Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

1822 Ind. Eng. Chem. Res.

2001, 40, 1822-1831

Kinetics and Mechanism of Cyclohexanol Dehydration in


High-Temperature Water
Naoko Akiya and Phillip E. Savage*
Department of Chemical Engineering, University of Michigan, Ann Arbor, Michigan 48109-2136

We examined cyclohexanol dehydration in pure water at temperatures of 250, 275, 300, 350,
and 380 °C with water densities ranging from 0.08 to 0.81 g/cm3. Under these conditions,
cyclohexanol dehydrates readily in the absence of added catalysts to form cyclohexene as the
major product. The most abundant minor products are 1- and 3-methyl cyclopentenes. The
reaction rate and product distribution at 380 °C show a remarkable sensitivity to the water
density. At low densities, the reaction is slow, and cyclohexene is the only product. At high
densities, the reaction is nearly complete, and methyl cyclopentenes appear along with
cyclohexene. The experimental results implied a reaction mechanism that comprises two
pathways: (1) reversible cyclohexanol dehydration to form cyclohexene through an E2 mecha-
nism, and (2) subsequent cyclohexene protonation to form the cyclohexyl cation, which rapidly
rearranges to form methyl cyclopentyl cations, which then lose a proton to form methyl
cyclopentenes. A kinetics model based on the proposed mechanism was able to predict the striking
effect of the water density on the product yields at 380 °C and, thereby, to demonstrate that the
proposed mechanism captures the trends in the experimental data. An analysis of mechanistic
issues regarding cyclohexanol dehydration in high-temperature water (HTW) revealed three
roles for water. Water participates in elementary reaction steps as a reactant and as a product,
water is the source of the acid catalyst (H3O+), and water also drives the mechanism toward E2
by favoring, through solvation, the oxonium ion rather than the carbocation as the reaction
intermediate. This study provides further evidence that acid-catalyzed reactions can be
accomplished readily in HTW in the absence of added acid and that HTW has potential
applications in environmentally benign industrial chemistry.

Introduction temperature of 375 °C for 20 min, but that the addition


of a solid metal catalyst (PtO2), acid (HCl), or base (NH4-
High-temperature water (HTW), including both liquid OH) leads to a variety of products. Kuhlmann et al.,4
water (T > ∼200 °C) and supercritical water (T > 374 on the other hand, reported that the dehydration to
°C, P > 218 atm), is attracting increased attention as a
cyclohexene occurs readily at 300 °C after 60 min
medium for organic chemistry.1,2 Water near its critical
without the deliberate addition of any catalysts. They
point exhibits many properties that are distinctly dif-
ferent from those of ambient liquid water, such as high also reported that cyclohexene is the only product
solubility for organic and gaseous compounds, high ion observed in their experiments. They speculated that the
product, low dielectric constant, and weak hydrogen dehydration was acid-catalyzed, with H3O+ generated
bonds. These and other properties of HTW vary over a from the dissociation of water being the likely catalyst.
wide range of values as a function of temperature and This hypothesis is consistent with the work by Antal
pressure. This strong temperature and pressure depen- and co-workers, who showed that tert-butyl alcohol
dence of the properties provides opportunities to tune dehydration occurs in pure HTW at 250 °C without the
the reaction environment to optimal conditions for the addition of any catalyst.5,6
chemical transformation of interest. Antal and co-workers have been very active in the
The reactions of cyclohexanol and its products in field of alcohol (but not cyclohexanol) dehydration in
HTW provide an opportunity for the convenient exami- HTW. They have reported mechanisms and kinetics for
nation of several different types of chemical transforma-
the dehydration of ethanol, 1- and 2-propanol, and tert-
tions (dehydration, dehydrogenation, aromatization,
butyl alcohol to the corresponding alkenes.5-8 The
rearrangement) in HTW. As such, it is a good model
system. The goal of the study described herein is to quantitative kinetics data have facilitated the concep-
obtain kinetics data for cyclohexanol dehydration in tual design and economic evaluation of chemical pro-
HTW and to use them to gain mechanistic insights. This cesses based on the new chemistry in HTW.9,10
study of cyclohexanol dehydration is the first step In contrast to the kinetics and mechanistic informa-
toward a better understanding of cyclohexanol chemis- tion available for the dehydration of the C2-C4 alcohols
try in HTW. in HTW, there are no kinetics or mechanisms available
Crittendon and Parsons3 reported that cyclohexanol for the dehydration of cyclohexanol in HTW. Moreover,
is unreactive in pure HTW at a nominal reaction the limited data that are available are contradictory.
The present investigation of cyclohexanol dehydration
* Corresponding author. E-mail: psavage@umich.edu. in HTW resolves this conflict and fills this gap in the
Phone: 734 764-3386. Fax: 734 763-0459. literature on alcohol dehydration in HTW.
10.1021/ie000964z CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/27/2001
Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001 1823

Experimental Section temperatures and pressures. Simulations of cyclohex-


anol in water and cyclohexene in water predicted a
We performed experiments at different batch holding single liquid phase at the experimental conditions.
times, temperatures, and water densities to determine Judging from these calculations and the experimental
the effects of these variables on the product yields and results of Kuhlmann et al., we believe that the reactor
reaction rates. The conditions examined are tempera- contents were always homogeneous under the reaction
tures of 250, 275, 300, 350, and 380 °C, times ranging conditions used.
from 15 to 180 min, and water densities ranging from The loaded and sealed reactors were placed in a
0.08 g/cm3 (4.7 mol/L) to 0.81 g/cm3 (45 mol/L). The preheated, fluidized sand bath (Techne SBL-2) set at
cyclohexanol concentration was 0.3 mol/L (at reaction the desired reaction temperature. The sand bath was
conditions) in all experiments. kept isothermal to within (1 °C with a temperature
We used batch reactors fashioned from nominal controller (Techne TC-8D). The reactor heat-up time has
1/ -in. stainless steel Swagelok tube fittings (one port
4 been measured to be 2-3 min,13 which is short com-
connector and two caps). The reactor volume is 0.59 cm3. pared to the typical batch holding times used in this
Preliminary experiments indicated that the use of new study. Upon reaching the desired holding time, the
reactors (as received) produced results that differ from reactors were removed from the sand bath, and the
those obtained with reactors that have been used at reaction was quenched immediately by immersing the
least once. New reactors led to higher cyclohexanol reactors in a cold water bath. The reactor temperature
conversions and the formation of additional (unidenti- dropped to room temperature in less than 1 min. The
fied) byproducts. To eliminate these effects, we seasoned reactors were further cooled in a freezer for up to 1 h to
the new reactors prior to use. The reactors were loaded condense any volatile components. The reactors were
with water and heated in a sand bath at 300 °C for 30 opened, and the contents were recovered through the
min, after which they were cooled gradually at ambient addition of acetone. It was essential to complete this last
conditions. The reactors were then washed with acetone step quickly to minimize the loss of volatile compounds.
and dried. All of the data reported herein were obtained We used two Hewlett-Packard model 5890 gas chro-
from seasoned reactors. matographs equipped with either a flame ionization
The reactors were loaded with a carefully measured detector or a mass spectrometric detector to analyze the
amount of cyclohexanol and placed in a glovebox filled reaction products. The sample constituents were sepa-
with purified helium. In the same glovebox, we vigor- rated with a Hewlett-Packard HP-5 fused silica capillary
ously bubbled helium through distilled and deionized column. Product identification was accomplished by
water to eliminate dissolved oxygen and carbon dioxide, comparing retention times with those of authentic
gases that might influence the reactions. The reactors standards and by inspecting mass spectra. The number
were then loaded with a measured amount of deaerated of moles of each product present was obtained from the
water and sealed under a stream of helium in the chromatographic analyses with experimentally deter-
glovebox. In some experiments, cyclohexene was the mined detector response factors for each component. We
reactant. Because cyclohexene is very volatile, it was used methyl cyclohexane as the standard. Product molar
added to the reactor in the glovebox immediately after yields were then computed by dividing the number of
the reactor was loaded with deaerated water. All moles of product formed by the number of moles of
chemicals were obtained commercially in high purity cyclohexanol initially loaded into the reactor.
and used as received.
The amount of water added to the reactor set the Experimental Results
volume of the vapor phase at subcritical reaction tem-
peratures and the water density at supercritical reaction Table 1 and Figures 1-7 present the experimental
temperatures. For experiments at subcritical tempera- product yields for cyclohexanol dehydration in pure
tures (250-350 °C), we chose water loadings that HTW. The uncertainties reported are 95% confidence
ensured that the reactors would be nearly filled with a intervals determined from multiple (typically 5-10)
single liquid phase. The required water loading for each runs under nominally identical reaction conditions.
temperature was determined by taking the density of Under the conditions employed in this study, cyclohex-
the reactor contents to be that of pure liquid water at anol readily dehydrates to form cyclohexene in HTW
the same temperature.11 without any added catalyst. The minor products include
Because cyclohexanol and cyclohexene are both in- 1- and 3-methyl cyclopentene, with yields as high as 22
soluble in water at room temperature, we were con- and 6%, respectively. Our identification of 3-methyl
cerned with the possibility that two immiscible liquid cyclopentene remains tentative because the mass spec-
phases might be present in the reactor during experi- trum for this compound is also consistent with the
ments at subcritical temperatures. Kuhlmann et al.4 fragmentation pattern for 2-methyl cyclopentene. Nev-
ascertained experimentally the solubility of cyclohexanol ertheless, we are confident that this product is a methyl
(and other organic reagents) in water up to 100 °C and cyclopentene. Methylene cyclopentane formed in some
concluded that mixtures that are 0.3-0.5 M in organics experiments but in yields of less than 1%. The combined
at room temperature are homogeneous at 300 °C. We yields of 1- and 3-methyl cyclopentene and methylene
conducted our experiments at temperatures of 250-350 cyclopentane appear as “methyl cyclopentenes” in the
°C and cyclohexanol concentrations of 0.4-0.6 M at figures and in Table 1. Cyclohexanone was also present
room temperature. To probe further the solubility of but in yields that were typically less than 1%. The
cyclohexanol and cyclohexene in HTW at the tempera- highest cyclohexanone yields (2-3%) were obtained at
tures and concentrations used in our experiments, we 380 °C and water densities of 0.20 g/cm3 and lower. The
performed calculations with the chemical process simu- carbon balance generally exceeded 90%, and the lowest
lator ASPEN PLUS.12 We used the RK-ASPEN ther- value was 79%. The carbon loss showed no notable trend
mophysical properties model, which is suitable for a with changes in temperature or water density. Back-
mixture of polar and nonpolar compounds at high ground experiments revealed that the failure to achieve
1824 Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001

Table 1. Experimental Molar Yields for Cyclohexanol Dehydration in Pure HTWa


T (°C) [H2O]0 (g/cm3) time (min) cyclohexanol cyclohexene methyl cyclopentenesb cyclohexanoneb
250 0.81 30 0.86 ( 0.15 0.077 ( 0.040 nd nd
60 0.70 ( 0.03 0.085 ( 0.043 nd nd
90 0.68 ( 0.08 0.20 ( 0.07 nd nd
120 0.70 ( 0.07 0.23 ( 0.07 nd nd
180 0.63 ( 0.11 0.24 ( 0.06 nd nd
275 0.76 15 0.71 ( 0.08 0.12 ( 0.04 nd nd
30 0.58 ( 0.09 0.27 ( 0.07 nd nd
45 0.43 ( 0.06 0.38 ( 0.06 nd nd
60 0.43 ( 0.08 0.41 ( 0.08 nd nd
90 0.40 ( 0.07 0.50 ( 0.07 nd nd
300 0.73 15 0.63 ( 0.08 0.33 ( 0.07 nd 0.001 ( 0.001
30 0.32 ( 0.07 0.59 ( 0.07 0.001 ( 0.001 0.001 ( 0.001
45 0.34 ( 0.08 0.64 ( 0.08 0.002 ( 0.001 0.001 ( 0.001
60 0.27 ( 0.07 0.67 ( 0.06 0.003 ( 0.001 0.002 ( 0.002
90 0.18 ( 0.03 0.71 ( 0.05 0.005 ( 0.001 0.002 ( 0.001
350 0.59 15 0.075 ( 0.015 0.80 ( 0.06 0.013 ( 0.003 0.004 ( 0.002
30 0.071 ( 0.006 0.80 ( 0.03 0.013 ( 0.003 0.003 ( 0.003
45 0.063 ( 0.023 0.82 ( 0.07 0.022 ( 0.012 0.002 ( 0.003
60 0.076 ( 0.013 0.76 ( 0.04 0.033 ( 0.006 0.004 ( 0.000
90 0.074 ( 0.019 0.78 ( 0.05 0.033 ( 0.014 0.003 ( 0.003
380 0.08 60 0.86 ( 0.07 0.070 ( 0.073 nd 0.021 ( 0.032
0.14 30 0.76 ( 0.08 0.12 ( 0.08 nd 0.004 ( 0.006
45 0.75 ( 0.08 0.14 ( 0.02 nd 0.011 ( 0.010
60 0.69 ( 0.09 0.15 ( 0.04 nd 0.020 ( 0.014
75 0.68 ( 0.06 0.17 ( 0.03 nd 0.007 ( 0.014
90 0.65 ( 0.08 0.28 ( 0.06 nd 0.021 ( 0.009
0.17 30 0.64 ( 0.05 0.23 ( 0.09 nd 0.007 ( 0.007
45 0.60 ( 0.10 0.28 ( 0.08 nd 0.010 ( 0.010
60 0.57 ( 0.07 0.34 ( 0.05 0.001 ( 0.001 0.009 ( 0.019
75 0.58 ( 0.08 0.38 ( 0.07 nd 0.011 ( 0.006
90 0.56 ( 0.07 0.34 ( 0.09 nd 0.012 ( 0.005
0.20 30 0.52 ( 0.06 0.35 ( 0.06 nd 0.006 ( 0.006
45 0.49 ( 0.09 0.46 ( 0.08 nd 0.012 ( 0.004
60 0.33 ( 0.09 0.58 ( 0.07 nd 0.028 ( 0.021
75 0.26 ( 0.05 0.68 ( 0.07 0.001 ( 0.002 0.009 ( 0.003
90 0.21 ( 0.05 0.71 ( 0.07 0.001 ( 0.002 0.011 ( 0.007
0.25 60 0.041 ( 0.055 0.91 ( 0.06 0.010 ( 0.003 0.010 ( 0.019
0.34 15 0.032 ( 0.004 0.87 ( 0.03 0.009 ( 0.001 0.005 ( 0.006
30 0.025 ( 0.010 0.87 ( 0.09 0.012 ( 0.003 0.004 ( 0.005
45 0.026 ( 0.004 0.86 ( 0.04 0.013 ( 0.002 nd
60 0.032 ( 0.016 0.87 ( 0.05 0.015 ( 0.013 0.013 ( 0.028
0.51 15 0.045 ( 0.010 0.80 ( 0.04 0.026 ( 0.011 0.008 ( 0.010
30 0.042 ( 0.006 0.81 ( 0.06 0.035 ( 0.046 0.004 ( 0.011
45 0.048 ( 0.016 0.84 ( 0.03 0.052 ( 0.024 0.008 ( 0.004
60 0.046 ( 0.006 0.82 ( 0.05 0.061 ( 0.032 0.009 ( 0.004
90 0.041 ( 0.009 0.74 ( 0.07 0.12 ( 0.05 0.006 ( 0.003
0.68 15 0.068 ( 0.007 0.82 ( 0.07 0.055 ( 0.014 nd
30 0.068 ( 0.013 0.74 ( 0.09 0.078 ( 0.025 nd
45 0.064 ( 0.010 0.70 ( 0.11 0.12 ( 0.04 nd
60 0.059 ( 0.007 0.60 ( 0.05 0.21 ( 0.01 0.005 ( 0.001
120 0.055 ( 0.018 0.48 ( 0.09 0.26 ( 0.08 0.006 ( 0.001
180 0.035 ( 0.013 0.48 ( 0.09 0.28 ( 0.04 0.001 ( 0.002
a [Cyclohexanol]0 ) 0.3 mol/L. b nd ) not detected.

Table 2. Molar Yields of Individual Methyl yields of methylene cyclopentane were always an order
Cyclopentenes at 380 °C and 0.68 g/cm3 of magnitude smaller than the yields of the 1- and
time 1-methyl 3-methyl methylene 3-methyl cyclopentenes.
(min) cyclopentene cyclopentene cyclopentane Figures 1-3 show the temporal variations of the
15 0.038 ( 0.012 0.016 ( 0.002 0.001 ( 0.001 product yields for cyclohexanol dehydration in HTW at
30 0.058 ( 0.021 0.020 ( 0.005 nd 250, 300, and 350 °C, respectively. At the lower tem-
45 0.092 ( 0.027 0.027 ( 0.009 0.001 ( 0.002
60 0.17 ( 0.01 0.037 ( 0.002 0.002 ( 0.001
peratures, the cyclohexanol yield decreases continuously
120 0.21 ( 0.06 0.045 ( 0.014 0.004 ( 0.003 with time, and this change is accompanied by a corre-
180 0.22 ( 0.04 0.057 ( 0.008 0.003 ( 0.002 sponding increase in the cyclohexene yield. At 250 °C,
cyclohexene is the only product, whereas at 300 °C,
100% carbon balance is most likely due to losses of methyl cyclopentenes appear in a low yield that in-
volatile products (cyclohexene and methyl cyclopentenes) creases with time. At 350 °C, the conversion of cyclo-
prior to analysis. hexanol is rapid and appears to reach an equilibrium
Table 2 shows the product yields for individual methyl value after no more than 15 min of reaction time. The
cyclopentenes at 380 °C and 0.68 g/cm3, the reaction cyclohexene yield, on the other hand, reaches 80% after
conditions for which the yields of these minor products 15 min but then gradually decreases at longer times.
were the highest. 1-Methyl cyclopentene always had the This decrease in the cyclohexene yield is accompanied
highest yield, followed by 3-methyl cyclopentene. The by a corresponding increase in the methyl cyclopentenes
Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001 1825

Figure 1. Yields of cyclohexanol and cyclohexene at T ) 250 °C Figure 4. Yields of cyclohexanol and cyclohexene at T ) 380 °C
and [H2O]0 ) 0.81 g/cm3. and [H2O]0 ) 0.14 g/cm3.

Figure 2. Yields of cyclohexanol, cyclohexene, and methyl cyclo- Figure 5. Yields of cyclohexanol and cyclohexene at T ) 380 °C
pentenes at T ) 300 °C and [H2O]0 ) 0.73 g/cm3. and [H2O]0 ) 0.20 g/cm3.

Figure 3. Yields of cyclohexanol, cyclohexene, and methyl cyclo- Figure 6. Yields of cyclohexanol, cyclohexene, and methyl cyclo-
pentenes at T ) 350 °C and [H2O]0 ) 0.59 g/cm3. pentenes at T ) 380 °C and [H2O]0 ) 0.51 g/cm3.

yield. Taken collectively, Figures 1-3 show that in- and the formation of methyl cyclopentenes. At these
creasing temperature increases the rate of cyclohexanol more liquidlike densities (0.51 and 0.68 g/cm3), the
disappearance and the selectivity toward methyl cyclo- methyl cyclopentenes yields at 380 °C are substantially
pentenes. higher than those observed at lower temperatures. This
Figures 4-7 show the temporal variations of the result is consistent with an increase in temperature
product yields at 380 °C and water densities of 0.14, shifting the selectivity toward methyl cyclopentenes.
0.20, 0.51, and 0.68 g/cm3, respectively. A remarkable We also conducted experiments in the absence of
influence of the water density on the yields is evident. water at 300 and 380 °C for 60 min. The cyclohexanol
At water densities of 0.14 and 0.20 g/cm3, cyclohexene conversion was 8.2 ( 2.0% at 300 °C and 8.5 ( 1.3% at
is the only product observed. The cyclohexanol yield 380 °C. The low cyclohexanol conversion in the absence
decreases continuously with time, and this decrease is of water confirms that water plays an important role
accompanied by a corresponding increase in the cyclo- in the reaction of cyclohexanol. Cyclohexanone was the
hexene yield. At water densities of 0.51 and 0.68 g/cm3, major product at both temperatures, but small quanti-
we observe a low and nearly time-invariant cyclohexanol ties of cyclohexene and methyl cyclopentenes were also
yield, a maximum in the cyclohexene yield, and the produced. At 300 °C, the product yields were 3.5 ( 0.6%
appearance of and a steady increase in the yield of for cyclohexanone and 0.2 ( 0.1% for cycloalkenes. At
methyl cyclopentenes. These data show that higher 380 °C, the product yields were 9.5 ( 1.1% for cyclo-
water densities facilitate the conversion of cyclohexanol hexanone and 0.5 ( 0.3% for cycloalkenes. Cyclohex-
1826 Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001

Figure 7. Yields of cyclohexanol, cyclohexene, and methyl cyclo- Figure 8. Yields of cyclohexanol, cyclohexene, and methyl cyclo-
pentenes at T ) 380 °C and [H2O]0 ) 0.68 g/cm3. pentenes at T ) 300 °C and [H2O]0 ) 0.73 g/cm3 with deaerated
(solid line) and untreated (dashed line) water.
anone likely forms as a result of thermally activated
free-radical dehydrogenation of cyclohexanol. The much Effect of Dissolved Air. We also examined the
lower yields of cyclohexanone from experiments in water effects of dissolved air on the reaction rates at 300 °C.
indicate that this pathway is either suppressed in HTW Dissolved oxygen can oxidize the organic compounds,
or supplanted by the faster dehydration pathway. and dissolved carbon dioxide forms carbonic acid, which
Cyclohexanol dehydration and formation of methyl might catalyze dehydration. Reactors loaded with cy-
cyclopentenes most likely occur by means of ionic clohexanol and unsparged (hence not deaerated) water
chemistry. were placed under a stream of purified helium and
We conducted experiments with cyclohexene in HTW sealed in the helium-filled glovebox to displace air from
to shed light on the fate of cyclohexene after its the reactor headspace.
formation via cyclohexanol dehydration. All experiments Figure 8 shows the product yields as functions of time
were conducted for the batch holding time of 60 min. for reactions at 300 °C in deaerated and unsparged
The cyclohexene conversion was 21.6 ( 9.6% at 300 °C water. The uncertainties from the experiments with
and 39.3 ( 5.7% at 380 °C. At 300 °C, the product yields unsparged water are larger because fewer replicates
were 14.3 ( 6.1% for cyclohexanol and 1.7 ( 1.6% for were done at each batch holding time. The cyclohexanol
methyl cyclopentenes. At 380 °C and 0.34 g/cm3 of conversion in Figure 8 is consistently higher in the
water, the product yields were 2.7 ( 0.5% for cyclohex- presence of dissolved air, and this rate-enhancing effect
anol and 15.3 ( 6.0% for methyl cyclopentenes. These appears to be stronger at shorter reaction times. For
results indicate that two reaction pathways are avail- example, the cyclohexanol conversions in the presence
able to cyclohexene in HTW. One pathway is hydration of dissolved air are 66% at 15 min and 85% at 90 min,
to generate cyclohexanol, and the other pathway is whereas in deaerated water, they are 37% and 82%,
rearrangement to form methyl cyclopentenes. The ob- respectively.
served product distribution indicates that hydration is The greater reaction rate is probably due to the lower
the preferred pathway at 300 °C whereas rearrange- pH produced by the dissolved CO2. Regardless of the
ment is the preferred pathway at 380 °C and 0.34 g/cm3. cause of this effect, however, it is clear that distilled
The higher yield of methyl cyclopentenes at higher and deionized water exposed to ambient air contained
temperature is consistent with the results obtained with sufficient quantities of dissolved gases to have measur-
cyclohexanol as the reactant. able effects on the reaction kinetics. Therefore, for
Effect of Metal Surface. Even though the reactors kinetics and mechanistic studies, it is important to
were treated hydrothermally prior to use, there remains remove dissolved air from water prior to use, even
a possibility that the treated metal surfaces impose though this necessity is not universally acknowledged
some catalytic or surface effects on the reaction. We in the studies of acid-catalyzed reactions in HTW found
investigated such wall effects by performing cyclohex- in the literature. From a practical point of view,
anol dehydration experiments under nominally identical however, the presence of dissolved air might be desir-
conditions but in reactors with varying surface-to- able precisely because of this rate enhancement.
volume ratios. This variation was achieved by sepa- Comparison with Previous Experiments. The
rately adding 7 and 22 mg of stainless steel filings to experimental results in Table 1 are in apparent conflict
the reactors. Reactions were carried out at 380 °C, 0.34 with those of Crittendon and Parsons,3 who reported
g/cm3 of water, and 60 min of batch holding time. In no reaction for cyclohexanol after 20 min in HTW (0.33
the absence of stainless steel filings, the product yields g/cm3) at 375 °C. We observed 97% conversion after 15
were 3.2 ( 1.6% for cyclohexanol, 86.9 ( 5.0% for min for reaction in HTW (0.34 g/cm3) at 380 °C. This
cyclohexene, and 1.5 ( 1.3% for methyl cyclopentenes. difference in the observed cyclohexanol reactivity can
With 7 mg of stainless steel filings, the product yields be easily explained, however. The reactor that Critten-
were 3.2% for cyclohexanol, 85.9% for cyclohexene, and don and Parsons used had a much larger thermal mass
3.0% for methyl cyclopentenes. With 22 mg of stainless than did the reactors that we used. As a result, the time
steel filings, the product yields were 4.7% for cyclohex- required to heat their reactor from room temperature
anol, 85.9% for cyclohexene, and 1.0% for methyl cyclo- to 375 ×bcC was much longer (>1 h). In fact, 20 min
pentenes. It is clear that the product yields are insen- after being placed in a tube furnace at 375 ×bcC, the
sitive to the presence of the added stainless steel filings. temperature inside their reactor was only 268 °C.14 Of
These results are consistent with negligible catalytic course, the average temperature of the reactor contents
effects due to the reactor walls. during the experiments would be much lower than 268
Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001 1827

°C. In contrast, the heat-up time is very short (2-3 min)


during our experiments. Another important distinction
is that, under the conditions employed by Crittendon
and Parsons, a large fraction of the reactor volume was
occupied by a vapor phase, so cyclohexanol could parti-
tion between the vapor and liquid phases. This phase
behavior could influence the reactivity. In contrast, we
conducted all of our experiments at subcritical temper-
atures in such a way that the reactor was nearly filled
with a single liquid phase. We speculate that the long
heat-up time and the partitioning of cyclohexanol
between the vapor and liquid phases contributed to the
apparent lack of reactivity of cyclohexanol in pure HTW
reported by Crittendon and Parsons.
The facile formation of cyclohexene that we observe Figure 9. Ion product for water at T ) 380 °C.
in the absence of added catalysts is consistent with the
results of Kuhlmann et al.,4 who reported that cyclo- experiments, Kw decreases from 10-11 to 10-12 (mol/kg)2
hexanol dehydration occurred readily in HTW at 300 as the temperature increases from 250 to 350 °C,18 yet
°C. They reported a cyclohexene yield of 33% at 60 min, the data show that the reaction rate increases with
which is lower, however, than the 67% yield listed in temperature. Again, the role of thermal activation is
Table 1 for the same conditions. Kuhlmann et al. did evident.
not report the formation of any other products. Thus, Alcohol dehydration is a classic organic reaction that
the present results, which identify 1-methyl cyclopen- has been studied extensively. The established mecha-
tene and 3-methyl cyclopentene as byproducts at 300 nisms for alcohol dehydration are E1cB, E1, and E2.
°C, albeit in small quantities, are also not in complete These notations stand for elimination (“E”) with a rate-
accord with the results of Kuhlmann et al. limiting step that is either unimolecular (“1”) or bimo-
Recall from the Experimental Section that we cooled lecular (“2”). A given alcohol can undergo dehydration
the reactors in a freezer after the experiments and by any one of these mechanisms, depending on the
completed the product recovery step quickly to minimize reaction conditions. The dominant mechanism is deter-
the loss of volatile compounds, i.e. cyclohexene and mined by various factors, including the acid/base prop-
methyl cyclopentenes. Exploratory experiments con- erties of the catalyst, the interactions of the reactant
ducted without cooling the reactors in the freezer and and intermediate species with the reaction medium, the
with a longer time (by ∼2 min) taken for the product reaction temperature, and the structural features of the
recovery step resulted in much lower cyclohexene yields alcohol.19-22
(30% after 45 min and 36% after 75 min) and no The balance of the acidity and basicity of the catalyst
recovery of methyl cyclopentenes. These results are fully often determines the dominant mechanism. In the
consistent with those reported by Kuhlmann et al. presence of a strong base and a weak acid, alcohol
Therefore, the likely source of the difference between dehydration occurs primarily via a carbanion intermedi-
the present results and those of Kuhlmann et al. is the ate, which is formed when the base abstracts a proton
possible loss of volatile components during their experi- from the β carbon (E1cB mechanism). In the presence
ments. of a strong acid and a weak base, alcohol dehydration
The paragraphs above provide a means of reconciling occurs primarily via a carbocation intermediate, which
the apparently contradictory results that had been is formed by the loss of the protonated hydroxy group
published for cyclohexanol dehydration in HTW. The (-OH2+) as water (E1 mechanism). In the E1cB and E1
discrepancies can be traced to the different experimental mechanisms, the (unimolecular) formation of the ionic
procedures and apparatuses that were used. intermediate is the rate-limiting step, and the subse-
quent formation of the alkene is facile. In the E2
Reaction Mechanism mechanism, alcohol dehydration occurs by β elimination,
in which the proton on the β carbon and the hydroxy
Previous studies have shown that HTW has a suf- group are eliminated in a concerted fashion. The reac-
ficiently high concentration of H3O+ and OH- ions that tion is bimolecular because a nucleophile is required to
some acid- and base-catalyzed organic reactions occur initiate elimination. Although the E2 mechanism is
in the absence of any added catalysts.5,6,15-17 Figure 9 typically observed in the presence of an acid and a base
shows Kw, the ion product for water, as a function of of balanced strength, the reaction is acid-catalyzed,
water density at 380 °C.18 The range of Kw is from 10-24 since the protonated hydroxy group is an excellent
(mol/kg)2 at 0.08 g/cm3 to 10-11 (mol/kg)2 at 0.68 g/cm3. leaving group.
This monotonic increase in the ion product, and hence In pure water, no base is present that is strong
in the H3O+ concentration, with increasing water den- enough to abstract a proton from the carbon backbone.
sity at 380 °C might account for the strong influence of Therefore, the E1cB mechanism is highly unlikely. The
the water density on the cyclohexanol conversion and remaining candidates are acid-catalyzed E1 and E2
product yields at 380 °C. Interestingly, the dehydration mechanisms, with H3O+ from water being the catalyst.
reaction occurs at 380 °C even at low water densities, Figure 10 shows the reaction network for acid-catalyzed
where Kw is many orders of magnitude smaller than its cyclohexanol dehydration. Formation of the oxonium ion
value for ambient liquid water [Kw ) 10-14 (mol/kg)2]. is the necessary first step for both the E1 and E2
This observation suggests that both thermal activation mechanisms. What distinguishes the two mechanisms
and acid catalysis are important for the rate enhance- is that the carbocation is the key intermediate in the
ment of cyclohexanol dehydration in HTW. At temper- E1 mechanism.
atures and water densities used in the liquid-phase There is only one previously reported attempt to
1828 Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001

Figure 10. Reaction network for cyclohexanol dehydration in


HTW.

elucidate the reaction mechanism of cyclohexanol de-


hydration. Whittaker and co-workers performed deute-
rium-labeling studies for the gas-phase dehydration of
cyclohexanol over zirconium phosphate.23,24 Dehydration
experiments at 350 °C using 2,2′,6,6′-[2H4]cyclohexanol Figure 11. Proposed reaction mechanism for cyclohexanol dehy-
as the starting material generated cyclohexene with dration in HTW.
significant deuterium scrambling. The authors attrib-
uted this outcome to the reaction intermediate being a the formation of carbocations. The experimental data
carbocation with highly mobile positive charge, in which in Table 1 support this hypothesis. Cyclohexene is the
case the dehydration mechanism would be E1. major product under all reaction conditions examined
There is some question, however, as to whether the in this study. At 250 and 275 °C, cyclohexene is the only
reaction mechanism observed for the gas-phase experi- product, but at higher temperatures, methyl cyclo-
ments would also operate in HTW. Solvation can play pentenes are also produced, in a yield that increases
an important role in reactions in water, modifying the with increasing temperature. At 380 °C, the methyl
reactivity of the species involved in ionic chemistry.19,25 cyclopentene yield also increases with increasing water
There are numerous examples of direct observation of density. If the secondary cyclohexyl carbocation is
carbocations by spectroscopic methods in (anhydrous) generated during cyclohexanol dehydration, as required
superacids,26 but not in weaker acids such as sulfuric in the E1 mechanism, it should quickly rearrange to the
acid solutions, because of the interactions of cations with tertiary methyl cyclopentyl carbocation, which is much
solvating molecules.25 This literature suggests, then, more stable.26 Thus, the absence of methyl cyclo-
that carbocations (R+) in water are solvated by water pentenes at 250 and 275 °C suggests that the E2
molecules and exist largely as oxonium ions (ROH2+),25 mechanism operates at these temperatures. The forma-
in which case the dehydration mechanism should be tion of methyl cyclopentenes at higher temperatures
predominantly E2. This hypothesis is supported by a suggests that the formation of carbocations becomes
recent ab initio study that showed that, in the presence increasingly more favorable as the reaction temperature
of room-temperature water, the oxonium ion is the increases. Because cyclohexene is consistently the major
stable intermediate, even for secondary (i-propyl) and product, however, the E2 mechanism probably remains
tertiary (tert-butyl) alcohols.19 A recent 13C NMR study the dominant dehydration mechanism, even at temper-
also showed that, even in a moderately concentrated atures of 300 °C and above.
solution of sulfuric acid, dehydration of tert-butyl alcohol On the basis of the experimental data reported herein
occurs via the oxonium ion intermediate and not the and supporting evidence from the literature as discussed
carbocation intermediate.27 in the preceding paragraphs, we propose the following
The spectroscopic evidences noted above were ob- scenario. Cyclohexanol dehydration to cyclohexene in
tained at room temperature, and there is always a pure HTW occurs through the E2 mechanism. Some of
possibility that a different mechanism operates at the cyclohexene molecules are then protonated to form
higher temperatures. Dabbagh and Mohammad Salehi28 the secondary cyclohexyl carbocation. This carbocation
observed a dramatic shift in the product distribution very quickly rearranges to form the tertiary methyl
with changes in temperature for the dehydration of 1,2- cyclopentyl carbocation, which then loses a proton to
diphenyl-2-propanol over alumina. At 200-260 °C, the form 1-methyl cyclopentene. Charge migration within
Hofmann (least branched) alkene is the major product, the methyl cyclopentyl carbocation can yield structures
whereas at 300 °C, the Saytzeff (most branched) alkene that deprotonate to form 3-methyl cyclopentene and
is the major product. The authors attributed this result methylene cyclopentane, both of which were observed
to a shift in the dehydration mechanism from E2 to E1 experimentally. Both the increase in temperature and
as the reaction temperature increased. Carrizosa and the isothermal increase in water density favor the
Munuera29 observed a similar shift in the product carbocation formation. Figure 11 shows the detailed
distribution with temperature (at about 100 °C) for reaction mechanism that we propose to account for
t-pentanol dehydration over TiO2. Thomke30,31 found cyclohexanol dehydration in HTW. Species appearing
that the mechanism of 2-butanol dehydration over ThO2 above the arrow in a reaction step are co-reactants, and
shifts from E1cB below 400 °C to E2 and even E1 above those appearing below the arrow are coproducts. The
400 °C. Although none of these studies was conducted mechanism described in Figure 11 is similar to the E2/
in aqueous environments, they suggested that the AdE3;E1/AdE3/Uni mechanism proposed by Antal et al.
formation of carbocations in water might also be favored to account for the reactions of 1- and 2-propanols in
at higher temperatures. HTW.7 This mechanism includes the conversion of
To summarize, this inspection of the relevant litera- 1-propanol to propylene via the E2 mechanism, followed
ture suggests that an oxonium ion intermediate (E2 by the conversion of propylene to a secondary carboca-
mechanism) would be anticipated for dehydration in tion, from which 2-propanol is formed.
HTW, but that the elevated temperature might admit The consideration of mechanistic issues related to
Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001 1829

cyclohexanol dehydration in HTW reveals that the role Table 3. Parameter Estimates for the Model in
of water in this reaction is threefold. First, water Equations 7-9 Optimized for Experimental Data at 380
participates in the elementary reaction steps as a °C and 0.68 g/cm3
reactant and as a product. Second, water is the source parameter value units
of H3O+, which catalyzes the formation of both cyclo- K1 2.05 × 106 L mol-1 s-1
hexene and methyl cyclopentenes. Third, water drives K2 4.39 × 103 L2 mol-2 s-1
the reaction mechanism toward E2, rather than E1, by K3 4.51 × 101 L mol-1 s-1
favoring, through solvation, the oxonium ion (“OXO6”
in Figure 11) rather than the carbocation (“CAT6” in This result suggests that the formation of the cyclohexyl
Figure 11) as the reaction intermediate. We expect cation from cyclohexene is essentially irreversible. One
water to make similar contributions to the dehydration could rationalize this outcome by recognizing that the
of other alcohols in HTW, although the extent to which formation of a more stable tertiary cation from a
solvation influences the reaction mechanism should secondary cation is highly favorable. It is likely that the
depend on both the nature of the intermediate species cyclohexyl cation rapidly rearranges to the tertiary
and the reaction conditions. methyl cyclopentyl cation as soon as it is formed via
cyclohexene protonation.
Mechanism-Based Model Taking one of the steps to be irreversible greatly
simplifies the kinetics model. If one assumes that the
Having proposed a mechanism for cyclohexanol de- formation of the cyclohexyl cation is in fact irreversible
hydration in HTW, we next sought to demonstrate that and, furthermore, employs the quasi-stationary state
the mechanism was consistent with the experimental approximation for the charged, reactive intermediates
results. Therefore, we constructed a detailed chemical (OXO6, CAT6, CAT5), one obtains the following system
kinetics model based on the reaction mechanism shown of equations from the reaction mechanism:

( )
in Figure 11. The model equations, which apply to
reactions in a constant-volume batch reactor, are d[NOL6] k01k12
)- [NOL6][H2O] +
dt k10 + k12

( )
d[NOL6]
) -k01[H3O+][NOL6] + k10[H2O][OXO6] k10k21
dt [ENE6][H2O][H3O+] (7)
(1) k10 + k12
d[OXO6] ) -K1[NOL6][H2O] +
) k01[H3O+][NOL6] -
dt K2[ENE6][H2O][H3O+]
(k10 + k12)[H2O][OXO6] + k21[H2O][H3O+][ENE6]
d[ENE5]
) k23[ENE6][H3O+] ) K3[ENE6][H3O+]
(2) dt
(8)
d[ENE6]
) k12[H2O][OXO6] - [ENE6] ) [NOL6]0 - [NOL6] - [ENE5] (9)
dt
(k21[H2O] + k23)[H3O+][ENE6] + k32[H2O][CAT6] where K1, K2, and K3 are lumped parameters. As one
can see, the number of parameters in the model is
(3) reduced from 10 to 3, which simplifies the parameter
estimation and reduces the likelihood that the data can
d[CAT6] be fit simply because there are many parameters rather
) k23[H3O+][ENE6] - than because the model is correct. We used Scientist to
dt
solve simultaneously eqs 7-9 and perform parameter
(k32[H2O] + k34)[CAT6] + k43[CAT5] (4)
estimation to obtain values for the parameters K1, K2,
and K3. We fit the model to a single concentration
d[CAT5] profile, which was acquired at 380 °C and 0.68 g/cm3.
) k34[CAT6] - (k43 + k45[H2O])[CAT5] +
dt Table 3 lists the optimized values for K1, K2, and K3
k54[H3O+][ENE5] (5) under these conditions.
If the proposed mechanism is correct, then the model
d[ENE5] in eqs 7-9 with the parameters in Table 3 should
) k45[H2O][CAT5] - k54[H3O+][ENE5] provide reliable predictions of the product yields at 380
dt °C and all other water densities. To test this expectation,
(6)
we used the parameter estimates to predict the product
We used a commercial modeling package, Scientist,32 yields at other water densities at 380 °C. Figure 12
to solve simultaneously the ordinary differential equa- compares the experimental data and the model predic-
tions above (eqs 1-6) and perform parameter estima- tions for reactions at 380 °C and 60 min as a function
tions to obtain the rate constants k01, k10, k12, k21, k23, of water density. The mechanism-based model is clearly
k32, k34, k43, k45, and k54. The minimization algorithm capable of predicting the strong water density effects
used by Scientist for parameter estimation is a Powell observed experimentally, even though the rate constants
variant of the Levenberg-Marquardt approach.33 We were optimized at only one density (0.68 g/cm3). Note
used the experimental concentration profiles acquired that the model employs no empiricism to account for
at 380 °C and water densities between 0.08 and 0.68 the water density dependence. The reaction kinetics
g/cm3 for parameter estimation. We fit the experimental depend on the water density in two ways. Water is both
data at each density independently and found that k32 a reactant and a product in different steps of the
was consistently near zero (∼10-15-10-20 L mol-1sec-1). mechanism, so the reaction rate is a direct function of
1830 Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001

Acknowledgment
We thank Michelle Osinski for performing prelimi-
nary experiments and Jianli Yu for assistance in the
laboratory. Financial support from the National Science
Foundation (CTS-9985456), the donors of the ACS
Petroleum Research Fund (34644-AC9), and the U.S.
Environmental Protection Agency (STAR Fellowship for
N.A.) is gratefully acknowledged.

Literature Cited
(1) Savage, P. E. Organic Chemical Reactions in Supercritical
Water. Chem. Rev. 1999, 99, 603-621.
(2) Savage, P. E.; Gopalan, S.; Mizan, T. I.; Martino, C. J.;
Figure 12. Comparison of experimental data and model predic- Brock, E. E. Reactions at Supercritical ConditionssApplications
tion for yields at T ) 380 °C and t ) 60 min. and Fundamentals. AIChE J. 1995, 41, 1723-1778.
(3) Crittendon, R. C.; Parsons, E. J. Transformations of Cyclo-
hexane Derivatives in Supercritical Water. Organometallics 1994,
the water concentration. Water is also the source of the 13, 2587-2591.
acid catalyst, H3O+, the concentration of which is a (4) Kuhlmann, B.; Arnett, E. M.; Siskin, M. Classical Organic
function of both temperature and water density.18 It is Reactions in Pure Superheated Water. J. Org. Chem. 1994, 59,
this latter effect that is the more important one in this 3098-3101.
(5) Xu, X.; Antal, M. J., Jr.; Anderson, D. G. M. Mechanism
model. This kinetics modeling exercise clearly demon- and Temperature-Dependent Kinetics of the Dehydration of tert-
strates that the proposed mechanism is consistent with Butyl Alcohol in Hot Compressed Liquid Water. Ind. Eng. Chem.
the experimental data. Res. 1997, 36, 23-41.
(6) Xu, X.; Antal, M. J., Jr. Kinetics and Mechanism of
Isobutene Formation from t-Butanol in Hot Liquid Water. AIChE
Conclusions J. 1994, 40, 1524-1534.
(7) Antal, M. J., Jr.; Carlsson, M.; Xu, X.; Anderson, D. G. M.
We studied cyclohexanol dehydration in HTW at Mechanism and Kinetics of the Acid-Catalyzed Dehydration of 1-
temperatures of 250-380 °C, reaction times of 15-180 and 2-Propanol in Hot Compressed Liquid Water. Ind. Eng. Chem.
Res. 1998, 37, 3820-3829.
min, and water densities of 0.08-0.68 g/cm3. Under (8) Xu, X. D.; DeAlmeida, C. P.; Antal, M. J. Mechanism and
these experimental conditions, cyclohexanol dehydrates Kinetics of the Acid-Catalyzed Formation of Ethene and Diethyl
readily in the absence of added catalysts to form Ether from Ethanol in Supercritical Water. Ind. Eng. Chem. Res.
cyclohexene as the major product. The most abundant 1991, 30, 1478-1485.
minor products are 1-methyl cyclopentene and 3-methyl (9) Gu, J. Thermodynamics and Conceptual Design of a Su-
percritical Ethylene from Ethanol Process. M. S. E. Thesis,
cyclopentene. Increasing temperature and water density University of Nebraska, Lincoln, NE, 1991.
increase the rate of cyclohexanol disappearance and the (10) Halvorsen, J. Process Evaluation for the Production of
selectivity toward methyl cyclopentenes over cyclohex- Ethylene from Ethanol. M. S. E. Thesis, University of Nebraska,
ene. These results provide further evidence that acid- Lincoln, NE, 1990.
catalyzed reactions can be accomplished readily in HTW (11) Lemmon, E. W.; McLinden, M. O.; Friend, D. G. Thermo-
physical Properties of Fluid Systems. In NIST Chemistry Webbook;
in the absence of added acid. The implications for Mallard, W. G., Linstrom, P. G., Eds.; NIST Standard Reference
environmentally benign industrial chemistry are clear. Database Number 69; National Institute of Standards and Tech-
The reaction mechanism for cyclohexanol dehydration nology: Gaithersberg, MD, February 2000 (http://webbook.nist-
in HTW comprises the following steps. Cyclohexanol .gov/chemistry/fluid/).
(12) ASPEN PLUS, release 9.3-1; Aspen Technology: Cam-
dehydrates reversibly to form cyclohexene by the E2 bridge, MA, September 1996.
mechanism. Cyclohexene is then protonated irreversibly (13) Lawson, J. R.; Klein, M. T. Influence of Water on Guaiacol
to form the cyclohexyl cation. This cation rapidly rear- Pyrolysis. Ind. Eng. Chem. Fundam. 1985, 24, 203-208.
ranges to form methyl cyclopentyl cations, which then (14) Holliday, R. L.; Jong, B. Y. M.; Kolis, J. W. Organic
lose a proton to form methyl cyclopentenes. A detailed Synthesis in Subcritical Water: Oxidation of Alkyl Aromatics. J.
Supercrit. Fluids 1998, 12, 255-260.
chemical kinetics model based on this mechanism (15) Lesutis, H. P.; Glaser, R.; Liotta, C. L.; Eckert, C. A. Acid/
captured the trends in the experimental data at 380 °C. Base-Catalyzed Ester Hydrolysis in Near-Critical Water. Chem.
The model was able to predict the striking effect of the Commun. 1999, 20, 2063-2064.
water density on the product yields, thereby supporting (16) Chandler, K.; Deng, F. H.; Dillow, A. K.; Liotta, C. L.;
the proposed reaction mechanism. Eckert, C. A. Alkylation Reactions in Near-Critical Water in the
Absence of Acid Catalysts. Ind. Eng. Chem. Res. 1997, 36, 5175-
We elucidated three roles for water in cyclohexanol 5179.
dehydration in HTW. First, water participates in the (17) Ikushima, Y.; Hatakeda, K.; Sato, O.; Yokoyama, T.; Arai,
elementary reaction steps as a reactant and as a M. Acceleration of Synthetic Organic Reactions Using Supercritical
product. Second, water is the source of H3O+, which Water: Noncatalytic Beckmann and Pinacol Rearrangements. J.
Am. Chem. Soc. 2000, 122, 1908-1918.
catalyzes the formation of both cyclohexene and methyl
(18) Marshall, W. L.; Franck, E. U. Ion Product of Water
cyclopentenes. Third, water drives the reaction mech- Substance, 0-1000 °C, 1-1000 bar: New International Formula-
anism toward E2 by favoring, through solvation, the tion and Its Background. J. Phys. Chem. Ref. Data 1981, 10, 295-
oxonium ion rather than the carbocation as the reaction 304.
intermediate. We expect water to make similar contri- (19) Uggerud, E.; Bache-Andreassen, L. Theoretical Models and
butions in the dehydration of other alcohols in HTW, Experimental Data for Reactions Between Water and Protonated
Alcohols: Substitution and Elimination Mechanisms. Chem. Eur.
although the extent to which solvation influences the J. 1999, 5, 1917-1930.
reaction mechanism should depend on both the nature (20) Luy, J. C.; Parera, J. M. Acidity Control in Alcohol
of the intermediate species and the reaction conditions. Dehydration. Appl. Catal. 1986, 26, 295-304.
Ind. Eng. Chem. Res., Vol. 40, No. 8, 2001 1831

(21) Halász, I.; Vinek, H.; Thomke, K.; Noller, H. Rate Deter- Alcohols over Aluminum Oxide. J. Org. Chem. 1998, 63, 7619-
mining Step in Alcohol Dehydration on La2O3, ThO2 and MoO3, 7627.
and Relations to Double Bond Shift in Olefins. Z. Phys. Chem. (29) Carrizosa, I.; Munuera, G. Study of the Interaction of
Neue Folge 1985, 144, 157-163. Aliphatic Alcohols with TiO2. II. On the Mechanism of Alcohol
(22) Winterbottom, J. M. Hydration and Dehydration by Het- Dehydration on Anatase. J. Catal. 1977, 49, 189-200.
erogeneous Catalysts. Catalysis 1981, 4, 141-174. (30) Thomke, K. Hydrogen-Exchange in Dehydration of Deu-
(23) Cruz Costa, M. C.; Hodson, L. F.; Johnstone, R. A. W.; Liu, terated 2-Butanols on BPO4 (Mechanism E1), Ba3(PO4)2 (Mecha-
J.-Y.; Whittaker, D. The Mechanism of Gas-Phase Dehydration nism E2) and ThO2 (Mechanism E1cB). Z. Phys. Chem. Neue Folge
of Cyclohexanol and the Methylcyclohexanols Catalysed by Zir- 1977, 105, 75-86.
conium Phosphate and Zirconium Phosphite. J. Mol. Catal. A (31) Thomke, K. Study on Mechanisms in Dehydration of
1999, 142, 349-360. Deuterated 2-Butanols on ThO2 (with Different Surfaces) and on
(24) Johnstone, R. A. W.; Liu, J.-Y.; Whittaker, D. Mechanism La2O3. Z. Phys. Chem. Neue Folge 1977, 105, 87-100.
of Cyclohexanol Dehydration Catalysed by Zirconium Phosphate. (32) Scientist, version 2.01; MicroMath, Inc.: Salt Lake City,
J. Chem. Soc., Perkin Trans. 2 1998, 6, 1287-1288. UT, 1995.
(25) Kazansky, V. B. Adsorbed Carbocations as Transition (33) Powell, M. J. D. A FORTRAN Subroutine for Solving
States in Heterogeneous Acid-Catalyzed Transformations of Hy- System of Nonlinear Algebraic Equations. In Numerical Methods
drocarbons. Catal. Today 1999, 51, 419-434. for Nonlinear Algebraic Equations; Robinowitz, P., Ed.; Gordon
(26) Olah, G. A. Carbocations and Electrophilic Reactions. and Breach Science Publishers: New York, 1970.
Angew. Chem. 1973, 12, 173-212.
(27) Kazansky, V. B.; Figueras, F.; de Ménorval, L. C. In Situ Received for review November 13, 2000
13C NMR Study of tert-Butanol Interaction with Moderately
Revised manuscript received February 15, 2001
Concentrated Sulfuric Acid. Catal. Lett. 1994, 29, 311-323. Accepted February 20, 2001
(28) Dabbagh, H. A.; Mohammad Salehi, J. New Transition-
State Models and Kinetics of Elimination Reactions of Tertiary IE000964Z

You might also like