Gas-phase rate coefficients for the OH + n -, i -, s -, and t-butanol reactions measured between 220 and 380 K: Non-arrhenius behavior and site-specific reactivity

Article

pubs.acs.org/JPCA

Gas-Phase Rate Coeﬃcients for the OH + n‑, i‑, s‑, and t‑Butanol

Reactions Measured Between 220 and 380 K: Non-Arrhenius

Behavior and Site-Speciﬁc Reactivity

Max R. McGillen,^†^,^‡Munkhbayar Baasandorj,^†^,^‡^,^§and James B. Burkholder*^,^†

^†Chemical Sciences Division, Earth System Research Laboratory, National Oceanic and Atmospheric Administration, 325 Broadway,

Boulder, Colorado 80305, United States

‡

Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, Colorado 80309, United States

*

S Supporting Information

ABSTRACT: Butanol (C H OH) is a potential biofuel alternative in fossil fuel gasoline and

4

9

diesel formulations. The usage of butanol would necessarily lead to direct emissions into the

atmosphere; thus, an understanding of its atmospheric processing and environmental impact

is desired. Reaction with the OH radical is expected to be the predominant atmospheric

removal process for the four aliphatic isomers of butanol. In this work, rate coeﬃcients, k, for

the gas-phase reaction of the n-, i-, s-, and t-butanol isomers with the OH radical were

measured under pseudo-ﬁrst-order conditions in OH using pulsed laser photolysis to produce

OH radicals and laser induced ﬂuorescence to monitor its temporal proﬁle. Rate coeﬃcients

were measured over the temperature range 221−381 K at total pressures between 50 and 200

Torr (He). The reactions exhibited non-Arrhenius behavior over this temperature range and

no dependence on total pressure with k(296 K) values of (9.68 ± 0.75), (9.72 ± 0.72), (8.88

0.69), and (1.04 ± 0.08) (in units of 10⁻¹²cm molecule s ) for n-, i-, s-, and t-butanol,

3

−1 −1

±

respectively. The quoted uncertainties are at the 2σ level and include estimated systematic

errors. The observed non-Arrhenius behavior is interpreted here to result from a competition between the available H-atom

abstraction reactive sites, which have diﬀerent activation energies and pre-exponential factors. The present results are compared

with results from previous kinetic studies, structure−activity relationships (SARs), and theoretical calculations and the

discrepancies are discussed. Results from this work were combined with available high temperature (1200−1800 K) rate

coeﬃcient data and room temperature reaction end-product yields, where available, to derive a self-consistent site-speciﬁc set of

n

reaction rate coeﬃcients of the form AT exp(−E/RT) for use in atmospheric and combustion chemistry modeling.

1

. INTRODUCTION

butanols under combustion conditions, their oxidation under

atmospheric conditions warrants attention as part of a thorough

biofuel life-cycle assessment.

The OH radical reaction is expected to be the predominant

atmospheric loss process for the aliphatic isomers of butanol

As a result of their clean-burning properties and high energy

density, butanols (C H OH) are considered as potential fuel

7

4

9

replacements and additives as the desire for alternative gasoline

and diesel formulations increases. As a consequence of fuel

usage, release of butanol into the atmosphere is necessary; a

correlation has been shown that exists between the composition

of a fuel formulation and the trace volatile organic compound

composition of the atmosphere. Butanols are also emitted

into the atmosphere from natural processes where n-, i-, and s-

butanol have been identiﬁed in the emission proﬁles of a variety

OH + CH3CH2CH2CH2OH(n‐butanol) → products

OH + (CH ) CHCH OH(i‐butanol) → products

(1)

(2)

(3)

(4)

3 2

2

1

,2

OH + CH CH CH(OH)CH (s‐butanol) → products

3

2

3

OH + (CH ) COH(t‐butanol) → products

3

3 3

of plants. Alcohols have been observed to be a dominant

biogenic oxygenated volatile organic carbon (OVOC) emission

in certain environments and quantiﬁcation of butanol

where their atmospheric lifetimes are relatively short with

values in the range of ∼0.5−6 days; the actual butanol lifetime

depends on the time and location of its emission. Wet

deposition is an additional possible loss process for the more

soluble s- and t-butanol isomers, which would reduce their

atmospheric lifetimes further. On the basis of their relatively

4

emissions from the biosphere is important for air quality

emission inventories; based on observations from plant

enclosure studies, butanol emission rates have been observed

to be greater than those of monoterpenes for some common

3

plant species. n- and t-butanol have been observed in the

atmosphere in a limited number of studies with reported

concentrations of up to 5.4 and 0.63 ppb, respectively.

Therefore, in addition to understanding the oxidation of

Received: March 18, 2013

Revised: April 24, 2013

5

6

©

XXXX American Chemical Society

A

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

short atmospheric lifetimes, butanol emissions are expected to

have the greatest impact on local and regional chemistry and air

quality.

The site-speciﬁc reactivity in reactions 1−4 is of critical

importance to atmospheric chemistry, since the site at which an

abstraction reaction proceeds ultimately determines the stable

end-product(s) formed in most cases. The possible sites for H-

atom abstraction in reactions 1−4 and the nomenclature used

throughout this paper are given in Figure 1. The butanol

more accurate representation of the rate coeﬃcients at

atmospherically relevant temperatures.

The temperature dependence of site-speciﬁc reactivity for

13−16

hydrocarbons has been studied extensively,

whereas the

database for site-speciﬁc reactivity for alcohols (OVOCs) is

1

7,18

much more limited. Tully and co-workers

investigated the

site-speciﬁc reactivity of i-propanol, CH CH(OH)CH , and

3

neopentyl alcohol, (CH ) CCH OH, between 293−745 and

3

2

287−903 K, respectively. In the case of the neopentyl alcohol

reaction, separation of site reactivity was achieved by comparing

the phenomenological rate coeﬃcient with that of neopentane,

(

CH ) C, whereas for i-propanol, experiments were conducted

3 4

with partially deuterated alcohols (the primary isotope eﬀect

reduces the site reactivity signiﬁcantly), from which the site-

speciﬁc temperature dependence for each reactive site (except

for the hydroxyl hydrogen atom) was obtained. In a more

19

recent study, Carr et al. also used partial deuteration to derive

site-speciﬁc reactivities for the OH + ethanol (CH CH OH)

3

2

reaction sites over the temperature range 298 to 864 K. The i-

propanol and ethanol studies report clear diﬀerences between

the temperature dependences of each reactive site and non-

Arrhenius temperature dependence for the site-speciﬁc rate

coeﬃcients that constitute the phenomenological rate coef-

ﬁcient. The butanol isomers studied here have several diﬀerent

reactive sites, see Figure 1, that are expected to have similarly

complex temperature dependences analogous to the alcohols

studied previously.

Figure 1. Aliphatic isomers of butanol (C H OH). α, β, γ and δ

4

9

preﬁxes denote C−H bonds that are 0, 1, 2, and 3 carbon atoms away

from the OH containing carbon atom, respectively. prim, sec, and tert

denote primary (CH −), secondary (−CH −), and tertiary (>CH−)

3

2

H atoms.

In this study, phenomenological rate coeﬃcients for the OH

radical reaction with the four aliphatic isomers of butanol (n, i,

s, and t), see Figure 1, were measured over the temperature

range 221−381 K using a pulsed laser photolysis−laser induced

ﬂuorescence (PLP-LIF) technique. This work extends the low

temperature range over that available in the current literature.

Experiments were also performed using isotopically labeled OH

radicals, OH, to quantify possible OH elimination from the

initial radicals formed in reactions 1−4; rapid OH elimination

was identiﬁed and quantiﬁed for the n-butanol reaction. The

present results are compared with the results from previous

studies and the discrepancies as well as the methods used to ﬁt

the temperature dependent rate coeﬃcients are discussed.

Finally, the present high-precision phenomenological rate

coeﬃcient results are combined with literature high temper-

ature rate coeﬃcient data, structure−activity relationship (SAR)

results, and previously reported end-product yields to develop a

semiempirical self-consistent set of temperature-dependent site-

speciﬁc rate coeﬃcients for each of the OH + butanol isomer

reactions over the temperature range 220−1800 K relevant to

both atmospheric and combustion chemistry.

isomers have unique combinations of H-atom abstraction

reactive sites that are expected to exhibit diﬀerent room

temperature rate coeﬃcients, k(298 K), and temperature

dependences. Site-speciﬁc rate coeﬃcients are, however, not

routinely measured experimentally. Instead phenomenological

rate coeﬃcients, which are the sum of the contributions of each

of the reactive sites at a given temperature, are more commonly

measured experimentally, e.g., by observing the time-resolved

loss of the OH radical. Phenomenological rate coeﬃcients for

hydrocarbon and OVOC reactions are expected to exhibit non-

Arrhenius behavior, i.e., curvature in ln(k) vs 1/T, due in part

to the combination of site-speciﬁc rate coeﬃcient temperature

dependences, although high precision rate coeﬃcient data over

a range of temperatures may be needed to clearly observe such

behavior.

18

Phenomenological rate coeﬃcients for reactions 1−4 at

temperatures <450 K have been measured in a limited number

of studies to date. The n-butanol reaction was studied by Yujing

8

and Mellouki (263−372 K), the s-butanol reaction by Jimen

́

ez

9

10

et al. (263−354 K), i-butanol by Mellouki et al. (241−370

11

́

K), and t-butanol by Teton et al. (253−372 K) and

12

Wallington et al. (240−440 K), where the temperature

ranges covered in the studies are given in parentheses. The

precision of the rate coeﬃcient measurements in these studies,

2

. EXPERIMENTAL DETAILS

In this study, rate coeﬃcients for the OH radical gas-phase

reaction with the n-, i-, s-, and t- butanol isomers were measured

over the temperature range 221−381 K using a pulsed laser

photolysis−laser induced ﬂuorescence technique. Rate coef-

ﬁcients were measured under pseudo ﬁrst-order conditions in

OH, [butanol] ≫ [OH] with OH radicals produced using 248

nm pulsed laser photolysis of a precursor with a high OH

radical quantum yield or via rapid secondary chemistry. The

experimental apparatus used in this work has been used

extensively in previous studies from this laboratory and is

11

́

with the exception of the Teton et al. study, was, however,

not suﬃcient to observe the expected non-Arrhenius behavior.

As a consequence, Arrhenius expressions were used to ﬁt the

observed phenomenological rate coeﬃcient temperature

dependence data to within the measurement precision,

although it is not necessarily physically meaningful to do so.

The available experimental rate coeﬃcient data does not cover

the full range of atmospherically relevant temperatures, and the

reported Arrhenius expressions may lead to systematic errors

when extrapolated beyond the temperature range of the

measurements. It is therefore desirable to obtain higher

precision data over a broader temperature range to enable a

2

0,21

described in detail elsewhere.

A brief description of the

experimental apparatus and methods used as well as details

speciﬁc to the present study are given below.

B

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

.1. OH Rate Coeﬃcient Measurements. The pulsed

Article

2

3

2

Sander et al. otherwise. [OH] was varied over the range

0

10

−3

photolysis and probe laser beams passed through a jacketed

(0.3−100) × 10 molecules cm in this study.

OH radical temporal proﬁles obeyed pseudo ﬁrst-order

kinetics

3

multi-port Pyrex LIF reactor (150 cm internal volume) that

was mounted inside a vacuum housing. The laser beams were at

right angles and intersected in the middle of the LIF reactor.

OH ﬂuorescence was detected orthogonal to the plane of the

laser beams. The reactor temperature was maintained by

circulating ﬂuid from a temperature-regulated reservoir through

the reactor’s outer jacket. The temperature of the gas in the LIF

reactor was measured directly, to within ±0.5 K, using a

retractable thermocouple. The low-temperature limit (∼220 K)

of the kinetic measurements was determined by the butanol

vapor pressure and the overall precision of the OH temporal

proﬁle measurements.

⎛

⎞

⎛

⎞

[

^O^H^]t

^St

ln⎜

⎟ = ln⎜ ⎟ = −(k[butanol] + k )t = −k′t

d

⎝[OH] ⎠

⎝S ⎠

0

(II)

where S

[OH] , and k′ and k

coeﬃcients in the presence and absence of butanol,

respectively. k represents the loss of OH radicals resulting

is the OH signal at time t, which is proportional to

t

are pseudo ﬁrst-order decay rate

t

d

from its reaction with the precursor and other possible reactive

impurities as well as diﬀusion out of the detection volume.

Typical values of k were 100 to 500 s . The rate coeﬃcients

for reactions 1−4 were determined as the slope of a weighted

linear least-squares ﬁt of k′ vs [butanol].

−1

OH radicals were produced by 248 nm (KrF excimer laser)

pulsed laser photolysis of H O , HNO , or (CH ) COOH

d

2

3

3 3

H O + hν → 2OH

(5)

(6)

(7)

2

The butanol concentration in the LIF reactor was

determined using online infrared and UV absorption measure-

ments. UV absorption measurements at 184.95 nm (Hg line)

were made before the sample entered the LIF reactor in a 100

cm long absorption cell. Infrared absorption measurements

were made using Fourier transform infrared spectroscopy

HNO + hν → OH + NO

3

2

(

CH ) COOH + hν → OH + products

3

The H O source was used at temperatures >250 K, whereas

2

(

FTIR) either before or after the LIF reactor. The absorption

the (CH ) COOH source was used over the full range of

3

cross sections (band strengths) used to quantify the butanol

concentration were determined as part of this work as described

below. The absorption measurements were performed at 296 K

and the butanol concentration in the LIF reactor was scaled to

account for diﬀerences in pressure and temperature. The

butanol concentration was varied over the range (0.14−30.7) ×

temperatures included in this study and the HNO source was

3

used at temperatures >245 K. In addition to direct photolysis,

OH radicals were produced via

1

O + hν → O( D) + O

(8a)

(8b)

3

2

3

14

−3

→

^O⁽^P⁾⁺^O2

10 molecules cm

for n-, i-, and s-butanol and (0.32−48.5) ×

−3

14

1

0

molecules cm for t-butanol. The absorption measure-

1

O( D) + H O → 2OH

(9)

ments yielded consistent butanol concentrations under all

experimental conditions indicating that there was negligible loss

of butanol in the gas ﬂow through the apparatus and LIF

reactor.

2

and ¹⁸OH radicals were produced using isotopically labeled

^w^a^t^e^r^,^H2¹⁸

O

O( D) + H₂O → OH + 16_O_H

1

18

2.2. Butanol UV and Infrared Spectra. Infrared band

(10)

strengths (IBSs), at 296 K, for each of the butanol isomers were

−1

OH radical ﬂuorescence was detected following pulsed laser

determined over the range ∼2500−3300 cm (C−H stretch

region) as part of this work. The integrated band strengths were

used to determine the butanol concentration online in the

kinetic measurements using the same FTIR and absorption cell

used in the band strength determinations. Online infrared

absorption measurements were used in this work to avoid

potential systematic errors in the preparation of butanol gas-

phase sample standards. A total combustion method similar to

2

+

2

excitation in the A Σ ← X Π (ν = 0) transition at 282 nm

1

8

using a frequency doubled Nd:YAG pumped dye laser ( OH

was excited at 282.07 nm). The OH ﬂuorescence was collected

using a pair of convex lenses and detected by a photomultiplier

tube (PMT). A band-pass ﬁlter (308 nm, fwhm of ∼10 nm)

mounted in front of the PMT isolated the OH ﬂuorescence.

The PMT signal was averaged for 100 shots with a gated charge

integrator. Temporal proﬁles of OH were obtained by varying

the delay time (reaction time) between the photolysis and

probe lasers between 10 μs and 10 ms.

24

that used by Veres et al. was used to quantify the butanol

concentrations and determine the IBSs of each of the butanol

isomers relative to CO absorption. The combustion reactor

2

consisted of a 1/4″ O.D. stainless steel tube (2″ long) packed

with a palladium catalyst; the palladium catalyst, 10% powder,

was suspended on a refractory mineral wool substrate. The

reactor was in thermal contact with an insulated brass block

heated to 623 K using a cartridge heater that was ∼50% duty

cycled using a PID controller. Butanol was catalytically

combusted in a He carrier gas ﬂow in the presence of excess O₂.

The butanol IBS was determined using the relationship

The initial OH concentration, [OH] , was estimated from

the precursor concentration and photolysis laser ﬂuence, F

0

[

OH] = σΦF[precursor]

(I)

0

where σ is the precursor cross section and Φ its OH quantum

yield at the photolysis wavelength. The photolysis laser ﬂuence

was measured at the exit of the LIF reactor using a calibrated

−

2

power meter and was varied over the range 1.4−10.8 mJ cm

−

1

pulse

over the course of the study. The precursor

^∫CH ^A⁼^S^C^H

L[CO ]/4

concentration was estimated from the measured pseudo ﬁrst-

order loss rate of OH in the absence of butanol and the rate

coeﬃcient for the OH reaction with the precursor molecule.

Rate coeﬃcients, absorption cross sections, and quantum yields

2

(III)

where ∫ A is the integrated absorbance over the butanol C−

CH

H stretch band, L is the path length of the infrared absorption

cell (376 cm), S_C_His the butanol IBS, [CO ] is CO

2

22

were taken from Baasandorj et al. for (CH ) COOH and

3

2

C

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 1. Summary of Experimental Conditions and Rate Coeﬃcients, k(T), Obtained in This Work for the OH + n-Butanol

(

CH CH CH CH OH) Reaction

3

2

T

P

v

OH radical

source

photolysis laser

k′ range

k

d

−1

K) (Torr) [O₂] (cm s⁻¹)

a

[source]

b

ﬂuence

c

[OH]₀

d

[n-butanol]

e

(s

−1

)

(s

)

k(T)

f

(

¹⁶OH + n-Butanol

221

230

231

232

235

255

266

275

296

297

298

310

334

348

363

380

381

98

5.5

6.5

(CH ) COOH

11

9.5

10

13

7.6

27

30

27

12

16

910

19

12

6.5

5.8

20

5.3

19

23

13

6.1

5.8

2.3

4.7

4.1

2.4

4.4

3.8

2.9

3.2

3.4

3.8

4.1

4.7

6.1

6.7

7.0

6.4

7.6

6.8

0.67

0.55

1.2

1.3

0.45

3.0

26

0.58−1.11

0.19−3.73

0.53−4.22

0.38−3.64

0.36−7.16

0.43−17.9

1.49−20.3

1.54−23.1

3.07−19.4

1.91−30.7

0.30−13.9

0.93−18.4

1.21−19.2

0.34−9.23

0.33−9.50

0.32−10.2

0.27−9.66

0.36−5.87

0.15−6.26

0.34−2.81

0.26−12.0

1600−3100

800−6100

710

540

580

730

420

1280

480

430

210

270

120

680

210

220

170

370

140

370

580

250

130

1.865 ± 0.194

1.328 ± 0.048

1.324 ± 0.042

1.310 ± 0.056

1.292 ± 0.052

0.991 ± 0.013

0.903 ± 0.014

0.854 ± 0.017

0.808 ± 0.018

0.845 ± 0.017

0.871 ± 0.014

0.826 ± 0.017

0.783 ± 0.012

0.872 ± 0.018

0.844 ± 0.018

0.870 ± 0.021

0.867 ± 0.022

0.861 ± 0.028

0.893 ± 0.024

0.893 ± 0.048

0.857 ± 0.022

3

99

(CH ) COOH

3 3

211

216

102

99

5.6

(CH ) COOH

1300−6600

1200−5800

900−10500

1600−19100

1800−19400

1700−19700

2800−16100

1900−27500

300−12300

1400−15700

1400−15600

600−8400

500−8100

700−9200

400−8600

700−5200

800−6200

600−2700

400−10300

3

6.93

5.5

(CH ) COOH

3 3

6.5

(CH ) COOH

3 3

6.7

(CH ) COOH

3

2

3

93

6.9

H O

2

99

7.6

H O

2

17

104

102

99

6.8

H O

2

8.8

12

4.4

H O

2

8.7

HNO₃

8.4

2.0

13

106

203

117

119

102

125

134

51

9.2

(CH ) COOH

3

2

3

5.3

H O

2

11.3

12.0

11.5

12.8

12.6

23.6

13.0

10.4

H O

2

8.9

7.9

30

H O

2

H O

2

H O

2

8.0

30

5.37

H O

2

H O

2

33

131

120

H O

2

22

HNO₃

23

18

OH + n-Butanol

1

8

2

3

63

76

97

22

49

80

81

123

121

127

144

117

131

117

126

8.3

8.0

O /H

O

0.68

5.0

6.4

8.8

7.8

7.0

9.8

8.1

7.6

5.8

52

39

37

14

26

13

0.58−2.78

0.27−7.47

0.31−6.68

0.39−3.84

0.22−8.75

0.26−2.45

0.35−4.81

800−3000

500−7700

500−6400

600−3700

300−7400

400−2200

400−4500

300−5400

80

70

60

70

1.130 ± 0.046

1.079 ± 0.036

0.970 ± 0.026

0.967 ± 0.034

0.889 ± 0.031

0.912 ± 0.053

0.858 ± 0.044

0.905 ± 0.034

3

2

O /H

3

8.5

O /H

3

0.12

0.2

8.5

O /H

3

8.5

O /H

3

10.6

12.9

12.4

11.3

O /H

3

5.0

O /H

3

0.27

0.14

5.6

O /H

3

O /H

3

70

0.22−6.20

10

a

f

_U_n_i_t_s_o_f₁₀16 molecules cm . Units of 10 molecules cm . mJ cm pulse . Units of 10 molecules cm . Units of 10_m_o_l_e_c_u_l_e_s_c_m−3_.

−3

b

13

−3

c

−2

−1

d

−3

e

14

−11

3

−1 −1

Units of 10 cm molecule s .

concentration measured using infrared absorption, and the

factor of 4 accounts for the number of CO₂molecules

produced in the complete combustion of butanol.

are provided in the Supporting Information. The IBS values

measured in this work were used in the kinetic data analysis.

Butanol isomer absorption cross sections at 184.95 nm,

σ(184.95 nm), were measured relative to the infrared band

strengths described above. The 185 nm absorption by butanol

was used as a secondary measurement of the butanol

concentration in the kinetic experiments as well a measure of

possible butanol losses in the gas ﬂow through the apparatus

and LIF reactor. The UV absorption setup consisted of a 100

cm long absorption cell with quartz windows, a Hg Pen-Ray

A stable ﬂow of the butanol sample in the He carrier gas

through the combustion reactor was established and the CO₂

concentration was quantiﬁed online using FTIR absorption

−1

measured at 0.25 cm resolution. Once the CO concentration

2

had been determined the sample gas ﬂow bypassed the

combustion reactor and the butanol infrared absorption was

measured using the FTIR. CO calibration measurements were

2

lamp light source (housed in a N -purged assembly), and a

2

performed using commercial (50.1 ppm) and in-house gas

standards.

solar-blind photodiode detector. The 185 nm Hg line was

isolated at the lamp and detector using narrow band-pass ﬁlters.

σ(184.95 nm) was determined using the Beer−Lambert law

The measured integrated band strengths were reproducible

1

4

for CO concentrations in the range (3.0−150) × 10

2

−3

molecules cm . The measured IBSs are in good agreement

⎛

⎝

⎞

I

A = −ln⎜ ⎟ = σ(184.95 nm)L[butanol]

with the values available from the Paciﬁc Northwest National

FTIR

I ⎠

0

(IV)

25

Laboratory (PNNL), where agreement was to within 0.4−

.1% except for n-butanol, which was 10.4% less than reported

2

where [butanol]_F_T_I_Ris the concentration of butanol determined

by PNNL. The PNNL values are reported to have 2σ estimated

uncertainties of ∼7% that include estimated systematic errors.

Plots of the present band strength data and a table summarizing

the IBS results obtained here and the values reported by PNNL

by FTIR absorption, L is the path length of the UV absorption

cell, and I and I are the light source intensities with and

0

without butanol present in the cell, respectively. σ(184.95 nm)

values were determined to be (9.79 ± 0.22) × 10⁻, (7.69 ±

19

D

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 2. Summary of Experimental Conditions and Rate Coeﬃcients, k(T), Obtained in This Work for the OH + i-Butanol

(

(CH ) CHCH OH) Reaction

3 2 2

T

P

v

OH radical

source

photolysis laser

k′ range

k

d

−1

K) (Torr) [O₂] (cm s⁻¹)

a

[source]

b

ﬂuence

c

[OH]₀

d

[i-butanol]

e

(s

−1

)

(s

)

k(T)

f

(

¹⁶OH + i-Butanol

224

234

239

244

258

261

268

273

283

297

298

302

324

335

349

365

380

381

216

102

99

8.86

5.5

7.9

(CH ) COOH

6.1

4.4

14

7.3

4.7

5.8

6.4

6.7

3.5

3.8

7.9

4.1

7.4

7.3

6.4

1.5

7.0

3.5

6.4

7.9

5.8

4.7

5.8

5.2

7.0

6.4

7.3

1.1

0.51

2.0

1.6

18

0.39−3.94

1.13−9.62

1.08−8.02

1.47−8.49

1.46−9.64

1.56−17.1

1.23−12.2

1.46−10.4

0.89−14.4

1.60−13.8

1.43−7.20

0.44−16.7

0.64−11.0

0.89−17.0

0.65−11.5

0.50−5.70

0.66−15.4

0.44−7.80

0.69−11.4

0.50−9.02

0.62−13.4

0.61−18.8

0.39−8.38

0.63−9.43

800−8300

370

250

750

520

230

480

370

210

600

170

40

1.725 ± 0.133

1.433 ± 0.028

1.453 ± 0.041

1.286 ± 0.041

1.230 ± 0.024

1.137 ± 0.025

1.162 ± 0.028

1.122 ± 0.027

1.070 ± 0.027

1.054 ± 0.020

0.963 ± 0.016

0.974 ± 0.026

0.987 ± 0.034

0.968 ± 0.020

0.966 ± 0.021

1.030 ± 0.036

0.976 ± 0.021

0.965 ± 0.040

0.930 ± 0.040

0.910 ± 0.031

0.892 ± 0.034

0.859 ± 0.025

0.846 ± 0.028

0.852 ± 0.038

3

(CH ) COOH

1700−14100

2200−13000

2700−11400

1600−12600

2400−19500

1600−14200

1600−12000

1500−15500

1800−14800

1400−6900

1000−16500

800−11200

1100−16100

1200−11400

600−5800

800−15000

500−7500

800−10800

500−8500

700−12700

800−15200

400−7100

3

8.0

(CH ) COOH

3 3

114

101

60

8.9

(CH ) COOH

10

3

8.3

HNO₃

110

31

5.7

H O

24

2

101

103

98

13.0

7.2

H O

23

20

2

8.5

HNO₃

H O

110

37

21

9.5

34

2

100

137

119

124

117

101

142

99

8.5

HNO₃

110

2.2

14

20

8.2

O /H₂¹⁶O

3

9.8

(CH ) COOH

9.9

4.2

28

20

24

19

18

16

8.7

16

17

36

490

220

300

430

70

3

2

3

9.6

H O

13

2

7.8

H O

2

18

10.2

10.1

9.7

H O

2

26

O /H₂¹⁶O

0.15

97

3

HNO₃

130

60

164

114

165

114

110

141

126

9.9

O /H₂¹⁶O

0.13

17

3

10.0

10.8

11.0

13.6

13.4

11.6

H O

300

60

2

O /H₂¹⁶O

0.11

7.4

10

3

H O

140

260

70

2

H O

2

O /H₂¹⁶O

0.11

22

3

6.67

H O

1000−7700

410

2

18

OH + i-Butanol

1

8

2

3

97

81

137

153

8.2

O /H

O

2.8

7.3

6.4

1.08−7.65

0.43−3.24

0.51−2.93

10

1200−7200

500−2900

500−2600

50

70

0.967 ± 0.019

0.879 ± 0.063

0.846 ± 0.060

3

2

13.6

O /H

3

0.11

8.4

13.6

−3

O /H

3

0.11

8.4

a

f

_U_n_i_t_s_o_f₁₀16 molecules cm . Units of 10 molecules cm . mJ cm pulse . Units of 10 molecules cm . Units of 10_m_o_l_e_c_u_l_e_s_c_m−3_.

b

13

−3

c

−2

−1

d

−3

e

14

−11

3

−1 −1

Units of 10 cm molecule s .

−

0

.17) × 10 ¹⁹, (1.80 ± 0.04) × 10⁻¹⁸, and (4.75 ± 0.10) ×

straightforward t-butanol isomer is presented ﬁrst and the

presentation concludes with the more complex n-butanol

isomer, which possesses 5 diﬀerent hydrogen abstraction sites

that contribute to the overall measured phenomenological rate

coeﬃcient, k(T). The discussion for each isomer includes

results obtained for the ¹⁸OH radical reaction for each butanol

−18

10

for n-, i-, s-, and t-butanol, respectively.

.3. Materials. n-Butanol (≥99.7%), i-butanol (≥99.5%), s-

2

butanol (99.5%), and t-butanol (≥99.5%) samples were

degassed in several freeze−pump−thaw cycles prior to use.

Butanol was introduced into the reactor gas ﬂow by passing a

small ﬂow of He (UHP, >99.9995%) bath gas over a liquid

reservoir held at constant temperature in a water or ice−water

1

6

isomer that was used to evaluate possible OH radical

1

6

generation on a time scale of ∼10 ms or less. OH radical

formation was observed for the n-butanol reaction, but not for

the other butanol isomer reactions. In the Global Site-Speciﬁc

Reactivity Analysis section the present phenomenological rate

coeﬃcient data were used in an analysis of the butanol isomer

site-speciﬁc reactivity and its temperature dependence.

bath. In the combustion measurements He and O (UHP,

9.99%) passed through a CO absorbent trap before entering

the apparatus. Concentrated H O (>95%) was prepared by

2

9

2

bubbling N through an initially ∼60% mole fraction sample for

2

several days. HNO was added to the gas ﬂow from a 2:1 ratio,

3

by volume, HNO /H SO mixture in a Pyrex vacuum reservoir

kept in an ice−water bath. (CH ) COOH (70 wt %) was

obtained commercially and used without puriﬁcation. H O

3

2

4

Tables 1−4 summarize the experimental conditions

employed in this work and rate coeﬃcients obtained for

reactions 1−4, respectively. In general, the measured OH

temporal proﬁles were ﬁrst-order (single exponential decays)

over at least two OH radical decay lifetimes and k′ was

measured with high precision under all experimental con-

ditions. Typically, the precision of a least-squares ﬁt of the data

to eq II was ±2% or better. Representative second-order kinetic

3

2

1

8

18

(

18MΩ) and commercial H O (97%; O ≥99.9%) samples

2

were used without puriﬁcation. Gas ﬂows were measured with

calibrated electronic mass ﬂow meters and pressures were

measured using 100 and 1000 Torr capacitance manometers.

The uncertainties quoted throughout the paper are at the 2σ

conﬁdence level unless stated otherwise.

data, (k′ − k ) vs [butanol], obtained over the range of

0

temperatures included in this study are given in Figure 2 for the

3

. RESULTS AND DISCUSSION

t-butanol reaction; [t-butanol] was varied over the range (3.2−

13

−3

Rate coeﬃcient results for each of the four aliphatic butanol

480) × 10 molecules cm . The second-order data for each

isomers are reported separately below. The kinetically most

isomer showed good linearity at all temperatures with the

E

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 3. Summary of Experimental Conditions and Rate Coeﬃcients, k(T), Obtained in This Work for the OH + s-Butanol

(

CH CH CH(OH)CH ) Reaction

3

2

3

T

P

v

OH radical

source

photolysis laser

k′ range

k

d

−1

K) (Torr) [O₂] (cm s⁻¹)

a

[source]

b

ﬂuence

c

[OH]₀

d

[s-butanol]

e

(s

−1

)

(s

)

k(T)

f

(

¹⁶OH + s-Butanol

223

226

227

233

234

237

245

249

254

258

264

272

289

296

297

320

351

365

380

381

201

106

209

99

4.4

7.9

(CH ) COOH

5.0

3.2

7.6

26

7.0

4.2

6.7

4.4

4.2

4.5

3.4

3.5

3.7

4.1

6.7

4.0

8.4

8.7

4.0

8.4

6.7

7.0

1.4

3.1

6.1

7.3

7.6

6.4

5.8

7.3

6.1

7.0

0.87

0.33

1.3

2.9

0.58

4.1

2.5

5.8

7.8

9.2

23

0.19−1.39

0.30−5.29

0.22−2.75

0.59−9.41

0.43−2.06

0.61−9.02

0.64−11.0

0.88−8.95

0.71−19.0

0.45−16.7

0.14−4.53

0.54−13.0

0.59−5.16

0.46−8.45

0.26−16.1

0.47−7.69

0.37−4.86

0.82−4.49

0.46−13.2

0.93−12.9

0.24−9.17

0.26−6.36

0.46−5.72

0.86−17.6

0.22−4.48

0.52−4.56

0.79−13.2

0.22−4.73

500−2400

600−6800

700−3900

3100−14400

1000−3000

2500−12500

2200−13300

1200−10500

1000−20400

700−16900

500−5000

700−13800

700−5000

600−8200

500−15700

600−7200

500−4400

1000−4300

700−11400

900−11200

600−8600

400−5800

500−5100

700−14900

300−3700

600−3800

800−10700

400−4100

310

190

450

1.292 ± 0.147

1.260 ± 0.078

1.229 ± 0.094

1.273 ± 0.054

1.241 ± 0.033

1.169 ± 0.059

1.086 ± 0.039

1.146 ± 0.039

1.077 ± 0.023

1.057 ± 0.031

1.045 ± 0.050

1.024 ± 0.024

0.913 ± 0.037

0.970 ± 0.031

0.947 ± 0.019

0.897 ± 0.027

0.912 ± 0.064

0.879 ± 0.043

0.877 ± 0.034

0.876 ± 0.019

0.919 ± 0.027

0.864 ± 0.033

0.835 ± 0.030

0.850 ± 0.015

0.783 ± 0.039

0.806 ± 0.038

0.829 ± 0.020

0.831 ± 0.045

3

(CH ) COOH

3

9.29

5.8

(CH ) COOH

3 3

7.9

(CH ) COOH

1480

310

1970

1480

170

200

240

190

270

120

160

130

220

210

220

370

160

140

130

140

150

110

200

3

101

99

5.5

(CH ) COOH

27

3

8.1

(CH ) COOH

36

3

100

96

8.4

(CH ) COOH

29

3

7.4

HNO₃

H O

67

98

7.6

86

100

107

99

7.6

91

9.1

16

2

7.7

HNO₃

97

9.7

31

63

7.0

HNO₃

150

66

198

100

85

6.2

HNO₃

14

7.9

HNO₃

89

9.0

22

9.0

HNO₃

100

3.6

13

206

112

95

5.64

7.6

(CH ) COOH

0.60

20

3

10.1

9.3

H O

2

HNO₃

H O

150

160

22

5.1

12

95

9.2

218

133

141

129

148

126

134

61

7.3

30

2

11.6

12.3

10.3

14.0

13.5

10.6

26.9

H O

2

5.1

5.2

140

7.5

6.0

120

7.9

8.3

8. 8

22

H O

2

HNO₃

H O

9.8

19

2

5.15

H O

2

HNO₃

H O

12

2

18

OH + s-Butanol

1

8

2

3

97

28

41

81

122

118

131

103

8.0

11.6

9.2

O /H

O

0.18

0.25

2.2

5.8

5.5

6.7

13

16.3

3.3

0.37−5.46

0.85−4.33

1.32−15.5

0.98−4.45

10

500−4600

900−4000

1200−13300

900−3900

60

70

60

70

0.896 ± 0.053

0.888 ± 0.052

0.880 ± 0.027

0.854 ± 0.051

3

2

O /H

3

O /H

3

12.1

−3

O /H

3

0.36

29

a

f

_U_n_i_t_s_o_f₁₀16 molecules cm . Units of 10 molecules cm . mJ cm pulse . Units of 10 molecules cm . Units of 10_m_o_l_e_c_u_l_e_s_c_m−3_.

b

13

−3

c

−2

−1

d

−3

e

14

−11

3

−1 −1

Units of 10 cm molecule s .

precision of the least-squares ﬁts having typical uncertainties of

relatively unreactive sites and, thus, reaction 4 has the lowest

4

% or less.

rate coeﬃcient of the four aliphatic isomers with a room

−12

3

Rate coeﬃcients obtained at all temperatures were found to

temperature rate coeﬃcient of (1.04 ± 0.08) × 10

cm

be independent of the total pressure, which was varied over the

range 55−205 Torr (He). The rate coeﬃcient data from this

work is plotted in Figures 3−5 for reactions 4−2, respectively,

and Figures 6 and 7 for the n-butanol reaction. All of the

butanol isomer reactions exhibited non-Arrhenius behavior, i.e.,

curvature in ln(k) vs 1/T, over the temperature range 221 to

−1 −1

molecule s . A summary of the experimental conditions and

obtained rate coeﬃcients is given in Table 4. The rate

coeﬃcient data is displayed in Figure 3. The measured rate

coeﬃcients were independent of variations in the experimental

conditions, OH radical concentration and source, and the

addition of O to the reaction mixture (see Table 4 and Figure

3

81 K, whereas the magnitude and temperature dependence of

the non-Arrhenius behavior diﬀered for each butanol isomer

discussed below).

.1. OH + t-Butanol ((CH ) C(OH)). The OH + t-butanol

2

3) to within the precision of the measurements. Experiments

performed for the ¹OH + t-butanol reaction at 297, 341, and

380 K had ﬁrst-order ¹⁸OH decays that agreed with the OH

results to within the measurement precision. This indicated that

8

(

16

3

reaction has the fewest H-atom abstraction channels available of

the four aliphatic butanol isomers with only two diﬀerent sites

16

no measurable, <∼5%, rapid OH radical generation from

(

see Figure 1)

nascent reaction products was observed. Nevertheless, non-

Arrhenius behavior was clearly observed in the high precision

data (Figure 3), although the rate coeﬃcients measured over

the temperature range 220 to 380 K diﬀered by less than 40%,

with a negative temperature dependence observed for temper-

atures <∼250 K.

•

OH + (CH ) C(OH) → (CH ) C H C(OH) + H O

(4a)

(4b)

3

3 2

2

•

→

(CH ) CO +H O

3

2

There are 9 equivalent β_p_r_i_msites (4a) and the OH reactive

site (4b). The t-butanol abstraction channels are limited to

F

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 4. Summary of Experimental Conditions and Rate Coeﬃcients, k(T), Obtained in This Work for the OH + t-Butanol

(

(CH ) COH) Reaction

3 3

T

K)

P

v

OH radical

source

photolysis laser

k′ range

k

d

−1

a

(cm s⁻¹)

b

c

d

e

−1

f

(

(Torr) [O₂]

[source]

ﬂuence

[OH]₀

[t-butanol]

(s

)

(s

)

k(T)

¹⁶OH + t-Butanol

2

3

21

23

24

35

53

55

73

83

97

98

10

22

45

59

70

80

81

95

205

113

6.5

5.7

(CH ) COOH

5.6

2.9

9.8

3.9

5.4

8.8

6.1

11

5.8

5.5

6.4

5.2

6.1

7.0

8.1

6.4

8.7

8.1

6.7

8.3

2.6

7.0

9.8

8.8

8.0

6.8

7.8

10.8

9.0

5.8

0.81

0.40

1.6

0.32−3.87

0.68−8.47

1.63−8.40

0.96−8.18

0.79−6.99

4.02−21.4

0.65−17.8

2.68−36.2

2.55−23.2

2.54−43.0

3.45−42.7

4.66−48.5

1.94−44.9

3.71−16.6

0.78−45.8

2.41−25.0

1.76−28.4

1.61−34.6

1.24−26.3

0.63−23.6

1.24−31.3

0.94−27.2

1.56−22.2

400−700

350

180

600

240

300

140

90

1.028 ± 0.169

1.053 ± 0.059

1.117 ± 0.136

1.075 ± 0.095

0.982 ± 0.073

0.981 ± 0.032

0.984 ± 0.028

0.944 ± 0.014

1.000 ± 0.030

0.970 ± 0.017

0.974 ± 0.016

1.011 ± 0.020

0.950 ± 0.021

1.023 ± 0.026

1.072 ± 0.010

1.062 ± 0.122

1.142 ± 0.037

1.204 ± 0.034

1.310 ± 0.038

1.371 ± 0.044

1.424 ± 0.036

1.364 ± 0.060

1.416 ± 0.039

3

(CH ) COOH

300−1000

400−1500

300−1100

400−1000

600−2200

200−1800

500−3500

500−2600

500−4000

500−4200

600−4800

400−4100

500−1800

400−5400

500−3000

300−3600

400−4100

600−4100

500−3200

300−4400

300−3800

400−3200

3

7.6

(CH ) COOH

3 3

201

95

5.39

6.5

(CH ) COOH

0.51

0.83

3

6.9

(CH ) COOH

3

2

3

66

6.5

H O

14

9.6

21

0.81

16

19

0.85

2

89

9.20

7.7

H O

2

100

105

103

101

108

109

102

204

146

139

125

137

55

8.9

H O

2

180

240

130

170

180

160

140

240

80

6.9

(CH ) COOH

5.1

8.4

10.3

5.1

4.5

8.5

14

3

2

3

9.3

H O

2

8.7

H O

2

7.3

(CH ) COOH

3

2

3

9.7

H O

0.93

5.0

2

9.2

H O

2

3.8

H O

2

22

10.2

8.6

H O

2

4.7

9.0

8.6

22.4

17

4.3

9.4

H O

2

160

420

320

110

180

170

10.9

12.2

24.0

12.6

11.6

13.5

H O

2

11

H O

2

26

18

H O

2

142

128

136

HNO₃

H O

5.9

9.3

8.7

1.0

4.36

12

11

2

H O

2

18

OH + t-Butanol

1

8

297

341

380

151

136

135

8.6

7.3

O /H

O

0.9

4.4

4.9

3.5

6.7

6.4

100

0.38−7.75

3.74−35.8

1.75−30.7

10

200−900

900−4800

400−4300

30

80

70

1.051 ± 0.134

1.243 ± 0.052

1.447 ± 0.096

3

2

1

O /H

1.0

29

3

2

12.7

−3

O /H

3

2

a

f

_U_n_i_t_s_o_f₁₀16 molecules cm . Units of 10 molecules cm . mJ cm pulse . Units of 10 molecules cm . Units of 10_m_o_l_e_c_u_l_e_s_c_m−3_.

Units of 10 cm molecule s .

b

13

−3

c

−2

−1

d

−3

e

14

−12

3

−1 −1

2

7

the reported measurement precisions. Saunders et al.

reported a value for k (298 K) from a low-pressure (1 Torr)

4

discharge ﬂow resonance ﬂuorescence study that is 21% less

27

than the present results; Saunders et al. is the only study that

was performed at low-pressure. Note, however, that rate

coeﬃcients reported in the Saunders et al. study for other

reactions are also less than literature values, which may be

suggestive of a general systematic bias in their study.

1

12

Tet

́

on et al. (253−372 K) and Wallington et al. (240−

4

40 K) measured absolute rate coeﬃcients over the range of

temperatures given in parentheses. Over the temperature range

common to the present work the reported k (T) values agree

4

well within the combined estimated uncertainties of this work,

1

±

́

7%, and the errors quoted by Teton et al. and Wallington et

12

al. Wallington et al. reported an Arrhenius ﬁt to their data,

due primarily to their limited number of k (T) data points,

4

1

2 −E/RT

́

while Teton et al. reported a ﬁt using a AT e

Figure 2. Representative second-order kinetic data (symbols) for the

OH + t-butanol reaction obtain over a range of temperatures (see

legend). The lines are linear-least-squares ﬁts of the data to eq II.

parametrization to account for the observed non-Arrhenius

behavior in their data (expressions are summarized in Table 5).

There are presently no direct experimental measurements of

k (T) at combustion temperatures available in the literature,

4

The present results are compared with results from previous

studies conducted over a similar temperature range in Table 5

and all available kinetic data for the OH + t-butanol reaction are

included in the lower panel of Figure 3. The room temperature

28

29

although Moss et al. and Sarathy et al. reported rate

coeﬃcient data for temperatures >1000 K obtained indirectly

from shock tube data that will be discussed along with

structure−activity relationship (SAR) estimates in the Global

Site-Speciﬁc Reactivity Analysis section below.

12

11

rate coeﬃcient data from Wallington et al., Tet

́

on et al., and

2

6

Wu et al. are in agreement with the present results to within

G

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 3. Rate coeﬃcient data for the OH + t-butanol reaction. Upper

panel: Experimental results from this work (see Table 4 and legend,

Figure 4. Rate coeﬃcient data for the OH + i-butanol reaction. Upper

panel: Experimental results from this work (see Table 2 and legend,

blue symbols: O added). The line is from the global rate coeﬃcient

2

analysis derived in this work (see text). Lower panel: Comparison of

the present results with results from previous studies (see legend).

blue symbols: O added). The line is from the global rate coeﬃcient

2

analysis derived in this work (see text). Lower panel: Comparison of

the present results with results from previous studies (see legend).

3

.2. OH + i-Butanol ((CH ) CHCH OH). i-Butanol has four

3 2 2

diﬀerent hydrogen abstraction sites: 6 γ_p_r_i_msites, 1 β_t_e_r_tsite, 2

α_s_e_csites, and 1 OH site (see Figure 1)

∼

−400 K. The non-Arrhenius behavior for this reaction is

•

OH + (CH ) CHCH OH → (CH ) CHC HOH + H O

subtle, more so than for the OH + t-butanol reaction discussed

earlier, and it is only through high-precision kinetic measure-

ments over an extended temperature range that this behavior is

3

2

3 2

2

(2a)

•

→

(CH ) C CH OH + H O

(2b)

(2c)

(2d)

18

3

2

apparent. Rate coeﬃcients measured for the OH + i-butanol

reaction at 297 and 381 K agreed with the results obtained for

the OH reaction to within the measurement precision. The

addition of O to the reaction mixture at 224, 261, and 381 K

•

C H (CH )CHCH OH + H O

16

2

3

2

•

2

(CH ) CHCH O +H O

3

2

showed no statistical diﬀerence from experiments performed

The experimental conditions used and the rate coeﬃcients

obtained in this work over the temperature range 224 to 381 K

are summarized in Table 2; the room temperature rate

without O . This result implies that the radical products formed

in reaction 2 do not lead to the formation of secondary OH

radicals on the time scale of our experiments.

The present results are compared with results from previous

studies conducted over a similar temperature range in Table 6

and all the available literature data is included in the lower

2

−12

3

coeﬃcient was determined to be (9.72 ± 0.72) × 10 cm

molecule⁻s . The second-order kinetic plots were linear at all

1

−1

1

4

temperatures for i-butanol concentrations between 0.39 × 10

molecules cm⁻and 18.8 × 10 molecules cm . Experiments

performed under similar conditions using diﬀerent radical

sources typically agreed to within the precision of the individual

measurements. The rate coeﬃcient data, plotted in Figure 4,

has a negative temperature dependence over the range of

temperatures included in this work with a slight non-Arrhenius

behavior where the higher temperature data displays weaker

temperature dependence. For reference, a simple Arrhenius

expression, k(T) = A exp(−E/RT), ﬁt of the data yielded E/R =

3

14

−3

panel of Figure 4. There are three previous studies of k (298 K)

2

26

to compare with the present work. Wu et al. and Andersen et

30

al. used relative rate methods and reported k (298 K) values

2

that bracket the present room temperature results, 4% lower

10

and 17% higher, respectively. Mellouki et al. reported k (298

2

K) values obtained using relative rate and absolute kinetic

methods that agree to within their measurement precision,

∼15%; the spread in their k (298 K) results overlaps the

2

present results.

H

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 7. Rate coeﬃcient data for the OH + n-butanol reaction. Upper

panel: Experimental results from this work (see Table 1 and legend,

Figure 5. Rate coeﬃcient data for the OH + s-butanol reaction. Upper

panel: Experimental results from this work (see Table 3 and legend,

blue symbols: O added). Note the greater overall rate coeﬃcient

2

blue symbols: O added). The line is from the global rate coeﬃcient

18

2

obtained for the OH + n-butanol reaction (see text). The line is from

analysis derived in this work (see text). Lower panel: Comparison of

the present results with results from previous studies (see legend).

the global analysis rate coeﬃcient derived in this work (see text).

Lower panel: Comparison of the present results with results from

previous studies (see legend). The results from this work are for the

18

OH + n-butanol reaction.

similar to the present work. Their rate coeﬃcient data typically

agrees to within the experimental error of the present results as

10

shown in Figure 4. Mellouki et al. reported an Arrhenius ﬁt to

their data, see Table 6, although the present high-precision data

over a slightly greater range of temperature clearly shows a non-

31

Arrhenius behavior for this reaction. Pang et al. reported rate

coeﬃcient data at temperatures between 907 and 1147 K that

will be discussed in the Global Site-Speciﬁc Reactivity Analysis

section.

3

.3. OH + s-Butanol (CH CH(OH)CH CH ). The OH + s-

3 2 3

butanol has 5 diﬀerent hydrogen atom abstraction sites: 3 γ_p_r_i_m

sites, 3 β_p_r_i_msites, 2 β_s_e_csites, 1 α_t_e_r_tsite, and 1 OH site (see

Figure 1)

OH + CH CH(OH)CH CH

(3a)

3

2

3

•

→

CH C (OH)CH CH + H O

Figure 6. Comparison of second order kinetic data obtained in this

work at 296 K for the ¹⁶OH and OH + n-butanol reactions (see

Table 1). The lines are linear least-squares ﬁts of the data. The inset

shows an example of the measured ¹⁶OH and OH temporal proﬁles

obtained under identical experimental conditions (see text).

3

2

3

2

18

•

C H CH(OH)CH CH + H O

(3b)

(3c)

(3d)

(3e)

2

3

2

18

•

CH CH(OH)C HCH + H O

3

2

•

CH CH(OH)CH C H + H O

3

2

10

Mellouki et al. is the only previous study to report

•

CH CH(O )CH CH + H O

3

2

3

2

temperature dependent rate coeﬃcient data at temperatures

I

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 5. Summary of Rate Coeﬃcient Data for the OH + t-Butanol ((CH ) COH) Reaction from This Work and Previous

3

Studies

n

k (T) = AT exp(−E/RT)F (T)

i

x

reaction

site

b

a

ref

method

P (Torr)

55−205

T (K)

k(298 K)

A

n

E/R

F_x(T)

this work

PLP-LIF

221−381

total

1.04 ± 0.08

see Table 9

(He)

1

−12

18

Tet

́

on et al.

PLP-LIF

∼100 (He)

253−372

total

1.0 ± 0.2

(2.66 ± 0.48) × 10

3.86 ± 0.55) × 10

(3.3 ± 1.6) × 10

0

2

0

−(270 ± 130)

335 ± 88

−(310 ± 150)

−

(

−

12

Wallington et

al.

Wu et al.

FP-RF

25−50 (Ar)

240−440

295 ± 2

298

total

1.07 ± 0.08

1.11 ± 0.07

0.81 ± 0.17

1

2

6

c

RR-

760

GCFID

Saunders et

DF-LIF

ST-CM

1

2

7

al.

2

8

−17

Moss et al.

760−3040

1200−1800

β

1.35 × 10

2

227

1.48 × 10⁻

18

(

O / Ar)

OH

β

2

9

−23

Sarathy et al.

ST-CM

200−1800

0.975

1.15 × 10

3.70

2.82

−1217

−295

9.76 × 10⁻

22

OH

0.025

k_T_o_t_a_l(298 K)

=

1.0

5

2

−18

Kwok et al.

SAR

β

0.503

3 × 4.49 × 10

2.1 × 10⁻

2

320 exp(62/T)

18

OH

0.14

85 exp(62/T)

k_T_o_t_a_l(298 K)

=

0.64

3

9

Bethel et al.

SAR

β

1.06

3 × 4.49 × 10

2.1 × 10⁻

2

320 exp(285/T)

85 exp(62/T)

18

OH

0.14

k_T_o_t_a_l(298 K)

=

1.20

a

Units are 10⁻¹²cm molecule s . PLP-LIF, pulsed laser photolysis-laser induced ﬂuorescence; FP-RF, ﬂash photolysis-resonance ﬂuorescence;

3

−1 −1

b

RR-GCFID, relative rate-gas chromatography ﬂame ionization detection; DF-LIF, discharge-ﬂow resonance ﬂuorescence; ST-CM, shock tube-

c

combustion modeling; SAR, structure−activity relationship. Reported as atmospheric pressure.

Table 3 summarizes the experimental conditions and

obtained rate coeﬃcients from this work over the temperature

range 223 to 381 K where the room temperature rate

included in the lower panel of Figure 5 for comparison with the

present work. Chew and Atkinson and Baxley and Wells

32

33

reported values of k (298 K) obtained using relative rate

3

−12

3

coeﬃcient was determined to be (8.88 ± 0.69) × 10 cm

methods. The results from these studies agree with the present

work to within their relatively large reported uncertainty limits,

molecule⁻s . The second-order kinetic plots were linear at all

1

−1

1

4

temperatures for s-butanol concentrations between 0.14 × 10

± 26% and ±25%, respectively.

molecules cm⁻and 19.0 × 10 molecules cm . Experiments

performed under similar conditions using diﬀerent radical

sources agreed very well, to within 3%. Rate coeﬃcient

measurements for the OH + s-butanol reaction at 297, 328,

41, and 381 K yielded results systematically greater than

obtained for the OH reaction by up to ∼5%, but are in

agreement to within the measurement precision. It was

therefore concluded the radical products formed in reaction 3

do not lead to a rapid formation of secondary OH radicals,

3

14

−3

Jimen

́

ez et al. is the only previous study to report k

9

3

temperature dependent data. Their results are in good

agreement with the present study at the low temperatures of

their work (263−273 K) but systematically diverge at higher

1

8

3

temperatures by as much as ∼20% at 338 K, the maximum

1

6

9

́

temperature of their study, see Figure 5. Jimenez et al.

reported an Arrhenius ﬁt to their data that is given in Table 7.

The present study extends the minimum temperature range for

k₃(T) considerably, from 263 to 226 K, with high precision

measurements that clearly show a non-Arrhenius behavior for

<

5%, on the time scale of our experiments. Additionally, when

O was added to the reaction mixture no measurable diﬀerence

this reaction.₃

2

4

in the measured rate coeﬃcient was observed indicating that

OH elimination from a possible β-scission mechanism (the β

hydrogen is abstracted, reactions 3b or 3c, followed by

elimination of OH and formation of an alkene coproduct)

does not occur within the temperature range of the present

experiments. The rate coeﬃcient data is plotted in Figure 5 and

shows a negative temperature dependence at temperatures

Pang et al. reported a direct measurement of k (T) over the

3

temperature range 888−1178 K. According to the analysis of

34

Pang et al., k (T) is notably smaller than observed for the

3

other butanol isomers at comparable temperatures, to the

extent that there is no apparent activating eﬀect of the alcohol

substitution upon the rate coeﬃcient when compared with n-

35

butane. This contradicts the general reactivity trend observed

35−38

<

300 K, an Arrhenius expression, k(T) = A exp(−E/RT), ﬁt to

between alcohols and their alkane analogs.

noted that due to OH regeneration, the OH + s-butanol

It should be

the <300 K data yields an E/R of approximately −1400 K.

However, over the full temperature range of the present study,

k (T) displays a non-Arrhenius behavior, where k (T) over the

34

reaction rate coeﬃcient reported in the Pang et al. study was

subject to large corrections from secondary chemistry of the β

site radicals; the rate coeﬃcients were also sensitive to the

combustion mechanism employed. A substantial correction to

estimate the OH regenerating reaction channel was also applied

to obtain the overall rate coeﬃcient, which may have resulted in

3

temperature range 320 to 400 K is nearly temperature

independent.

Rate coeﬃcient results for the OH + s-butanol reaction from

four previous kinetic studies is given in Table 7 and the data is

J

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 6. Summary of Rate Coeﬃcient Data for the OH + i-Butanol ((CH ) CHCH OH) Reaction from This Work and

3

2

Previous Studies

n

k

i

(T) = AT exp(−E/RT)F

x

(T)

reaction

site

c

a,b

ref

method

PLP-LIF

RR-

P (Torr)

T (K)

k(298 K)

A

n

E/R (K)

x

F (T)

this work

60−216 (He)

760 (air)

224−381

295

total

0.972 ± 0.72

0.908 ± 0.35

see Table 9

2

6

Wu et al.

GCFID

RR-

GCFID

total

0.959 ± 0.45

1.05 ± 0.09

0.85 ± 0.01

0.98 ± 0.01

(3.1 ± 0.9) × 10⁻

12

0

−(352 ± 82)

Mellouki et

al.

PLP-LIF

101−111

241−373

298

1

0

(He)

RR-

GCFID

760 (air)

RR-

GCFID

298

d

total

0.92 ± 0.15

Andersen et

RR-FTIR 700

296

1.14 ± 0.17

3

0

al.

3

1

1.65× 10⁻²¹

Pang et al.

ST-LIF

ST-CM

760−912

760−3040

O / Ar)

907−1147

1200−1800

total

α

3.18

1304

−1108

−939

227

2

8

−18

Moss et al.

3.99 × 10

2

−18

(

β

1.83 × 10

8.97 × 10

1.48 × 10

2

γ

OH

227

k_T_o_t_a_l(298 K) = 1.86

6.00 × 10⁻

21

2.89

−1174

Sarathy et

al.

ST-CM

200−1800

α

0.436

2

9

−

24

17

22

β

0.772

0.058

0.0025

1.28 × 10

1.27 × 10

9.76 × 10

3.70

1.84

2.82

−2488

−75

−295

−

γ

−

OH

k_T_o_t_a_l(298 K) = 1.27

3

9

−18

Bethel et al.

SAR

α

0.333

0.506

0.035

0.014

4.50 × 10

2.12 × 10

2 × 4.49 × 10

2.10 × 10⁻

2

−253

−696

303

exp(62/T)exp(317/T)

exp(285/T)

−18

β

−18

γ

exp(62/T)

18

OH

85

exp(62/T)

k_T_o_t_a_l(298 K) = 0.888

a

Units are 10⁻¹¹cm molecule s . Errors are quoted as 2σ unless stated otherwise. PLP-LIF, pulsed laser photolysis-laser induced ﬂuorescence;

3

−1 −1

b

c

RR-GCFID, relative rate-gas chromatography ﬂame ionization detection; ST-CM, shock tube-combustion modeling; SAR, structure−activity

d

relationship. Average reported relative rate measurement with 2σ error.

an underestimation of the reported overall rate coeﬃcient. The

rate coeﬃcients estimated using the updated SAR of Bethel et

conditions, diﬀerent OH radical sources, and the addition of

16

−3

7 × 10 molecules cm O to the reaction mixture agreed very

2

3

9

28

18

al. and the analysis given by Moss et al. for temperatures

1000 K will be discussed in the Global Site-Speciﬁc Reactivity

Analysis section.

.4. OH + n-Butanol (CH CH CH CH OH). The OH + n-

well, to within 3%. However, rate coeﬃcients for the OH + n-

>

butanol reaction obtained at 263, 276, 297, 322, and 349 K

yielded results systematically greater than obtained for the

16

3

OH reaction. This result implies that the radical products

3

2

butanol reaction has 5 diﬀerent hydrogen abstraction sites: 3

δ_p_r_i_msites, 2 γ_s_e_csites, 2 β_s_e_csites, 2 α_s_e_csites, and 1 OH site (see

Figure 1)

formed in reaction 1 lead to a ∼10% rapid formation of

secondary ¹⁶OH radicals at 296 K on a time scale shorter than

the measured OH decay rates.

In addition to the measurements mentioned above, experi-

OH + CH CH CH CH OH

(1a)

1

18

3

2

ments were also performed using the O( D) + H O source in

2

•

16

18

→

CH CH CH C HOH + H O

which the OH and OH radical temporal proﬁles were

3

2

monitored nearly simultaneously in the same experiment by

•

CH CH C HCH OH + H O

(1b)

(1c)

(1d)

(1e)

16

18

3

2

switching the probe laser between OH and OH excitation

•

transitions. The inset in Figure 6 shows temporal proﬁles for

CH C HCH CH OH + H O

3

2

1

6

18

OH and OH measured at 297 K and 131 Torr (He) in such

•

C H CH CH CH OH + H O

an experiment. Although the diﬀerences in the OH decays are

2

•

relatively small, the high precision of the measured temporal

CH CH CH CH O +H O

18

16

3

2

proﬁles enables a distinct diﬀerence in the OH and OH

radical decays to be observed; the diﬀerence in this example

Table 1 summarizes the experimental conditions and

obtained rate coeﬃcients from this work over the temperature

range 221 to 381 K. The second-order kinetic plots were linear

at all temperatures for n-butanol concentrations between 1.5 ×

1

6

data is ∼10%. We attribute the diﬀerence in the OH and

18

16

OH radical decays to the rapid elimination of OH from a

reaction product containing an alcohol moiety, which leads to a

OH ﬁrst-order decay rate coeﬃcient that is slower than for

OH. Figure 6 also includes a comparison of the second-order

1

3

−3

15

−3

16

1

0

molecules cm and 3.07 × 10 molecules cm .

18

Experiments performed using a range of experimental

K

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 7. Summary of Rate Coeﬃcient Data for the OH + s-Butanol (CH CH CH(OH)CH ) Reaction from This Work and

3

2

3

Previous Studies

n

k (T) = AT exp(−E/RT)F (T)

i

x

reaction

site

c

a,b

ref

method

P (Torr)

T (K)

k(298 K)

A

n

E/R

F_x(T)

this work

PLP-LIF

ST-LIF

61−218 (He)

41−193 (He/Ar/N₂)

722−942

226−381

263−354

888−1178

297

total

α

8.88 ± 0.69

8.77 ± 1.46

8.18

see Table 9

9

Jimen

́

ez et al.

3

4

d

4.95 × 10⁻²⁰

Pang et al.

2.66

1123

3

Baxley et al.

RR-GCFID 760 (air)

8.1 ± 2.0

9.2 ± 2.4

3

2

Chew et al.

RR-GCFID 760 (air)

296

2

8

−18

Moss et al.

ST-CM

760−3040 (O / Ar)

1200−1800

1.99 × 10

2

−1561

227

2

−

18

β_p_r_i_m

β_s_e_c

γ

4.48 × 10

2

4.32 × 10⁻

4.48 × 10⁻

1.48 × 10⁻

2

−385

227

2

OH

α

2

227

2

9

−21

Sarathy et al.

200−1800

3.50

3.00 × 10

2.89

3.70

1.81

2.82

−1315

−1483

−1881

6

3.84 × 10⁻

2.56 × 10⁻

24

β_p_r_i_m

β_s_e_c

γ

0.795

2.01

24

18

22

−

0.253

0.025

8.58 × 10

OH

9.76 × 10⁻

−295

k_T_o_t_a_l(298 K) = 6.58

3

9

−18

Bethel et al.

SAR

α

6.94

2.12 × 10

4.49 × 10

4.5 × 10⁻

2

−696

320

exp(317/T) exp(62/T)

exp(285/T)

−18

β_p_r_i_m

β_s_e_c

γ

0.354

2.43

18

−253

320

exp(285/T)

4.49 × 10⁻

2.10 × 10⁻

18

exp(62/T)

0.168

0.140

OH

85

exp(62/T)

k_T_o_t_a_l(298 K) = 10.0

a

Units are 10⁻¹²cm molecule s . Errors are quoted as 2σ unless stated otherwise. PLP-LIF, pulsed laser photolysis-laser induced ﬂuorescence;

3

−1 −1

b

c

RR-GCFID, relative rate-gas chromatography ﬂame ionization detection; ST-CM, shock tube-combustion modeling; SAR, structure−activity

d

9

́

relationship; Estimate of Pang et al. of the k(298 K) rate coeﬃcient based on a ﬁt that includes the data of Jimenez et al.

kinetic data obtained at 297 K that shows a systematic

Scheme 1

16

18

diﬀerence in the OH and OH radical data used to obtain the

rate coeﬃcient data in Figure 7.

16

OH + n-butanol experiments were also performed with O₂

added to the reaction mixture and yielded identical pseudo ﬁrst-

order decay rate coeﬃcients, to within the measurement

precision, which implies that the regeneration of the OH radical

occurs with a ﬁrst-order (unimolecular) rate coeﬃcient greater

5

−1

than ∼1 × 10 s . The lack of O dependence also implies that

2

OH regeneration is not facilitated by a peroxy radical

intermediate which may eliminate an OH radical through a

Waddington-type mechanism. The fact that OH regeneration

was not observed in the reactions of the other butanol isomers

gives us additional conﬁdence in the methods used and that

OH is regenerated in the n-butanol reaction. Thus, it seems that

products have not been observed as products in the OH + n-

butanol reaction to date. As discussed further below, the major

end-products observed in the OH + n-butanol reaction are not

necessarily associated with OH elimination.

40

16

The OH + n-butanol reaction rate coeﬃcient has received a

great deal of attention, particularly at room temperature, and

the results from previous studies are listed in Table 8 and all the

available rate coeﬃcient data is included in Figure 7 for

comparison with the results from the present work. The room

the absolute measurements of k in the literature are biased

1

toward rate coeﬃcients that are too low as a consequence of a

1

8

OH regeneration channel and that the OH measurements of

this study is more representative of the actual overall rate

coeﬃcient for this reaction.

2

6,41−46

temperature studies

are for the most part in agreement

with the results from the present work, to within the combined

The coproduct and reactive site leading to the regenerated

OH radical are currently unknown. Some potential pathways

and OH regeneration coproducts in this system are provided in

Scheme 1

A possible mechanism would include 1-butene formation

following the initial H-atom abstraction at the β_s_e_csite of n-

butanol. Another possible coproduct may be cycloalkane

formation. The fact that only n-butanol was observed to exhibit

OH regeneration may be a result from its longer chain length, a

property that is generally more conducive to cyclization

reactions. Note, however, that 1-butene and cycloalkane end-

error limits. The exception is the rate coeﬃcient from the

41

Campbell et al. study, which is signiﬁcantly less than the other

2

6,42−45

studies. The relative rate studies

should, in principle, not

be inﬂuenced by OH regeneration and, therefore, in better

agreement with our ¹⁸OH results. The study of Nelson et al.

reported rate coeﬃcients that were obtained using both relative

and absolute methods and the rate coeﬃcient determined using

the relative rate technique was ∼10% greater than the value

measured using their absolute method.

44

8

Yujing and Mellouki is the only previous study to report

temperature dependent rate coeﬃcients over a temperature

L

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 8. Summary of Rate Coeﬃcient Data for the OH + n-Butanol (CH CH CH CH OH) Reaction from This Work and

3

2

Previous Studies

n

k

i

(T) = AT exp(−E/RT)F

x

(T)

reaction

site

c

a,b

ref

method

P (Torr)

T (K)

k(298 K)

A

n

E/R

x

F (T)

this work

PLP-LIF

ST-LIF

51−216

30−300

760−1520

∼1710

700

221−381

263−372

896−1197

1017−1269

296

9.68 ± 0.75 see Table 9

8

8.47 ± 0.34 (5.3 ± 1.62) × 10⁻¹²

3.24 × 10⁻¹⁰

Yujing et al.

−

0

−(146 ± 92)

4

7

Pang et al.

2505

4

8

Vasu et al.

Hurley et

RR-FTIR

_d8.86 ± 0.85

8.66 ± 0.66

4

3

al.

2

6

Wu et al.

RR-

GCFID

760

295

4

2

Cavalli et al.

RR-FTIR

740

760

298

8.28 ± 0.85

9.3 ± 0.4

4

5

Oh et al.

RR-

GCMS

9

.1 ± 0.3

Nelson et

al.

PR-KS

730−750

298

7.8 ± 0.2

8.56 ± 0.7

4

RR-

GCFID

Wallington et FP-RF

al.

25−50

100

296

8.31 ± 0.63

_e6.81 ± 1.03

4

6

Campbell et

GC−CO₂

292

4

1

al.

2

8

−18

Moss et al.

ST-CM

760−3040

1200−1800

α

3.99 × 10

4.32 × 10

4.48 × 10

1.48 × 10

6.00 × 10

2.56 × 10

1.89 × 10

8.77 × 10

9.76 × 10

2

−1108

−385

227

−18

−21

β

γ

δ

OH

α

227

Sarathy et

al.

ST-CM

200−1800

4.08

2.01

3.34

0.151

2.89

3.70

2.87

0.97

2.82

−1154

−1881

−1473

799

2

9

−24

β

−21

−15

−22

γ

δ

OH

0.025

k_T_o_t_a_l(298

K) = 9.61

−295

3

9

−18

Bethel et al.

SAR

α

3.33

2.99

1.15

0.168

4.50 × 10

4.49 × 10

2.10 × 10

2

−253

320

exp(62/T)exp(317/T)

exp(62/T)exp(285/T)

exp(62/T)

−18

β

γ

δ

exp(62/T)

OH

0.140

85

exp(62/T)

k_T_o_t_a_l(298

K) = 7.78

f

2.14 × 10⁻¹²

Galano et

al.

theory

(CCSD)

290−500

total

6.88

0

−440

4

9

g

9

.46

5

0

6.69 × 10⁻²³

2.89 × 10⁻

h

Zhou et al.

theory

500−2000

total

3.57

3.69

−2128

−1703

(G3)

23

theory

(CCSD)

5

1

Seal et al.

theory

200−2400

6.07

(MS-

VTST)

a

Units are 10⁻¹²cm molecule s . Errors are quoted as 2σ unless stated otherwise. PLP-LIF, pulsed laser photolysis-laser induced ﬂuorescence;

3

−1 −1

b

c

ST-LIF, shock-tube laser induced ﬂuorescence; RR-FTIR, relative rate-Fourier transform infrared; RR-GCFID, relative rate-gas chromatography

ﬂame ionization detection; RR-GCMS, relative rate-gas chromatograph; PR-KS, pulse radiolysis-kinetic spectroscopy; FP-RF, ﬂash photolysis-

resonance ﬂuorescence; GC−CO , measurement of CO yield from the reaction using gas chromatography; ST-CM, shock tube combustion

2

modeling; SAR, structure−activity relationship; CCSD, quantum chemical calculation using coupled-cluster with single and double excitations; G3,

d

quantum chemical calculation using the G3 potential energy surface. Quoted errors represent the extremes of two diﬀerent determinations.

e

f

g

Quoted errors are unknown. Speciﬁc calculation at room temperature. Room temperature rate coeﬃcient based on calculated Arrhenius

h

51

53

parameters. Seal et al. employ the expression of Zheng to describe the temperature dependence of these reactions and for clarity of presentation

are omitted in this table, and readers are referred to Seal et al. for those expressions.

5

1

2

8

47

range similar to the range used in the present work. Their rate

221 K. Moss et al. and Pang et al. have reported rate

coeﬃcient data at temperatures >1000 K that are discussed in

the Global Site-Speciﬁc Reactivity Analysis section.

1

6

coeﬃcient data is in good agreement with the OH

experiments of this study, although as discussed above the

rate coeﬃcient results are subject to a systematic error due to

OH regeneration. Our study has extended the temperature

range of the ¹⁶OH rate coeﬃcient measurements from 263 to

In summary, the OH radical reaction with the four aliphatic

isomers of butanol was measured over the temperature range

221 to 381 K in this work. The use of multiple OH radical

M

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Table 9. Summary of OH + Butanol Rate Coeﬃcient Global Fit Results Obtained from This Work

n

k (T) = AT exp(−E/RT)

i

k (298 K) (10⁻cm molecule s⁻¹

12

3

−1

)

branching ratio (298 K) A (cm molecule s⁻¹

3

−1

)

n

E/R (K)

molecule

n-butanol

CH CH CH CH OH

reactive site

i

6.38 × 10⁻

3.31 × 10⁻

2.43 × 10⁻

4.49 × 10

2.10 × 10

18

1.93

4.11

2.67

2.00

−792

−1858

−797

258

α_s_e_c

β_s_e_c

γ_s_e_c

5.42

2.49

1.42

0.17

0.05

0.57

0.26

0.15

0.02

0.01

25

20

3

2

a

−18

δ_p_r_i_m

a

−18

OH

23

k_T_o_t_a_l(298 K) = 9.56

i-butanol

CH ) CHCH OH

α_s_e_c

β_t_e_r_t

γ_p_r_i_m

5.45

4.19

0.34

0.05

0.54

0.42

0.03

<0.01

9.91 × 10⁻

1.63 × 10⁻

8.98 × 10⁻

19

2.18

2.64

2.00

−924

−1289

258

20

18

(

3

2

a

−18

OH

2.10 × 10

23

k_T_o_t_a_l(298 K) = 10.0

s-butanol

CH CH CH(OH)CH

3

α_t_e_r_t

5.59

1.51

1.86

0.05

0.63

0.17

0.20

<0.01

4.91 × 10⁻

9.71 × 10⁻

2.75 × 10⁻

19

2.20

2.52

2.41

2.00

−1107

−656

−595

23

20

19

β_s_e_c

3

2

β_p_r_i_m+ γ_p_r_i_m

a

−18

OH

2.10 × 10

k_T_o_t_a_l(298 K) = 9.01

k_T_o_t_a_l(298 K) = 1.02

17

t-butanol

CH ) COH

β_p_r_i_m

0.56

0.43

0.04

0.54

0.42

0.04

1.34 × 10⁻

6.83 × 10⁻

2

227

−548

23

14

(

β

3

complex

a

−18

OH

2.1 × 10

a

39

Modiﬁed Arrhenius parameters taken from the Bethel et al. structure−activity relationship (SAR).

sources and variation of experimental conditions combined

with the online determination of the butanol concentration

enabled highly precise and accurate rate coeﬃcient data to be

obtained. In general, the agreement with previously reported

k(T) is good, although the present results are of higher

precision and enable the identiﬁcation of non-Arrhenius

behavior (curvature); t-butanol and s-butanol showing the

clearest non-Arrhenius behavior. Evidence for rapid OH radical

regeneration from the nascent products in the n-butanol

reaction was observed, where the regeneration was ∼10% at

butanol isomers have many similarities an interpretation of site-

speciﬁc reactivity for each isomer needs to be considered

individually. The general approach used in the present site-

speciﬁc reactivity analysis was to take direct kinetic measure-

ments of the total (phenomenological) rate coeﬃcient over as

broad a temperature range as possible and minimize the

diﬀerence between parametrized and measured values. The

derived site-speciﬁc parametrization was subject to several

constraints, notably the reaction branching ratios were

constrained by room temperature end-product yield studies

where possible. The temperature dependence of the site-

speciﬁc rate coeﬃcient was deﬁned by a modiﬁed Arrhenius 3-

parameter expression for each reactive site and the total

(phenomenological) rate coeﬃcient was assumed to be the sum

of the independent site-speciﬁc rate coeﬃcients

2

96 K of the total reaction and decreased at higher

temperatures.

4. GLOBAL SITE-SPECIFIC REACTIVITY ANALYSIS

Site-speciﬁc reactivity is of particular interest to both

combustion and atmospheric chemistry. In combustion, the

reaction of butanol with the OH radical is a major fuel

oxidation pathway and hydrogen abstraction at diﬀerent

reaction sites will produce diﬀerent outcomes. For example,

OH attack at the β site of n-butanol is expected to recycle OH

n

−E /RT

i

k_t_o_t_a_l(T) =

∑

A T e

i

(V)

where A , n , and E are parameters for reactive site i, R is the gas

i

constant, and k_t_o_t_a_l(T) is the total rate coeﬃcient.

4

7

radicals and regenerating reactions such as this directly aﬀect

the eﬃciency of the combustion process. Under atmospheric

conditions, the stable end-products formed following the initial

OH + butanol reaction, mainly aldehyde or ketone end-

products, have diﬀerent environmental impacts. Ketones are, in

general, transported over larger distances before being oxidized

or photolyzed and, therefore, represent a trace species that

impacts atmospheric chemistry away from the emission source.

Aldehydes, on the other hand, have higher reactivity and,

therefore, have a greater inﬂuence on local and regional air

quality.

Reaction end-product yields provide a valuable constraint on

site-speciﬁc reactivity. Quantitative end-product data for the

OH + butanol isomer reactions, however, is limited and the

available studies were only performed at room temperature.

42

43

Cavalli et al. and Hurley et al. performed environmental

chamber studies of the OH + n-butanol reaction and reported

end-products associated with the α, β, and γ site H-atom

30

abstraction. Andersen et al. studied the OH + i-butanol

reaction and reported an acetone (CH C(O)CH ) end-product

3

yield that results from reaction at the β site of i-butanol. Baxley

33

and Wells studied the OH + s-butanol reaction and reported

The n-, i-, s-, and t-butanol isomers have diﬀerent H-atom

abstraction sites and rate coeﬃcients each of which is expected

to have a diﬀerent temperature dependence. The site-speciﬁc

reactivity depends on the degree of alkyl substitution as well as

the separation from the OH functional group. Therefore,

although the rate coeﬃcients for the OH reaction with the

yields of methyl ethyl ketone (CH CH C(O)CH , MEK) and

3

2

3

acetaldehyde (CH CHO), which are produced from reaction at

3

the α and β sites of s-butanol, respectively. Chew and

32

Atkinson also reported an MEK yield for the s-butanol

reaction. The accuracy of end-product yields is typically ±10%,

whereas the end-product mass balance is often more uncertain.

N

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

The present analysis was optimized to reproduce the end-

product yields, i.e., site-speciﬁc reaction branching ratios, to

within this level of uncertainty.

At high- and low-temperatures, site-speciﬁc reactivity

information is indirect or only available through predictive/

estimation methods such as structure−activity relationships

28

(

SARs). Moss et al. modeled shock tube data, over the

temperature range 1200 to 1800 K, using Evans−Polanyi

relationships to estimate site-speciﬁc OH reaction rate

coeﬃcients for the butanol isomers. In general, there is good

agreement between the rate coeﬃcients estimated by Moss et

Figure 8. α, β, and γ site pre-reactive complex geometries possible in

the OH + butanol reaction.

2

8

31,47,48

al.,

available experimental data,

and the SAR

39

predictions of Bethel et al. An extrapolation of their derived

site-speciﬁc rate coeﬃcient expressions to lower temperatures,

however, is in poor agreement with the experimental rate

coeﬃcient data from this work and previous studies. In another

study, Sarathy et al. estimated site-speciﬁc rate coeﬃcients

indirectly over a broad range of temperatures that included

both atmospheric and combustion temperature regimes. The

Sarathy et al. study drew on a variety of data sources in their

analysis including high-temperature shock tube data, theoretical

calculations (published and unpublished), and experimentally

available total rate coeﬃcient data. The rate coeﬃcient

expressions provided by Sarathy et al. reproduce the

experimental observations qualitatively but do not provide the

quantitative level of precision sought in the present work. The

positive temperature dependence and resemble alkane-like

reactivity.

The site-speciﬁc rate coeﬃcient analysis presented below was

constrained by results from high temperature studies, SAR

estimates, and results from end-product yield studies. The

accuracy of the site-speciﬁc temperature dependent expressions

was determined primarily by the accuracy and availability of the

kinetic data. In general, the total and site-speciﬁc rate

coeﬃcient analysis presented here is considered to be more

accurate over the entire temperature range than previously

reported rate coeﬃcient expressions derived from experimental

data over narrower temperature ranges, SAR estimates, and

theoretical calculations, which are included in Figures 9−12 for

comparison. The absence of experimental rate coeﬃcient and

product yield data in the temperature range ∼400−1000 K

increases the uncertainty of the derived parametrization in this

temperature region. As a result of limitations in the available

data, the parametrizations derived here are not a unique

solution to eq V, but does provide a physically based solution

that is self-consistent with the available data. The extrapolation

of the site-speciﬁc rate coeﬃcients outside the range of the

experimental data, particularly in the low temperature regime, is

expected to be more uncertain.

29

2

9

2

9

28

29

results from the Moss et al. and Sarathy et al. studies,

although not used in the present analysis, are included in

Figures 9−12 and Tables 6−9 for comparison purposes.

The SAR estimated site-speciﬁc temperature dependent rate

coeﬃcients were used to constrain the high temperature,

>∼600 K, parametrization in our analysis. SAR estimated rate

coeﬃcients are expected to be most reliable at high temperature

since the less reactive sites of butanol are expected to behave

more like ordinary alkyl hydrogen sites than hydrogen sites that

are activated by an alcohol moiety. The total rate coeﬃcients

4

.1. OH + n-Butanol (CH CH CH CH OH) Site-Speciﬁc

3 2 2 2

39

Reactivity Analysis. The results of the global site-speciﬁc rate

coeﬃcient analysis for the OH + n-butanol reaction are given in

Table 9 and plotted in Figure 9. Including the high temperature

data in the Arrhenius plot clearly illustrates the non-Arrhenius

behavior of the total rate coeﬃcient over the 200−2000 K

temperature range for the butanol isomer reactions, see Figures

obtained using the SAR of Bethel et al. are included in Figures

39

9

−12. The SAR of Bethel et al. was also used to estimate the

rate coeﬃcients of the minor reaction channels since no

product branching ratios are available for these reaction

pathways and the present analysis was not sensitive to the

minor channels.

9

−12. The n-butanol parametrization reproduces the measured

At low temperatures k_t_o_t_a_l(T) for each of the butanol isomers

has a negative temperature dependent component that is not

always well represented in the SAR estimates, in particular for

the OH + t-butanol reaction. The negative temperature

dependence implies the formation of a pre-reactive complex

in the entrance channel of the reactions potential energy surface

total rate coeﬃcient data and end-product yields to within 14%

over the temperature range 221−1800 K.

The determination of stable end-product yields greatly

constrained the site-speciﬁc reactivity analysis. A simpliﬁed n-

butanol atmospheric degradation mechanism, initiated at each

of the possible H-atom abstraction sites, under high NO

conditions, given in Scheme 2 illustrates that unique end-

products result from the diﬀerent reaction channels.

(PES). The formation of pre-reactive complexes is known to be

a common characteristic of OH radical reactions with

4

9,50

oxygenated species.

Possible pre-reactive complex geo-

Butanal (CH CH CH CHO), propanal (CH CH CHO),

3

2

3

2

metries for the OH + butanol reaction are illustrated in Figure

acetaldehyde (CH CHO), and glycolaldehyde (HOCH CHO)

3

2

8^.^T^h^e^f^o^r^m^a^tⁱ^oⁿ^o^f^p^r^e^-^r^e^a^c^tⁱ^v^e^c^o^m^p^l^e^x^e^s^eⁿ^h^aⁿ^c^e^s^ktotal(T),

yields provide valuable information for all branching reaction

pathways except for the reaction at the OH site, which is

expected to give a negligibly small yield of butanal, and the

reaction at the δ site, which is expected to yield a hydroxy

carbonyl end-product. As mentioned earlier, two end-product

chamber studies are available in the literature for the OH + n-

at low temperature, while the magnitude and temperature

dependence of the enhancement depends on the stability of the

complex, the reaction activation barrier, and the tunneling

coeﬃcient. At present there are no experimental or theoretical

studies of the OH + butanol reaction that quantitatively address

the low-temperature reaction regime. The contribution of

complex formation to k_t_o_t_a_l(T) is expected to diminish at high

temperature where complex formation is less favorable; under

high temperature conditions k_t_o_t_a_l(T) is expected to have a

42

butanol reaction. Cavalli et al. quantiﬁed the yields of butanal

(CH CH CH CHO), propanal (CH CH CHO), and acetalde-

3

2

3

2

hyde (CH CHO), which are associated with H-atom

3

abstraction at the α, β, and γ sites, to be 52 ± 7%, 23 ± 4%,

O

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

uncertainties in each case. The yields and uncertainties of

butanal, propanal, and acetaldehyde used in the present analysis

42

were taken from Cavalli et al. to constrain the branching

ratios of the α, β, and γ reaction pathways at room temperature.

39

For reaction at the δ and OH sites, the SAR of Bethel et al.

was used to ﬁx these minor branching ratios. The sum of the

reaction channels was scaled to a 100% total yield. Note that

these studies have not identiﬁed a coproduct consistent with

the OH radical regeneration identiﬁed in the present study

discussed earlier.

39

The updated SAR of Bethel et al. was used to ﬁx the total

rate coeﬃcient at temperatures >1200 K and was also used to

estimate reaction channel branching ratios in this temperature

47

range. The results from the shock tube study of Pang et al.,

over the temperature range 900−1200 K, were also included in

the analysis. Pang et al. reanalyzed data reported in an earlier

48

study from the same group and their recommended rate

coeﬃcient temperature dependence includes estimated rate

coeﬃcient data for the β site reaction; at high temperature the β

site reaction leads to rapid regeneration of OH radicals and,

therefore, their experiments were not sensitive to this reactive

channel. Note that the mechanism for OH radical regeneration

at combustion temperatures may diﬀer from the mechanism

leading the formation of OH radicals near room temperature

that was observed in our work. As shown in Figure 9 the results

from the Pang et al. study are in reasonable agreement with the

SAR prediction. Below ∼400 K the SAR underestimates the

total rate coeﬃcient considerably and fails to capture the

observed negative temperature dependence and therefore was

not used in this temperature range.

The results of the site-speciﬁc analysis are included in the

lower panel of Figure 9 together with total rate coeﬃcient data

Figure 9. Global site-speciﬁc rate coeﬃcient analysis results for the

OH + n-butanol reaction between 200 and 2000 K. Upper panel: Total

rate coeﬃcient from this work (solid line, see Table 9) with results

from previous studies, structure−activity relationships (SAR), and

theoretical calculations (see Table 8) included for comparison (see

legend). Lower panel: Total and site-speciﬁc reaction rate coeﬃcients

derived in the present work (see text; Figure 1 for nomenclature).

47

from this work and Pang et al. and the room temperature

end-product branching ratio data. The total rate coeﬃcient data

in this analysis is considered to be most reliable, with direct

kinetic measurements available over a broad temperature range.

The region between the highest temperature measurements of

this study, 381 K, and the lowest temperature of the Pang et

47

End-product yields of acetaldehyde, CH CHO, (γ-sec. site reaction,

3

al. study are based on a smoothed interpolation. The site-

speciﬁc reactivity below room temperature is only constrained

by total rate coeﬃcient data and is therefore considered less

certain, since there are no reliable estimates or product studies

available in this temperature region.

green circles), propanal, CH CH CHO, (β-sec. site reaction, magenta

3

2

circles), and butanal, CH CH CH CHO, (α-sec. site reaction, blue

3

2

circles), are included with closed and open circles the results of Cavalli

42

43

et al. and Hurley et al., respectively.

To the best of our knowledge the OH + n-butanol is the only

butanol isomer reaction that has been studied theoret-

Scheme 2

49−51

ically.

The theoretical results are included in Figure 9

and Table 8 for comparison with the experimental data and

present analysis. In general, the theoretically calculated rate

coeﬃcients are in reasonably good agreement at high

temperatures, >1000 K, but deviate from each other at lower

4

9

temperatures. Galano et al.

calculated reaction rate

coeﬃcients between 290 and 500 K using conventional

transition state theory and their results are in reasonable

agreement with experimental data near room temperature.

However, systematic deviations from the experimental data are

evident at higher temperatures with the calculated values being

systematically low, i.e. a greater calculated negative temperature

50

dependence. Zhou et al. calculated reaction rate coeﬃcients

between 500 and 2000 K using G3 and CCSD potential energy

surfaces and variable reaction coordinate transition state theory.

The CCSD calculations of Zhou et al. are in better overall

agreement with the global analysis of the present work, despite

the slightly better performance of the G3 calculations at 1000

4

3

and 13 ± 2%, respectively. Hurley et al. reported yields for

these same products to be 44 ± 4%, 19 ± 2%, and 12 ± 3%,

respectively, which are slightly lower than the yields of Cavalli

50

4

2

et al., but are within the combined reported experimental

P

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

5

1

K. More recently, Seal et al. reported a theoretical

investigation using multistructural variational transition state

theory (MS-VTST) with which they probed the eﬀects of

diﬀerent conformers on the transition state. At the temper-

atures of the present experimental work the calculations of Seal

et al. underestimate the experimental rate coeﬃcient data by a

factor of 1.6, or more, and report a weaker negative

temperature dependent component than observed experimen-

tally.

Scheme 3

4

.2. OH + i-Butanol ((CH ) CHCH OH) Site-Speciﬁc

3 2 2

Reactivity Analysis. The results of the global site-speciﬁc rate

coeﬃcient analysis for the OH + i-butanol reaction are given in

Table 9 and plotted in Figure 10. The analysis reproduces the

Scheme 3 shows that several reaction pathways lead to

unique end-products that enable identiﬁcation of the site-

speciﬁc reactivity. Isobutanal ((CH ) CHCHO), acetone, and

3

2

acetaldehyde yields provide valuable information on all

branching reaction pathways except for the reaction at the

OH site, which is expected to give a negligibly small yield of

30

isobutanal. Andersen et al. reported an end-product yield

study for this reaction and a 61 ± 4% acetone yield, which was

attributed to H-atom abstraction at the β_t_e_r_tsite. A yet

unpublished chamber study conducted by us constrained the

acetone yield to a value closer to 40%, which was used in the

present analysis. We conclude that the higher acetone yield in

30

the Andersen et al. study results from a contribution of

secondary acetone formation from isobutanal oxidation. For

39

reaction at the δ and OH sites, the SAR of Bethel et al. was

used to ﬁx the branching ratios and the remainder of the total

rate coeﬃcient was attributed to reaction at the α site.

39

The SAR of Bethel et al. was used to ﬁx the total rate

coeﬃcient at temperatures >1200 K and was also used to

estimate reaction channel branching ratios in this temperature

31

range. The results from the shock tube study of Pang et al.

over the temperature range 900−1150 K were also included in

the analysis. As shown in Figure 10, the results from the Pang et

31

al. study are in good agreement with the SAR prediction. The

SAR provides a good estimate of the rate coeﬃcient at ∼400 K,

but underestimates the negative temperature dependence of

this reaction at lower temperatures.

The results of the site-speciﬁc analysis are shown in the lower

panel of Figure 10 together with total rate coeﬃcient data and

end-product branching ratios. The total rate coeﬃcient data is

considered to be most reliable with direct kinetic measurements

available over a broad temperature range. The site-speciﬁc

reaction rate coeﬃcients below room temperature are less

certain as there are no reliable estimates or product studies at

these temperatures.

Figure 10. Global site-speciﬁc rate coeﬃcient analysis results for the

OH + i-butanol reaction between 200 and 2000 K. Upper panel: Total

rate coeﬃcient from this work (solid line, see Table 9) with results

from previous studies and structure−activity relationships (SAR; see

Table 6) included for comparison (see legend). Lower panel: Total

and site-speciﬁc reaction rate coeﬃcients derived in the present work

4.3. OH + s-Butanol (CH CH(OH)CH CH ) Site-Speciﬁc

3

2

3

Reactivity Analysis. The results of the global site-speciﬁc rate

coeﬃcient analysis for the OH + s-butanol reaction are given in

Table 9 and plotted in Figure 11. The analysis reproduces the

experimental data to within experimental uncertainty over the

temperature range 221−1800 K, which is more accurate than

previously reported rate coeﬃcient expressions included in

Figure 11.

A simpliﬁed s-butanol atmospheric degradation mechanism,

initiated at each of the possible H-atom abstraction sites, under

high NO conditions, is given in Scheme 4

(

see text; Figure 1 for nomenclature). Acetone (CH C(O)CH ) end-

3 3

30

product yield (β site reaction) reported by Andersen et al. (ﬁlled

magenta circle) is included.

experimental data to within experimental error over the

temperature range 224−1800 K, which is more accurate than

previously reported rate coeﬃcient expressions derived from

experimental data over narrower temperature ranges and SAR

estimates, see Figure 10.

A simpliﬁed i-butanol atmospheric degradation mechanism,

initiated at each of the possible H-atom abstraction sites, under

high NO conditions is given in Scheme 3

Scheme 4 shows that several reaction pathways possess

unique end-products for the oxidation of s-butanol by the OH

radical. 2-butanone (CH C(O)CH CH ), acetaldehyde

3

2

3

Q

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

32

2

-butanone. Chew and Atkinson reported a 2-butanone yield

of 69.5 ± 7.3%, which is attributed to reaction at the α site.

tert

33

Baxley and Wells reported a 2-butanone yield of 60% ± 2% in

addition to an acetaldehyde yield of 29 ± 4% that constrains

the site-speciﬁc reactivity of the ethyl β reaction channel. End-

product yields for the other reaction channels are currently not

available in the literature. For reaction at the γ and OH sites,

39

the SAR of Bethel et al. was used for the branching ratios and

the remainder of the total rate coeﬃcient was attributed to

39

reaction at the methyl β site. The SAR of Bethel et al. was

used for the total rate coeﬃcient at temperatures >1200 K and

to estimate reaction channel branching ratios in this temper-

ature range. The SAR provides a good estimate of the rate

coeﬃcient at ∼400 K but overestimates the total rate coeﬃcient

at lower temperatures.

The results of the site-speciﬁc analysis are shown in the lower

panel of Figure 11, where the temperature dependence of each

channel are provided together with total rate coeﬃcient data

and the available experimental room temperature product

branching ratios. The total rate coeﬃcient data <400 K is

considered the most reliable as only SAR estimates are available

in the high temperature regime, although the SAR estimates

have been shown to be in very good agreement with

experimental data at high temperatures for the n- and i-butanol

reactions. The branching ratios were constrained by the end-

product measurements at room temperature and the site-

speciﬁc reactivity at other temperatures was guided by SARs

and is expected to be less certain than the total rate coeﬃcient

analysis.

4

.4. OH + t-Butanol ((CH ) C(OH)) Site-Speciﬁc

3 3

Reactivity Analysis. The results of the global site-speciﬁc

rate coeﬃcient analysis for the OH + t-butanol reaction are

given in Table 9 and plotted in Figure 12. The rate coeﬃcient

parametrization reproduces the experimental data to within the

estimated experimental uncertainties. The analysis for the OH

Figure 11. Global site-speciﬁc rate coeﬃcient analysis results for the

OH + s-butanol reaction between 200 and 2000 K. Upper panel: Total

rate coeﬃcient from this work (solid line, see Table 9) with results

from previous studies and structure−activity relationships (SAR) (see

Table 7) included for comparison (see legend). Lower panel: Total

and site-speciﬁc reaction rate coeﬃcients derived in the present work

+

t-butanol reaction diﬀers from that of the other isomer

reactions in several key aspects. Notably, there is only one

major reaction site available in the t-butanol molecule, that is, 9

^e^q^uⁱ^v^a^l^eⁿ^t^βprim sites. As a result, SAR estimates the total rate

coeﬃcient to have a weak non-Arrhenius behavior (see upper

panel in Figure 12). However, the extension of the rate

coeﬃcient measurements in this study to 221 K has more

clearly identiﬁed a much stronger negative temperature

dependence at low temperature. Since a relatively strong non-

Arrhenius behavior was observed, the relative contributions of

direct abstraction and a complex-forming reaction channel were

estimated. As mentioned earlier there are several possible cyclic

complexes that could be formed, t-butanol does not, however,

possess an α reaction site and hence cannot form a 5-

membered ring complex (see Figure 8). This restriction may

account, in part, for the lower overall rate coeﬃcient for the t-

butanol reaction when compared with the n-, i-, and s-butanol

reactions.

(

see text; Figure 1 for nomenclature). End-product yields of

acetaldehyde, CH CHO, (β-sec. site reaction, magenta circle) and 2-

butanone, CH C(O)CH CH , (α-tert. site reaction, blue circles) are

included with the closed and open circles the results of Baxley and

Wells and Chew and Atkinson, respectively.

3

2

3

33

32

Scheme 4

In addition, stable end-product studies for the OH + t-

butanol reaction would not provide useful insight to the rate

coeﬃcient parametrization since both the β and OH sites are

expected to produce formaldehyde and acetone equally, see

Scheme 5 for a simpliﬁed degradation mechanism under high

NO conditions.

(

CH CHO), and propanal (CH CH CHO) yields provide

In order to estimate the abstraction and complex formation

contributions in our analysis the abstraction channel of the t-

butanol reaction was assumed to behave like an alkane as

3

2

useful information on the site-speciﬁcity, except for the γ and

OH sites. The γ site is expected to lead to minor end-products,

while the OH site is expected to make a negligibly small yield of

39

deﬁned by the SAR of Bethel et al; the activation of the β site

R

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

given here below 220 K may lead to systematic error. Direct

kinetic measurements at lower temperatures are desirable to

reﬁne the present analysis further.

4

.5. Summary. The site-speciﬁc rate coeﬃcient para-

metrizations derived in this work reproduce the available

experimental kinetic data more accurately over the temperature

range 220−1800 K than the empirical and theoretical

expressions currently available in the literature. Given the

accuracy and precision of the measurements at temperatures

below 400 K and the predictability of the overall rate coeﬃcient

in the temperature range 1000−1800 K, the summed

expression provides a method for estimating the total rate

coeﬃcient in the temperature range 400−1000 K where no

measurements are presently available. The semiempirical

treatment provided here is considered preferable in terms of

accuracy to an extrapolation of expressions derived solely from

high or low temperature kinetic data. The formulation of site-

speciﬁc and total reactivity developed here is appropriate for

use in atmospheric and combustion modeling. By using the

above approach to ﬁtting the entire kinetic data set, it was

shown that the diﬀerent reaction sites of the butanol isomers

exhibit pronounced non-Arrhenius behavior. This implies that

these reactive sites have contributions from both direct and

complex forming reaction mechanisms.

5

. CONCLUSIONS

In this study, rate coeﬃcients for the reaction of the OH radical

with n-, i-, s-, and t-butanol were measured over the

temperature range 220 to 380 K. The precision of the present

measurements and extension to lower temperatures than

available in previous studies enabled non-Arrhenius behavior

to be observed for each isomer reaction. The present results

were combined with available OH + butanol end-product

Figure 12. Global site-speciﬁc rate coeﬃcient analysis results for the

OH + t-butanol reaction between 200 and 2000 K. Upper panel: Total

rate coeﬃcient from this work (solid line, see Table 9) with results

from previous studies and structure−activity relationships (SAR) (see

Table 5) included for comparison (see legend). Lower panel: Total

and site-speciﬁc reaction rate coeﬃcients derived in the present work

3

1,34,47,48

results, combustion temperature kinetic studies,

and

rate coeﬃcients estimated from structure−activity relationships

39

(SARs) to develop a self-consistent semiempirical set of site-

speciﬁc reactivities for each butanol isomer (see Table 9) over

the temperature range 220−1800 K. The non-Arrhenius

behavior in the total (phenomenological) reaction rate

coeﬃcient is clearly apparent when data over the entire 220−

(

see text; Figure 1 for nomenclature).

Scheme 5

1

800 K temperature range is considered. The transition of the

total reaction rate coeﬃcient from a positive temperature

dependence at high temperatures, >600 K, to a negative

temperature dependence at temperatures <250 K is observed

for all of the butanol isomers. The transition corresponds to a

transition in reaction mechanism, from a more complex-

forming regime at lower temperatures to a primary abstraction

reaction at high temperatures.

The results from this study are considered to be of use in a

number of research areas. In addition to deﬁning the total rate

coeﬃcient over a broad temperature range, stable end-product

branching ratios can be estimated for use in atmospheric

chemical models, e.g. in the determination of photochemical

ozone creation potentials (POCPs) and evaluation of the

eﬀects of n-butanol emissions on local and regional air quality.

In addition, the total and site-speciﬁc reaction rate coeﬃcients

of the OH + n-butanol reaction are of direct interest to the

by the alcohol substituent was neglected. The total rate

coeﬃcient was optimized very well by adding an Arrhenius

expression with a negative activation energy to represent the

contribution of the complex forming reactive channel. The

results are shown in the lower panel of Figure 12 and the

parameters are given in Table 9, where E/R for the complex

forming channel was found to be ∼−550 K. The interpretation

of the reaction mechanism presented in this analysis is not

unique, although the experimentally available rate coeﬃcient

data obtained over the temperature range 221−370 K is

reproduced to within the precision of the measurements. At

present only SAR estimates are available for the rate coeﬃcient

at temperatures greater than ∼400 K and should be considered

less certain. Extrapolation of the rate coeﬃcient expressions

2

8,29

combustion community

in the development of low-

temperature models where OH radical reactions aﬀect

autoignition behavior of these fuels. The present analysis has

also highlighted some limitations in the use of currently

39

available structure−activity relationships (SARs) for the

butanol reactions, particularly at temperatures <400 K. Finally,

S

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

the results of this study indicate that OH radical regeneration

occurs in the OH + n-butanol reaction at low temperatures,

Butanol and 3-Methyl-2-Butanol Between 241 and 373 K. Phys. Chem.

Chem. Phys. 2004, 6, 2951−2955.

(

́

11) Teton, S.; Mellouki, A.; Le Bras, G.; Sidebottom, H. Rate

<

380 K, which indicates a generally higher level of reactivity of

Constants for Reactions of OH Radicals with a Series of Asymmetrical

n-butanol.

Ethers and tert-Butyl Alcohol. Int. J. Chem. Kinet. 1996, 28, 291−297.

(12) Wallington, T. J.; Dagaut, P.; Liu, R.; Kurylo, M. J. Gas-Phase

ASSOCIATED CONTENT

* Supporting Information

■

Reactions of Hydroxyl Radicals with the Fuel Additives Methyl and

tert-Butyl Ether and tert-Butyl Alcohol Over the Temperature Range

S

Infrared band strength data for the butanol isomers measured

in this work; graphical summaries of the site-speciﬁc branching

ratios for the OH + butanol reactions determined in this work.

This material is available free of charge via the Internet at

http://pubs.acs.org.

2

(

40−440 K. Environ. Sci. Technol. 1988, 22, 842−844.

13) Droege, A. T.; Tully, F. P. Hydrogen-Atom Abstraction from

Alkanes by OH. 5. n-Butane. J. Phys. Chem. 1986, 90, 5937−5941.

(14) Greiner, N. R. Hydroxyl Radical Kinetics by Kinetic Spectros-

copy. VI. Reactions with Alkanes in the Range 300−500 K. J. Chem.

Phys. 1970, 53, 1070−1076.

(15) Tully, F. P.; Goldsmith, J. E. M.; Droege, A. T. Hydrogen-Atom

Abstraction from Alkanes by OH. 4. Isobutane. J. Phys. Chem. 1986,

AUTHOR INFORMATION

Corresponding Author

E-mail: James.B.Burkholder@noaa.gov.

■

9

(

0, 5932−5937.

*

16) Tully, F. P.; Koszykowski, M. L.; Binkley, J. S. Hydrogen-Atom

Present Address

Abstraction from Alkanes by OH. I. Neopentane and Neooctane. 20th

Symp. (Int.) Combust./Combus. Inst. 1984, 715−721.

§

Department of Soil, Water, and Climate, University of

Minnesota, St. Paul, Minnesota 55108-6028, United States.

(17) Dunlop, J. R.; Tully, F. P. Catalytic Dehydration of Alcohols by

Notes

OH. 2-Propanol: An Intermediate Case. J. Phys. Chem. 1993, 97,

457−6464.

18) Tully, F. P. Catalytic Dehydration of Alcohols by OH.

H C) CCH OH: A Limiting Case. 23rd Symp. (Int.) Combust./

6

(

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

This work was supported in part by NOAA’s Health of the

Atmosphere program.

3

2

■

Combust. Inst. 1990, 147−153.

(19) Carr, S. A.; Blitz, M. A.; Seakins, P. W. Site-Specific Rate

Coefficients for Reaction of OH with Ethanol from 298 to 900 K. J.

Phys. Chem. A 2011, 115, 3335−3345.

REFERENCES

■

(20) Papadimitriou, V. C.; Talukdar, R. K.; Portmann, R. W.;

Ravishankara, A. R.; Burkholder, J. B. CF CFCH and (Z)-

(

1) Derwent, R. G.; Jenkin, M. E.; Saunders, S. M. Photochemical

Ozone Creation Potentials for a Large Number of Reactive

3

2

CF CFCHF: Temperature Dependent OH Rate Coefficients and

3

Hydrocarbons Under European Conditions. Atmos. Environ. 1996,

Global Warming Potentials. Phys. Chem. Chem. Phys. 2008, 10, 808−

3

(

0, 181−199.

8

20.

2) Fortin, T. J.; Howard, B. J.; Parrish, D. D.; Goldan, P. D.; Kuster,

(21) Vaghjiani, G. L.; Ravishankara, A. R. Kinetics and Mechanism of

W. C.; Atlas, E. L.; Harley, R. A. Temporal Changes in U.S. Benzene

Emissions Inferred from Atmospheric Measurements. Environ. Sci.

Technol. 2005, 39, 1403−1408.

OH Reaction with CH OOH. J. Phys. Chem. 1989, 93, 1948−1959.

3

(22) Baasandorj, M.; Papanastasiou, D. K.; Talukdar, R. K.; Hasson,

A. S.; Burkholder, J. B. (CH ) COOH (tert-Butyl Hydroperoxide):

3

(

̈

3) Konig, G.; Brunda, M.; Puxbaum, H.; Hewitt, C. N.; Duckham, S.

OH Reaction Rate Coefficients Between 206 and 375 K and the OH

Photolysis Quantum Yield at 248 nm. Phys. Chem. Chem. Phys. 2010,

C.; Rudolph, J. Relative Contribution of Oxygenated Hydrocarbons to

the Total Biogenic VOC Emissions of Selected Mid-European

Agricultural and Natural Plant Species. Atmos. Environ. 1995, 29,

1

(

2, 12101−12111.

23) Sander, S. P.; et al. Chemical Kinetics and Photochemical Data

8

(

61−874.

for Use in Atmospheric Studies, Evaluation Number 17, JPL

Publication 10−6, Jet Propulsion Laboratory; California Institute of

Technology: Pasadena, CA, 2011, http://jpldataeval.jpl.nasa.gov.

4) Goldan, P. D.; Kuster, W. C.; Fehsenfeld, F. C.; Montzka, S. A.

The Observation of a C Alcohol Emission in a North-American Pine

5

Forest. Geophys. Res. Lett. 1993, 20, 1039−1042.

(24) Veres, P.; Gilman, J. B.; Roberts, J. M.; Kuster, W. C.; Warneke,

(

5) Kelly, T. J.; Callahan, P. J.; Plell, J.; Evans, G. F. Method

C.; Burling, I. R.; de Gouw, J. Development and Validation of a

Portable Gas Phase Standard Generation and Calibration System for

Volatile Organic Compounds. Atmos. Meas. Technol. 2010, 3, 683−

Development and Field-Measurements for Polar Volatile Organic

Compounds In Ambient Air. Environ. Sci. Technol. 1993, 27, 1146−

1153.

6

(

91.

25) “Paciﬁc Northwest National Laboratory (PNNL) Infrared

Database”, 2013, nwir.pnl.gov.

26) Wu, H.; Mu, Y. J.; Zhang, X. S.; Jiang, G. B. Relative Rate

(

6) Pankow, J. F.; Luo, W. T.; Bender, D. A.; Isabelle, L. M.;

Hollingsworth, J. S.; Chen, C.; Asher, W. E.; Zogorski, J. S.

Concentrations and Co-Occurrence Correlations of 88 Volatile

Organic Compounds (VOCs) in the Ambient Air of 13 Semi-Rural

to Urban Locations in the United States. Atmos. Environ. 2003, 37,

(

Constants for the Reactions of Hydroxyl Radicals and Chlorine Atoms

with a Series of Aliphatic Alcohols. Int. J. Chem. Kinet. 2003, 35, 81−

5

(

023−5046.

8

7.

7) Levasseur, A.; Lesage, P.; Margni, M.; Deschenes, L.; Samson, R.

(27) Saunders, S. M.; Baulch, D. L.; Cooke, K. M.; Pilling, M. J.;

Considering Time in LCA: Dynamic LCA and Its Application to

Global Warming Impact Assessments. Environ. Sci. Technol. 2010, 44,

Smurthwaite, P. I. Kinetics and Mechanisms of the Reactions of OH

with Some Oxygenated Compounds of Importance in Tropospheric

Chemistry. Int. J. Chem. Kinet. 1994, 26, 113−130.

3

(

169−3174.

8) Yujing, M.; Mellouki, A. Temperature Dependence for the Rate

(28) Moss, J. T.; Berkowitz, A. M.; Oehlschlaeger, M. A.; Biet, J.;

Constants of the Reaction of OH Radicals with Selected Alcohols.

Warth, V.; Glaude, P.-A.; Battin-Leclerc, F. An Experimental and

Kinetic Modeling Study of the Oxidation of the Four Isomers of

Butanol. J. Phys. Chem. A 2008, 112, 10843−10855.

Chem. Phys. Lett. 2001, 333, 63−68.

(

́ ́

9) Jimenez, E.; Lanza, B.; Garzon, A.; Ballesteros, B.; Albaladejo, J.

Atmospheric Degradation of 2-Butanol, 2-Methyl-2-Butanol, and 2,3-

Dimethyl-2-Butanol: OH Kinetics and UV Absorption Cross Sections.

J. Phys. Chem. A 2005, 109, 10903−10909.

(29) Sarathy, S. M.; Vranckx, S.; Yasunaga, K.; Mehl, M.; Oßwald, P.;

̈

Westbrook, C. K.; Pitz, W. J.; Kohse-Hoinghaus, K.; Fernandes, R. X.;

Curran, H. J. A Comprehensive Chemical Kinetic Combustion Model

(

10) Mellouki, A.; Oussar, F.; Lun, X.; Chakir, A. Kinetics of the

Reactions of the OH Radical with 2-Methyl-1-Propanol, 3-Methyl-1-

for the Four Butanol Isomers. Combust. Flame 2012, 159, 2028−2055.

T

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

(

30) Andersen, V. F.; Wallington, T. J.; Nielsen, O. J. Atmospheric

Chemistry of i-Butanol. J. Phys. Chem. A 2010, 114, 12462−12469.

31) Pang, G. A.; Hanson, R. K.; Golden, D. M.; Bowman, C. T.

Quantum Mechanical Approach. Phys. Chem. Chem. Phys. 2002, 4,

4648−4662.

(

(50) Zhou, C. W.; Simmie, J. M.; Curran, H. J. Rate Constants for

Hydrogen-Abstraction by OH from n-butanol. Combust. Flame 2011,

158, 726−731.

High-Temperature Rate Constant Determination for the Reaction of

OH with iso-Butanol. J. Phys. Chem. A 2012, 116, 4720−4725.

(

32) Chew, A. A.; Atkinson, R. OH Radical Formation Yields from

(51) Seal, P.; Oyedepo, G.; Truhlar, D. G. Kinetics of the Hydrogen

Atom Abstraction Reactions from 1-Butanol by Hydroxyl Radical:

Theory Matches Experiment and More. J. Phys. Chem. 2012, 117,

275−282.

the Gas-Phase Reactions of O with Alkenes and Monoterpenes. J.

3

Geophys. Res. 1996, 101, 28649−28653.

(

33) Baxley, J. S.; Wells, J. R. The Hydroxyl Radical Reaction Rate

Constant and Atmospheric Transformation Products of 2-Butanol and

-Pentanol. Int. J. Chem. Kinet. 1998, 30, 745−752.

34) Pang, G. A.; Hanson, R. K.; Golden, D. M.; Bowman, C. T.

Experimental Determination of the High-Temperature Rate Constant

for the Reaction of OH with sec-Butanol. J. Phys. Chem. 2012, 116,

(52) Kwok, E. S. C.; Atkinson, R. Estimation of Hydroxyl Radical

Reaction Rate Constants for Gas-Phase Organic Compounds Using a

Structure-Reactivity Relationship: An Update. Atmos. Environ. 1995,

29, 1685−1695.

2

(

(53) Zheng, J. J.; Seal, P.; Truhlar, D. G. Role of Conformational

Structures and Torsional Anharmonicity in Controlling Chemical

9

(

607−9613.

35) Atkinson, R. Kinetics of the Gas-Phase Reactions of OH

Radicals with Alkanes and Cycloalkanes. Atmos. Chem. Phys. 2003, 3,

233−2307.

36) Cheema, S. A.; Holbrook, K. A.; Oldershaw, G. A.; Walker, R.

Reaction Rates and Relative Yields: Butanal + HO Reactions. Chem.

2

Sci. 2013, 4, 200−212.

2

(

W. Kinetics and Mechanism Associated with the Reactions of

Hydroxyl Radicals and of Chlorine Atoms with 1-Propanol Under

Near-Tropospheric Conditions Between 273 and 343 K. Int. J. Chem.

Kinet. 2001, 34, 110−121.

(

37) Bott, J. F.; Cohen, N. A Shock-Tube Study of the Reactions of

the Hydroxyl Radical with Several Combustion Species. Int. J. Chem.

Kinet. 1991, 23, 1075−1094.

(

38) Sivaramakrishnan, R.; Su, M. C.; Michael, J. V.; Klippenstein, S.

J.; Harding, L. B.; Ruscic, B. Rate Constants for the Thermal

Decomposition of Ethanol and Its Bimolecular Reactions with OH and

D: Reflected Shock Tube and Theoretical Studies. J. Phys. Chem. A

2

(

010, 114, 9425−9439.

39) Bethel, H. L.; Atkinson, R.; Arey, J. Kinetics and Products of the

Reactions of Selected Diols with the OH Radical. Int. J. Chem. Kinet.

001, 33, 310−316.

40) Ray, D. J. M.; Redfearn, A.; Waddington, D. J. Gas-Phase

2

(

Oxidation of Alkenes: Decomposition of Hydroxy-Substituted Peroxy

Radicals. J. Chem. Soc. Perkin Trans. 2 1973, 5, 540−543.

(

41) Campbell, I. M.; McLaughlin, D. F.; Handy, B. J. Rate Constants

for Reactions of Hydroxyl Radicals with Alcohol Vapours at 292 K.

Chem. Phys. Lett. 1976, 38, 362−364.

(

42) Cavalli, F.; Geiger, H.; Barnes, I.; Becker, K. H. FTIR Kinetic,

Product, and Modeling Study of the OH-Initiated Oxidation of 1-

Butanol in Air. Environ. Sci. Technol. 2002, 36, 1263−1270.

(

43) Hurley, M. D.; Wallington, T. J.; Laursen, L.; Javadi, M. S.;

Nielsen, O. J.; Yamanaka, T.; Kawasaki, M. Atmospheric Chemistry of

n-Butanol: Kinetics, Mechanisms, and Products of Cl Atom and OH

Radical Initiated Oxidation in the Presence and Absence of NO . J.

x

Phys. Chem. A 2009, 113, 7011−7020.

(

44) Nelson, L.; Rattigan, O.; Neavyn, R.; Sidebottom, H.; Treacy, J.;

Nielsen, O. J. Absolute and Relative Rate Constants for the Reactions

of Hydroxyl Radicals and Chlorine Atoms with a Series of Aliphatic

Alcohols and Ethers at 298 K. Int. J. Chem. Kinet. 1990, 22, 1111−

1

(

126.

45) Oh, S.; Andino, J. M. Kinetics of the Gas-Phase Reactions of

Hydroxyl Radicals with C −C Aliphatic Alcohols in the Presence of

1

6

Ammonium Sulfate Aerosols. Int. J. Chem. Kinet. 2001, 33, 422−430.

(

46) Wallington, T. J.; Kurylo, M. J. The Gas-Phase Reactions of

Hydroxyl Radicals with a Series of Aliphatic Alcohols Over the

Temperature Range 240−440 K. Int. J. Chem. Kinet. 1987, 19, 1015−

1023.

(

47) Pang, G. A.; Hanson, R. K.; Golden, D. M.; Bowman, C. T. Rate

Constant Measurements for the Overall Reaction of OH + 1-Butanol

→

Products from 900 to 1200 K. J. Phys. Chem. A 2012, 116, 2475−

2483.

(

48) Vasu, S. S.; Davidson, D. F.; Hanson, R. K.; Golden, D. M.

Measurements of the Reaction of OH with n-Butanol at High-

Temperatures. Chem. Phys. Lett. 2010, 497, 26−29.

(

́

49) Galano, A.; Alvarez-Idaboy, J. R.; Bravo-Perez, G.; Ruiz-Santoyo,

M. E. Gas-Phase Reactions of C −C Alcohols with the OH Radical: A

1

4

U

dx.doi.org/10.1021/jp402702u | J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article Doi

Gas-phase rate coefficients for the OH + n -, i -, s -, and t-butanol reactions measured between 220 and 380 K: Non-arrhenius behavior and site-specific reactivity

DOI: 10.1021/jp402702u

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jp402702u

Full text of DOI:10.1021/jp402702u

Products guided by the article

R&D Labs maybe for 123-38-6

Relevant to this article

Hot Product