Advantages of synthesizing trans-1,2-cyclohexanediol in a continuous flow microreactor over a standard glass apparatus

Article abstract of DOI:10.1021/jo701758p

(Figure Presented) Comparison is made between the preparation of trans-1,2-cyclohexanediol in standard glassware (conventional batch production) and in a microreactor (continuous flow production). The reaction sequence involved two exothermic steps where the standard procedure demands slow reagent addition and careful temperature control. In the microreactor, the reaction could be carried out safely with up to 3 times higher reagent concentration. Synthetic benefits were a faster reaction rate and a higher purity product free of colored impurities (a feature of the batch procedure).

Full text of DOI:10.1021/jo701758p

chemical synthesis are preferably made from glass, silicon, or
steel substrates which require specialist equipment for fabrica-
tion. Not surprisingly, the top-of-the-range commercial microre-
actors preferred in industry can be quite expensive. All this has
prevented a more widespread use of microreactors in the
synthetic organic laboratory despite their numerous advantages.
McQuade and co-workers have shown that a much simpler
version of a microreactorsconsisting of PVC tubing into which
reactants are delivered by syringe pumpsscan provide a
relatively cheap and viable alternative.1,6 Even combination of
two or more reactant streams is easily achieved through
micromixers, such as those used in standard low-pressure liquid
chromatography systems.
Advantages of Synthesizing
trans-1,2-Cyclohexanediol in a Continuous Flow
Microreactor over a Standard Glass Apparatus
Andreas Hartung,† Mark A. Keane,‡ and Arno Kraft*,†
Chemistry and Chemical Engineering, School of Engineering &
Physical Sciences, Heriot-Watt UniVersity, Riccarton,
Edinburgh EH14 4AS, United Kingdom
a.kraft@hw.ac.uk
ReceiVed August 10, 2007
A number of organic reactions have already been adapted
successfully to microreactors.1-5 Reactions that are exothermic
and/or involve unstable reagents (recent examples include Swern
oxidations7 or the generation of benzyne-free o-bromophenyl-
lithium)8 benefit most from being run in a microreactor.
Improved mass transfer and the use of higher temperatures and/
or concentrations ensure that chemical reactions proceed at faster
rates in microreactors when compared with a batch system.
Small (e1 mm) channel diameters guarantee a high surface-
to-volume ratio that facilitates effective heat transfer. This has
the advantage that even highly exothermic reactions can be
carried out more safely in a microreactor, allowing otherwise
hazardous conditions to be circumvented. Improved temperature
control leads to better reproducibility compared to reactions in
conventional flasks or in batch reactors where fluctuations in
temperature and hot spots are often responsible for poor
selectivity and the formation of byproducts. Scale-up, on the
other hand, is generally accomplished by running a microreactor
continuously or by operating (“numbering up”) several microre-
actors in parallel.
Here, we report an adaption of a 2-step synthesis of trans-
1,2-cyclohexanediol (5) to a simple continuous flow microre-
actor. To illustrate the scope of this approach, we have scaled
up the synthesis to a laboratory scale typical for starting
materials in a synthetic sequence.9 The results in a microreactor
were subsequently compared with those obtained in a round-
bottom glass flask (batch reaction) based on an Organikum and
Organic Synthesis protocol.10
Scheme 1 shows the synthetic sequence. We anticipated that
a switch of this 2-step synthesis from batch to continuous
microreactor operation would have the following advantages:
(1) Both steps 1 and 2 are exothermic. Whereas the Organi-
kum suggested a reaction temperature of 70 °C for step 1,
Organic Synthesis recommended that the temperature is main-
tained at 40 °C by adjusting the rate of addition of cyclohexene
Comparison is made between the preparation of trans-1,2-
cyclohexanediol in standard glassware (conventional batch
production) and in a microreactor (continuous flow produc-
tion). The reaction sequence involved two exothermic steps
where the standard procedure demands slow reagent addition
and careful temperature control. In the microreactor, the
reaction could be carried out safely with up to 3 times higher
reagent concentration. Synthetic benefits were a faster
reaction rate and a higher purity product free of colored
impurities (a feature of the batch procedure).
Increased attention is currently paid at the R&D level to
performing synthetic reactions in microreactors, especially by
chemical suppliers and pharmaceutical companies.1-5 A mi-
croreactor is a miniaturized flow reactor with channel diameters
between 10 µm and 1 mm. Under these conditions, fluid flow
is laminar while mass transfer is dominated by molecular
diffusion. Despite the success of custom-made microfluidic
devices, they have the disadvantage of being generally based
on silicone rubber, a polymer incompatible with most organic
solvents and reagents. For this reason, microreactors for
(6) (a) Quevedo, E.; Steinbacher, J.; McQuade, D. T. J. Am. Chem. Soc.
2005, 127, 10498-10499. (b) Poe, S. L.; Cummings, M. A.; Haaf, M. P.;
McQuade, D. T. Angew. Chem., Int. Ed. 2006, 45, 1544-1548.
(7) Kawaguchi, T.; Miyata, H.; Ataka, K.; Mae, K.; Yoshida, J. Angew.
Chem., Int. Ed. 2005, 44, 2413-2416.
† Chemistry.
‡ Chemical Engineering.
(1) (a) Steinbacher, J. L.; McQuade, D. T. J. Polym. Sci., Polym. Chem.
2006, 44, 6505-6533. (b) Mason, B. P.; Price, K. E.; Steinbacher, J. L.;
Bogdan, A. R.; McQuade, D. T. Chem. ReV. 2007, 107, 2300-2318.
(2) Watts, P.; Wiles, C. Chem. Commun. 2007, 443-467.
(3) Ahmed-Omer, B.; Brandt, J. C.; Wirth, T. Org. Biomol. Chem. 2007,
5, 733-740.
(4) Pennemann, H.; Watts, P.; Haswell, S. J.; Hessel, V.; Lo¨we, H. Org.
Proc. Res. DeV. 2004, 8, 422-439.
(5) Brivio, M.; Verboom, W.; Reinhoudt, D. N. Lab Chip 2006, 6, 329-
344.
(8) Usutani, H.; Tomida, Y.; Nagaki, A.; Okamoto, H.; Nokami, T.;
Yoshida, J. J. Am. Chem. Soc. 2007, 129, 3046-3047.
(9) Diol 5 is used as the starting material for the synthesis of artemisinin
model compounds.
(10) (a) Roebuck, A.; Adkins, H. Organic Syntheses; Wiley: New York,
1955; Collect. Vol. III, pp 217-219. (b) Autorenkollektiv. Organikum:
Organisch-chemisches Grundpraktikum, 19th ed.; Deutscher Verlag der
Wissenschaften: Leipzig, 1993; p 273.
10.1021/jo701758p CCC: $37.00 © 2007 American Chemical Society
Published on Web 11/15/2007
J. Org. Chem. 2007, 72, 10235-10238
10235

SCHEME 1. Synthesis of trans-1,2-Cyclohexanediol (5)
FIGURE 1. Schematic diagram of the microreactor setup for step 1
(flow rates and temperatures are given in parentheses). The flow rates
were adjusted to achieve a stoichiometric ratio of the reactants.
to the oxidant through a dropping funnel and, if necessary, by
cooling in an ice-water bath.
(2) Although the overall sequence is relatively free of side
reactions, diol 5 made by the batch procedure tended to be
discolored and therefore had to be purified further. We believed
that this is due to the extended reaction times used under batch
conditions (step 1: 2 h; step 2: 45 min)10b and that a much
shorter contact time (i.e., the time that reactants would be
exposed to elevated temperatures) in a microreactor would
provide a higher purity crude product, thus eliminating the need
for an extra recrystallization or distillation step.
The batch reaction was carried out following the Organikum
procedure10b on a scale designed to give an overall yield of ca.
10-14 g of crude diol 5. Two main problems became im-
mediately apparent with the reported batch procedure when
standard glassware was used. First, step 1 was highly exothermic
as evident from a sharp rise in temperature from 20 °C to above
90 °C when cyclohexene (12 mL) was added to a 1 M solution
of H2O2 in formic acid (100 mL). Second, the large amount of
solvent (formic acid) was essential to dilute all reactants in the
batch system in order to keep the reaction under control. A
further increase in the concentration of H2O2, in order to make
the reaction faster, would therefore be inadvisable for the batch
reaction. Both issues had to be addressed when the reaction
was switched to a microreactor. While the microreactor would
cope well with the exothermicity of the reaction, we were also
keen to increase the concentration of the reactants. Not only
would this reduce the amount of waste solvent generated but it
would also ensure a higher reaction rate and guarantee a
sufficient throughput when the reaction was to be carried out
on a 40-50 mL scale.
TABLE 1. Conditions for Epoxidation/Ring-Opening: Step 1
conditions
batch reaction10b
microreactor
reaction temp, °C
H2O2 concn, mol L-1
contact time, min
reaction time, min
70
50
2.9
1
1.0
120
120
82
through a 0.80 m long coil of PTFE tubing immersed in a heated
poly(ethylene glycol) bath. The reaction mixture then passed
through a short piece of tubing cooled in an ice-water bath
before finally being collected in a flask.
Reaction conditions such as temperature and contact time
were optimized initially on a small scale (corresponding to 0.4
mL of cyclohexene). On a 30-fold larger laboratory scale
equivalent to 12 mL of cyclohexene, essentially the same
conditions were maintained, only the reactor had to be run for
a longer period. A temperature of 50 °C was deemed most
appropriate as it was slightly higher than that recommended in
the original Organic Synthesis procedure10a but still well below
the boiling points of cyclohexene and formic acid.11 A contact
time of only 1 min was sufficient to achieve full conversion in
the microreactor. By carrying out step 1 inside thin PTFE tubing
we were able to increase the H2O2 concentration in formic acid
to 2.9 M, almost 3-fold compared to the original literature
procedure.10b The conditions for batch and optimized microre-
actor operation are summarized in Table 1.
GC analysis showed that three compounds had formed, and
cyclohexene (1) or epoxide 2 had reacted completely. After
destruction of excess H2O2 with sodium hydrogensulfite and
concentration in vacuum, an oil was isolated that was analyzed
by 1H and 13C NMR spectroscopy. The crude product from step
1 was found to consist of a mixture of monoester 3, diester 4,
and diol 5. The only byproducts present were traces of adipic
acid resulting from over-oxidation.12 The ratio of 3:4:5 after 1
min in a microreactor was 1:0.1:0.4, quite different from that
(1:1.3:0.1) obtained at prolonged contact times of up to 2 h in
the batch reaction. The increase in the ratio of diester 4 to diol
5 indicated that both monoester 3 and diol 5 underwent
esterification with time under the reaction conditions.
Our microreactor setup for step 1 made use of two syringe
pumps, one dispensing cyclohexene and the other a mixture of
formic acid and hydrogen peroxide, into poly(tetrafluoroethyl-
ene) (PTFE) tubings with an inner diameter of 1 mm (Figure
1). The two reactant streams were adjusted so that they
combined to generate the requisite 1:1 stoichiometry in a static
T mixer. A mixing unit was essential to facilitate rapid
dissolution of cyclohexene in formic acid.11 After leaving the
T mixer, the reactant stream was delivered at a constant rate
(11) Even in a microreactor, insufficient mixing can result in overheating
of the reaction mixture, albeit only on a small scale since the volume exposed
to higher temperature was less than 0.8 mL per meter of PTFE tubing.
Overheating was observed when a simple T coupler was used instead of a
static T mixer and the reaction temperature was raised to 70 °C. Under
these conditions, cyclohexene droplets several millimeters in length formed
inside the PTFE tubing and their subsequent reaction with H2O2 generated
so much heat that the formic acid mixture exhibited “bumping”. Surprisingly,
the overall yield of diol 5 was still virtually the same (88%) as that obtained
with the optimized procedure.
The batch reaction for step 2 revealed again two concerns.
A colored impurity formed within 10-15 min under the reaction
(12) Hydrogen peroxide is able to oxidize diol 5 to adipic acid under
acidic conditions, although a promoter such as sodium tungstate is needed
to make this conversion synthetically useful: Venturello, C.; Ricci, M. J.
Org. Chem. 1986, 51, 1599-1602.
10236 J. Org. Chem., Vol. 72, No. 26, 2007

TABLE 2. Conditions for Saponification: Step 2
conditions
batch reaction10b
microreactor
reaction temp, °C
90
70
equiv of NaOH
2.5
1.2
contact time, min
45
0.5
reaction time, min
45
35
yield of crude diol 5, %
mp of crude diol 5, °C
appearance of crude product
86
100.2
brown
88
102.0
colorless
FIGURE 3. Differential scanning calorimetry traces (heating rate 10
deg min-1, arrow indicates direction of endothermic transitions) of diol
5 prepared by a batch reaction versus reaction in a microreactor. The
photographs of the two samples show the discoloration (impurity) of
the product obtained in the batch reaction.
We were thus able to reduce the amount of waste salts (NaCl
and sodium formate) that were obtained after neutralization
with hydrochloric acid and from which product 5 was to be
extracted.
Diol 5 was isolated by extracting the concentrated residue
from step 2 with chloroform. The overall yield of the crude 5
obtained by the 2-step batch reaction (86%) differed little from
the yield in the microreactor (88%). However, the product from
the batch reaction was noticeably discolored (Figure 3). When
the reaction sequence was conducted in a microreactor, a
colorless crude diol 5 was obtained instead. Differential
scanning calorimetry of crude samples produced in the batch
and microreactor reaction showed a melting temperature of
100.2 vs 102.0 °C (Figure 3). The higher melting temperature
(comparable to recrystallized batch product), elemental analysis,
and sharper melting peak indicated a higher purity for the
product produced in the microreactor.
FIGURE 2. Schematic diagram of a continuous flow recycle reactor
consisting of a syringe pump and a peristaltic pump for step 2 (flow
rates and temperatures are given in parentheses).
conditions,13 leading us to infer that short contact times would
be advisable. Although the literature procedure suggested the
use of 2.5 equiv of sodium hydroxide, a large excess of inorganic
salts was likely to interfere with the extraction of diol 5 upon
workup. In adapting the saponification of the ester mixture to
a microreactor, we opted for a cyclic flow system. A continuous
flow recycle reactor is not usually considered in microreactor
technology for reasons of reduced space-time yield. It does,
however, have a role to play in special cases, for example, in
feeding microwave reactors where the closed loop arrangement
allows a short residence time from a single pass to be extended
until the conversion has reached satisfactory levels.14 In our case,
it allowed us to monitor the extent of saponification by thin-
layer chromatography and keep, at the same time, the excess
of NaOH to a minimum. This was more easily achieved in a
cyclic flow system since the required stoichiometry could vary
depending on the amount and ratio of esters 3 and 4 (as well as
residual formic acid) present. Conditions for batch and microre-
actor reaction are summarized in Table 2.
A mixture of 3, 4, and 5 in aqueous methanol was placed in
a beaker from which it was continuously circulated through a
silicone tubing (inner diameter 1 mm) with use of a peristaltic
pump. A syringe pump dispensed aqueous NaOH through a
needle piercing the tubing wall just before the ester mixture
was heated to 70 °C. The starting material/product mixture was
finally fed back to the beaker that contained the original ester
mixture, allowing a continuous loop process (Figure 2).
Complete saponification required a contact time of only 0.5 min.
Controlled addition of NaOH to the ester mixture made it
furthermore possible to minimize the excess NaOH needed.
In summary, both the highly exothermic epoxidation/ring-
opening step 1 and the saponification step 2 benefitted from
being run in a continuous flow microreactor system. Our
microreactor was assembled simply from syringes, ordinary
laboratory tubing, a mixing unit, and either two syringe pumps
or one syringe pump and a peristaltic pump. Apart from
improved temperature control, reactions inside the thin tubing
could be carried out safely at much higher reagent concentra-
tions, thus increasing the reaction rate considerably while also
reducing the amount of solvent waste. Full conversion was
achieved with contact times of 1 (step 1) and 0.5 min (step 2).
Advantages of the microreactor reaction from an organic
chemist’s point of view were a noticeable improvement in purity
of diol 5 (circumventing the extra distillation or recrystallization
step required in the batch procedure), reproducible yields, and
slightly shorter overall reaction times. Moreover, since the
conditions remained identical whether the reaction was carried
out on a small or a larger scale, the microreactor facilitated the
scaling up of this 2-step synthesis of a typical starting material
just by running the reaction for a longer period. These benefits
were achieved with a simple and relatively inexpensive mi-
croreactor setup (using common plastic syringes, laboratory
tubing, HPLC mixing unit and fittings, and syringe/peristaltic
pumps) and should be applicable to many other exothermic
reactions that similarly demand careful temperature control and
slow reagent addition.
(13) Although the identity of the colored impurity remains unclear,
possible sources for such an impurity are trace residues of stabilizer present
or oxidation byproducts formed over time in commercial cyclohexene.
(14) Chemat, F.; Poux, M.; di Martino, J.-L.; Berlan, J. Chem. Eng.
Technol. 1996, 19, 420-424.
J. Org. Chem, Vol. 72, No. 26, 2007 10237

through a 2.20 m long silicone tube (inner diameter 1 mm) by using
a peristaltic pump set at a flow rate of 58.8 mL h-1. A polypropylene
syringe was filled with an aqueous solution of NaOH (20%, 24
mL, 0.12 mol) and mounted on a syringe pump. The NaOH solution
was dispensed into the reaction stream, through a 27-gauge blunt-
edged needle inserted halfway into the silicone tubing, at a flow
rate of 48.0 mL h-1. A 0.80 m section of the tubing was immersed
in a heating bath kept at 70 °C, while 0.10 m of tubing nearer the
outlet was cooled in an ice-water bath. At the outlet, the reaction
solution was allowed to flow back into the beaker containing the
starting material, as shown schematically in Figure 2. The reaction
was stopped once the NaOH solution had been completely dispensed
and 3 or 4 could be detected no longer by thin-layer chromatog-
raphy. The mixture was neutralized with concentrated HCl and the
solvent was removed in a vacuum. The residue was extracted with
chloroform (5 × 30 mL) and the organic layers were combined.
After removing the solvent in a vacuum, a colorless solid was
Experimental Section
Microreactor Reaction of Cyclohexene with Hydrogen Per-
oxide in Formic Acid. The epoxidation/ring-opening step 1 was
carried out in a microreactor assembled from two syringe pumps,
PTFE tubing (1 mm inner diameter, 1/16 in. outer diameter), and
a static T mixer as shown in Figure 1. Luer fittings were used to
connect PTFE tubing with syringes and mixing tee. A 50 mL glass
syringe was filled with aqueous hydrogen peroxide (47%, 8.69 g,
0.12 mol) in formic acid (98%, 33 mL) and mounted on a syringe
pump. A 20 mL polypropylene syringe was filled with cyclohexene
(9.86 g, 12.2 mL, 0.12 mol) and mounted on a second syringe pump.
Flow rates were adjusted so that the two reagent streams were
combined in an equimolar ratio in a static mixing tee (flow rates:
29 mL h-1 for aqueous H2O2, 9 mL h-1 for cyclohexene). Upon
leaving the mixer, the resulting solution was allowed to pass through
0.80 m of PTFE tubing immersed in a heating bath (50 °C) followed
by a 0.10 m section of tubing cooled in an ice-water bath. The
resulting product solution was collected in a round-bottomed flask.
After removing the solvent in a vacuum, a clear oil (17.7 g) was
obtained and identified as a mixture of trans-1,2-cyclohexanediol
diformate (3), trans-1,2-cyclohexanediol monoformate (4), and
trans-1,2-cyclohexanediol (5). 3: 1H NMR (200 MHz, CDCl3) δ
1.18-1.47 (m, 4H), 1.64-1.78 (m, 2H), 1.98-2.12 (m, 2H), 3.55
(∼td, J ) 9.7, 4.6 Hz, 1H), 4.62 (td, J ) 9.1, 5.0 Hz, 1H), 5.44 (br
s, 1H, OH), 8.10 (s, 1H, OCHO); 13C NMR (50 MHz, CDCl3) δ
23.6 (CH2), 23.7 (CH2), 29.9 (CH2), 30.0 (CH2), 72.2 (CH), 78.0
(CH), 161.3 (CH); Rf (ethyl acetate) 0.50. 4: 1H NMR (200 MHz,
CDCl3) δ 1.18-1.47 (m, 4H), 1.64-1.78 (m, 2H), 1.98-2.12 (m,
2H), 4.85-5.00 (m, 2H), 8.02 (s, 2H, OCHO); 13C NMR (50 MHz,
CDCl3) δ 23.2 (CH2), 32.8 (CH2), 73.2 (CH), 160.4 (CH); Rf (ethyl
acetate) 0.64.
isolated and dried. Yield 12.3 g (88%); DSC 102.0 °C (99.3 J g-1
,
lit.10a mp 101.5-103 °C); H NMR (200 MHz, CDCl3) δ 1.18-
1.47 (m, 4H), 1.64-1.78 (m, 2H), 1.98-2.12 (m, 2H), 3.26-3.27
(m, 2H, CHOH), 3.55 (br s, 2H, OH); 13C NMR (50 MHz, CDCl3)
δ 24.2 (CH2), 32.8 (CH2), 75.7 (CH); Rf (ethyl acetate) 0.32. Anal.
Calcd for C6H12O2: C, 62.04; H, 10.41. Found (after drying): C,
61.19; H, 10.29. Found (after sublimation): C, 62.09; H, 10.59.
1
Acknowledgment. We thank the Royal Society, the School
of Engineering & Physical Sciences at Heriot-Watt University,
and ScotCHEM for support.
Supporting Information Available: Detailed experimental
procedures, NMR and GC analyses, and compound characterization
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
Microreactor Saponification Step. The oily mixture of 3, 4,
and 5 from step 1 was dissolved in a water-methanol mixture (5
mL:1 mL) and placed in a beaker. The solution was circulated
JO701758P
10238 J. Org. Chem., Vol. 72, No. 26, 2007