Anisole hydrolysis in high temperature water

Article abstract of DOI:10.1039/c3cp43877e

We investigated the hydrolysis of anisole to phenol in high-temperature water with and without water-tolerant Lewis acid catalysis. With no catalyst present, anisole hydrolyzes to phenol in 97% yield after 24 hours at 365 °C, our experimentally determined optimal temperature and time. Experiments with varied water density and analysis of comparable literature data suggest that anisole hydrolysis is almost third order in water, when the S_N2 mechanism dominates. Of the water-tolerant Lewis acid catalysts studied, In(OTf)₃ offered the best phenol yield. Anisole hydrolysis was first order in catalyst and first order in substrate. Introducing In(OTf)₃ catalysis lowered the activation energy for anisole hydrolysis to 31 ± 1 kcal mol^-1. Anisole hydrolysis in high-temperature water with In(OTf)₃ catalysis is competitive with other techniques in the literature based on rate and yield. In the presence of 5 mol% In(OTf) ₃ catalyst, anisole hydrolyzes to phenol in 97% yield after 90 minutes at 300 °C. This journal is

Full text of DOI:10.1039/c3cp43877e

PCCP
View Article Online
View Journal | View Issue
PAPER
Anisole hydrolysis in high temperature water
Natalie A. Rebacza and Phillip E. Savage*b
Cite this: Phys. Chem. Chem. Phys., 2013,
15, 3562
We investigated the hydrolysis of anisole to phenol in high-temperature water with and without water-
tolerant Lewis acid catalysis. With no catalyst present, anisole hydrolyzes to phenol in 97% yield after
24 hours at 365 1C, our experimentally determined optimal temperature and time. Experiments with
varied water density and analysis of comparable literature data suggest that anisole hydrolysis is almost
third order in water, when the SN2 mechanism dominates. Of the water-tolerant Lewis acid catalysts
studied, In(OTf)3 offered the best phenol yield. Anisole hydrolysis was first order in catalyst and first
order in substrate. Introducing In(OTf)3 catalysis lowered the activation energy for anisole hydrolysis to
31 ꢀ 1 kcal molꢁ1. Anisole hydrolysis in high-temperature water with In(OTf)3 catalysis is competitive
with other techniques in the literature based on rate and yield. In the presence of 5 mol% In(OTf)3
catalyst, anisole hydrolyzes to phenol in 97% yield after 90 minutes at 300 1C.
Received 1st November 2012,
Accepted 22nd January 2013
DOI: 10.1039/c3cp43877e
www.rsc.org/pccp
reaction of base with PhSH.3 Others researched ionic liquids of
1-n-butyl-3-methylimidazolium bromide and p-toluenesulfonic
1 Introduction
The hydrolysis of aryl alkyl ethers is an important transforma-
tion due to its use in deprotection. Protection is used during
multi-step organic syntheses to mask a functional group, such
as an alcohol, from reactions desired elsewhere on the mole-
cule. The formation of a methyl ether in place of a hydroxy
group is one example of protection. Later in the multi-step
synthesis, the protecting ether must be removed to present the
alcohol anew. Most procedures for this transformation call for
the use of chlorinated solvents and strong Lewis acids, such as
BF3 or BCl3. New methods are continuously developed with
improvements in selectivities, yields, reaction times, simplicity,
and/or cost. For example, researchers developed a boron-
trichloride tetra-n-butylammonium iodide system for readily
deprotecting methyl-, ethyl-, and benzylnaphthyl ethers at low
temperatures (ꢁ78 1C warmed to rt) over 1–2 hours.1 Notably,
benzyl ethers are selectively cleaved in the presence of methyl
ethers, and basic functional groups are well tolerated with
additional equivalents of BCl3.1 In another deprotection study,
acid (or some other Brønsted acid) at 115 1C over 12 to 48
hours.4 Research efforts have also focused on merely comparing
the behavior of deprotection reagents. For example, Hwu et al.
compared the behaviour and effectiveness of two different alkali
organoamides (NaN(SiMe3)2 and LiN(i-Pr)2) under similar
conditions.5 Konieczny et al. focused efforts on demonstrating
the extent to which BF3ꢂSMe2 can selectively cleave allyl or methyl
phenyl ethers depending on reaction conditions and the ether’s
position on the ring.6
Research has shown that anisole hydrolyzes in high-
temperature water (HTW) to form phenol without any added
catalyst.7,8 Hydrolysis in HTW was broadly reviewed elsewhere.9
If it could be made practical for deprotection, ether hydrolysis
in HTW could avoid the use of chlorinated solvents and strong
Lewis acids. Together, the work of Klein et al.7 and Patrick
et al.8 show that the reaction in HTW and in supercritical water
(SCW) is most likely SN2, catalyzed by water. Both authors
determine a positive Hammett constant for the reaction,
corroborating their mechanistic assignment. In this paper, we
integrate our own experiments on uncatalyzed anisole
hydrolysis in HTW with those of previous researchers, and
further show the benefits of applying catalytic quantities of
water-tolerant Lewis acids.
lithium–ethylenediamine in
a THF-mediated cold system
successfully deprotected sterically hindered alkyl ethers such
as 2,6-di-tert-butyl-anisole, a feat the boron-trichloride tetra-
n-butylammonium iodide system failed to accomplish.2
However, the reagent was so aggressive that reduction of the
aromatic ring often reduced yields.2 Chakraborti et al. explored
aryl alkyl cleavage by thiolates that were generated in situ by
2 Experimental methods
We carried out reactions with two different mini batch reactors:
quartz capillary and stainless steel Swagelok. The quartz capillary
reactors measure 6 mm outer diameter, 2 mm inner diameter,
a Phillips 66, Hwy 60 & 123, Bartlesville, OK 74003, USA
b University of Michigan, 2300 Hayward St, Ann Arbor, MI 48109, USA.
E-mail: psavage@umich.edu; Fax: +1 734 763 0459; Tel: +1 734 764 3386
c
3562 Phys. Chem. Chem. Phys., 2013, 15, 3562--3569
This journal is the Owner Societies 2013

View Article Online
Paper
PCCP
and had an internal volume of approximately 590 mL. We estimated the desired hydrolysis product, were formed. At 365 1C and
their maximum allowable pressure as 1000 psi. Stainless steel above, side products were formed. Only benzene was observed
Swagelok reactors were constructed from one 1/400 port connector at 365 1C; benzene and benzyl alcohol were detected at 380 1C.
and two 1/400 end caps. The nominal internal diameter is 0.1700. We At 400 1C, toluene and benzaldehyde were observed in addition
estimate the internal volume of the 1/400 Swagelok reactors to be to benzene and benzyl alcohol.
590 mL. Reactors were loaded at room temperature such that the
The data are plotted in Fig. 1 as the temporal variation of the
expansion of water at reaction conditions rendered them 95% full % molar yield of phenol and % recovery of anisole. Fig. 1 also
of liquid at the reaction temperature, when subcritical; the thermo- shows the result of data fitting, which will be explained further.
physical properties of the system were approximated as the thermo- In supercritical water at 380 1C with a water density of
physical properties of water alone. Steam tables were used to 0.4 g mLꢁ1, Fig. 1c, a maximum phenol yield of 88% was
estimate the liquid density at reaction conditions.10 A single liquid achieved after 70 hours. If near-critical water is used instead,
phase is assumed for all temperatures studied because anisole and and the reaction temperature is lowered to 365 1C, 97% yield of
water were found to be completely miscible at 280 1C and above,8 phenol is achieved after 24 hours, Fig. 1b. Lowering the
and because our typical reactor loadings are lower in concentration temperature another 15 1C causes the rate to slow dramatically,
than the solubility of benzene in water,11–13 and anisole is much achieving only 60% yield at 65 hours, Fig. 1a, though the yield is
more hydrophilic than is benzene. Each reactor was first loaded still increasing with time. Increasing the temperature to 400 1C,
with deionized, distilled water and then loaded with the desired Fig. 1d, increased the rate of conversion of anisole, but
amount of anisole, typically 10 mL, before being sealed. The quartz decreased the phenol yield due to thermal degradation reactions.
reactors were sealed over a 1500 K flame; the stainless steel The present results show that uncatalyzed anisole hydrolysis in
Swagelok reactors were capped according to the manufacturer’s HTW at 365 1C gives the highest yield by maximizing hydrolysis
instructions. Reactors were heated in a Techne fluidized sand bath rates while keeping decomposition rates low. At 400 1C, a
with temperature control to ꢀ2 1C. The quartz reactors were cooled significant fraction of the original starting material decomposed
by forced convection with air at ambient conditions, while the or over-oxidized to form benzene, benzaldehyde, benzyl alcohol,
stainless steel reactors were cooled by immersion in a cold water and toluene, in that order of abundance.
bath. After thermal quenching, quartz reactors were scored and
Based on the experimental data, 365 1C achieves the best
snapped open about one to two inches from an end; Swagelok yield in short times. Higher temperatures promote greater
reactors were opened with a wrench. The reactors were then thermal degradation products. Lower temperatures have
unloaded and rinsed with acetone (Aldrich, reagent grade). The significantly slower kinetics. Previously, the HTW literature
reactor contents and rinses were collected in a 10-mL class A on anisole hydrolysis included temperatures of only 300 and
volumetric flask, which was then diluted to volume with acetone.
GC/MS was used to identify the products, and GC-FID was
380 1C, and no search for the optimum temperature.
A model of the reaction kinetics is properly determined
used to quantify the components. Gas chromatography was alongside a reaction pathway. Gas chromatograms from experi-
carried out on a 6890 Agilent system equipped with an HP-5 ments at 350 1C show only recovery of anisole and phenol from
(GC-FID) or HP-5 ms (GC/MS) capillary column (50 m ꢃ 0.2 mm ꢃ reaction samples; no side products are detected. Hence, for this
0.33 mm) and split/splitless injector. Standardization was data set, a simple reaction pathway, Fig. 2, sufficiently describes
achieved for the FID by preparing calibration standards for the kinetic data. The set of ordinary differential equations that
each analyte. For the uncatalyzed reactions, the calibration describes the elementary model of Fig. 2 is given in eqn (1),
standards were prepared using acetone as diluent. The where CA denotes the concentration of anisole; CP, the concen-
presence of water in the reaction samples suppressed the FID tration of phenol; and CW, the concentration of water. The 2.6-order
output relative to the calibration standards. This was corrected dependence upon the concentration of water will be corroborated
for by normalizing the concentration data after calibration. with evidence later in this section.
However, for the catalyzed reactions, calibration standards
were prepared to contain the same proportion of water to
acetone as the reaction samples, eliminating the need for
normalization.
dCA
¼ ꢁkHCW 2:6CA
dt
(1)
dCP
dt
¼ kHCW 2:6CA
3 Anisole hydrolysis without added catalyst
All of the data fitting to kinetic rate models was done using
Reactions were carried out at different batch holding times at Scientistt, created by Micromath Research. Differential
350, 365, 380, and 400 1C. Since the critical temperature of equations were solved by the Runge–Kutta iterative method of
water is 374 1C, the last two temperatures are within the numerical analysis, and least squares was used to estimate
supercritical regime. The 380 1C experiments were carried out parameters by non-linear regression. During parameter estima-
with a water density of 0.40 g mLꢁ1; the 400 1C experiments, tion, the data were weighed with a weight inversely propor-
0.20 g mLꢁ1. For each temperature, one batch holding time was tional to the number of replicates. The inverse proportionality
replicated three times to obtain a sense for the level of to the number of replicates prevented the replicated time point
experimental error. At 350 1C, no products other than phenol, from having an undue influence over the parameter estimation.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 3562--3569 3563

View Article Online
PCCP
Paper
Fig. 1 Hydrolysis of anisole to phenol. Filled circles represent molar yield of phenol; open circles represent % anisole unreacted. Dashed lines represent fit to eqn (2),
with kD equal to zero for reaction at 350 1C, as no decomposition products were detected at this temperature. (a) Top left, 350 1C. (b) Top right, 365 1C. (c) Bottom left,
380 1C. (d) Bottom right, 400 1C.
The data collected for reaction at 380 1C were modeled in the
same way as was the data collected for reaction at 365 1C, with
two changes. One, CD now represents a lumped concentration
of all two decomposition products: benzene and benzyl alcohol.
Fig. 2 A simple pathway describes the kinetic data at 350 1C.
Two, during parameter estimation, we needed to give more
weight to the data at longer batch holding times so that the
model would accurately fit these late data points. The dashed
lines in Fig. 1c show the model fit to the presented data and 380 1C.
The result of fitting the data collected at 350 1C to the model
equations of eqn (1) are shown as dashed lines in Fig. 1a.
Experiments at 365 1C produced benzene as a side product,
presumably via decomposition of anisole. Hence, the data
Finally, the data collected at 400 1C were modeled similarly
as for reaction at 365, except that D now included toluene and
collected for reaction at 365 1C were modeled by the set of
benzaldehyde as decomposition products. Values for kH, the
ordinary differential equations in eqn (2). Term CD signifies the
concentration of the decomposition product, benzene, and kD
signifies the kinetic rate constant for the formation of benzene.
The dashed lines in Fig. 1b show the model fit to the 365 1C
hydrolysis rate constant, are presented in Table 1.
One beautiful aspect of supercritical water is the ability to
change water density and hence water concentration, without
changing temperature. We tested the effect of water density in
the supercritical regime, at 380 1C, calculated the pseudo-first-
data set.
order hydrolysis rate constant kH, and plotted it against water
dCA
¼ ꢁkHCW 2:6CA ꢁ kDCA
concentration on log–log coordinates in Fig. 3. Alongside our
dt
data, we plotted the analogous data from the Klein lab.7 We
dCP
¼ kHCW 2:6CA
used the data fitting software Scientistt to estimate, with
nonlinear regression, the slope of the line formed by our data
set to be 2.5 ꢀ 0.2. Similarly, we estimated the slope formed by
the data set from Klein et al. to be 2.8 ꢀ 0.1. (They reported
(2)
dt
dCD
dt
¼ kDCA
c
3564 Phys. Chem. Chem. Phys., 2013, 15, 3562--3569
This journal is the Owner Societies 2013

View Article Online
Paper
PCCP
Table 1 Anisole hydrolysis rate constants, kH for reaction in HTW with no added
catalyst. The rate constants reported are first-order in anisole and 2.6-order in
water. The reported error is the 95% confidence interval
Temperature (1C)
kH (L2.6 molꢁ2.6 sꢁ1
)
350
365
380
400
(4.6 ꢀ 0.6) ꢃ 10ꢁ10
(4.1 ꢀ 0.4) ꢃ 10ꢁ9
(6.4 ꢀ 0.8) ꢃ 10ꢁ9
(9.6 ꢀ 1.3) ꢃ 10ꢁ9
Fig. 4 Arrhenius plot for hydrolysis of anisole to phenol. Data are taken from
Patrick et al.8 (&), Klein et al.7 (ꢃ), and the present study (J).
rate equation that is first order in anisole and 2.6-order in
water. The agreement among all the data was good, as shown in
Fig. 4. We again employed nonlinear regression with Scientist
to determine activation energy to be 38 ꢀ 6 kcal molꢁ1 and
frequency factor to be 104.3ꢀ2.0 sꢁ1
M
ꢁ2.6. Out of curiosity, we
repeated the Arrhenius fit to the 2.79-apparent reaction order in
water that Klein et al. determined, and computed activation
energy and frequency factor to be 39 kcal molꢁ1 and 104.5 ꢁ1 Mꢁ2.6
s
respectively, well within the error bars of our analysis. Thus,
we conclude that anisole hydrolysis follows a constant almost
third-order rate law in water concentration in both the sub-critical
Fig. 3 Hydrolysis rate constant kH vs. water concentration at 380 1C. Data are
from Klein et al.7 (ꢃ) and the present study (J).
a value of 2.79.) Given the scatter in the data, we judge these and super-critical regimes.
results to be in good agreement.
The Arrhenius plot, including nine points at five tempera-
The two data sets are slightly offset from one another, likely tures from three labs, indicates that a single line explains the
due to systematic differences in experimental apparatus. data from 300 to 400 1C. This linearity indicates that the
Klein’s experiments were conducted in 1/400 stainless steel activation energy, and presumably, the mechanism, do not
reactors, while the present data set was obtained with 2 mm change for anisole hydrolysis as the reaction medium moves
quartz capillary reactors, which have a shorter heat-up time in through the critical point. Since Patrick et al. have established
the sand bath. Perhaps more importantly, in both labs, the the SN2 mechanism for anisole hydrolysis in sub-critical water,8
reaction pressure is not measured. The pressure is calculated it can be deduced that the mechanism for anisole hydrolysis in
based upon the estimated reactor volume, the amount of water high-temperature water more generally, is also SN2.
loaded, the reaction temperature, and the thermophysical
For a number of substituted anisoles, Klein et al. found an
properties of water, which are assumed to determine the apparent order in water concentration, ranging from 2.04 to
thermophysical behavior of the reaction mixture. Since the 3.33.7 The researchers appraised the apparent order in water
offset is relatively small from the viewpoint of kinetic data, concentration as a measure of the solvent effect of super-critical
only e0.5 = 1.6, it is likely attributable to such systematic error water upon hydrolysis of anisoles. Our work supports extending
arising from the differences in equipment. However, in each set the applicability of such solvent effects into the near-critical
of laboratory experiments, it is only the water density that is regime. The modeling work of Tucker and Gibbons further
being varied. Hence, the slope of each line still offers order in demonstrates the important role played by water in anisole
water, largely decoupled from other experimental factors. We hydrolysis in HTW. Tucker and Gibbons found that solvent
took a democratic approach and settled upon an apparent clustering effects were an important component in determining
order of water that is the weighed averaged of that determined the entropy of activation.14 They modeled solute–solvent inter-
by the two sets of experimental data: 2.6 – weighed by the actions with a pressure-dependent dielectric constant. Taking
number of data points in each data set.
these studies together, it seems plausible that water stabilizes
This result suggests that hydrolysis of anisole is almost third the transition state structure in anisole hydrolysis. We posit
order in water. We supposed that this should also be the case that the key aspect of the role played by water is the stabili-
for subcritical water. We combined our data with the other zation of the incipient phenoxide anion through hydrogen
literature data for anisole hydrolysis in HTW, and plotted all on bonding, as portrayed in Fig. 5. In fact, when Klein et al. order
an Arrhenius plot, with rate constant k being calculated from a the substituted anisoles of their study in terms of decreasing
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 3562--3569 3565

View Article Online
PCCP
Paper
Fig. 5 The SN2 hydrolysis of anisole in HTW may proceed through a phenoxide anion that is stabilized by hydrogen bonding with water molecules.
‘‘solvent effect’’, with one exception, they have also ordered the Second, the pressure required to maintain a saturated liquid
compounds in terms of increasing potential to withdraw elec- condition for water at 365 1C is almost 200 atm. Third, decom-
tron density from the methoxy group, e.g., in terms of ability to position at this temperature is not negligible. Lastly, the rate is
stabilize a negatively charged transition state. Hence, an still slow. The reaction is complete, with 97% yield, after one
increased ‘‘solvent effect’’ compensates for a decreased ability day. Other methods proceed at milder conditions with better
to stabilize a negatively charged transition state. The single selectivity and shorter reaction times.1,6
exception to this rule is m-(trifluoromethyl)anisole, whose
solvent effect is larger than anticipated given its substituent
effect. This may be due to a disruption of hydrogen bonding
5 Anisole hydrolysis catalyzed by
water-tolerant Lewis acids
due to the halogen atoms. Through their studies, Tucker and
Gibbons indicated the importance of including the effects of
We supposed that since the customary procedure for aryl alkyl
variable clustering, which is strongest at the reaction inter-
ether hydrolysis involved the use of a Lewis acid, that perhaps a
mediate. Given the important role played by water in this
water-tolerant Lewis acid (WTLA) could improve the reactivity
reaction, we wonder if theoretical models could be improved
of anisole toward hydrolysis in near critical water. A survey of
by explicitly including at least three water molecules for
four different WTLAs alongside mineral acids in HTW at 275 1C
reaction and phenoxide stabilization through hydrogen bonding;
and that the macroscopic–microscopic connection be made at the
was carried out in 1/400 stainless steel reactors. Each experiment
was replicated two to seven times, and the resulting 95%
interface of water molecules with a solvent continuum that
confidence intervals recorded. The room temperature pH of
captures the fluid’s dielectric constant.
each aqueous solution was measured with an Accumet
pH meter calibrated with standard buffer solutions at
pH 2, 4, and 7.
4 Assessment of uncatalyzed anisole
Fig. 6 shows molar yield of phenol after 2.75 hours at
hydrolysis in HTW
reaction temperature for the different acid catalysts tested.
For each experiment, the initial loading of anisole was
Given the data presented thus far, there seems to be three
factors that are most important in determining the optimal
conditions for uncatalyzed anisole hydrolysis in HTW. Since
order in water concentration is almost three, conditions should
160 mM and catalyst loading was 8.0 ꢀ 0.3 mM. Error bars
represent 95% confidence. At these conditions, uncatalyzed
anisole hydrolysis resulted in negligible phenol yield. The
results demonstrate that WTLAs are successful in catalyzing
be chosen to maximize water density. However, the activation
energy for this system is also quite large, 38 kcal molꢁ1
,
indicating that high temperatures are necessary for adequate
rates. For saturated liquid water below the critical point,
as temperature increases, water density decreases. However,
the temperature dependence of the rate is exponential while
the water density dependence of the rate is almost third power,
so higher temperatures will generally increase the reaction
rate, at least until decomposition reactions become important.
Hence the experimentally determined optimum temperature
of 365 1C for uncatalyzed anisole hydrolysis may be the
temperature that best balances hydrolysis with decom-
position.
However, there are drawbacks to reaction at this condition.
First, the temperature is very high. Since the deprotection of
aryl methyl ethers is a reaction of interest to the pharmaceutical
and fine chemical industry, a high temperature of 365 1C
would preclude applications with thermally labile moieties.
Fig. 6 Anisole hydrolysis to phenol at 250 1C and 2.75 hours; comparison of
yield using different catalysts.
c
3566 Phys. Chem. Chem. Phys., 2013, 15, 3562--3569
This journal is the Owner Societies 2013

View Article Online
Paper
PCCP
anisole hydrolysis. Some of the WTLAs tested – namely,
In(OTf)3 and Sc(OTf)3 were superior to a commensurate loading
(by mole) of mineral acids HCl and H2SO4. These results
were promising, and inspired us to continue studying anisole
hydrolysis catalyzed by In(OTf)3.
We began by confirming reaction order in substrate and in
catalyst. The analytical procedure for these analyses is exactly
analogous to that described previously for 1-phenyl-1-propyne
hydration in HTW.15 To recapitulate the methodology in brief,
initial rates are measured for substrate loadings that differ
greatly. Here, almost four orders of e were spanned in anisole
concentration, from 8.2 to 650 mM. If the reaction is first
order, then the rate constant will not change if substrate
concentration changes. Fig. 7 shows the result of the pseudo-
first-order analysis for anisole, where k0 is the pseudo first
order rate constant in substrate, and the mean concentration of
anisole, CA, is found by taking the arithmetic mean of anisole
concentration before and after the reaction. Note that
anisole concentration before reaction is calculated based on
the amount of anisole added to the reactors (10 mL). Each
experiment was replicated 2 to 8 times, with the average
number of replicates being 4.2 per datum. The error bars
Fig. 8 Pseudo-first-order plot for catalyst in anisole hydrolysis at 275 1C with
In(OTf)3. Initial loading of anisole was 160 mM. Error bars represent 95%
confidence intervals.
determined slope from Fig. 8 is 1.33 ꢀ 0.08. However, without
additional compelling evidence for a fractional order, we
yielded to simplicity and adopted a rate law that was first-order
in indium triflate catalyst.
Armed with knowledge of order in substrate and in catalyst,
we proceeded to collect data at different times for different
temperatures to determine activation energy and frequency
factor for In(OTf)3-catalyzed anisole hydrolysis. We performed
a series of anisole hydrolysis reactions with an initial anisole
loading of 160 mM, and 5 mol% In(OTf)3 catalyst at tempera-
tures 200, 225, 250, 275, and 300 1C. The results of our study
are depicted as data points in Fig. 9. Once again, we used
Scientistt to fit the data to a kinetic rate law that was first-order
in anisole and first-order in catalyst. The set of two differential
equations defining the kinetics, eqn (3) was solved numerically
using the Runge–Kutta algorithm, and least squares methodology
was used for parameter estimation with non-linear regression. In
contrast to the uncatalyzed experiments, no side products were
observed in any of the catalyzed experiments. Hence, a system of
shown signify 95% confidence intervals in both ln(k0)
ÀÂ ÃÁ
and ln CA
. Linear regression estimates the slope as
0.05 ꢀ 0.15, which has no statistically significant difference
from zero. The slope of zero indicates that the reaction is first
order in substrate.
To find order in catalyst, a similar analysis was carried out,
wherein catalyst concentration is varied by almost four orders
of e, from 1.8 to 65 mM, and the effect upon rate is measured by
calculating the pseudo first order rate constant k0. The concen-
tration of catalyst is assumed to be constant over the period of
reaction, and to be given by the amount added to the reactor.
Initial anisole concentration is held constant at 160 mM for all
of these experimental runs. Experiments were conducted in
replicates of two to eight per datum, with the average number
of replicates being 4.4. A slope of unity would indicate that the
reaction is first order in catalyst. The experimentally
Fig. 7 Pseudo-first-order plot for substrate in anisole hydrolysis at 275 1C with
catalysis by In(OTf)3. Catalyst loading was 8.2 mM. Error bars represent 95%
confidence intervals.
Fig. 9 Kinetics of anisole hydrolysis at different temperatures with In(OTf)3
catalyst, represented as conversion vs. time. Error bars represent 95% confidence
intervals.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 3562--3569 3567

View Article Online
PCCP
Paper
Table 2 Anisole hydrolysis rate constant, k for reaction in HTW with 5 mol%
In(OTf)3 catalyst. The rate constants reported are first-order in anisole and first-
order in catalyst. The reported error is the 95% confidence interval
365 1C for HTW hydrolysis, and compared it to the pseudo-first
order rate constant at 300 1C for In(OTf)3-catalyzed hydrolysis
in HTW. These values are shown in Table 3. Compared to the
uncatalyzed route in HTW, anisole hydrolysis with In(OTf)3
catalysis in HTW is much faster.
We wanted to compare our best reaction conditions with
another literature method for the same synthesis, so we calculated
the pseudo-first order rate constant for the hydrolysis of anisole to
phenol for a technique that uses 6 equivalents of BF3ꢂSMe2 catalyst
in CH2Cl2.6 Our estimate of k for this method, in Table 3, is based
on a datum presented by Konieczny et al.6 Our method of using
In(OTf)3 catalyst in HTW compares favorably to the BF3ꢂSMe2
method. This demonstrates that the use of WTLAs in HTW could
offer rates and yields that are comparable to those provided by
conventional techniques.
The activation energy determined for In(OTf)3 catalyzed
anisole hydrolysis was 31 ꢀ 1 kcal molꢁ1, while the uncatalyzed
method for anisole hydrolysis in HTW gave an activation energy
of 38 ꢀ 6 kcal molꢁ1. Nominally, the use of In(OTf)3 catalyst
lowered the energy barrier for hydrolysis – the generally
expected outcome for a catalyzed reaction; this result should
be interpreted within the context of the measured error.
The advantages of using the WTLA-catalyzed method in HTW
include the complete avoidance of organic solvent and the use of
catalytic quantities of catalyst rather than stoichiometric. If the
water can be recycled, then this method can lead to reduced waste
generation. If needed, faster rates than those reported herein are
expected if more catalyst is used; we used only 5 mol% relative to
anisole in our studies. A higher catalyst loading could also decrease
the reaction temperature, if needed.
Temperature (1C)
k (L molꢁ1 sꢁ1
)
200
225
250
275
300
0.000198 ꢀ 0.000026
0.00147 ꢀ 0.00014
0.00394 ꢀ 0.00075
0.0122 ꢀ 0.0033
0.0446 ꢀ 0.0065
only two differential equations is sufficient to describe the data
set. The results of data fitting are tabulated in Table 2 and shown
as smooth lines in Fig. 9.
dCA
¼ ꢁkCACcatalyt
dt
(3)
dCP
¼ kCACcatalyst
dt
We could now determine the activation energy and frequency
factor for this reaction with an Arrhenius plot. From Fig. 10, the
activation energy was determined to be 31 ꢀ 1 kcal molꢁ1
;
the frequency factor, 1010.6ꢀ0.5 sꢁ1 L molꢁ1. If we had modeled
the data as 1.33-order in catalyst, the activation energy, of course,
would be the same 31 ꢀ 1 kcal molꢁ1; only the frequency factor
would change slightly to 1011.3ꢀ0.4 ꢁ1 L1.33 molꢁ1.33 to reflect the
s
slightly more demanding collision requirement.
6 Assessment of anisole hydrolysis with
indium triflate catalysis for deprotection
We have now studied anisole hydrolysis under uncatalyzed
HTW conditions, and under WTLA catalysis in HTW. We
calculated the pseudo-first order rate constant k (sꢁ1) at
The WTLA method introduces some disadvantages, too. The
higher temperatures and pressures required incur a higher
capital cost and higher operating costs relative to a room
temperature reaction at atmospheric pressure. Further, the
presented method is not expected to provide selectivity, while
conventional methods may offer selectivity with the right
catalyst. A selective catalyst has not yet been identified for
hydrolysis in HTW. WTLA-catalyzed hydrolysis in HTW
will likely not be the choice method of deprotection for every
multi-step synthesis, but for the right reaction, it can be the
right tool.
Acknowledgements
We gratefully acknowledge Dongil Kang, Alex Cohen, and
Stephanie Fraley for their assistance in completing experi-
ments. This work was funded by the National Science Founda-
tion (CTS-0625641).
Fig. 10 Arrhenius plot for anisole hydrolysis with In(OTf)3 catalyst in HTW.
Table 3 Comparison of anisole hydrolysis rate among different methods. The kinetics reported are pseudo-first order in anisole hydrolysis rate constants
Method
Temperature (1C)
Time (h)
Equiv. of catalyst
Phenol % yield
k (sꢁ1
)
Uncatalyzed HTW
In(OTf)3 in HTW
BF3ꢂSMe2 in CH2Cl2
365
300
0
24
2
3
+
0.05
6
83
88
80
0.000023 ꢀ 0.000001
0.00035 ꢀ 0.00003
0.00015
c
This journal is the Owner Societies 2013
3568 Phys. Chem. Chem. Phys., 2013, 15, 3562--3569

View Article Online
Paper
PCCP
9 P. Savage and N. Rebacz, Green Solvents Vol. 5: Reactions in
Water, John Wiley & Sons, Inc., 2010, ch. Water Under
Extreme Conditions for Green Chemistry.
10 E. W. Lemmon, M. O. McLinden and D. G. Friend, in
NIST Chemistry WebBook, NIST Standard Reference
Database Number 69, ed. P. J. Linstrom and W. G.
Mallard, National Institute of Standards and Technology,
Gaithersburg MD, 2008, p. 20899.
11 F. E. Anderson and J. M. Prausnitz, Fluid Phase Equilib.,
1986, 32, 63–76.
12 H. Chen and J. Wagner, J. Chem. Eng. Data, 1994, 39,
470–474.
13 K. Chandler, B. Eason, C. L. Liotta and C. A. Eckert, Ind. Eng.
Chem. Res., 1998, 37, 3515–3518.
14 S. C. Tucker and E. M. Gibbons, Structure and reactivity in
aqueous solution: characterization of chemical and biological
systems, 1994, vol. 568, p. 196.
References
1 P. R. Brooks, M. C. Wirtz, M. G. Vetelino, D. M. Rescek,
G. F. Woodworth, B. P. Morgan and J. W. Coe, J. Org. Chem.,
1999, 64, 9719–9721.
2 T. Shindo, Y. Fukuyama and T. Sugai, Synthesis, 2004,
692–700.
3 A. K. Chakraborti, M. K. Nayak and L. Sharma, J. Org. Chem.,
2002, 67, 1776–1780.
4 S. K. Boovanahalli, D. W. Kim and D. Y. Chi, J. Org. Chem.,
2004, 69, 3340–3344.
5 J. R. Hwu, F. F. Wong, J. J. Huang and S. C. Tsay, J. Org.
Chem., 1997, 62, 4097–4104.
6 M. T. Konieczny, G. Maciejewski and W. Konieczny, Synthesis,
2005, 1575–1577.
7 M. T. Klein, Y. G. Mentha and L. A. Torry, Ind. Eng. Chem.
Res., 1992, 31, 182–187.
8 H. R. Patrick, K. Griffith, C. L. Liotta, C. A. Eckert and
R. Glaser, Ind. Eng. Chem. Res., 2001, 40, 6063–6067.
15 N. A. Rebacz and P. E. Savage, Ind. Eng. Chem. Res., 2010, 49,
535–540.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 3562--3569 3569