Kinetics and mechanism of tetrahydrofuran synthesis via 1,4-butanediol dehydration in high-temperature water

Article abstract of DOI:10.1021/jo061017o

We conducted an experimental investigation into the kinetics and mechanism of tetrahydrofuran synthesis from 1,4-butanediol via dehydration in high-temperature liquid water (HTW) without added catalyst at 200-350 °C. The reaction was reversible, with tetrahydrofuran being produced at an equilibrium yield of 84% (at 200 °C) to 94% (at 350 °C). The addition of CO₂ to the reaction mixture increased the reaction rate by a factor of 1.9-2.9, because of the increase in acidity resulting from the formation and dissociation of carbonic acid. This increase was much less than that expected (factor of 37-60) from a previously suggested acid-catalyzed mechanism. This disagreement prompted experiments with added acid (HCl) and base (NaOH) to investigate the influence of pH on the reaction rate. These experiments revealed three distinct regions of pH dependence. At high and low pH, the dehydration rate increased with increasing acidity. At near-neutral pH, however, the rate was essentially insensitive to changes in pH. This behavior is consistent with a mechanism where H₂O, in addition to H₊, serves as a proton donor. This work indicates that the relatively high native concentration of H₊ (large K_w), which has commonly been thought to lead to the occurrence of acid-catalyzed reactions in HTW without added catalyst, does not explain the dehydration of 1,4-butanediol in HTW without catalyst. Rather, H₂O serves directly as the proton donor for the reaction.

Full text of DOI:10.1021/jo061017o

Kinetics and Mechanism of Tetrahydrofuran Synthesis via
1,4-Butanediol Dehydration in High-Temperature Water
Shawn E. Hunter, Carolyn E. Ehrenberger, and Phillip E. Savage*
Department of Chemical Engineering, The UniVersity of Michigan, Ann Arbor, Michigan 48109-2136
E-mail: psaVage@umich.edu.
ReceiVed May 17, 2006
We conducted an experimental investigation into the kinetics and mechanism of tetrahydrofuran synthesis
from 1,4-butanediol via dehydration in high-temperature liquid water (HTW) without added catalyst at
2
00-350 °C. The reaction was reversible, with tetrahydrofuran being produced at an equilibrium yield
of 84% (at 200 °C) to 94% (at 350 °C). The addition of CO to the reaction mixture increased the
2
reaction rate by a factor of 1.9-2.9, because of the increase in acidity resulting from the formation and
dissociation of carbonic acid. This increase was much less than that expected (factor of 37-60) from a
previously suggested acid-catalyzed mechanism. This disagreement prompted experiments with added
acid (HCl) and base (NaOH) to investigate the influence of pH on the reaction rate. These experiments
revealed three distinct regions of pH dependence. At high and low pH, the dehydration rate increased
with increasing acidity. At near-neutral pH, however, the rate was essentially insensitive to changes in
+
pH. This behavior is consistent with a mechanism where H
2
O, in addition to H , serves as a proton
+
donor. This work indicates that the relatively high native concentration of H (large K
W
), which has
commonly been thought to lead to the occurrence of acid-catalyzed reactions in HTW without added
catalyst, does not explain the dehydration of 1,4-butanediol in HTW without catalyst. Rather, H
directly as the proton donor for the reaction.
2
O serves
Introduction
THF can be obtained via several synthetic routes, with
acetylene and formaldehyde (Reppe process), propylene oxide,
n-butane, and furfural each serving as different feedstocks.1
The most commonly applied commercial route is the Reppe
Tetrahydrofuran (THF) is a cyclic ether used in a wide range
of industrial processes. Almost 400,000 metric tons of THF were
,2
1
4
produced in 2003. Roughly /5 of THF produced is polymerized
to form polytetramethylene glycol (PTMEG), which is used to
produce elastic (Spandex) fibers, polyurethane elastomers, and
1
,2
process, wherein THF is obtained directly from the acid-
catalyzed dehydration of 1,4-butanediol (BD). BD dehydration
is also a key reaction in the propylene oxide process, which is
copolyester-ether elastomers.1 The remaining /5 of THF is used
,2
1
1
1-3
also applied commercially. In other processes
a mixture of
as a solvent in the production of poly(vinyl chloride) (PVC)
cements and coatings, pharmaceuticals, magnetic tapes and
films, and in Grignard reactions.1,2 THF may also be used as
BD and THF is formed in the final step, and hence BD
dehydration can be used to increase overall selectivity to THF.
BD dehydration is thus a significant reaction for many industrial
processes that produce THF.
the starting material for producing succinic acid, adipic acid,
and γ-valerolactone.2
In the Reppe and propylene oxide processes, BD dehydration
is achieved in the presence of an acid catalyst at temperatures
*
Fax: +1 734 763 0459. Tel: +1 734 764 3386.
1) Ring, K. L.; Kalin, T.; Yokose, K. In Chemical Economics Handbook;
Stanford Research Institute: Menlo Park, CA, 2004; p 695.3000C-F.
2) Muller, H. In Ullmann’s Encyclopedia of Industrial Chemistry, 6th
ed.; Wiley-VCH: Weinheim, Germany, 2003; Vol. 35, p 671.
(
2
above 100 °C and near atmospheric pressure. Effective catalytic
(
(3) Bertola, A. U.S. Patent 6,248,906, 2001.
1
0.1021/jo061017o CCC: $33.50 © 2006 American Chemical Society
Published on Web 07/12/2006
J. Org. Chem. 2006, 71, 6229-6239
6229

Hunter et al.
agents include strong mineral acids, heteropolyacids, zeolites,7
4
5,6
8
,9
10
sulfonic acids, and dimethyl sulfoxide (DMSO). Strong
mineral acids, aluminum silicates, and ion-exchange resins are
the preferred commercial catalysts,
1,2,9
and THF is typically
produced at 95% yield.1
The acid catalysts in the current processes create undesired
environmental and economic burdens. For example, mineral
acids are corrosive and often necessitate the use of corrosion-
resistant materials. Additionally, these acids are typically
neutralized after the reaction, which consumes both the catalyst
and the neutralizing agent and generates waste salt. Solid acids
avoid the corrosion and neutralization issues of mineral acids,
but often require frequent regeneration or replacement owing
to catalyst deactivation.
FIGURE 1. Molar yield of 1,4-butanediol and THF in HTW at 200
A cleaner, more environmentally benign approach for syn-
thesizing THF from BD may be the use of high-temperature
liquid water (HTW) as the reaction medium. Liquid water at
elevated temperatures (T > 200 °C) has received much recent
attention as an alternative reaction medium for organic
°
C. CBD,0 ) 0.3 mol/kg. Curves represent the fit of eqs 3 and 4 to the
data.
BD dehydration in HTW has been investigated previously.
26
Richter and Vogel conducted BD dehydration experiments at
00-400 °C and residence times under three minutes using a
flow reactor. They found that THF could be produced in high
11-14
synthesis.
Under these conditions, water exhibits increased
3
solubility for small organic compounds typically thought of as
being insoluble in water.15 In addition, HTW can facilitate the
occurrence of some classically acid- and base-catalyzed reac-
(near 100%) selectivity from BD in HTW and supercritical water
(
SCW, T > 374 °C) without catalyst. They reported that the
1
1,12,16
tions, even in the absence of any added catalyst.
This
reaction was irreversible, but did not investigate conversions
greater than about 60%. Nagai et al. examined the dehydration
ability has often been attributed1
7-24
to the relatively large value
27
of the ion product KW in HTW, and associated high native
of BD using glass tube batch reactors, in both high-temperature
+
-
concentrations of H and OH ions, coupled with the elevated
temperature of HTW. Further, conducting the reaction in HTW
enables the possibility of accelerating the rate through the
liquid D2O at 150-365 °C and supercritical D2O at 385-400
°
C and with added mineral acid. On the basis of a kinetic
analysis, they concluded that a “water-induced” reaction was
significant for BD dehydration in neutral HTW. This conclusion
differed from that of Richter and Vogel, who proposed catalysis
1
2,25
addition of CO2, an environmentally benign additive.
Enhancement of BD dehydration via added CO2 could provide
competitive reaction rates for a HTW-based process, without
the environmental concerns associated with mineral acid.
+
by H as the dominant mechanism.
To further investigate the applicability of HTW as a reaction
medium for THF synthesis, and especially to elucidate the
mechanism of BD dehydration in HTW, we conducted a detailed
study of THF synthesis via BD dehydration in HTW. Here, we
present the results of THF synthesis experiments at 200-350
(
(
4) Hudson, B. G.; Barker, R. J. Org. Chem. 1967, 32, 3650-3658.
5) T o¨ r o¨ k, B.; Bucsi, I.; Beregsz a´ szi, T.; Kapocsi, I.; Moln a´ r, A. J. Mol.
Catal. A: Chem. 1996, 107, 305-311.
(
6) Baba, T.; Ono, Y. J. Mol. Catal. 1986, 37, 317-326.
(7) Rao, Y. V. S.; Kulkarni, S. J.; Subrahmanyam, M.; Rao, A. V. R. J.
°
C for batch holding times of up to 115 h. We also examine
Org. Chem. 1994, 59, 3998-4000.
8) Bucsi, I.; Moln a´ r, A.; Bart o´ k, M. Tetrahedron 1995, 51, 3319-3326.
the potential for added CO2 to accelerate the dehydration
reaction. We then explore the kinetics and mechanism of BD
dehydration in HTW, over a wide range of pH, and provide
Sci. 2001, 56, 2171-2178.
(
(
(
(
10) Moln a´ r, A.; Bart o´ k, M. HelV. Chim. Acta 1981, 64, 389-398.
+
evidence confirming that an H -catalyzed mechanism for BD
11) Savage, P. E. Chem. ReV. 1999, 99, 603-621.
12) Hunter, S. E.; Savage, P. E. Chem. Eng. Sci. 2004, 59, 4903-4909.
13) Broll, D.; Kaul, C.; Kramer, A.; Krammer, P.; Richter, T.; Jung,
dehydration is not the dominant mechanism in near-neutral
HTW. Rather, we demonstrate that H2O serves as the proton
donor and catalyst for BD dehydration in HTW without added
catalyst.
M.; Vogel, H.; Zehner, P. Angew. Chem., Int. Ed. 1999, 38, 2998-3014.
(
14) Akiya, N.; Savage, P. E. Chem. ReV. 2002, 102, 2725-2750.
(15) Chandler, K.; Eason, B.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem.
Res. 1998, 37, 3515-3518.
16) Kuhlmann, B.; Arnett, E. M.; Siskin, M. J. Org. Chem. 1994, 59,
098-3101.
17) Chamblee, T. S.; Weikel, R. R.; Nolen, S. A.; Liotta, C. L.; Eckert,
C. A. Green Chem. 2004, 6, 382-386.
18) Chandler, K.; Deng, F. H.; Dillow, A. K.; Liotta, C. L.; Eckert, C.
A. Ind. Eng. Chem. Res. 1997, 36, 5175-5179.
19) Chandler, K.; Liotta, C. L.; Eckert, C. A.; Schiraldi, D. AIChE J.
998, 44, 2080-2087.
20) Glaser, R.; Brown, J. S.; Nolen, S. A.; Liotta, C. L.; Eckert, C. A.
Abstr. Pap. Am. Chem. Soc. 1999, 217, U820-U820.
21) Lesutis, H. P.; Glaser, R.; Liotta, C. L.; Eckert, C. A. Chem.
Commun. 1999, 2063-2064.
22) Nolen, S. A.; Liotta, C. L.; Eckert, C. A.; Glaser, R. Green Chem.
003, 5, 663-669.
23) Patrick, H. R.; Griffith, K.; Liotta, C. L.; Eckert, C. A.; Glaser, R.
Ind. Eng. Chem. Res. 2001, 40, 6063-6067.
24) Sato, K.; Kishimoto, T.; Morimoto, M.; Saimoto, H.; Shigemasa,
Y. Tetrahedron Lett. 2003, 44, 8623-8625.
25) Hunter, S. E.; Savage, P. E. Ind. Eng. Chem. Res. 2003, 42, 290-
94.
(
3
Results and Discussion
(
This section provides information about the influence of
temperature, added CO2, and pH on the reaction kinetics. We
then use this information to draw inferences about the mech-
anism governing THF synthesis in HTW.
(
(
1
(
Kinetics. Figures 1, 2, 3, and 4 show yield versus time
profiles for experiments at 200, 250, 300, and 350 °C,
respectively. Under the conditions examined, BD dehydrates
in HTW without added catalyst to selectively yield THF. Figures
3 and 4 demonstrate that the reaction is reversible, with the
(
(
2
(
(
(26) Richter, T.; Vogel, H. Chem. Eng. Technol. 2001, 24, 340-343.
(27) Nagai, Y.; Matubayasi, N.; Nakahara, M. Bull. Chem. Soc. Jpn.
2004, 77, 691-697.
(
2
6230 J. Org. Chem., Vol. 71, No. 16, 2006

Mechanism of 1,4-Butanediol Dehydration in HTW
d[THF]
)
rTHF ) k [BD] - k [THF]
(2)
F
R
dt
These equations can be recast in terms of molar yield Yi ) Ci/
CBD,0 as
dYBD
)
-kFYBD + kRYTHF
(3)
(4)
dt
dYTHF
)
k Y - kRYTHF
F BD
dt
We fit the experimental yield versus time data in Figures 1-4
to eqs 3 and 4 to estimate kF and kR at each temperature. We
FIGURE 2. Molar yield of 1,4-butanediol and THF in HTW at 250
2
8
employed the modeling software Scientist to perform an
unweighted, nonlinear least-squares analysis. Table 1 lists the
estimated values of kF and kR at each temperature. Fitting the
data at 200 °C led to a value of zero for kR, because of the
relatively low conversion at this temperature. Thus, at 200 °C
we report only kF. The agreement between the model and
experimental data is shown at each temperature by the curves
in Figures 1-4. These figures demonstrate that the model
provides a good description of the data over the entire
experimental temperature range of 200-350 °C.
°
C. CBD,0 ) 0.3 mol/kg. Curves represent the fit of eqs 3 and 4 to the
autoclave data. Tubing bomb data are shown as open circles (BD yield)
and filled diamonds (THF yield).
Figure 2 displays the BD and THF yields obtained in
preliminary experiments at 250 °C with tubing bomb reactors.
The agreement of these data with the autoclave reactor data
demonstrates that consistent kinetics were obtained in both types
of reactors.
4
Hudson and Barker provide kF in 2 M HCl at 99 °C as 5.1
-
5
-1
×
10 s . Table 1 shows that kF is on the same order of
FIGURE 3. Molar yield of 1,4-butanediol and THF in HTW at 300
magnitude in HTW at 250 °C, and is 1 and 2 orders of
magnitude greater at 300 °C and 350 °C, respectively. Thus,
the rate of BD dehydration in HTW without added catalyst is
comparable and even superior to that in strong acid solutions
at lower temperature.
°
C. CBD,0 ) 0.4 mol/kg. Curves represent the fit of eqs 3 and 4 to the
data.
The concentration of H2O, which is a product of the reaction,
is lumped into the reverse rate constant kR. To remove the effect
of the temperature-dependent water concentration from the
estimated reverse rate constant, we also obtained estimates for
the second order reverse rate constant k′R, defined as k′R ) kR/
(H2O). Table 1 displays the estimated values of k′R.
Table 1 also provides the Arrhenius activation energy and
preexponential factor for each rate constant. Hudson and Barker4
report the activation energy for the dehydration of BD at 90-
-
1
27
1
23 °C in 2 M HCl to be 31 kcal mol . Nagai et al. report
the activation energy for the dehydration in HTW (D2O) without
-
1
catalyst to be 29 kcal mol . Our estimate of 32.8 ( 2.0 kcal
-
1
FIGURE 4. Molar yield of 1,4-butanediol and THF in HTW at 350
mol agrees well with these previous values.
°
C. CBD,0 ) 0.5 mol/kg. Curves represent the fit of eqs 3 and 4 to the
Having obtained rate constant estimates for the conversion
of BD to THF, we next estimated the equilibrium constant and
equilibrium yield of THF at each experimental temperature. The
equilibrium yield YTHF,EQ can be obtained by setting eq 3 equal
to zero, noting that YBD ) 1 - YTHF, and combining the two
equations to solve for YTHF:
data.
equilibrium yield of THF being greater than 90% at these
temperatures.
We performed a kinetic analysis to obtain forward (kF) and
reverse (kR) rate constants for the conversion of BD to THF.
For a reversible reaction with pseudo-first-order kinetics in a
constant-volume batch reactor, the following differential equa-
tions apply for BD and THF:
2
6
K
+ K
YTHF,EQ
)
(5)
1
where the equilibrium constant K is defined as K ) kF/kR )
YTHF/YBD. The equilibrium constants and equilibrium yields of
d[BD]
dt
)
rBD ) -k [BD] + k [THF]
(1)
F
R
(28) Scientist, version 2.01; MicroMath, Inc.: Salt Lake City, UT, 1995.
J. Org. Chem, Vol. 71, No. 16, 2006 6231

Hunter et al.
TABLE 1. Parameter Estimates for the Forward and Reverse Rate Constants for 1,4-Butanediol Dehydration in HTWa
pH
k
F
× 10 (s-1)
5
k
R
× 10 (s-1)
5
k′
R
× 10 (L mol s-1
7
-1
)
K
K′ (mol L-1
)
THF EQ
Y ,
T ) 200 (°C)
T ) 250 (°C)
T ) 300 (°C)
5.7
5.7
5.9
6.4
0.091 ( 0.005
2.9 ( 0.3
45 ( 3
405 ( 22
9.1 ( 0.8
32.8 ( 2.0
0.84b
0.90
0.92
0.94
0.31 ( 1.07
4.0 ( 1.5
27 ( 8
6.6 ( 5.2
28.8 ( 13.5
0.71 ( 2.5
10 ( 4
9.4
11
15
415
441
482
T ) 350 (°C)
84 ( 24
log A (s-
1
)
5.8 ( 2.8
31.0 ( 7.4
E
a
(kcal mol-1)
a
Uncertainties represent the 95% confidence interval. b The equilibrium THF molar yield at 200 °C was calculated based on an extrapolated value of kR,
which was obtained using the Arrhenius equation.
again decouple water concentration effects from the estimated
parameter. By performing linear regression of ln K′ vs 1/T data
-
1
at 250-350 °C, we estimated ∆Hrxn to be 1.0 kcal mol with
-
1
a standard error of 0.2 kcal mol .
To compare this experimental value for ∆Hrxn in HTW with
that at lower temperatures, we calculated the standard heat of
0
rxn
reaction ∆H for 1,4-butanediol dehydration in the liquid
phase as
0
rxn
0
f
0
f
0
f
∆H
) ∆H (THF, l) + ∆H (H O, l) - ∆H (BD, l) (7)
2
0
where ∆H (i, l) is the standard heat of formation of species i
f
0
f
0
f
in the liquid phase. We obtained ∆H (BD, l) and ∆H (H2O, l)
from Knauth and Sabbah
∆H (THF, l) has not been reported previously, so we esti-
3
0
31
and Chase,
respectively.
0
f
FIGURE 5. Equilbrium constant for BD dehydration in HTW as a
function of temperature.
mated its value from the standard heat of combustion
0
32
∆
H (THF, l) of THF using
c
THF appear in Table 1. The equilibrium THF yield is 84% at
2
00 °C and increases to 94% at 350 °C, indicating that the
0
f
0
c
0
f
∆
H (THF, l) ) -∆H (THF, l) + 4∆H (H O, l) +
2
reaction is endothermic.
0
Figure 5 shows a comparison of the equilibrium constants
estimated from the data in Figures 1-4 with those reported by
Nagai et al.27 for this reaction in HTW. The figure shows good
agreement between the two sets of data, with the previous data
being slightly lower than the new data reported here. One
explanation for this slight difference may be the experimental
technique. In the present work, liquid-phase samples were
acquired from a 440 mL reactor using a dip tube. Nagai et al.,
however, performed experiments using 1.0 mL quartz tubes.
The entire contents of those tubes were recovered for analysis,
and the tubes were loaded with a specified amount of water at
room temperature (for example, 0.5 mL at T < 350 °C). At the
water loadings specified in Nagai et al., a considerable amount
of vapor space would have occupied the 1.0 mL reactor at the
reaction temperature. For example, at 250 °C, the vapor phase
would have occupied roughly 40% of the reactor volume, based
on the density of saturated water.29 In the autoclave reactor used
in our work, a much smaller portion of the reactor volume
4∆H (CO , g) (8)
f 2
0
f
where ∆H (CO , g) is the standard heat of formation of gas-
phase CO , also obtained from Chase. Using eq 7, we
calculated the standard heat of reaction for BD dehydration in
the liquid phase to be 2 kcal mol . This value is close to the
experimental value of 1.0 kcal mol determined for the reaction
at higher temperatures.
Effect of Added CO . We performed 1,4-butanediol dehy-
dration experiments in HTW with added CO at 250 and 300
°C. Figure 6 shows that at both temperatures, the reaction was
accelerated by the addition of CO . For example, at 250 °C
and 60 min of reaction time, the experimental THF yield
increased from 11% to 25% with the addition of CO . At 300
°C, the THF yield in HTW at 20 min was 38%, but it increased
to 59% at only 18 min with added CO . Hence, CO addition
is effective for accelerating the synthesis of THF from 1,4-
butanediol dehydration in HTW.
2
3
1
2
-
1
-1
2
2
2
2
2
2
(
∼5%) was available for the coexisting vapor phase. Thus, there
To quantify the effect of added CO on the reaction, we
2
may have been more partitioning of organics into the vapor
phase during the Nagai et al. experiments, leading to a lower
observed BD conversion based on the total amount of BD
loaded.
To estimate the heat of reaction (∆Hrxn) for BD dehydration,
which has not been reported previously, we applied the van’t
Hoff equation. Assuming the heat of reaction to be independent
of temperature over the range investigated, one obtains
estimated k and k using the data in Figure 6. Table 2 provides
F
R
these rate constant estimates, obtained with Scientist using an
unweighted least-squares regression. We next compared the
experimentally observed increase in rate constant due to added
2
CO to that expected on the basis of the mechanism suggested
26
previously by Richter and Vogel for BD dehydration in HTW.
Scheme 1 illustrates their acid-catalyzed mechanism. The
reaction is initiated by protonation of an alcohol group by a
-
∆Hrxn
1
T
(29) Harvey, A. H. NIST/ASME Steam Properties, version 2.2; National
Institute of Standards and Technology: Boulder, CO, 1996.
ln K′ )
+ b
(6)
(
)
R
(
(
(
30) Knauth, P.; Sabbah, R. Struct. Chem. 1990, 1, 43-46.
31) Chase, M. W. J. Phys. Chem. Ref. Data Monograph 9 1998, 1-1951.
where K′ ) kF/k′R ) (THF)(H2O)/(BD) and b is a constant.
We utilized K′, rather than K, in our estimation of ∆Hrxn to
32) Cox, J. D.; Pilcher, G. Thermochemistry of Organic and Organo-
metallic Compounds; Academic Press: New York, 1970.
6232 J. Org. Chem., Vol. 71, No. 16, 2006

Mechanism of 1,4-Butanediol Dehydration in HTW
mechanism has also been advanced for this reaction in strong
4,33
acid solutions at lower temperatures. Taking the second step
to be rate determining and the other two to be in quasi-
4
equilibrium, as suggested by Hudson and Barker, leads to the
+
following rate law for the disappearance of BD via H catalysis:
+
+
-
rBD ) k [H ][BD] - k [H ][THF]
(9)
1
-1
where -rBD is the rate of disappearance of BD, and k1 and k-1
are the observable second-order forward and reverse rate con-
stants, respectively, for the reaction. In this analysis, we will
consider only the forward reaction for BD dehydration, given
by
+
rBDfTHF ) k [H ][BD] ) k [BD]
(10)
1
F
+
The pseudo-first-order rate constant kF (kF ) k1[H ]) should
+
be directly proportional to the concentration of H ions, for
the mechanism shown in Scheme 1.
To quantitatively assess the effect of added CO2 on the rate
of dehydration, we define a rate increase factor (RIF) as the
ratio of kF in the presence of CO2 to that without added CO2.
For the mechanism in Scheme 1, the expected increase in
+
rate with added CO2 can be estimated as the ratio of [H ] in
HTW with added CO2 to that in pure HTW. The pH of neutral
3
4
HTW is 5.7 at 250 °C and 5.9 at 300 °C. For the CO2-added
experiments, we estimated the pH to be 3.9 and 4.3 at the same
temperatures. On the basis of these pH values and the mecha-
nism in Scheme 1, one expects the rate to increase by factors
of 60 and 37 for the CO2-added experiments at 250 °C and 300
°C, respectively. The experimentally determined RIFs were only
2.9 and 1.9 though, as summarized in Table 2. Thus, the addition
of CO2 was far less effective than predicted by eq 10. These
FIGURE 6. Molar yields of 1,4-butanediol and THF in HTW with
added CO
THF in HTW without added CO
2
. The dashed lines represent the molar yields of BD and
, calculated using eqs 3 and 4 and
2
the parameters in Table 1: (a) 250 °C, CBD,0 ) 0.4 mol/kg, and the
average reactor pressure was 152 bar; (b) 300 °C, CBD,0 ) 0.4 mol/kg,
and the average reactor pressure was 183 bar.
+
results imply that the reaction is not first order in [H ] at these
TABLE 2. Comparison of Experimental and Expected Rate
Increase Factor (RIF) for the CO2-Added BD Dehydration
Experiments at 250 and 300 °C
temperatures, but rather exhibits an apparent reaction order less
than one. These results further imply that that Scheme 1 does
not accurately describe the reaction in HTW at near-neutral
conditions.
estimated
pH
exptl
RIF
expected
RIF
5
T (°C)
PCO (bar)
kF × 10 (s-1)
2
Effect of pH. To resolve the disagreement between the low
2
3
50
00
112
97
3.9
4.3
8.5 ( 0.8
85 ( 2
2.9
1.9
60
37
+
H reaction order implied by the added-CO2 experiments and
the mechanism-based order of unity expected from the literature,
we examined the effect of pH on the pseudo-first-order rate
constant for BD dehydration. We conducted BD dehydration
experiments in HTW using tubing bomb reactors with either
added HCl or NaOH at 200, 250, and 300 °C, and over a pH
range of 2.1 to 10.0. Table 3 summarizes these experimental
conditions and the results. Only BD and THF were observed
as reaction products, and their sum accounted for 94 ( 2% of
the moles of BD loaded into the reactors for all tubing bomb
experiments.
SCHEME 1. Acid-catalyzed Mechanism for 1,4-Butanediol
Dehydration with H Serving as the Catalyst4
+
For these experiments with added acid or base, we calculated
the pseudo-first-order rate constant kF as
1
t
k ) - ln(1 - X)
(11)
F
where X is the conversion of BD and t is the batch holding
time. Equation 11 is exact for an irreversible first-order reaction,
and it provides a reasonable estimate for a reversible first-order
reaction far from equilibrium. For example, using eq 11 to
hydronium ion. The resulting oxonium ion undergoes a rate-
limiting intramolecular nucleophilic substitution to liberate H2O
and form a protonated THF intermediate. The protonated ether
then quickly deprotonates to yield the product THF. This
(
33) Wisniewski, A.; Szafranek, J.; Sokolowski, J. Carbohydr. Res. 1981,
7, 229-234.
34) Marshall, W. L.; Franck, E. U. J. Phys. Chem. Ref. Data 1981, 10,
295-304.
9
(
J. Org. Chem, Vol. 71, No. 16, 2006 6233

Hunter et al.
TABLE 3. Summary of Experiments with Added HCl and NaOH
SCHEME 2. Acid-catalyzed Mechanism for 1,4-Butanediol
Dehydration with Water Serving as the Catalyst
log(kF/s-1
)
T (°C)
pH
t (min)
X
200
2.1
1060
1060
7260
0.57 ( 0.43
0.16 ( 0.10
0.15 ( 0.10
-5.0 ( 0.7
-5.6 ( 0.3
-6.5 ( 0.3
3.8
.1
7
250
3.2
15
20
180
180
990
0.53 ( 0.11
0.30 ( 0.02
0.49 ( 0.05
0.14 ( 0.06
0.10 ( 0.06
-3.1 ( 0.1
3.6
4.0
7.4
8.1
-3.53 ( 0.03
-4.2 ( 0.1
-4.9 ( 0.1
-5.8 ( 0.6
300
4.3
15
20
25
0.33 ( 0.04
0.51 ( 0.05
0.07 ( 0.11
0.019 ( 0.004
-3.4 ( 0.1
-3.2 ( 0.1
-4.4 ( 0.7
-5.8 ( 0.1
4
8
.9
.0
a
1
0.0
180
a
THF yield used as estimate of conversion.
CO2 is consistent with the rates observed with added mineral
acid at similar pH. This agreement supports the notion that the
increases in rate observed with added CO2 were due to lower
values of pH in those experiments, which resulted from the
formation and dissociation of carbonic acid. This agreement
further shows that the treatment of the kinetics data from the
autoclave and the tubing bombs led to consistent results.
Mechanism. To explain the effect of pH on log kF, we
examined the mechanism of BD dehydration. Scheme 1 displays
the classical acid-catalyzed reaction, wherein the catalytic action
is initiated by the transfer of a proton from a hydronium ion to
the reactant. However, in an aqueous medium, it is possible
that water may also serve as a proton donor. Participation by
2
7
water was previously suggested but without any mechanistic
details. We propose that water can act as the proton source to
catalyze the dehydration of BD, as shown in Scheme 2. This
mechanism proceeds through the same organic intermediates
as Scheme 1.
FIGURE 7. Effect of pH on pseudo-first-order rate constant for BD
dehydration in HTW at 200, 250, and 300 °C. Added-CO rate constants
2
are indicated as open symbols.
estimate kF for a reversible reaction with equilibrium yield of
0% from an experiment with 50% conversion overestimates
We use Schemes 1 and 2, together, to describe the pH
9
+
dependence of the reaction in HTW. Both H and H2O serve
kF and log kF by only 5% and 1%, respectively. Over the
temperature range of 200-300 °C, the equilibrium conversion
for BD dehydration varies from 84% to 92%. In these added-
acid and added-base experiments, the conversion ranged from
as proton-donating acid catalysts. Taking step 2 to be rate
determining and steps 1A, 1B, 3A, and 3B to be in quasi-
equilibrium leads to the pseudo-first-order rate equation below.
2
to 57% and was typically less than 40%. Thus, eq 11 provides
-
rBD ) k [BD] - k [THF]
(12)
a reliable estimate for log kF in these experiments.
F
R
Figure 7 illustrates the effect of pH on kF at 200, 250, and
3
00 °C. Over large changes in pH, log kF generally increases
where
and
with increasing acidity. Over smaller changes in pH, however,
the effect of added acid or base appears to vary depending on
pH regime. For example, at 250 °C, the rate constant displays
a strong dependence on pH at pH < 4 and at pH > 7. At near-
neutral pH (4 < pH < 7), however, the rate constant is less
strongly affected by added acid or base.
+
k [H ] + k [H O]
1
A
1B
2
kF ) k2
) k KF
2
(13)
k-1A + k-1B[OH-]
The behavior displayed by kF in Figure 7 is not consistent
with the reaction being first order in H , as required by the
+
+
k-3A[H ] + k [H O]
-3B
2
mechanism in the literature. Rather, the apparent reaction orders
kR ) k-2
) k K
(14)
-2
R
-
+
k3A + k [OH ]
3B
for H are 0.30, 0.44, and 0.44 at 200, 250, and 300 °C,
respectively. Moreover, Figure 7 shows that log kF is not a linear
function of pH, but rather depends on pH in a more complicated
manner.
To assess the consistency between the added CO2 and added
HCl experiments, which were done in two different types of
batch reactors and had rate constants determined two different
ways, we included the kF values from Table 2 in Figure 7. These
values show that the rate of dehydration observed with added
Positive and negative numerical coefficients of the rate constant
subscripts in eqs 13 and 14 indicate reaction in the clockwise
and counterclockwise directions, respectively, in Schemes 1 and
2. KF and KR represent quantities akin to equilibrium constants
for the quasi-equilibrated protonation of BD and THF, respec-
tively, and are herein referred to as apparent equilibrium
constants.
6234 J. Org. Chem., Vol. 71, No. 16, 2006

Mechanism of 1,4-Butanediol Dehydration in HTW
TABLE 4. Estimates for the Parameters in Equation 15a
for water in the hydrothermal cleavage of bisphenol A,35 where
water, rather than the native hydroxyl ion, serves as a proton
acceptor at near-neutral pH. The kinetic significance of the
water-catalyzed path for BD dehydration in HTW can be
appreciated by comparing the rate constant kF observed at neutral
T (°C)
log kA
log kB
log kC
200
250
300
-2.3 ( 0.5
-0.33 ( 0.06
0.87 ( 0.07
-6.01 ( 0.02
-4.54 ( 0.06
-3.34 ( 0.05
-6.8 ( 0.2
-7.1 ( 0.4
-7.6 ( 0.2
+
a
pH (Table 1) with the value of the H -catalyzed rate constant
Uncertainties represent one standard deviation. Units of kA, kB, and kC
+
-
1
-1 -1
-1
calculated at neutral pH, which is given by kA[H ] and is the
are L mol
s , s , and mol L , respectively.
hypothetical rate constant that would be observed if H2O did
not behave as a proton donor in the reaction. The observed rate
+
constant, kF, is 30-90 times greater than the hypothetical H -
catalyzed rate constant over this temperature range. This result
suggests that the dehydration of BD in HTW without catalyst
would be 1 to 2 orders of magnitude slower, if not for the ability
of H2O to transfer a proton directly to BD.
The pH dependence of the dehydration reaction in HTW can
+
be discussed in terms of H catalysis (reactions 1A and -1A
in Scheme 1) and H2O catalysis (reactions 1B and -1B in
Scheme 2) and can be described based on eq 13 as follows.
Under strongly acidic conditions, both the protonation and
+
+
deprotonation of BD are dominated by H catalysis (k1A[H ]
-
.
k1B[H2O] and k-1A . k-1B[OH ]). In this scenario, eq 13
+
+
simplifies to kF ) k2K1A[H ] ) kA[H ]. As the pH increases,
the protonation of BD by H2O becomes increasingly significant
FIGURE 8. Agreement between eq 15 and log k
00 °C.
F
vs pH data at 200-
+
3
compared to protonation by H . At near-neutral pH, the
protonation is dominated by H2O catalysis (k1B[H2O] . k1A-
+
Equation 13, an expression for kF based on the mechanisms
presented in Schemes 1 and 2, can be written as a three-
parameter equation that can be used to fit log kF versus pH data:
[H ]). The deprotonation, however, still occurs primarily via
+
-
the H -catalyzed path (k-1A . k-1B[OH ]). Under these
conditions, eq 13 simplifies to kF ) (k2k1B/k-1A)[H2O] ) kB.
As the pH increases further, deprotonation to BD via H2O
catalysis becomes faster than that via H catalysis (k-1B[OH ]
k-1A), because of the increase in OH concentration. Thus,
under strongly basic conditions, H2O catalysis dominates the
+
-
+
k [H ] + k
A
B
-
.
kF )
(15)
kC
1
+
+
-
+
reaction, (k1B[H2O] . k1A[H ] and k-1B[OH ] . k-1A), and
[
H ]
+
eq 13 can be simplified to kF ) (k2K1B/Kw)[H2O][H ] ) kB-
+
[
H ]/kC. This analysis suggests that the effect of pH on the
where
dehydration reaction is to alter the form of the apparent
equilibrium constant KF, rather than to promote fundamentally
k1A
36,37
different reaction mechanisms
or changes in rate determining
kA ) k
(16)
(17)
2k
38
-1A
step, as observed in other systems. Table 5 summarizes the
effect of pH on the apparent equilibrium constant and dehydra-
tion reaction.
k1B
kB ) k
[H O]
2
2k
Implications for CO2 Addition. The addition of CO2 to
HTW may be an environmentally benign method of achieving
increased rates of acid-catalyzed reactions, and may therefore
be a technique for making HTW more industrially competitive
as a reaction medium. The curves in Figure 8 suggest that, for
BD dehydration in HTW, the addition of sufficient quantities
of CO2 to the reaction medium will lead to increased rates of
reaction. Quantitatively, however, these data suggest that
significant increases in the reaction rate will only be achieved
with added CO when the addition of CO yields a pH that is
-
1A
and
k-1B
kC ) KW
(18)
k-1A
We fit eq 15 to the log kF versus pH data at 200-300 °C,
including data from the added-CO2 experiments, using Scientist.
We performed a weighted, nonlinear-least-squares fit, using the
reciprocal of the variance of each point as the statistical weight.
Table 4 provides the resulting estimates for kA, kB, and kC at
each temperature. Figure 8 shows that eq 15 provides a good
description of the experimental data over the entire range of
pH and temperature. Additionally, Figure 8 shows that the
reaction exhibits general acid catalysis at near-neutral pH, and
specific acid catalysis at high and low pH.
2
2
at least 2 to 3 units lower than the neutral pH. Hence, at
moderate CO loadings (P
2
CO2
< 100 bar), CO addition may
2
be ineffective for achieving significant rate increases for BD
dehydration in HTW. High pressures (100 < PCO2 < 1000 bar)39
(
(
35) Hunter, S. E.; Savage, P. E. J. Org. Chem. 2004, 69, 4724-4731.
36) Buurma, N. J.; Blandamer, M. J.; Engberts, J. B. F. N. J. Phys.
Org. Chem. 2003, 16, 438-449.
37) Salvestrini, S.; Di Cerbo, P.; Capasso, S. J. Chem. Soc., Perkin
Trans. 2 2002, 1889-1893.
38) Fett, R.; Simionatto, E. L.; Yunes, R. A. J. Phys. Org. Chem. 1990,
, 620-626.
39) Hunter, S. E. Ph.D. Thesis, The University of Michigan, Ann Arbor,
MI, 2005.
(
The ability of eq 15 to capture the trends in the data at each
temperature suggests that water, in addition to the hydronium
ion, functions as a proton donor in the dehydration of BD in
HTW. Similar acid/base behavior has been noted previously
(
3
(
J. Org. Chem, Vol. 71, No. 16, 2006 6235

Hunter et al.
TABLE 5. Summary of Effect of pH on 1,4-Butanediol Dehydration in HTW
dominant
BD
protonation
step
dominant
BDH
deprotonation
step
apparent
equilibrium
constant
+
H+ reaction
pH
low
expression
simplified kF
order
k1A[H+]
+
k ) k [H ]
F
A
1A
1B
-1A
-1A
K )
1st
F
k-1A
k [H O]
kF ) kB
1B
2
intermed
high
0th
1st
K )
F
k-1A
k [H O]
kB
1B
2
+
1B
-1B
K )
kF ) [H ]
F
k-1B[OH-]
kC
are required to achieve the low pH conditions where the reaction
rate will be significantly greater than the neutral reaction rate.
These high-pressure conditions may render CO2 addition an
impractical method of achieving accelerated rates of BD
dehydration in HTW.
1-11 and a temperature range of 50-350 °C, calculated using
eq 15 and the parameters in Table 6.
At each temperature, the intermediate pH regime, as defined
in Table 5, can be estimated as having lower and upper bounds
+
+
of [H ] ) kB/kA and [H ] ) kC, respectively. These bounds
are plotted in Figure 12. As shown in the figure, this intermediate
regime expands and shifts to higher pH with increasing
temperature. Additionally, Figure 12 shows the neutral pH of
H2O at each temperature. At low temperature, the neutral pH
is greater than the upper bound of the intermediate regime. As
temperature increases and KW approaches its maximum, the
neutral pH moves within the intermediate regime. The location
of the intermediate region with respect to the neutral pH is
Comparison with the Literature. Nagai et al. proposed eq
+
1
9 for BD dehydration based on their kF versus [H ] data,
+
kF ) kacid[H ] + k
(19)
water
+
where kacid is the H -catalyzed rate constant and kwater is the
“
water-induced” rate constant.27 Empirical eq 19 can be
recovered from mechanism-based eq 15 as a limiting case. When
-
reaction at high pH is neglected (k-1B(OH ) , k-1A), then kF
+
)
kA(H ) + kB.
Equation 19 implies that a linear relationship should exist
+
between kF and the H concentration. The data in Figure 8 and
+
eq 15 demonstrate that kF is only a linear function of [H ] for
the limiting case of intermediate and acidic pH, the region
explored by Nagai et al. Over the larger range of pH examined
+
here, kF depends on [H ] in a more complicated manner. When
+
considering acid/base-catalyzed reactions, where the H and
-
OH concentrations and observed rate constant can vary by
several orders of magnitude, it is advantageous to examine kF
+
40
versus [H ] data on a logarithmic scale. Significant informa-
tion regarding acid/base-catalyzed reactions in HTW can be
obtained by examining the observed rate constant over a wide
range of pH, including acidic-, basic-, and neutral-pH conditions.
We compared the data obtained for kA and kB in this
investigation with that of Nagai et al.27 and Hudson and Barker.
Nagai et al. provide kA and kB in HTW at 150-290 °C, whereas
Hudson and Barker provide kA at 99 °C. Figure 9 shows good
consistency among the kA data from 99 to 300 °C. Additionally,
Figure 10 demonstrates consistency between the kB data of Nagai
et al. and that of the present work, over a temperature range of
4
A
FIGURE 9. Temperature dependence of k .
1
50-300 °C. Figure 11 demonstrates an inverse relationship
of kC with temperature, which is permissible for such a lumped,
or composite, rate constant. We performed linear regression (ln k
vs 1/T) of the data in Figures 9-11 to obtain the apparent
activation energy and preexponential factor of each lumped rate
constant. Table 6 displays the estimated values.
Having obtained Arrhenius parameters for the rate constants
in eq 15, we next used the mechanism-based model to examine
the behavior of BD dehydration as a function of pH over a wide
range of temperature. Figure 12 shows kF over the pH range of
(
40) Gates, B. C. Catalytic Chemistry; John Wiley & Sons: New York,
1
992.
B
FIGURE 10. Temperature dependence of k .
6236 J. Org. Chem., Vol. 71, No. 16, 2006

Mechanism of 1,4-Butanediol Dehydration in HTW
FIGURE 13. Schematic representation of autoclave batch reactor used
for neutral HTW and CO -added BD dehydration experiments.
2
CO2 will accelerate the rate. The reaction follows pseudo-first-
order kinetics in reactant and product, and equilibrium THF
yields are greater than 90% at temperatures above 250 °C. The
dehydration reaction is endothermic, exhibiting a heat of reaction
of 1.0 kcal/mol.
C
FIGURE 11. Temperature dependence of k .
TABLE 6. Arrhenius Parameters for kA, kB, and kCa
rate
constant
log A
EA(kcal mol-1)
The use of HTW for the synthesis of THF via 1,4-butanediol
dehydration has potential advantages over commercial synthesis
in strong acid solutions or over solid acid catalysts. In HTW,
no catalyst is necessary for the reaction, and dehydration rates
exceed those at lower temperature in the presence of strong
mineral acid catalyst. The equilibrium yield of THF (>90%) is
comparable with that typically achieved in the Reppe and
propylene oxide processes (95%). Hence, THF synthesis in
HTW may be a green and practical alternative to conventional
catalytic processes.
kA
kB
kC
12.2 ( 2.4
6.1 ( 2.8
29.3 ( 5.4
25.1 ( 6.4
-9.8 ( 21.7
-11.3 ( 9.2
a
Units of kA, kB, and kC are L mol-1 s-1, s-1, and mol L-1, respectively.
The dehydration of 1,4-butandediol occurs with a pH-
independent rate at near-neutral pH in HTW. In contrast, the
+
dehydration is first-order in H concentration at strongly acidic
or strongly basic pH. This behavior is consistent with an acid-
+
catalyzed dehydration mechanism, where not only H , but also
+
H2O, serve as proton donors. At low and high pH, H and H2O
serve as the primary catalysts, respectively. However, at near-
neutral pH, butanediol is protonated via reaction with H2O
(
water catalysis), and deprotonation of the protonated butanediol
+
+
produces H (H catalysis). Hence, the elevated values of KW
for water at high temperatures, and the associated H concentra-
+
tions, are not primarily responsible for the occurrence of BD
dehydration in HTW in the absence of added catalyst. Rather,
water itself plays the more important role as a proton donor.
This work thus highlights the importance of verifying the role
of KW in acid- and base-catalyzed reactions in neutral HTW
through experimentation with added acid and base over a wide
range of pH.
FIGURE 12. Effect of pH and temperature for BD dehydration in
O at elevated temperatures. The intermediate pH range, as defined
H
2
in Table 5, is located at each temperature between the two dashed lines.
Each dashed line represents the pH around which the reaction transitions
+
from 1st to 0th order in H .
Experimental Section
significant when attempting to use pH to tune the reaction to
achieve higher rates. At lower temperatures, the low pH regime,
as defined in Table 5, begins at lower pH, and is further away
from neutral pH. At higher temperatures, the low pH regime
begins at higher pH, and is generally closer to the neutral pH.
Hence, the effect of decreasing pH (e.g., adding acid) on the
rate will be a function of temperature and for BD dehydration
in H2O will generally increase with increasing temperature.
Materials. Deionized water was obtained from an in-house water
purification system, which contained water softener (ion exchange),
reverse osmosis, high-capacity ion exchange, UV sterilization,
ultrapure ion exchange, and submicron filtration units. BD, THF,
0.2 M HCl, and 0.5 M NaOH were obtained from a commercial
supplier. All chemicals were used as received.
Reactors. We conducted THF synthesis experiments in the two
different batch reactor systems described in detail below.
1
. Autoclave Reactor. To obtain yield versus time profiles in
Conclusions
HTW and in HTW with added CO , we employed a 440 mL
2
hastelloy batch reactor from a commercial supplier. Figure 13
1
,4-Butanediol dehydrates reversibly in HTW at 200-350
C to selectively yield tetrahydrofuran. No added catalyst is
required for the reaction, although addition of mineral acid or
depicts the autoclave apparatus. In this system, deionized water
was first loaded into the reactor body, and the reactor was sealed.
The quantity of water loaded was determined such that the expanded
°
J. Org. Chem, Vol. 71, No. 16, 2006 6237

Hunter et al.
reaching equilibrium, they can be viewed as multiple samples at
the same conditions. The average YTHF was equal to 0.929 with a
95% confidence interval of 0.003. According to this analysis, then,
the uncertainty for molar yields obtained from the autoclave reactor
can be estimated as less than 0.01 at the 95% confidence level.
The measured value of C
initial BD concentration CBD,0, which was calculated as the number
of moles of BD loaded divided by the mass of H O loaded. This
result indicates that nearly all of the BD resided initially in the
liquid phase. Additionally, C was time invariant at 200 and 250
C, but it decreased by roughly 15% at 300 °C and roughly 30%
at 350 °C over the course of the reaction. This decrease in C is
T
at short times was 95 to 99% of the
2
T
°
FIGURE 14. Exploded-view illustration of septum-containing attach-
ment utilized during autoclave sample recovery.
T
not due to the formation of additional reaction products. We
observed no other product peaks in the chromatograms, and no
reported the formation of species other
than BD or THF at temperatures up to 350 °C. Rather, the decrease
previous investigator2
6,27
water volume at reaction temperature occupied roughly 95% of the
reactor volume, based on the saturation density of pure liquid
water.29 To degas the water, the reactor headspace was filled with
in C at the higher temperatures is likely due to the transfer of
T
2
roughly 69 bar of N , agitated for 15 min at roughly 800 rpm with
organic species, especially the more volatile THF, from the liquid
an internal stirrer, and then depressurized. This process was repeated
at least two additional times.
phase to the vapor phase during the experiment. The larger change
in C at higher temperatures is a reflection of the higher specific
T
When required, CO
through a port in the top of the reactor head. The initial (nonequi-
librium) partial pressure of CO in the reactor just after loading
was roughly 55 bar, and it decreased with time before heating, as
the gas dissolved in the water. To promote dissolution of the CO
2
was next loaded into the reactor headspace
volume of H O at those conditions, a larger number of samples
2
and sample line rinses being taken during those reactions, and more
THF being present. Because the reaction is first order in BD and
2
THF concentration, though, and because the decrease in C was
T
2
minimal during acquisition of the first 6-8 samples, we do not
in the water, the reactor agitator was set to roughly 1000 rpm during
the reactor heating process. The heat-up time was always greater
expect these decreases in C to affect the observed reaction kinetics.
T
This expectation was verified by the kinetics determined in tubing
than 1 h, which provided adequate time41 to reach CO
vapor-liquid equilibrium prior to initiating the reaction.
-H
O
2
2
bomb reactors, which did not suffer from this sampling-induced
potential problem and yielded results consistent with those deter-
mined from autoclave experiments.
The reactor was next heated to the desired reaction temperature,
which was stable to within 3 °C of the set point in all experiments
and typically did not vary by more than 2 °C during a reaction. To
initiate the reaction, roughly 10 mL of BD were loaded into the
hot, pressurized reactor through a dip tube extending into the reactor
body using a syringe pump set to operate in constant pressure mode
at 34-69 bar greater than the reactor pressure. The end of this
loading procedure, which took 5-10 s, was regarded as the initial
time (t ) 0) for the reaction. The reaction mixture was agitated
throughout the reaction at roughly 1000 rpm.
Liquid phase samples were removed from the reactor periodically
through a dip tube and collected in stainless steel sample bulbs
that had an internal volume of roughly 0.9 mL. To ensure that the
sample acquired was representative of the reactor contents at the
time at which it was taken, at least one presample was taken just
prior to the desired sample time to purge the sampling line of any
residual material. All sample bulbs were cooled in a refrigerator
for at least 30 min before opening to maximize product recovery.
To recover the reactor samples for analysis, sample bulbs were
first depressurized using a septum-containing attachment, repre-
sented in Figure 14. The device was attached to the sample bulb,
and the bulb contents were allowed to expand into a 5 or 10 mL
syringe, containing roughly 1 mL of acetone, inserted through the
septum in the attachment. This process was repeated until the bulb
was no longer pressurized. The bulb was then inverted, and opened.
Remaining bulb contents were rinsed with acetone. All acetone
rinses, including the solutions obtained during depressurization,
were collected in a 10 mL flask. After recovery, the flasks were
filled to the 10 mL line with solvent, and samples were removed
2. Tubing Bomb Reactors. For experiments with added acid
and added base, and for preliminary THF synthesis experiments,
we used 0.6 mL tubing bomb reactors. Detailed information
concerning the reactor assembly, reaction procedure, and product
42
recovery appeared in a previous paper. Roughly 15 mg of BD
was carefully measured and then loaded into each tubing bomb
reactor. The required amounts (3-180 µL) of HCl or NaOH
solutions were added after loading degassed water. Tubing bomb
reactors that had been used for HTW experiments with added acid
or base exhibited residual catalytic effects when later used to
perform experiments without added catalyst. Hence, once a set of
reactors had been exposed to strong acid, that set was exclusively
used for experiments with added strong acid, and likewise for those
reactors with added base. Uncertainty for results obtained using
these tubing bomb reactors was determined as the 95% confidence
interval, on the basis of at least three independent experiments
conducted under identical conditions.
Product Analysis. We used a gas chromatograph, equipped with
a flame ionization detector, autosampler, and 50 m × 0.2 mm ×
0.33 µm HP-5 capillary column, for product analysis. The temper-
ature program comprised a 9 min soak at 70 °C, followed by a 70
C/min ramp to 240 °C, which was held for 9 min. Products were
°
identified by matching retention times with those of known
standards and quantified using calibration curves, which were
calculated from the chromatographic results of four to six standards.
Calculation of pH. The pH of neutral HTW and HTW with
34
added acid or base was calculated using literature values for K
W
43
44
and recognizing that HCl and NaOH are essentially completely
dissociated in HTW. To estimate the pH in the added-CO
from the flask and placed in vials for analysis. The molar yield Y
of species i was calculated as Y ) C /C , where C is the molal
concentration of species i and C is the total concentration of organic
species in the reactor. This quantity was taken to be equal to CBD
THF since the reactant and product were the only species detected
i
2
25
i
i
T
i
experiments, we applied the development of Hunter and Savage,
wherein the reactor contents are modeled as a two-phase, two-
component mixture with vapor-liquid equilibrium between H
and CO . Recognizing that the second dissociation of carbonic acid,
T
2
O
+
C
2
throughout the reaction.
We calculated the 95% confidence interval for the yield of THF
using four samples that were obtained at 350 °C after the reaction
had reached equilibrium. Because these samples were taken after
(
42) Hunter, S. E.; Felczak, C. A.; Savage, P. E. Green Chem. 2004, 6,
2
22-226.
(
43) Frantz, J. D.; Marshall, W. L. Am. J. Sci. 1984, 284, 651-667.
(44) Chen, X. M.; Gillespie, S. E.; Oscarson, J. L.; Izatt, R. M. J. Solution
(41) Takenouchi, S.; Kennedy, G. C. Am. J. Sci. 1964, 262, 1055-1074.
Chem. 1992, 21, 803-824.
6238 J. Org. Chem., Vol. 71, No. 16, 2006

Mechanism of 1,4-Butanediol Dehydration in HTW
vapor- and liquid-phase nonidealities, and the influence of the small
concentration. In the autoclave experiments, the value of f changes
as samples are withdrawn. In the experiments presented here,
however, the effect of sample withdrawal on pH was less than 0.05
pH units.
2
amount of dissolved CO on the aqueous-phase density are
negligible,39 one can estimate the molal H concentration according
+
to
In this paper, we express the concentration of H+ in units of
molarity to be consistent with all other concentrations used in the
T
CO2
C
KW
Ka1
+
+
kinetics analysis. Thus our definition of pH herein is as -log [H ],
(
H ) )
+
(20)
+
(
+
)
1
- f
(
H )
(H )
rather than the more rigorous definition on the molal scale as pH
F
f +
H O
2
+
(
H O)
) -log (H ), where the square and curved brackets represent molar
2
RT
and molal concentrations, respectively. The density of liquid water
at high temperatures is less than 1 kg/L, and hence molal and molar
concentrations are not numerically equal in HTW as they are in
ambient liquid water. The quantities are related as pHmolar ) pHmolal
H
4
5
where Ka1 is the first dissociation constant of carbonic acid,
F
2
H O
29
is the density of saturated liquid water, H is the Henry’s constant
T
-
log FH2O (kg/L).
46
for CO
2
,
C
is the total number of moles of CO
2
added to the
CO2
reactor divided by the reactor volume, f is the fraction of the reactor
volume occupied by the aqueous phase, R is the gas constant, T is
absolute temperature, and the curved brackets represent molal
Acknowledgment. We thank Brendon Webb for his as-
sistance in conducting tubing bomb reactor experiments. We
also acknowledge the financial support from the ACS-PRF
(
Grant 37603-AC9) and NSF (graduate fellowship to S.E.H.
(
45) Patterson, C. S.; Slocum, G. H.; Busey, R. H.; Mesmer, R. E.
Geochim. Cosmochim. Acta 1982, 46, 1653-1663.
46) Crovetto, R. J. Phys. Chem. Ref. Data 1991, 20, 575-589.
and Grant CTS-0218772).
(
JO061017O
J. Org. Chem, Vol. 71, No. 16, 2006 6239