Hydride reduction of a lactate ester: Optimisation and scale-up

Article abstract of DOI:10.1021/op0200064

The two-step synthesis of (2S)-2-(tetrahydropyran-2-yloxy)-propane-1-ol was selected for fine chemical scale-up. The first step, an acid-catalyzed protection of ethyl (S)(-)-lactate with 3,4-dihydro-2H-pyran, could be performed without solvent. The reaction enthalpy was determined to be -48 kJ/mol ethyl (S)-(-)-lactate, ensuring safe scale-up. The crude reaction mixture was used in the second step, the hydride reduction. For the reduction of the protected ester a clear solution of 1 M LiAlH₄ in THF was used. The reaction enthalpy was determined to be -313 kJ/mol ethyl (2S)-2-(tetrahydropyran-2-yloxy)propionate, and the work-up was optimised with experimental design techniques. Combination of all experimental and theoretical results resulted in a master recipe for a 10-dm³ scale synthesis of (2S)-2-(tetrahydropyran-2-yloxy)propane-1-ol.

Full text of DOI:10.1021/op0200064

Organic Process Research & Development 2002, 6, 606−610
Hydride Reduction of a Lactate Ester: Optimisation and Scale-Up
M. P. J. Donners, M. C. Hersmis, J. P. A. Custers, J. Meuldijk, J. A. J. M. Vekemans, and L. A. Hulshof*
EindhoVen UniVersity of Technology, Laboratory of Macromolecular and Organic Chemistry and Process DeVelopment
Group, Department of Chemical Engineering and Chemistry, P.O. Box 513, 5600 MB EindhoVen, The Netherlands
Scheme 1
Abstract:
The two-step synthesis of (2S)-2-(tetrahydropyran-2-yloxy)-
propane-1-ol was selected for fine chemical scale-up. The first
step, an acid-catalyzed protection of ethyl (S)(-)-lactate with
3,4-dihydro-2H-pyran, could be performed without solvent. The
reaction enthalpy was determined to be -48 kJ/mol ethyl (S)-
(-)-lactate, ensuring safe scale-up. The crude reaction mixture
was used in the second step, the hydride reduction. For the
reduction of the protected ester a clear solution of 1 M LiAlH4
in THF was used. The reaction enthalpy was determined to be
-313 kJ/mol ethyl (2S)-2-(tetrahydropyran-2-yloxy)propionate,
and the work-up was optimised with experimental design
techniques. Combination of all experimental and theoretical
results resulted in a master recipe for a 10-dm3 scale synthesis
of (2S)-2-(tetrahydropyran-2-yloxy)propane-1-ol.
absence of hydrogen production, lower solubility of inter-
mediates in water, and less consumption of the expensive
LiAlH4.
An extensive study was initiated to design the optimal
process for the production of (2S)-2-(tetrahydropyran-2-
yloxy)propane-1-ol (3) on a 10-dm3 scale in a fully automated
(semi)batch-wise operated miniplant. In this design study the
reactions are optimised, taking into account process safety
as well as environmental considerations.
Introduction
Fine chemical process research and development limits
itself in various aspects. In contrast to bulk chemicals, fine
chemicals are usually produced with multi-purpose (semi)
batch-wise operated equipment and have a relatively short
lifetime. The demands for a short time-to-market of new fine
chemicals asks for a challenging batch process design. The
fine chemical process research and development requires a
fast and a systematic approach in the scale-up of fine
chemical processes leading to an early recognition of scale-
up surprises. This can lead to thorough understanding of the
failures and key issues for scaling-up processes prior to (pilot)
plant implementation.
To design tools and a methodology for fine chemical
scale-up the synthesis of a chiral synthon, (2S)-2-(tetrahy-
dropyran-2-yloxy)propane-1-ol (3), was selected. This chiral
building block can be used in the synthesis of liquid
crystalline disc-shaped molecules1
For the two-step synthesis2 of (3), as depicted in Scheme
1, ethyl (S)(-)-lactate (1) was used as starting material. In
the first step 1 is protected with 3,4-dihydro-2H-pyran using
p-toluene sulphonic acid as the acid catalyst. In the second
step the protected lactate ester 2 is reduced with LiAlH4.
(S)(+)-Propane-1,2-diol being an interesting chiral C3-
synthon for other applications can also be synthesised by
deprotection of 3. This indirect route is preferred over the
direct reduction of ethyl (S)(-)-lactate (1). Advantages of
the indirect synthesis of (S)(+)-propane-1,2-diol are the
Synthesis
For the optimisation and scale-up study literature recipes
had to be adjusted and further investigated. Diethyl ether is
generally used2-4 as a solvent in both acid-catalyzed protec-
tion reactions and reduction reactions with LiAlH4. However,
in a large-scale protection (step 1) and reduction reaction
(step 2) diethyl ether is not a suitable solvent due to its
flammability and potential peroxide formation. Since both
the protection as well as the reduction with LiAlH4 are
exothermic reaction steps, the reaction enthalpy for both steps
had to be determined.
To select the best solvent for the acid-catalyzed protection
of ethyl (S)(-)-lactate (1) tetrahydrofuran (THF), tert-butyl
methyl ether (MTBE), and toluene were screened as alterna-
tive solvents on small scale (100 mL). The concentration of
ethyl (S)(-)-lactate (1) was 5 kmol/m3 solvent and that of
the catalyst was 0.02 mol % based on 1. The reaction
temperature was kept at 50 °C by dropwise addition of 3,4-
dihydro-2H-pyran. Thus, the reaction proceeded under
(3) (a) Asakawa, M,; Ashton, P. R.; Hayes, W.; Janssen, H. M.; Meijer, E. W.;
Menzer, S.; Pasini, D.; Stoddart, J. F.; White, A. J. P.; Williams, D. J. J.
Am. Chem. Soc. 1998, 120, 920. (b) Buchwald, S. L.; Pliura, D. H.; Knowles,
J. R. J. Am. Chem. Soc. 1984, 106, 4916.
(4) (a) Perkins, M. V.; Kitching, W.; Ko¨nig, W. A.; Drew, R. A. I. J. Chem.
Soc., Perkin Trans. 1 1990, 2501. (b) Cowie, J. M. G.; Hunter, H. W.
Makromol. Chem. 1990, 191, 1393. (c) Chiellini, E.; Galli, G.; Carrozzino,
S.; Gallot, B. Macromolecules 1990, 23, 2106. (d) Cheskis, B. A.; Shpiro,
N. A.; Moiseenkov, A. M. J. Org. Chem. USSR (Engl. Transl.) 1990, 26,
1613. (e) Huszthy, P.; Oue, M.; Bradshaw, J. S.; Zhu, C. Y.; Wang, T.;
Dalley, N. K.; Curtis, J. C.; Izatt, R. M., J. Org. Chem. 1992, 57, 5383. (f)
Neubert, M. E.; Sabol-Keast, S., Mol. Cryst. Liq. Cryst. 1990, 188, 67.
* Author for correspondence: E-mail: L.A.Hulshof@tue.nl.
(1) Brunsveld, L.; Zhang, H.; Glasbeek, M.; Vekemans, J. A. J. M.; Meijer, E.
W. J. Am. Chem. Soc. 2000, 122, 6175.
(2) Ghirardelli, R. G. J. Am. Chem. Soc. 1973, 95, 4987.
606
•
Vol. 6, No. 5, 2002 / Organic Process Research & Development
10.1021/op0200064 CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/21/2002

Table 1. Variables and minimum and maximum levels used
in the factorial design
“starving conditions”. To force the reaction to complete
conversion of 1 an excess of 3% 3,4-dihydro-2H-pyran was
added. Reactions using THF or MTBE as solvent were much
slower than those using toluene. This presumably results from
hydrogen bonding between the catalyst and the two former
solvents.
To reduce the number of operations and the amount of
solvent the protection reaction was also investigated in the
absence of solvent. In that case the crude reaction mixture
might be used as such in the subsequent reduction step.
Performing the reaction without solvent did not induce
product loss. The crude yield determined by GC analysis
was 97%. The only difference with the reactions in a solvent
was a stronger temperature rise during the reaction due to
the higher concentration.
To ensure proper chemical control of the reaction on a
larger scale, the reaction enthalpy was determined for a 1-dm3
isothermal reaction at 323 K to be -48 kJ/mol ethyl (S)-
(-)-lactate (1). Since the reaction takes place under “starving
conditions”, the maximum allowed reaction temperature of
50 °C will be maintained by the addition rate of 3,4-dihydro-
2H-pyran.
Xi
variable
(+)
0
(-)
X1 stirring speed (rpm)
X2 base
X3 temperature (°C)
400
300
200
LiOH LiOH/NaOH (1:1) NaOH
50
37
2
25
2
X4 stirring afterward (min) 30
Table 2. Fractional factorial design: experimental matrix
and results
variables
responses
tfilt
(min)
WEff
%
Ytot
%
expt
X1
X2
X3
X4
1
2
3
4
5
6
7
8
9
2*
-
+
-
+
-
+
-
+
0
-
-
+
+
-
-
+
+
0
-
-
-
-
+
+
+
+
0
-
+
+
-
-
+
-
+
-
+
3
3
6
4
7
2
2.75
2.5
2
85
92
84
84
93
89
85
88
88
91
95
93
88
92
85
95
97
84
95
92
+
-
-
3
Optimisation of the reduction step started with 100-mL-
scale experiments. Addition of a clear solution of 1 M LiAlH4
in THF to the crude reaction mixture gathered from the
protection reaction was chosen due to three main advantages.
First it is commercially available, second the addition of a
transparent solution is easier than the addition of a suspen-
sion, and finally, the risk of accumulation of the reductant
during the addition is limited. The initial temperature for
the reduction step was 0 °C. The feed rate of the LiAlH4
solution determined the temperature rise of the reaction
mixture. To safely scale up this instantaneous reaction the
reaction enthalpy of the reduction was determined first. It
amounted to be -313 kJ/mol protected ester (2) for a 1 dm3
reaction at 50 °C. Since the reaction takes place under
“starving conditions”, the maximum reaction temperature will
be governed by the addition rate of the LiAlH4 solution. To
force the reaction to completion an excess of 1 mol % of
LiAlH4 was used.
To obtain pure (2S)-2-(tetrahydropyran-2-yloxy)propane-
1-ol (3), the reaction mixture was subjected to an effective
work-up and subsequent purification by distillation. The
method to work up the reaction mixture described in the
literature is by adding per kilogram consumed LiAlH4: 1
dm3 water, 1 dm3 15% caustic soda, and 3 dm3 water, in the
given sequence.5 Although gel formation was observed with
this work-up procedure, a perfectly filterable suspension was
obtained after vigorous stirring. The solid particles were
separated by filtration, and the solid residue was washed with
four equal amounts of THF. The solvent was evaporated,
and the residual product was purified by distillation. The
yield after distillation was 96%.
effect of the stirring speed, the temperature, and the used
base on the yield and the filtration time was not known. To
maximize the yield, implying a good filterability, the process
conditions during work-up were optimised using a fractional
factorial design.
The selected variable process parameters for a 24-1
fractional factorial design were stirring speed (N) of the
pitched-blade stirrer, temperature (T), and the base used in
the work-up (B). Neglecting the three-factor interaction of
N, B, and T, the stirring time R was introduced as a fourth
variable in an eight experiments factorial design. The stirring
time R is the time between the visual change of the gel into
a suspension and the discharge of the reaction vessel.
The chosen variables and the minimum and maximum
considered values of these variables are depicted in Table
1. In the central column the “centre point” experiments are
shown. The values are used for 1-dm3-scale reactions.
Because of practical problems during scale-up the values of
the variables were partially fixed. A stirring speed lower than
200 rpm was expected to cause problems due to the thickness
of the gel, while a translation of energy input to larger scale
was considered impossible with a stirring speed exceeding
400 rpm. Temperatures above 50 °C were considered not
suitable in view of the close boiling point of THF. NaOH
was chosen as a base since it is generally used for the work-
up of lithium aluminum hydride reductions and it is cheap.
LiOH was chosen as a substitute since lithium salts are
already present in the reaction mixture.
In the four variables fractional factorial design three
responses were studied: the mean filtration time of four
washing steps when triturating the filter cake after every
washing step (tfilt), the washing efficiency after the first
washing step (WEff), and the total yield after four washing
steps (Ytot).
Fractional Factorial Design
An instantaneous reaction has no advantages if the work-
up is slow or difficult or when a lot of product is lost. The
(5) Mic´ovic´, V. M.; Michailovic´, M. L. J. Org. Chem. 1953, 18, 1190.
Vol. 6, No. 5, 2002 / Organic Process Research & Development
•
607

Table 3. Model equations per responsea
model equation
R2
Q2
RSD
tfilt
WEff
Ytot
3.6 - 0.9X1 + 0.2X2 - 0.1X3 + 1.0X4 - 0.4X1X3 - 0.9X2X3
87.8 + 0.8X1 - 2.2X2 + 1.2X3 - 1.5X4 - 0.9X1X3
91.3 - 1.2X2 - 0.8X3 - 3.9X4 - 2.1X1X2 + 1.2X2X3
0.93
0.95
0.98
0.31
0.75
0.62
0.77
1.07
1.09
a R2 ) correlation coefficient; Q2 ) predictability of model; RSD ) residual standard deviation.
Table 4. Optimal configuration per response
predicted
T (°C) R (min) optimum
N (rpm)
B
t
filt (min)
WEff (%)
tot (%)
400
300
400
LiOH
NaOH
NaOH
50
50
25
2
30
2
0.4
93.0
100.4
Y
For large-scale experiments it is of upmost importance
that the filtration time (tfilt) can be influenced. The washing
efficiency after one washing step (WEff) multiplied by the
isolated yield (Ytot) demonstrates the yield after one washing
step. All three responses determine the productivity of the
process during scale-up. However, reduction of washing steps
or volume of washing fluids gives a more sustainable process.
Therefore, the most important response is the yield after one
washing step.
The influence of the four variables was studied at 1-dm3-
scale with eight experiments, a centre point experiment, and
a duplicate experiment. In Table 2 the experimental matrix
and the results are depicted showing the measured values
for the five chosen responses. Experiment 9 is the “centre
point” experiment and experiment 2* is the duplicate
experiment of experiment 2.
Figure 1. Response surface X1 (stirring speed N from 400 to
200 rpm) versus X4 (stirring time R from 2 to 30 min) for the
filtration time. Other variables Xi ) 0.
The regression analysissusing the Modde program6sof
this factorial design resulted into models that would fit the
experimental data as given in the model equations shown in
Table 3. It can be concluded that the equations for filtration
time, washing efficiency, and total yield describe the
experimental values best and that the predictability of the
model is varying. The most optimal configuration per
response according to the model equations is depicted in
Table 4 and Figures 1 and 2.
The experiment using a stirring speed of 400 rpm, NaOH
as the base, a temperature of 50 °C and a stirring time of 2
min (X1 ) 1, X2 ) -1, X3 ) 1, X4 ) -1) is the theoretically
best experiment when the filtration time (tfilt), and both yields
(WEff and Ytot) are assumed to be the crucial responses. The
experimental and calculated results of the optimal experi-
ment7 are shown in Table 5. The experimental values
correspond well with the calculated values.
Large-Scale Synthesis. The insights and results of the
small-scale experiments and the theoretically best work-up
experiment for the reduction step (N ) 400 rpm, B ) NaOH,
T ) 50 °C, R ) 2 min) were translated into a master recipe
for the 10-dm3-scale process.
Figure 2. Response surface X2 (nature of the base B, NaOH
(-1) and LiOH (+1)) versus X4 (stirring time R from 2 to 30
min) for the total yield. Other variables Xi ) 0.
10-dm3 reactor, charged with a mixture of ethyl (S)(-)-lactate
and p-toluene sulphonic acid. For the hydride reduction step,
the reactor was filled with the crude reaction mixture from
the first step. The hydride solution was also added with the
aid of a membrane pump under “starving conditions” within
3 h. To complete the reaction the reaction mixture was stirred
for 1 h at 50 °C. The reaction mixture was worked up by
adding consecutively water, caustic soda, and water. After
addition of the final amount of water a perfectly filterable
suspension was obtained. The suspension was filtered on a
For the first step, 3,4-dihydro-2H-pyran was added
automatically within 2 h under “starving conditions” to the
(6) Modde 4.0, Umetri AB, Umeå, Sweden.
(7) The optimal experiment was selected as the best practical one in our case.
608
•
Vol. 6, No. 5, 2002 / Organic Process Research & Development

equipped with a 2-dm3 HP60 stainless steel high-pressure
reactor with a pitched-blade impeller. The 1-dm3-scale
reactions were performed in a glass reactor with a 45°
pitched-blade impeller with six blades (diameter impeller )
5 cm; diameter reactor ) 10 cm; clearance ) 1 cm). The
large-scale reaction was carried out in a 10-dm3 fully
automated (semi)batch-operated Belatec reactor equipped
with a 45° pitched-blade impeller with six blades (diameter
impeller ) 10 cm; diameter reactor ) 20 cm; clearance )
2 cm). The reactor is controlled by a PLC, a special computer
monitoring physical data (batch history), and securing
optimal process and safety conditions.
Ethyl (2S)-2-(tetrahydropyran-2-yloxy)propionate (2).
Ethyl (S)(-)-lactate (1) (228.5 mL, 236.3 g, 2.00 mol) and
p-toluene sulphonic acid monohydrate (80 mg, 0.42 mmol,
0.02 mol %) were added to toluene (400 mL). The reaction
mixture was heated to 50 °C. Within 60 min 3,4-dihydro-
2H-pyran (188.0 mL, 173.3 g, 2.06 mol) was added dropwise
to the reaction mixture. Then, the reaction mixture was
washed with saturated NaHCO3 solution (100 mL). Drying
of the toluene layer with Na2SO4 (10 g, 0.07 mol) and
evaporation of the toluene yielded the crude product.
Distillation (75-77°C; 1.6 mbar) yielded pure (2) as a
colourless oil (385 g, 1.90 mol, 95%).
Table 5. Validation optimal experiment
experiment
(1,-1,1,-1)
predicted
RSD
experimental
t
filt (min)
WEff (%)
tot (%)
2.0
92.5
96.4
0.77
1.24
1.09
2
89
95
Y
Table 6. Filterability and cake resistance
scale
experiment
(1,-1,1,-1)
pre-
dicted RSD
1 dm3
10 dm3
t
Y
filt (min)
tot (%)
2.0 0.77
96.4 1.09 95
2
7
99
filter area (m2)
6.36 × 10-3 4.91 × 10-2
cake resistance (1/m2)
filterability (m3/kg)
5 × 1013
9.9 × 1013
7.3 × 10-11 3.6 × 10-11
Bu¨chner filter, and the residue was washed with four equal
amounts of THF. The filtration time was 7 min, and the
filterability was calculated to be 3.6 × 10-11 m3/kg. The
washing efficiency (WEff) was 87%. To obtain the crude
product the filtrate was evaporated. The product was isolated
in 99% yield.
The cake resistance and filterability on 10-dm3-scale were
compared with the 1 dm3 results showing a longer filtration
time for reactions on 10-dm3-scale, as can be seen in Table
6. Isolation by centrifuge on 6-m3-scale would not be suitable
in view of these results (relatively high cake resistance and
a factor 2 deterioration in filtration performance) and instead
a belt filter is recommended for large-scale synthesis.8
1H NMR (CDCl3, 300 MHz) δ 4.65 (1H, m, O-CHR-O),
4.35 (0.7H, q, O-CH′-(COOEt)CH3), 4.2-4.1 (2.3H, m,
O-CH2-CH3 and O-CH′′-(COOEt)CH3), 3.9-3.35 (4H, m,
RCH2-CH2-O of DHP-group), 1.95-1.45 (12H, m, O-CH2-
(CH2)3-CHO2 of the THP-group), 1.4-1.3 (6H, dd, CH-CH3-
(COOEt)), 1.25 (6H, m, O-CH2-CH3); 13C NMR (CDCl3,
100 MHz) δ 173.2 and 173.1 (CdO), 98.1 and 97.4 (O-
CHR-O), 72.4 and 69.8 (O-CH(COOEt)CH3), 62.3 and 62.1
(CH2-O-CHR-O), 60.7 and 60.6 (O-CH2-CH3), 30.3 and 30.2
(CH-CH2), 25.3 and 25.2 (CH2-CH2-O), 19.0 and 18.9 (CH-
CH2-CH2-CH2-CH2O), 18.6 and 17.9 (O-CH(COOEt)CH3),
14.1 and 14.0 (O-CH2-CH3).
Conclusions
The synthesis of (2S)-2-(tetrahydropyran-2-yloxy)propan-
1-ol (3) was successfully optimised and scaled up to a 10-
dm3 reactor. The work-up of the hydride reduction was
optimised by using a fractional factorial design. Experience
with the centrifuge performance of many large-scale pro-
cesses gave valuable insight in the best choice of the proper
isolation equipment.8 The above example illustrates that
important scale-up features can be identified on lab-scale
and can be reliably translated to larger scale.
The results of this research are useful to expand the
current fine chemical scale-up methodology to industrial
scale.
[R]20 ) -61.7° (c 3.19, CHCl3) (lit. [R]20 ) -60.7°
D
D
(c 3.19, CHCl3)).9
The reaction conditions of the reaction for the 1-dm3-
scale reactor without solvent are identical. The amounts of
reactants are: ethyl (S)(-)-lactate (430.0 mL, 444.6 g, 3.76
mol); p-toluene sulphonic acid (0.1433 g, 0.753 mmol, 0.02
mol %); 3,4-dihydro-2H-pyran (354.0 mL, 326.4 g, 3.88
mol).
(2S)-2-(Tetrahydropyran-2-yloxy)propan-1-ol(3). Through
crude (2) (195.0 g, containing 0.96 mol ethyl (2S)-2-
tetrahydropyran-2-yloxy)propionate, 0.01 mol 3,4-dihydro-
2H-pyran and 0.19 mmol p-toluene sulphonic acid) argon
was bubbled for half an hour. Then, 1 M LiAlH4 solution in
THF (500 mL, 0.50 mol) was added dropwise to the reaction
mixture within 1 h. The clear reaction mixture was allowed
to react at room temperature for an additional hour after the
addition of the LiAlH4 solution. The reaction mixture was
quenched by adding water (19.0 mL), 15% caustic soda (19.0
mL), and water (57.0 mL), in that order. In the beginning,
the reaction mixture turned white and formed a gel-like mass.
After thorough stirring a filterable suspension was formed.
The reaction mixture was filtered, and the residue was
Experimental Section
All reagents and solvents were used without further
purification. All NMR spectra were recorded on a Varian
300 MHz spectrometer with TMS as internal standard. GLC
was conducted with a Perkin-Elmer Autosystem instrument
with FID equipped with a CP-SIL 8 column (15 m × 0.25
mm × 0.1 µm). Optical rotations were determined with a
Jasco polarimeter at 20 °C. Reaction calorimetry was
performed with a Mettler Toledo RC1e reaction calorimeter
(8) Hulshof, L. A. Challenges in Fine Chemical Scale-Up: The Battle of
Tweedledee and Tweedledum. Fifth International Conference on Organic
Process Research and Development, New Orleans, LA, U.S.A., November,
2001.
Vol. 6, No. 5, 2002 / Organic Process Research & Development
•
609

washed with THF (150 mL) four times. The filtrate was
evaporated, and the crude product was distilled (83-85 °C;
6.1 mbar) from Na2CO3 (20 g), to obtain the pure colourless
product (147.5 g, 0.92 mol, 96%).
obtain maximum heat transfer, additional solvent (THF, 2.0
dm3) was added, and the mixture was stirred at 400 rpm.
The reactor was cooled to 15 °C, and then a clear solution
of LiAlH4 (4260 g, 1.0 M in THF, 4.73 mol) was added
with the aid of a membrane pump under “starving conditions”
within 3 h. The reaction mixture was stirred to completion
for 1 h at 50 °C. Water (60 mL) was added in 10 min via a
membrane-dosing pump under argon ventilation, and little
hydrogen gas production was observed. Consecutively, water
(120 mL) in 2 min, 15% NaOH (180 mL) in 2 min, and
water (540 mL) were added, while maintaining the temper-
ature at 50 °C. During the addition of the first portion of
water, a white gel-like mass was formed. Continuous stirring
for 2 min during the last phase of the work-up gave a good
filterable suspension. The suspension was filtered using a
Bu¨chner filter in 7 min, and the residue was washed with
THF (1500 mL) four times. The washing efficiency was
87%. The filtrate was evaporated, and the pure colourless
product (3) was obtained in 99% yield (based on 1) in a
1H NMR (CDCl3, 300 MHz) δ 4.65 and 4.45 (2H, m,
O-CHR-O), 4.0-3.6 (4H, m), 3.55-3.3 (6H, m), 2.55 (1H,
m, OH), 1.85-1.4 (13H, m, OH and O-CH2-(CH2)3-CH of
the DHP-group), 1.1-0.9 (6H, dd, O-CH(CH3)-(CH2OH)).
13C NMR (CDCl3, 100 MHz) δ 99.6 and 98.9 (O-CHR-O),
77.3 and 74.6 (O-CH(CH2OH)CH3), 67.0 and 66.0 (CH2-
O-CHR-O), 64.3 and 62.9 (O-CH(CH2OH)CH3), 31.4 and
30.9 (CH-CH2), 25.2 and 24.9 (CH2-CH2-O), 20.7 and 19.9
(CH-CH2-CH2-CH2-CH2O), 17.6 and 17.0 (O-CH2-CH3).
[R]20 ) -8.3° (c 3.15, CHCl3) (lit. [R]20 ) -8.6°
D
D
(c 3.15, CHCl3).9
Master Recipe To Synthesise (2S)-2-(Tetrahydropy-
ran-2-yloxy)propan-1-ol (3) on 10-dm3 Scale. A 10-dm3
fully automated Belatec reactor was charged with ethyl (S)-
(-)-lactate (1) (2521 g, 21.34 mol) and p-toluene sulphonic
acid monohydrate (1.0 g, 5.26 mmol). The reactor was heated
to 50 °C and stirred at 400 rpm. With the aid of a membrane
pump 3,4-dihydro-2H-pyran (1813 g, 21.55 mol) was added
under “starving conditions” within 2 h. The acidic, colourless,
clear solution was neutralised by addition of LiAlH4 (200
mL 1.0 M in THF, 0.2 mol). The reactor was discharged
and rinsed with THF (50 mL). The slightly pale mixture
contained 92.0% (2) and 7.5% THF according to GLC and
1H NMR analysis.
1
purity of 98% according to GC and H NMR analysis.
Acknowledgment
This research is supported by Professor E. W. Meijer
(laboratory of Macromolecular and Organic Chemistry,
TU/e, Netherlands) and DSM Fine Chemicals, Netherlands.
We gratefully acknowledge J. G. M. Moerel, H. G. Fiedler,
P. H. van Eeten, and A. J. Bombeeck for all their efforts
and support in modificating the 10-dm3 Belatec equipment.
S. V. W. de Lima and C. A. Scholtens are acknowledged
for their support in the RC1e measurements.
For the reduction the reactor was charged with the crude
reaction mixture from the protection reaction (1999.7 g). To
(9) Janssen, H. Synthetic Coiled Analogues Based on Modified Ethylene Oxides;
Report of the Post Graduate Research Education, Eindhoven University of
Technology, 1994.
Received for review January 15, 2002.
OP0200064
610
•
Vol. 6, No. 5, 2002 / Organic Process Research & Development