Hydroxyl radical reaction rate constant and atmospheric transformation products of 2-butanol and 2-pentanol

Article abstract of DOI:10.1002/(SICI)1097-4601(1998)30:10<745::AID-KIN7>3.0.CO;2-V

The relative rate technique has been used to measure the hydroxyl radical (OH) reaction rate constant of +2-butanol (2BU, CH₃CH₂CH(OH)CH₃) and 2-pentanol (2PE, CH₃CH₂CH₂CH(OH)CH₃). 2BU and 2PE react with OH yielding bimolecular rate constants of (8.1±2.0)×10^-12 cm³molecule^-1s^-1 and (11.9±3.0)×10^-12 cm³molecule^-1s^-1, respectively, at 297±3 K and 1 atmosphere total pressure. Both 2BU and 2PE OH rate constants reported here are in agreement with previously reported values [1-4]. In order to more clearly define these alcohols' atmospheric reaction mechanisms, an investigation into the OH+alcohol reaction products was also conducted. The OH+2BU reaction products and yields observed were: methyl ethyl ketone (MEK, (60±2)%, CH₃CH₂C(to-O double bond)CH₃) and acetaldehyde ((29±4)% HC(to-O double bond)CH₃). The OH+2PE reaction products and yields observed were: 2-pentanone (2PO, (41±4)%, CH₃C(to-O double bond)CH₂CH₂CH₃), propionaldehyde ((14±2)% HC(to-O double bond)CH₂CH₃), and acetaldehyde ((40±4)%, HC(to-O double bond)CH₃). The alcohols' reaction mechanisms are discussed in light of current understanding of oxygenated hydrocarbon atmospheric chemistry. Labeled (¹⁸O) 2BU/OH reactions were conducted to investigate 2BU's atmospheric transformation mechanism details. The findings reported here can be related to other structurally similar alcohols and may impact regulatory tools such as ground level ozone-forming potential calculations (incremental reactivity).

Full text of DOI:10.1002/(SICI)1097-4601(1998)30:10<745::AID-KIN7>3.0.CO;2-V

The Hydroxyl Radical
Reaction Rate Constant
and Atmospheric
Transformation Products of
2
-Butanol and 2-Pentanol
1 2
J. STEVEN BAXLEY , J. R. WELLS
1Applied Research Associates, Gulf Coast Division, P.O. Box 40128, Bldg. 1142, Mississippi Road, Tyndall Air Force Base,
Florida 32403
2
Air Force Research Laboratory, AFRL/MLQR, 139 Barnes Drive, Tyndall AFB, Florida 32403-5323
Received 20 October 1997; accepted 17 April 1998
ABSTRACT: The relative rate technique has been used to measure the hydroxyl radical (OH)
reaction rate constant of ϩ2-butanol (2BU, CH CH CH(OH)CH ) and 2-pentanol (2PE,
3
2
3
CH CH CH CH(OH)CH ). 2BU and 2PE react with OH yielding bimolecular rate constants of
3
2
2
3
3
Ϫ12
Ϫ1
Ϫ1
Ϫ12
3
Ϫ1 Ϫ1
(
8.1 Ϯ 2.0) ϫ 10 cm molecule s
and (11.9 Ϯ 3.0) ϫ 10 cm molecule s , respec-
tively, at 297 Ϯ 3 K and 1 atmosphere total pressure. Both 2BU and 2PE OH rate constants
reported here are in agreement with previously reported values [1–4]. In order to more clearly
define these alcohols’ atmospheric reaction mechanisms, an investigation into the OH ϩ
alcohol reaction products was also conducted. The OH ϩ 2BU reaction products and yields
observed were: methyl ethyl ketone (MEK, (60 Ϯ 2)%, CH CH C("O)CH ) and acetaldehyde
3
2
3
(
(29 Ϯ 4)% HC("O)CH ). The OH ϩ 2PE reaction products and yields observed were: 2-pen-
3
tanone (2PO, (41 Ϯ 4)%, CH C("O)CH CH CH ), propionaldehyde ((14 Ϯ 2)% HC("
3
2
2
3
O)CH CH ), and acetaldehyde ((40 Ϯ 4)%, HC("O)CH ). The alcohols’ reaction mechanisms
2
3
3
are discussed in light of current understanding of oxygenated hydrocarbon atmospheric chem-
18
istry. Labeled ( O) 2BU/OH reactions were conducted to investigate 2BU’s atmospheric trans-
formation mechanism details. The findings reported here can be related to other structurally
similar alcohols and may impact regulatory tools such as ground level ozone-forming potential
calculations (incremental reactivity) [5]. ᭧ 1998 John Wiley & Sons, Inc. Int J Chem Kinet 30: 745–
7
52, 1998
INTRODUCTION
are also being used as solvents in paints and paint
strippers. With the increased usage of these com-
pounds a better understanding of their environmental
impact is necessary. Several hydroxyl radical
(OH) ϩ oxygenated organic bimolecular rate con-
stants are well known, but details pertaining to the
reaction mechanisms are sparse and incomplete. Cur-
rent studies of the products of OH ϩ oxygenated or-
ganic compounds have illustrated the complexity of
their atmospheric reaction mechanisms [6–10]. Re-
Oxygenated organic compounds are being used in fu-
els to promote better combustion and to increase oc-
tane without using more aromatic compounds, thereby
reducing carbon monoxide and photochemically ac-
tive volatile exhaust emissions. Oxygenated organics
Correspondence to: J. R. Wells
1998 John Wiley & Sons, Inc.
᭧
CCC 0538-8066/98/100745-08

7
46
BAXLEY AND WELLS
cently, structural reactivity, which uses factors based
on neighboring molecular groups and molecular struc-
ture to predict hydroxyl radical rate constants, has be-
come a useful tool for initially assessing molecules’
complex atmospheric transformations [11].
Poorly understood atmospheric reaction mecha-
nisms seriously affect air quality analysis and regula-
tory decisions. Calculation of a compound’s incre-
mental reactivity [5], tendency to generate
[6,14,15]. Therefore, a brief description is provided
here. Reactants were introduced and samples were
withdrawn through a 6.4-mm Swagelok fitting at-
tached to ca. 100 L Teflon௡ film chamber. Dry com-
pressed air was added as a diluent to the reaction
Ϫ1
chambers and measured with a 0–100 L-min mass
flow controller. The filler system was equipped with a
syringe injection port facilitating the injection of both
liquid and gaseous reactants into the chambers in a
flowing airstream. All reactant mixtures and calibra-
tion standards were generated by this system. Irradi-
ations were carried out in a light tight chamber using
2-mil FEP Teflon௡-film bags (ca. 100 liters), and the
following mix of lamps: 6-Philips TL40W/03; 1-GE
F40BL; 2-QPANEL UV351 and 7-QPANEL UV340
which were shown to accurately simulate solar radia-
tion in both frequency and intensity from 420–
300 nm.
All samples (ca. 100 mL) were quantitatively mon-
itored using an Hewlett–Packard (HP) gas chromat-
ograph (GC) 5890 with a flame ionization detector
(FID) and HP series Chem Station software. Gas sam-
ples for GC analysis were cryogenically collected on
a Hastelloy C sample loop (ca. 1.3 mL) and injected
onto the GC column (Restek Rtx 200 column
(0.53 mm i.d., 30-m, 1.0 ␮m film thickness)) with a
heated rotary valve [16]. The GC oven temperature
program used was as follows: 35ЊC for 5 min then
ramped temperature up 10ЊC/min to 210ЊC and held
for two min. Helium (UHP grade) supplied by Air
Products, was passed through drying and hydrocarbon
traps before use.
Product identification experiments were also per-
formed and sampled as described above except the
chromatographic sample volume was up to twice as
large (ca. 200mL) and analyzed with an HP 5890 Se-
ries II Plus GC/HP 5971 mass selective detector/HP
5965B infrared detector (GC/MS/FTIR) system. Com-
pound separation was achieved using a Restek Rtx 200
(0.53 mm i.d., 30-m length, 1.0-␮m film thickness)
column.
Carbonyl compounds produced by gas-phase re-
actions were measured by sampling the contents of the
reaction chamber through a 2,4-dinitrophenylhydra-
zine (DNPH) cartridge or impinger. Hydrazones
formed by derivatization were separated and quanti-
tatively measured by HPLC (HP 1050) using a three
component gradient solvent program as described pre-
viously [17].
tropospheric ozone (O ), requires detailed knowledge
3
of a compound’s reactive and mechanistic properties.
While not emitted directly, O is a by-product of the
3
photooxidation of hydrocarbons [12] and a regulated
pollutant. Sources that contribute to its formation in
the troposphere are also regulated. Currently, the re-
activity models incorporate educated assumptions
about a compound’s unknown atmospheric mecha-
nism, but in order to minimize uncertainties and to
calculate a compound’s incremental reactivity more
accurately, the hydroxyl radical rate constant and the
compound’s atmospheric degradation mechanism
must be well understood. As a side benefit, under-
standing the compounds’ atmospheric mechanisms in
detail can assist in chemical selection based on chem-
ical structure. Kinetic and mechanistic investigations
also expand the possibility of synthesizing new com-
pounds incorporating environmentally and technically
beneficial molecular structures. The information
gained from the type of research presented here can
lead to more beneficial use of organic compounds in
the future.
Alcohols form a subset of oxygenated organics that
are finding more widespread use. Both 2-butanol and
2
-pentanol are present in coating formulations, and to
date, there has been only one report of the OH/2-bu-
tanol rate constant and reaction products [4]. In the
work presented here, the rate constant of OH with the
alcohols was measured by the relative rate method [13]
and the products of the OH ϩ alcohols reaction are
reported and used to derive the alcohols’ atmospheric
reaction mechanisms. This article presents information
regarding the atmospheric impact of alcohols that is
consistent with currently proposed transformation
mechanisms.
EXPERIMENTAL METHODS
Apparatus and Materials
Experiments to measure the gas-phase rate constant of
OH, the primary oxidizing radical in the atmo-
sphere, was generated from the photolysis of methyl
the OH ϩ 2-butanol (2BU, CH CH CH(OH)CH ) and
3
2
3
2
-pentanol (2PE, CH CH CH CH(OH)CH ) reactions
nitrite (CH ONO) in the presence of nitric oxide (NO)
3
2
2
3
3
were conducted with a previously described apparatus
in air [13].

HYDROXYL RADICAL REACTION RATE CONSTANT
747
CH ONO ϩ hv !: CH O ϩ NO
(1)
(2)
(3)
and 2PE), and hexane (2PE). The use of two different
reference compounds with different OH rate constants
more definitively assured the accuracy of the alcohol/
OH rate constant.
The typical concentrations of the pertinent species
in the 100 liter Teflon௡ chamber were 3–5 ppm al-
3
3
CH O ϩ O !: CH O ϩ HO
2
3
2
2
HO ϩ NO !: OH ϩ NO
2
2
CH ONO was prepared in gram quantities using the
cohols, 2–4 ppm reference, 10 ppm CH ONO, and
3
3
method of Taylor et al. [18] and stored in a lecture
1 ppm NO in air. These mixtures were allowed to
stand for 30–60 min before background chamber
samples were collected. The samples were collected.
The samples were collected on a cryogenic (Ϫ65ЊC)
sample loop (described above) for 4 min at
25 mL min and then flash injected (260ЊC) onto the
GC column [16]. Typically, 15–30 s irradiation inter-
vals were used on the reaction mixture for a combined
total photolysis time of approximately 90 s. The flame
ionization detector (FID) signal was used to determine
alcohols and reference concentrations.
bottle at room temperature. The CH ONO purity
3
(
Ͼ95%) was verified by GC/MS/FTIR.
-Pentanol was obtained from Theta Corporation
and used as received. 2-Butanol, methyl ethyl ketone,
-pentanone, propionaldehyde, dodecane, n-nonane,
2
Ϫ1
2
and heptane (all 99%ϩ) were used as received from
Aldrich Chemical. Methanol and acetonitrile, HPLC
grade, were used as received from Aldrich Chemical.
The water was distilled, deionized to 18 megaohm,
and filtered using a Milli-Q௡ filter system. 2,4-Dini-
trophenylhydrazine (DNPH) standards were used as
received from Radian Corporation. The nitric oxide
Alcohols ϩ OH reaction product experimental
methods and component concentrations were similar
to those for reaction rate experiments, except that the
reference compound was excluded from the reaction
mixture. For the GC/MS/FTIR measurements the mix-
ture was irradiated for timed (10, 30, 50 s) intervals
followed by two separate sample collections. The mass
selective detector was tuned using perfluorotributy-
lamine (FC-43). Full scan electron impact ionization
spectra were collected from 25 to 220 mass units. The
infrared detector of GC/MS/FTIR system was oper-
(
NO) (0.662% in N ) was obtained from Matheson
2
Gases and used as received. Nitric oxide was added to
facilitate the generation of OH and to minimize ozone
(
O ) and NO formation therefore preventing other
3 3
1
8
possible radical reactions. The O labeled 2-butanol
98ϩ%) was synthesized by Cambridge Isotope Lab-
(
oratories, Inc. and used as received. Experiments were
carried out at 297 Ϯ 3 K at about 1 atmosphere.
Ϫ1
ated at eight cm resolution with each four scans av-
eraged to give a single IR spectrum every 1.5 s. Pre-
liminary compound identifications from the GC/MS/
FTIR data sets were made by searching the Wiley/
NBS Mass Spectra Library and the EPA vapor library.
Pure samples of the identified products were obtained
to check for matching spectra (MS and FTIR) and re-
tention time.
The loss of the parent alcohols was plotted against
the formation of products yielding a linear relationship
with a slope equal to the product yield. Product quan-
tification was performed using similar GC, elution pa-
rameters, and column described for the reaction rate
determination with the column effluent analyzed by an
FID. Pure samples of identified products were ob-
tained to check for matching retention times and to
generate a quantitative calibration curve.
All measurements were at least duplicated. A rel-
ative standard deviation (the data set standard devia-
tion divided by the data set average) of approximately
2.5% was achieved with the described sampling
method. All compounds were tested for chromato-
graphic interferences and photolytic stability. Methyl
nitrite, NO, reference, and alcohols were injected into
Experimental Procedures
The experimental procedures for determining the
OH ϩ alcohols reaction kinetics were similar to those
described previously [14,15].
k
Alcohol
Alcohol ϩ OH 9
9
: Products
(4)
(5)
k
R
Reference ϩ OH 99 : Products
The rate equations for reactions 4 and 5 are combined
and integrated resulting in the following equation:
[
Alcohol]0
k
[R]
0
Alcohol
ln
ϭ
ln
(I)
ͩ
ͪ ͩ ͪ
[
Alcohol]t
k
[R]t
R
If reaction with OH is the only removal mechanism
for the alcohol and reference, plot of
ln ([Alcohol] /[Alcohol] ) vs. ln ([R] /[R] ) yields a
a
0
t
0
t
straight line with an intercept of zero. Multiplying the
slope of this linear plot by k yields kAlcohol. The OH
R
rate constant experiments employed the use of the ref-
erence compounds n-nonane (2BU), dodecane (2BU

7
48
BAXLEY AND WELLS
Figure 1 Relative rate data for 2 butanol with both n-non-
ane (᭿) and dodecane (᭺) as reference compounds. The
Figure 2 Relative rate data for 2 pentanol with both hep-
tane (᭜) and dodecane (᭺) as reference compounds. The
OH ϩ 2PE rate constant, k , measured is (11.89 Ϯ
OH ϩ 2BU rate constant, k BU, measured is (8.09 Ϯ
2
2
PE
1
2
3
Ϫ1 Ϫ1
0
.36) ϫ 10 cm molecule s .
Ϫ12
3
Ϫ1 Ϫ1
0
.66) ϫ 10 cm molecule s .
the bag and left for a minimum of 6 h to determine
stability. None of these preliminary experiments
yielded observable wall losses, chromatographic peak
overlaps, chemical photolysis, or observable losses
due to dark reactions. At the end of each run, the Tef-
lon௡ bag was cleaned by flushing the bag a minimum
of 6 times with zero-air. Measurements of an air-filled
bag showed no cross contamination between runs.
Hydroxyl Radical/Alcohols Reaction Rate
Constant (k2BU and k2PE
)
The individual alcohol/reference data sets yielded
the following hydroxyl radical rate constant results:
Ϫ
12
cm3
2
-butanol/n-nonane ϭ (9.26 Ϯ 0.14) ϫ 10
Ϫ1 Ϫ1
molecule s
,
2-butanol/dodecane ϭ (8.14 Ϯ
Ϫ12
3
Ϫ1 Ϫ1
0.47) ϫ 10
cm molecule s
,
3
2-pentanol/
Ϫ12
Ϫ1 Ϫ1
heptane ϭ (9.93 Ϯ 1.21) ϫ 10 cm molecule s ,
Ϫ
12
and 2-pentanol/dodecane ϭ (12.46 Ϯ 0.88) ϫ 10
RESULTS
3
Ϫ1 Ϫ1
cm molecule s . The combination of both alcohol/
reference data sets are shown in Figures 1 (2BU) and
Hydroxyl Radical/Alcohol Reaction Rate
Constant
2
(2PE). The slopes of the lines in Figures 1 and 2
yield an hydroxyl radical bimolecular rate constants
Ϫ
12
3
Ϫ1 Ϫ1
The OH rate constants for 2-butanol (2BU,
of (8.09 Ϯ 0.36) ϫ 10 cm molecule s
and
Ϫ12
3
Ϫ1 Ϫ1
CH CH CH(OH)CH )
and
2-pentanol
(2PE,
(11.89 Ϯ 0.66) ϫ 10 cm molecule s for 2-bu-
tanol, k2BU, and for 2-pentanol, k2PE, respectively. The
data points at the origin are experimental points be-
cause preirradiation, t ϭ 0, data showed no detectable
loss of 2BU, 2PE or references. The error in the rate
constant stated above is the 95% confidence level from
the random uncertainty in the slope. Incorporating the
uncertainties associated with the reference rate con-
3
2
3
CH CH CH CH(OH)CH ) were obtained using the
3
2
2
3
relative rate method described above. Typically five
experimental runs were conducted on each alcohol/
reference pair. The plots of a modified version of eq.
(I) are shown in Figures 1 (2BU) and 2 (2PE). The
ln([Alcohol] /[Alcohol] ) term is multiplied by the re-
0
t
spective recommended reference rate constant (dode-
Ϫ
12
3
Ϫ1 Ϫ1
cane (14.2 Ϯ 3.6) ϫ 10 cm molecule s and n-
stants used (Ϯ25%) yields final values for k
of
2BU
Ϫ12
3
Ϫ1
Ϫ1
Ϫ12
3
Ϫ1 Ϫ1
nonane (10.2 Ϯ 2.6) ϫ 10 cm molecule s and
(8.1 Ϯ 2.0) ϫ 10 cm molecule s and k2PE of
Ϫ
12
3
Ϫ1
Ϫ1
Ϫ12
3
Ϫ1 Ϫ1
heptane (7.15 Ϯ 1.79) ϫ 10 cm molecule s )
(11.9 Ϯ 3.0) ϫ 10 cm molecule s
which are
Ϫ
12
3
Ϫ1 Ϫ1
[
19] and divided by 10 cm molecule s resulting
within experimental error of the individual alcohol/
reference data set results. Assuming an OH concentra-
in a unitless number and yielding a slope that is equal
6
Ϫ3
to the hydroxyl radical/alcohol rate constant, kAlcohol,
tion of 1 ϫ 10 molecules cm , the atmospheric (1/
Ϫ12
3
Ϫ1 Ϫ1
divided by 10 cm molecule s . This modification
allows for simultaneous comparison of the two refer-
ence compound/alcohol data sets to demonstrate data
consistency.
e) lifetimes calculated for 2BU and 2PE are 34 and 23
hours, respectively. The 2BU/OH rate constant, k2BU
has been previously measured, using the relative rate
method, by Chew et al. [4] yielding a value of
,

HYDROXYL RADICAL REACTION RATE CONSTANT
749
Ϫ
12
3
Ϫ1 Ϫ1
(
9.2 Ϯ 2.4) ϫ 10 cm molecule s , within error
(kAlcohol Ϫ kP)
F ϭ
limits of our reported value. Wallington et al., using
flash-photolysis resonance fluorescence, reported a
k2PE of (11.8 Ϯ 0.8) ϫ 10 cm molecule s , also
kAlcohol
[
Alcohol]t
Ϫ12
3
Ϫ1
Ϫ1
1 Ϫ
[
Alcohol]0
within error limits of our reported value [1]. Using
ϫ
(II)
k
[
Alcohol]
[Alcohol]t
t
ͩ ͪ
P/kAlcohol
structure reactivity, the calculated k
of 9.8 ϫ
Ϫ
2
BU
Ϫ12
3
Ϫ1
Ϫ1
Ϫ
12
[Alcohol]
0
[Alcohol]
0
1
0
cm molecule s
and k2PE of 11 ϫ 10
3
Ϫ1 Ϫ1
cm molecule s are within error limits of our and
the previously reported values. [11]
F, the correction factor, was multiplied by the product
concentration data; kAlcohol is the OH ϩ alcohol rate
constant, and k is the rate constant for the reaction of
OH/Alcohols Reaction Products
P
OH with reaction product. The value for k reported
P
The observed OH/alcohol reaction products are con-
sistent with previously observed hydroxyl radical at-
mospheric chemistry reaction mechanisms for organic
compounds [6–9]. For 2BU and 2PE, the major OH/
alcohol reaction products observed were ketones of the
same length as the alcohol and the carbonyl group in
the same position as the hydroxyl group of the parent
alcohol. The specific product results for each alcohol
are described in separate sections below. The reported
yields are based on the slopes of the alcohol reacted
versus product formed plots (Figure 3). The reported
error in the product yield is the 95% confidence level
from the random uncertainty in the slope of these
plots.
Because the OH/alcohol reaction products could
subsequently react with OH, the observed product
concentrations had to be corrected for OH ϩ
reaction product reactions. This correction has been
described in detail [7,20] and has the following
form:
in ref. [19] was used. It should be noted that all of the
products exhibited linear concentration profiles; the
lack of curvature strongly suggests no unusual side
reactions that generate or remove primary reaction
products. For completeness, the k values are pre-
P
sented with the respective product in Table I.
2-Butanol/OH Reaction Product Results
Two major products were observed from the 2-buta-
nol/OH reaction: methyl ethyl ketone (MEK) and ac-
etaldehyde. Quantification of MEK yields (Figure 3)
was achieved with GC/FID and DNPH derivatization
was used to quantify acetaldehyde yields. The product
results data are summarized in Table I.
The large correction for acetaldehyde was due
mainly to the large value of kAcetaldehyde relative to
k2BU
.
2-Pentanol/OH Reaction Product Results
Three major products were observed from the 2-pen-
tanol/OH reaction: 2-pentanone, propionaldehyde, and
acetaldehyde. Quantification of both 2-pentanone and
propionaldehyde yields was achieved with GC/FID
and DNPH derivatization was used to quantify acet-
aldehyde yields. The product results data are summa-
rized in Table I.
The large corrections for propionaldehyde and ac-
etaldehyde were due mainly to the large value of
kProduct relative to k2PE
.
DISCUSSION
OH reacts with alcohols by H-atom abstraction and
both 2BU and 2PE are molecules with several possible
abstraction sites. However, the products of the reaction
of OH with alcohols strongly suggest that the OH ab-
stracts hydrogen principally from the alcoholic carbon.
This is consistent with the reaction “hot spot” pro-
Figure 3 MEK yield (determined from slope of plot) from
reaction of OH with 2BU. Yield data are corrected for re-
action of OH with MEK. A yield of 0.60 Ϯ 0.02 was ob-
served.

7
50
BAXLEY AND WELLS
Table I 2-Butanol/OH and 2-Pentanol/OH Reaction Product Data and Correction Factors
kProduct
Yield
%
(Units of
10 molecule s )
Favg
%
FMaximum
Ϫ12
Ϫ1 Ϫ1
Product
-Butanol
Structure
%
2
Methyl Ethyl
Kentone
Acetaldehyde
CH C("O)CH CH
HC("O)CH3
60 Ϯ 2
29 Ϯ 4
1.15 [19]
15.8 [19,21–23]
1.8
41
3.6
127
3
2
3
2
-Pentanol
2
-Pentanone
CH C("O)(CH ) CH
3
41 Ϯ 4
14 Ϯ 2
40 Ϯ 4
4.9 [19]
19.6 [19]
15.8 [19,21–23]
7
30
24
12
86
67
3
2 2
Propionaldehyde HC("O)CH CH
Acetaldehyde
2
3
HC("O)CH3
kProduct is the OH/product reaction bimolecular rate constant, Favg is the average correction factor calculated
using eq. (II) for the yield data set, and FMaximum is the largest correction factor calculated using Eq. (II) for the
yield data set.
posed by structure reactivity analysis of both 2BU and
PE. [11] Also, the agreement between the calculated
mechanism proposed, MEK is a product of the hydro-
gen abstraction from the alcoholic carbon. Using 18
2
O
vs. the measured kAlcohols supports these “reactive site”
assignments. However, the structure reactivity calcu-
lated MEK and 2-pentanone yields were 84% and
labeled 2BU and mass spectral analysis of OH ϩ
2BU reaction products, insights into the details of the
1
8
CH CH C ؒ ( OH)CH radical reaction pathway
3
2
3
7
6
6%, respectively. The lower observed ketone yields,
0% (MEK) and 41% (2-pentanone), reported here
could be investigated. The MEK product from reaction
1
8
7a was pure O labeled. This result confirms retention
of the alcoholic oxygen in the product and means that
and in ref. [4] suggests the 9OH activating effects
are less than assumed by structure reactivity calcula-
tions and/or due to competition for acetaldehyde and
propionaldehyde formation.
The proposed OH reaction mechanisms for 2BU
and 2PE are shown below. Depending on the nature
of the radical formed in Reaction (4): For 2BU:
1
6
atmospheric oxygen (mainly O ) does not play a role
2
in the reaction mechanism. If atmospheric oxygen was
involved in MEK formation, the MEK product would
1
8
16
16
either be mixed O/ O or pure O, neither of which
was observed. This finding is instructive for assessing
the incremental reactivity (ozone-forming potential) of
2BU and possibly other aliphatic alcohols. Tropo-
CH CH CH(OH)CH ϩ OH !:
spheric ozone is a by-product of the oxidation of vol-
3
2
3
1
8
CH CH C ؒ (OH)CH ϩ H O (6a)
atile organic compounds (VOCs) [12]. The O exper-
iment provides definitive evidence that the CH CH C
3
2
3
2
3
2
CH CH CH(OH)CH ϩ OH !:
ؒ (OH)CH radical/atmospheric oxygen reaction does
3
2
3
3
CH C ؒ HCH(OH)CH ϩ H O (6b)
not yield a tropospheric ozone-forming RO ؒ type
3
3
2
2
radical. [12] To our knowledge, this is the first direct
confirmation of the retention of the alcoholic oxygen
in the reaction product.
CH CH C·(OH)CH ϩ O !:
3
2
3
2
CH CH C("O)CH (MEK) ϩ HO
(7a)
3
2
3
2
The other reaction product, acetaldehyde, was ob-
served by DNPH derivatization. The acetaldehyde
could have come from decomposition (Reactions 7b
and 9) and reaction of the CH CH ؒ radical (Reaction
CH CH C ؒ (OH)CH (decompose) !:
3
2
3
CH CH ؒ ϩ HC("O)CH (Acetaldehyde) (7b)
3
2
3
3
2
CH CH ؒ ϩ NO ϩ O !: !: !:
8). The detection method used could not distinguish
between different acetaldehyde pathways.
3
2
2
HC("O)CH (Acetaldehyde)
3
(
via an RO ؒ intermediate) ϩ NO
(8)
2-Pentanol (2PE) more than likely has similar
mechanistic pathways as 2BU discussed above.
2
2
CH C ؒ HCH(OH)CH ϩ NO ϩ O !: !: !:
3
3
2
HC("O)CH (Acetaldehyde)
3
For 2PE:
(
via an RO ؒ intermediate) ϩ NO
(9)
2
2
The major product of the OH/2BU reaction was
CH CH CH CH(OH)CH ϩ OH !:
3
2
2
3
methyl ethyl ketone (MEK (60 Ϯ 2)%). From the
CH CH CH C ؒ (OH)CH ϩ H O (10a)
3 2 2 3 2

HYDROXYL RADICAL REACTION RATE CONSTANT
751
CH CH CH CH(OH)CH ϩ OH !:
native reaction paths resulting in products that are not
simple straight chain aldehydes and ketones such as
cyclic compounds, diols and others. As can be seen
from the mechanism proposed, there are several routes
for propionaldehyde and acetaldehyde formation.
More labeling experiments to sort these transformation
pathways are needed to clarify these mechanisms.
Also, experiments to determine nitrate reaction prod-
uct transformation pathways are needed to more fully
understand organic compounds’ atmospheric mecha-
nisms.
3
2
2
3
CH CH C ؒ HCH(OH)CH ϩ H O (10b)
3
2
3
2
CH CH CH CH(OH)CH ϩ OH !:
3
2
2
3
CH C ؒ HCH CH(OH)CH ϩ H O (10c)
3
2
3
2
CH CH CH C ؒ (OH)CH ϩ O !:
3
2
2
3
2
CH CH CH C("O)CH
3
2
2
3
(2-Pentanone) ϩ HO2 (11)
CH CH CH C ؒ (OH)CH ϩ NO ϩ O
3
2
2
3
2
(
decompose) !: !: !: CH CH C("O)H
3 2
(
Propionaldehyde)(via an RO ؒ intermediate)
2
CONCLUSION
ϩ CH C("O)H(Acetaldehyde)
3
(
via an RO ؒ intermediate) ϩ NO
(12)
2
2
The OH radical abstracts hydrogen principally from
the alcoholic carbon on aliphatic alcohols. The 2BU/
OH and 2PE/OH bimolecular rate constants, k2BU
and k2PE, were measured using the relative rate tech-
CH CH C ؒ HCH(OH)CH ϩ NO ϩ O (decompose)
3
2
3
2
!: !: !: CH CH C("O)H
3 2
(
Propionaldehyde) (via an RO ؒ intermediate)
2
Ϫ12
nique. The values for k
((8.1 Ϯ 2.0) ϫ 10
2
BU
ϩ CH C("O)H (Acetaldehyde)
3
3
Ϫ1 Ϫ2
Ϫ
12
cm molecule s ) and k ((11.9 Ϯ 3.0) ϫ 10
2
PE
(via an RO ؒ intermediate) ϩ NO2 (13)
2
3
Ϫ1 Ϫ1
cm molecule s ) are consistent with the single pre-
viously measured values for each alcohol; using an
CH C ؒ HCH CH(OH)CH ϩ NO ϩ O
3
2
3
2
6
Ϫ3
OH radical concentration of 1 ϫ 10 molecules cm
(
(
decompose) !: !: !: CH C("O)H
Acetaldehyde) (via an RO ؒ intermediate)
3
the atmospheric (1/e) lifetimes for 2-butanol and 2-
pentanol are approximately 34 and 23 h, respectively.
The identification and quantification of the 2BU/
OH and 2PE/OH reaction products allowed a tentative
development of atmospheric reaction mechanisms for
these alcohols. For 2BU, the observed yields for
methyl ethyl ketone and acetaldehyde were (60 Ϯ
2
ϩ NO ϩ unobserved products (14)
2
The major product, 2-pentanone, is expected to be
formed the same way as the MEK product in the 2BU
mechanism above. The smaller yield of 2-pentanone,
compared to structure reactivity calculations, could be
due in part to the multiple reaction/decomposition
2)% and (29 Ϯ 4)%, respectively. For 2PE, the ob-
served yields for 2-pentanone, acetaldehyde, and pro-
pionaldehyde were (41 Ϯ 4)%, (40 Ϯ 4)%, and
pathways for the CH CH CH C ؒ (OH)CH radical
3
2
2
3
and more OH reactive sites on 2PE. The multiple rad-
ical pathways leading to product are probably due to
the stability of this larger radical. The same implica-
tion regarding the effect of ketone yield data on incre-
mental reactivity that was made for 2BU above can be
made for 2PE.
The pathways for propionaldehyde and acetalde-
hyde formation are not as straightforward as 2-pentan-
one’s. The combination of acetaldehyde and propi-
onaldehyde carbon structures yields a five carbon
backbone like 2PE. At first inspection, the yields of
acetaldehyde and propionaldehyde might be expected
to be the same if both were formed from decomposi-
tion of the same 2PE molecule. However, the yields
of propionaldehyde ((14 Ϯ 2)%) and acetaldehyde
(
14 Ϯ 2)%, respectively.
Using O labeled 2BU, transformation product ex-
18
periments led to the verification that the alcoholic oxy-
gen is retained in the principal ketone product. This
finding coupled with product yield information has po-
tential impact on the incremental reactivity (ozone
forming potential) calculations of this class of com-
pounds. However, as more methylene groups are
added to the carbon backbone isomerization pathways
that result in ringed, multi carbonyl, or other oxygen-
ated reaction products become more likely. More in-
vestigations into the detailed atmospheric transfor-
mation chemistry of alcohols are needed to more
accurately describe the air quality impact of these vol-
atile organic compounds.
((40 Ϯ 4)%) are significantly different to suggest each
product may have multiple formation routes. As the
carbon backbone lengthens the possibility of isomer-
ization increases. [24, 25] The primary radical can self
abstract hydrogen and open the door to several alter-
The authors wish to thank Dr. Jean Andino, University of
Florida, for helpful comments and discussions.

7
52
BAXLEY AND WELLS
BIBLIOGRAPHY
13. R. Atkinson, W. P. L. Carter, A. M. Winer, and J. N.
Pitts, Jr. J. Air Pol. Control, 31, 1090 (1981).
1
1
1
4. L. N. O’rji and D. A. Stone, Int. J. Chem Kinet., 24, 703
1992).
5. D. C. Williams, L. N. O’rji, and D. A. Stone, Int. J.
Chem. Kinet., 25, 539 (1993).
6. D. F. Smith, T. E. Kleindienst, E. E. Hudgens, and J. J.
Bufalini, Intern. J. Environ. Anal. Chem., 54, 265
1
2
3
. T. J. Wallington, P. Dagaut, R. Liu, and M. J. Kurylo,
Int. J. Chem. Kinet., 20, 541 (1988).
. T. J. Wallington and M. J. Kurylo, Int. J. Chem. Kinet.,
(
1
9, 1051 (1987).
. L. Nelson, O. Rattigar, R. Neavyn, H. Sidebottom, J.
Treacy, and O. J. Nielsen, Int. J. Chem Kinet., 22, 1111
(
(1994).
1990).
. A. A. Chew and R. Atkinson, J. of Geophys. Res., 101,
8649 (1996).
. W. P. L. Carter, J. Air and Waste Man. Assoc., 44, 881
1994).
1
1
7. D. F. Smith, T. E. Kleindienst, and E. E. Hudgens, J.
Chromatog., 483, 431 (1989).
8. W. D. Taylor, D. Allston, M. J. Moscato, G. D. Fazekas,
R. Kozlowski, and G. A. Takacs, Int. J. Chem. Kinet.,
4
5
6
2
(
1
2, 231 (1980).
. J. R. Wells, F. L. Wiseman, D. C. Williams, J. S. Bax-
1
9. R. Atkinson, J. Phys. Chem. Ref Data, Monograph No.
ley, and D. F. Smith, Int. J. Chem. Kinet., 28, 475
(1996).
2, (1995).
7
8
. D. F. Smith, C. D. McIver, and T. E. Kleindienst, Int.
J. Chem. Kinet., 27, 453 (1995).
. D. F. Smith, T. E. Kleindienst, E. E. Hudgens, C. D.
McIver, J. J. Bufalini, Int. J. Chem. Kinet., 24, 199
20. R. Atkinson, S. M. Aschmann, W. P. L. Carter, A. M.
Winer, and J. N. Pitts Jr., J. Phys. Chem., 86, 4563
(1982).
21. H. Niki, P. D. Maker, C. M. Savage, and L. P. Breiten-
bach, J. Phys. Chem., 82, 132 (1978).
(1992).
9
. T. J. Wallington, J. M. Andino, A. R. Potts, S. J. Rudy,
W. O. Siegl, Z. Zhang, M. J. Kurylo, and R. E. Huie,
Environ. Sci. Technol., 27, 98 (1993).
22. R. Atkinson and J. N. Pitts Jr., J. Chem. Phys., 68, 3581
(1978).
23. J. V. Michael, D. G. Keil, and R. B. Klemm, J. Chem.
Phys., 83, 1630 (1985).
24. R. Atkinson, E. S. C. Kwok, J. Arey, and S. M. Asch-
mann, Faraday Discuss., 100, 23 (1995).
25. J. Eberhard, C. M u¨ ller, D. W. Stocker, and J. A. Kerr,
Environ. Sci. Technol. 29, 232 (1995).
1
1
1
0. M. P. Foster, R. Guillermo, and J. C. Galloo, Int. J.
Chem. Kinet., 28, 235 (1996).
1. E. S. C. Kwok, and R. Atkinson, Atmos. Environ., 29,
1
685 (1995).
2. J. H. Seinfeld, Science, 243, 745 (1989).