Studies on the Cl + C2H5I reaction; Site specific abstraction reactions and thermodynamics of adduct formation studied by observation of HCL product

Article abstract of DOI:10.1039/b907793f

The reaction of chlorine atoms with alkyl iodides can play a role in the chemistry of the marine boundary layer. Previous studies have shown that at room temperature the reaction takes place via a complex mechanism including adduct formation. For the Cl + ethyl iodide reaction results on the thermodynamics of adduct formation and on the product yields are inconsistent. The kinetics of the reaction Cl + C₂H₅I have been studied by the direct observation of the HCl product in real time flash photolysis/IR absorption experiments as a function of temperature from 273 to 450 K. At temperatures above 375 K kinetic measurements confirm a direct process and the rate coefficient determined (4.85 ± 0.55) × 10^-11 exp((-363 ± 51)/T) cm³ molecule^-1 s^-1 is in good agreement with previous direct determinations. Product yield studies have also been undertaken by comparing the HCl signal from Cl + C₂H ₅I with that from a calibration reaction which shows that HCl is the sole product of the reaction at these temperatures. Yield studies with selectively deuterated ethyl iodide demonstrate that abstraction occurs predominantly from the α site, with the selectivity decreasing with temperature. Extrapolation of the yield data to 298 K predicts an α: β ratio of 0.68: 0.32. At temperatures between 273 and 325 K a biexponential growth was observed for the HCl signal consistent with adduct formation. Analysis of the HCl time profiles allowed the extractions of the forward and reverse rate coefficients for adduct formation and hence the calculations of the thermodynamic properties of adduct formation. A third law analysis yields a value of Δ_rH = (-54 ± 4) kJ mol ^-1. The value of Δ_rH is in good agreement with a previous third law determination (J. J. Orlando, C. A. Piety, J. M. Nicovich, M. L. McKee, P. H. Wine, J. Phys. Chem. A, 2005, 109, 6659).

Full text of DOI:10.1039/b907793f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Studies on the Cl + C H I reaction; site specific abstraction reactions
2
5
and thermodynamics of adduct formation studied by observation
of HCL product
R. Wada, R. C. Sharma, M. A. Blitz and P. W. Seakins*
Received 17th April 2009, Accepted 20th August 2009
First published as an Advance Article on the web 23rd September 2009
DOI: 10.1039/b907793f
The reaction of chlorine atoms with alkyl iodides can play a role in the chemistry of the marine
boundary layer. Previous studies have shown that at room temperature the reaction takes place
via a complex mechanism including adduct formation. For the Cl + ethyl iodide reaction results
on the thermodynamics of adduct formation and on the product yields are inconsistent. The
kinetics of the reaction Cl + C H I have been studied by the direct observation of the HCl
2
5
product in real time flash photolysis/IR absorption experiments as a function of temperature from
2
73 to 450 K. At temperatures above 375 K kinetic measurements confirm a direct process and
ꢂ11
3
ꢂ1 ꢂ1
the rate coefficient determined (4.85 ꢀ0.55) ꢁ10
exp((ꢂ363 ꢀ51)/T) cm molecule
s
is in good agreement with previous direct determinations. Product yield studies have also been
2 5
undertaken by comparing the HCl signal from Cl + C H I with that from a calibration reaction
which shows that HCl is the sole product of the reaction at these temperatures. Yield studies with
selectively deuterated ethyl iodide demonstrate that abstraction occurs predominantly from the
a site, with the selectivity decreasing with temperature. Extrapolation of the yield data to 298 K
predicts an a : b ratio of 0.68 : 0.32. At temperatures between 273 and 325 K a biexponential
growth was observed for the HCl signal consistent with adduct formation. Analysis of the HCl
time profiles allowed the extractions of the forward and reverse rate coefficients for adduct
formation and hence the calculations of the thermodynamic properties of adduct formation.
ꢂ1
A third law analysis yields a value of D
agreement with a previous third law determination (J. J. Orlando, C. A. Piety, J. M. Nicovich,
r
H = (ꢂ54 ꢀ4) kJ mol . The value of D
r
H is in good
M. L. McKee, P. H. Wine, J. Phys. Chem. A, 2005, 109, 6659).
Introduction
At lower temperatures (typically o250 K) adduct formation
becomes irreversible and at higher temperatures (typically 4350 K)
adduct decomposition is so rapid that abstraction is the only
effective removal process. The focus of this paper is on the
reaction of Cl with ethyl iodide (C H I).
Alkyl iodides have been observed in the marine boundary
,2
1
layer, MBL, and when liberated by reaction or photolysis,
the iodine contained plays an important role in particle
2
5
3
,4
formation. Whilst photolysis is the dominant removal pro-
cess, radical reactions play a significant role and under certain
conditions the reaction with Cl atoms can effectively compete
k2a;kꢂ2a
Cl þC2H5I
!
C2H5I ꢃꢃꢃCl
ðR2aÞ
5
with other removal processes. Methyl iodide is the most
abundant alkyl iodide but significant concentrations of higher
1
k2b
0
Cl þRI ꢂ!HCl þR I
ðR2bÞ
,2
alkyl iodides have been observed. The potential importance
of the Cl atom reaction with alkyl iodides has led to a number
Orlando et al. studied both the simpler and more complex
regimes of reaction (2) by flash photolytic generation of Cl
atoms in an excess of ethyl iodide, monitoring Cl removal by
5
–12
of kinetic and product
studies. Such studies have
5
ꢂ3
confirmed that under elevated [Cl] (410 molecule cm ) Cl
reactions will compete with photolytic removal, but more
interestingly have shown that the reaction mechanism is
complex; at ambient temperatures Cl atoms are removed by
reversible formation of an adduct or by abstraction:
6
resonance fluorescence. Under simple conditions, where
either irreversible addition (R2a) or abstractions (R2b)
dominate, Cl decayed exponentially according to pseudo first
order kinetics, but under conditions where adduct formation
occurred, the Cl decays were biexponential. Analysis of such
decays allowed extraction of k2a and kꢂ2a and hence the
equilibrium constant for the reaction. The thermodynamic
k1a;kꢂ1a
Cl þRI
!
RI ꢃꢃꢃCl
ðR1aÞ
k1b
Cl þRI ꢂ!HCl þR0I
parameters D
r
H and D
r
S could then be obtained from a van’t
ꢂ1
ðR1bÞ
Hoff plot (ln K vs. 1/T) yielding D H = (ꢂ63.4 ꢀ4.0) kJ mol
p
r
ꢂ1 ꢂ1
and D S = (ꢂ114 ꢀ15) J mol
K
(for the reaction
School of Chemistry, University of Leeds, Leeds, UK LS2 9JT.
E-mail: p.w.seakins@leeds.ac.uk
r
Cl + C D I at 298 K) or D S could be calculated allowing
2
5
r
This journal is ꢄc the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 10417–10426 | 10417

View Article Online
D
r
H to be obtained by the third law method giving D
r
H =
with irreversible formation of the adduct at temperatures
around 200 K and with reversible formation of the adduct at
room temperature. Analysis of the adduct temporal profiles
ꢂ1
6
(
ꢂ53.3 ꢀ4.0) kJ mol at 298 K.
6
In the same study Orlando et al. also observed the stable
3
ꢂ1
products formed at approximately 720 Torr in a chamber
experiment with FTIR analysis. They concluded that abstraction
from the a position was favored with the a : b ratio being
yielded an equilibrium constant, (70 ꢀ20) ꢁ10 atm , that is
in good agreement with earlier observations. No reaction of
the adduct with oxygen was observed, and at low temperatures
(200 K) adduct removal was dominated by rapid reactions
approximately 1.5 : 1, but under their conditions, [C
2
H
5
I]
molecule cm , there appeared to be significant reaction
of the adduct with the ethyl iodide with a rate coefficient of
0
E
1
5
ꢂ3
1
0
3
with radical species such as CH .
Orlando and co-workers initial direct studies on the
equilibration of reaction (2a) were based on observation of
ꢂ13
3
ꢂ1 ꢂ1
(
2–5) ꢁ 10
cm molecule s . The corresponding
6
the Cl atom, more recently they have directly observed the
reaction for methyl iodide had been proposed by Bilde and
8
Wallington.
14
adduct using UV absorption. In these studies the reaction
5
,9
The reactions have also been studied by Cotter et al. The
was carried out at low temperatures, 250 K, where the adduct
formation is essentially irreversible. The rate coefficient for
reaction (2a) obtained from monitoring the adduct was
in good agreement with the earlier studies monitoring Cl
removal. Interestingly the cross sections for the adduct
between 405–500 nm are approximately a factor of four below
room temperature kinetics were measured at low pressures
(
1.5–10.0 Torr) in a flow tube experiment reporting a value,
ꢂ11
ꢂ3 ꢂ1 5
k2 = (1.60 ꢀ0.17) ꢁ10
molecule cm
s , that is
consistent with the extrapolations of the higher temperature
9
direct data of Orlando et al. However, Cotter et al. draw
1
0
different conclusions from their end product studies carried
out in bag reactor at 760 Torr also with FTIR analysis. Firstly,
the earlier CRDS measurements.
Whilst the room temperature kinetics measured by Cotter
et al. and Orlando et al. for the removal of Cl atoms appear to
be in reasonable agreement, the site specific rate coefficients
are very different and the two groups assign differing roles to
the halogen atom displacement channel (2c). Additionally
there is uncertainty about the reactivity of the adduct with
2 5
as with Orlando et al., C H Cl was observed in significant
amounts (46%), but this was attributed to a primary channel:
k2c
Cl þC2H5I ꢂ!C2H5CI þI
ðR2cÞ
6
rather than to secondary reactions involving the adduct.
Support for a direct substitution process in such systems
comes from recent theoretical calculations that have reported
the chamber studies of Orlando et al. and Bilde and
8
Wallington suggesting significant reactivity with the parent
1
1
iodide, while Gravestock et al.’s direct measurement of the
complex was consistent with adduct removal by radical adduct
reactions. Clearly considerable uncertainties remain as to the
mechanism of the reaction and the reactivity of the adduct. In
this paper we take a complementary approach to the study of
reaction (2) observing the production of HCl product in real
time following pulsed laser generation of Cl atoms in the
presence of excess ethyl iodide. The kinetics and HCl yields
of selectively deuterated ethyl iodides (CD CH I and
3
a 13% CH Br yield from the corresponding reaction of Br
1
3
with CH
CD CH I replaced C
abstraction occurred exclusively at the a position.
Direct observation of the adduct, C
IꢃꢃꢃCl, has been
reported in absorption by Enami et al. and more recently
3
I. Secondly, as no DCl was observed when
3
2
2 5
H I, Cotter et al. concluded that
2
H
5
1
0
1
1
for CH IꢃꢃꢃCl by fluorescence by Gravestock et al. Enami
3
et al. studied a range of adducts via cavity ring down spectro-
3
2
scopy (CRDS) following the reaction of Cl atoms with alkyl
0,12
CH CD I) have been studied to address the site specificity of
3 2
1
halides.
The absorption spectrum of C
2
H
5
IꢃꢃꢃCl in
the abstraction process.
their experimental region of 400–650 nm was characterized
by a smooth monotonic decrease in cross section from
Experimental
ꢂ17
2
cm molecule at 400 nm to approximately
ꢂ1
about 2.5 ꢁ10
ꢂ17
2
cm molecule at 632 nm. Absorption data were
ꢂ1
0
.1 ꢁ10
The reactions were studied in a slow flow flash photolysis
reaction cell with generation of Cl radicals by excimer laser
photolysis and observation of the HCl product from abstraction
by tunable IR diode laser absorption. The reaction cell has
used to characterize the formation of the adduct and to
investigate possible reactions with oxygen. Reaction with
oxygen could have atmospheric significance but the temporal
profiles of adduct growth and decay were unaffected by the
change from nitrogen to synthetic air, suggesting a very slow
reaction of the adduct with oxygen that would be of no
relevance in the atmosphere. Experimental studies on the
adduct were supplemented by theoretical calculations
15–17
been described in previous publications.
It is an adaptation
1
8
of a Herriot cell design by Pilgrim et al. that provides a
region of overlap between the photolysis and probe laser
beams in the central portion of the cylindrical cell facilitating
accurate temperature control in the observation region and
still allowing for efficient overlap of pump and probe beams.
Gas flows in the cell were controlled by calibrated mass flow
controllers. Precursor and substrates, diluted in helium, were
premixed with the helium buffer gas prior to entering the
reaction cell. Pressures were measured using a baratron gauge
and flow rates were adjusted to ensure replenishment of the
gas mix between photolysis pulses. The central portion of the
stainless steel reactor was surrounded by a ceramic heater
which provided uniform heating across the interaction region.
(
B3LYP with 6-311++G(d,p) basis sets for C and H and
6
-311G(d,p) for I and Cl) which yielded a binding energy
ꢂ1
of 61.7 kJ mol in good agreement with the second law
experimental determination of Orlando et al.
1
1
Gravestock et al. have made the first observation of the
adduct of methyl iodide and atomic chlorine via laser induced
fluorescence. Monitoring the time dependence of the adduct
gives similar kinetic behavior to the earlier Cl observations of
7
10
Ayhens et al. and the absorption studies of Enami et al.,
1
0418 | Phys. Chem. Chem. Phys., 2009, 11, 10417–10426
This journal is ꢄc the Owner Societies 2009

View Article Online
Temperature profiles along the cell were measured using a
moveable thermocouple.
the samples was checked by NMR and FT-IR after every
experiment.
Cl atoms were generated from two precursors. A majority of
the equilibrium studies were carried out using the photolysis of
molecular chlorine at 351 nm (XeF mix in a Lambda Physik,
LPX200 laser). Laser energies were varied between 20 and
Results and discussion
(
1) High temperature kinetics (350–450 K)
The reactions of Cl atoms with C
CD CH I have been studied under pseudo first order conditions
ꢂ2
ꢂ1
8
0 mJ cm pulse and a central 12 mm diameter portion of
2
5 3 2
H I, CH CD I and
the beam was passed into the cell. Other experiments were
3
2
ꢂ2
ꢂ1
)
carried out at 193 nm (ArF mix, typically 50 mJ cm pulse
1
at temperatures greater than 350 K, where adduct formation is
insignificant. Under these conditions the decay of Cl atoms is
controlled by the following processes (illustrated for C H I):
9
with oxalyl chloride as the precursor. Neither precursor is
ideal. Initial experiments were conducted at 193 nm as there
2
5
2
was concern about possible chain reactions with a Cl source.
k2
However significant photolysis of the alkyl iodides at 193 nm
complicates the analysis of the equilibrium studies, but does
provide information on the reactivity of the adduct to radical
species. Subsequent experiments showed that chain reactions
Cl þC2H5I ꢂ!Products ðincluding HCIÞ ðR2bÞ
kloss
Cl ꢂ!
where kloss is the rate coefficient for loss processes that are
ðR3Þ
involving Cl do not appear to be important in these systems
2
and a majority of the equilibrium and yield experiments have
been reported with molecular chlorine as the precursor.
HCl product was analyzed by IR absorption on the R(2)
independent of C
2
H
5
I, primarily diffusive loss which can be
0
approximated as a first order process and k
2
is the pseudo
first order loss rate coefficient for reaction (2). The time
dependence of the Cl atom concentration is then given by:
ꢂ1
line (2944.91377 cm ). IR radiation was generated from a
tunable diode laser (Laser Components, L5736) with a mono-
chromator to ensure single mode operation. The frequency
stability of the output was not sufficient to allow simple
absorption experiments to be performed. The output of the
probe laser was frequency locked to the correct wavelength by
modulating the frequency at 4 kHz and monitoring the output
from a lock-in amplifier (EG&G Brookdeal, 9503-SC). The
modulation was switched off 1 ms before the excimer fired and
remained at line centre for the duration of the kinetic trace
[
Cl]
t
= [Cl]
0
exp(ꢂk0
2
+ kloss)t
(E1)
The differential equation controlling the concentration of
HCl is:
d½HClꢅ
0
¼
k ½Clꢅꢂkdiff ½HClꢅ
ðE2Þ
tot
dt
and the corresponding relationship for the monitored HCl
product is:
(
typically 1–5 ms) before being switched on again between
0
Ak
0
t
tot
ðeꢂktot ꢂe Þ
ꢂkdiff t
excimer pulses. This technique simplifies the analysis and
allows for the use of relatively slow lock-in amplifiers. Further
½HClꢅ¼
ðE3Þ
t
0
0
tot
k ꢂk
diff
1
6
details can be found in Qian et al.
The IR radiation was coupled into the chamber by a CaF
where A is a constant related to the initial Cl atom concentration
0
0
2
and k
(=k + kloss) can be found from the appropriate
2
tot
window and passed into the Herriot cell through a small hole
drilled into one of the gold coated spherical mirrors of the
Herriot optical arrangement. The IR beam was reflected from
the outer region of the Herriot mirrors (40 mm diameter); the
central portions of the mirrors have been drilled out and allow
the photolysis beam to enter the cell. The Herriot cell was set
for 27 passes giving a total path length of approximately 27 m,
but only approximately 30% of the path overlapped with the
photolysis beam. A small flow of buffer gas was passed over
the mirrors to prevent the build up of deposits or attack by Cl
precursors. On exciting the Herriot cell through another small
hole in the opposite mirror to entry, the IR beam was focused
onto the element of an MCT detector (EG&G Judson,
J15012). The signal was captured and averaged using in house
software mounted in a PC.
non-linear least squares fit to the time dependence of the HCl
signal (inset in Fig. 1). k can then be found by varying the
2
0
ethyl iodide concentration and plotting k tot against [C
H
2 5
I] as
shown in Fig. 1.
The rate coefficients were determined to be independent of
radical concentration, laser power and pressure (20–70 Torr).
A potential problem for measurements with molecular chlorine
as the radical precursor is the potential for chain chemistry
with the radical products of the primary reaction reacting with
chlorine to regenerate molecular chlorine, e.g.
k4
CH CHI þCl ꢂ!CH CHClI þCl
3
2
3
ðR4Þ
Chain chemistry can usually be suppressed by carrying out the
reaction with a significant oxygen concentration to intercept
the primary product, generating a peroxy radical that is
unreactive on the timescale of the measurements:
He (BOC, CP Grade) and O
used as supplied. Cl (Aldrich, 99.5%) was diluted by He to
approximately 5% and stored in a darkened glass bulb.
Mixtures of C I (Sigma-Aldrich, 99%) were prepared in a
glass bulb after freeze-and-vacuum operation to purify and
diluted by helium to typically 5%. The C I was prepared on
a daily basis. The deuterated samples, CD CH I (CIL, 99.4% d)
and CH CD I (CIL, 99.8% d), were prepared in a glass bulb
2
(BOC, Zero Grade) gases were
2
k5
CH3CHI þO2 ꢂ!CH3CHIO2
ðR5Þ
2
H
5
Experiments were carried out with an oxygen concentration of
6
1
ꢂ3
2
H
5
9 ꢁ10 molecule cm . Typically k
5 4
/k is in the range
2
0
3
2
0.1–1.0 and therefore the added oxygen was sufficient to
capture over 95% of the primary radicals. No change was
observed in the recorded rate coefficients suggesting that
3
2
in a same way with C H I. The exchange of D with H of
2
5
This journal is ꢄc the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 10417–10426 | 10419

View Article Online
contact times accessible to Cotter et al. However, adduct
formation does occur under these conditions and the rate
coefficient associated with Cl loss cannot be directly attributed
to the abstraction process.
The reactions of Cl atoms with the partially deuterated
compounds CD CH I and CH CD I were also studied at
3
2
3
2
temperatures at which abstraction dominates in order to provide
information on site specific abstraction rate coefficients. The
site specific rate coefficients are defined as below:
kaH
Cl þCD3CH2I ꢂ!HCl þCD3CHI
kbD
Cl þCD3CH2I ꢂ!DCl þCD2CH2I
kaD
Cl þCH3CD2I ꢂ!DCl þCH3CDI
Fig. 1 Pseudo first order plot for reaction (2) at high temperature
T = 405 K) with a typical HCl time profile and fit as an inset.
kbH
Cl þCH3CD2I ꢂ!HCl þCH2CD2I
(
The total reaction rate coefficients kaH + kbD and kaD + kbH
are obtained under pseudo first order conditions over the
temperature range of 360–450 K for reaction of Cl atoms with
CD CH I and CH CD I. The resulting Arrhenius expressions
reactions such as R4 are ineffective at regenerating Cl atoms.
Molecular chlorine was also the source of Cl radicals used by
6
Orlando et al. in the flash photolysis component of
3
2
3
2
ꢂ11
that study.
are kaH + kbD = (6.72 ꢀ0.25) ꢁ10 exp(ꢂ(572 ꢀ16)/T)
ꢂ11
Experiments were then repeated over a range of temperatures
from 376 to 436 K with the highest temperature being determined
by the onset of thermal decomposition of the ethyl iodide and
the results for the rate coefficient for reaction (2) are compared
with previous work in Table 1 and Fig. 2.
and kaD + kbH = (5.8 ꢀ3.2) ꢁ10 exp(ꢂ(576 ꢀ311)/T).
The Arrhenius plots and least square fitting line are shown in
Fig. 3. The significant errors in the Arrhenius parameters for
kaD + kbH appear to arise from the rate coefficient determination
at 425 K, removing this from the analysis produces a revised
The current work is in very good agreement with the direct
study of Orlando et al. both in terms of the absolute measurements
at higher temperatures and the extrapolation to room
temperature where adduct formation prevents direct measurements.
Given that adduct formation is significant at room temperature,
Arrhenius equation of kaD + kbH = (1.1 ꢀ 0.5) ꢁ
ꢂ10
1
0
exp(ꢂ(800 ꢀ170)/T). Other than the fact that this point
lies some way from the best fit line, there is no obvious reason
why it should be ignored. Calculations based on both Arrhenius
parameterizations are presented below.
5
it may initially appear surprising that Cotter et al. reported
The reaction rate coefficients kaH + k and k + kbH at
aD
bD
ꢂ12
3
ꢂ1 ꢂ1
exponential decays in their studies at room temperature
and pressures of B2–12 Torr. Simulations using the adduct
formation parameters determined in the subsequent sections
show that the Cl decays are expected to be linear over the
2
98 K are estimated as (9.85 ꢀ0.37) ꢁ10
ꢂ1 ꢂ1
cm molecule s by extrapolation of
cm molecule s
ꢂ12
3
and (8.5 ꢀ4.6) ꢁ10
the high temperature data. Using kaD + kbH and kaH + kbD
with a reaction rate coefficients of Cl + C
2 5
H I, the a and b
site H atom abstraction rate coefficients are estimated as
ꢂ12
3
ꢂ1 ꢂ1
k
a
= 8.1 ꢁ10
cm molecule s
and k
ꢂ1 ꢂ1
cm molecule s . The branching ratio for the H atom
b
= 6.2 ꢁ
ꢂ12
3
1
0
abstraction pathways at 298 K is 0.57 and 0.43 for channels
a and b sites. The obtained branching ratio is in reasonably
good agreement with a reported value by Orlando et al.:
a : b = 0.63 : 0.37. Using the modified Arrhenius parameters
without the 425 K data for kaD + kbH gives a : b = 0.61 : 0.39.
(
2) High temperature (370–450 K) HCl yields
The HCl yield from reaction (2) has been studied at high
temperatures using the reaction of Cl atoms with n-butane as a
calibration reaction. Although it is possible to calculate
absolute HCl concentrations from absorption measurements,
uncertainties in the initial chlorine atom concentration and in
the effective path length result in significant uncertainties in
the HCl yield. An alternative method is to use a calibration
Fig. 2 Arrhenius plots for the reaction of C
2
H
5
I with Cl at T Z 376 K.
’) represent data obtained by this work, (K) Orlando et al.
absolute rate study), (m) Orlando et al. (relative rate study) and (.)
(
(
2
1
reaction. In this method the amount of HCl produced from a
Cotter et al.
reaction with an excess of a reagent (n-butane) with well
1
0420 | Phys. Chem. Chem. Phys., 2009, 11, 10417–10426
This journal is ꢄc the Owner Societies 2009

View Article Online
2 5
Table 1 Summary of kinetic data on Cl + C H I
a
a
Reference
k(T)
k
298
Comments
—
ꢂ11
ꢂ11
ꢂ11
b
ꢂ11 c
ꢂ11 c
This work
(4.85 ꢀ0.55) ꢁ10
exp((ꢂ363 ꢀ51)/T)
exp((ꢂ430 ꢀ210)/T)
1.43 ꢁ10
1.55 ꢁ10
—
6
6
6
Orlando et al.
Orlando et al.
Orlando et al.
(6.5 ꢀ3.4) ꢁ10
Laser flash photolysis study monitoring Cl removal
Cl + C
(2.21 ꢀ0.44) ꢁ10
exp((ꢂ317 ꢀ76)/T)
2 5
D I
ꢂ11
ꢂ11
—
—
(2.5 ꢀ0.4) ꢁ10
Relative rate study
Discharge flow
5
Cotter et al.
(1.60 ꢀ0.17) ꢁ10
a
3
ꢂ1 ꢂ1
b
c
Units of cm molecule
parameterization.
s
.
Errors in this work represent 1s uncertainties from the Arrhenius fit. Extrapolated from Arrhenius
The results are consistent with 100% abstraction to form
HCl with two important implications. Firstly, as expected
from the previous measurements on the adduct formation by
Orlando et al. and the measurements reported below, the
lifetime of the adduct is too short for removal by any
secondary processes. Secondly, there appears to be no
evidence for the direct substitution of Cl for I at these higher
temperatures. Although both the end product studies of
9
6
Cotter et al. and Orlando et al. report the production of
Cl at 298 K, Orlando et al. attribute this to secondary
reactions involving the adduct, whereas Cotter et al. suggest
that C Cl is a direct product of reaction (2); channel (2c).
Whilst the current results appear to preclude direct formation
of C Cl, it is still possible that this process does occur at
2 5
C H
2 5
H
2 5
H
lower temperatures as an alternative decomposition product of
the adduct.
Fig. 3 Arrhenius plots for the reaction of C
CD CH I (K, this work), CH CD I (m, this work) and C
., reported by Orlando et al.) with Cl atoms.
2 5
H I (’, this work),
3
2
3
2
2 5
D I
Similar yield studies, based on the calibration reaction of Cl
with n-butane, were carried out for the partially deuterated
ethyl iodide compounds over the temperature range 375–450 K.
The results are summarized in Fig. 5 and Table 2. It can clearly
be seen that abstraction from the a position dominates, and
that the specificity of the abstraction decreases with temperature.
A linear extrapolation of the variation of yield with temperature
predicts that the room temperature ratio is 0.68 : 0.32
consistent with both our kinetics based determinations
(
determined product yield (100%) is measured and then the
calibrant is replaced with an excess of the test substrate and
the experiment repeated with identical initial chlorine
concentrations and path length. In practice corrections need
to be made for any absorbance of the probe laser by the
substrate or calibrant. These small correction factors can be
calculated by observing the modulation of a chopped probe
beam in the presence of varying concentrations of substrate or
calibrant.
(especially using the modified Arrhenius parameters for reaction
of Cl with CH CD I) and the end product studies of Orlando
3
2
et al. The results are in contrast to the work of Cotter et al.
Fig. 4 shows a plot of the HCl yield from reaction (2) as a
function of temperature.
who observed no DCl production from the reaction of Cl with
CD
3 2
CH I, and hence concluded that abstraction only took
place at the a position.
(
3) Adduct formation and yields from 273 to 325 K:
thermochemistry of adduct formation
Below 350 K the growth of the HCl signal can no longer be
represented by a single exponential, indicative of the onset of
significant adduct formation. An example of such a trace is
shown in Fig. 6. Adduct formation was observed by both
6
10
Orlando et al. (T = 256–305 K) and Enami et al. monitoring
Cl atom removal and adduct formation, respectively.
However, the recent determinations of the cross sections for
1
4
the adduct suggest that there is some uncertainty surrounding
adduct properties which suggests that further studies on the
adduct are required. As we are dependent on HCl observation
we can only carry out such investigations in the temperature
region where adduct formation and abstraction are in
competition.
Fig. 4 HCl yield from reaction (2) as a function of temperature. The
solid line is the average value with an error reported at the 2s level.
This journal is ꢄc the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 10417–10426 | 10421

View Article Online
Fig. 5 HCl yield from the reaction of Cl with CD
CH CD I (K).
3 2
CH I (’), and
3
2
Temporal profiles of HCl were monitored over a variety of
conditions of temperature, pressure, radical and ethyl iodide
concentration and details of the experimental conditions can
be found in Table 3. The results from the fits for k2a and k
ꢂ2a
and the resulting equilibrium constant (k /k ) are tabulated
2
a ꢂ2a
in Table 3. Fig. 7a and b show the invariance of the equilibrium
constant with radical and substrate concentration. Yield
experiments were also performed at room temperature.
A comparison of the HCl signals from reaction (2) and from
the n-butane calibrant gives a yield of 0.80 ꢀ0.15 for the
HCl from reaction (2). This yield is inconsistent with the
mechanism used in Table 4 to analyze the triexponential
HCl profiles which would predict a yield of only slightly less
than unity due to diffusive loss of Cl and the adduct. If the loss
of Cl and adduct is purely diffusive then k and k can be
Fig. 6 Typical HCl time profile at room temperature with different
time scales; (a) short time scale and (b) longer time scale. The HCl
profile in (a) clearly shows a biexponential rise, which contains H atom
abstraction reaction and adduct formation HCl profile in (b) shows the
HCl signal decay mainly caused by HCl diffusion.
3
6
calculated from the well constrained HCl loss process (k7).
Although Fig. 7 shows that the equilibrium constant is invariant
6
with radical concentration, the loss rate for the adduct (k )
The HCl profiles have therefore been reanalyzed with the
expanded mechanism using the Facsimile numerical integration
does increase with radical concentration suggesting that
radical adduct reactions of the form
2
2
package. The best fit values for k
8
ꢂ10
and k
9
are 2.7 ꢁ
ꢂ10
3
cm molecule
ꢂ1 ꢂ1
3
cm molecule
ꢂ1 ꢂ1
s .
10
s
and 1.6 ꢁ10
adduct + adduct - product
adduct + C I - product
adduct + Cl - product
(R8)
(R9)
Similar values are obtained for k when the uncertainties in
0
1
2
H
4
k are taken into account. These values are strongly correlated
3
to each other and cannot be determined precisely; however k8
is in agreement with the values reported by Orlando et al. and
(R10)
1
4
ꢂ10
3
ꢂ1 ꢂ1
Dookwah-Roberts et al. ((2.0 ꢀ0.8) ꢁ10 cm molecule
s ).
do play a role.
Some basic sensitivity analysis has been carried out to
investigate the influence of reaction (8) and (9) on the returned
Temporal HCl profiles such as that shown in Fig. 6 were
analyzed numerically based on the model described in Table 4.
equilibrium constants. k
8
and k
9
were constrained to values of a
Table 2 a : b branching ratios from kinetics and yield studies
a
ab
Temperature/K
a : b from HCl yields
a : b from kinetics
a : b from kinetics
c
298
368
388
426
450
0.68 : 0.32
0.57 : 0.43
0.56 : 0.44
0.56 : 0.44
0.55 : 0.45
0.54 : 0.46
0.61 : 0.39
0.60 : 0.40
0.59 : 0.41
0.58 : 0.42
0.55 : 0.45
0.62 : 0.38
0.64 : 0.36
0.59 : 0.41
0.58 : 0.42
a
b
3 2
Based on an Arrhenius fit to the kinetic data and extrapolated to 298 K. Arrhenius parameters for Cl + CH CD I with data point at 425 K
c
removed. Linear extrapolation of yield data.
1
0422 | Phys. Chem. Chem. Phys., 2009, 11, 10417–10426
This journal is ꢄc the Owner Societies 2009

View Article Online
Table 3 Experimental conditions and results for adduct formation studies
a
a
a
I]
b
ass
c
Temperature/K
Press/Torr
[Cl]
0
[Cl
2
]
[C
2
H
5
k
k
diss
K
p
2
2
2
2
3
3
3
3
3
3
3
3
73
73
94
94
02
02
07
07
14
14
25
25
20
50
20
50
20
50
20
50
20
50
20
50
4.2
4.2
8.3
8.3
8.3
8.3
8.3
8.3
4.2
4.2
5.0
5.0
120
120
110
110
110
110
110
110
120
120
100
100
301
302
302
307
309
310
302
307
300
299
301
303
3.04
1.95
1.32
1.95
1.33
2.50
0.93
1.69
1.35
2.23
0.92
1.25
488
327
851
1 680 000
1 610 000
388 000
306 000
313 000
220 000
166 000
178 000
198 000
125 000
82 000
1592
1036
2769
1342
2275
1594
4173
2528
4538
62 000
a
12
Units of 10 molecule cm
ꢂ3
b
ꢂ11
3
cm molecule
ꢂ1 ꢂ1
c
ꢂ1
Units of s
.
Units of 10
s
.
.
Table 4 Mechanism for numerical analysis of HCl profiles with adduct formation
Reaction Value Comment
k2b
Cl þC2H5I ꢂ!HCl þC2H4I
4.85 ꢀ0.55) ꢁ10ꢂ11 exp((ꢂ363 ꢀ51)/T) Extrapolated from higher temperature measurements
(
k2a
Cl þC2H5I ꢂ!adduct
Floated
—
kꢂ2a
Adduct ꢂ!Cl þC2H5I
Floated
Floated
Floated
Floated
—
k3
Cl ꢂ!
Diffusive loss, reaction with adduct and any impurities
Diffusive loss process
k6
Adduct ꢂ!
k7
HCl ꢂ!
Diffusive loss and well constrained by the long time measurements
Fig. 7 Variation of the measured equilibrium constant with (a) radical and (b) substrate concentration. (a) Filled square: [Cl
photolysis laser power, open circles: photolysis laser power varied at constant [Cl ] to generate different [Cl].
2
] varied at constant
2
factor two above and below the best fit values reported above
k2a and kꢂ2a are primarily determined by the initial rapid rise in
and then the HCl traces were reanalyzed. The effect on k2a and
[HCl] whereas the adduct–adduct and adduct–product reactions
kꢂ2a was minimal, with the values changing by less than 1%.
are more significant during the slower rise and subsequent loss of
This journal is ꢄc the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 10417–10426 | 10423

View Article Online
HCl. Therefore whilst secondary reactions do introduce some
uncertainty into the role of adduct reactions, they appear to have
little effect on the parameters required for the calculation of the
thermodynamics of adduct formation.
In theory both D H and D S can be obtained from a van’t
r
r
Hoff plot, however, second law plots over relatively small
temperature ranges are subject to significant errors. In
many cases the entropy change for the reaction is known to
quite high levels allowing the intercept of the van’t Hoff plot
to be constrained (third law analysis). For reaction (2a),
there is some uncertainty in the calculated entropy, DS =
ꢂ1 ꢂ1
K , for the deuterated system. For the
(
ꢂ77.0 ꢀ8.0) J mol
hydrogenated system there is a slight change in the entropy to
ꢂ1 ꢂ1
mainly due to changes in the rotational
ꢂ78.0 J mol
K
entropy (the vibrations in the ethyl fragment are essentially
unchanged on formation of the adduct) and we have taken
the same error limits for DS. Fig. 9 shows the third law
analysis for the calculated entropy and the maximum
Fig. 9 Plots of ln K
p 2 5
vs. 1/T for reaction of C H I with Cl at
2
73 K r T r 325 K. (K) This work, (’) Orlando et al.
and minimum entropies based on the value of ln K
5
p
at 294 K;
ꢂ1
(
3.5 ꢀ0.4) ꢁ10 atm
.
similar to those observed by Orlando et al., see Table 5,
Fig. 8 shows the variation in the gradient (and hence D
r
H)
although in the opposite direction. Given that our third law
from the stated uncertainties in the entropy; D
r
H varies
p
calculations are based on K values where we have carried out
ꢂ1
by ꢀ2.2 kJ mol . The third law analysis returns a value of
systematic tests to check for the influence of radical–radical
and other reactions, we would recommend this value.
ꢂ1
D
deviation) is made up of the uncertainty from the scatter in
r
H = (ꢂ54 ꢀ4) kJ mol where the error (1 standard
Our third law determinations are in excellent agreement
with the third law determinations of Orlando et al. The third
law determinations from this work and from Orlando et al. lie
in between the values determined by the B3LYP calculations
of both Orlando et al. and Enami et al. and the higher level,
the data and from the uncertainties in the entropy.
Fig. 9 shows a van’t Hoff plot of the equilibrium data
ꢂ1
which yields D H = (ꢂ44 ꢀ 4) kJ mol
and D S =
r
from a second law analysis, where
r
ꢂ1 ꢂ1
(
ꢂ41 ꢀ13) J mol
K
2
4
the statistical components of the errors from the linear fit are
G2, calculations performed by Orlando et al. Given that
at the one standard deviation level. The corresponding
ꢂ1
there is now more evidence to support a value of D H around
r
ꢂ1
2 5
enthalpy of formation of C H ICl at 298 K would be 70 kJ mol
5
3 kJ mol further high level calculations might be useful.
ꢂ1 23
using heats of formation for Cl (121 kJ mol
ꢂ1 23
2 5
) and C H I
(
ꢂ7 kJ mol
). The errors in our determination of D
made up of a 2.5 kJ mol contribution of statistical errors
r
H are
ꢂ1
(4) Reactions of the adduct
ꢂ1
from the linear fit and an additional 1.5 kJ mol arising from
There have been several investigations into the reactivity of the
adduct. One possible reaction is with the parent iodide, first
8
proposed by Bilde and Wallington and supported by the end
uncertainties in the numerical model used to determine k2a and
k
. Fig. 9 also shows the values determined by Orlando et al.
ꢂ2a
and the thermodynamic data are summarized in Table 5.
product analysis studies of Orlando et al. A rate coefficient
ꢂ13
3
cm molecule
ꢂ1 ꢂ1
6
has been proposed which
Our values for D H from second and third law calculations
r
of B2 ꢁ10
s
are in agreement at the two standard deviation level. The
difference between our second and third law calculations are
under our conditions would result in less than 10% removal of
adduct. More recent direct studies on adduct removal rates by
1
Dookwah-Roberts et al. could not confirm reaction of the
4
adduct with ethyl iodide and reported a slightly higher upper
ꢂ13
3
cm molecule
ꢂ1 ꢂ1
limit of 5 ꢁ10
s . The invariance of K
p
with [C H I] (Fig. 7b) confirms that this reaction is not
5 0
2
significant under our conditions.
1
0
Several previous studies, particularly those of Enami et al.,
1
1
14
Gravestock et al. and Dookwah-Roberts et al. with direct
observation of the adduct, have demonstrated that the adduct
does not appear to react with molecular oxygen. The addition
of molecular oxygen under conditions where we observe
adduct formation does change the shape of the HCl formation
profiles, with the fast and slower rises being replaced with an
approximately single exponential rise. However, variations in
2
the yield of HCl with [O ] at room temperature are not
consistent with reaction of the adduct with molecular oxygen
in agreement with the previously mentioned studies. The shape
of HCl profiles is laser power dependent, which in conjunction
Fig. 8 Third law plot of our data using the entropy change based on
6
the calculations of Orlando et al.
1
0424 | Phys. Chem. Chem. Phys., 2009, 11, 10417–10426
This journal is ꢄc the Owner Societies 2009

View Article Online
Table 5 Summary of thermodynamic parameters for adduct formation
ꢂ1
ꢂ1
ꢂ1
K
Reference
D
r
H/kJ mol
D
r
S/ J mol
Comments
This work
This work
ꢂ44 ꢀ4
ꢂ41 ꢀ —1 3
Second law determination
Third law
Second law
ꢂ53.8 ꢀ3.8
Orlando et al.
Orlando et al.
Orlando et al.
Orlando et al.
Orlando et al.
Enami et al.
ꢂ63.4 ꢀ4.0
ꢂ53.3 ꢀ4.0
ꢂ114 ꢀ15
a
ꢂ77 ꢀ8
Third law
ꢂ57 ꢀ10
—
—
—
—
Recommended value from Orlando et al.
B3LYP/ECP calculation
G2, [ECP(S)]
b
ꢂ60
ꢂ39
b
ꢂ62
B3LYP/ECP calculation
a
b
Calculated using B3LYP. Corrected from 0 K calculations of Orlando et al.
with previous studies suggests that it is the reaction of the
approximately 80% were recorded under conditions where
the adduct is formed. The loss of adduct is consistent with
removal via a combination of radical–radical (adduct +
adduct and adduct + radicals), adduct + ethyl iodide and
diffusive loss.
adduct with radicals (e.g. CH CH(O )I, the peroxy radical
2
3
formed following a abstraction) that is responsible for the
change in HCl profile.
Summary
1
. Rate coefficients for the reaction of Cl atoms with
Acknowledgements
ethyl iodide have been measured over the temperature
range 350–450 K by observing HCl production. The rate
This work was funded by an EPSRC grant GR/T28560/01.
The authors appreciate discussions with Prof. Mike Pilling and
Prof. Dwayne Heard and with Dr David Glowacki and
Dr Baoshan Wang for some calculations. The authors also
acknowledge the helpful suggestions of the referees.
coefficient can be parameterized in Arrhenius format, k =
2
ꢂ11
(
4.85 ꢀ0.55) ꢁ10
exp((ꢂ363 ꢀ51)/T), and the values are
in excellent agreement with a previous study. Comparing the
yield of HCl with that from a calibration reaction shows that
over this temperature range, abstraction to form HCl accounts
for 100% of the reaction.
References
2
. The kinetics and HCl yields of selectively deuterated ethyl
1
Y. Yokouchi, H. Mukai, H. Yamamoto, A. Otsuki, C. Saitoh and
Y. Nojiri, J. Geophys. Res., [Atmos.], 1997, 102, 8805.
L. J. Carpenter, W. T. Sturges, S. A. Penkett, P. S. Liss, B. Alicke,
K. Hebestreit and U. Platt, J. Geophys. Res., [Atmos.], 1999, 104,
1679.
S. Pechtl, E. R. Lovejoy, J. B. Burkholder and R. von Glasow,
Atmos. Chem. Phys., 2006, 6, 505.
A. Saiz-Lopez, J. M. C. Plane, G. McFiggans, P. I. Williams,
S. M. Ball, M. Bitter, R. L. Jones, C. Hongwei and T. Hoffmann,
Atmos. Chem. Phys., 2006, 6, 883.
iodides have been studied over the same temperature range.
Branching ratios have been determined from an analysis of the
kinetics and more directly from the HCl yields. The two
approaches produce consistent results with the a channel
dominating. Extrapolation to room temperature yields values
of the a : b ratio of 0.57 : 0.43 and 0.68 : 0.32. We prefer to give
more weight to the more direct yield determinations. Both
estimates of the yield are in good agreement with the end
product studies of Orlando et al. (0.63 : 0.37) but contrast with
the 100% a abstraction reported by Cotter et al.
2
3
4
5
E. S. N. Cotter, N. J. Booth, C. E. Canosa-Mas, D. J. Gray,
D. E. Shallcross and R. P. Wayne, Phys. Chem. Chem. Phys., 2001,
, 402.
3
6 J. J. Orlando, C. A. Piety, J. M. Nicovich, M. L. McKee and
3. At lower temperatures (273–325 K) the HCl time profiles
P. H. Wine, J. Phys. Chem. A, 2005, 109, 6659.
Y. V. Ayhens, J. M. Nicovich, M. L. McKee and P. H. Wine,
J. Phys. Chem. A, 1997, 101, 9382.
8 M. Bilde and T. J. Wallington, J. Phys. Chem. A, 1998, 102, 1550.
9 E. S. N. Cotter, N. J. Booth, C. E. Canosa-Mas and R. P. Wayne,
Atmos. Environ., 2001, 35, 2169.
0 S. Enami, S. Hashimoto, M. Kawasaki, Y. Nakano, T. Ishiwata,
K. Tonokura and T. J. Wallington, J. Phys. Chem. A, 2005, 109,
1587.
1 T. J. Gravestock, M. A. Blitz and D. E. Heard, J. Phys. Chem. A,
2008, 112, 9544.
are consistent with the reversible formation of an adduct
confirming earlier studies monitoring both Cl removal and
the direct observation of a species presumed to be the adduct.
7
4. Measurement of the forward and reverse rate coefficients
allows for the determination of the thermodynamic
parameters of radical formation. We have determined
1
that the enthalpy for the adduct formation reaction
ꢂ1
1
(
(ꢂ54 ꢀ4) kJ mol ) is in excellent agreement with a previous
ꢂ1
6
experimental third law determination ((ꢂ53.3 ꢀ4) kJ mol ).
1
2 S. Enami, T. Yamanaka, S. Hashimoto, M. Kawasaki and
K. Tonokura, J. Phys. Chem. A, 2005, 109, 6066.
The third law determinations lie in between previous G2
and B3LYP calculations.
13 K. K. Irikura and J. S. Francisco, J. Phys. Chem. A, 2007, 111,
852.
6
5
. HCl yields under conditions where adduct formation is
1
4 V. Dookwah-Roberts, J. M. Nicovich and P. H. Wine, J. Phys.
Chem. A, 2008, 112, 9535.
15 N. Choi, M. J. Pilling, P. W. Seakins and L. Wang, Phys. Chem.
Chem. Phys., 2006, 8, 2172.
6 H. B. Qian, D. Turton, P. W. Seakins and M. J. Pilling, Chem.
Phys. Lett., 2000, 322, 57.
7 H. B. Qian, D. Turton, P. W. Seakins and M. J. Pilling, Int. J.
Chem. Kinet., 2002, 34, 86.
occurring preclude the direct formation of C Cl at the levels
2 5
H
9
proposed by Cotter et al. (46%), however, we cannot rule the
possibility that rearrangement of the adduct leads to products
such as C H Cl at levels of less than 10%.
1
2
5
6
. The adduct does not appear to react directly with oxygen
1
confirming the results of previous studies. HCl yields of
This journal is ꢄc the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 10417–10426 | 10425

View Article Online
1
8 J. S. Pilgrim, R. T. Jennings and C. A. Taatjes, Rev. Sci. Instrum.,
23 S. P. Sander, R. R. Friedl, A. R. Ravishankara, D. M. Golden,
M. J. Kurylo, R. E. Huie, V. L. Orkin, M. J. Molina,
G. K. Moortgat and B. J. Findlayson-Pitts, Chemical Kinetics
and Photochemical Data for Use in Atmospheric Studies, Evaluation
#14, Jet Propulsion Laboratory, Pasadena, 2003.
1997, 68, 1875.
1
9 N. Hemmi and A. G. Suits, J. Phys. Chem. A, 1997, 101, 6633.
0 J. A. Seetula, D. Gutman, P. D. Lightfoot, M. T. Rayes and
S. M. Senkan, J. Phys. Chem., 1991, 95, 10688.
2
2
2
1 P. W. Seakins, Annu. Rep. Prog. Chem., Sect. C, 2007, 103.
2 Facsimile. MCPA Software: Harwell, 2003.
24 M. N. Glukhovtsev, A. Pross, M. P. McGrath and L. Radom,
J. Chem. Phys., 1995, 103, 1873.
1
0426 | Phys. Chem. Chem. Phys., 2009, 11, 10417–10426
This journal is ꢄc the Owner Societies 2009