Reactions of Three Lactones with Cl, OD, and O3: Atmospheric Impact and Trends in Furan Reactivity

Article abstract of DOI:10.1021/acs.jpca.7b02325

Lactones, cyclic esters of hydroxycarboxylic acids, are interesting biofuel candidates as they can be made from cellulosic biomass and have favorable physical and chemical properties for distribution and use. The reactions of γ-valerolactone (GVL), γ-crotonolactone (2(5H)-F), and α-methyl-γ-crotonolactone (3M-2(5H)-F) with Cl, OD, and O₃ were investigated in a static chamber at 700 Torr and 298 ± 2 K. The relative rate method was used to determine k_GVL+Cl = (4.56 ± 0.51) × 10^-11, k_GVL+OD = (2.94 ± 0.41) × 10^-11, k_2(5H)-F+Cl = (2.94 ± 0.41) × 10^-11, k_2(5H)-F+OD = (4.06 ± 0.073) × 10^-12, k_{3M-2(5H)-F+Cl} = (16.1 ± 1.8) × 10^-11, and k_{3M-2(5H)-F+OD} = (12.6 ± 0.52) × 10^-12, all rate coefficients in units of cm³ molecule^-1 s^-1. An absolute rate method was used to determine k_2(5H)-F+O3 = (6.73 ± 0.18) × 10^-20 and k_{3M-2(5H)-F+O3} = (5.42 ± 1.23) × 10^-19 in units of cm³ molecule^-1 s^-1. Products were identified for reactions of the lactones with Cl. In the presence of O₂ the products are formic acid (HCOOH), formyl chloride (CHClO), and phosgene (CCl₂O), and also maleic anhydride (C₂H₂(CO)₂O) for 2(5H)-F. In addition both reactions produced a number of unidentified products that likely belong to molecules with the ring-structure intact. A review of literature data for reactions of other furans show that the reactivity of the lactones are generally lower compared to that of corresponding compounds without the carbonyl group.

Full text of DOI:10.1021/acs.jpca.7b02325

Subscriber access provided by University of Colorado Boulder
Article
3
Reactions of Three Lactones With Cl, OD and O:
Atmospheric Impact and Trends in Furan Reactivity
Stina Ausmeel, Christina Andersen, Ole John Nielsen, Freja From
Østerstrøm, Matthew S. Johnson, and Elna Johanna Kristina Nilsson
J. Phys. Chem. A, Just Accepted Manuscript • Publication Date (Web): 28 Apr 2017
Downloaded from http://pubs.acs.org on May 3, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Physical Chemistry A is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 36
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
2
Fig. 1. Reference spectra of GVL, 2(5H)-F and 3M-2(5H)-F in 700 Torr N at 298±2
K.
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 2 of 36
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
Fig. 2. Decay of GVL versus C
2 2 2 6
H (diamonds) and C H (circles) in the presence of
Cl atoms in 700 Torr N at 298±2 K. Dashed linear regressions belong to white
2
symbols and solid to black symbols.
ꢀ
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 3 of 36
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
Fig. 3. Decay of GVL versus C
OD-radicals in 700 Torr N at 298±2 K. Dashed linear regressions belong to white
symbols and solid to black symbols.
2 2 2 4
H (diamonds) and C H (circles) in the presence of
2
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 4 of 36
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
Fig. 4. Pseudo-first order loss of 2(5H)-F (left) and 3M-2(5H)-F (right) vs. O
concentration. The inserts show the decay of 2(5H)F in 595 ppm (green triangles),
3
80 ppm (cyan squares), 1209 ppm (blue diamonds) and 1461 ppm (red circles) of O ,
3
7
respectively and the decay of 3M2(5H)F in 536 ppm (green triangles), 741 ppm (cyan
squares), 985 ppm (blue diamonds), 1108 ppm (red circles) and 1891 ppm (red
circles) of O
3
, respectively 700 Torr air at 298±2 K.
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 5 of 36
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
Fig. 5. Decay of 2(5H)-F vs. C
2 2 3
H in the presence of O and cyclohexane in 700 Torr
air at 298±2 K.ꢀ
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 6 of 36
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
Fig. 6. IR spectra before and after UV irradiation for experiment for 2(5H)-F reacting
with Cl. The inset in the lower panel shows features in the carbonyl region.
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 7 of 36
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
Fig. 7. Available furan relativities ordered by increasing rate constant, for OH, Cl
3
and O respectively.
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 8 of 36
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
TOCꢀGraphicꢀ
ACS Paragon Plus Environment

Page 9 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
4
1
2
ꢀ
ꢀ
Reactions of Three Lactones with Cl, OD and O3:
Atmospheric Impact and Trends in Furan Reactivity
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a, †
b, ‡
b
b
b, *
3ꢀ S. Ausmeel , C. Andersen , O. J. Nielsen , F. F. Østerstrøm , M. S. Johnson , E. J. K.
a, *
4
5
6
ꢀ
ꢀ
ꢀ
Nilsson
a
Combustion Physics, Lund University, Box 118, SE-221 00 Lund, Sweden.
b
Copenhagen Center for Atmospheric Research, Department of Chemistry, University of
7ꢀ Copenhagen, Universitetsparken 5, DK-2100 Copenhagen Ø, Denmark
†
8
9
ꢀ
ꢀ
Current affiliation: Division of Nuclear Physics, Lund University, SE-22100, Lund, Sweden
Current affiliation: Ergonomics and Aerosol Technology, Lund University, Box 118 SE-221 00
‡
10ꢀ Lund, Sweden
*
1
1
1ꢀ
Corresponding authors.
2ꢀ msj@chem.ku.dk (M. S. Johnson)
13ꢀ elna.heimdal_nilsson@forbrf.lth.se (E. J. K. Nilsson)
1
4ꢀ
1
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 10 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
5ꢀ ABSTRACT
6ꢀ
Lactones, cyclic esters of hydroxycarboxylic acids, are interesting biofuel candidates as
17ꢀ they can be made from cellulosic biomass and have favorable physical and chemical properties
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
8ꢀ for distribution and use. The reactions of γ-valerolactone (GVL), γ-crotonolactone (2(5H)-F)
9ꢀ and α-methyl-γ-crotonolactone (3M-2(5H)-F) with Cl, OD and O were investigated in a static
3
20ꢀ chamber at 700 Torr and 298 K. The relative rate method was used to determine kGVL + Cl = (4.56
-
11
-11
-11
2
2
2
2
2
2
1ꢀ ± 0.51)·10 , kGVL + OD = (2.94 ± 0.41)·10
,
k
2(5H)-F + Cl = (2.94 ± 0.41)·10 , k2(5H)-F + OD = (4.06
-12
-11
-12
2ꢀ ± 0.073)·10 , k3M-2(5H)-F + Cl = (16.1 ± 1.8)·10 and k3M-2(5H)-F + OD = (12.6 ± 0.52)·10 , all rates
3
-1 -1
3ꢀ with dimensions of cm molecule s . An absolute rate method was used to determine k2(5H)-F + O3
-
20
-19
3
-1 -1
4ꢀ = (6.73 ± 0.18)·10 and k3M-2(5H)-F + O3 = (5.42 ± 1.23)·10 in units of cm molecule s .
5ꢀ Products were identified for reactions of the lactones with Cl. In the presence of O the products
6ꢀ are formic acid (HCOOH), formyl chloride (CHClO) and phosgene (CCl O), and also maleic
(CO) O) for 2(5H)-F. In addition both reactions produced a number of
8ꢀ unidentified products that likely belong to molecules with the ring-structure intact. A review of
9ꢀ literature data for reactions of other furans show that the reactivity of the lactones are generally
2
2
27ꢀ anhydride (C
2
H
2
2
2
2
30ꢀ lower compared to that of corresponding compounds without the carbonyl group.
3
3
1ꢀ
2ꢀ
2
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 11 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
3
3
3ꢀ 1. Introduction
In recent years cyclic ethers with a five-membered ring structure, generally called
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4ꢀ
35ꢀ furans, have been identified as desirable future biofuels as they have favorable physical and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1,2,3
3
6ꢀ chemical properties for distribution and use, and can be produced from cellulosic biomass.
As
4
3
7ꢀ a result of this, research concerning the combustion properties as well as the environmental
5
,6,7
38ꢀ effects related to potential release of furans has increased.
The compound commonly
3
4
4
4
4
4
9ꢀ considered as most promising as a fuel is 2,5-dimethylfuran (C H O) and understanding of its
6 8
5 6
0ꢀ reactivity is quickly improving. A few other compounds e.g. 2-methylfuran (C H O) and γ-
2
1ꢀ valerolactone (C
5
H
8
O
2
) are also promising candidates. Large scale use of these fuels would
2ꢀ inevitably result in some amounts being released into the atmosphere during fueling of vehicles,
3ꢀ and as a result of leakage from distribution systems. In addition vehicle exhaust will contain
4ꢀ fractions of unburned fuel and reaction intermediates that will escape into the surrounding
45ꢀ atmosphere. The presence of larger amounts of the furans in the atmosphere would alter its
4
4
6ꢀ chemical composition and reactivity relative to the situation today where the fossil fuels being
7ꢀ emitted are mainly long chain alkanes, with smaller amounts of cycloalkanes, alkenes and
48ꢀ aromatic hydrocarbons. Research is needed concerning the reaction rates and products of new
4
5
5
5
5
5
9ꢀ potential biofuels like the furans in order to understand changes, both good and bad, that may
0ꢀ occur due to their introduction.
1ꢀ
As a furan fuel is burned, oxygenated intermediates are produced, ranging from small
2ꢀ molecules like formaldehyde and acetic acid, to rings with structures of typically the same size as
3ꢀ the fuel. During high temperature oxidation in an engine the furan molecules can react or
4ꢀ isomerize to form a variety of other furans, changing the degree of saturation and the number and
8
,9
55ꢀ position of substituents on the ring structure. One important group of oxygenated intermediates
3
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 12 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
5
6ꢀ with the furan ring structure intact are the lactones, also called furanones, cyclic ethers with a
7ꢀ double bonded oxygen transforming one of the alpha-carbons into a carbonyl group. An example
4 6 2
58ꢀ is that one of the simplest lactones, γ-butyrolactone (C H O ), has been shown to be produced in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
10
5
6
4 8
9ꢀ flames of the saturated tetrahydrofuran (C H O) and has been identified in static reactor
1
1
0ꢀ experiments at high temperatures. A second example is that in low pressure flames 2-
4 4 2
61ꢀ methylfuran 2(3H)-furanone (C H O ) was found to be produced by isomerization of 2-
12
6
6
6
6
6
2ꢀ hydroxylfuran.
3ꢀ
Non-furan compounds can also produce lactones; γ-butyrolactone and γ-valerolactone
1
3
4ꢀ are intermediates in the oxidation of toluene (C
7
H
8
) in the intermediate temperature regime. In
),
) are implemented in
5ꢀ the atmosphere toluene oxidation can also result in lactone formation; 2(5H)-furanone (C
6ꢀ 1-methyl-2(5H)-furanone (C ) and 2-methyl-2(5H)-furanone (C
4 4 2
H O
5
H
6
O
2
5 6 2
H O
1
4
67ꢀ the Master Chemical Mechanism for atmospheric oxidation of toluene. Studies of lactone
6
6
8ꢀ production at atmospherically relevant conditions are scarce, but it has been shown that they,
9ꢀ apart from the mentioned production from toluene, can be produced from various furans. Various
70ꢀ substituted furanones are produced in 2-methylfuran reactions with Cl in the presence of air, for
7
7
7
7
7
7
5 6 2
1ꢀ example 3-methyl-2(3H)-furanone (C H O ) has a (10±2)% yield and isomers of 3-methyl-
2ꢀ 2(5H)-furanone (C
5
H
8
O
2
) have a (2±1)% yield at room temperature. Chlorinated and hydroxyl
7
3ꢀ substituted lactones are significant and detected products, respectively. Other atmospheric
1
5
4ꢀ sources of lactones are via cyclisation processes of saturated di-carbonyls.
During daytime in the atmosphere the lactones can be expected to react mainly with OH
6ꢀ radicals, and to some extent with Cl and O depending on concentration. The reactivity of cyclic
5ꢀ
3
77ꢀ ethers depends on the degree of saturation, the position of the double bonds, the presence and
8ꢀ position of alkyl substituents, and the presence of other functional groups like the carbonyl group
7
4
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 13 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
7
8
9ꢀ in the lactones. Cyclic ethers have been shown to react faster with OH than the corresponding
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
6
0ꢀ cyclic alkanes, but significantly slower than the linear ether analogues. Cyclic ethers are
81ꢀ expected to react via direct H abstraction and to a much smaller extent via a hydrogen bonded
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
8
2ꢀ adduct as is the case for linear ethers; according to Moriarty et al. this explains the lower rate
1
6
3ꢀ constants. Reactions of the lactone 2-(5H)-furanone with ozone was shown by Grosjean and
1
7
84ꢀ Grosjean to be significantly slower than for other furans, as a result of the electron-
8
8
8
8
5ꢀ withdrawing effects of the C=O group.
6ꢀ
Available rate coefficients for furan compounds with various substituents are compiled
7ꢀ in Table 1.
8ꢀ
8
9
9ꢀ Table 1
0ꢀ Literature reaction rate coefficients of different furan compounds.
1
1
10
18
Abbrev.
k
OH × 10
k
Cl × 10
kO3 × 10
3
-1 -
3
-1 -
3
-1 -1
Name
(cm molecule s
(cm molecule s
(cm molecule s )
1
1
)
)
a
Maleic anhydride
γ-valerolactone
Tetrahydrofuran
MA
GVL
THF
0.145 ± 0.01
0.281 ± 0.034
0.88 ± 0.18
.80 ± 0.07
.78 ± 0.16
b
c
d
f
2.50 ± 0.39
2.71 ± 0.34
e
1
1
1
g
h
.59 ± 0.39
i
2
-methyltetrahydrofuran
2-MTHF
Furan
2.2
j
k
l
Furan
4.19 ± 0.21
.01 ± 0.30
4.45 ± 0.26
6.45 ± 1.69
6.90 ± 0.46
7.31 ± 0.35
2.0 ± 0.2
2.42 ± 0.28
l
4
a
d
a
m
j
2
2
5
2
(3H)-furanone
2(3H)-F
2,3-DHF
5M2(3H)-F
2-MF
d
d
,5-dihydrofuran
-methyl-2(3H)-furanone
-methylfuran
4.48 ± 0.59
16.55 ± 3.1
k
4.1 ± 0.2
6
.19 ± 0.30
n
k
3
-Methylfuran
3-MF
9.1 ± 0.3
4.2 ± 0.3
o
1
8
1.3 ± 2.2
.73 ± 0.18
10.77 ± 2.11
11.95 ± 2.79
12.6 ± 0.4
m
j
k
2
2
2
2
-ethylfuran
2-EF
4.6 ± 0.3
d
d
d
,3-dihydrofuran
,3-dimethylfuran
,5-dimethylfuran
2,3-DHF
2,3DMF
2,5-DMF
4.52 ± 0.99
4432.0 ± 790
m
m
k
p
12.5 ± 0.4
3.21 ± 0.92
5.7 ± 0.3
420 ± 90
j
1
q
2
(5H)-furanone
2(5H)-F
0.22 ± 0.05
9
9
9
1ꢀ
2ꢀ
3ꢀ
a. Reference 15; b. Reference 18; c. Reference 19; d. Reference 20; e. Reference 16; f. Reference 21; g. Reference
22; h. Reference 23; i. Reference 24; j. Reference 25; k. Reference 26; l. Reference 27; m. Reference 5; n. Reference
28; o. Reference 7; p. Reference 6; q. Reference 17.
5
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 14 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
9
9
4ꢀ
5ꢀ
In the present work reactions of the three lactones: γ-valerolactone (GVL); γ-
96ꢀ crotonolactone or 2(5H)-furanone (2(5H)-F); and α-methyl-γ-crotonolactone or 3-methyl-2(5H)-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
9
9
7ꢀ furanone (3M-2(5H)-F) with OD, Cl and O
3
have been investigated experimentally. Literature
1
8
17
8ꢀ data are available for the reaction of GVL with Cl and 2(5H)-F with O
3
,
for the other
99ꢀ reactions no previous experimental data exist. Products from the Cl initiated oxidation of the
00ꢀ lactones are determined and used together with atmospheric lifetimes calculated from the
01ꢀ reaction rate constants to assess atmospheric impact of the lactones. Trends in reactivity as a
1
1
102ꢀ result of presence and position of double bonds and substituents are discussed.
1
1
03ꢀ
04ꢀ 2. Experimental methods
105ꢀ 2.1 Experimental setup
1
1
1
06ꢀ
The experimental system used in this work has been described in detail by Nilsson et
29
07ꢀ al. and is only briefly outlined in the following. The static reaction chamber is a 100 L quartz
08ꢀ reactor equipped with a Bruker IFS 66V/s FTIR spectrometer for monitoring gas mixture
109ꢀ composition. In the present study a SiC globar was used as a broadband MIR light source and the
1
1
10ꢀ detector was a liquid nitrogen cooled MCT. The reaction chamber is enclosed in an insulating
11ꢀ box equipped with a temperature control system to maintain stable temperatures (298 ± 2 K)
112ꢀ during experiments. All experiments were conducted at a pressure of 700 Torr. UV-A or UV-C
1
1
13ꢀ lamps surrounding the reactor were used to photochemically initiate reactions.
An absorption path length of 66.8 m, 50.8 m or 42.9 m was obtained using a White type
14ꢀ
−
1
115ꢀ multipass optical system in the reaction chamber. IR spectra with a resolution of 0.25 cm were
1
16ꢀ obtained by co-adding 32 interferograms. Reference spectra were recorded using the same
6
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 15 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
1
1
17ꢀ conditions (700 Torr, 298 ± 2 K) employed in the kinetics experiments with a resolution of 0.125
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
−
1
18ꢀ cm and co-addition of 64 interferograms. The reference spectra of the three lactones are shown
119ꢀ in Fig. 1. Reactants were introduced into the chamber using a Pyrex gas handling manifold.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
20ꢀ Concentrations of compounds introduced into the chamber were monitored via the pressure of a
21ꢀ calibrated volume.
122ꢀ
1
1
1
1
1
1
23ꢀ 2.2 Materials and methods
The experimental procedures are summarized below and more detailed descriptions can
24ꢀ
3
0
25ꢀ be found elsewhere. The reactions investigated in the work are between lactones and chlorine
26ꢀ atoms, deuterated hydroxyl radicals or ozone. Chlorine atoms were produced in the
27ꢀ photochemical reactor by UV-A or UV-C photolysis of Cl
2
.
28ꢀ
In the present study OD was used as reactant instead of OH to avoid spectral overlap
129ꢀ from the precursor molecule. Since the hydroxyl-hydrogen does not have an active role in the
1
1
30ꢀ reaction mechanism, the lactone reaction rates in this study are not expected to be significantly
31ꢀ changed by deuterium substitution. The validity of using OD in these types of reactions was
3
1
132ꢀ investigated by Greiner who compared reactions of C
2
H
6
with OH and OD, and concluded that
33ꢀ the isotopic change in the rate was much smaller than the experimental error bars. In the present
34ꢀ study, deuterated hydroxyl radicals, OD, were produced in the photochemical reactor by UV-C
35ꢀ photolysis of CD ONO in the presence of O . CH ONO is a suitable source of OH according to
1
1
1
1
1
1
3
2
3
3
2
36ꢀ Atkinson et al.
37ꢀ
Ozone was produced using a commercially available ozone generator from Infuser.
/O -ratios were obtained by pre-concentration of ozone using a silica gel trap at dry ice
38ꢀ High O
3
2
139ꢀ temperature.
7
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 16 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
40ꢀ
Samples of γ-valerolactone (≥ 98 %), 2(5H)-furanone (98 %) and 3-methyl-2(5H)-
3
41ꢀ furanone (90 %) were from Sigma-Aldrich. CD ONO was synthesized at the Department of
142ꢀ Chemistry, University of Copenhagen. The samples were further purified through freeze-pump-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
43ꢀ thaw cycling.
44ꢀ kinetics and chlorine kinetics experiments were conducted in 700 Torr total pressure
diluent. OD kinetics experiments were conducted in 700 Torr total pressure of air diluent.
46ꢀ Experimental details as well as results for each individual experiment are presented in Table 2.
O
3
145ꢀ of N
2
1
1
1
47ꢀ
48ꢀ Table 2
149ꢀ
1
Experimental details and results from relative rate measurements. Shaded rows indicate
50ꢀ averages of the measured results.
-11
Compound
ppm)
Radical precursor
(ppm)
Reference
(ppm)
Photolysis, lamps Slope,
and total time
klactone /10
3
-1 -1
(
k
k
lactone
/
(cm molec s )
reference
GVL (3.50)
GVL (3.81)
GVL (3.81)
GVL (3.74)
GVL + Cl
Cl
Cl
Cl
Cl
2
2
2
2
(47.3)
(76.7)
(76.7)
(53.4)
C
C
C
C
2
2
2
2
H
H
H
H
2
2
6
6
(8.97)
(9.76)
(10.1)
(10.1)
1 UVA 75 s
2 UVA 120 s
2 UVA 120 s
1 UVA 15 s
0.946
0.914
0.752
0.786
4.82
4.66
4.29
4.48
4.56
3.14
3.02
2.87
2.72
2.94
15.0
16.2
16.4
16.8
16.1
2.70
3.19
2.30
3.49
2.92
4.21
4.58
4.72
3.33
3.46
2
(5H)F (1.81)
Cl
Cl
Cl
Cl
2
2
2
2
(52.1)
(59.6)
(77.3)
(80.9)
C
C
C
C
2
2
2
2
H
H
H
H
2
6
2
6
(9.69)
(10.6)
(11.4)
(12.9)
1 UVA 75 s
1 UVA 40 s
4 UVC 150 s
4 UVC 90s
0.616
0.529
0.563
0.477
2(5H)F (2.09)
2(5H)F (2.50)
2(5H)F (2.33)
2
3
3
3
3
3
(5H)F + Cl
M2(5H)F (5.66)
M2(5H)F (4.74)
M2(5H)F (5.60)
M2(5H)F (5.60)
M2(5H)F + Cl
Cl
Cl
Cl
Cl
2
2
2
2
(66.4)
(66.3)
(76.4)
(76.4)
C
C
C
C
2
2
2
2
H
H
H
H
6
2
2
6
(8.56)
(9.09)
(10.8)
(11.6)
1 UVA 50 s
1 UVA 50 s
4 UVC 140 s
4 UVC 140 s
2.64
3.18
3.22
2.940
GVL (6.10)
GVL (5.94)
GVL (5.89)
GVL (5.89)
GVL + OD
CD
CD
CD
CD
3
ONO (140)
ONO (136)
ONO (140)
ONO (140)
C
C
C
C
2
2
2
2
H
H
H
H
4
2
4
2
(15.6)
(15.4)
(19.1)
(15.1)
4 UVC 165 s
4 UVC 215 s
4 UVC 460 s
4 UVC 460 s
0.358
3.90
3
3
3
0.305
4.26
2
2
2
(5H)F (1.94)
(5H)F (1.94)
(5H)F (2.03)
CD
CD
CD
CD
CD
3
3
3
3
3
ONO (144)
ONO (144)
ONO (163)
ONO (163)
ONO (139)
C
C
C
C
C
2
2
2
2
2
H
H
H
H
H
2
6
2
4
4
(16.3)
(11.2)
(13.8)
(9.93)
(11.8)
4 UVC 110 s
4 UVC 110 s
4 UVC 80 s
4 UVC 80 s
4 UVC 145 s
5.14
18.3
5.76
2(5H)F (2.03)
2(5H)F (1.70)
0.442
0.459
8
ꢀ
ꢀ
ACS Paragon Plus Environment

Page 17 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
2
3
3
3
3
(5H)F + OD
4.06
10.4
10.2
17.7
12.0
12.6
M2(5H)F (2.69)
M2(5H)F (2.69)
M2(5H)F (2.56)
M2(5H)F (2.56)
CD
CD
CD
CD
3
3
3
3
ONO (157)
ONO (157)
ONO (140)
ONO (140)
C
C
C
C
2
2
2
2
H
H
H
H
2
4
2
4
(12.4)
(24.0)
(~7)
4 UVC 55 s
4 UVC 55 s
4 UVC 105 s
4 UVC 105 s
12.7
1.35
21.7
1.60
(~13)
3M2(5H)F + OD
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
51ꢀ
52ꢀ
1
1
The relative rate method is a well-established technique for measuring the rate coefficients
53ꢀ of Cl atoms or OH radicals with organic compounds. The decay of reactants can be expressed
154ꢀ as:
ꢀ
ꢁꢂꢃꢄꢅꢆ ꢇꢈ
ꢉꢊꢋꢌꢇꢍꢎꢏ
ꢒꢆꢓꢆꢒꢆꢅꢂꢆ ꢇꢈ
.
ꢒꢆꢓꢆꢒꢆꢅꢂꢆ ꢇ
1
1
55ꢀ
56ꢀ
ln
= ꢉ
ln
ꢀ
ꢁꢂꢃꢄꢅꢆ ꢇ
ꢐꢏꢑꢏꢐꢏꢎꢌꢏ
Linear
regressions
) are expected to pass through the origin, giving a line with slope
lactone/kreference. Relative changes in concentration were determined by scaling subsequent to
59ꢀ initial experimental spectra using the software OPUS. Absolute concentrations are obtained by
of
plots
of
ln[lactone]t0/[lactone]
t
)
vs.
157ꢀ ln([reference]t0/[reference]
t
1
1
58ꢀ
k
160ꢀ comparing with reference spectra with known concentrations, Fig. 1.
9
ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 18 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
61ꢀ
2
162ꢀ Fig. 1. Reference spectra of GVL, 2(5H)-F and 3M-2(5H)-F in 700 Torr N at 298±2 K.
1
63ꢀ
164ꢀ
2 2 2 6 2 2 2 4
The reference compounds used were C H and C H in the Cl study, C H , C H and
1
1
65ꢀ
C
2
H
6
in the OD study and C
2
H
2
in the O
3
study. Reaction rate coefficients for the references
33
34
66ꢀ were obtained from Sander et al. or Iwasaki et al. as summarized in Table 3.
167ꢀ
1
1
68ꢀ Table 3
69ꢀ Reference compounds used and rate coefficients for reactions with Cl, OH and O
3
, respectively.
11
12
20
Reference
k
Cl × 10
k
OH × 10
k
O3 × 10
3
-1 -1
3
-1 -1
3
-1 -1
(
cm molecule s ) cm molecule s ) cm molecule s )
(
(
b
a
a
a
a
C
C
C
2
2
2
H
H
H
2
4
6
5.1
8.19 ± 0.82
1.0 ± 2.0
Not used
75.4 ± 15.1
2.50 ± 0.18
Not used
Not used
a
5.7 ± 0.4
1
0ꢀ
ꢀ
ACS Paragon Plus Environment

Page 19 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
4
b
C
3
H
6
26
Not used
Not used
1
70ꢀ
a. Reference 33; b. Reference 34.
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
171ꢀ
1
1
1
72ꢀ
For the reactions of 2(5H)-F and 3-methyl-2(5H)-F with O
73ꢀ was also determined using an absolute rate method in which the pseudo first-order loss of the
74ꢀ lactones was measured in the presence of excess O . Cyclohexane was added to avoid loss of the
3
the reaction rate constant
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
175ꢀ lactones via reaction with OH radicals formed from ozone reactions in the chamber. The
76ꢀ experiments were carried out using 130 ppm cyclohexane and 500–1900 ppm O in 700 Torr of
177ꢀ air diluent. The loss of the lactones followed pseudo first-order kinetics in the experiments with
1
3
1
1
78ꢀ 2(5H)-F and 3M-2(5H)-F. It was not the case for GVL, which reacted with OH radicals in the
79ꢀ chamber despite the addition of cyclohexane or other OH scavengers.
180ꢀ
Product studies were carried out for chlorine reactions in both N
2
and air diluent,
1
1
1
81ꢀ respectively. In product studies no reference compounds were present.
82ꢀ Additional loss mechanisms. To check for unwanted loss of reactants and reference compounds
83ꢀ reaction mixtures were left to stand in the chamber for 30 min and were then exposed to 5 min of
184ꢀ UV-A or UV-C irradiation. Absorption spectra were collected after each of the photolysis steps
1
1
85ꢀ and compared to the reference spectra. The maximum available number of lamps was used
86ꢀ during these experiments. In the experiments varying sets of lamps were used, but always less or
187ꢀ equal to what was used in the tests. For all lactones photolysis tests showed negligible losses
1
1
88ꢀ from exposure to UV-A radiation (≤ 2 %), while UV-C radiation resulted in 3 % loss of GVL
89ꢀ and 25-30 % loss of 2(5H)-F and 3M-2(5H)-F. No dark reactions or wall adsorption were
190ꢀ observed for GVL or 3M-2(5H)-F but to some extent for 2(5H)-F (5 %). There was no
1
1
91ꢀ observable loss of the reference compounds, neither to photolysis nor to dark reactions.
92ꢀ
1
1ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 20 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
93ꢀ 3. Results and discussion
94ꢀ 3.1 Reaction rates
1
195ꢀ
The individual relative rate experiments are presented in Table 2, and the final results
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
96ꢀ for all studied reactions are summarized in Table 4. Uncertainties include the standard deviation
97ꢀ from linear regression and a fixed estimated error related to data analysis and spectral
198ꢀ subtraction. If this calculated error is smaller than the variation in the results, the spread in the
1
2
2
2
2
2
99ꢀ measured rate coefficients is used as the error estimate. Comparing the results from this work
00ꢀ (Table 4) with literature values (Table 1), kGVL + Cl is in agreement with the previously
01ꢀ determined value. k2(5H)-F+O3 from the present work, determined from relative as well as absolute
02ꢀ rate methods, is about a third of the previously determined value, as discussed in section 3.1.3
17
03ꢀ below.
04ꢀ
In the following subsections, results for each of the atmospheric oxidants are discussed.
205ꢀ
2
2
2
06ꢀ Table 4
07ꢀ Reaction rate coefficients determined in this work for the eight reactions with lactones in 700
08ꢀ Torr N
Reaction
GVL + Cl → products
2
at 298±2 K.
3
-1 -1
k
lactone (cm molecule s )
-11
(4.6 ± 0.5)·10
(2.9 ± 0.4)·10
-
11
2
3
(5H)F + Cl → products
-11
M-2(5H)F + Cl → products
(16 ± 2)·10
-
-
12
12
GVL + OH → products
(2.9 ± 0.6)·10
2
3
(5H)F + OH → products
(4.1 ± 0.7)·10
-
12
M-2(5H)F + OH → products (13 ± 5)·10
-
-
20
19
2(5H)F + O
3
→ products
→ products
(6.7 ± 0.7)·10
(5.4 ± 1.2)·10
3
M-2(5H)F + O
3
209ꢀ
2
10ꢀ 3.1.1 Cl experiments
1
2ꢀ
ꢀ
ACS Paragon Plus Environment

Page 21 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
2
2
11ꢀ
In each reaction of a lactone and chlorine, C
2 2 2 6
H and C H were used as reference
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
12ꢀ compounds, as explained in experimental section. As examples of results the relative rate plots
213ꢀ for the chlorine reaction of GVL are shown in Fig. 2. Similar plots were made for the other
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
14ꢀ lactones. The measured slopes from relative rate experiments and calculated lactone rate
15ꢀ coefficients of the reactions with Cl atoms are found in Table 2.
216ꢀ
To check the consistency of the results the relative rates of the reference reactions were
3
3,34
2
2
17ꢀ determined. Based on literature data the expected ratio k(C2H6+Cl)/
k
(C2H2+Cl) is 1.1 ± 0.1.
In
18ꢀ experiments performed with C
2
H
6
and C
2
H
2
, the relative rate ratios 1.22 and 1.10 were obtained.
2
19ꢀ
ꢀ
220ꢀ Fig. 2. Decay of GVL versus C
2
H
2
2 6
(diamonds) and C H (circles) in the presence of Cl atoms in
2
2
2
21ꢀ 700 Torr N
22ꢀ symbols.
23ꢀ
2
at 298±2 K. Dashed linear regressions belong to white symbols and solid to black
2
2
24ꢀ
25ꢀ 3.1.2 OD experiments
1
3ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 22 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
26ꢀ
For each reaction of lactones with OD, C
2 2 2 4
H and C H were used as reference
27ꢀ compounds and additionally C was used as a reference in 2(5H)-F measurements. The
2
H
6
228ꢀ relative rate plots for the OD reactions of GVL are shown in Fig. 3. The measured slopes from
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
29ꢀ relative rate measurements and calculated lactone rate coefficients of the OD reactions are found
30ꢀ in Table 2.
231ꢀ
The expected ratio k(C2H4+OH)/k(C2H2+OH) is 9.21 ± 3 and k(C2H6+OH)/ k(C2H2+OH) is 0.31 ±
ꢀ
3
3
2
2
2
2
2
2
2 4 2 2 2 6 2 2
32ꢀ 0.05. From experiments performed with C H and C H or C H and C H present the relative
33ꢀ rate ratios were k(C2H4+OD)/ k(C2H2+OD) = 13.1 ± 2.2 and k(C2H6+OD)/ k(C2H2+OD) = 0.28. While in
ꢀ
34ꢀ both cases the present results fall within the mutual error bars, there appears to be an
35ꢀ overestimation of the rate of reaction of C with OH in this work compared to the literature
36ꢀ value. At present we do not know the origin of the deviation. The GVL + OH rate, determined
37ꢀ here in an experiment that used C as reference, is in excellent agreement with the literature
2
H
4
2
H
4
238ꢀ value (cf. Tables 1 and 4).
17
2
2
39ꢀ
In an earlier experiment, an excess of 2(5H)F (ca. 2-3 ppm) and a smaller
40ꢀ concentration of O (ca. 60-200 ppb) was used relative to the current study. In that work, the O
3
3
241ꢀ decrease was used to derive the pseudo first order rate constant. In the present study the opposite
2
2
2
2
2
42ꢀ approach was used i.e. an O
3
concentration of 500 to 1900 ppm and monitoring the decrease of
43ꢀ 2(5H)F. This may have made the previous work vulnerable to other reactions that degrade O
3
44ꢀ including catalytic cycles and surface reactions. In the present work two independent methods
45ꢀ are used, absolute rate and relative rate and the results are in agreement, in both cases about three
46ꢀ times lower than in the previous work.
1
4ꢀ
ꢀ
ACS Paragon Plus Environment

Page 23 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
47ꢀ
48ꢀ Fig. 3. Decay of GVL versus C
2
H
2
(diamonds) and C
2 4
H (circles) in the presence of OD-radicals
2
2
2
49ꢀ in 700 Torr N
50ꢀ symbols.
51ꢀ
2
at 298±2 K. Dashed linear regressions belong to white symbols and solid to black
2
52ꢀ 3.1.3 O
3
experiments
253ꢀ
The kinetics of the reactions of a set of lactones with ozone were studied by observing
2
2
2
54ꢀ the decay of each lactone when exposed to excess O
55ꢀ quantified and no rate constant was determined. The plots of the pseudo-first order decay as a
56ꢀ function of O concentrations for 2(5H)-F and 3M-2(5H)-F are shown in Fig. 4. Linear least
3
. For GVL, the reaction was too slow to be
3
257ꢀ squares fitting of the data was performed and the resulting reaction rate coefficients are k2(5H)-F +
-20
3
-1 -1
-19
3
-1 -1
2
2
58ꢀ O3 = (6.73 ± 0.18)·10 cm molecule s and k3M-2(5H)-F + O3 = (5.42 ± 1.23)·10 cm molecule s .
59ꢀ In addition a relative rate measurement of the rate of the reaction of 2(5H)-F with O was
as reference compound. The relative rate plot is shown in Fig. 5. The
3
2 2
260ꢀ performed with C H
-20
3
-1 -1
2
2
61ꢀ measured slope is 7.44, corresponding to a rate coefficient of 7.44·10 cm molecule s , in
62ꢀ agreement with the result from the absolute rate study.
1
5ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 24 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
63ꢀ
264ꢀ Fig. 4. Pseudo-first order loss of 2(5H)-F (left) and 3M-2(5H)-F (right) vs. O
3
concentration. The
65ꢀ inserts show the decay of 2(5H)F in 595 ppm (green triangles), 780 ppm (cyan squares), 1209
66ꢀ ppm (blue diamonds) and 1461 ppm (red circles) of O , respectively and the decay of 3M2(5H)F
2
2
3
267ꢀ in 536 ppm (green triangles), 741 ppm (cyan squares), 985 ppm (blue diamonds), 1108 ppm (red
2
3
68ꢀ circles) and 1891 ppm (red circles) of O , respectively 700 Torr air at 298±2 K.
2
2
69ꢀ
2 2 3
70ꢀ Fig. 5. Decay of 2(5H)-F vs. C H in the presence of O and cyclohexane in 700 Torr air at
271ꢀ 298±2 K.ꢀ
1
6ꢀ
ꢀ
ACS Paragon Plus Environment

Page 25 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
72ꢀ
ꢀ
73ꢀ 3.2 Product studies with Cl
Products were determined for the reactions of lactones with Cl atoms in both N
) diluent. The identified products for each lactone are shown in Table 5. IR-spectra from
76ꢀ experiments are shown for a case of 2(5H)-F + Cl in 700 Torr air, Fig. 6. The top panel is a
2
74ꢀ
2
and air
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2 2
275ꢀ (N /O
2
2
2
77ꢀ spectrum of the initial reaction mixture, showing the same main spectral features as the reference
78ꢀ spectra of 2(5H)-F in Fig. 1, but with some HCl bands appearing, indicating that there is a dark
279ꢀ reaction with Cl
2
. Lower panel is a spectrum after irradiation. Here the HCl bands are stronger
-1
2
2
80ꢀ and several new spectral features appear, for example around 1100 cm , upon comparison with
81ꢀ reference spectra this product was identified as formic acid.
2
82ꢀ
1
7ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 26 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
83ꢀ Fig. 6. IR spectra before and after UV irradiation for experiment for 2(5H)-F reacting with Cl.
284ꢀ The inset in the lower panel shows features in the carbonyl region.
2
2
85ꢀ
86ꢀ
2
In N diluent few reaction products are identified and there are unidentified features in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
87ꢀ the spectrum similar to the initial spectrum of the lactone with most changes in the fingerprint
288ꢀ region. For all lactones in both diluents, HCl is seen as a product which is expected as a result of
2
2
89ꢀ hydrogen abstraction. CO
2 2 2
and CO increase as the reaction progresses, more so in N /O diluent,
90ꢀ leaving other products with only relatively weak spectral features. As shown in Table 5 the
291ꢀ products identified are close to identical for the three lactones: formic acid (HCOOH), formyl
2
2
2
2
2
2
2
92ꢀ chloride (CHClO) and phosgene (CCl O). Regarding phosgene it is not expected to form under
93ꢀ standard atmospheric conditions, but it is a result of the high Cl concentration in the
94ꢀ experimental reaction mixture. For the reaction of 2(5H)-F with Cl the cyclic molecule maleic
95ꢀ anhydride (C
96ꢀ region indicating products with an intact ring structure, see insert in lower panel of Figure 6.
In general terms, the formation of the products HCOOH, CHClO and CCl O in the
4 2 3
H O ) was also identified. There are unidentified spectral features in the carbonyl
97ꢀ
2
1
8
298ꢀ chlorine reaction systems occurs as follows. Hydrogen abstraction from the C5 position can be
2
3
99ꢀ followed by 1) addition of oxygen to C5 yielding a peroxy- or alkoxy- radical, 2) formation of a
00ꢀ double bond, resulting for example in maleic anhydride, 3) addition of Cl atoms or 4) opening of
301ꢀ the ring structure. There are several reactions which can form HCOOH and CHClO by further
3
3
3
3
02ꢀ breaking the carbon chain and adding Cl, O or OH. COCl
2
can form when CO reacts with Cl in
03ꢀ the presence of Cl . For 2(5H)-F and 3M-2(5H)-F the addition of Cl to the double bond is also a
2
4
0
04ꢀ possible initial reaction. As described by Villanueva et al. it is the dominant path for the furan
05ꢀ + Cl reaction.
1
8ꢀ
ꢀ
ACS Paragon Plus Environment

Page 27 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
3
3
3
06ꢀ Table 5
07ꢀ Products identified from studying chlorine reactions of three lactones in the presence and
08ꢀ absence of O
Reaction
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
at 298±2 K and a total pressure of 700 Torr.
Diluent Identified products
GVL + Cl
N
N
N
N
N
N
2
2
2
2
2
2
HCl, CO
HCl, CO
HCl, CO
HCl, CO
HCl, CO
HCl, CO
2
2
2
2
2
2
GVL + Cl (+ O
2
)
/O
/O
/O
2
2
2
, CO, HCOOH, CHClO, CCl
, CO
2
2
2
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
(5H)-F + Cl
(5H)-F + Cl (+ O
2
)
, CO, HCOOH, CHClO, CCl
, CO
O, Maleic anhydride
O
3M-2(5H)-F + Cl
M-2(5H)-F + Cl (+ O
3
2
)
, CO, HCOOH, CHClO, CCl
3
3
09ꢀ
10ꢀ
311ꢀ 3.3 Atmospheric implications
3
3
3
3
3
3
12ꢀ 3.3.1 Lifetimes
13ꢀ Natural chemical lifetimes with respect to each reactant are calculated. For chlorine and
14ꢀ ozone, different concentrations are used representing realistic scenarios of different levels for
15ꢀ urban and local conditions. The lifetimes of GVL, 2(5H)-F and 3M-2(5H)-F are shown in Table
16ꢀ 6. The lifetimes for OH reactions are based on the OD measurements and an atmospheric
6
−3 35
17ꢀ concentration of OH radicals of 0.04 ppt (10 cm ) and are on the order of tens of days. The
318ꢀ lifetimes obtained for Cl reactions are calculated using the atmospheric average concentrations of
2
5
−3 36
3
3
19ꢀ chlorine atoms of 10 to 10 cm . Results are on the order of 10-100 years for typical
20ꢀ conditions. In areas where Cl is abundant in high concentrations (0.004 ppt), which could occur
3
7
38
321ꢀ in coastal areas due to sea spray or in the presence of industrial emissions, Cl lifetimes could
3
3
3
3
22ꢀ be as short as those of OH. The lifetimes of all lactones for O
23ꢀ several years even in high O environments (100 ppb).
The studied lactones are less reactive than many other furan compounds and therefore
25ꢀ have longer lifetimes and are transported farther away from a source. The reaction rates are
3
reaction are on the order of
3
24ꢀ
1
9ꢀ
ꢀ
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 28 of 36
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
3
26ꢀ similar to those typically found in many common components of gasoline and diesel fuel. As for
27ꢀ most hydrocarbons the lifetime is mainly determined by reactivity towards OH radicals, while in
328ꢀ areas with high Cl concentration this may also be an important oxidant.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
29ꢀ
3
3
30ꢀ Table 6
31ꢀ Lifetimes of GVL, 2(5H)-F and 3M-2(5H)-F against some common atmospheric oxidants.
Compound OH
Cl
Cl
O
3
O
3
-
2
-6
-3
(4.07·10 ppt) (4.07·10 ppt) (4.07·10 ppt) (10 ppb) (100 ppb)
GVL
40 days
29 days
70 years
110 years
20 years
25 days
39 days
7 days
-
-
2
3
(5H)-F
19 years 1.9 years
2.4 years 0.24 years
M-2(5H)-F 9 days
3
32ꢀ
333ꢀ 3.3.2 Trends in reactivity
3
3
34ꢀ
It is possible to predict the reactivity within a series of structurally similar compounds
35ꢀ as a result of the number and position of e.g. double bonds, alkyl substituents, topology and
336ꢀ carbonyl groups. However, many classes of compounds have not been examined. Compounds
3
3
3
37ꢀ involving oxygen and nitrogen can be especially difficult, due in part to interactions of lone pair
38ꢀ electrons with radicals.
39ꢀ
Knowledge of the reactivity of cyclic ethers has been improved recently by
340ꢀ experimental and theoretical studies. The data for furans with different substituents and degrees
3
3
41ꢀ of saturation, summarized in Table 1, is becoming extensive enough to show some general trends
42ꢀ in reactivity. Fig. 7 shows trends in reactivity for furan derivatives, ordered by increasing
3
343ꢀ reaction rate coefficient for OH, Cl and O , respectively. For OH reactions the rate coefficients
-
11
3
-1 -1
3
3
3
,
44ꢀ of the molecules in Fig. 7 span a range of (0.145-13.21)·10 cm molecule s for chlorine
-10
3
-1 -1
45ꢀ reactions the rate coefficients are in the range of (0.294-5.7)·10 cm molecule s , and for
-
5
-10
3
-1 -1
46ꢀ ozone a larger variation of (4.4·10 -6.7)·10 cm molecule s is observed.
2
0ꢀ
ꢀ
ACS Paragon Plus Environment

Page 29 of 36
The Journal of Physical Chemistry
ꢀ
ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
47ꢀ
ꢀ
2
1ꢀ
ꢀ
ACS Paragon Plus Environment