Prilezhaev dihydroxylation of olefins in a continuous flow process

Article abstract of DOI:10.1002/cssc.201100342

Epoxidation of both terminal and non-terminal olefins with peroxy acids is a well-established and powerful tool in a wide variety of chemical processes. In an additional step, the epoxide can be readily converted into the corresponding trans-diol. Batch-wise scale-up, however, is often troublesome because of the thermal instability and explosive character of the peroxy acids involved. This article describes the design and semi-automated optimization of a continuous flow process and subsequent scale-up to preparative production volumes in an intrinsically safe manner. Olefins go with the flow: Prilezhaev dihydroxylation can be performed on a large scale in continuous flow microreactor systems in the oxidation of terminal and internal olefins. Major drivers for a continuous flow process include better control, improved safety, and a faster overall process, leading to a significantly higher throughput. Copyright

Full text of DOI:10.1002/cssc.201100342

DOI: 10.1002/cssc.201100342
Prilezhaev Dihydroxylation of Olefins in a Continuous
Flow Process
Bas A. M. W. van den Broek,[a] Renꢀ Becker,[a] Florian Kçssl,[a] Mariꢁlle M. E. Delville,[b]
Pieter J. Nieuwland,*[a] Kaspar Koch,[a] and Floris P. J. T. Rutjes*[b]
Epoxidation of both terminal and non-terminal olefins with
peroxy acids is a well-established and powerful tool in a wide
variety of chemical processes. In an additional step, the epox-
ide can be readily converted into the corresponding trans-diol.
Batch-wise scale-up, however, is often troublesome because of
the thermal instability and explosive character of the peroxy
acids involved. This article describes the design and semi-auto-
mated optimization of
a continuous flow process and
subsequent scale-up to preparative production volumes in an
intrinsically safe manner.
Introduction
The Prilezhaev dihydroxylation is a transformation often used
in organic synthesis for the epoxidation of olefins and subse-
quent hydrolysis into the corresponding trans-diol.[1]
In this reaction, a peroxy acid is formed in situ by mixing the
carboxylic acid with hydrogen peroxide and sulfuric acid.[2]
After addition of the olefin, basic hydrolysis to the diol is per-
formed by addition of sodium hydroxide. This method, howev-
er, is laborious, and the thermal instability and explosive char-
acter of the peroxides, especially in a basic environment,
render the scale-up difficult. Since its first publication in 1909,
several attempts have been made to simplify the oxidation,
and several alternative oxidizing agents, such as meta-chloro-
peroxybenzoic acid (m-CPBA), have become commercially
available.[3–10] In addition to oxidation with peroxy acids, selec-
tive epoxidation of olefins by alkyl hydroperoxidases catalyzed
by d0-metal complexes (MoVI, VV, and TiIV) has been developed
for the manufacture of propylene oxide.[11] Both types of oxi-
dizing agents and catalysts are rather expensive, and many of
these oxidation processes suffer from the same thermal
instability,[12] making them unsuitable for most industrial appli-
cations. Continuous-flow microreactors can circumvent the
aforementioned problems. In a microreactor, reactions take
place on a microliter scale so that only small amounts of
peroxy acid are present during the process.[13] We envisioned
that a flow-chemistry approach would eventually enable us to
perform the Prilezhaev dihydroxylation on an industrial scale
with enhanced safety because the actual active reaction
volume remains within several milliliters.
reaction time, molar ratio (MR), and solvent. Secondly, a flow
method for epoxide hydrolysis has been investigated. The
combined process has also been successfully applied to a vari-
ety of substrates.
Results and Discussion
We chose to investigate the dihydroxylation of cyclohexene (1)
(Scheme 1) as a generally applicable procedure for the dihy-
droxylation of olefins using peroxy acetic acid (2) as the oxi-
dant. In a later stage, the resulting epoxide 3 was then con-
verted into trans-cyclohexane-1,2-diol (4).
Scheme 1. Prilezhaev dihydroxylation of cyclohexene (1).
In the batch procedure, peroxy acetic acid and cyclohexene
(MR=1:1) were stirred at 608C (Scheme 2); after 4 h, the reac-
tion was quenched by using an aqueous solution of sodium
sulfite (1m). A quantitative conversion of cyclohexene was ob-
served resulting in epoxide 3 (54%) and trans-diol 4 (36%), as
analyzed by using GC. In addition, diester 5 and monoester 6
In 2007, Hartung, Keane, and Kraft demonstrated the feasi-
bility of the oxidation of cyclohexene with in situ-prepared
peroxy formic acid in a qualitative continuous flow set up
using HPLC-tubing. In an additional step, the epoxide was con-
verted into trans-cyclohexane-1,2-diol.[14] Based on these re-
sults, we decided to develop a more widely applicable method
for the oxidation of olefins in flow by using the commercially
available peroxy acetic acid. This quantitative method involves
full optimization of reaction parameters including temperature,
[a] S. A. M. W. van den Broek, R. Becker, F. Kçssl, P. J. Nieuwland, K. Koch
FutureChemistry Holding BV, Toernooiveld 100
6525 EC Nijmegen (The Netherlands)
E-mail: p.nieuwland@futurechemistry.com
[b] M. M. E. Delville, Prof. Dr. F. P. J. T. Rutjes
Radboud University Nijmegen, Institute for Molecules and Materials
Heyendaalseweg 135, 6525 AJ Nijmegen (The Netherlands)
E-mail: f.rutjes@science.ru.nl
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201100342.
ChemSusChem 2012, 5, 289 – 292
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
289

P. J. Nieuwland, F. P. J. T. Rutjes, et al.
polynomial model. In house-developed FlowFit software[18] was
used to calculate the best possible model fit to provide a set
of optimal values for the reaction parameters. The results are
visualized in 2D-contour plots (Figure 2). The optimal reaction
conditions according to the model fit led to full conversion of
cyclohexene in a reaction time of 300 s, a temperature of
608C, and a molar ratio of 1.2. However, from this model, vari-
Scheme 2. Epoxidation of cyclohexene in-flow.
were obtained as minor side-
products. The large amount of
diol observed was probably
caused by acid hydrolysis during
the reaction. In flow, a similar
yield of the epoxide (67%) was
obtained, but because of effec-
tive heat transfer in the reactor,
the reaction time could be sig-
nificantly decreased to 5 min
when heated to 608C. Moreover,
only 10% of the diol together
with traceable amounts of 5 and
6
was observed under these
Figure 2. Contour plots of reaction-optimization model fit.
conditions.
In a continuous flow process
(Figure 1), cyclohexene and tolu-
ene (internal standard; syringe A) were mixed (M) with a
peroxy acetic acid solution (syringe B) for the oxidation reac-
tion to proceed. The reaction was quenched by using an aque-
ous solution of sodium sulfite (syringe Q), and the reaction
mixture was collected in a mixture of acetone/H2O (1:1, v/v).
The reactor had an internal volume of 92 mL, a channel width
of 600 mm, a channel depth of 500 mm, and an effective chan-
nel length of 360 mm. The channel layout contained two
mixing units (M) of the folding-flow type.[15] The reactor tem-
perature was controlled by using Peltier elements and sensed
by using a Pt1000 temperature sensor.
ous optimal parameter sets can be selected, depending on the
demands that have to be satisfied.
An in-line flow approach for epoxide hydrolysis and ester
saponification was investigated (Figure 3). Quenching was not
required as the reaction was driven to full completion; the
quench inlet was used to add an aqueous solution of sodium
hydroxide (5 equiv, 3.5m) to the crude reaction mixture to hy-
drolyze the epoxide along with saponification of the mono-
and diesters. Although hydrolysis proceeded quantitatively, the
large amount of salts formed in the second reaction led to
repetitive clogging of the reactor. Therefore, it was decided to
collect the epoxide in a solution of sodium hydroxide (3.5m)
(Figure 3). A fast hydrolysis to a nearly quantitative conversion
of trans-cyclohexane-1,2-diol (4) was observed when starting
from cyclohexene (1). Traceable amounts of mono- and diester
were also obtained, but a full conversion to the diol was ac-
complished after standing for an additional 2 h (Scheme 3).
Scale-up of batch processes generally poses a variety of
problems, as heat conduction and mixing greatly differ with re-
actor size. In batch chemistry, small-scale reactions are usually
A full optimization study was performed based on a reaction
time range (tr =60, 108, 180, 300 s), temperature range (T=25,
36, 48, 608C), and molar ratios of peroxy acetic acid to cyclo-
hexene (MR=0.6, 1.4, 2.2) spread across the optimization
region.[16,17] From these ranges, 50 data points (including dupli-
cates and triplicates) were selected using a D-optimal algo-
rithm. All experiments were performed in random order, and
the collected samples were analyzed at-line by using GC.
The GC results were normalized and fitted to a third-order
Figure 1. Microreactor design for epoxide optimization.
Figure 3. Microreactor design for olefin dihydroxylation.
290
www.chemsuschem.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2012, 5, 289 – 292

Dihydroxylation of Olefins
Conversely, the terminal olefins 1-octene (7, entry 1) and
methyl eugenol (17, entry 6) appeared slightly less reactive,
resulting in somewhat lower yields of 25 and 31%, respective-
ly, together with a substantial amount of starting material.
Oxidation of 1-octadecene (15, entry 5) gave only trace
amounts of the product, which can be probably attributed to
a low solubility of the lipophilic olefin in the aqueous solvent
mixture. The internal olefins (Z)-4-octene (9, entry 2) and oleic
acid (13, entry 4) were readily dihydroxylated in good to high
isolated yields of 65 and 99%, respectively.
Scheme 3. Epoxide hydrolysis in-flow.
conducted by organic chemists, whereas large-scale reactions
are performed by chemical engineers. By using flow chemistry,
scale-up is a one-to-one process in most cases, eliminating the
need for a different viewpoint on reaction conditions and prac-
tical concerns between small- and large-scale.
Conclusions
It has been demonstrated that the Prilezhaev dihydroxylation
can be performed on a large scale in a continuous-flow micro-
reactor system for the oxidation of both terminal and internal
olefins. A full optimization study for the epoxidation of cyclo-
hexene was performed by using a D-optimal algorithm for the
parametric optimization of reaction time (tr), temperature (T),
and MR. Optimal conditions were reached in a reaction time of
300 s at 608C with a molar ratio of 1.2. With the optimal condi-
tions found, an at-line procedure was developed to convert
the epoxide in the trans-diol. The newly developed method
was used to successfully scale-up the dihydroxylation of cyclo-
hexene to continuously produce 2.46 ghÀ1 of trans-cyclohex-
ane-1,2-diol (4). The overall yield of 82% showed that the
continuous process performs similar to the conventional batch
procedure. The major benefits, however, of performing this
flow process are better control and therefore less safety risks
and a faster overall process, leading to a significantly higher
throughput. The process also shows good applicability to a
wider range of substituted olefins.
By using the optimal conditions obtained from the optimiza-
tion data, a preparative synthesis of the target compound
(trans-cyclohexane-1,2-diol) resulted in continuous production
at a 2.46 ghÀ1 rate when using Uniqsis FlowSyn with a reactor
having an internal volume of 650 mL. If required, the produc-
tion can readily be scaled-up to higher rates by using parallel
reactors.
Based on the optimal conditions obtained in the dihydroxy-
lation of cyclohexene (1), a range of different substrates were
screened, containing both terminal and internal olefins
(Table 1). Initially, substrates 7, 9, and 11 were oxidized by
using the previously found optimal conditions. The latter sub-
strates, however, appeared to be less reactive than cyclohex-
ene, which results in low to moderate yields ranging from 5–
64%. Extended reaction times and slightly higher temperatures
were used for the preparative-scale experiments. Dihydroxyla-
tion of the terminal olefin 4-phenyl-1-butene (11, entry 3) on a
100 mg scale resulted in a satisfactory isolated yield of 74%.
Table 1. Screening of terminal and non-terminal olefins.
Experimental Section
Entry Substrate
Product
Isolated
yield
Reaction optimization
The FutureChemistry FlowScreen (C-300) automated reaction opti-
mization platform was used to perform the screening of reaction
parameters. The set up was identical to the FlowStart B-200, but
equipped with a sample collector and software to run automated
reaction parameter sequences. A glass syringe with an internal
volume of 5 mL and plastic syringes (NORM-JECT) of 10 mL were
used, as indicated in Figure 1. Pumps A and B were loaded with cy-
clohexene (5 mL) and peroxy acetic acid (10 mL), respectively.
Pump Q contained an aqueous solution of sodium sulfite (1m,
10 mL). The quenching and sampling procedure were left un-
changed.
25%
t=10 min
T=658C
MR=1.2
65%
t=10 min
T=658C
MR=1.2
74%
t=5 min
T=758C
MR=1.2
>99%
t=5 min
T=608C
MR=1.2
<5%
t=5 min
T=608C
MR=1.2
31%
t=5 min
T=608C
MR=1.2
1
2
3
4
5
6
Scale-up reaction
A scale-up experiment was performed by using a Uniqsis FlowSyn
(UQ-1020) equipped with a glass microreactor (0.65 mL) containing
folding flow-type mixing units. The general setup was similar to
the one depicted in Figure 1, but the quenching feed was omitted.
With a flow rate A (cyclohexene) of 39.1 mLminÀ1 and a flow rate B
(peroxy acetic acid, 32% w/w in acetic acid/water) of 90.8 mLminÀ1
,
a reaction time of 5 min was set to obtain full cyclohexene conver-
sion. In contrast to the optimization, no in-line quenching was
ChemSusChem 2012, 5, 289 – 292
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
291

P. J. Nieuwland, F. P. J. T. Rutjes, et al.
used, because the reaction was driven to completion. The reaction
mixture was collected for 157 min. Excess peroxy acetic acid from
the collected reaction mixture was removed under reduced pres-
sure, and the residue was treated with NaOH (5m, 20 mL) at 608C
for 45 min and neutralized with aqueous HCl. The solvent was re-
moved under reduced pressure, and the product was extracted
from the residue by using ethyl acetate (4ꢃ30 mL) and concentrat-
ed to yield trans-cyclohexane-1,2-diol (5.81 g, 50.5 mmol) in a 82%
isolated yield, which was comparable to the yield obtained in the
batch process (86%).
standard) were mixed together and filled into a 5 mL syringe (solu-
tion A). The syringes were connected to the FlowStart B-200
system and the flows were actuated. For a quantitative analysis,
the reaction mixture (100 mL) was collected in acetone/water
(400 mL, 1:1).
Keywords: alkenes
· dihydroxylation · flow chemistry ·
microreactors · oxidation
[1] N. Prileschajev, Ber. Dtsch. Chem. Ges. 1909, 42, 4811–4815.
[2] D. A. Singleton, S. R. Merrigan, J. Liu, K. N. Houk, J. Am. Chem. Soc. 1997,
119, 3385–3387.
[3] D. Swern, G. N. Billen, J. T. Scanlan, J. Am. Chem. Soc. 1946, 68, 1504–
1507.
[4] W. Hꢄckel, H. Schlee, Chem. Ber. 1955, 88, 346–348.
[5] R. B. Woodward, F. E. Bader, H. Bickel, A. J. Frey, R. W. Kierstead, Tetrahe-
dron 1958, 2, 1–57.
GC-analysis
All product mixtures were analyzed by using GC; the retention
times were according to Table 1. Analysis was performed by using
a Shimadzu GC2010 using a Quadrex 007 1701 apolar column
(length 15.0 m, inner diameter 0.10 mm) and a flame-ionization
detector (T=3258C, hydrogen flow rate 60 mLminÀ1, air flow rate
400 mLminÀ1), a temperature program (0–0.8 min: 358C; 0.8–
4.9 min: 35–2008C; 4.9–5.4 min: 2008C), and a 1.0 mL injection with
a split ratio of 200 (2508C injection temperature).
[6] B. Rickborn, S.-Y. Lwo, J. Org. Chem. 1965, 30, 2212–2216.
[7] R. C. Ewins, H. B. Henbest, M. A. McKervey, Chem. Commun. (London)
1967, 1085–1086.
[8] B. A. Carlson, N. S. Behn, J. Org. Chem. 1967, 32, 1363–1366.
[9] H. C. Brown, J. H. Kawakami, S. Ikegami, J. Am. Chem. Soc. 1970, 92,
6914–6918.
Experimental procedures
[10] W. Knçll, E. K. W. V. Tamm, Helv. Chim. Acta 1975, 58, 1162–1165.
[11] H. Mimoun, M. Mignard, P. Brechot, L. Saussine, J. Am. Chem. Soc. 1986,
108, 3711–3718.
[12] J. Howarth, 2003, “Decay Kinetics of Peroxyacetic Acid (PAA) and Hydro-
gen Peroxide (Perasan, EPA # 63838–2) in a Variety of Water Matrices.
http://www.envirotech.com/pdf/Perasan%20Decay.pdf.
[13] www.futurechemistry.com.
[14] A. Hartung, M. Keane, A. Kraft, J. Org. Chem. 2007, 72, 10235–10238.
[15] J. M. MacInnes, A. Vikhansky, R. W. K. Allen, Chem. Eng. Sci. 2007, 62,
2718–2727.
[16] P. J. Nieuwland, K. Koch, N. van Harskamp, R. Wehrens, J. C. M. van Hest,
F. P. J. T. Rutjes, Chem. Asian J. 2010, 5, 799–805.
[17] M. M. E. Delville, P. J. Nieuwland, P. Janssen, K. Koch, J. C. M. van Hest,
F. P. J. T. Rutjes, Chem. Eng. J. 2011, 167, 556–559.
[18] FlowFit is commercially available from FutureChemistry.
[19] Y. Hayashi, J. Yamaguchi, T. Sumiya, K. Hibino, M. Shoji, J. Org. Chem.
2004, 69, 5966–5973.
A typical procedure for the oxidation in batch: Cyclohexene
(5.61 mL, 55.0 mmol) was slowly added to peroxy acetic acid
(12.83 mL, 61.0 mmol, 32% w/w) at room temperature. The reac-
tion mixture was stirred at 608C for 4 h. For a quantitative analysis,
the reaction was quenched with an aqueous sodium sulfite solu-
tion (1m). The remaining peroxy acetic acid was evaporated, and
the residue was treated with an aqueous solution of sodium hy-
droxide (5m, 20 mL) at 608C for 45 min. The solution was neutral-
ized using hydrochloric acid (1N) and concentrated in vacuo.
The residue was washed with ethyl acetate (4ꢃ30 mL) to extract
the diol and recrystallized from cyclohexene. Yield: 86%; m.p.:
1
1018C; H NMR: d=1.10–1.33 (m, 4H), 1.60–1.75 (m, 2H), 1.90–2.05
1
(m, 2H), 2.10–2.60 (brs, 2H), 2.35–2.40 ppm (m, 2H). H NMR data
are in agreement with literature.[19]
A typical procedure for the oxidation in flow: Peroxy acetic acid
(32% w/w; solution B) and an aqueous solution of sodium sulfite
(1m, solution Q) were filled into a 10 mL syringe. Cyclohexene
(3.95 mL, 39.0 mmol) and toluene (803 mL, 7.55 mmol) (internal
Received: July 6, 2011
Revised: September 23, 2011
Published online on December 1, 2011
292
www.chemsuschem.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2012, 5, 289 – 292