Pathways and kinetics of anisole pyrolysis studied by NMR and selective 13C labeling. Heterolytic carbon monoxide generation

Article abstract of DOI:10.1246/bcsj.20110334

By applying ¹³C and ¹HNMR spectroscopy the pyrolysis of site-selectively ¹³C-enriched (H₃¹³CO ₁₂C₆H₅) and normal anisole compounds was studied in the dark at 0.001-1.0M (M, mol dm^-3) and at 400-600°C (supercritical conditions). Conversion of the ¹³C-labeled methyl group was confined to the methoxy-originated fragments, ¹³CO and ¹³CH₄, and the reactive intermediate, H¹³CHO. The normal phenyl group, ¹²C₆H₅-was converted to benzene, ¹²C6H6 and phenol, ¹²C₆H ₅OH without ring disintegration. The pyrolysis consists of two elementary steps: (1) the ratedetermining unimolecular ether-bond fission (k1) to generate the fragmented product C6H6 and energized intermediate H ¹³CHO* through the intramolecular proton transfer from the methoxy group to the phenyl, and (2) the fast bimolecular disproportionation (k2) through the intermolecular proton/hydride transfer from H ¹³CHO* to H₃¹³COC₆H ₅ to produce ¹³CO, ¹³CH₄, and C ₆H₅OH. CO is generation by the heterolytic (ionic) mechanism in contrast to the homolytic (radical) one via the phenoxy radical intermediate (C₆H₅O?) in the literature despite the agreement of the rate constant (k1) and the activation energy.

Full text of DOI:10.1246/bcsj.20110334

124
Bull. Chem. Soc. Jpn. Vol. 85, No. 1, 124-132 (2012)
© 2012 The Chemical Society of Japan
Pathways and Kinetics of Anisole Pyrolysis Studied by NMR and
13
Selective C Labeling. Heterolytic Carbon Monoxide Generation
³
Yasuo Tsujino, Yoshiro Yasaka, Nobuyuki Matubayasi, and Masaru Nakahara*
Institute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011
Received October 31, 2011; E-mail: nakahara@scl.kyoto-u.ac.jp
By applying 13C and 1H NMR spectroscopy the pyrolysis of site-selectively 13C-enriched (H313CO12C6H5) and
normal anisole compounds was studied in the dark at 0.001-1.0 M (M, mol dm¹3) and at 400-600 °C (supercritical
conditions). Conversion of the 13C-labeled methyl group was confined to the methoxy-originated fragments, 13CO and
13CH4, and the reactive intermediate, H13CHO*. The normal phenyl group, 12C6H5- was converted to benzene, 12C6H6
and phenol, 12C6H5OH without ring disintegration. The pyrolysis consists of two elementary steps: (1) the rate-
determining unimolecular ether-bond fission (k1) to generate the fragmented product C6H6 and energized intermediate
H13CHO* through the intramolecular proton transfer from the methoxy group to the phenyl, and (2) the fast bimolecular
disproportionation (k2) through the intermolecular proton/hydride transfer from H13CHO* to H313COC6H5 to produce
13CO, 13CH4, and C6H5OH. CO is generation by the heterolytic (ionic) mechanism in contrast to the homolytic (radical)
one via the phenoxy radical intermediate (C6H5O•) in the literature despite the agreement of the rate constant (k1) and the
activation energy.
It is important to investigate the kinetics and mechanisms of
ethers; This is called “hinge reaction” at high temperature.3,4
When the elementary steps assumed previously (see below)
were the case, CO carbon should come from the methoxy but
not from the phenoxy. To test this, here we have selectively
labeled the asymmetric ether, anisole, by 13C as C6H5-O-13CH3.
In all of the earlier papers,5-12 the following radical CO
pathway was postulated:
the thermal reaction process of ethers in order to understand
more deeply high-temperature reactions and to develop more
efficient high-temperature cracking and processing of fossil (oil
and coal) and biomass fuels. In the previous letter,1 we reported
a kinetic study on the noncatalytic pyrolysis of the asymmetric
ether, anisole (methyl phenyl ether), at 1 M and at 400-430 °C.
By the product analysis with the gas- and liquid-phase NMR
we have succeeded in revealing that the thermal fragmentation
of the ether is induced by not homolytic (radical) but
heterolytic (ionic) C-H bond fission, and that the reaction
consists of the slow and the fast elementary steps: (1) The slow
step is the unimolecular C-H bond fission that is initiated
by the intramolecular proton transfer from the methyl to the
phenyl group to generate the reactive intermediate formalde-
hyde (HCHO*). (2) The fast step is composed of the successive
intermolecular proton and hydride transfers over the electro-
negative oxygen atom from the thermally excited intermediate
HCHO*, respectively, to the phenoxy and methyl groups of the
parent anisole molecule. They are expressed as:
C6H5OCH3 ! C6H5O• þ CH3•
C6H5O• ! C5H5• þ CO
ð3Þ
ð4Þ
The homolytic bond breakage is assumed to take place between
the ether oxygen (eq 3) and the methyl carbon to generate the
methyl and phenoxy radicals, and furthermore, the ring-size
reduction is presumed to give rise to CO as in eq 4. The thermal
fragmentation of dimethyl ether, which is regarded as the
prototype of the “unimolecular reaction,” has been believed to
have the radical mechanism since the pioneering work by
Hinshelwood and co-workers in the 1920s.13 The methoxy
group is a key characteristic of these symmetric and asymmetric
ethers, and the reactive intermediate formaldehyde* is expected
to be generated for both cases. The asymmetric ether can be a
promising candidate to distinguish the reaction pathways for the
generated hydrocarbon fragments: benzene and methane for
anisole and only methane twins for dimethyl ether. Convincing
evidence is wanted to resolve the discrepancy in the mechanism
of the fundamental thermal reaction. To this end we have
scrutinized the CO generation pathway by applying high-
resolution NMR spectroscopy to the labeled anisole studied in
the dark to carefully avoid the interference by photochemical
radical processes. In this work the quantitative overall product
analysis with the mass balance strictly confirmed by NMR is
carried out to demonstrate that the reaction mechanism is cor-
rectly represented by the elementary steps given by eqs 1 and 2.
C6H5OCH3 ! C6H6 þ HCHOꢀ ðslowÞ
ð1Þ
C6H5OCH3 þ HCHOꢀ
! C6H5OH þ CH4 þ CO ðfastÞ
ð2Þ
As represented here, the major products of the high-pressure
(concentration) pyrolysis are benzene, phenol, methane, and
carbon monoxide, all in almost equal amounts. The intra-
molecular proton transfer induced by the C-O-C bending mode
was also observed for acetaldehyde,2 and dimethyl and diethyl
³ Present address: Institute for Computational Molecular Science,
Temple University, Philadelphia, Pennsylvania 19122-6078,
USA
Published on the web December 30, 2011; doi:10.1246/bcsj.20110334

Y. Tsujino et al.
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
Results and Discussion
125
Our1 and the other6-12 investigations focus on the pyrolysis,
respectively, at high (1.0 M) and low (0.1-10 mM) concen-
trations (pressures). In these works some products are common
but others are different probably due to differences in method
for product analysis. Here we show that the mechanistic
discrepancy does not come from differences in the reaction
conditions, such as the concentration and the temperature. We
have scrutinized in detail the rate laws and kinetics for
supercritical anisole reaction by covering wide enough ranges
of concentrations and temperatures. Analytical connectivity of
the kinetics between the high- and low-concentration limits
has been also demonstrated. The anisole depletion rate
constants in this work and those in the literature6-9 are in good
agreement, showing the same Arrhenius relation over the wide
temperature range of 400-700 °C. Thus the unimolecular bond
fission mechanism is the same at the low and high pressures,
see eq 1.
We elucidate reaction pathways and mechanisms by mon-
itoring all of the intermediate and final products of 13C-labeled
and normal anisole reactants. High-resolution 1H and 13C NMR
spectra taken in both the gas and the liquid phases are used
extensively for the elemental and structural analyses under the
H and C mass balances. To confirm the newly found heterolytic
proton-transferred mechanism we carry out a comprehensive
kinetic analysis. We show that the nonradical mechanism
operates not only at high but also at low concentrations over a
wide temperature range.
Product Distribution and Reaction Pathways. Figure 1
shows the H and 13C NMR spectra taken at room temperature
1
for the study on the thermal fragmentation of anisole in the
supercritical condition at 500 °C.16 In panel a, there is given the
13C spectrum of the reactant C6H5-O-13CH3. Only one strong
peak is observed in the high magnetic field. This is assigned to
the enriched methyl carbon; in fact, the signal is a quartet due
to the spin coupling of the 13C with the three methyl 1H’s. Only
when the intensity axis is expanded (not shown here) the
naturally abundant phenyl carbons are observable reflecting the
low ratio of 13C (μ1%); the relative intensity of each phenyl
carbon was determined to be (0.98 « 0.02)% according to
13C NMR analysis in high precision. Correspondingly the
isotope side-bands were observed as weak ones in the 1H
spectra, which are rather complicated by the fine structures of
various ring-proton signals. The 13C enrichment is so selective
and extensive (>99%) that we can unambiguously discuss the
CO reaction pathway.
First we discuss which products are in the liquid phase in the
NMR measurement of the reaction system quenched to room
temperature. Of the four major products, benzene, phenol,
methane, and carbon monoxide, the ring compounds, benzene
and phenol, exist in the liquid phase at room temperature as
well as the reactant anisole. These ring compounds collectively
serve as liquid “solvent” in the sealed quartz tube reactor of
150-mm height for the other gaseous products, CO and CH4. In
the reaction system these small nonpolar molecules quenched
are partitioned between the gas and the liquid phases depending
on the solubility or Henry’s constant. The CO partition ratio of
the gas to the liquid phases is high as mentioned below in the
closed system where the volume ratio of the gas to the liquid
phases is μ13/2; the liquid volume slightly increases with the
fragmentation reaction time at most by 10%. To scrutinize the
CO pathway in a reliable manner we should pay attention to the
spectrum in the gas phase rather than that in the liquid phase.
In panel b we see the methyl carbon destinations at the
reaction time of 20 min where the pyrolysis is almost
completed (yield > 95%). It turns out that the end products
are 13CO and 13CH4. The former and the latter are given,
respectively, by the oxidation and the reduction of the methoxy
group. We can conclude that the CO comes from the methoxy
in harmony with the scheme given by eqs 1 and 2. The
prevailing phenoxy radical mechanism given by eqs 3 and 4 is
not supported by the present investigation.
In the following section, the experimental features are
described. In the third section, the NMR spectroscopic results
are analyzed quantitatively as much as possible to discuss
the pyrolysis mechanism. Conclusions are given in the last
section.
Experimental
The labeled anisole (H313COC6H5, purity >99%)
was synthesized from phenol (Nacalai; purity >99%) and
13C-enriched iodomethane (ISOTEC; purity >99%) by the
Williamson synthesis.14 The deuterated solvents, benzene-d6
and toluene-d5 were supplied by ISOTEC. Anisole was loaded
into a quartz tube of 1.5-mm i.d. and 3.0-mm o.d, and 150-mm
height to set the desired concentration (1.0 M) in the homo-
geneous gas phase at 400-600 °C; for example, the liquid
reactant was 20 mm high at room temperature for the super-
critical reaction of anisole at 1.0 M, and the pressure of the
sample gasified at 500 °C corresponds to 6.3 MPa according to
the ideal-gas law before pyrolysis. For the reaction at a low
concentration of 1.0 mM a high-precision pyrex NMR tube
(Shigemi Co., Ltd.) with a diameter of 5 or 10 mm was
employed. Oxygen dissolved and/or adsorbed in the sealed
sample was completely removed by argon gas. In order to
avoid photochemical processes, the sample was put into a steel
pipe kept in advance at a desired temperature in a program-
mable electric furnace; the temperature was controlled within
«1 °C. Without opening the sample tube, the quenched gas and
liquid phases were subjected to quantitative 13C and H NMR
1
measurements at room temperature using ECA400 (JEOL) and
600-MHz spectrometer (ECA, JEOL). 13C NMR signals in the
gas phase were collected with sufficient delay times without
proton irradiation in order to quantify the concentrations of
the reactant and products. At high temperatures (>500 °C) a
number of reactor tubes were simultaneously put into the
furnace and taken out as a function of reaction time. To make a
correction for the effect of the dead time required for the
heating of the tube the early time portion (μ1 min) was cut
from all of the kinetic data.15 The sample setup and relevant
techniques for NMR measurements in the liquid and the gas
phases have been described elsewhere.1-4 Yields for the
reactant and products were obtained from the peak intensity
within an error of a few percent.
The peak intensity of 13CO in panel b is slightly larger than
that of 13CH4 because the solubility of 13CO is lower than
that of 13CH4 in the liquid “solvent;” see the 13C spectrum in

126
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
Heterolytic Anisole Pyrolysis Studied by NMR
13C NMR, 0 min, liquid
panel c. When the signal intensities of the species in panel b
are compared with those in panel c, the partition mole ratios
are determined to be 20(gas)/1(liquid) for 13CO and 10(gas)/
1(liquid) for 13CH4 in the present volume ratio of the gas to the
liquid phase (13/2). In panel c the 13C signal intensities of
benzene and phenol are quite low as a consequence of
the natural abundance (0.98% « 0.02% by NMR), and the
intensities are coincidentally comparable with those of spar-
ingly soluble 13CO or 13CH4. In any event, the ring carbons of
benzene and phenol keep the natural abundance of 13C without
any scrambling by the enriched methyl carbon (13C) in anisole.
Thus the phenyl ring of anisole is not disintegrated during the
reaction course, in disagreement with eq 4.
(a)
C6H5O13CH3
200
150
100
50
0
13C NMR, 20 min, gas
(b)
13CO
1.0
13CH4
0.9
200
150
100
50
0
13C NMR, 20 min, liquid
The heterolytic CO pathway, not from the phenoxy but from
the “methoxy” as described mechanistically in Figure 2, can
be confirmed by using the H spectra in panels d and e at
(c)
13CH4
C6H513CH3
1
C6H5O13CH3
intermediate and almost final reaction times, respectively. The
ring products, benzene and phenol, are “equally” generated as
in the case of the methyl-originated products, methane and
13CH3OH
13CO
1
OH
4
2.5
C6H6
2
3
1.0
1
1
carbon monoxide; note the integrated H signal intensity ratio
2
2
of benzene to phenol that is not 1 but 6/5 = 1.2 as shown in
the expanded chart between 7.00 to 7.50 ppm in panel e, just
corresponding to the ratio of the ring protons. The essential
counterpart of the CO pathway from the methoxy is the
retention of the phenyl ring during the proton-transferred ether
bond fission. In fact, the total number of ring protons
contributed from benzene (eq 1) and phenol (eq 2) does not
decrease but increases as shown by the time-dependent 1H mass
balance in Figure 3. The increase is the evidence for the proton
transfer from the methyl to the phenyl to produce benzene
through eq 1, which results in the increase in the number of the
3
1
4
200
150
100
50
0
1H NMR, 4 min, liquid
(d)
C6H5O13CH3
H213CO
C6H6
13CH3OH
13CH4
10
9.5
C6H513CH3
1
1
ring H’s from 5 to 6 with the total H mass balanced before
and after the reaction (Figure 3). After the proton transfer, there
is generated 13C-labeled formaldehyde whose proton is actually
observed as the doublet peak (see panel d). Energized formal-
dehyde*, which is conjectured to be vibronically excited by the
excess energy generated by the bond dissociation, is originated
as a reactive intermediate from the enriched methoxy through
the intramolecular proton transfer.17 The thermal decomposi-
tion of anisole consists of the bond breakage induced by the
intramolecular proton-transfer step (eq 1) and the intermolec-
ular proton and hydride transfer step (eq 2) as pictorially
summarized in Figure 2.
Intra- and intermolecular proton-transfer mechanisms pro-
posed above are justified in view of the charge distribution in
anisole. The phenyl ring and the oxygen atom in anisole have a
negative partial charge, respectively, due to the ³ and the lone-
pair electrons, and the hydrogen atom of the methyl group has a
weakly positive charge (μ0.1e). Hence the methyl hydrogen
can slide on the oxygen atom surface as a proton from the
methyl site to the phenyl surface without leaving the molecular
surface. The proton migration can be driven by the thermal
excitation of the internal vibration and/or rotation modes. In
eq 2 the proton migration is followed by the hydride transfer
from the state negatively charged by the proton leaving as the
first step (Figure 2).
10
7.5
5.0
2.5
0
1H NMR, 20 min, liquid
(e)
1.2
1.0
C6H6
C6H5OH
C6H513CH3
ortho
para
meta
7.75
7.50
7.25
7.00
13CH3OH
13CH4
C6H513CH3
10
7.5
5.0
2.5
0
Chemical Shift / ppm
Figure 1. 13C and 1H NMR spectra for the reactant and
products of labeled anisole treated at 500 °C. (a) 13C
spectrum before reaction, (b) gas 13C spectrum after
20 min, (c) liquid 13C spectrum after 20 min, (d) liquid
1H spectrum after 4 min, and (e) liquid H spectrum after
20 min. In panels (b) and (c), the numerical value above the
curve indicates the integrated intensity for each peak. The
weak peak at 3.2 ppm was assigned to water.
1
Mass Balances and Mechanisms. Figure 3 shows the
plots of mass balances for the elements 1H and 13C and
conversion fractions of the four major products and the reactive

Y. Tsujino et al.
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
127
H+
Proton Transfer
H
O
O
slow
+ H13CHO*
(eq 1)
13CH2
H+
Proton Transfer
OH
+
fast
13CH4 + 13CO
(eq 2)
13CH3 + 13CO
H-
Hydride Transfer
Figure 2. Reaction scheme for anisole pyrolysis.
CH3OH
CO
H2O
C6H5CH3
1.0
0.8
0.6
0.4
0.2
1.0
0.8
0.6
0.4
0.2
0.0
(eq 6)
(eq 7)
HCHO
C6H6
Ring Proton Mass Balance
Proton Mass Balance
Carbon Mass Balance
C6H5OCH3
(eq 1)
CO
(eq 2)
CH4
C6H5OH
Benzene
Phenol
Methane
Figure 4. Reaction pathways of anisole pyrolysis. Solid-
line arrows represent main pathways and broken lines the
minor. Main products are underlined. For the explanation
see the text.
Carbon Monoxide
Formaldehyde (1H NMR)
Formaldehyde (13C NMR)
reactions generating toluene and methanol. As seen in Figure 3,
each conversion fraction is slightly less than the ideal value
(1/2), as a result of the conversion of the reactive intermediate,
formaldehyde*, into the other by-products in addition to one
of the major products, phenol. These reaction pathways for
the prototypical pyrolysis are summarized in Figure 4.18 Let us
discuss slight deviations from the above-mentioned reference
value (1/2) in terms of the side-reactions. Toluene can be
formed by
0.0
0
5
10
15
20
Time / min
Figure 3. Time dependencies of conversion fractions for
the four major products and formaldehyde intermediate,
and those of the H and 13C total mass balances together
1
1
with the ring H mass balance at 500 °C.
12C6H5O13CH3 þ H13CHOꢀ
intermediate against the reaction time. The conversion fraction
f is defined by the product yield (¦[product]) normalized by
the amount of anisole (¹2¦[anisole]) consumed by the two
pyrolysis steps given by eqs 1 and 2; f = ¹¦[product]/
2¦[anisole]. The comprehensive analysis of the products in
both the gas and liquid phases enables us to examine in
!
!
12C6H513CH3 þ H13COOH
12C6H513CH3 þ 13CO þ H2O
ð6Þ
where the migration of the methyl cation (not radical) is
assumed. Methanol can be produced by the following self-
disproportionation reaction15 of formaldehyde* via the proton
and hydride transfers:
more detail reaction pathways; in Figure 3 note the 1H and 13
C
mass balances maintained in a high degree at each reaction
time. An ideal reference value of f for the major products is
obtained by adding eqs 1 and 2 to have the following overall
reaction:
2H13CHOꢀ ! 13CH3OH þ 13CO
ð7Þ
The conversion fraction of the intermediate formaldehyde* is
consumed more by these side-reactions. The conversion
fractions (parenthesized numbers) of the major products
(underlined below), the by-products, and the intermediate at
1 min (early reaction stage) are in the following sequence:
212C6H5O13CH3
!
12C6H6 þ 12C6H5OH þ 13CH4 þ 13CO
ð5Þ
Here the intermediate and by-products are dropped out, and the
bimolecular conversion (stoichiometric) fraction defined above
takes the ideal value, 1/2.
The deviation of the conversion fraction from the ideal is
a measure for the competition extent (weight) of the side-
benzene ð0:485Þ > CO ð0:45Þ > methane ð0:44Þ
ꢁ phenol ð0:43Þ > formaldehydeꢀ ð0:04Þ > toluene ð0:015Þ
ꢁ H2O ð0:015Þ > methanol ð0:005Þ
ð8Þ

128
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
Heterolytic Anisole Pyrolysis Studied by NMR
Now we can confirm the competition of the side-reactions
consuming the intermediate formaldehyde* as shown in
Figure 4 by checking the following mass balance:
on the basis of the elemental and structural analysis by NMR
(Figure 3). In contrast, the necessary checking of the mass
balance was not done in previous reports.5-12 Reaction
mechanism is subject to changes depending on new informa-
tion on products and intermediate so that the kinetic self-
consistency can be maintained. It is indispensable to show that
the new proton-transferred mechanism is not specific to the
high concentration of 1.0 M at 500 °C. In the following section
we proceed to a detailed kinetic analysis of the time evolution
of the chemical species involved in order to scrutinize whether
a new reaction channel is opened or not in low concentrations
so far studied in the literature.
fðbenzeneÞ ¼ fðformaldehydeꢀÞ þ ½fðphenolÞ
þ fðtolueneÞ þ 2fðmethanolÞꢂ
ð9Þ
On the right-hand side of eq 9 the 1st term is the remained
intermediate and the bracketed terms are corresponding to the
consumed one (Figure 4). The sum value for the right-hand
side at the early reaction stage is equal to 0.489, which is very
close to the conversion fraction for benzene (0.485). A small
amount of deviation (0.011) of the conversion fraction of
benzene from the ideal is approximately equal to the value for
toluene (0.015), which corresponds to the branched conversion
of anisole to toluene as can be seen from the reaction pathways
summarized in Figure 4. At the almost final reaction stage
(20 min) the sum value for the right-hand side of eq 9 is equal
to 0.47, which is very close to the conversion fraction for
benzene (0.46). The slightly increased deviation (0.03) of the
conversion fraction of benzene from the ideal is equal to the
value for toluene (0.03) as in the case of the early reaction
stage. For the same reason the conversion fraction for benzene
is the largest as seen in Figure 3.19 The product distribution is
neither broadened nor flattened at all, in contrast to the radical
mechanism where the branching and propagation take place
after a long reaction run. Thus the quantitative overall product
analysis based on the NMR elemental and structural analysis is
in harmony with the reaction pathways established in Figure 4.
Now let us compare the products identified here with those
in the literature.5-12 The four major products we identified and
quantified are ring-containing phenol and benzene, and the
counter products, carbon monoxide and methane. A full set
of these products were reported along with some other ring
compounds in several papers.1,7,9,11 The presence of phenol as
one of the ring products, which is a clear indication of the
retention of the C-O bond between the phenyl carbon and the
oxygen, was reported in all of the papers1,6-12 except for an MS
(mass spectroscopy) study of anisole pyrolysis.5 This earliest
work at a very high temperature of 950 °C referred to the
possibility of formation and decomposition of phenoxy to
carbon monoxide. The key product, carbon monoxide was
actually reported in all of the papers1,5-12 at the lower
temperatures of 400-600 °C.1,6-9,12 Methane was found in most
of the papers.1,6,7,9,11,12 In several papers,1,7,9-11 benzene was
found as the ring product, always accompanied by phenol and
carbon monoxide. All of the information on the product
distribution implies the phenyl ring retention and is not
contradictory to the bimolecular proton-transfer step (eq 2)
followed by the unimolecular proton-transfer step (eq 1). Thus
the present results on the products qualitatively agree with
those previously reported despite the difference in the proposed
mechanism.
Kinetic Analysis. The reaction pathways in Figure 4 lead
us to the rate laws expressed by the following differential
equation:
d½Aꢂ
¼ ꢃk1½Aꢂ ꢃ k2½Aꢂ½Fꢀꢂ
ð10Þ
dt
where t is the time, k1 and k2 are the rate constants for the
elementary steps given by eqs 1 and 2, respectively, and the
symbols [A] and [F*] denote the concentrations of the reactant
anisole and intermediate formaldehyde*, respectively. The first
term on the right-hand side of eq 10 presents the unimolecular
step induced by the intramolecular proton transfer from the
methyl group to generate benzene and active intermediate
formaldehyde. The second term presents the successive
bimolecular step induced by the intermolecular proton transfer
from the intermediate F* to the reactant A to give such products
as phenol, methane, and carbon monoxide.
Given a functional relationship between [F*] and [A], eq 10
can be integrated. As already noted in Figure 3 [F*] is not
constant but somehow related to [A], finally dropping to zero.
In Figure 5 it is shown how [F*] depends on [A] at 400 and
500 °C in the neat and carrier or solvent gases added in excess.
Benzene is one of the main products and toluene is alien. The
ring protons of benzene and toluene are fully deuterated in
order to distinguish products of anisole from those of the
supercritical solvation shell.20,21 It turns out that [F*] is linearly
related to [A] at both temperatures, and that the linear region is
separated into the high (HC) and the low (LC) concentration
(pressure) regions; the crossover occurs, respectively, at [A] =
ca. 650 and ca. 250 mM at 400 and 500 °C in the supercritical
conditions.16 As the temperature rises [F*] significantly
increases as a result of the acceleration of the slow unim-
olecular ether-bond fission on the phenyl side (eq 1), and
in consequence, the crossover concentration decreases with
increasing temperature. Although supercritical benzene-d6 has
a negligibly small effect, supercritical toluene-d5, which is
composed of the phenyl and methyl moieties common to
anisole, lowers the crossover point.22 The relation in the HC
region is expressed as:
HC
HC
HC
½Fꢀꢂ ¼ ¡HC½Aꢂ þ ¢
ð11Þ
Such traditional methods as GC (gas chromatography) and
MS for the product analysis lack the power to differentiate
molecular structures compared with NMR; GC relies upon the
retention-time identity, and MS invasively destroys the mo-
lecular structure because of the use of photons or electrons to
fragment original molecules into ionized species. One of the
major points in this work is to discern the reaction mechanism
where ¡HC and ¢HC are the constant parameters. In the LC
region the relation is just proportional with a much larger slope:
LC
½Fꢀꢂ ¼ ¡LC½Aꢂ
LC
ð12Þ
The linear parameters used for the integration are summarized
in Table 1.

Y. Tsujino et al.
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
129
Given the linear relation between [F*] and [A], we
can reduce the number of the variables in eq 10 and the
general solution is obtained separately in the HC and LC
regions:
intercept, ¢HCA, may be taken. The rate of the reactant
depletion due to the steady-state approximation is combined to
the present linear-relation approach. For the intermediate F* the
rate law is:
d½Fꢀꢂ
d½Aꢂ
2
¼ k1½Aꢂ ꢃ k2½Aꢂ½Fꢀꢂ
ð16Þ
¼ ꢃðk1 þ ¢k2Þ½Aꢂ ꢃ ¡k2½Aꢂ
ð13Þ
dt
dt
ðk1 þ ¢HCk2Þ½Aꢂ0 exp½ꢃðk1 þ ¢HCk2Þtꢂ
HC
In the steady-state approximation d[F*]/dt = 0 and [F*] =
½Aꢂ
¼
k1 þ ¢HCk2 þ ¡HCk2½Aꢂ0ð1 ꢃ exp½ꢃðk1 þ ¢HCk2ÞtꢂÞ
HCA
¢
are assumed. Hence it follows that
ðHCÞ ð14Þ
ð15Þ
k2 ¼ k1=¢HCA ðHCÞ
By inserting eq 17 into eq 10 we have the limiting expression
for [A]HCSS under the steady-state condition:
ð17Þ
k1½Aꢂ0 expðꢃk1tÞ
LC
½Aꢂ
¼
ðLCÞ
k1 þ ¡LCk2½Aꢂ0½1 ꢃ expðꢃk1tÞꢂ
The conventional steady-state approximation corresponds
to the limiting case, ¡HC = 0 with ¢ unknown. When an
intermediate is observed as in the present case an average
HCSS
½Aꢂ
¼ ½Aꢂ0 expðꢃ2k1tÞ ðHCSSÞ
ð18Þ
The apparent depletion constant in the HCSS condition is
double the rate constant for the slow unimolecular step (eq 1)
due to the assumption for the extremely rapid bimolecular step
(eq 2). When the intermediate is accumulated up to the critical
concentration ¢HCA the reactant begins to be quickly con-
sumed, and the depletion rate appears to be in the 1st-order
with the rate constant two times faster than that of the
intrinsically unimolecular step (the first term on the right-hand
side of eq 10). The HCSS values of k1 and k2 can be obtained
respectively from eqs 18 and 17. Surprisingly, the difference
in the determined rate constant k1 between eqs 14 and 18 is
confirmed to be negligibly small simply because of the small
slope of eq 11.
25
20
Neat (500 °C)
Neat (13C)(500 °C)
in Benzene-d6 (500 °C)
in Toluene-d5 (500 °C)
Neat (400 °C)
15
It is to be noticed that the low anisole concentration, [A]LC
,
can be set initially or approached inevitably at the late stage of
the reaction. In the LC limit eq 15 can be switched to:
10
4
LC
½Aꢂ ¼ ½Aꢂ0 expðꢃk1tÞ ðLCÞ
ð19Þ
3
The corresponding rate law is obtained from eq 13 as follows:
the [A]2 term much (exponentially) smaller than the [A] term
is dropped away from the right-hand side where ¢ = 0. Thus
the depletion rate slows down in the LC limit compared to eq 18
because the much faster bimolecular step (k2, eq 2) disappears.
This is ascribed to the dramatic reduction of the encounter
probability (density multiplication) of dilute A and F* to form a
reactive complex between them in the supercritical conditions.21
Probably F* is deactivated too much nearer to the ground state
in a long mean-free-path.22-24 In this LC limit the product
distribution is therefore narrow with benzene and formalde-
hyde* or its transformed species like methanol (eq 7).
2
5
1
0
0
100 200 300 400
[A] / mM
0
0
200 400 600 800 1000 1200 1400
[A] / mM
Figure 5. Relation between [F*] and [A] in supercritical
anisole, anisole/1.0-M benzene-d6 and anisole/1.0-M
toluene-d5 (inset) at 400 and 500 °C.
Table 1. Rate Constants, k1 and k2, Determined at 500 °C at High and Low Concentrations together with the
Linear Parameters, ¡ and ¢, in eqs 11 and 12
¹1
¹1
HC
HC
LC
concn Region
103k1/s
k2/s¹1 M
¡
¢
¡
Neat
HC
LC
1.0 mM
HC
LC
1.0 mM
HC
LC
1.9 « 0.2
1.9 « 0.2
2.0 « 0.3
1.8 « 0.3
2.1 « 0.3
1.8 « 0.4
2.0 « 0.4
2.2 « 0.4
1.8 « 0.4
0.12 « 0.02
0.12 « 0.02
0.0040
0.0035
0.0056
17.0
17.5
0.064
0.062
0.0189
Benzene-d6
Toluene-d5
0.11 « 0.03
0.12 « 0.03
0.40 « 0.08
0.70 « 0.08
1.41
1.0 mM

130
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
Heterolytic Anisole Pyrolysis Studied by NMR
1.0
1.0
0.8
0.6
0.4
0.2
0.0
Proton Mass Balance
Benzene
Phenol
Fitting (High Concentration)
Fitting (Low Concentration)
0.8
0.6
0.4
0.2
Methane
Carbon Monoxide
Formaldehyde
Bibenzyl
0
5
10
15
20
Reaction Time / min
0
5
10
15
20
Reaction Time / min
Figure 7. Time dependencies of conversion fractions
f (closed squares) for the four major products and
formaldehyde (closed circle) in the pyrolysis of anisole
([A]0 = 280 mM) in supercritical toluene (1.0 M) at
500 °C; f = ¹¦[product]/2¦[A]. Open squares denote
[product]/[A]0. Solid triangles denote [product]/[toluene]0.
Figure 6. Time dependence of anisole concentration at
500 °C; the initial concentration is 1.0 M. Equations 14 and
15 are respectively fitted to the high-concentration (early-
time) and the low-concentration (late-time) data; see the
text.
Figure 6 shows how the initial depletion of anisole (A) at a
high initial concentration (1.0 M) is expressed by eq 14 as long
as [A] remains larger than the crossover concentration (>ca.
250 mM). When anisole is depleted below this point eq 14
cannot be used to reproduce the kinetic data; see the broken-
curve in Figure 6. The long-time portion of the anisole
depletion data are represented by eq 15 (red curve) but not
by eq 14. The values of k1 and k2 thus determined are
summarized in Table 1. The values obtained in different ways
are in good agreement. Now we are ready to demonstrate the
agreement of the rate constant k1 between the HC and LC limits
in a more direct manner. As compared in Table 1, the k1 value
thus calculated for neat anisole at the initial concentration of
1.0 M is in excellent agreement with that directly obtained
through eq 19 in the LC region (1.0 mM). Thus the generality
of the heterolytic reaction mechanism is confirmed by the self-
consistency of the kinetics. No new channel is opened in the
LC region.
One subject remaining to be scrutinized is the effect of
supercritical solvent or carrier gases on the reaction pathways
and the product distribution. The effect of toluene was
discussed in terms of the radical mechanism in the previous
works.6,9 All of our main products, benzene, phenol, and
methane except for carbon monoxide were also reported in
Ref. 9, whereas only phenol and methane were identified in
Ref. 6. We reinvestigated the effect of toluene in the anisole
(0.001-1.0 M) and toluene (1.0 M) mixtures. Figure 7 displays
the 1H mass balance with respect to the anisole reaction and the
time evolution of all of the species involved in the pyrolysis
of 13C-labeled anisole (H313CO12C6H5; 280 mM, double the
crossover concentration in the HC region) in normal toluene
(H312C12C6H5; 1.0 M) at 500 °C. As shown, in fact, there are
the four main products as well as formaldehyde*. They are all
in almost equal amounts at any time as in the case of neat
anisole (1 M) shown in Figure 3. The characteristic profiles of
Figure 7 are quite common to those of Figure 3. Although the
minor products, toluene (eq 6) and methanol (eq 7), are also
observed, respectively, at μ1 and μ2 mM (similar to the
concentration of formaldehyde*) in toluene, they are omitted to
avoid the overlapping in the plots. The minor products are
generated less than those in the neat reaction because of the
larger value of k2. In addition to 13C-labeled methane (H413C;
1
doublet H signal as seen in Figure 1) that comes from the
enriched solute anisole, normal methane (H412C; singlet 1H
signal) and bibenzyl were observed in almost equal amounts.
They are thought to be formed by the following overall reaction
of supercritical solvent toluene:
3H312C12C6H5
! H412C þ 12C6H6 þ 12C6H512CH2H212C12C6H5 ð20Þ
In any species here is no insertion of 13C isotope. The
identification of these solvent products is also ascertained by
using the site-selectively deuterated anisole and toluene. The
evolution of bibenzyl seems to be almost saturated when
anisole is consumed, which indicates the influence of anisole
on the solvent side-reaction. The reaction mechanism of this
solvent side-reaction is unknown.
When toluene is added in excess [F*] is lower than that in
benzene. As observed thus the toluene system can give the
main products at a lower anisole concentration. These are
reflected in the quantitative values of k1 and k2 summarized in
Table 1. The agreement of the k1 values in the HC and LC
limits are excellent. The k1 value obtained here for anisole at
high and low concentrations in toluene is in harmony with the
literature (interpolated) value, 1.9 © 10¹3 s¹1, reported for
anisole at low concentrations (0.2-0.8 mM) in toluene by Paul

Y. Tsujino et al.
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
131
800 °C
700 °C
600 °C
500 °C
400 °C
product analysis that the four major products, benzene, phenol,
methane, and carbon monoxide as well as the reactive
intermediate formaldehyde* given by these elementary steps.
In addition to the main reaction pathways, the side-reaction
pathways leading to the formation of toluene and methanol by
the consumption of the reactive intermediate formaldehyde*
are disclosed. The conclusion drawn here on the CO genera-
tion pathway and the phenyl ring retention disfavors the
prevailing mechanism that CO is generated through the
phenoxy radical.5-12 The pyrolysis studied is not homolytic
but heterolytic over wide enough ranges of concentrations and
temperatures. The self-consistency of the kinetics has been
firmly confirmed. Further study on the pyrolysis of the methoxy
ether, anisole is in progress by using multinuclear NMR
spectroscopy and site-selectively deuterated anisole isotopes, in
order to obtain more direct evidence for the proton-transfer
mechanism. It is suggested that in situ vibrational spectroscopic
study could provide us with more detailed quantum states
involved in the pyrolysis. What are found here would make a
contribution to advances in science and technology for non-
catalytic thermal cracking of fuel-related compounds at high
temperatures.
5
0
ref 8
-5
ref 7
ref 9
ref 6
This work
-10
0.9
1.0
1.1
1.2
1.3
1.4
1000 (K/T)
Figure 8. Arrhenius plots of rate constants, k1, for the
unimolecular decomposition of anisole; for the k1 defi-
nition see the text. The experimental data shown in Refs. 6,
7, and 9 were graphically read. Their anisole concen-
trations were more or less than 1.0 mM.
This work is supported by the Grants-in-Aid for Scientific
Research (Nos. 18350004 and 21300111) from the Japan
Society for the Promotion of Science and by the Grants-in-Aid
for Scientific Research on Priority Areas (Nos. 15076205
and 20038034), the Grant-in-Aid for Scientific Research on
Innovative Areas (No. 20118002), and the Next-Generation
Integrated Nanoscience Simulation Software Project from the
MEXT. N. M. is also grateful to the CREST project from Japan
Science and Technology Agency (JST) and to the grant from
the Association for the Progress of New Chemistry, the grant
from the Suntory Institute for Bioorganic Research, and the
Supercomputer Laboratory of Institute for Chemical Research,
Kyoto University. M. N. and Y. T. acknowledge the donation
of The Water Chemistry Energy Laboratory (AGC) from Asahi
Glass Co., Ltd.
and Back;6 recall that their products are quite similar to ours as
mentioned above. Even when the alien gas is added in excess
the reaction mechanism does not change from the heterolytic to
the homolytic.
One of the most important remaining problems is the effect
of temperature on the rate constant of k1. The Arrhenius plot of
the unimolecular process is shown in Figure 8 for comparison.
It is clearly shown that all of the rate constants determined by
different groups at different temperatures are all lying in the
single straight line within experimental uncertainties over the
wide range of temperature. The general form of the Arrhenius
plot for the pyrolysis of neat anisole is expressed by
k1 ¼ A expðꢃEa=RTÞ
ð21Þ
References
Here the frequency factor A = (6.8 « 0.5) © 1014 s and the
activation energy Ea is 259 « 8 kJ mol¹1; R is the gas constant
and T the temperature. The Ea value obtained from the wide
temperature range is in reasonable agreement with those
reported in the narrower temperature range.6,7,9 As seen in
Figure 8, however, one paper8 reports a considerably smaller
Ea value.25 Thus the anisole pyrolysis reaction discussed here
is essentially the same as that so far studied.5-12 The reaction
mechanism proposed here is heterolytic for the reason
discussed above.
¹1
1
Y. Tsujino, C. Wakai, N. Matubayasi, M. Nakahara, Chem.
Lett. 2006, 35, 1334.
2
Y. Nagai, S. Morooka, N. Matubayasi, M. Nakahara,
J. Phys. Chem. A 2004, 108, 11635.
3
Y. Nagai, N. Matubayasi, M. Nakahara, J. Phys. Chem. A
2005, 109, 3550, and papers cited therein.
Y. Nagai, N. Matubayasi, M. Nakahara, J. Phys. Chem. A
2005, 109, 3558.
4
5
A. G. Harrison, L. R. Honnen, H. J. Dauben, Jr., F. P.
Lossing, J. Am. Chem. Soc. 1960, 82, 5593.
6
7
S. Paul, M. H. Back, Can. J. Chem. 1975, 53, 3330.
J. C. Mackie, K. R. Doolan, P. F. Nelson, J. Phys. Chem.
Conclusion
1989, 93, 664.
We have focused on the combined application of site-
selective labeling and high-resolution NMR spectroscopy in
the gas and liquid phases in order to establish the reaction
pathways and mechanism of the pyrolysis of the asymmetric
ether, anisole. The reaction is composed of the elementary steps
expressed by eqs 1 and 2. It is confirmed by the comprehensive
8
M. M. Suryan, S. A. Kafafi, S. E. Stein, J. Am. Chem. Soc.
1989, 111, 1423.
9
I. W. C. E. Arends, R. Louw, P. Mulder, J. Phys. Chem.
1993, 97, 7914.
10 M. Pecullan, K. Brezinsky, I. Glassman, J. Phys. Chem. A
1997, 101, 3305.

132
Bull. Chem. Soc. Jpn. Vol. 85, No. 1 (2012)
Heterolytic Anisole Pyrolysis Studied by NMR
11 V. V. Platonov, V. A. Proskuryakov, S. V. Ryl’tsova, Yu. N.
Popova, Russ. J. Appl. Chem. 2001, 74, 1047.
12 A. V. Friderichsen, E.-J. Shin, R. J. Evans, M. R. Nimlos,
D. C. Dayton, G. B. Ellison, Fuel 2001, 80, 1747.
13 C. N. Hinshelwood, P. J. Askey, Proc. R. Soc. London, Ser.
A 1927, 115, 215.
14 M. B. Smith, J. March, Advanced Organic Chemistry:
Reaction, Mechanism, and Structure, 5th ed., John Wiley & Sons,
New York, 2001, p. 477.
bibenzyl is generated not from anisole but from the high-
concentration supercritical solvation shell itself,21 contrary to the
phenoxy-radical mechanism.6,9 Furthermore, when benzene-d6 or
toluene-d5 is added to anisole to change the molecular environment
no propagation products of methyl and hydrogen radicals assumed
in the literature are detected. In other words, no 1H-methyl-
substituted derivatives of the deuterated solvents are detected.
21 Y. Yoshida, N. Matubayasi, M. Nakahara, J. Chem. Phys.
2007, 127, 174509.
15 S. Morooka, N. Matubayasi, M. Nakahara, J. Phys. Chem.
A 2007, 111, 2697.
22 Some complex or cluster formation can be induced by
the attractive polarization effect on the phenyl ³-electrons
provided by the strong dipole of F* and/or by the repulsion effect
by the blocking cage molecules.
16 The critical temperature and density of anisole were
¹3
determined to be 393 « 1 °C and 0.32 « 0.03 g cm
(3.0 «
0.3 M), respectively. The critical temperatures for benzene and
toluene are 289 and 319 °C, respectively. Thus all of the anisole
reactions studied here occurred under supercritical conditions.
17 High-velocity collisions can excite C-H vibrations com-
posed of the very light H atoms. The following kinetic data led us
to assume the vibronically excited formaldehyde* as a reactive
intermediate in eq 1 that is to cross over dissociative curves onto
which the fragmentation reaction manifold is projected. The
reaction rate of anisole depletion at 1.0 M was not affected by
addition of a formaldehyde generator, 1,3,5-trioxane that can be
completely decomposed and decayed into ground-state monomers
at 200-250 °C. The initial depletion rate constants with and
without ground-state formaldehyde at 100 mM in excess at 500 °C
23 It is interesting to compare the proton-transferred reaction
mechanism (heterolytic) proposed here and the traditional unimo-
lecular reaction mechanism (homolytic) given by the Lindemann-
Hinshelwood theory because a work8 heavily relies upon the
RRKM theory:24 A + A ꢀ A + A* and A* ¼ P
for which the rate constants for the activation and the deactivation
of the activated species A* are defined, respectively, by k+a and
k
¹a, and that for the transformation of A* into the product P is
denote by kb. The steady-state approximation leads us to:
d[P]/dt = kb[A*] = kakb[A]2/(kb + k¹a[A]),
which is reduced to:
d[P]/dt = (kakb/k¹a) [A]
for the high-pressure condition, k¹a[A*][A] º kb[A*], and
d[P]/dt = ka [A]2
¹3
are (1.9 « 0.1) © 10 and (1.6 « 0.3) © 10¹3 s¹1, respectively.
These values are almost identical.
for the low-pressure condition, k¹a[A*][A] ¹ kb[A*].
Common to the two mechanisms, the reaction order is controlled
by the magnitude of the concentration [A], leading to the high- and
low-pressure (concentration) limits. The essential difference,
however, exists in the reaction order in the low-concentration
limit. The present and the Lindemann-Hinshelwood mechanisms
respectively give the 1st- and 2nd-order reactions; see eq 19 and
the derivation for the rate law. Also in the high-concentration limit,
our expression by eq 13 or 14 (solution) is more accurate and
useful for the concentration dependence over a wide concentration
range including the high.
18 For the case of the asymmetric ether (anisole) the reaction
pathways are fully separated. For the case of the symmetric ether,
dimethyl ether, however, the products as well as the pathways are
not completely distinguished. Compare the fragmentation products
of anisole and dimethyl ether,4 and we have the following
correspondence: benzene , methane, phenol , methanol, tolu-
ene , ethane. Thus the products overlap significantly in the case
of dimethyl ether whose mechanism has been discussed in the
same way based on the heterolytic mechanism. Thus ethane is not
formed from the combination of methyl radicals.
19 The CO conversion fraction is always larger than that
for phenol because of the contributions from the side-reactions,
eqs 6 and 7 as well as eq 2; see Figure 4. The corresponding
mass balance is: [carbon monoxide](overall) = [phenol](eq 2) +
[toluene](eq 6) + [methanol](eq 7).
20 It is important to show no appearance of products related to
the homolysis-originated methyl radical, that is the counterpart of
phenoxy radical. The 13C, 2H, and 1H NMR spectroscopic study on
the toluene-d5 system clearly indicates that the minor product,
24 J. I. Steinfeld, J. S. Francisco, W. L. Hase, Chemical
Kinetics and Dynamics, 2nd ed., Prentice Hall, New Jersey, 1999,
Chap. 11.
25 Although the rate constants experimentally determined8 at
temperatures higher than 600 °C are smaller than those provided by
the Arrhenius line, those extrapolated with the help of the RRKM
theory to the temperature range of 450-550 °C are said to be in
good agreement in Ref. 9 (Figure 3).