies for the dissociative ionisation of SF₅CF₃

Observed appearance energy/eV

Fragment ion Neutral fragment(s) Reaction Predicted appearance energy/eV Photoionisation (TPEPICO¹⁰

)

Electron impact ionisation²⁶

12.9^a

13.40

18.90

14.27

17.37

11.76

17.86

12.26

12.9 ꢁ 0.4

13.9 ꢁ 1.5

12.9 ꢁ 0.1

13.2 ꢁ 0.6

+

CF₃

+

SF₅

CF₃

(1)

(2)

(3)

(10)

(4)

(8)

(5)

(9)

+

CF₂

F + SF₅

SF₆

16.0 ꢁ 2.0

13.5 ꢁ 1.5

14.94 ꢁ 0.13

17.8 ꢁ 0.5

12.1 ꢁ 0.3

14.5 ꢁ 0.1

+

SF₄

SF₃

F + CF₃

CF₄

+

F + F + CF₃

F + CF₄

a

Obtained from the TPEPICO study.¹⁰

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2213

View Article Online

to react eﬃciently by charge transfer, and instead a chemical

pathway in which bonds are broken and formed leads to the

major product:

combining the TPEPICO determination of the dissociative

ionisation limit for reaction (1), with the enthalpies of forma-

tion for CF₃⁺²⁴and SF₅²⁵obtained from recent ab initio calcu-

lations. The failure to observe SF₄CF₃as a product of the

reaction of N⁺with SF₅CF₃provides a tentative lower bound

+

H₂O^þþ SF₅CF₃! SF₃^þþ CF₄þ HF þ OH

It is diﬃcult to discern trends in the small TPEPICO branch-

ð11Þ

D_fH^o(SF₄CF₃⁺) > ꢀ348 kJ mol^ꢀ1

.

0

ing ratios to SF₅⁺, SF₄⁺, and CF₂⁺, but it does appear that the

yield of SF₅increases with photon energy. The charge-trans-

The product ion mass spectra were carefully inspected for

the presence of SF₂⁺, which has a mass only 1 u greater than

that of CF₃⁺. The lowest energy dissociative ionisation path-

+

fer data show that the reactions produce an increasing propor-

tion of SF₅as the recombination energy of the reagent ion

+

way to SF₂is:

SF₅CF₃! SF₂^þþ F₂þ CF₄þ e^ꢀ

ð16Þ

increases. This can be viewed as a consequence of competition

between channels (1) and (2). As the available energy increases,

the dissociation of (SF₅CF₃⁺)* produces a greater proportion

for which the thermodynamic limit is 14.87 eV. SF₂⁺was only

observed as a product of the reaction of SF₅CF₃with Ne⁺. It

was not seen in the TPEPICO spectra of SF₅CF₃.

+

of SF₅(thermodynamic threshold 13.40 eV) compared to

CF₃(12.90 eV).

+

Inspection of Table 1 shows that there are reactions between

SF₅CF₃and certain reagent ions whose recombination ener-

gies are below the calculated adiabatic threshold for the lowest

energy dissociative ionisation channel of SF₅CF₃(!

SF₄⁺+ CF₄+ e^ꢀ, 11.76 eV). An example is the reaction

In addition to product ion branching ratios, the reaction

rate coeﬃcient can be a useful diagnostic of the mechanism.

For long-range charge-transfer to be eﬃcient, identiﬁed by

k

_exp/k_cꢅ 1, it is necessary that there is good vibrational over-

lap between the neutral reactant M and the ion (M⁺)* initially

formed by charge transfer:

+

between CF₃and SF₅CF₃:

CF₃^þþ SF₅CF₃! SF₃^þþ CF₄þ CF₄

ð17Þ

ꢄ

A^þþ M ! ðM^þÞ þA ! . . .

ð12Þ

(this is the most exothermic pathway to SF₃⁺from these reac-

tants). In contrast to the charge-transfer reactions (12), where

there is no exchange of atoms between A and M, and A is a spec-

tator to the dissociation of (M⁺)*, the pathway to these pro-

posed products involves the formation of a new chemical

bond between the reagent ion (CF₃⁺) and a fragment abstracted

from SF₅CF₃(F^ꢀ). For the reaction above this is coupled with

fragmentation of SF₄CF₃⁺to release the observed ion product

SF₃⁺. Since SF₄CF₃⁺is not observed as a product of the reac-

tion (which is consistent with the proposed bounds on

D_fH^o₀(SF₄CF₃⁺)), the overall reaction may well be concerted:

This can be judged by inspecting the photoelectron spectrum

of M. For the generic charge-transfer reaction (12), energy bal-

ance requires that:

h

i

ꢀ

ꢁ

ꢄ

E ðM^þÞ ꢀ EðMÞ ¼ RE A^þ

ð13Þ

so if there is a strong signal in the photoelectron spectrum of

M at an ionisation energy equal to RE(A⁺), eﬃcient charge-

transfer (reaction (12)) can be anticipated. The photoelectron

spectrum of SF₅CF₃has not been reported. However, a TPES

(Threshold PhotoElectron Spectrum) of SF₅CF₃is available.¹⁰

It shows a signiﬁcant yield of electrons for all photon energies

from 12.5 eV to greater than 23 eV. This suggests that if the

mechanism of reaction for all reagent ions with recombination

energies greater than 12.5 eV is long-range charge-transfer,

then the reactions should all be eﬃcient. This is conﬁrmed

by the data in Table 1. The combination of reaction rate coeﬃ-

cients equal to the capture estimates, and product ion yields

matching TPEPICO results conﬁrms that the reactions of

SF₅CF₃with reagent ions possessing recombination energies

greater than 12.5 eV can be viewed as long-range charge trans-

fer reactions.

ð18Þ

(CF₃⁺has been replaced by the generic reagent ion A⁺, because

+

we believe the several other reactions in which SF₃is formed

proceed via this mechanism, see below)

The reactions of SF₅CF₃do not furnish any examples of

detected ion products which contain new bonds between units

from SF₅CF₃and the reagent ion. The existence of pathways

such as (18) is inferred from the requirement that the overall

proposed reaction must be close to thermoneutral or exother-

mic, in order for it to proceed at an observable rate. The reac-

The reactions with Ar⁺and N₂⁺resulted in one product ion

not observed from other charge-transfer reactions. This pro-

duct ion has a mass of 177 u, and could be either SF₄CF₃

+

tions of SF₅CF₃with SF₄⁺, CF₂⁺, SF⁺, SF₅⁺, CF⁺, and CF₃

or SF₅CF₂(or both). Since the bond dissociation energy of

S–F in SF₆is less than that of C–F in CF₄, we propose that

+

provide further examples of reactions proceeding via chemical

rather than charge-transfer pathways. Table 1 indicates the

likely neutral products, identiﬁed from the exothermicity cri-

+

SF₄CF₃is the product ion. SF₄CF₃(and/or SF₅CF₂

+

)

was not a reported product of either the electron impact ioni-

sation²⁶or the photoionisation of SF₅CF₃.¹⁰The observation

of the charge-transfer reactions:

+

terion. For all of the reactions, SF₃is one of the major pro-

duct ions, and the inferred product neutrals point to reaction

via mechanism (18). A more exothermic pathway exists for

the reaction with CF⁺, with C₂F₆as the neutral product,

rather than CF₂+ CF₄as anticipated for reaction (18).

The other major product ion seen from the reactions of

CF₂⁺, SF⁺, SF₅⁺, and CF⁺is CF₃⁺, and reaction (19) shows

a possible mechanism:

Ar^þþ SF₅CF₃! SF₄CF₃^þþ F þ Ar

ð14Þ

and

N₂^þþ SF₅CF₃! SF₄CF₃^þþ F þ N₂

can be used to estimate an upper bound for the enthalpy of for-

ð15Þ

mation of SF₄CF₃, D_fH₀^o(SF₄CF₃⁺). However, the possibility

+

that the Ar⁺(²P_1/2and P_3/2) and N₂(v ¼ 1 and 0) reagent

ions may not be fully thermalised increases the uncertainty

in the estimate. If the SF₄CF₃⁺, for which the branching ratio

2

+

ð19Þ

is only 1 or 2%, is formed from the reactions of Ar⁺(²P_1/2) and

As Table 1 shows, for both CF₂and CF⁺there is a more

+

N₂(v ¼ 1), then the bound obtained is <ꢀ221 kJ mol^ꢀ1

,

exothermic route to CF₃⁺, with CF₄as one of the neutral pro-

ducts. The lowest energy pathways to CF₃⁺from the reactions

while if this ion can be formed from Ar⁺(²P_3/2) and N₂

+

(v ¼ 0), then the bound is <ꢀ247 kJ mol^ꢀ1. These limits use

+

of SF₄⁺, SF₂⁺, and SF₃have substantial endothermicities,

a value for D_fH^o₀(SF₅CF₃) ¼ ꢀ1673 kJ mol^ꢀ1, obtained by

2214

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

For reagent ions with recombination energies greater than

12.5 eV, there is only one case where it is necessary to invoke

a chemical pathway to obtain an exothermic reaction to one of

the observed ion products. This is:

ꢂ

CF₃O þ N₂

N₂O^þþ SF₅CF₃! SF₅^þþ

ð24Þ

CF₂O þ F þ N₂

For O₂⁺, given the large uncertainty in the dissociative ioni-

sation energy of SF₅CF₃, it is not possible to be certain as to

which pathways are exothermic. The situation is further com-

+

.

plicated by the likely presence of vibrationally excited O₂

+

However, plots of ln(O₂counts) against SF₅CF₃concentra-

tion show no curvature, indicating that the rate coeﬃcient is

+

the same for O₂in v ¼ 0, 1, and 2. For O₂the production

+

of CF₃and SF₅is likely to be via chemical routes:

+

ð25Þ

ð26Þ

Fig. 2 Plot of reaction eﬃciencies (k_exp/k_c) against ﬂuoride ion aﬃ-

+

nity for the reactions of the CF_m(m ¼ 1–3) and SF_n(n ¼ 1–5)

reagent ions with SF₅CF₃, SF₅Cl, and SF₆.

ꢀ

+

which accounts for the non-observation of CF₃as a product

of any of these reactions.

(reaction (26) is CF₃abstraction, analogous to (21)), with

SF₄coming from a dissociative charge-transfer reaction (8).

+

Surveying the thermochemistry in Table 1, for which the

major source of uncertainty is ꢁ0.4 eV (ꢁ39 kJ mol^ꢀ1) in

the dissociative ionisation energy of SF₅CF₃(reaction (1)),

shows that all of the observed products can be formed in reac-

tions which are close to thermoneutral or exothermic. Thus

application of the exothermicity criterion indicates that there

are no obvious inconsistencies in the thermochemical data

used in the construction of Table 1. By combining the dissocia-

tive ionisation energy of SF₅CF₃¹⁰with the ab initio values for

Reaction schemes (18) and (19) involve the abstraction of

F^ꢀfrom SF₅CF₃by the reagent ion, so some correlation of

the reaction eﬃciencies with the F^ꢀaﬃnities of the reagent

ions can be anticipated. The ﬂuoride aﬃnities of the reagent

ions, expressed as minus the 0 K enthalpy change for the reac-

tion:

A^þþ F^ꢀ! AF

are given in Table 1. The reaction eﬃciencies are plotted

against ﬂuoride aﬃnities in Fig. 2. Comparing the CF_m

ð20Þ

D_fH^o₀(CF₃), D_fH^o₀(CF₃),²⁴and D_fH₀^o(SF₅⁺),²⁵the following

+

thermochemical

data

for

SF₅CF₃

are

obtained:

+

(m ¼ 1–3) and SF_n(n ¼ 1–5) reagent ions, the CF_mions

have higher ﬂuoride aﬃnities, and react more eﬃciently. For

CF₂⁺, which has the highest ﬂuoride aﬃnity, F^ꢀabstraction

D_fH^o₀(SF₅CF₃) ¼ ꢀ1673 kJ mol^ꢀ1and D₀(SF₅–CF₃) ¼ 374

kJ mol^ꢀ1

.

+

is directly observed, through the identiﬁcation of SF₄CF₃

as one of the product ions.

3.2 SF₅Cl reactions

+

SF₅was identiﬁed as a product of the reactions of CF₂

+

,

The experimental reaction rate coeﬃcients for the reactions of

SF₅Cl with Ne⁺, F⁺, Ar⁺, N₂⁺, N⁺, CO⁺, CO₂⁺, O⁺, N₂O⁺,

H₂O⁺, O₂⁺, SF₄⁺, CF₂⁺, SF⁺, SF₂⁺, NO₂⁺, SF₅⁺, NO⁺, CF⁺,

CF₃⁺, SF₃⁺, and H₃O⁺, together with the observed product

ions, and their branching ratios are given in Table 2. Before

discussing these results, some basic thermochemistry needs to

be evaluated.

CF⁺, and CF₃with SF₅CF₃._ꢀFor CF₂and CF₃⁺, the most

+

likely reaction pathway is CF₃abstraction:

ð21Þ

However, a more complicated pathway is required for CF⁺,

because in order to satisfy the exothermicity criterion, the neu-

tral product must be CF₂=CF₂, and not CF–CF₃. Despite

The JANAF tables²⁷provide a value for the enthalpy of for-

mation of SF₅Cl (D_fH₀^o(SF₅Cl) ¼ ꢀ1026 kJ mol^ꢀ1) derived

from a thermochemical study of the reaction:²⁸

+

this, the SF₅branching ratio from CF⁺+ SF₅CF₃is about

+

the same (2%) as for the reactions of CF₂and CF₃

+

.

Some conﬁrmation of the proposed mechanisms for the

SF₅Cl_ðlÞþ 8NaOH

! Na₂SO_4ðaqÞþ 5NaF

þ NaCl_ðaqÞþ 4H₂O

ðaqÞ

+

reactions of CF_mand SF_ncould be obtained by studying

the reactions of the isotopically labelled reagent ions,

ð27Þ

ðlÞ

+

¹³CF_mand ³⁴SF_n⁺. Thus reaction scheme (19) predicts that

When this is combined with the enthalpy of formation of SF₅

obtained from the high level ab initio calculations of Bauschli-

cher and Ricca,²⁵the 0 K adiabatic energy for the lowest dis-

sociative ionisation limit of SF₅Cl:

+

the CF₃product will not incorporate the ¹³C present in

+

labelled ¹³CF_mreagent ions, and according to scheme (18)

+

the use of labelled ³⁴SF_nwill not lead to the production of

³⁴SF₃⁺. CF₃⁺may undergo two reactions, which are not listed

SF₅Cl ! SF₅^þþ Cl þ e^ꢀ

ð28Þ

+

:

in Table 1, because they do not consume CF₃

is predicted to be 1230 kJ mol^ꢀ1or 12.75 eV. This is not consis-

tent with the reported photoionisation data for SF₅Cl. The

photoelectron spectrum of SF₅Cl²⁹places an upper bound on

the adiabatic ionisation energy of SF₅Cl of 12.33 eV. Further,

the photoionisation mass spectrum (PIMS) of SF₅Cl³⁰shows

an onset for the formation of SF₅⁺of 12.32 eV, with other frag-

ment ions appearing at higher energies. It seems that either

D_fH^o(SF₅Cl) or D_fH^o(SF₅⁺) must be wrong. Recent years have

þ

CF₃^þþ SF₅CF₃! SF₅CF₃þ CF₃

! CF₄þ SF₄þ CF₃

ð22Þ

D_rH₀^o¼ ꢀ15 kJ mol^ꢀ1ð23Þ

The occurrence of these reactions would be revealed through

þ

+

the use of labelled ¹³CF₃as the reagent ion.

+

In accord with the observation that NO⁺and NO₂do not

react with SF₅CF₃, a search of the possible ion-molecule reac-

tions failed to identify any exothermic reactions. Since H₃O⁺

does not react with SF₅CF₃, it is concluded that the proton

aﬃnity of SF₅CF₃is less than that of H₂O.

0

seen considerable eﬀorts, both experimental and theoretical, to

establish a reliable value for D_fH^o(SF₅⁺). A recommended

0

experimental value due to Fisher et al.³¹of 11 kJ mol^ꢀ1would

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2215

View Article Online

place the dissociative ionisation limit (28) at 11.99 eV, and give

consistency with the observed upper bound to this limit of 12.32

eV. However, theoretical calculations^25,32,33consistently give

then D_rH^o₀(29) < 1165 kJ mol^ꢀ1. The other possible route is F^ꢀ

+

abstraction by O₂

:

O₂^þþ SF₅Cl ! SF₄Cl^þþ F O₂

ð31Þ

rather larger values, for example D_fH^o(SF₅⁺) ¼ 84 kJ mol^{ꢀ1 25}

,

and some potential problems in the experimental determination

0

taking D₀(F–O₂) ¼ 50 kJ mol^ꢀ127leads to D_rH^o(29) < 1215 kJ

of D_fH^o(SF₅⁺) have been identiﬁed.³⁴In the absence of an inde-

0

mol^ꢀ1. These bounds assume that SF₄Cl⁺can be formed from

O₂⁺(v ¼ 0). The neutral products, F + O₂or FO₂, are

unknown, so other data must be used to choose between these

two bounds. The bond dissociation energy D₀(SF₄⁺–F) is 383

0

pendent test of D_fH^o₀(SF₅Cl), and the need to use a consistent set

of thermochemistry for SF₅Cl, SF_nand SF_n⁺, we have chosen to

use the observed onset for the dissociative ionisation (28) com-

bined with D_fH^o(SF₅⁺) ¼ 84 kJ mol

,

^{ꢀ1 25}to determine a lower

kJ mol^{ꢀ1 25}it is expected that D₀(SF₄⁺–Cl) will be lower.

,

0

limit to D_fH^o₀(SF₅Cl) ꢂ ꢀ985 kJ mol^ꢀ1. With the exception that

this value is a lower limit, the procedure matches that adopted to

determine D_fH^o₀(SF₅CF₃). In Table 2 and the remainder of the

Adopting D_rH^o(29) < 1215 kJ mol^ꢀ1gives D₀(SF₄⁺–Cl) > 356

0

kJ mol^ꢀ1, which is reasonable, while D_rH^o₀(29) < 1165 kJ mol^ꢀ1

leads to D₀(SF₄⁺–Cl) > 406 kJ mol^ꢀ1. At the onset of photo-

text, we have used D_fH^o(SF₅Cl) ¼ ꢀ985 kJ mol^ꢀ1. Calculated

0

reaction enthalpies are thus upper bounds. When the revised

+30

ionisation (12.32 eV) the only observed ion product is SF₅

(reaction 28), a value for D_rH^o₀(29) below 1189 kJ mol^ꢀ1is then

unlikely to be correct. The recommended bound is thus

D_rH^o₀(29) < 1215 kJ mol^ꢀ1, and the value is expected to be

quite close to this upper bound. The calculation of the 0 K

enthalpy changes for ion-molecule reactions is facilitated by

combining D_rH^o(29) with D_rH^o(SF₅Cl) ¼ ꢀ985 kJ mol^ꢀ1, to

value for D_fH^o₀(SF₅Cl) is combined with thermochemistry for

+

SF_nand SF_nfrom the ab initio calculations of Bauschlicher

and Ricca,²⁵and for other neutrals and ions from the JANAF

tables,²⁷Fig. 3, showing the relative energies of many of the

lower dissociative ionisation limits of SF₅Cl, can be constructed.

SF₄Cl⁺is a product of many of the ion-molecule reactions

of SF₅Cl, so in analysing the reactions it would be useful to

have a value for D_fH₀^o(SF₄Cl⁺). SF₄Cl⁺is seen in the electron

impact mass spectrum of SF₅Cl,³⁵with an appearance energy

of 15.9 eV, and is known to be a product of the dissociative

photoionisation of SF₅Cl,^30,36the ion is ﬁrst observed at a

photon energy of 14.76 eV. This places an upper bound on

the energy for:

0

give D_fH^o₀(SF₄Cl⁺) < 152 kJ mol^ꢀ1; the calculations shown in

Table 2 used D_fH^o(SF₄Cl⁺) ¼ 150 kJ mol^ꢀ1. Since SF₄Cl⁺is

0

a minor product of the reaction between O₂and SF₅Cl, it

+

could be formed exclusively from the vibrationally excited

O₂known to be present. For example, if all the SF₄Cl⁺is

+

produced by the reaction of O₂⁺(v ¼ 2) with SF₅Cl:

O₂^þðv ¼ 2Þ þ SF₅Cl ! SF₄Cl^þþ F O₂

ð32Þ

SF₅Cl ! SF₄Cl^þþ F þ e^ꢀD_rH₀^oð29Þ ꢆ 1424 kJ mol^ꢀ1

ð29Þ

The exothermicity criterion (D_rH^o₀< 0 for an observed reac-

tion) can be applied to reﬁne the upper bound to D_rH^o₀(29).

In particular, the observation of SF₄Cl⁺from the reaction of

O₂⁺with SF₅Cl produces the lowest upper bound. If the path-

way to SF₄Cl⁺is through dissociative charge transfer:

then D_rH^o(29) < 1259 kJ mol^ꢀ1, D₀(SF₄⁺–Cl) > 312 kJ mol^ꢀ1

,

0

and D_fH^o₀(SF₄Cl⁺) < 197 kJ mol^ꢀ1. This rather higher upper

bound for D_rH^o₀(29) is not ruled out by application of the

exothermicity criterion to any of the other reactions observed

to form SF₄Cl⁺.

SF₄⁺and SF₃⁺are observed as products of the photoionisa-

tion of SF₅Cl,³⁰with appearance energies of 15.87 and 16.2 eV,

respectively. Some preliminary work on the TPEPICO spec-

trum of SF₅Cl³⁶gives slightly higher values (at 298 K) of

O₂^þþ SF₅Cl ! SF₄Cl^þþ F þ O₂

ð30Þ

+

16.2 ꢁ 0.3 and 16.8 ꢁ 0.3 eV, respectively. SF₄and SF₃are

observed in the electron impact mass spectrum of SF₅Cl,³⁵

but their appearance energies have not been reported. The

fragmentation of SF₅Cl⁺may occur by step-wise loss of Cl

and F atoms:

ꢄ

SF₅Cl^þ! ðSF₅^þÞ þ Cl ! ðSF₄^þÞ þ F þ Cl

ð33Þ

ꢄ

! ðSF₃^þÞ þ 2F þ Cl !

+

in which case the adiabatic limits to SF₄and SF₃will be

16.28 and 16.78 eV, respectively. The other possibility is a con-

certed fragmentation leading to the formation of ClF:

ꢄ

SF₅Cl^þ! ðSF₄^þÞ þ ClF ! ðSF₃^þÞ þ F þ ClF ! ð34Þ

the predicted limits are now 13.72 eV (SF₄⁺) and 14.22 eV

(SF₃⁺). The limit for SF₃could be even lower, if ClF₂is

formed as the neutral product. The fragment ions SF₄and

+

37

+

SF₃are not observed at energies well below the limits for

their formation by step-wise loss of Cl and F from SF₅Cl⁺

(reaction scheme (33)). Indeed the preliminary TPEPICO data

suggest that their appearance may coincide with the adiabatic

energies for simple fragmentation.

Ion product branching ratio comparisons between the ion-

molecule reactions (Table 2) and preliminary results from a

TPEPICO study of energy selected SF₅Cl⁺³⁶provide strong

evidence that the reactions of SF₅Cl with Ne⁺, F⁺, Ar⁺, and

N₂all occur via long-range charge transfer.³⁸Fig. 4 shows

+

Fig. 3 Adiabatic dissociative ionisation limits below 22 eV for SF₅Cl.

The limits are divided into those attained by simple fragmentation, and

those involving the migration of an F atom across the S–Cl bond. The

recombination energies of the reagent ions with RE > 12 eV are also

presented in the ﬁgure. The electrons have been omitted from the label-

ling of the various limits.

the mass spectrum obtained for the reaction of Ne⁺with

SF₅Cl. A large amount of energy (9.25 eV, relative to the low-

est dissociative ionisation limit to SF₅⁺+ Cl) is available to

dissociate (SF₅Cl⁺)*, and this is reﬂected in the range of pro-

2216

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

between 75 and 133 kJ mol^{ꢀ1 37}

.

An experimental determina-

,

tion of D₀(F–CO₂) has yielded a value of 48 kJ mol^{ꢀ1 39}

calculations providing a slightly higher value of 61 kJ

with

mol^{ꢀ1 40–42}

Thus (38) and (39) are both exothermic routes to

.

SF₃⁺, and both involve the formation of a relatively weakly

bound product neutral.

+

The CO₂reaction generates two ion products, in addition

to SF₄Cl⁺, for which enthalpies of formation have not been

reported. They are SFCl⁺and SF₂Cl⁺, but it should be noted

that only traces of SF₂Cl⁺were observed. Estimates of their

enthalpies of formation can be obtained, starting from the cal-

culated thermochemistry²⁵of SF₂and SF₃⁺, and the knowl-

+

edge that S–Cl bonds are weaker than S–F bonds (for example

D₀(SF₅–F) > D₀(SF₅–Cl), and D₀(S–F) > D₀(S–Cl)). Assuming

that for the exchange reaction:

Fig. 4 A typical mass spectrum obtained from the charge-transfer

reaction of Ne⁺with SF₅Cl, illustrating the considerable degree of

fragmentation, that results from the high recombination energy of

Ne⁺(21.56 eV) compared to the other reagent ions used in this study.

SFCl^þþ F ! SF₂^þþ Cl

ð40Þ

D_rH^o₀¼ ꢀ50 kJ mol^ꢀ1, we obtain D_fH^o₀(SFCl⁺) ¼ 782 kJ

+

mol^ꢀ1

,

which is compatible with the bound of

Four product ions are identiﬁed in the ﬁgure; Cl⁺, SF₃(dominant),

D_fH^o₀(SFCl⁺) < 878 kJ mol^ꢀ1, implied by the observation of

the reaction:

SF₅and SF₄Cl⁺. SF₂and SF₄are also observed products of the

reaction, but their branching ratios are very small, resulting in product

ion intensities which are too small to be discernible on the ﬁgure.

+

CO₂^þþ SF₅Cl ! SFCl^þþ CF₄þ O₂

ð41Þ

(other neutral products can be excluded, for pathways leading

to them cannot satisfy the exothermicity criterion). Adopting

the procedure described above to generate an estimate for

duct fragment ions seen in the mass spectrum. Indeed, Cl⁺and

SF₂are only observed from the reaction of SF₅Cl with Ne⁺.

+

For photon energies below 14.8 eV, the only product ion

observed in the TPEPICO experiments is SF₅⁺. The results

in Table 2 show that the reactions of SF₅Cl with ions having

the

enthalpy

of

formation

of

SF₂Cl⁺

yields

D_fH^o(SF₂Cl⁺) ¼ 453 kJ mol^ꢀ1. Even with the aid of this esti-

0

mate it is not possible to identify conclusively the pathway that

recombination energies in the range 12.3 to 14.8 eV are eﬃ-

+

leads to the production of SF₂Cl⁺from the reaction of CO₂

+

cient, but yield various ion products in addition to SF₅

.

with SF₅Cl, and two possible sets of neutral products are listed

in Table 2. Dissociative charge transfer is not the route to

SF₂Cl⁺. SF₂Cl⁺is also an observed product of the reaction

with H₂O⁺, and a chemical pathway such as:

SF₄Cl⁺is a product of all these reactions. It could be the pro-

duct of charge transfer and fragmentation, induced by short-

range interactions:

A

^þþ SF₅Cl ! SF₄Cl^þþ F þ A

ð35Þ

H₂O^þþ SF₅Cl ! SF₂Cl^þþ 2HF þ FO

ð42Þ

and/or the formation of SF₄Cl⁺may be linked to chemical

reaction between F and A, such as:

is required to satisfy the exothermicity criterion. The reaction

with H₂O⁺also produces SF₂OH⁺, providing direct conﬁrma-

tion that reactions involving the making as well as breaking of

bonds do occur. An estimate for D_fH^o₀(SF₂OH⁺) can be

obtained by assuming that D₀(SF₂⁺–OH) ꢇ D₀(SF₂⁺–F),

which leads to D_fH₀^o(SF₂OH⁺) ꢇ 321 kJ mol^ꢀ1. This is compa-

tible with the formation of SF₂OH⁺in the reaction of H₂O⁺

with SF₅Cl:

A

^þþ SF₅Cl ! SF₄Cl^þþ AF

ð36Þ

Either (or both) of reactions (35) and (36) may occur when

RE(A⁺) > D_rH^o₀(29), schemes of the type shown in reaction

(36) will be necessary when RE(A⁺) < D_rH₀^o(29). The relative

branching ratios for SF₅and SF₄Cl⁺vary widely with the

+

reagent ion. There is no correlation of the SF₅⁺/SF₄Cl⁺ratio

with either the recombination energies of the reagent ions, or

with the enthalpies of reactions in which bonds are formed

and broken. Speciﬁc short-range interactions within the initi-

ally formed ion-molecule complex are evidently important to

the outcome of the reactions. Without further information,

particularly the identiﬁcation of the neutral products, the

mechanisms of the reactions must remain uncertain.

H₂O^þþ SF₅Cl ! SF₂OH^þþ ClF₂þ HF

ð43Þ

For the CF_m⁺and SF_n⁺reagent ions, which all have recom-

bination energies less than 12.0 eV, i.e. well below the energy

needed for dissociative ionisation of SF₅Cl, the majority of

+

The reaction of CO₂with SF₅Cl produces a number of

+

interesting products. The recombination energy of CO₂

(13.77 eV) is well below the minimum energy (14.22 eV)

required for the dissociative ionisation:

SF₅Cl ! SF₃^þþ ClF þ F þ e^ꢀ

ð37Þ

D_rH₀^o(37) ꢀ RE(CO₂⁺) ¼ 43 kJ mol^ꢀ1, but SF₃⁺is an observed

(trace) product from the reaction of CO₂⁺. Two possible

routes to SF₃are dissociative charge transfer with ClF₂as

the neutral product:

+

þ

SF₅Cl þ CO₂! SF₃^þþ ClF₂þ CO₂

ð38Þ

and F^ꢀabstraction by CO₂coupled with fragmentation of

+

the intermediate SF₄Cl⁺:

þ

SF₅Cl þ CO₂! SF₃^þþ ClF þ FCO₂

ClF₂is a stable species, with D₀(FCl–F) calculated to lie

ð39Þ

Fig. 5 A typical mass spectrum obtained from the chemical reaction

of CF₃with SF₅Cl. Two product ions are observed; SF₅and

SF₄Cl⁺.

+

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2217

View Article Online

the observed ion products are the result of F^ꢀand Cl^ꢀabstrac-

tion:

strengths of the S–Cl and S–F bonds in SF₅Cl. SF₃⁺(+CClF₃)

may be the result of CF₂Cl⁺abstracting F^ꢀfrom SF₄before

the ion-molecule complex breaks up following reaction. An

ꢂ

SF₄Cl^þþ ZF_nþ1

+

ZF_n^þþ SF₅Cl !

ð44Þ

alternative pathway to SF₃is:

SF₅^þþ ZF_nCl

+

Fig. 5 shows a mass spectrum from the reaction of CF₃with

SF₅Cl, both SF₄Cl⁺and SF₅⁺are seen, together with unreacted

CF₃⁺. There are wide variations in both the eﬃciencies of the

reactions, and in the relative yields of SF₄Cl⁺(F^ꢀabstraction)

and SF₅⁺(Cl^ꢀabstraction). Fig. 2 shows how the reaction eﬃ-

ciencies vary with the ﬂuoride aﬃnities of the reagent ions. SF⁺

reacts with a much higher eﬃciency than the other SF_n⁺reagent

ions. The major ion product for the reactions of each of the

ð47Þ

+

This route to SF₃from the reaction of CF⁺with SF₅Cl is

much less exothermic (see Table 2), but without identiﬁcation

of the neutral products, a conclusive identiﬁcation of the path-

way cannot be established.

CF_mand SF_nreagent ions is SF₄Cl⁺, but for SF⁺it is

+

SF₂was observed not to react with SF₅Cl, but it appears

+

SF₅(Cl^ꢀabstraction). The overall reactions can be viewed

likely that F atom transfer (!SF₃⁺+ SF₄Cl) and Cl atom

transfer (!SF₂Cl⁺+ SF₅) are exothermic pathways. These

neutral atom transfer pathways are also predicted to be

as two linked steps - formation of an ion pair from SF₅Cl,

SF₅⁺+ Cl^ꢀfor Cl^ꢀabstraction and SF₄Cl⁺+ F^ꢀfor F^ꢀ

abstraction, coupled with combining the released anion with

the reagent ion. SF₄Cl⁺+ F^ꢀlies 45 kJ mol^ꢀ1higher in energy

than SF₅⁺+ Cl^ꢀ. This may be expected to favour the produc-

exothermic for SF⁺, but SFCl⁺and SF₂were not formed.

+

For SF₃⁺, NO⁺, and NO₂⁺, an exploration of the possible

reaction pathways, including transfer of F^ꢀ, Cl^ꢀ, F, and Cl,

did not identify any overall exothermic reactions, which

accounts for the failure of these ions to react with SF₅Cl.

H₃O⁺also did not react, showing that the proton aﬃnity of

SF₅Cl is less than that of H₂O.

+

tion of SF₅over SF₄Cl⁺. However, bond formation between

the reagent ion and Cl^ꢀ/F^ꢀis occurring simultaneously. The

reagent ions all have higher aﬃnities for F^ꢀthan for Cl^ꢀ, which

will tend to lead to the formation of SF₄Cl⁺in preference to

SF₅⁺. The overall outcome is a balance between these two com-

peting factors. The diﬀerence between the F^ꢀand Cl^ꢀaﬃnities

is smallest for SF⁺, and this is reﬂected in the product branching

3.3 SF₆reactions

ratios, for the reaction with SF⁺gives the highest yield of SF₅

.

+

There have been several past studies of the gas-phase reactions

under thermal conditions of SF₆with positive ions.^1–4,6–8The

results are summarised in Table 3. We have extended the pre-

vious studies by investigating the reactions, under thermal con-

+

Indeed in the case of SF⁺, the branching ratio for SF₅is

greater than for SF₄Cl⁺, even though the F^ꢀabstraction path-

way has a higher overall exothermicity.

In addition to Cl^ꢀand F^ꢀabstraction, leading to SF₅⁺and

+

ditions, of SF_n(n ¼ 1 to 5) with SF₆. In each case we

+

SF₄Cl⁺, respectively, CF₂reacts to produce CF₂Cl⁺, by trans-

observed no depletion of the reagent ion signals over the full

range of SF₆concentrations accessible to us. This implies that

for each of these reactions the rate coeﬃcient is less than 10^ꢀ13

cm³molecule^ꢀ1s^ꢀ1. As for Tables 1 and 2, the thermochemis-

fer of a Cl atom from SF₅Cl to CF₂⁺. It is perhaps surprising that

CF₃⁺, arising from F atom transfer was not one of the observed

products. It is unlikely that the F abstraction pathway is

endothermic. To show this a value for D_fH^o₀(SF₄Cl) is required.

D_fH^o₀(SF₄Cl) ¼ D_fH₀^o(SF₄) + D_fH₀^o(Cl)–D₀(SF₄–Cl), assuming

that D₀(SF₄–Cl) ¼ 120 kJ mol^ꢀ1(a little less than D₀(SF₄–

F) ¼ 153 kJ mol^ꢀ125) leads to D_fH^o(SF₄Cl) ¼ ꢀ761 kJ mol^ꢀ1

,

0

from which D_rH^o₀(CF₂⁺+ SF₅Cl ! CF₃⁺+ SF₄Cl) ¼ ꢀ277 kJ

mol^ꢀ1. Based on overall exothermicity, F abstraction should

be favoured over Cl abstraction. The atom abstraction reactions

begin with stretching an S–F or S–Cl bond, the greater strength

of the S–F bond compared to the S–Cl bond may account for the

+

.

formation of CF₂Cl⁺rather than CF₃

The reaction with CF⁺produces an ion product, CF₂Cl⁺,

corresponding to the transfer of two neutral atoms to the

reagent ion. Transfer of a single Cl or F atom to CF⁺is not

observed. This is consistent with estimates of D_rH^o₀for:

ꢂ

D_rH₀^o¼ 42 kJ mol^ꢀ1

CFCl^þþ SF₅

CF^þþ SF₅Cl !

ð45Þ

CF₂^þþ SF Cl

ꢀ1

o

D_rH₀¼ 11 kJ mol

4

which suggest that both these reactions are endothermic. The

endothermicities are not large (<50 kJ mol^ꢀ1), but they do

imply that CF₂Cl⁺will not be formed by the sequential trans-

fer of two atoms, but rather by a concerted mechanism:

ð46Þ

+

CF₃could also be produced by a mechanism of this type.

The estimated exothermicity (ꢀ349 kJ mol^ꢀ1

,

using

Fig. 6 Adiabatic dissociative ionisation limits below 22 eV for SF₆.

The limits are divided into those attained by simple fragmentation,

and those involving the migration of an F atom across an S–F bond.

The recombination energies of the reagent ions with RE > 14 eV are

also presented in the ﬁgure. The electrons have been omitted from

the labelling of the various limits.

D_fH^o₀(SF₃Cl) ¼ ꢀ619 kJ mol^ꢀ1) is similar to that calculated

+

for the production of CF₂Cl⁺, but CF₃was not observed.

+

This parallels the reaction of CF₂⁺, where CF₃was not one

of the observed products, and may be explained by the relative

2218

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

try has been constructed using enthalpies of formation for

CF_m, CF_m⁺²⁴and SF_n, SF_n⁺²⁵from recent high level ab initio

calculations. These ab initio values place the dissociative ioni-

sation energy for:

consistent with an observable rate coeﬃcient. Two possibilities

are:

N₂O^þþ SF₆! SF₅^þþ F N₂O

ð53Þ

and

SF₆! SF₅^þþ F þ e^ꢀ

ð48Þ

N₂O^þþ SF₆! SF₅^þþ FO N₂

ð54Þ

at 14.18 eV. The dissociative ionisation limits of SF₆that lie

below 22 eV are shown in Fig. 6. The high energies of the limits

mean that only Ne⁺, F⁺, Ar⁺, N₂⁺, and N⁺may react with

SF₆by dissociative charge transfer. All other reactions must

occur by chemical pathways, for example:

with 0 K enthalpy changes given by:

½D_rH₀^oð52Þ=kJ mol^ꢀ1ꢈ ¼ 124 ꢀ ½D₀ðF

and

N₂OÞ=kJ mol^ꢀ1ꢈ ð55Þ

A

^þþ SF₆! SF₅^þþ A F

ð49Þ

½D_rH₀^oð53Þ=kJ mol^ꢀ1ꢈ ¼ 69 ꢀ ½D₀ðFO N₂Þ=kJ mol^ꢀ1

ꢈ

ð56Þ

Williams et al.⁷have demonstrated that the reactions of SF₆

with positive ions possessing recombination energies greater

than 14.2 eV conform well to the expectations of models of

long-range charge-transfer reactions. They also found evidence

to indicate that the reaction of SF₆with Ne⁺is mediated by

short-range interactions. On thermochemical evidence they

There appears to be no independent evidence for the existence

of any isomer of FN₂O, and D₀(F–N₂O) and D₀(FO–N₂) are

unknown.

F^ꢀabstraction from SF₆to form SF₅is exothermic for

+

CF⁺, CF₂⁺, and CF₃⁺, but for each of these reactions the

experimental rate coeﬃcient (k_exp) is less than the calculated

+

concluded that the minor SF₃product from the reaction of

N⁺with SF₆is the result of a chemical pathway leading to

the formation of NF as one of the neutral products.

value from Langevin capture theory (k_c), and for CF₃

k_expꢇ 0.25 ꢃ k_c. The eﬃciencies of the reactions, deﬁned as

k

_exp/k_c, are plotted in Fig. 2 against the ﬂuoride aﬃnities of

+

The slow reaction with CO₂is interesting. The obvious

the reagent ions, to show the correlation between reaction eﬃ-

ciency and reagent ion ﬂuoride aﬃnity.

+

pathway to the observed SF₅product is dissociative charge

transfer:

A minor product from the reaction of CF⁺is CF₃⁺. Sequen-

tial transfer of two F atoms is unlikely to be the mechanism for

the ﬁrst step (to form CF₂⁺+ SF₅) is highly endothermic

CO₂^þþ SF₆! SF₅^þþ CO₂þ F

ð50Þ

(D_rH^o¼ 156 kJ mol^ꢀ1). Instead the transfer must be con-

0

certed:

but this pathway is so endothermic (D_rH^o₀(50) ¼ 39 kJ mol^ꢀ1

)

that this reaction should not be observable with our SIFT

apparatus at 300 K. It seems improbable that even a small

fraction of the CO₂reagent ions possess suﬃcient internal

ð57Þ

+

energy to surmount the endothermicity. Instead, following

the suggestion of Babcock and Streit,³the proposed pathway

The greater complexity of this scheme, compared to the much

less exothermic F^ꢀabstraction leading to SF₅⁺, provides an

explanation for the higher branching ratio to form SF₅com-

for the formation of SF₅is F^ꢀtransfer:

+

pared to CF₃⁺. With CF₂⁺as the reagent ion, there are several

pathways of greater exothermicity than F^ꢀabstraction:

CO₂^þþ SF₆! SF₅^þþ FCO₂

ð51Þ

to form the weakly bound molecule FCO₂, providing a path-

way to SF₅with a lower overall enthalpy change:

CF₂^þþ SF₆

+

8

D_rH₀^o¼ ꢀ132 kJ mol^ꢀ1

CF ^þþ SF₅

ð58Þ

ð59Þ

ð60Þ

>

3

<

ꢀ1

!

SF₃^þþ F þ CF₄

SF₄^þþ CF₄

D_rH₀^oð50Þ ¼ D_rH₀^oð49Þ ꢀ D₀ðF CO₂Þ

ð52Þ

D_rH₀^o¼ ꢀ193 kJ mol

>

:

D_rH₀^o¼ ꢀ241 kJ mol^ꢀ1

FC(O)O has been characterised experimentally by a variety of

spectroscopic methods.^39,43,44The measurement of the electron

aﬃnity of FCO₂has led to the determination of D₀(F–

yet the ion products of these reactions were not observed.

+

For the SF_nreagent ions, only one exothermic pathway

was identiﬁed:

C(O)O) ¼ 48 ꢁ 13 kJ mol^{ꢀ1 39}

A very similar value (D_fH^o₀(FC(O)O) ¼ ꢀ376 ꢁ 13 kJ mol^ꢀ1

.

Theoretical calculations_ꢀa₁t₄₀th_,4e₁

G2 level have given D_fH^o₀(FC(O)O) ¼ ꢀ378 ꢁ 8 kJ mol

.

SF^þþ SF₆! SF₃^þþ SF₄D_rH₀^o¼ ꢀ187 kJ mol^ꢀ1ð61Þ

)

has resulted from the use of density functional theory to calcu-

late the enthalpy changes for a number of isodesmic reactions

involving FC(O)O.⁴²Combining these theoretical values with

D_fH₀^o(CO₂) and D_fH₀^o(F) from the JANAF tables²⁷gives

D₀(F–C(O)O) ¼ 61 ꢁ 15 kJ mol^ꢀ1. Either value for D₀(F–

C(O)O) leads to the conclusion that reaction (51) is exothermic.

The observed low rate coeﬃcient for the reaction of CO₂⁺with

SF₆(k_exp/k_c¼ 0.008) may reﬂect a barrier associated with F^ꢀ

abstraction from SF₆, which is not fully removed by the small

overall exothermicity of the reaction. Several other reactions in

which F^ꢀis abstracted from SF₆also have rate constants less

than predicted by capture theory (see below).

but this reaction, involving the transfer of two F atoms, was

not observed. In experiments using non-thermal reagent

ions,^5,45evidence has been presented for the occurrence of

the endothermic reaction:

SF₅^þþ SF₆! SF₃^þþ F₂þ SF₆D_rH₀^o¼ 276 kJ mol^ꢀ1

ð62Þ

and for:

SF₃^þþ SF₆! S₂F₇^þþ F₂

ð63Þ

The slow reaction of N₂O⁺with SF₆to form SF₅⁺provides

another example of a reaction where it is necessary to postulate

the formation of a weakly bound product in order to provide a

pathway with a suﬃciently low overall endothermicity to be

The reaction products are only accessible for reagent ions

which are translationally and/or internally excited, and the

reactions are not observed in SIFT experiments, where the ions

have been thermalised before they can react.

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2219

View Article Online

F^ꢀabstraction is calculated to be highly endothermic for

NO⁺and NO₂⁺, and SF₆does not react with either of these

reagent ions. SF₆has a much lower proton aﬃnity than

H₂O,⁴⁶and as a consequence there is no reaction between

H₃O⁺and SF₆.

sideration of the energetics indicates that in some cases the

appearance of certain fragment ions is only possible if new

bonds are formed. The great majority of these cases occur

for SF₅CF₃, where exothermic charge-transfer routes to

SF₃and SF₄in general require the formation of CF₄as

one of the neutral products:

+

A

^þþ SF₅CF₃! SF₄^þþ CF₄þ A

! SF₃^þþ F þ CF₄þ A

ð68Þ

ð69Þ

3.4 Comparisons

The parent cation SF₅X⁺(X ¼ CF₃, Cl, F) was not observed as

a product from any of the reactions. When it is formed by charge

transfer, it immediately dissociates. The most obvious pathways

for dissociative ionisation involve scission of one bond:

Reactions of this type can be identiﬁed in Table 1. The high

bond dissociation energy D₀(CF₃–F) ( ¼ 374 kJ mol^ꢀ1) is a

strong thermochemical driving force for these reactions of

+

SF₅CF₃. For SF₅Cl and SF₆, the fragment ions SF₄and

SF₃are not observed at energies well below the limits for

SF₅X ! SF₅^þþ X þ e^ꢀ

! X^þþ SF₅þ e^ꢀ

ð64Þ

ð65Þ

ð66Þ

ð67Þ

+

their formation by stepwise loss of atoms from SF₅Cl⁺and

+

SF₆(reaction scheme (33)). D₀(Cl–F) ( ¼ 247 kJ mol^ꢀ1) and

! SF₄X^þþ F þ e^ꢀ

! F^þþ SF₄X þ e^ꢀ

D₀(F–F) ( ¼ 154 kJ mol^ꢀ1) are much lower than D₀(CF₃–F),

and there are few examples of reactions requiring the forma-

tion of ClF or F₂. In addition, the XF forming pathways to

+

SF₄and SF₃can be viewed as dissociation of SF₅X⁺

(!SF₅+ X⁺) coupled with abstraction of F^ꢀfrom SF₅by

X⁺. This will be a much higher energy pathway for SF₅Cl

and SF₆than for SF₅CF₃, because of the higher ionisation

energy of Cl (12.97 eV) and F (17.42 eV) compared to CF₃

with smaller SF_p⁺(p< 5) and SF_qX⁺(q< 4) ions resulting from

+

stepwise fragmentation of SF₅and SF₄X⁺. F⁺was not

observed among the ion products for any of the reactions. Even

with Ne⁺, the reagent ion with the highest recombination

energy, the products of pathway (67) are energetically inaccessi-

ble for SF₆, and barely accessible for SF₅CF₃and SF₅Cl. Disso-

ciative channels (64) and (65) diﬀer only in the location of the

positive charge. For X ¼ Cl and F, channel (65) lies higher in

energy than (64), by 3.42 eV (330 kJ mol^ꢀ1) for X ¼ Cl, and

by 7.88 eV (760 kJ mol^ꢀ1) for X ¼ F, but for X ¼ CF₃channel

(65) is below (64) by 0.50 eV (48 kJ mol^ꢀ1). For channel (64) the

S–X bond is broken, channel (66) corresponds to breaking an S–

F bond. The energy diﬀerences between (64) and (66) can be

(9.04 eV). The formation of SF₃from the reactions of N⁺,

+

CO₂⁺, and H₂O⁺with SF₅Cl probably involves the formation

+

of ClF or ClF₂. In the case of SF₆, the formation of SF₃

from the reaction with F⁺requires the formation of F₂. Other

charge-transfer reactions of each of the molecules may occur

with the formation of new bonds, but without identiﬁcation

of the neutral products this cannot be demonstrated.

Almost all of the (presumed) charge-transfer reactions pro-

ceed with rate constants close to the values calculated by cap-

ture theories. The only clear exception is the reaction of Ne⁺

with SF₆. According to the long-range charge-transfer model

k_expꢇ k_cimplies that there is a match between the recombina-

tion energy of the reagent ion (RE) and a band in the photo-

electron spectrum of the molecule. A photoelectron spectrum

is not available for SF₅CF₃. For SF₅Cl there is a good match

between a band in the photoelectron spectrum²⁹and the recom-

bination energy of each of the reagent ions, including Ne⁺, for

which charge-transfer reactions are energetically possible (i.e.

RE > 12.3 eV). In the case of SF₆, the match for Ne⁺is poor,

which provides an explanation for the low rate coeﬃcient for

the reaction, compared to the capture theory prediction.⁷

Since the product neutrals from the reactions cannot be

detected, the identiﬁcation of charge-transfer reactions is based

on a combination of knowledge of the product ions and basic

energetic data. Without energetic data on the dissociative ioni-

sation limits of the molecules, obtained from various photoio-

nisation experiments^10,30,36and theoretical calculations,^24,25it

would be impossible to discern when the reaction mechanism

must change from charge transfer to a chemical pathway.

obtained using our estimates of D_fH^o(SF₄CF₃⁺) ( ¼ ꢀ221 to

0

ꢀ348 kJ mol^ꢀ1) and D_fH₀^o(SF₄Cl⁺) (ꢅ150 kJ mol^ꢀ1). For

SF₅CF₃the products of (60) lie considerably below those of

(66), the diﬀerence is between 1.13 and 2.45 eV (109 and 236

kJ mol^ꢀ1). In the case of SF₅Cl, the diﬀerence between (64)

and (66) is small, 0.24 eV (23 kJ mol^ꢀ1), with (66) being the lower

energy channel.

These energy diﬀerences play a signiﬁcant role in determin-

ing the products observed from the charge-transfer reactions

of the SF₅X molecules. In the case of SF₅CF₃, for reagent ions

with recombination energies in excess of 12.9 eV, the experi-

mentally determined limit for (65),¹⁰CF₃is the dominant

+

product ion. The ratio of SF₅to CF₃increases with the

recombination energy of the reagent ion, reﬂecting the higher

energy of (64) compared to (65). The energy diﬀerence between

+

(66) and (65) is quite large, and SF₄CF₃was observed as a

trace product from only a few of the reactions. The pattern

of the energies of the channels is quite diﬀerent for SF₅Cl

(compare Figs. 1 and 3), (64) is the lowest (12.32 eV^30,36),

(66) is only slightly higher (12.56 eV), with (65) much higher

(15.74 eV). The relative branching ratios for SF₅⁺and SF₄Cl⁺

vary greatly, and there is no obvious correlation with the

recombination energy of the reagent ion. The TPEPICO

experiments on the fragmentation of energy-selected SF₅Cl⁺

ions³⁶do not show these variations. The results thus show a

deviation from the predictions of long-range charge-transfer

theory. This suggests that when the energy diﬀerence between

two channels is small, short-range interactions within the ion-

molecule complex between formation by capture and break-up

leading to the release of the products can have a large eﬀect on

the relative yields of diﬀerent product ions. For SF₆the only

energetically accessible channel is (64), which is identical to

(66). Available energy appears to be the most important factor

in determining the outcome of the charge-transfer reactions.

+

The reagent ions CF_m(m ¼ 1–3) and SF_n(n ¼ 1–5) all

have recombination energies below 12.0 eV, so their reactions

with SF₅CF₃, SF₅Cl, and SF₆must occur via chemical path-

ways. The observed ion products of the reactions of these ions

with SF₅CF₃(mainly SF₃and CF₃⁺) show that they are

+

mechanistically quite diﬀerent from their reactions with SF₅Cl

(SF₄Cl⁺and SF₅⁺are the major products) and SF₆(primarily

yield SF₅⁺). While the major reaction pathways for SF₅Cl and

SF₆are simple Cl^ꢀand F^ꢀabstractions, for SF₅CF₃transfer

of F^ꢀto the reagent ion is combined with breaking the S–C

bond (see reaction schemes (18) and (19)). In large part the dif-

ferences between SF₅CF₃and the other two reactants appear

to follow the thermochemistry of the reaction pathways.

Mechanism (18) is the proposed pathway to SF₃⁺, which is an

observed product from the reactions of SF₅CF₃with each of the

+

Various SF_p(p < 5) and SF_qX⁺(q < 4) ions were observed

from the charge-transfer reactions. In many cases these can be

formed by fragmentation of SF₅and SF₄X⁺. However, con-

CF_m⁺and SF_n⁺reagent ions, except SF₃⁺and SF₂⁺. For SF₃

+

2220

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

it is not possible to distinguish product SF₃⁺from the reagent

ions. For SF₂⁺, the calculated enthalpy change for the produc-

to the other CF_m⁺and SF_n⁺reagent ions, CF₂⁺is expected to

have the highest aﬃnity for CF₃and Cl. The only example of

transfer of a Cl atom is the reaction of CF₂⁺with SF₅Cl, where

it is highly exothermic (see Table 2). The reaction is facilitated

by the relatively low SF₅–Cl bond strength (D₀(SF₅–Cl) ¼ 268

kJ mol^ꢀ1, compared to D₀(SF₅–CF₃) ¼ 374 kJ mol^ꢀ1, and

tion of SF₃(D_rH^o₀¼ ꢀ21 kJ mol^ꢀ1) reveals that pathway (18)

+

is less exothermic than for the reactions of other reagent ions

with SF₅CF₃in which SF₃is an observed product (see Table

+

1). This must be an important factor in making the reaction

of SF₂too slow to be detected. As the group X of SF₅X

D₀(SF₅–F) ¼ 447 kJ mol^ꢀ1). Using D_fH₀^o(CF₃CF₂⁺) ¼ 14 kJ

+

the transfer of CF₃from SF₅CF₃to CF₂is calcu-

changes from CF₃to Cl and then F pathway (18) becomes

increasingly less exothermic, due mainly to the decreasing

strength of the new X–F bond. In the case of SF₅Cl, pathway

mol^{ꢀ1 47}

,

lated to be exothermic (D_rH^o₀¼ ꢀ60 kJ mol^ꢀ1), but CF₃CF₂

+

was not observed. Transfer of CF₃(Cl^ꢀ) leads to the forma-

ꢀ

+

(18) is only calculated to be exothermic for the CF_mreagent

ions. The SF₃⁺product ion was only observed from the reaction

of CF⁺. With CF⁺as the reagent ion, the calculated exothermi-

tion of SF₅⁺. SF₅was detected as an ion product from a

ꢀ

number of reactions, and in each case the CF₃(Cl^ꢀ) transfer

pathway is calculated to be exothermic (for CF⁺reacting with

SF₅CF₃a rearrangement of the neutral product is required).

The only reactions which do not fall into the above cate-

gories are those of CF⁺with SF₅Cl (!CF₂Cl⁺), and SF₆

city (ꢀ23 kJ mol^ꢀ1) is less than for CF₃(ꢀ41 kJ mol^ꢀ1) and

+

CF₂(ꢀ80 kJ mol^ꢀ1) which suggests that the CF⁺reaction

may proceed by a diﬀerent, more exothermic pathway:

+

(!CF₃⁺). For SF₅CF₃, the CF₃product can result from

CF^þþ SF₅Cl ! SF₃^þþ CF₃Cl D_rH₀^o¼ ꢀ482 kJ mol^ꢀ1

ð70Þ

reaction scheme (19), but a very diﬀerent mechanism must

operate for SF₅Cl and SF₆. An exothermic route to CF₂Cl⁺

and CF₃⁺, respectively, is obtained by transferring Cl + F

and F + F, respectively, (mechanism (46) for SF₅Cl, and (57)

for SF₆) to the reagent CF⁺. This mechanism might also be

operative for SF₅CF₃, in which case the neutral product would

be SF₃–CF₃. The suggestion that the production of CF₃⁺from

SF₅CF₃occurs by a diﬀerent mechanism from the formation of

CF₂Cl⁺from SF₅Cl, and CF₃⁺from SF₆, is supported by the

+

There is a highly exothermic pathway to SF₃for CF₂

+

:

CF₂^þþ SF₅Cl ! SF₃^þþ CF₄þ Cl

D_rH₀^o¼ ꢀ373 kJ mol^ꢀ1ð71Þ

+

but SF₃was not one of the detected ion products from the

+

reaction of CF₂with SF₅Cl. For SF₆, pathway (18) to

+

very diﬀerent product ion yields (47% CF₃(SF₅CF₃), 3%

CF₂Cl⁺(SF₅Cl), and 5% CF₃(SF₆)).

SF₃⁺is in every case endothermic. There are exothermic path-

ways, analogous to (70) and (71) above, when the reagent ion

+

The major ion products of the reactions with CF_mand

is CF⁺and CF₂

:

+

SF_nare formed through mechanisms that commence with

the abstraction of F^ꢀ(Cl^ꢀin the case of the reactions of SF₅Cl

which form SF₅⁺). Fig. 2 shows how the eﬃciencies of the

reactions vary with the F^ꢀaﬃnity of the reagent ion. In addi-

tion to the anticipated general increase in overall eﬃciency

with F^ꢀaﬃnity, it is observed that for each of the reagent ions

the eﬃciencies are in the sequence SF₅Cl > SF₅CF₃> SF₆.

This does not follow the order of the energies needed for F^ꢀ

abstraction from the molecules (SF₅Cl < SF₆< SF₅CF₃),

which provides further evidence of mechanistic diﬀerences

for the reactions of SF₅CF₃compared to those of SF₅Cl and

SF₆. The reaction of SF⁺with SF₅Cl has an unexpectedly high

eﬃciency (0.67 for SF₅Cl, compared to 0.025 for SF₅CF₃). As

discussed earlier, on the basis of the observed ion products,

this reaction proceeds primarily by Cl^ꢀrather than F^ꢀabstrac-

CF^þþ SF₆! SF₃^þþ CF₄D_rH₀^o¼ ꢀ484 kJ mol^ꢀ1ð72Þ

CF₂^þþ SF₆! SF₃^þþ CF₄þ F D_rH₀^o¼ ꢀ194 kJ mol^ꢀ1

ð73Þ

+

but SF₃was not one of the detected product ions from the

reactions of CF⁺and CF₂with SF₆. Exothermicity alone is

+

not suﬃcient to predict the outcome of these ‘chemical’ ion-

molecule reactions.

Most of the reactions of SF₅CF₃lead to the formation of

+

CF₃as one of the ion products. Reaction scheme (19) is the

most likely pathway. Whenever (19) is exothermic or close to

+

thermoneutral, CF₃is detected. Cl⁺and F⁺were not

+

tion. The trends between CF⁺, CF₂⁺, and CF₃are the same

+

observed from the reactions of SF₅Cl and SF₆with CF_m

for each of the reactant molecules, with CF₃⁺in each case hav-

ing a signiﬁcantly lower reaction eﬃciency than both CF⁺and

CF₂⁺, even though the F^ꢀaﬃnity of CF₃⁺is greater than that

of CF⁺. This is suggestive of some steric factor impeding the

and SF_n⁺. This is due to the high ionisation energies of Cl

and F, compared to CF₃, making all pathways to Cl⁺and

F⁺endothermic.

The transfer of F^ꢀto the reagent ion to form SF₄X⁺was

observed whenever this pathway is exothermic. The occurrence

of this reaction for particular reagent ion/neutral reactant

combinations then depends on the diﬀerence between the ﬂuor-

+

reaction of the relatively bulky CF₃to a greater extent than

.

+

for CF⁺and CF₂

ide aﬃnity of the reagent ion and of SF₄X⁺. Using our bounds

for D_fH^o(SF₄CF₃⁺), the ﬂuoride aﬃnity of SF₄CF₃lies

+

4. Conclusions

0

between 1074 and 1201 kJ mol^ꢀ1, which is greater than for

+

SF₅(1040 kJ mol^ꢀ1) and SF₄Cl⁺(884 kJ mol^ꢀ1). This is

Rate coeﬃcients and product ion distributions are reported for

the reactions of twenty-two atomic and molecular reagent

cations with two derivatives of SF₆; SF₅CF₃and SF₅Cl. The

reagent ions have a wide range of recombination energies, so

that both charge transfer and chemical ion-molecule reactions

are observed. Most of the reactions proceed with rate coeﬃ-

cients close to the predictions of capture theories. There are

notable exceptions to this, as can be seen by inspecting the data

in Tables 1 to 3. A variety of reaction processes are evident,

including dissociative charge transfer and various abstraction

routes. SF₅X⁺(X ¼ F, Cl, CF₃) ions were not observed as

products of any of the reactions. Although there may well be

bound regions on the potential energy surfaces of the parent

ions, these are not accessed in charge-transfer reactions.

Instead, as seen in studies of the photoionisation of SF₅X,

reﬂected in the observation that the ion product of F^ꢀtransfer

is seen for many of the reactions of SF₅Cl, and for the reac-

tions of SF₆with each of the CF_nreagent ions. In contrast,

+

SF₄CF₃was seen only as a trace product from the reaction

of SF₅CF₃with CF₂⁺. F atom transfer may occur in the

reaction of CF₂with SF₅CF₃, but the observed CF₃pro-

duct may result from a number of other pathways, including

(19) and:

+

CF₂^þþ SF₅CF₃! CF₃^þþ SF₃þ CF₄

D_rH₀^o¼ ꢀ193 kJ mol^ꢀ1ð74Þ

For SF₅CF₃and SF₅Cl there is the possibility of transferring

ꢀ

CF₃(Cl^ꢀ) and CF₃(Cl) in addition to F^ꢀand F. Compared

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2221

View Article Online

the initially formed SF₅X⁺ions dissociate very rapidly over

repulsive energy surfaces.

and shown to be consistent with the observation of SF_pCl⁺

from some of the reactions of SF₅Cl. These include:

D_fH^o₀(SFCl⁺) ¼ 782 kJ mol^ꢀ1, and D_fH₀^o(SF₂Cl⁺) ¼ 453 kJ

mol^ꢀ1. Other estimated values are D_fH₀^o(SF₂OH⁺) ¼ 321 kJ

mol^ꢀ1, and D_fH₀^o(SF₄Cl) ¼ ꢀ761 kJ mol^ꢀ1. With one excep-

tion there are no observed reactions for which, allowing for

uncertainties in the values of the enthalpies of formation, the

enthalpy of reaction is greater than +20 kJ mol^ꢀ1. The one

exception is the slow reaction between N₂O⁺and SF₆, to form

The observed ion products show that the reactions of

SF₅CF₃are quite diﬀerent from those of SF₅Cl and SF₆.

The major products from the charge-transfer reactions of

+

reagent ions A⁺with SF₅CF₃are SF₃(+CF₄+ F + A) and

+

CF₃

(+SF₅+ A), while SF₅

(+Cl + A) and SF₄Cl⁺

+

(+F + A) are formed from SF₅Cl, and SF₅(+F + A) from

SF₆. These diﬀerences can be traced to the thermochemistry

of the reactions, the energies of the dissociative ionisation

pathways are shown in Figs. 1, 3, and 6. The high dissociation

energy of CF₃–F (374 kJ mol^ꢀ1) compared to F–Cl (247 kJ

mol^ꢀ1) and F–F (154 kJ mol^ꢀ1) accounts for the appearance

+

SF₅⁺. The least endothermic pathway to SF₅is:

N₂O^þþ SF₆! SF₅^þþ FO þ N₂D_rH₀^o¼ 69 kJ mol^ꢀ1

ð75Þ

+

of SF₃from SF₅CF₃. The low ionisation energy of CF₃

(9.04 eV) compared to SF₅(9.52 eV), Cl (12.97 eV), and F

but, as calculated, this is too endothermic to account for the

formation of SF₅⁺. It is unlikely that the weak binding

between FO–N₂can be suﬃciently strong to make the reac-

tions proceed with a high enough rate coeﬃcient to be obser-

vable by the SIFT technique at 300 K. Thus the observation

+

(17.42 eV) is responsible for the formation of CF₃and little

+

SF₅from SF₅CF₃, while the ion products from SF₅Cl are

SF₅⁺and SF₄Cl⁺, and SF₅⁺from SF₆. These energetic factors

also inﬂuence the products of chemical ion-molecule reactions

of the molecules. Thus the major ion products from SF₅CF₃

+

of SF₅from the reaction of SF₆with N₂O⁺suggests that

+

D_fH^o₀(SF₅⁺) ¼ 84 kJ mol^ꢀ125is too high. The experimental

are again SF₃and CF₃⁺, SF₅Cl produces mainly SF₅and

data on SF₅Cl may also point to a problem with D_fH^o(SF₅⁺).

If the observed appearance energy of 12.32 eV for the

SF₄Cl⁺, and SF₆yields SF₅

.

+

0

The charge-transfer reactions generally conform well to the

expectations of models of long-range charge-transfer pro-

cesses. However, it was noted that some of the reactions of

SF₅CF₃, speciﬁcally those with N⁺, CO⁺, CO₂⁺, O⁺, and

dissociative photoionisation (28)³⁰is combined with

D_fH^o₀(SF₅Cl) ¼ ꢀ1026 kJ mol

,

^{ꢀ1 27}obtained from calorimetric

measurements,²⁸then D_fH^o(SF₅⁺) ꢆ 43 kJ mol^ꢀ1. Chim et al.¹⁰

0

+

have sought to determine the dissociative ionisation energy of

SF₆(!SF₅⁺+ F + e^ꢀ) by extrapolation of TPEPICO data to

zero kinetic energy release. From a rather long extrapolation

they determined the dissociative ionisation energy to be 13.6

(ꢁ0.1) eV, when combined with standard literature values

N₂O⁺, produced more SF₃than expected on the basis of

TPEPICO experiments on the fragmentation of isolated energy

selected (SF₅CF₃⁺)*. Also the wide variations in the SF₄Cl⁺/

+

SF₅ratio with the recombination energy of the reagent ion

are not compatible with other data on the ionisation of SF₅Cl.

There is thus evidence that short-range interactions can inﬂu-

ence the outcome of fast charge-transfer reactions.

for

D_fH^o₀(SF₆)

and

D_fH^o₀(F)²⁷

this

result

gives

D_fH^o₀(SF₅⁺) ¼ 28 kJ mol^ꢀ1. D_fH^o₀(SF₅⁺) ¼ 84 kJ mol^ꢀ125is

strongly supported by other theoretical calculations.^48,49Iri-

kura⁵⁰has proposed that DG^o( ¼ DH^oꢀ TDS^o) rather than

DH^o₀might be used to predict the spontaneity (observation)

of gas-phase ion-molecule reactions. This introduces entropic

factors, which we have ignored in our analysis of the thermo-

dynamics of the reactions. For reactions in which there is an

increase in the number of molecules (e.g. (75)), DS^o> 0, mak-

ing DG^o< DH^o, so that if DG^o< 0 is the criterion, then

endothermic reactions may become observable. For the two

F^ꢀabstraction and processes in which F^ꢀabstraction is

coupled to other chemical transformations dominate the reac-

+

tions of ions such as CF_mand SF_n⁺, which have recombina-

tion energies below the lowest dissociative ionisation limits of

SF₅CF₃, SF₅Cl, and SF₆. Many of these reactions have rate

coeﬃcients below those predicted by capture theories, the eﬃ-

ciencies generally do correlate with the ﬂuoride aﬃnity of the

reagent ion. Lower than expected eﬃciencies for the reactions

of CF₃⁺, when compared against CF⁺and CF₂⁺, may be indi-

cative of steric factors associated with the relatively bulky

reactions DS^o₂₉₈(75) ¼ 180.76 J K^ꢀ1mol^{ꢀ1 27}

,

neglecting the

+

small temperature dependence of DH^o, DG₂^o₉₈(75) ¼ 15 kJ

CF₃reagent ion. As for the charge-transfer reactions, the

+

mol^ꢀ1, suggesting that SF₅might be produced by reaction

ion products show that SF₅CF₃reacts by pathways that are

mechanistically distinct from the routes followed by SF₅Cl

and SF₆. For SF₅CF₃the major products from the chemical

(75), and that the reaction would have a low eﬃciency. How-

ever, for reactions in which there is a change in the number

of molecules, both DS and DG change with the partial pres-

sures of the reactants and products. It is then not clear that

DG^o₂₉₈will be the correct quantity from which to predict

whether a particular ion-molecule reaction will be observable

using the SIFT technique.

+

reaction with A⁺are CF₃

(+SF₄+ AF) and SF₃

+

(+CF₄+ AF), SF₅Cl forms SF₅(+ACl) and SF₄Cl⁺

(+AF), and SF₆yields SF₅(+AF). These diﬀerences are

+

again attributed to the energetics - the low ionisation potential

of CF₃, and the high CF₃–F bond dissociation energy are

major factors.

We have chosen to base our thermochemical analysis of the

reactions on enthalpies of formation for CF_m⁺and SF_n⁺taken

from recent ab initio calculations.^24,25This immediately

impacts on the enthalpies of formation of SF₅CF₃and SF₅Cl,

Acknowledgement

We are grateful to the Technological Plasma Initiative Pro-

gram, EPSRC, (Grant Reference: GR/L82083) for the ﬁnan-

cial support of this study. Miss Clair Atterbury thanks

EPSRC for a research studentship. Professor P. J. Knowles

(School of Chemical Sciences, University of Birmingham) is

thanked for calculating the dipole moments of SF₅CF₃and

SF₅Cl.

+

which can be derived from appearance energies of SF₅+ CF₃

and SF₅⁺+ Cl, respectively, obtained from photoionisation

experiments.^10,30We have adopted D_fH₀^o(SF₅CF₃) ¼ ꢀ1673

kJ mol^ꢀ1and D_fH₀^o(SF₅Cl) ꢂ ꢀ985 kJ mol^ꢀ1

.

Central to our analysis of the results is the exothermicity cri-

terion, that for a particular ion product to be observed in a

SIFT study of a reaction, there must be a pathway that is

exothermic, or at least close to thermoneutral. Use of this cri-

terion has allowed us to propose values for the enthalpies of

formation of some of the product ions, these include: ꢀ348

References

1

2

3

4

F. C. Fehsenfeld, J. Chem. Phys., 1971, 54, 438.

M. T. Bowers and M. Chau, J. Phys. Chem., 1976, 80, 1739.

L. M. Babcock and G. E. Streit, J. Chem. Phys., 1981, 74, 5700.

R. J. Shul, B. L. Upschulte, R. Passarella, R. G. Keesee and A. W.

Castlemann, Jr., J. Phys. Chem., 1987, 91, 2556.

kJ

mol^ꢀ1< D_fH^o(SF₄CF₃⁺) < ꢀ221

kJ

mol^ꢀ1

,

and

0

D_fH^o₀(SF₄Cl⁺) ¼ 150 kJ mol^ꢀ1. Values for a number of

SF_pCl⁺ions were suggested by comparisons with SF_p+1⁺ions,

2222

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

View Article Online

5

6

7

8

9

Z. A. Talib and M. Saporoschenko, Int. J. Mass Spectrom. Ion

Phys., 1992, 116, 1.

R. A. Morris, T. M. Miller, A. A. Viggiano and J. F. Paulson, J.

Geophys. Res., 1995, 100, 1287.

T. L. Williams, L. M. Babcock and N. G. Adams, Int. J. Mass

Spectrom., 1999, 185–187, 759.

G. K. Jarvis, R. A. Kennedy, C. A. Mayhew and R. P. Tuckett, J.

Phys. Chem. A, 2000, 104, 10 766.

32 K. K. Irikura, J. Chem. Phys., 1995, 102, 5357.

33 Y.-S. Cheung, Y.-J. Chen, C. Y. Ng, S.-W. Chiu and W.-K. Liu,

J. Am. Chem. Soc., 117, 9725.

34 P. B. Armentrout, Adv. At. Mol. Opt. Phys., 2000, 43, 187.

35 P. Harland and J. C. J. Thynne, J. Phys. Chem., 1969, 73, 4031.

36 R. Y. L. Chim and R. P. Tuckett, personal communication,

2001.

37 T. J. Van Huis, J. M. Galbraith and H. F. Schaefer, Mol. Phys.,

1996, 89, 607.

38 G. K. Jarvis, R. A. Kennedy, C. A. Mayhew and R. P. Tuckett,

Int. J. Mass Spectrom., 2000, 202, 323.

W. T. Sturges, T. J. Wallington, M. D. Hurley, K. P. Shine, K.

Sihra, A. Engel, D. E. Oram, S. A. Penkett, R. Mulvaney and

C. A. M. Brenninkmeijer, Science, 2000, 289, 611.

10 R. Y. L. Chim, R. A. Kennedy, R. P. Tuckett, W. Zhou, G. K.

Jarvis, D. J. Collins and P. A. Hatherly, J. Phys. Chem. A,

2001, 105, 8403.

11 R. A. Kennedy and C. A. Mayhew, Int. J. Mass Spectrom., 2001,

206, vii–x.

39 D. W. Arnold, S. E. Bradforth, E. H. Kim and D. M. Neumark, J.

Chem. Phys., 1995, 102, 2493.

40 T. S. Dibble and J. S. Francisco, J. Phys. Chem., 1994, 98, 11 694.

41 W. F. Schneider, M. M. Maricq and J. S. Francisco, J. Chem.

Phys., 1995, 103, 6601.

12 C. Atterbury, R. A. Kennedy, C. A. Mayhew and R. P. Tuckett,

Phys. Chem. Chem. Phys., 2001, 3, 1949.

42 M. E. Tucceri, M. P. Badenes, A. E. Croce and C. J. Cobos, Phys.

Chem. Chem. Phys., 2001, 3, 1832.

13 D. Smith and N. G. Adams, Adv. At. Mol. Phys., 1988, 24, 1.

14 N. G. Adams and D. Smith, in Techniques for the study of ion-

molecule Reactions, ed. J. M. Farrar, and W. H. Saunders, Wiley,

New York, 1988, p. 165.

43 M. M. Maricq, J. Szente, Z. Li and J. S. Francisco, J. Chem.

Phys., 1993, 98, 784.

44 G. A. Arguello, H. Grothe, M. Kronberg, H. Willner and H.-G.

¨

Mack, J. Phys. Chem., 1995, 99, 17 525.

15 D. Smith and N. G. Adams, Int. J. Mass Spectrom. Ion Phys.,

1976, 21, 349.

45 K. Bederski and L. Wojcik, Int. J. Mass Spectrom. Ion Processes,

1996, 154, 145.

16 G. K. Jarvis, C. A. Mayhew and R. P. Tuckett, J. Phys. Chem.,

1996, 100, 17 166.

17 K. J. Miller and J. A. Savchik, J. Am. Chem. Soc., 1979, 101, 7206.

18 T. Su and W. J. Chesnavich, J. Chem. Phys., 1982, 76, 5183.

19 G. Gioumousis and D. P. Stevenson, J. Chem. Phys., 1959, 29,

294.

46 (a) D. R. Latimer and M. A. Smith, J. Chem. Phys., 1994, 101,

3410; (b) Erratum D. R. Latimer and M. A. Smith, J. Chem.

Phys., 1994, 101, 10 197.

47 C. W. Bauschlicher and A. Ricca, J. Phys. Chem. A, 2000, 104,

4581.

48 K. K. Irikura, J. Chem. Phys., 1995, 102, 5357.

49 Y.-S. Cheung, Y.-J. Chen, C. Y. Ng, S. W. Chiu and W.-K. Li, J.

Am. Chem. Soc., 1995, 117, 9725.

20 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC

Press Inc., 80th edn., 1999.

21 D. S. Peterka, M. Ahmed, C.-Y. Ng and A. G. Suits, Chem. Phys.

Lett., 1999, 312, 108.

22 E. R. Fisher, M. E. Weber and P. B. Armentrout, J. Chem. Phys.,

1990, 92, 2296.

23 J. C. Creasey, I. R. Lambert, R. P. Tuckett, K. Codling, L. J. Fra-

sinski, P. A. Hatherly and M. Stankiewicz, J. Chem. Soc., Faraday

Trans., 1991, 87, 1287.

24 A. Ricca, J. Phys. Chem. A, 1999, 103, 1876.

25 C. W. Bauschlicher and A. Ricca, J. Phys. Chem. A, 1998, 102,

4722.

26 G. Hansel, B. Gstir, M. Probst, P. Scheier, N. J. Mason and T. D.

Ma¨rk, J. Phys. B, At. Mol. Opt. Phys., in press.

27 M. W. Chase, Jr., NIST-JANAF Thermochemical Tables, J.

Phys. Chem. Ref. Data, Monograph, 4th edn., 2000, vol. 9.

28 H. L. Leach and H. L. Roberts, J. Chem. Soc., 1960, 4693.

29 R. L. DeKock, B. R. Higginson and D. R. Lloyd, Faraday Dis-

cuss. Chem. Soc., 1972, 54, 84.

50 K. K. Irikura, J. Am. Chem. Soc., 1999, 121, 7689.

51 NIST Chemistry WebBook, NIST Standard Reference Database

Number 69, eds. W. G. Mallard and P. J. Linstrom, February

2000, National Institute of Standards and Technology, Gaithers-

burg MD, 20 899 (http://webbook.nist.gov).

52 D. A. Dixon and D. Feller, J. Phys. Chem. A, 1998, 102, 8209.

53 F. Louis, D. R. Burgess, M.-T. Rayez and J.-P. Sawerysyn, Phys.

Chem. Chem. Phys., 1999, 1, 5087.

54 K. P. Huber and G. Herzberg, Constants of Diatomic Molecules

(data prepared by J. W. Gallagher and R. D. Johnson, III) in

NIST Chemistry WebBook, NIST Standard Reference Database

Number 69, eds. P. J. Linstrom and W. G. Mallard, July 2001,

National Institute of Standards and Technology, Gaithersburg

MD, 20 899 (http://webbook.nist.gov).

55 J. Sehested, T. Ellerman, O. J. Nielsen and T. J. Wallington, Int.

J. Chem. Kinet., 1994, 26, 615.

56 S. J. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R. D.

Levin and W. G. Mallard, J. Phys. Chem. Ref. Data, Suppl.,

1988, 1, 17.

30 H. Baumga¨rtel, H.-W. Jochims, E. Ruhl, O. Lo¨sking and H. Will-

¨

ner, Z. Naturforsch. B, 1989, 44, 21.

31 E. R. Fisher, B. L. Kickel and P. B. Armentrout, J. Chem. Phys.,

1992, 97, 4859.

57 N. L. Haworth, M. H. Smith, G. B. Bacskay and J. C. Mackie, J.

Phys. Chem. A, 2000, 104, 7600.

Phys. Chem. Chem. Phys., 2002, 4, 2206–2223

2223

Article Doi

A study of the gas phase reactions of various cations with two derivatives of SF6; SF5CF3 and SF5Cl

DOI: 10.1039/b109567f

Source and publish data:

Authors:

Article abstract of DOI:10.1039/b109567f

Full text of DOI:10.1039/b109567f

Products guided by the article

R&D Labs maybe for 75-73-0

Relevant to this article

Hot Product