Formation of hydroxyl and hydroperoxy radicals in the gas-phase ozonolysis of ethene

Article abstract of DOI:10.1016/S0009-2614(99)00057-3

Ozonolysis of alkenes is thought be a significant source of free radicals (OH and HO₂) in the atmosphere. Although studied for many years, the reaction mechanism and the product yields are still very much under discussion. We report measurements of the production of HO₂ radicals from the reaction of O₃ with ethene using matrix isolation and electron-spin-resonance spectroscopy (MIESR). Formation of OH radicals was established by conversion of OH to HO₂ via reaction with CO. The OH yield is 20±2%, the HO₂ yield is 39±3%. Our measurements suggest that the Criegee intermediate in the ground state is not a radical.

Full text of DOI:10.1016/S0009-2614(99)00057-3

5
March 1999
Chemical Physics Letters 301 Ž1999. 559–564
Formation of hydroxyl and hydroperoxy radicals in the gas-phase
ozonolysis of ethene
)
D. Mihelcic , M. Heitlinger, D. Kley, P. M u¨ sgen, A. Volz-Thomas
Institut f u¨ r Chemie und Dynamik der Geosph a¨ re (ICG-2) Forschungszentrum J u¨ lich, D-52425 J u¨ lich, Germany
Received 5 October 1998; in final form 20 December 1998
Abstract
Ozonolysis of alkenes is thought be a significant source of free radicals ŽOH and HO . in the atmosphere. Although
2
studied for many years, the reaction mechanism and the product yields are still very much under discussion. We report
measurements of the production of HO2 radicals from the reaction of O3 with ethene using matrix isolation and
electron-spin–resonance spectroscopy ŽMIESR.. Formation of OH radicals was established by conversion of OH to HO via
2
reaction with CO. The OH yield is 20"2%, the HO2 yield is 39"3%. Our measurements suggest that the Criegee
intermediate in the ground state is not a radical. q 1999 Elsevier Science B.V. All rights reserved.
1
. Introduction
The formation of free radicals in the ozonolysis of olefinic hydrocarbons, which are emitted in large
quantities from both anthropogenic and biogenic sources w1,2x, is of considerable interest to atmospheric
chemists, as it may significantly influence the radical budgets in rural and urban environments w3x. Although the
mechanism of gas-phase ozonolysis has been intensively studied, there is still considerable uncertainty about the
exact mechanism of ozonolysis and in particular the radical yields, even for the simplest alkene, C H , which
2
4
has been most widely studied. Reaction of O with C H in the gas phase is believed to occur, in analogy to the
3
2
4
Criegee mechanism in the liquid phase w4x, by addition of O to the double bond ŽR1.. The five-membered ring
3
intermediate with an excess energy of up to 60 kcalrmol w5–7x is believed to decompose into formaldehyde and
Ø
Ø
an excited carbonyl oxide, often referred to as Criegee bi-radical ŽH C OO .:
2
)
Corresponding author. Fax: q49 2461 615346; e-mail: d.mihelcic@fz-juelich.de
0
009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
PII: S0009-2614Ž99.00057-3

5
60
D. Mihelcic et al.rChemical Physics Letters 301 (1999) 559–564
The excited bi-radical is thought to be either stabilized by collision with O and N ŽR2a. or to undergo
2
2
further decomposition ŽR2b-f., thereby producing free radicals ŽH, OH and HCO. and stable molecules, such as
H O, CO, CO , H and HCŽO.OH w7–10x. The postulated reactions are listed below, together with the
2
2
2
measured branching ratios w10x.
)
R2a wH COOx qM™H COOqM ;40%
2
2
)
R2b wH COOx ™HCOqOHr
12–40%
2
)
R2c wH COOx ™CO q2H
3–5%
2
2
)
R2d wH COOx ™COqH O
31–58%
;13%
;7%
2
2
)
R2e wH COOx ™CO qH
2
2
2
)
R2f wH COOx ™HC
Ž
O
.
OH
2
The H-atoms and HCO radicals are instantaneously converted to HO in the presence of O at atmospheric
2
2
conditions ŽR3 and R4..
R3 HCOqO ™HO qCO
2
2
R4 HqO qM™HO qM
2
2
R5 OHqC H qO ™HOCH CH O
2
4
2
2
2
2
R6 OHqCOqO ™HO qCO
2
2
2
There is little direct spectroscopic evidence for the proposed mechanism. In most studies, the branching
ratios and radical yields were estimated from the observed stable products w10x. The yield of OH radicals was
inferred from the observed excess consumption of C H due to R5 as compared to that of ozone or from the
2
4
decay of OH scavengers which were added to the O –C H mixture or from the products of the reaction of OH
3
2
4
with the scavenger w10–13x. The only direct measurement of OH formation w7x was done at low pressure Ž5–8
hPa.. While OH was detected in these experiments, the yields were much larger than those determined from
indirect measurements. Recently, the formation of OH was called into question since the relative decay rates of
two different OH scavengers which were added to O ralkene mixtures were inconsistent with OH kinetics w14x.
3
Peroxy radicals have been identified by ESR spectroscopy in Ar-matrices w9x, however, without sufficient
spectral resolution to allow the distinction between HO from R2b and R2c and the HOCH CH O radicals
2
2
2
2
formed in R5. No spectroscopic evidence exists for the stabilized Criegee bi-radical.
2
. Experimental
We have used matrix isolation and electron spin resonance ŽMIESR, w15x. to study the radical products that
emerge from the ozonolysis of ethene. The experiments were performed at 1000 hPa and 295 K in a flow
reactor with a movable injector. The flow reactor consisted of a pyrex glass tube with an internal diameter of 1.8
cm ŽFig. 1.. Ozone was generated in an O Calibrator ŽThermo. Instr., Model TE 49. by irradiation of pure O2
3
ŽLinde, purity 99.999%. with a low-pressure mercury lamp. Pure ethene ŽLinde, purity 99.95%. was metered
with quartz capillaries into a flow of synthetic air ŽMesser–Griesheim, purity 99.998%.. The flow rates of
ethene Ž1–3 mlrmin., synthetic air Ž3 lrmin. and the O rO mixture Ž3 lrmin. were determined volumetri-
3
2
cally with an automated soap-film flow meter ŽGillibrator.. The C H rair mixture was introduced into the flow
2
4
of O rO through an 8 mm long tubular glass frit at the tip of a movable injector ŽFig. 1.. In order to achieve
3
2
efficient mixing of the reagent gases, the total gas flow rate was divided evenly between the injector and the
surrounding tube. The injector could be moved about 12 cm, corresponding to a maximum reaction time of

D. Mihelcic et al.rChemical Physics Letters 301 (1999) 559–564
561
Fig. 1. Schematic diagram of the flow system and the MIESR cryosampler used for the experiments.
0
.4 s. The end of the flow tube was attached to the MIESR sampling nozzle, which sampled about 3% of the
total gas flow. The rest was vented through a coaxial slit around the nozzle.
The initial concentrations of C H and O and the maximum reaction time Žsee Table 1. were chosen such
2
4
3
that less than 0.5% of the reactants was consumed in the flow tube. Under these conditions, reactions of the
radicals with themselves or with O were negligible Ž-1% of the primary reaction rate., and the OH radicals
3
produced in R2b reacted exclusively with C H ŽR5., thereby producing hydroxyethylperoxy radicals
2
4
ŽHOCH CH O .. In separate experiments, carbon monoxide Žpurity 99.997%, further purified by passage over
2
2
2
iodineractivated charcoal. was supplied together with the C H through the movable injector at sufficiently
2
4
large concentrations to convert )99% of the OH radicals to HO .
2
Samples for the MIESR are collected under vacuum Ž p-10y2 hPa. and at 77 K on a gold-plated cold
finger in a polycrystalline matrix formed by the controlled addition of D O. The sample flow Ž0.2 lrmin. is
2
controlled by a critical orifice. Total sample volume is 5 l ŽSTP.. The sampling efficiency as determined for
NO , NO , HO , HOCH CH O and several other RO is )95% w15x. The samples are transferred under
2
3
2
2
2
2
2
vacuum and at 77 K into the cavity of a Bruker, Model ESP 300E, ESR spectrometer. The method allows the
speciation of NO , NO , HO , CH COO , and the sum of organic peroxy radicals ŽS RO . with a detection
2
3
2
3
2
2
limit of 2–3 ppt and an accuracy of "5%. For further details of the method see Ref. w15x. The radicals are
trapped in a D O matrix, which provides sufficient spectral resolution to discriminate between HO2 and
2
HOCH CH O w16x.
2
2
2
Table 1
Experimental conditions and measured radical concentrations
Sample O3 Ethene Dt CO
No.
O consumption HO2
RO2
3
Žmolecules cmy3. Žmolecules cmy3. Žs. Žmolecules cmy3. Ž% of O3.
Žmolecules cmy3. Žmolecules cmy3.
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
15
15
15
15
15
15
15
15
15
9
9
9
9
1
2
3
4
5
6
7
8
9
6.1=10
6.1=10
6.0=10
5.7=10
5.7=10
6.0=10
6.0=10
5.7=10
6.3=10
8.3=10
8.3=10
3.3=10
3.1=10
3.1=10
3.3=10
3.3=10
3.1=10
8.8=10
0.28 0.0
0.15 0.0
0.15 0.0
0.15 1.4=10
0.28 1.4=10
0.29 0.0
0.42 0.0
0.40 1.4=10
0.43 0.0
0.3
0.2
0.1
0.1
0.1
0.1
0.2
0.2
0.5
7.5=10
3.9=10
2.2=10
2.7=10
4.9=10
3.5=10
5.0=10
5.9=10
1.1=10
3.4=10
2.0=10
1.0=10
0.0
9
9
18
18
9
9
0.0
9
9
9
1.7=10
2.5=10
0.0
9
18
9
10
9
6.3=10

5
62
D. Mihelcic et al.rChemical Physics Letters 301 (1999) 559–564
3
. Results and discussion
Fig. 2 shows two typical examples of the ESR-spectra recorded from samples collected in the experiments. In
the presence of CO, the ESR spectrum is exclusively due to HO ŽFig. 2A.. In the samples collected from
2
experiments without the addition of CO ŽFig. 2B., the HO signal is reduced and the spectra contain a strong
2
signature from hydroxyethylperoxy radicals HOCH CH O . After normalizing to identical reaction conditions,
2
2
2
the total peroxy radical concentration ŽHO and HOCH CH O . obtained in the absence of CO agrees within
2
2
2
2
5
% with the HO2 concentration when CO is present. The hydroxyethylperoxy radicals must have been
produced, therefore, by reaction of C H with the OH radicals formed in reaction R2b.
2
4
The measured HO and HOCH CH O concentrations are plotted in Fig. 3 versus the ozone consumption
2
2
2
2
due to R1. Since the initial concentrations of O and C H changed by less than 0.5% in our experiments, the
3
2
4
ozone consumption was not measured but calculated from the reaction time and the initial reaction rate, i.e.
HwO xwC H xk dt. For the rate coefficient k , we adopted the value obtained from Ref. w17x, where cyclohexane
3
2
4
1
1
was added as a scavenger for OH and which is 40% smaller than the currently recommended value. Since the
free radicals formed in R1 can react further with the reactants, experiments in which only the disappearance of
the educts is monitored tend to overestimate the rate coefficient for R1, if no measures are taken to scavenge the
OH radicals.
In the absence of CO, the slopes of the regressions correspond to stoichiometric radical yields of 39"3% for
HO and 20"2% for HOCH CH O , which corresponds to the OH yield from R2b. Since one HO molecule
2
2
2
2
2
is formed in R2b, the branching ratio for R2c is calculated to 10"4%, about 2–3 times larger than what was
Fig. 2. ŽA. ESR spectrum of a sample collected from the reaction of C H qO in the presence of excess CO. The sample was measured at
2
4
3
7
7 K Ž2 gauss modulation amplitude, 200 gauss scan range, 50 scans averaged.. The middle trace shows the HO reference spectrum
2
9
y3
produced by the reaction of H-atoms with O . Comparison of the two spectra gave a concentration of 5.9=10 molecules cm . The
2
lowest trace shows the residuals after subtraction of the reference spectrum from the original spectrum magnified by 100. ŽB. ESR spectrum
of a sample collected from the reaction of C H qO . The HOCH CH O reference spectrum was produced by the reaction of OH
2
4
3
2
2
2
radicals with ethene. The lowest trace shows the residuals after subtraction of the fitted reference spectra from the original spectrum,
magnified by 100. No spectroscopic evidence for biradicals was found in any of the samples.

D. Mihelcic et al.rChemical Physics Letters 301 (1999) 559–564
563
Fig. 3. HO2 and HOCH CH O radical concentrations measured by MIESR during the ozonolysis of ethene plotted versus the time
2
2
2
integrated reaction rate of R1 Žozone consumption.. Filled and open circles denote the HO2 concentrations from experiments with and
without addition of CO, respectively. Crosses give the HOCH CH O concentrations from experiments without CO. Error bars include the
2
2
2
uncertainties from the determination of the concentrations of O3 Ž5%. and C H Ž5–8%. and the residence time in the reactor Ž5–8%., as
2
4
well as the measurement accuracy of the MIESR Ž5%.. The regression lines are calculated by minimizing errors in both coordinates. The
offsets reflect the uncertainty about the time which is required for complete mixing of the reagents Ž20 ms..
suggested from indirect experiments w18x. The experiments with CO give an HO yield of 66"9%, which
2
confirms the total radical yield of 59"5% ŽHO qOH. from the experiments without CO within the
2
experimental uncertainties.
Because of the increased spectral resolution that is obtained in a D O–matrix, our results unequivocally
2
confirm the formation of HO during the ozonolysis of ethene. The HO yield is significantly larger than that
2
2
estimated from stable product analysis w10,18x. Part of the difference is due to the fact that we adopt a smaller
rate coefficient for R1 in our calculations w17x. With the currently accepted value for k , the radical yields would
1
come out 40% smaller. Loss of radicals to the walls of the flow tube would lead to an underestimate of the
radical yields. However, since experiments conducted at different reaction times but comparable O consump-
3
tion Ždue to higher wC H x. yield nearly identical results Žsee Table 1, samples a2 and a6., we conclude that
2
4
wall losses are negligible.
The formation of OH is strongly implied by the formation of HOCH CH O in the experiments without CO
2
2
2
and is confirmed by the additional production of HO in the experiments with excess CO. Donahue et al. w7x
2
measured a larger OH yield of 40"20%. The large uncertainty in the latter experiment makes it difficult to
claim a significant difference to our 20"2%. A possible explanation for the difference is that those experiments
were conducted at reduced pressure, at which the decomposition of the excited carbonyl oxide is favored
because of the less efficient thermalisation w19x.
The ESR spectra also provide important information on the structure of the thermalized Criegee intermediate
formed in R2a, for which three possible mesomeric structures exist, with the peroxymethylene having the
highest and the dioxirane the lowest internal energy.
Only the dioxirane has been identified spectroscopically as a reaction product in the reaction of O with
3
ethene w20x, whereas the existence of the peroxy form is suggested from the observed formation of H O and
2
2

5
64
D. Mihelcic et al.rChemical Physics Letters 301 (1999) 559–564
organic peroxides when H O or alcohols are added to the ozonolysis mixture w21–23x. Hatakeyama et al. w19x
2
determined the yield of the thermalized Criegee intermediate at 1013 mbar via reaction with SO Žforming
2
H SO . to 39"5%.
2
4
The ESR spectra ŽFig. 2. contain no significant signature besides those from HO and HOCH CH O . This
2
2
2
2
rules out the presence of peroxymethylene and methylene bis oxy, which should have a significantly different
spectrum. Most important is the absence of a spectral signature at half of the magnetic field. For a bi-radical, as
the postulated Criegee intermediate, we expect a forbidden transition ŽD M s2. at half of the magnetic field of
s
the allowed transition w24x. From the absence of such a spectroscopic signature and the estimated relative signal
intensities Ž0.19 because of the higher microwave power that can be used for the forbidden transition., we obtain
an upper limit for the possible concentration of the stabilized bi-radical intermediate of 4% of the HOCH CH O
2
2
2
concentration, which corresponds to -1% of R1. Isomerisation of the thermalized methylene bis oxy to
dioxirane in the D O matrix is unlikely due to the low temperature Ž77 K. and the required energy w6,7x.
2
Theoretical calculations w6,7x have shown that the most likely reaction of the excited intermediate is
isomerisation to the vibrationally excited dioxirane, followed by rearrangement to methylene bis oxy, which
further decomposes into molecular products such as formic acid, H , H O, CO and CO . Complete decomposi-
2
2
2
tion of the intermediate, however, would be in conflict with the observed formation of H O in the presence of
2
2
water vapor w21–23x and also the formation of H SO in the presence of SO w19x. We hence conclude that the
2
4
2
most likely form for the stabilized Criegee intermediate is the dioxirane, which was first observed by Ref. w20x.
References
w1x S.D. Piccot, J.J. Watson, J.W. Jones, J. Geophys. Res. 97 Ž1992. 9897.
w2x A. Guenther, C.N. Hewitt, D. Erickson, R. Fall, C. Geron, T. Graedel, P. Harley, L. Klinger, M. Lerdau, W.A. Mckay, T. Pierce, B.
Scholes, R. Steinbrecher, R. Tallamraju, J. Taylor, P. Zimmerman, J. Geophys. Res. 100 Ž1995. 8873.
w3x S.E. Paulson, J.J. Orlando, Geophys. Res. Lett. 23 Ž1996. 3727.
w4x R. Criegee, G. Wenner, Justus Liebigs Ann. Chem. 564 Ž1949. 9.
w5x L.B. Harding, W.A. Goddard, J. Am. Chem. Soc. 100 Ž1978. 7180.
w6x R. Gutbrod, R.N. Schindler, E. Kraka, D. Cremer, Chem. Phys. Lett. 252 Ž1996. 221.
w7x N.M. Donahue, J.H. Kroll, J.G. Anderson, K.L. Demerjian, Geophys. Res. Lett. 25 Ž1998. 59.
w8x B.J. Finlayson-Pitts, J.N. Pitts, Atmospheric Chemistry: Fundamentals and Experimental Techniques, John Wiley, New York, 1986.
w9x F.J. Rakoczi, T.-K. Ha, H.H. G u¨ nthard, Chem. Phys. 74 Ž1983. 273.
w10x R. Atkinson, J. Phys. Chem. Ref. Data 26 Ž1997. 215.
w11x A.A. Chew, R. Atkinson, J. Geophys. Res. 101 Ž1996. 28649.
w12x S. Paulson, A.D. Sen, P. Liu, J.D. Fenske, M.J. Fox, Geophys. Res. Lett. 24 Ž1997. 3193.
w13x G. Marston, C.D. McGill, A.R. Rickard, Geophys. Res. Lett. 25 Ž1998. 2177.
w14x C. Sch a¨ fer, O. Horie, J.N. Crowley, G.K. Moortgat, Geophys. Res. Lett. 24 Ž1997. 1611.
w15x D. Mihelcic, P. M u¨ sgen, D.H. Ehhalt, J. Atmos. Chem. 3 Ž1985. 341.
w16x D. Mihelcic, A. Volz-Thomas, H.W. P a¨ tz, D. Kley, M. Mihelcic, J. Atmos. Chem. 11 Ž1990. 271.
w17x K.-H. Becker, I. Barnes, A. Bierbach, K.J. Brockmann, F. Kirchner, B. Klotz, H.G. Libuda, A. Mayer-Figge, S. M o¨ nninghoff, L.
Ruppert, W. Thomas, E. Wiesen, K. Wirtz, F. Zabel, in: G. LeBras ŽEd.., Chemical Processes in Atmospheric Oxidation, Springer,
Heidelberg, 1997.
w18x J.T. Herron, R.E. Huie, J. Am. Chem. Soc. 99 Ž1977. 5430.
w19x S. Hatakeyama, H. Kobayashi, Z. Lin, H. Takagi, H. Akimoto, J. Phys. Chem. 90 Ž1986. 4131.
w20x R.D. Suenram, F.J. Lovas, J. Am. Chem. Soc. 100 Ž1978. 5117.
w21x K.H. Becker, K.H. Brockmann, J. Bechara, Nature 346 Ž1990. 256.
w22x S. G a¨ b, W.V. Turner, S. Wolff, K.H. Becker, L. Ruppert, K.J. Brockmann, Atmos. Environ. 29 Ž1995. 2401.
w23x S. Wolff, A. Boddenberg, J. Thamm, W.V. Turner, S. G a¨ b, Atmos. Environ. 31 Ž1997. 2965.
w24x S.S. Eaton, K.M. More, B.M. Sawant, G.R. Eaton, J. Am. Chem. Soc. 105 Ž1983. 6560.