Turbulent ion flow tube study of the cluster-mediated reactions of SF6- with H2O, CH3OH, and C2H5OH from 50 to 500 Torr

Article abstract of DOI:10.1021/jp003967y

A class of cluster-mediated reactions that includes the reactions of SF₆ ^- with H₂O, CH₃OH, and C₂H₅OH was studied at 298 K from 50 to 500 torr in a newly constructed turbulent ion flow tube. These reactions were cluster-mediated reactions, where the reactants are in equilibrium with a stable association complex and then products are formed via a bimolecular reaction of the complex with a second neutral reactant. Primary product ions for the H₂O reaction include SOF₄^- and F^-(HF)₂, and primary products for the CH₃OH and C₂H₅OH reactions include F^-(HF)(ROH) and F^-(HF)₂. SF₆^- was proposed as a chemical ionization agent for detection of atmospheric trace neutrals, and it is presently being used as such in the dry environment of the South Pole.

Full text of DOI:10.1021/jp003967y

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
J. Phys. Chem. A 2001, 105, 3527-3531
3527
-
Turbulent Ion Flow Tube Study of the Cluster-Mediated Reactions of SF6 with H2O,
CH3OH, and C2H5OH from 50 to 500 Torr
Susan T. Arnold* and A. A. Viggiano
Air Force Research Laboratory, Space Vehicles Directorate, 29 Randolph Road,
Hanscom Air Force Base, Massachusetts 01731-3010
ReceiVed: October 26, 2000; In Final Form: January 16, 2001
-
Reactions of SF6 with H2O, CH3OH, and C2H5OH were studied at 298 K from 50 to 500 Torr in a newly
constructed turbulent ion flow tube (TIFT). These reactions were found to be cluster-mediated reactions,
where the reactants are in equilibrium with a stable association complex and then products are formed via a
bimolecular reaction of the complex with a second neutral reactant. By use of known equilibrium constants,
rate constants were derived for the reaction of the complex with the second solvent molecule. The rate constants
for these reactions, which are essentially pressure-independent, are 3.2 × 10-14 cm3 s-1, 1.8 × 10-13 cm3 s-1,
and 2.2 × 10-13 cm3 s-1 for H2O, CH3OH, and C2H5OH, respectively. Primary product ions for the H2O
-
reaction include SOF4 and F-(HF)2, and primary products for the CH3OH and C2H5OH reactions include
F-(HF)(ROH) and F-(HF)2. The observed ions are frequently solvated. The product distributions for the
alcohol reactions show pressure dependencies from 50 to 500 Torr. These results have implications for the
-
modeling of low-temperature corona or glow-type discharges, as well as for the use of SF6 as a chemical
ionization mass spectrometry agent in humid environments.
-
Introduction
500 Torr. We have also examined the related reactions of SF6
with CH3OH and C2H5OH over this pressure range.
It is well-known that the high-voltage gaseous dielectric, SF6,
reacts in the presence of trace oxygen or water vapor in electrical
discharges to form neutral byproducts that include such species
as SOF2, SOF4, SO2F2, SO2, and HF.1,2 A variety of ionic by-
products are also observed, including H3O+(H2O)n, SF5+(H2O)n,
SOF3+(H2O)n, SOFx-, F-(HF)n, OH-(H2O)n, SF6-(HF)n, and
SF6-(H2O)n. Some of the proposed reaction pathways for the
A previous report by Knighton et al.9 indicated that at 2-3
-
Torr the reaction of SF6 with methanol is a cluster-mediated
reaction:
Keq
SF6- + CH3OH 7 8 SF6 (CH3OH) k [CH OH]8 products (2)
2
3
-
-
formation of various negative ions involve SF6 reacting with
where the reactants are in equilibrium with the association
complex and then a bimolecular reaction of the stable cluster
ion with another CH3OH molecule yields products. A class of
either the impurities or the byproducts of neutral reactions.
Unfortunately, kinetics for the fundamental ion reactions have
not been thoroughly investigated in many cases. For example,
-
cluster-mediated reactions that includes the reactions of SF6
-
there are numerous reports of SOF4 being observed in SF6
with H2O, CH3OH, and C2H5OH has been investigated in our
laboratory. We report rate constants and product distributions
for these reactions at 298 K over the pressure range from 50 to
500 Torr.
discharges1 and in the ion source of an atmospheric pressure
mass spectrometer,3 and it has been suggested that the ion results
from the following reaction:
SF6- + H2O f SOF4- + 2HF
(1)
Experimental Section
Measurements were made with the Air Force Research
Laboratory’s recently constructed turbulent ion flow tube (TIFT).
This instrument has been described previously in detail.10 For
However, the kinetics of reaction 1 has not been examined; only
-
the thermochemistry of the clustering reaction between SF6
and H2O has been reported.4 This reaction also suggests a role
for trace water vapor in the formation of HF, a species that
attacks insulators once it is formed in the discharge.
-
this study, SF6 reactant ions were formed and injected into
the flow tube by coflowing a mixture of 1% SF6 in helium, at
a rate of approximately 1 cm3 min-1, with buffer gas (typically
60 L min-1) into the flow tube through an in-line ionizer. An
additional flow of buffer gas, typically 15 L min-1, was
introduced into the flow tube via another inlet in order to
maintain turbulent flow. Most experiments utilized a nitrogen
buffer gas, which was obtained from the gas delivery port of a
high-pressure liquid N2 dewar. Some experiments also required
the use of a helium buffer (He, 99.997%), although the excessive
flow rates needed to maintain turbulent flow limited the use of
the helium buffer gas. Neutral reactants (absolute ethanol, U.S.
Industrial Chemicals; 99.9% methanol, Fisher; and distilled
Because H2O is often present as an impurity in SF6
discharges, a more thorough understanding of the reaction
between SF6- and H2O would have application to the modeling
of low-temperature corona, partial discharges, or glow-type
discharges. In addition, understanding the reaction of SF6- with
H2O is important for chemical ionization mass spectrometry
(CIMS) because H2O is present in large amounts (and is
presumed to be unreactive) when SF6- is used as a CIMS agent
in humid atmospheres.5-8 This study clarifies the interaction
of SF6- with H2O at 298 K over the pressure range from 50 to
10.1021/jp003967y This article not subject to U.S. Copyright. Published 2001 by the American Chemical Society
Published on Web 03/16/2001

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3528 J. Phys. Chem. A, Vol. 105, No. 14, 2001
Arnold and Viggiano
water) were introduced into the flow tube through a bubbler
system where nitrogen or helium carrier gas was passed through
a sample of the reactant at a rate of up to 1 L min-1. The reactant
flow rate was derived from the known vapor pressure of the
solvent along with the nitrogen flow rate and the total pressure
in the bubbler; the reactant flow rate was roughly proportional
to the square root of the nitrogen flow rate. The flow tube
temperature was maintained at 298 K; the tube pressure was
varied from 50 to 500 Torr by throttling the exhaust valve while
keeping the flow rate constant. A small fraction of the gas in
the tube flowed through a sampling orifice, and the reactant
and product ions in this flow were mass-analyzed in a quad-
rupole mass filter and detected by a particle multiplier. Kinetics
were measured by monitoring the reactant and product ion count
rates as a function of the flow rate of the neutral reactant gas.
The reaction times were determined from the neutral reactant
inlet distance and from the ratio of the ion velocity to the buffer
velocity, which has been characterized previously at 298 K by
time-of-flight techniques as functions of both the Reynolds
number and the tube pressure.
Figure 1. Demonstration of initial equilibrium for SF6- reacting with
H2O and C2H5OH at 300 Torr. K′ ) [SF6-(ROH)]/{[SF6-][ROH]}.
Results
-
Over the pressure range from 50 to 500 Torr, SF6 reacts
with ROH (R ) H, CH3, C2H5) in a manner similar to the
scheme reported by Knighton et al.9 that is shown in reaction
2. The only primary product ions observed in these reactions
are the stable association complexes, SF6-(ROH). The fact that
the reactants are in equilibrium is demonstrated by plotting
-
[SF6 (ROH)]
K′ )
(3)
-
[SF6 ][ROH]
as a function of the solvent concentration, as shown in Figure
1 for the H2O and C2H5OH reactions at 300 Torr. Relative ion
concentrations are provided by the measured ion intensities,
while the neutral density is obtained from the derived reactant
flow rate. The magnitude of K′ is found to be essentially
independent of the solvent concentration, except at very low
concentrations, indicating that the reactants rapidly establish
equilibrium with the association complex.
Evidence for the involvement of an additional solvent
molecule in the overall reaction is provided by the kinetics data.
For the complex mechanism shown in reaction 2, the integrated
rate equation obtained by use of the equilibrium approximation
is11
-
Figure 2. Example of data for reaction of SF6 with CH3OH at 200
[SF6 ] ) [SF6 ]0 exp(-Keqk2[ROH]2t)
(4)
-
-
Torr plotted as (a, top panel) pseudo-first-order kinetics and (b, bottom
panel) pseudo-second-order kinetics. The species are represented as
follows: (b) SF6; (O) SF6-(CH3OH); (arrowhead pointing right)
F-(HF)2; (4) F-(HF)2(CH3OH); (3) F-(HF)2(CH3OH)2; (+) F-(HF)-
(CH3OH); (×) F-(HF)(CH3OH)2; (0) F-(HF)(CH3OH)3.
under conditions where [ROH] . [SF6-]0. If this mechanism
is correct, the reaction is second-order with respect to the solvent
concentration. Note that if products were to form instead via a
unimolecular dissociation of the cluster complex, then the rate
expression would be
The applicable rate equation (eq 4) further indicates that if
Keq were known, then k2 could be determined from the linear
slope of the plot of ln [SF6-]/[SF6-]0 vs [ROH]2. Ideally, data
such as that shown in Figure 1 could be used to determine
equilibrium constants for these reactions. Unfortunately, the
“equilibrium constants” obtained in this manner are not inde-
pendent of the total buffer gas pressure; K′ is found to increase
with increasing buffer gas pressure from 50 to 500 Torr, as
shown in Figure 3 for the H2O reaction. This increase in K′ is
an experimental artifact that can be attributed to the formation
of additional SF6-(ROH) upon sampling the high-pressure flow
tube gases into vacuum through an orifice 50 µm in diameter.
-
-
[SF6 ] ) [SF6 ]0 exp(-Keqk1[ROH]t)
(5)
-
An example of the kinetics data for the reaction of SF6 with
CH3OH at 200 Torr is shown in Figure 2, plotted vs [CH3OH]
in the top panel and vs [CH3OH]2 in the bottom panel. The
data clearly demonstrate that the decays of SF6- are consistent
with the involvement of two solvent molecules. The [H2O]2
-
dependence for the SF6 and H2O reaction has recently been
confirmed by Huey et al.12

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
-
TIFT Study of Cluster-Mediated Reactions of SF6
J. Phys. Chem. A, Vol. 105, No. 14, 2001 3529
Figure 3. Pressure dependence of K′ for H2O reaction, shown for both
N2 and He buffers. Actual Keq, as determined by Sieck,4 is shown on
y-axis.
TABLE 1: Rate Constants, k2, and Product Distributions
from 50 to 500 Torr for the Reactions of SF6-(ROH) with
an Additional ROH Moleculea
product distributions
F-(HF)2-
(ROH)n
F-(HF)-
(ROH)n
SF4O--
(ROH)n
b
pressure
(Torr)
k2
(10-13 cm3 s-1
)
H2O
50
100
300
500
0.37
0.26
0.31
0.33
0.21
0.14
0.16
0.16
0.79
0.86
0.84
0.84
-
Figure 4. Mass spectra for reactions of SF6 with (a) H2O and (b)
CH3OH at 100 Torr.
to be pressure-independent and are larger than that for the H2O
reaction, 1.8 ((0.7) × 10-13 cm3 s-1 and 2.2 ((0.9) × 10-13
cm3 s-1 for CH3OH and C2H5OH, respectively. The reported
errors represent the combined uncertainties in the equilibrium
measurements and the current rate measurements. All the
measured reaction rate constants are significantly smaller than
the collisional rate constants calculated from the parametrized
trajectory model of Su and Chesnavich;14,15 for H2O, CH3OH,
and C2H5OH, the collisional rate constants are 2.2 × 10-9 cm3
s-1, 1.8 × 10-9 cm3 s-1, and 1.6 × 10-9 cm3 s-1, respectively.
The final products of these cluster-mediated reactions are
illustrated in the mass spectra shown in Figure 4. For the H2O
reaction (top panel), product ions include SOF4-(H2O)n and
F-(HF)2(H2O)n. For the CH3OH reaction (bottom panel),
products include F-(HF)(CH3OH)n and F-(HF)2(CH3OH)n. The
reaction of SF6- with C2H5OH is analogous to that of CH3OH,
yielding F-(HF)(C2H5OH)n and F-(HF)2(C2H5OH)n. The pri-
CH3OH
50
100
200
300
500
2.6
1.3
1.5
2.2
1.7
0.39
0.50
0.56
0.60
0.65
0.61
0.50
0.44
0.40
0.35
C2H5OH
50
100
200
300
500
1.5
1.6
2.4
2.6
2.9
0.20
0.57
0.71
0.73
0.80
0.43
0.29
0.27
a Collisional rate constants for H2O, CH3OH, and C2H5OH reactions
are 2.2 × 10-9 cm3 s-1, 1.8 × 10-9 cm3 s-1, and 1.6 × 10-9 cm3 s-1
,
respectively. b k2 was determined from Keq values of 1400, 3225, and
5320 atm-1 for the reactions of SF6- with H2O, CH3OH, and C2H5OH,
respectively, as reported in ref 4.
To illustrate this effect, several experiments were also carried
out with helium buffer gas, which should result in less cluster
formation upon sampling.13 As expected, a significantly smaller
increase in K′ is observed with the helium buffer gas. The value
of K′ is also affected by the fact that clusters may dissociate
upon sampling. These sampling complications make it impos-
sible to accurately determine equilibrium constants in this
instrument at this time. However, the sampling problems should
not affect rate constant or primary product determinations.
Clustering of the products with the solvent molecules is affected
by this problem.
Actual equilibrium constants, Keq, measured by high-pressure
mass spectrometry have been reported by Sieck4 for the reactions
of SF6- with H2O, CH3OH, and C2H5OH. Use of Sieck’s values
for Keq now allows k2 to be determined from the present kinetics
measurements made at 50-500 Torr. These results are shown
in Table 1. The reaction rate constant for SF6-(H2O) reacting
with another H2O is pressure-independent, within our experi-
mental uncertainty, and is approximately 3.2 ((1.3) × 10-14
cm3 s-1. The rate constants for both alcohol reactions also appear
mary product ions that are formed in the bimolecular reaction
-
of SF6-(ROH) with another solvent molecule are SOF4
,
F-(HF)2, and F-(HF)(ROH). Subsequently, these product ions
become further solvated by (a) additional reactions with the
solvent [most of which are in equilibrium, as evidenced by plots
such as that shown in Figure 1, where the ratio [A-(ROH)]/
([A-][ROH]) remains constant with increasing solvent concen-
tration] and by (b) sampling the high-pressure flow tube gas
into vacuum. Thus, to determine product branching fractions,
it is necessary to sum the solvated and unsolvated forms of each
product ion. Branching fractions are determined by recording
the product ion count rates as a function of the neutral reactant
flow rate. To account for the effects of additional reactions
between the product ions and the neutral reactants, the reported
branching fractions are determined by extrapolating the meas-
ured branching fractions to a neutral reactant flow rate of zero
as shown in Figure 5 for both H2O and C2H5OH reactions.
Product distributions determined in this manner are reported in
Table 1. In addition to yielding the product branching fractions,
these plots further demonstrate that, for CH3OH and C2H5OH,

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3530 J. Phys. Chem. A, Vol. 105, No. 14, 2001
Arnold and Viggiano
Figure 6. Pressure dependence of the product distributions. Only the
F-(HF)2(ROH)n product channel is shown. The remaining products are
SF4O-(H2O)n for the H2O reaction and F-(HF)(ROH)n for the
C2H5OH and CH3OH reactions.
Torr. However, given the range of the present data, it is difficult
to reconcile the product distributions that were reported previ-
ously, 72% F-(HF)2(CH3OH)n and 28% F-(HF)(CH3OH)n, with
the current measurements. This difference remains unexplained.
For all three solvents, the rate constants for the bimolecular
reactions that ultimately lead to products are significantly less
than the collisional rate constant. The small rate constants are
not surprising given that products must form from rather
complex reaction mechanisms involving numerous bonds being
broken and re-formed. In a previous study, Huey et al.5 proposed
-
a mechanism for the reaction of SF6 with a number of
Figure 5. Product plot for (a, top panel) C2H5OH reaction at 500 Torr
atmospherically important neutrals (ClNO3, HOCl, HCl, N2O5,
and HNO3). Their suggested mechanism involved the formation
of an ion-molecule collision complex, with a sufficient lifetime
to allow extensive rearrangement within the complex, followed
by unimolecular dissociation of the complex to the observed
product ions. The kinetics data presented here indicate clearly
that this is not the case for the reaction of SF6- with the protic
solvents water, methanol, and ethanol. At 298 K, these reactions
require two solvent molecules to form product ions other than
the association complexes.
and (b, bottom panel) H2O reaction at 300 Torr.
the F-(HF)2(ROH)n product ions undergo additional ligand
switching, e.g.
F-(HF)2(C2H5OH)n + C2H5OH f
F-(HF)(C2H5OH)n+1 + HF (6)
No ligand switching was found in the H2O reactions. The rate
constants for the switching reactions are much lower than the
collision rate constants, and the reactions are most likely
endothermic since ligand switching is generally rapid for
exothermic reactions.16 Comparing the reactant ion signal loss
to the sum of the product ion signals yields a maximum absolute
uncertainty in the branching fractions of 25%.
Unlike the reaction rate constants, the product distributions
for the reactions of SF6-(ROH) with an additional ROH
molecule are pressure-dependent, as shown in Figure 6. The
H2O reaction produces mainly SOF4-. The product distribution
has at most a weak pressure dependence. In contrast, the alcohol
reactions, which yield predominately F-(HF)2 at higher pres-
sures, are pressure-dependent over the entire range studied. The
C2H5OH reactions were found to have a more significant
pressure dependence than the CH3OH reactions.
The fact that multiple solvent molecules play an essential
role in a reaction is not without precedent; a dramatic example
is the reaction of SO3 with H2O to form H2SO4. Morokuma
and Muguruma17 reported that the energetically most favorable
transition state along the reaction pathway is actually a six-
centered cyclic transition state involving two water molecules.
Their proposed mechanism involves two water molecules
concertedly transferring their hydrogen-bonding protons to the
proton acceptor oxygen atoms, while a new S-O bond is formed
between the newly generated hydroxyl group and SO3. This
implies the formation of H2SO4 from SO3 and H2O should be
second-order with respect to the H2O pressure, which is
consistent with the experimental findings.18-20
-
In the case of SF6 reactions with alcohols, the ionic prod-
ucts of the cluster-mediated reactions include F-(HF)2 and
F-(HF)(ROH):
Discussion
SF6 (ROH) + ROH f F-(HF)2 + (SF3O2 + R2)
-
Qualitatively, reactions of SF6- with the protic solvents H2O,
CH3OH, and C2H5OH appear to behave according to the scheme
first outlined for methanol by Knighton et al.,9 where the
formation of product ions is mediated by the initial formation
of a stable solvent complex. The present measurement of the
rate constant for the reaction of SF6-(CH3OH) with CH3OH is
in relatively good agreement with the previous measurement
of 3.3 × 10-13 cm3 s-1 made at 35 °C in a CH4 buffer at 2-3
F-(HF)(ROH) + (SF3O + RF) (7)
The neutral products of these reactions, shown in parentheses,
are not characterized in this experiment; they are assumed, on
the basis of most favorable energetics.21,22 Given that SF6- might
be more appropriately represented as (SF5)F- rather than as a