Continuous Process for Phase-Transfer-Catalyzed Bisalkylation of Cyclopentadiene for the Synthesis of Spiro[2.4]hepta-4,6-diene

Article abstract of DOI:10.1021/acs.oprd.5b00046

The transfer of a highly exothermic phase-transfer-catalyzed bisalkylation of cyclopentadiene with 1,2-dichloroethane from batch to continuous mode using standard and widely available laboratory equipment is presented. Besides the optimization of the reaction temperature and residence time, the efficient mixing of the organic phase (cyclopentadiene, 1,2-dichloroethane, and MeBu₃NCl) and the aqueous phase (30% NaOH) was studied in detail and optimized by the use of a simple, homemade "PTFE Raschig ring static mixer" consisting of a PTFE tube filled with small pieces cut from the same PFTE tube. A flow set-up with two PTFE Raschig ring static mixers and three residence time units with a three-stage temperature profile allowed the synthesis of highly energetic spiro[2.4]hepta-4,6-diene in a yield of 95% with a productivity of 15 g/h under safe conditions.

Full text of DOI:10.1021/acs.oprd.5b00046

Article
pubs.acs.org/OPRD
Continuous Process for Phase-Transfer-Catalyzed Bisalkylation of
Cyclopentadiene for the Synthesis of Spiro[2.4]hepta-4,6-diene
Elia Kilcher, Sebastien Freymond, Ennio Vanoli, and Roger Marti*
́
HES-SO Haute ec
Technology, Bd Per
́
ole spec
́ ́ ́ ́
ialisee de Suisse occidentale, Haute ecole d’ingenierie et d’architecture de Fribourg, Institut Chemical
́
olles 80, CH-1700 Fribourg, Switzerland
Gunther Schmidt and Stefan Abele
Actelion Pharmaceuticals Ltd, Gewerbestrasse 16, CH-4123 Allschwil, Switzerland
S Supporting Information
*
ABSTRACT: The transfer of a highly exothermic phase-transfer-catalyzed bisalkylation of cyclopentadiene with 1,2-
dichloroethane from batch to continuous mode using standard and widely available laboratory equipment is presented. Besides
the optimization of the reaction temperature and residence time, the efficient mixing of the organic phase (cyclopentadiene, 1,2-
dichloroethane, and MeBu NCl) and the aqueous phase (30% NaOH) was studied in detail and optimized by the use of a simple,
3
homemade “PTFE Raschig ring static mixer” consisting of a PTFE tube filled with small pieces cut from the same PFTE tube. A
flow set-up with two PTFE Raschig ring static mixers and three residence time units with a three-stage temperature profile
allowed the synthesis of highly energetic spiro[2.4]hepta-4,6-diene in a yield of 95% with a productivity of 15 g/h under safe
conditions.
INTRODUCTION
possibilities to discover and exploit new chemical trans-
formations.
An important technical feature of this continuous reaction
■
8
Spiro[2.4]hepta-4,6-diene (1) is a key intermediate for the
synthesis of an active pharmaceutical ingredient at Actelion
1
technology is the mixer. For very fast reactions (so-called type
Pharmaceuticals Ltd. It can be prepared in semibatch mode by
9
A, based on the classification by Roberge ), the right mixer
phase-transfer-catalyzed alkylation of cyclopentadiene with 1,2-
design is crucial for a successful implementation of a
continuous process. Another field where the mixing has a
great impact on the process is the study of biphasic reactions,
dichloroethane using concentrated NaOH as the base at
2
5
0−60 °C. A detailed thermal risk assessment of this reaction
3
by Abele et al. points out the challenges of this synthesis: (a)
cyclopentadiene is volatile (bp 39−43 °C) and unstable, with
start of the decomposition reaction of −890 kJ/kg at
10
11
such as liquid/liquid or liquid/gas reactions, hence mass-
transfer-controlled reactions. Our interest is in liquid/liquid
reactions, as we aimed for the transfer of an alkylation reaction
of cyclopentadiene and 1,2-dichloroethane (organic phase)
with aqueous NaOH as the base from batch mode to a
continuous process. In microstructures, different flow pat-
4
approximately 60 °C; (b) the desired alkylation reaction itself
is highly exothermic with a high maximum heat release rate
80−100 W/kg) and an inherent risk of accumulation if the
(
addition of the reagents is done too fast or mixing is not
efficient; and (c) the exothermic decomposition reaction of the
final product 1 has a low onset temperature of 80 °C. This
requires careful process control of the semibatch reaction
andif the purity is not sufficient for the subsequent
reactiondistillation of 1.
12
terns for biphasic liquid/liquid mixtures can be observed
depending on the type of mixer and flow rates: from annular or
parallel flow, where the two immiscible liquids are just in
contact via an interlayer, to slug flow, where large drops or
segments of immiscible liquid are present in the liquid flow, to
bubbly flow, where small droplets are dispersed in the flow
These challenges and the fact that we have to deal with a
biphasic reaction motivated us to develop a flow process for the
synthesis of 1. Microreactors in the field of process chemistry
13
(
Figure 1). This bubbly flow is characterized by a high
2
−3
surface-to-volume ratio above 150 000 m m . By coalescence,
this bubbly flow goes back to slug flow with larger droplets. For
phase-transfer-catalyzed reactions, one aims at either reprodu-
cible, small-segmented slug flow or, even better, bubbly flow to
5
often offer the exploration of “novel process windows”, hence
their attractiveness. Better heat and mass transfers due to
miniaturization and changing from classical “reaction time”
known from classical batch reactions to “flow distance” in
continuous reaction systems on the one hand enable the study
of very short reaction times (so-called flash chemistry ) and on
the other hand allow robust and efficient chemistry to be
14
ensure efficient mixing and therefore fast reactions.
6
Special Issue: Continuous Processing, Microreactors and Flow
Chemistry
established. Flow chemistry also allows syntheses to be carried
out at very high temperatures and pressures and opens up
Received: February 13, 2015
7
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 1. Semibatch synthesis of spiro[2.4]hepta-2,4-diene
(
1)
entry
conc. of NaOH (%)
yield (%)
59
1
2
3
4
50
50
40
30
a
78
63
63
a
NaOH was dosed over 1 h to a mixture of cyclopentadiene, 1,2-
dichloroethane, and MeBu NCl.
3
5
0% aqueous NaOH solution over 1 h to a mixture of
cyclopentadiene, 1,2-dichloroethane, and MeBu NCl (entry 2),
3
the yield increased slightly from 59% to 78%. Again the
Figure 1. Schematic presentation of the different flow regimes.
exothermic reaction could be easily controlled (ΔT of T /T ≈
r
j
1
0 °C), and a second exotherm of approximately 15 min after
addition was detected. During the reaction, no problems with
precipitates were observed, nor were any byproducts such as
vinyl chloride (by an elimination reaction of 1,2-dichloro-
ethane) or dicyclopentadiene detected in the crude product by
NMR analysis. Next, we evaluated the use of less concentrated
aqueous NaOH solutions, keeping the 8 equiv constant (entry
1
5
Kobayashi reported a phase-transfer-catalyzed alkylation
reaction of β-keto esters in microreactors and observed faster
reaction using smaller channels (100 μm as opposed to 200/
4
00 μm), resulting in “smaller/uniform” segmented flow. By
careful optimization of the organic/aqueous phase ratio and the
flow rate, Schouten and Hessel obtained uniform segmented
slug flow and higher conversions and selectivites compared with
the batch reaction for the phase-transfer-catalyzed alkylation of
phenylacetonitrile. Working under bubbly-flow conditions, a
redispersion (remixing) is needed to avoid coalescence and re-
formation of slug flow. For example, Hessel and co-workers
used a redispersion capillary to ensure bubbly flow within the
whole capillary and therefore high and reproducible reaction
3
and 4). The yields remained the same, but the advantage was
that the reaction mixture was less viscous and the emulsion
separated more slowly at higher temperature compared with
the runs using 50% NaOH. Using less concentrated aqueous
NaOH solution also has the benefit of better compatibility (less
corrosive and less viscous) with the equipment such as pumps,
tubing, and the glass microreactors.
16
On the basis of these initial results from the semibatch
reaction, we designed a first flow system using an organic
stream composed of a mixture of cyclopentadiene, 1,2-
17
rates for phase-transfer-catalyzed biphasic esterification. To
ensure redispersion, Buchwald used a packed-bed reactor
consisting of a tube filled with stainless steel beads (60−125
μm diameter) for biphasic C−N coupling reactions to give high
dichloroethane, and MeBu NCl, which was mixed with a
3
stream of aqueous 50% NaOH solution at 40 °C in a
microreactor followed by a residence time unit (Teflon tube).
The reaction mixture was quenched at the exit of the flow
reactor with ice−water, and after aqueous workup and
extraction of 1 with heptane, the yield was determined by
NMR assay of the neat heptane solution with 1,4-dimethoxy-
benzene as an internal standard.
1
8
and reproducible yields. Similarly, McQuade employed
TEMPO-modified polymer beads in a catalytic packed-bed
reactor as a remixing module for the biphasic bleach oxidation
1
9
of alcohols.
RESULTS AND DISCUSSION
■
First tests were performed in semibatch mode to evaluate the
feasibility of transferring the reaction into a continuous process.
We especially checked for the formation of any solids in the
reaction mixture that might block the flow system. According to
20
The first continuous reactions were done using cyclo-
pentadiene, 1 equiv of 1,2-dichloroethane, and 0.01 equiv of
MeBu NCl as the organic stream and 8 equiv of 50% aqueous
3
NaOH solution as the aqueous stream, keeping the reaction
2
,3
the published procedure, freshly prepared cyclopentadiene
temperature constant at 40 °C. First we tested a simple T-
was added to 8 equiv of 50% aqueous NaOH and 0.01 equiv of
mixer. At a residence time (t ) of 10 min we got a yield of 25%,
R
the phase-transfer catalyst (PTC) MeBu NCl, after which 1,2-
which was significantly lower than that of the semibatch
reaction, but a clean and selective reaction was observed (Table
2, entry 1). Only the starting materials cyclopentadiene and 1,2-
dichloroethane could be determined (by NMR) in the crude
reaction mixture besides the desired product 1. Similar to the
batch reaction, no byproducts such as vinyl chloride or
dicyclopentadiene were observed. The reaction mixture showed
the typical reddish-brown color of the sodium cyclopentadiene
salt, and in the Teflon tube we clearly saw two separate phases,
indicating slug flow. Increasing the temperature to 50 °C to
speed up the reaction caused evaporation of the cyclo-
pentadiene (bp 42 °C) and bubble formation and “pulsation”
of the liquid phase. At a residence time of 30 min, the reaction
3
dichloroethane was dosed over 2 h at 30 °C (Table 1). When
the reaction was performed using Mettler EasyMax equipment,
the exothermic reaction could be easily controlled by
maintaining the ΔT between the temperature of the reaction
mixture (T ) and the temperature of the reactor jacket (T) at
r
j
approximately 10 °C. Increasing the stirrer speed increased the
reaction rate and ΔT accordingly. After the reaction, a dark-
brown emulsion was obtained but no precipitate was observed!
After aqueous workup and extraction of the product 1 with
heptane, the yield was determined by NMR assay of the crude
heptane solution with 1,4-dimethoxybenzene as an internal
standard. When the mode of addition was changed by dosing of
B
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
a
Table 2. Initial “simple” microreactor set-up
Table 3. Variation of the NaOH concentration, PTC loading,
and residence time with the initial flow reactor set-up
a
equiv of
PTC
conc. of NaOH
(%)
tR
(min)
yield
(%)
entry
mixer
1
2
3
4
5
6
LTF-MX
LFT-MX
LTF-MX
LTF-MX
LFT-MX
LTF-MX
0.01
0.1
50
50
50
30
30
30
10
5
31
32
b
0.1
10
10
20
30
−
0.1
43
49
46
0.1
0.05
entry
mixer type
tR (min)
yield (%)
25
a
Conditions: 1.0 equiv of 1,2-dichloroethane, 8 equiv of NaOH, 40
C; microreactor set-up as in Table 2. The flow system was blocked,
and no product could be isolated.
1
2
3
4
5
6
7
8
9
T-mixer
10
10
30
5
b
°
b
T-mixer
32
c
T-mixer
−
LTF-MX
LTF-MX
LTF-MX
LTF-MX
LTF-MX
Ehrfeld LH2
8
32
49
10
20
20
30
10
turned dark brown, and by the end of the residence time unit
the Teflon tube was blocked completely (entry 3). When the
residence time was shortened to 5 min (entry 2), the reaction
mixture remained “liquid” at the exit of the tube and the yield
was 32%, nearly quadruple that with 0.01 equiv of PTC (8%
yield; see Table 2, entry 4).
To avoid the blocking of the microreactor system by the
increased viscosity of the reaction mixture and/or formation of
solids, we changed to a more dilute 30% aqueous NaOH
solution. The 8 equiv of base was kept constant, which was
adjusted by the flow ratio of the pumps. The aspect of the
reaction mixture at the exit of the reactor was right away more
“liquid”, and residence times of 20 or even 30 min were
possible without blocking of the tubes (Table 3, entries 5 and
6). In all of the tests, the yield did not exceed 50%: even with
b
56
c
−
11
a
Conditions: 1.0 equiv of 1,2-dichloroethane, 0.01 equiv of
MeBu NCl, 8 equiv of 50% aq. NaOH, 40 °C; see the Experimental
3
b
Section for more details on the set-up of the flow reactor system. The
residence time unit was placed in an ultrasonic bath. The flow system
was blocked, and no product could be isolated.
c
mixture became very viscous and precipitates formed at the
interphases, which caused blocking of the system (entry 3).
With an MX mixer, a split-and-combine mixer made from glass
by Little Things Factory (LTF), better mixing of the biphasic
reaction mixture was observed: the slug flow in the tube
showed smaller segments, and at a residence time of 10 min the
yield was slightly improved compared with the simple T-mixer
0.1 equiv of MeBu NCl and a prolonged residence time,
3
starting material (cyclopentadiene and 1,2-dichloroethane)
remained in the crude product mixture. Small bubbles were
clearly discerned, indicating good mixing of the biphasic
reaction mixture after the LTF-MX mixer, but thereafter this
bubbly flow coalesced and turned into slug flow in the
residence time unit (as judged by visual inspection). This
observation prompted us to study a multi-mixer set-up that
forces this slug flow to remix again.
(
entry 5). At a residence time of 5 min (entry 4) the yield
dropped drastically (to 8%), but at 20 min (entry 6) a
reasonable yield of 49% was achieved, comparable to that of the
semibatch reaction at 30 °C with an exothermic postreaction of
approximately 15 min.
When the residence time was increased to 30 min, the system
again became blocked as a result of the very viscous reaction
mixture (Table 2, entry 8). The yield could be slightly increased
by applying ultrasonication (entries 2 and 7), but the effect
was not significant (the accuracy of the NMR assays is
approximately 1−2% for standard NMR ).
When an LH2 multilamination mixer from Ehrfeld was used
This new set-up featured a second mixer between two
residence time units (see Table 4). At a residence time of 10
min, this new set-up resulted in an increase in the yield by 10%
(entries 5 and 6), and at a residence time of 20 min the yield
increased by 5% (entries 7 and 8). The remixing had a
beneficial effect, but optical examination revealed that the
organic and aqueous phases remained segmented after the
additional LTF-MX mixer. To see whether an increased flow
rate at the same residence time would give better mixing, we
doubled the reactor volume from 10 to 20 mL by doubling the
length of the residence time unit and adjusted the flow rate by a
factor of 2 to keep the residence time constant (entry 3).
Obviously, this higher energy input by the pumps had no
significant impact on the yield (Figure 2). The graph in Figure
2 clearly illustrates that even with increased residence times the
yield stalled at approximately 50%. One can also see that the
new set-up with an additional mixer increased the yield
(compared with the data for the set-up with just one mixer,
indicated by ■). In order to improve the yield, the mixing
efficiency of the flow system had to be optimized. Increasing
the temperature to drive the reaction to completion was not
feasible, as at reaction temperatures above 80 °C the product 1
1
4c
21
(
Table 2, entry 9), the yield of the spiro compound 1 was only
about half of the yield obtained with the simple T-mixer. This
was an astonishing result, but checking the data sheet of the
LH2 mixer showed that we would need at least 10 times higher
flow rates to have the appropriate energy input to obtain good
mixing (≥15 mL/min compared with 1 mL/min as in our
case). We also tried a redispersion capillary as described by
1
7
Hessel. We modified a Teflon tube (i.d. = 2.8 mm, length =
0 mm) with inserts of small pieces of Teflon tube (o.d. = 2.8
9
mm, i.d. = 1.2 mm, length = 3 mm, distance between inserts = 5
mm; for more details see Supporting Information), but we
could not observe any improvement in mixing or yield because
the flow was too slow to create remixing (result not shown).
Next, we evaluated the influence of the concentration of the
aqueous NaOH, the amount of the PTC, and the residence
time (Table 3; microreactor set-up as in Table 2). A 10-fold
increase in the amount of the PTC at a residence time of 10
min blocked the flow system: the reaction mixture immediately
3
starts to decompose violently.
C
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
a
Table 4. Test of a multi-mixer set-up
time unit with an integrated snake mixer unit, and finally an
additional LTF-MX mixer as a remix unit) followed by three
residence time units held at increasing temperatures. The
second mixer was placed before the last residence time unit
(
7
Table 5). Already in the first run the yield of 1 increased to
5% (entry 1), and by visual examination one could easily
observe much better mixing right after the first series of mixers
and “smaller” segmentation of the two phases in the Teflon
tube coils. The yield increased by 10% when the amounts of
both the PTC and 1,2-dichloroethane were increased (entry 2).
This improved yield of 85% was maintained also with an
enlarged reactor volume (PTFE tubing with i.d. = 2.4 mm
instead of 1 mm) and flow rates (both by a factor of 3) while
the residence time was kept constant at 25 min (entry 3).
Compared with all of the proceeding tests (Tables 2−4), no
traces of the cyclopentadiene starting material were detected in
the crude product.
mixer type
entry
mixer i
mixer ii
tR (min)
yield (%)
1
2
3
4
5
6
7
8
LTF-MS
LTF-MS
LTF-MS
LTF-MS
LTF-MX
LTF-MS
LTF-MS
LTF-MS
LTF-MX
LTF-MX
LTF-MX
LTF-MX
−
0.5
2
20
35
b
2
34
5
49
43
53
49
54
10
10
20
20
LTF-MX
−
Corrosion of the LTF glass reactors and the aluminum case
led to breaking of the glass flow reactors. These problems
prompted us to develop our own “PTFE Raschig ring static
mixer” (RRSM) inspired by the packed-bead reactors of
LTF-MX
a
Conditions: 1.0 equiv of 1,2-dichloroethane, 0.1 equiv of MeBu NCl,
3
b
8
equiv of 30% NaOH, 40 °C. The reactor volume and flow rate were
18
19
doubled.
Buchwald and by McQuade. This was done by filling a
Teflon tube (i.d. = 2.4 mm, length = 50 cm) with self-made
Raschig rings made by cutting a Teflon tube (o.d. = 1.6 mm, i.d.
= 1.0 mm) perpendicular to the axis of the tube into small
pieces with lengths of 1−2 mm. This self-made RRSM was
integrated into the set-up of Table 5 by using a T-mixer
connected to the two exits of the pumps instead of the LTF-
MX/VS/MX mixer sequence. The yield for this new mixer set-
up was 80% and in a run over 7 h even 85% (entries 4 and 5).
Very efficient mixing resembling bubbly flow was observed at
the exit of the RRSM, which then coalesced back to segmented
flow in the residence time units. With an additional self-made
RRSM as a remixing unit (mixer ii), the yield of 1 even
increased to 95%. Operating the flow reactor for 1.3 h, we
produced 19.8 g of 1 as a solution in heptane (10.5% w/w by
NMR assay). The productivity for this final set-up was 14.8 g of
In the set-up in Table 4, we could observe good mixing in the
beginning, as just after the first LTF-MS mixer we had bubbly
flow. However, after a few centimeters the two phases started to
separate in the Teflon tube, and we observed slug flow with less
contact between the phases, thus limiting the reaction rate. To
overcome the coalescence of the two immiscible phases either
17
remixing as described by Hessel or a packed-bed reactor as
18
19
described by Buchwald and McQuade are conceivable
options. In addition, some bubbles were observed, most
probably due to evaporation of cyclopentadiene. Therefore, a
staggered temperature profile was deemed ideal to cope with
this evaporation and with the challenge of decomposition at
prolonged exposure to high temperature and stalled conversion.
The next flow reactor set-up (see Table 5) entailed a series of
mixers at the start (a LFT-MX mixer, an LTF-VS residence
1
1/h (entry 6). The only impurities detected by H NMR
Figure 2. Yield of 1 as a function of the residence time. Values from Table 4: ◆, set-up with two mixers; ▲, set-up with two mixers at the same tR
but double the flow rate; ■, set-up with one mixer.
D
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 5. Optimization of the multi-mixer set-up and the amounts of PTC (MeBu NCl) and 1,2-dichloroethane (DCE)
3
mixer type
entry
mixer i
mixer ii
equiv of PTC
equiv of DCE
tR (min)
yield (%)
1
2
3
4
5
6
LTF-MX/VS/MX
LTF-MX/VS/MX
LTF-MX/VS/MX
LTF-MX
LTF-MX
LTF-MX
−
0.12
0.14
0.14
0.14
0.14
0.14
1
20
20
25
25
25
25
75
85
85
80
1.2
1.4
1.4
1.4
1.4
a
T-mixer/RRSM
a
b
T-mixer/RRSM
−
85
a
a
c
T-mixer/RRSM
RRSM
95
a
b
c
Self-made “PTFE Raschig ring static mixer” (RRSM). Production run over 7 h. 19.8 g of 1 was produced in 1.3 h.
Scheme 1. Subsequent Diels−Alder reaction of 1
analysis were 1,2-dichloroethane and traces of dicyclopenta-
diene. Attempts to purify 1 by distillation from the crude
reaction mixture were not successful because of the similarity of
the boiling points of 1 and heptane. Instead, the crude product
Table 6. Comparison of the batch and flow modes
batch
a
b
reaction
flow reaction
1:1.4:0.14
cyclopentadiene/1,2-dichloroethane/
1:1:0.01
1
was sufficiently pure to be used directly in the subsequent
MeBu NCl molar ratio
3
c
Diels−Alder reaction, as shown in Scheme 1. The Diels−Alder
product 2 was formed in 98% yield with a diastereomeric ratio
of 94:6, which compares favorably with the results obtained
conc. of NaOH
50%
30%
reaction temperature
residence time
30 °C
−
25 °C/40 °C/60 °C
25 min
d
1
,3,22
using 1 purified by distillation (78% yield).
of the Diels−Alder adduct 2 yielded the dicarboxylic acid 3 in
1% yield as a crystalline material, and an enantiomeric ratio
Saponification
yield
78%
95%
purity (by NMR assay)
productivity of 1
95% w/w
95% w/w
14.8 g/h
−
9
a
b
(
e.r.) of 97:3 was determined by HPLC.
Data from the reaction in Table 1, entry 2. Data from the reaction in
c
d
The results obtained from the optimized flow reactor
Table 5, entry 6. 8 equiv for the batch and flow reactions. Dosing
over 1 h, postreaction 30 min.
synthesis (Table 5, entry 6) were compared with the data for
the batch reaction (Table 1, entry 2), as shown in Table 6. A
major difference was the larger excess of 1,2-dichloroethane and
e.r. of the dicarboxylic acid 3 was similar. The productivity of 15
g/h should be easily increased by the use of a larger flow
system, provided that the mixing is similarly efficient (for a
discussion of the scale-up of continuous processes, see refs 10a
and 23).
greater amount of MeBu NCl used in the flow synthesis. These
3
two parameters were decisive in reaching complete conversion
and therefore a high yield of 1 in the continuous process. It is
important to note that using larger amounts of these two
reagents, MeBu NCl and 1,2-dichloroethane, had no detri-
3
mental impact on the quality of the final Diels−Alder product.
The use of a less concentrated aqueous NaOH solution was
based just on practical reasons: a less viscous reaction mixture
and less corrosion at similar reactivity were achieved. The yields
and purities of 1 are comparable or even better for the flow
reaction than for the batch reaction. This was confirmed by the
subsequent Diels−Alder reaction (Scheme 1). When distilled 1
originating from a batch reaction was used, the overall yield and
CONCLUSION
The semibatch biphasic synthesis of spiro[2.4]hepta-4,6-diene
1) was successfully transferred into a high-yielding continuous
process. A key feature of the new process is a specifically
designed mixer to secure the desired flow regime within the
flow system. The careful choice of the concentration of the
aqueous NaOH solution and the adjusted amounts of 1,2-
■
(
dichloroethane and MeBu NCl as the phase-transfer catalyst
3
E
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
allowed a practical residence time of 25 min for complete
conversion. A well-adapted temperature profile of three
residence time units was required in order to reconcile the
constraints imposed by the boiling point of cyclopentadiene
and the increasing viscosity of the reaction mixture while
maximizing the conversion. By optimization of the reactor
design, the temperature ramp, and the reaction conditions, the
yield of 1 was improved from 32% using a simple T-mixer to
approximately 50% using two LTF-MX glass mixers and finally
to 95% using new, simple, and cheap “Raschig ring static
mixers” (RRSMs). Our self-made RRSMs were made from
simple Teflon tubing and proved to be efficient mixers for
biphasic and corrosive media. These RRSMs can be simply
adapted in their design (i.d., length) and materials (e.g., glass or
stainless steel) to any flow chemistry problem. Additional work
is currently ongoing for the characterization of the RRSM in
terms of pressure drop, mixing characteristics, and flow regime
of the mixer, residence time distribution, “scalability” of the
RRSM, and prolonged use.
dimethoxybenzene, this solution contained 11% w/w 1,
corresponding to a yield of 78%. The analytical data correspond
2
to the literature data.
1
H NMR (CDCl ): δ 1.67 (s, 4H, cyclopropyl); 6.1−6.15
(m, 2H, CH); 6.5−6.6 (m, 2H, CH).
3
“Simple” Flow Reactor (Table 2, Entry 6). The flow reactor
consisted of an LTF-MX mixer (0.2 mL) and coiled Teflon
tube (20 mL, length = 1050 cm, i.d. = 1.56 mm), which was
kept at 40 °C (Configuration shown in Table 2). Feed 1 (6.05
mmol of cyclopentadiene/mL) was prepared by mixing
cyclopentadiene (16.0 g, 20 mL, 242 mmol, 1 equiv), 1,2-
dichloroethane (24.0 g, 19.4 mL, 242 mmol, 1 equiv), BHT (30
mg), and MeBu NCl (570 mg, 2.4 mmol, 0.01 equiv) and was
3
pumped at a flow rate of 0.355 mL/min. Feed 2 consisted of
50% aqueous NaOH solution and was pumped at a flow rate of
0.645 mL/min. The total flow rate was 1.0 mL/min,
corresponding to a residence time of 20 min. The reactor
was run under these conditions for 35 min before the reaction
mixture was collected during 22 min (theoretical yield of 1:
4
7.3 mmol) by quenching into H O (150 mL) at 0 °C. The
2
EXPERIMENTAL SECTION
dark-brown emulsion was diluted with heptane (50 mL), and
the phases were separated. The organic layer was washed twice
with H O (40 mL), once with 1 M HCl solution (40 mL), and
twice with H O (40 mL). The organic layer was dried over
MgSO and filtered, and 1 was obtained as a solution in
■
General. The chemicals were reagent grade and used as
supplied, unless stated otherwise. The cracking of dicyclopen-
tadiene into cyclopentadiene was done at 160 °C (IT), and the
freshly distilled cyclopentadiene (bp 42 °C) was collected and
2
2
4
2
0
heptane (27.0 g). This solution contained 8% w/w 1 as
determined by H NMR assay, corresponding to a yield of 49%.
stored at −20 °C.
1
1
H (300 MHz) and 13C NMR (75 MHz) spectra were
“
Multi-Mixer” Flow Reactor (Table 4, Entry 8). The flow
recorded on a Bruker AVANCE 300 spectrometer. Chemical
reactor consisted of an LTF-MS mixer (0.2 mL), a coiled
Teflon tube (9.8 mL, length = 1250 cm, i.d. = 1.0 mm), an
additional LTF-MX mixer (0.2 mL), and an additional coiled
Teflon tube (9.8 mL, length = 1250 cm, i.d. = 1.0 mm). The
total volume was 20 mL, and the set-up was kept at 40 °C (the
configuration is shown in Table 4). Feed 1 (5.35 mmol of
cyclopentadiene/mL) was prepared by mixing cyclopentadiene
(7.1 g, 8.9 mL, 107 mmol, 1 equiv), 1,2-dichloroethane (10.6 g,
shifts (δ) are reported in parts per million relative to Me Si
4
(
0.00 ppm). The contents (% w/w) of all crude products and
reaction mixtures were assessed by NMR assay with 1,4-
dimethoxybenzene as an internal standard. For 1, integrals of
the signals of the cyclopentadiene hydrogens (two multiplets at
6
.1 and 6.5 ppm) were compared to the integral of the methyl
singlet of 1,4-dimethoxybenzene (at 3.56 ppm). In all cases, the
time between scans (the relaxation delay, D1) was set to 20 s.
Flow Reactor Set-Up. The T-mixer was purchased from
Bola (Germany); the LTF-MX (volume = 0.2 mL, channel size
8.6 mL, 107 mmol, 1 equiv), BHT (20 mg), and MeBu NCl
3
(2.54 g, 10.7 mmol, 0.1 equiv) and was pumped at a flow rate of
0.188 mL/min. Feed 2 consisted of 30% aqueous NaOH
solution and was pumped at a flow rate of 0.812 mL/min. The
total flow rate was 1.0 mL/min, corresponding to a residence
time of 20 min. The reactor was run under these conditions for
35 min before the reaction mixture was collected during 43 min
=
1 mm), LFT-MS (volume = 0.2 mL, channel size = 1 mm),
and LTF-VS (volume = 1.1 mL, channel size = 1 mm) were
purchased from Little Things Factory (www.ltf-gmbh.com).
The residence time units were prepared from Teflon tubing.
For heating, the reactors were immersed in a water bath. The
feed was pumped using an NE-1010 syringe pump from New
Era Pump Systems Inc. (www.syringepump.com).
Syntheses of Spiro[2.4]hepta-4,6-diene (1). Semibatch
Synthesis (Table 1, Entry 2). A reactor was charged with freshly
prepared cyclopentadiene (9.0 g, 7.2 mL, 136.2 mmol, 1 equiv),
(theoretical yield of 1: 43.2 mmol) by quenching into H O
2
(150 mL) at 0 °C. The dark-brown emulsion was diluted with
heptane (50 mL), and the phases were separated. The organic
layer washed twice with H
solution (50 mL), and twice with H
layer was dried over MgSO and filtered, and 1 was obtained as
O (50 mL), once with 1 M HCl
2
O (50 mL). The organic
2
MeBu NCl (320 mg, 1.36 mmol, 0.01 equiv), 2,6-di-tert-butyl-
4
3
4-methylphenol (BHT) (20 mg, 0.1 mmol), and 1,2-dichloro-
a solution in heptane (43.3 g). This solution contained 5% w/w
1
ethane (13.5 g, 16.7 mL, 136.2 mmol, 1 equiv). The
temperature was set to 30 °C and the stirrer speed to 500
rpm. To the clear solution was dosed 50% aqueous NaOH
solution (88.0 g, 58 mL, 1.1 mol, 8 equiv) over 1 h. During the
addition, the reaction mixture became dark brown and viscous.
After an additional 30 min, the reaction mixture was cooled to
1 as determined by H NMR assay, corresponding to a yield of
54%.
“RRSM” Flow Reactor (Table 5, Entry 6). The flow reactor
consisted of a T-mixer, an RRSM (1.5 mL, length = 50 cm, i.d.
= 2.4 mm), a coiled Teflon tube (18 mL, length = 400 cm, i.d.
= 2.4 mm) kept at 25 °C, an additional coiled Teflon tube (18
mL, length = 400 cm, i.d. = 2.4 mm) kept at 40 °C, a second
RRSM (1.5 mL, length = 50 cm, i.d. = 2.4 mm), and an
additional coiled Teflon tube (36 mL, length = 800 cm, i.d. =
2.4 mm) kept at 60 °C. The total volume was 75 mL (the
configuration is shown in Table 5). Feed 1 (4.4 mmol of
cyclopentadiene/mL) was prepared by mixing cyclopentadiene
(29.0 g, 36.3 mL, 439 mmol, 1 equiv), 1,2-dichloroethane (60.8
1
0 °C and diluted with H O (60 mL) while the temperature
2
rose to 23 °C. The dark-brown emulsion was diluted with
heptane (90 mL), and the phases were separated. The organic
layer was washed twice with H O (40 mL), once with 1 M HCl
2
solution (40 mL), and twice with H O (40 mL). The organic
2
layer was dried over MgSO and filtered, and 1 was obtained as
4
1
a solution in heptane (88.7 g). By H NMR assay with 1,4-
F
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
g, 49.2 mL, 614 mmol, 1.4 equiv), and MeBu NCl (14.5 g, 61.5
ASSOCIATED CONTENT
Supporting Information
H NMR spectra and pictures of flow reactors. This material is
3
■
mmol, 0.14 equiv) and was pumped at a flow rate of 0.663 mL/
min. Feed 2 consisted of 30% aqueous NaOH solution and was
pumped at a flow rate of 2.337 mL/min. The total flow rate was
*
S
1
available free of charge via the Internet at http://pubs.acs.org.
3.0 mL/min, corresponding to a residence time of 25 min. The
reactor was run under these conditions for 40 min before the
reaction mixture was collected during 1 h 20 min (theoretical
AUTHOR INFORMATION
Corresponding Author
E-mail: roger.marti@hefr.ch. Phone: +41 26 429 6703. Fax:
41 26 429 6600.
■
yield of 1: 232.5 mmol) by quenching into H O (100 mL) and
2
*
heptane (200 mL) at 0 °C, which afforded a dark-brown
emulsion. The phases were separated, and the organic layer was
+
Notes
washed twice with H O (50 mL), once with 1 M HCl solution
2
The authors declare no competing financial interest.
(
50 mL), and twice with H O (50 mL). The organic layer was
2
dried over MgSO and filtered, and 1 was obtained as a
4
ACKNOWLEDGMENTS
■
yellowish solution in heptane (188.3 g). This solution
1
We thank Daniel Meyer and Tamara Wyss for valuable lab
support for the synthesis of the Diels−Alder adduct as well as
tests with the new RRSM.
contained 10.5% w/w 1 as determined by H NMR assay,
corresponding to a yield of 95%.
Synthesis of (1R,4S,5R,6R)-Bis((S)-1-ethoxy-1-oxopro-
pan-2-yl) Spiro[bicyclo[2.2.1]hept-2-en-7,1′-cyclopro-
pan]-5,6-dicarboxylate (2). To a solution of 1 in heptane
REFERENCES
■
(
1) Bur, D.; Corminboeuf, O.; Cren, S.; Grisostomi, C.; Leroy, X.;
Richard-Bildstein S. WO/2010134014 A1, 2010.
2) (a) Coe, J. W.; Wirtz, M. C.; Bashore, C. G.; Candler, J. Org. Lett.
004, 6, 1589−1592. (b) van Beek, J. A. M.; Gruter, G. J. M.; Green,
R. U.S. Patent 6,232,516, 2001.
3) Abele, S.; Schmidt, G.; Funel, J.-A.; Schwaninger, M.; Wagner, S.
(
154.5 g, NMR assay 10.5% w/w, 176 mmol) was added
(−)-bis[(S)-1-(ethoxycarbonyl)ethyl] fumarate (66.8 g, 211
(
2
mmol, 1.2 equiv) at rt. The mixture was stirred for 48 h at rt
and then concentrated under reduced pressure to yield 2 (84.8
g, NMR assay 83% w/w, corresponding to an assay-corrected
yield of 98%, d.r. 94:6) as a yellow oil. For analytical purposes, a
(
ACS Symp. Ser. 2014, 1181, 189−210.
2
.00 g sample of crude 2 was purified by column
chromatography (EtOAc/heptane 15:85) to yield pure 2
1.58 g, 79%) as a colorless oil. R 0.30 (EtOAc/heptane
(4) am Ende, D. J.; Whritenour, D. C.; Coe, J. W. Org. Process Res.
Dev. 2007, 11, 1141−1146.
(
̈
5) (a) Illg, T.; Lob, P.; Hessel, V. Bioorg. Med. Chem. 2010, 18,
(
f
1
3707−3719. (b) Hessel, V.; Cortese, B.; de Croon, M. H. J. M. Chem.
1
0
7
5:85); H NMR (300 MHz, CDCl ) δ 0.36−0.58 (m, 3H),
3
Eng. Sci. 2011, 66, 1426−1448. (c) Hessel, V.; Kralisch, D.;
.60−0.76 (m, 1H), 1.27 (td, J = 7.1, 1.3 Hz, 6H), 1.49 (dd, J =
.1, 4.6 Hz, 6H), 2.74−2.82 (m, 1H), 2.82−2.87 (m, 1H), 2.90
̈
Kockmann, N.; Noel, T.; Wang, Q. ChemSusChem 2013, 6, 746−789.
(
6) Yoshida, J. Flash Chemistry: Fast Organic Synthesis in Microsystems;
Wiley-Blackwell: Chichester, U.K., 2008.
(7) (a) Abele, S.; Hock, S.; Schmidt, G.; Funel, J.-A.; Marti, R. Org.
(
d, J = 4.6 Hz, 1H), 3.74 (dd, J = 4.1, 4.3 Hz, 1H), 4.19 (qd, J =
7
2
0
1
6
1
.1, 0.8 Hz, 3H), 4.13−4.25 (m, 1H), 5.07 (dq, J = 15.8, 7.1 Hz,
̈
Process Res. Dev. 2012, 16, 1114−1120. (b) Reichart, B.; Tekautz, G.;
Kappe, C. O. Org. Process Res. Dev. 2013, 17, 152−157. (c) Zaborenko,
N.; Bedore, M. W.; Jamison, T. F.; Jensen, K. F. Org. Process Res. Dev.
H), 6.30 (ddd, J = 5.8, 2.8, 0.7 Hz, 1H), 6.42 (ddd, J = 5.8, 3.1,
.7 Hz, 1H); 13C NMR (75 MHz, CDCl ) δ 4.9, 8.5, 14.12,
3
4.14, 16.98, 16.99, 44.2, 47.2, 48.7, 51.9, 52.0, 61.25, 61.31,
8.6, 68.7, 135.4, 137.4, 170.7, 170.8, 172.5, 172.6; IR (neat)
017, 1048, 1092, 1131, 1168, 1450, 1734, 2942, 2988, 3070
2
011, 15, 131−139. (d) Hornung, C. H.; Mackley, M. R.; Baxendale, I.
R.; Ley, S. V. Org. Process Res. Dev. 2007, 11, 399−405. (e) White, T.
D.; Alt, C. A.; Cole, K. P.; McClary Groh, J.; Johnson, M. D.; Miller, R.
D. Org. Process Res. Dev. 2014, 18, 1482−1491. (f) Buono, F. G.;
Eriksson, M. C.; Yang, B.-S.; Kapadia, S. R.; Lee, H.; Brazzillo, J.;
Lorenz, J. C.; Nummy, L.; Busacca, C. A.; Yee, N.; Senanayake, C. Org.
Process Res. Dev. 2014, 18, 1527−1534.
−1
20
D
cm ; [α] −105.7 (c 2.0, EtOH).
Synthesis of (1R,4S,5R,6R)-Spiro[bicyclo[2.2.1]hept-2-
en-7,1′-cyclopropan]-5,6-dicarboxylic Acid (3). To a
solution of crude 2 (82.8 g, NMR assay 83% w/w, 0.168
mol) in MeOH (250 mL) was added a 16% aqueous solution of
NaOH (250 mL). The mixture was stirred for 30 min at 65 °C,
whereupon the starting emulsion disappeared. Methanol was
evaporated off, and the residue was diluted with water (150
mL). The mixture was treated with charcoal and filtered
through HyFlo. The product was precipitated by the addition
of 32% HCl (135 mL). The suspension was cooled to 5 °C,
filtered, washed with cold water, and dried under vacuum to
(
2
8) (a) Rodrigues, T.; Schneider, P.; Schneider, G. Angew. Chem.
014, 126, 5858−5866. (b) Hartwig, J.; Metternich, J. B.; Nikbin, N.;
Kirschning, A.; Ley, S. V. Org. Biomol. Chem. 2014, 12, 3611−3615.
(
c) Anderson, N. G. Org. Process Res. Dev. 2012, 16, 852−869.
(9) (a) Roberge, D. M.; Ducry, L.; Bieler, N.; Cretton, P.;
Zimmermann, B. Chem. Eng. Technol. 2005, 28, 318−323. (b) Plouffe,
P.; Macchi, A.; Roberge, D. M. Org. Process Res. Dev. 2014, 18, 1286−
1
(
294.
10) (a) Nagy, K. D.; Shen, B.; Jamison, T. F.; Jensen, K. F. Org.
Process Res. Dev. 2012, 16, 976−981. (b) Leduc, A. B.; Jamison, T. F.
yield 3 (31.7 g, 91%) as a white crystalline solid. R 0.30
f
Org. Process Res. Dev. 2012, 16, 1082−1089.
1
(
EtOAc/heptane 1:1); mp 161−163 °C; H NMR (300 MHz,
(
11) (a) Kashid, M. N.; Kiwi-Minsker, L. Ind. Eng. Chem. Res. 2009,
DMSO-d ) δ 1.30−1.39 (m, 2H), 0.40−0.52 (m, 2H), 2.54−
48, 6465−6485. (b) Pastre, J. C.; Browne, D. L.; O’Brien, M.; Ley, S.
6
2
1
.64 (m, 3H), 3.42 (t, J = 4.2 Hz, 1H), 6.11 (dd, J = 5.5, 2.6 Hz,
L. Org. Process Res. Dev. 2013, 17, 1183−1191.
H), 6.37 (dd, J = 5.5, 3.0 Hz, 1H), 12.23 (s, 2H); 13C NMR
(12) (a) Zhao, Y.; Chen, G.; Yuan, Q. AIChE J. 2006, 52, 4052−
4
060. (b) Nieves-Remacha, M. J.; Kulkarni, A. A.; Jensen, K. F. Ind.
(
75 MHz, DMSO-d ) δ 4.7, 8.2, 44.2, 47.5, 48.4, 50.5, 50.6,
6
Eng. Chem. Res. 2012, 51, 16251−16262.
13) Bannock, J. H.; Krishnadasan, S. H.; Heeney, M.; de Mello, J. C.
Mater. Horiz. 2014, 1, 373−378.
14) (a) Reddy, G. S.; Reddy, N. S.; Manudhane, K.; Krishna, M. V.
R.; Ramachandra, K. J. S.; Gangula, S. Org. Process Res. Dev. 2013, 17,
1272−1276. (b) Ufer, A.; Mendorf, M.; Ghaini, A.; Agar, D. W. Chem.
1
1
34.7, 137.6, 174.10, 174.12; IR (neat) 669, 695, 752, 1163,
(
−1
20
385, 1423, 1683, 1718, 2976, 3170 cm ; [α] −64.4 (c 2.0,
D
EtOH); HPLC (ChiralPak AD-H, 4.6 mm × 250 mm, 5 μm;
heptane/EtOH 0.1% TFA 60:40; 0.8 mL min ; 25 °C; 210
nm) t (minor) 8.6 min, t (major) 9.9 min, e.r. = 97:3.
(
−1
R
R
G
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Eng. Technol. 2011, 34, 353−360. (c) Ahmed-Omer, B.; Barrow, D.;
Wirth, T. Chem. Eng. J. 2008, 135, S280−S283.
(
15) Ueno, M.; Hisamoto, H.; Kitamori, T.; Kobayashi, S. Chem
Commun. 2003, 936−937.
16) Jovanovic, J.; Rebrov, E. V.; Nijhuis, T. A.; Hessel, V.; Schouten,
J. C. Ind. Eng. Chem. Res. 2010, 49, 2681−2687.
17) Jovanovic, J.; Hengeveld, W.; Rebrov, E. V.; Nijhuis, T. A.;
Hessel, V.; Schouten, J. C. Chem. Eng. Technol. 2011, 34, 1691−1699.
18) Naber, J. R.; Buchwald, S. L. Angew. Chem. 2010, 122, 9659−
664.
19) Bogdan, A.; McQuade, D. T. Beilstein J. Org. Chem. 2009, 5,
No. 17.
(
(
(
9
(
(
20) Moffett, R. B. In Organic Syntheses; Wiley: New York, 1963;
Collect. Vol. IV, pp 238−243.
(
21) Frank, O.; Kreissel, J. K.; Daschner, A.; Hofmann, T. J. Agric.
Food Chem. 2014, 62, 2506−2515.
22) Funel, J.-A.; Abele, S. Angew. Chem., Int. Ed. 2013, 52, 3822−
863.
23) (a) Kockmann, N.; Gottsponer, M.; Roberge, D. M. Chem. Eng.
(
3
(
J. 2011, 167, 718−726. (b) Allian, A. D.; Richter, S. M.; Kallemeyn, J.
M.; Robbins, T. A.; Kishore, V. Org. Process Res. Dev. 2011, 15, 91−97.
(
c) Mendorf, M.; Nachtrodt, H.; Mescher, A.; Ghaini, A.; Agar, D. W.
Ind. Eng. Chem. Res. 2010, 49, 10908−10916.
H
DOI: 10.1021/acs.oprd.5b00046
Org. Process Res. Dev. XXXX, XXX, XXX−XXX