152
J. S. Zhang et al.
induced by the electron withdrawing effect of the acid
group and thus can even undergo spontaneous rearrange-
ment without Brønsted acids [20, 21]. On the other hand,
caprolactam is more facile to be obtained by distillation or
solvent extraction due to its weak interaction with the acid.
In the early studies, acetic anhydride and acetyl chloride
have been employed as promoters in the oxime rear-
rangement [22]. Recently, progresses have been made on
the homogeneous Beckmann rearrangement using organ-
ocatalysis, such as cyanuric chloride [23, 24], sulfamic acid
[25], and trifluoroacetic acid [26, 27]. Excellent conver-
sion and selectivity were obtained. In particular, trifluoro-
acetic acid (TFA) can be a promising catalyst for its low
boiling point (345.5 K) and its facile separation
with amides by distillation [26, 27]. However, the low
selectivity (83–95 %) and slow reaction rate (1.2 9
10-5 mol L-1 s-1, 343 K) compared with the industrial
process still present a significant challenge.
heated magnetic stir-plate (IKA RCT basic). The vial is
closed by a bottle cap which has a cushion for sampling.
The hermetical environment can approximately ensure a
constant system pressure and the vapor phase composition
during the reaction process. TFA as the catalyst and solvent
has a majority of the reaction mixture and the volatilization
of TFA would bring nearly no effect on the reaction. The
TFA and additives (total about 15 mL) were first mixed in
the glass vial and then heated to the reaction temperature.
The COX was then added into the vial and the reaction
time was recorded with a stopwatch. Small amounts of the
reaction mixtures (80lL) were drawn at different time and
immediately diluted by large amounts of acetonitrile
(1 ml). Ronchin et al. [27] described the kinetics of the
rearrangement by TFA in acetonitrile at 90 °C. The con-
version of COX reached only 15 % with the reaction time
of 50 min and TFA concentration of 1 mol L-1. Thus we
can figure out that the reaction in the mixtures (TFA con-
centration: 0.9 mol L-1) is stopped after the dilution of
acetonitrile at room temperature. All the samples were
measured by gas chromatography (Shimadzu GC-2014)
with a flame ionization detector under following condi-
tions: the injection temperature, 260 °C; the column tem-
In our previous work, microreactors have been applied
to carry out the Beckmann rearrangement of COX with
oleum as the catalyst [28, 29]. 100 % conversion of COX
and higher than 99 % selectivity to CPL were reached with
the reaction time less than 40 s and molar ratios of acid to
COX as low as 0.8. It is much lower than the industrial
value of 1.2–1.7. However the side product of ammonium
sulfate has not been avoided completely. In this work, a
modified catalytic system, in which TFA acts as the cata-
lyst and solvent and organic solvents act as the additives,
was employed in the Beckmann rearrangement of COX to
CPL. The effect of different organic solvent and catalyst
composition on the reaction rate and the selectivity to CPL
were investigated. In addition, an immiscible two-phase
system was proposed to study the side reaction of oxime
hydrolysis which determined the selectivity. Based on the
results, a simplified reaction process was suggested and a
mathematical kinetic model was developed to represent the
reaction process. The reaction rate constants and activation
energy of the kinetic model were given.
perature, 140–260 °C, 30 °C min-1
;
the detector
temperature, 280 °C. The sample volume for all analysis
was 1 lL. Dimethyl terephthalate was used as the internal
standard. The concentrations of cyclohexanone (main side
product, CYC), COX and CPL could be determined by this
internal standard method within the measurement errors of
0.8, 1 and 0.5 %, respectively. The ester intermediates
(including trifluoroacetyl COX and trifluoroacetyl capro-
lactam) concentrations could not be quantified due to the
high interaction with TFA and they were calculated by the
mass balance. By this method, the concentration of esters
was determined within a measurement error of 1.5 %.
3 Results and Discussion
3.1 Modified Catalytic System Based on TFA
2 Experimental
3.1.1 The Influence of Additive on the Reaction Rate
and Selectivity
2.1 Materials
Reagents (COX, CPL and TFA) were purchased from J&K
Scientific Ltd. (Beijing). All the organic additives (hexane,
toluene, Dichloroethane, Benzonitrile, acetonitrile, DMSO)
were HPLC grade without further purification.
It was concluded that the reaction rate increases with the
increase of TFA (as the catalyst) concentration in the
previous paper [27]. To improve the catalytic system, it
first occurs to us that TFA is employed as the catalyst and
solvent.
2.2 Apparatus and Analysis
Table 1 shows influence of some organic solvents on the
reaction rate and selectivity. Pure TFA without any organic
solvents was first examined (entries 1 and 2). However,
COX conversion and CPL selectivity were 27 and 85 %
The experiments were performed in
a well-stirred
(1000 rpm) glass vial (20 mL) which was placed on a
123