Catalytic hydrogenation of cyanohydrin esters as a novel approach to N-acylated β-amino alcohols - Reaction optimisation by a design of experiment approach

Article abstract of DOI:10.1002/ejoc.200500870

The catalytic hydrogenation of acylated cyanohydrins and subsequent intramolecular migration of the acyl group to yield pharmaceutically interesting N-acyl β-amino alcohols is shown to be a successful one-pot preparation method. The combination of a multi

Full text of DOI:10.1002/ejoc.200500870

FULL PAPER
DOI: 10.1002/ejoc.200500870
Catalytic Hydrogenation of Cyanohydrin Esters as a Novel Approach to
N-Acylated β-Amino Alcohols – Reaction Optimisation by a Design of
Experiment Approach
Lars Veum,[a] Silvia R. M. Pereira,[a] Jan C. van der Waal,[a] and Ulf Hanefeld*[a]
Keywords: N-Acyl β-amino alcohols / β-Amino alcohols / Cyanohydrin esters / Hydrogenation / Design of experiments
The catalytic hydrogenation of acylated cyanohydrins and
subsequent intramolecular migration of the acyl group to
yield pharmaceutically interesting N-acyl β-amino alcohols
is shown to be a successful one-pot preparation method. The
combination of a multistep DoE approach and high-through-
put methodology proved to be an effective strategy for the
optimisation of the reaction. With the favoured catalyst/sol-
vent combination of nickel on alumina in dioxane, both hy-
drogenation and acyl group migration proceeded smoothly,
giving the N-acyl β-amino alcohols in yields (determined by
GC) of up to 90% for aliphatic substrates and up to 50% for
benzylic ones, the latter being more prone to side reactions.
No racemisation was found to occur at the chiral centre of an
aliphatic molecule when an enantiopure cyanohydrin ester
was used, though a minor decrease in ee was observed with
a benzylic substrate.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Introduction
N-Acylated β-amino alcohols such as aegeline (see Fig-
ure 1) occur in nature and can readily be converted into β-
sec-amino alcohols, an important class of compounds in the
pharmaceutical and agrochemical industries. Some repre-
sentative examples of the numerous biologically active β-
sec-amino alcohols are etilefrine, bamethane and denopam-
ine (Figure 1). An established route to N-acyl β-amino
alcohols is by the reduction of the free cyanohydrin, fol-
lowed by acylation of the amino group.[1] The reduction is
usually performed by use of stoichiometric amounts of
either LiAlH4 or BH3, but it can also be achieved by cata-
lytic hydrogenation under strongly acidic conditions.[2] If
enantiopure substrates are used, the stereocentre remains
intact during all these reactions. Given the low atom effi-
ciency of aluminium and boron hydride reductions, and the
strongly acidic conditions required for the catalytic hydro-
genations, a different approach has been investigated, with
the overall aim of integrating the reduction and acylation
steps in a one-pot procedure under mild conditions.
Figure 1. N-Acyl-β-amino alcohols and β-sec-amino alcohols
showing high biological activity.
drogenation of the nitrile group, the newly formed amine,
as a strong nucleophile, can immediately react with the
neighbouring acyl group via a five-membered transition
state to yield the N-acyl β-amino alcohol. This type of in-
tramolecular acyl migration has previously been described
in the NaBH3(OCOCF3) reduction of an acylated cyanohy-
drin to yield denopamine,[5] suggesting that it should pro-
ceed equally well after catalytic hydrogenation of the nitrile.
Earlier reports of catalytic hydrogenations of acylated cy-
anohydrins, and in particular of mandelonitrile esters, de-
scribe the application of Pd/C or PtO2 under strongly acidic
conditions.[6] The primary products obtained were not the
N-acyl β-amino alcohols but β-phenyl ethylamines, owing
to the facile hydrogenation of benzylic C–O bonds over
platinum or palladium catalysts. In this case the amine was
the desired product.[6a] In the current work, however, the
The unprotected cyanohydrins commonly used as start-
ing materials are relatively unstable and racemise easily. In
contrast to this, cyanohydrin esters are stable and do not
racemise. Moreover, they are readily prepared, both in their
racemic[3] and in their enantiopure forms.[4] In addition, the
acyl group of a protected cyanohydrin is a potential intra-
molecular acyl donor (see Scheme 1). After the catalytic hy-
[a] Gebouw voor Scheikunde, Technische Universiteit Delft,
Julianalaan 136, 2628 BL Delft, The Netherlands
Fax: +31-15-278-4289
E-mail: U.Hanefeld@tudelft.nl
1664
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 1664–1671

A Novel Approach to N-Acylated β-Amino Alcohols
FULL PAPER
are therefore well suited to reduce the search space in the
early stages of the research effort. A parameter has a signifi-
cant effect if it produces a positive or negative change that
is above the 95% confidence interval of the variance.
Though continuous refinement of the conditions in subse-
quent optimisation designs, the most influential parameters
can then be studied, and further optimisation achieved.
Since the number of parameters to be investigated has been
reduced, the number of experiments per parameter can be
increased. In this way, more information on the main ef-
fects, and especially on the interactions of the parameters,
can be obtained. In such designs more than 50% of the
possible number of reactions are typically performed. Such
a sequence of DoEs is believed to be a better strategy than
one large one, because the information obtained from one
design is used to improve the next.
Scheme 1. The hydrogenation of acylated cyanohydrins with subse-
quent acyl migration.
objective is to maximize the yields of the N-acylated β-
amino alcohols, and the reductive cleavage of the benzylic
C–O bonds needs to be avoided. This investigation of a se-
lective, catalytic route for the direct conversion of acylated
cyanohydrins (1) into N-acylated β-amino alcohols (2) em-
ployed high-throughput methods for the screening of cata-
lysts, solvents and reaction conditions.
The large number of parameters to be investigated sug-
gested a Design of Experiment (DoE) approach. DoE
methodologies[7] are superior to traditional methods involv-
ing the consecutive optimization of the various parameters;
they make it possible to maximize the amount of infor-
mation that can be obtained from the results while minimiz-
ing the number of experiments, and also increase the likeli-
hood of establishing the true optimum within the search
space. The experiments to be performed are chosen statistic-
ally in order to cover the whole search space as efficiently
as possible. The size of the design (selected number of reac-
tions) depends on the kind of information that is desired.
In the present case, a strategy of three sequential small de-
signs was adopted; this enables the information obtained
from the first to be used to improve the subsequent de-
Results and Discussion
As shown by Hartung,[6a] the hydrogenation of benzylic
cyanohydrin acetates easily yields products such as β-phenyl
ethylamines through reductive cleavage of benzylic C–O
bonds. In an aliphatic substrate, on the other hand, the
C–O bond is more stable and resistant to cleavage even un-
der drastic conditions. Different conditions are likely to be
required for the selective hydrogenation of aliphatic and
benzylic cyanohydrins, so it was decided to optimise the
reactions for mandelonitrile acetate (1a), representative of
benzylic substrates, and 2-cyanohexyl acetate (1c; see
Scheme 2), representative of the aliphatic substrates, sepa-
rately.
First Design
The use of a DoE strategy requires as the first step the
signs.[8] Preliminary small designs (typically less than 25% compilation of all potentially important parameters, based
of the possible number of reactions) are sufficient to distin- on previous experience, the literature or chemical intuition.
guish between significant and insignificant parameters and The initial search space should be broad enough to assure
Scheme 2. The catalytic hydrogenation of cyanohydrin esters.
Eur. J. Org. Chem. 2006, 1664–1671
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1665

L. Veum, S. R. M. Pereira, J. C. van der Waal, U. Hanefeld
FULL PAPER
that the real optimum of the reaction is included. In the of the two substrates by use of a D-Optimal algorithm.[12]
hydrogenation of nitriles the main factor influencing the re- This design is sufficient to study the main effects of each
action rate and the product distribution is the metal in the parameter, and a subsequent more detailed study would
hydrogenation catalyst, with the most commonly used being then allow for further optimisation. Acidic conditions were
Raney nickel, Raney cobalt, Pd/C, Pt/C, Ru/C and Rh/C.[9] not included, since any formation of the amino salts would
The same metals supported on SiO2 and Al2O3 are often prevent the intramolecular migration of the acyl group.
employed as well.[9] Normally Rh, Pd and Pt tend to give Furthermore, in contrast to the free cyanohydrins, the cya-
more secondary and tertiary amines than Co, Ni and Ru. nohydrin esters (1) are more stable towards possible basic
As the migration of the acyl group might suppress the for- side products such as the secondary amines.
mation of secondary and tertiary amines, these metals were
After the execution of the 2×24 reactions, N-acyl β-
also included in the investigation. For the initial screening, amino alcohols 2a and 2c were identified among the prod-
Ni, Pd, Rh, Pt and Ru, on carbon and Al2O3 as carriers, ucts in two of the experiments for each substrate, showing
were selected as representative catalysts.
the hydrogenation to have indeed been followed by intra-
The solvent is a second important parameter. The most molecular acyl migration in a one-pot procedure. The con-
commonly used solvents for the hydrogenation of nitriles ditions for the four successful reactions are given in Table 2.
are protic solvents such as methanol and ethanol. However, The results of the first screening show the advantage of
since the envisaged reaction involved acyl group migration, using DoE with successive small designs as an approach
solvents with a broader range of properties were selected: towards the optimisation of a new reaction. This made it
2-propanol (IPA, a protic solvent but less polar than meth- possible to investigate a large parameter space and to ident-
anol), dioxane (an aprotic, polar ether) and toluene (a rela- ify the region of interest for further exploration, even
tively apolar solvent).
though only 8% of the possible number of reactions had
It is known that the addition of ammonia and of water been executed. If a single large DoE design had been cho-
can change the distribution ratio of the products of nitrile sen, a large number of unnecessary reactions would have
hydrogenation.[9] Ammonia is a commonly used additive, been performed.
favouring the formation of primary amines, though in the
When using such a small design it is important to realise
present case reaction with the ester group is a possible side that each result is extremely influential for the calculation
reaction. Reports on the effect of water are conflicting; se- of the main effects of the parameters. These calculations
veral cases have been reported in which water is added to will become increasingly inaccurate with a growing number
promote the formation of both primary and secondary of “zero-yield” reactions or failed experiments. In this case
amines[10] but it has also been claimed that water does not the number of reactions affording the desired product is so
change the product distribution but instead increases the low (i.e., 2 per design) that a statistical evaluation of the
reaction rate.[11] The effects of both these additives were effect of the parameters on the yield would not be meaning-
studied in the initial screening.
ful. The results do, however, enable the identification of un-
The parameter space for the initial screen is summarised favourable factors and their exclusion from the next screen-
in Table 1. The reaction temperature was varied over two ing phase.
levels. The small reactors of the high-throughput unit did
The reactions in Table 2 were all run for 24 hours at
not permit independent variation in pressure, which was 120 °C, with either dioxane or 2-propanol as the solvent
kept constant at 20 bar H2. A selection of 24 reactions out and ammonia or water as the additive. The successful met-
of the total of 320 possible combinations was made for each als were Ni and Rh, supported on either carbon or alumina.
Table 1. The parameter space to be investigated for substrates 1a and 1c.
Parameter
Level 1
Level 2
Level 3
Level 4
Level 5
Temperature (°C)
Reaction time (h)
Support
Solvent
Additive
90
3
120
24
carbon
toluene
H2O
Pd
–
–
–
–
–
–
–
–
–
–
–
Ru
alumina
2-propanol
no additive
Ni
dioxane
NH3
Pt
methanol
H2O + NH3
Rh
Metal
Table 2. Conditions for the successful hydrogenation in the initial screening.
Substrate
Metal
Support
Temp. [°C]
Solvent
Additive
Conversion[a] of 1 [%]
Yield[a] 2 [%]
1a
1a
1c
1c
Ni
Rh
Ni
C
120
120
120
120
dioxane
dioxane
dioxane
2-propanol
NH3
NH3 + H2O
H2O
100
100
100
100
33
24
65
48
Al2O3
Al2O3
C
Rh
H2O
[a] According to GC.
1666
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 1664–1671

A Novel Approach to N-Acylated β-Amino Alcohols
FULL PAPER
From the reactions that did not yield the desired product, Table 3. Conditions and parameters in the second screening round.
the following trends could be observed: the reactions using
Parameter
Level 1
Level 2
Level 3
Ru or toluene gave low degrees of conversion, while the
reactions performed in MeOH in all cases gave complete
conversion, but with a wide range of side products. Since
Additive
Solvent
Metal
no additive
2-propanol
Ni
H2O
–
–
dioxane
Rh
Ru
silica
the intention of the first screening was to reduce the param- Support
alumina
carbon[a]
eter space, none of the side products of the reactions was
isolated. However, GC-MS enabled the identification of se-
veral side products (3 to 6; Figure 2).
[a] In the case of nickel, Raney nickel was used instead of nickel
on carbon.
The results of this screening are presented in Figure 3
and Figure 4. Conversion of 1a and 1c was 100% in all
cases, except for those reactions in which Ru–alumina and
Ru–silica were used. For those two catalysts no conversion
was observed. In sharp contrast to the first screening, in
which only a few reactions had yielded the N-acyl β-amino
alcohols, all active catalysts now yielded the desired prod-
ucts. A statistical evaluation of the results shown in Figure 3
and Figure 4, with respect to the significance of the main
effects of the parameters and their interactions, is presented
in Figure 5 and Figure 6.
Figure 2. Identified side products in the hydrogenation of 1a.
The presence of 5 shows that the secondary amine had
been formed in some cases. Products 3 to 5, in which the
benzylic alcohol group has been removed, were particularly
dominant when platinum was used as the catalyst, which is
to be expected from the application of platinum catalysts
for the cleavage of this type of bond. Equivalent byproducts
could also be identified in the case of the aliphatic substrate,
though in much smaller amounts, which is in accordance
with the greater stability of the C–O bond. The fact that
even the aliphatic C–O bond can be cleaved can be attrib-
uted to the stabilising effect of the nitrile group on the inter-
mediate radical formed during the cleavage.
Figure 3. Graphical representation of the results of the second
screening with substrate 1c.
In the successful reactions of 1a (see Table 2) ammonia
was present as additive, but analysis of the results did not
show unambiguously that the presence of ammonia was es-
sential. Since the formation of 6 indicates that ammonia
also reacts with the substrate it was decided, in order to
avoid this side reaction, to optimise the conditions further
in the absence of ammonia.
Samples taken after 3 hours showed only low degrees of
conversion and there was no formation of 2a or 2c in any
reaction other than those performed under the conditions
reported in Table 2. The long reaction time could be due to
an initial activation period for the catalyst but the study of
this was deferred to a later stage and a reaction time of 24 h
was maintained for the second design.
From the results of the first design, the second was con-
ducted with the parameters indicated in Table 3. In order
to study the effect of the carrier further, silica was included
in this design. Since the parameter space was now consider-
Figure 4. Graphical representation of the results of the second
screening with substrate 1a.
In the evaluation of the main parameters in Figure 5,
ably reduced, a full factorial design (i.e., 36 combinations) many similarities for the two different substrates can be
was now feasible for each substrate. All the reactions were noted. The most important parameter in both cases is the
performed at 120 °C and 20 bar H2, with a reaction time of type of metal, with nickel being the best, followed by rho-
24 hours.
dium. For ruthenium, the poor results from the initial
Eur. J. Org. Chem. 2006, 1664–1671
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1667

L. Veum, S. R. M. Pereira, J. C. van der Waal, U. Hanefeld
FULL PAPER
Figure 5. Effect of the main parameters in the second screening:
a) aliphatic substrate 1c, b) benzylic substrate 1a. The lengths of
the bars show the relative influence of the main parameters on the
yields of 2a and 2c. The bars directed to the right have a positive
relative effect and those to the left a negative one. The dotted lines
represent the 95% confidence interval calculated from the esti-
mated experimental variance. Effects higher than this confidence
interval are considered significant and are represented in black
(“Al” = alumina, “Si” = silica).
screening are confirmed. In the initial screening, the suc-
cessful reactions included those in which water had been
used as an additive. A small but statistically significant pos-
itive effect of water is indeed observed in the case of the
aliphatic substrate, though not with the benzylic one. Al-
though the difference between the two solvents is small, di-
oxane is statistically significantly better than 2-propanol for
both substrates. With respect to the effects of the catalyst
carriers, the apparent superiority of carbon is a function
solely of the fact that ruthenium gives the product only in
combination with carbon while, in addition, nickel on car-
bon was not available and Raney-Ni was used instead, so
no conclusions on carrier effects can be drawn from this
second design.
With respect to the interactions between the parameters
shown in Figure 6 it was noted that, for both substrates,
there are significant additive/solvent and metal/support in-
teractions, while in the case of the benzylic substrate a sol-
vent/metal interaction also exists. However, the effects of
these interactions are relatively small in comparison with
the main effect of the metal itself. For the aliphatic sub-
strates, nickel on silica, in dioxane as the solvent and with
water as additive, is the combination of choice, while for
the benzylic substrate Raney nickel is the indicated catalyst,
in dioxane without addition of water.
Figure 6. Interaction effects between the parameters of the second
screening: a) aliphatic substrate 1c, b) benzylic substrate 1a. The
lengths of the bars show the relative influence of the interaction
effects, between the different parameters, on the yields of 2a and
2c. The bars directed to the right have a positive relative effect
and to the left a negative one. The dotted lines represent the 95%
confidence interval calculated from the estimated experimental
variance. Effects higher than this confidence interval are considered
significant and are represented in black (“Al” = alumina, “Si” =
silica).
nient to use this catalyst for further optimisation. In the
case of the aliphatic substrate the reaction time was reduced
to two hours, and for the benzylic substrate to three hours.
Once again, dioxane proved to be slightly superior to 2-
propanol As a result of this, it was decided to perform the
third round of screening, for the optimisation of tempera-
ture and pressure, with activated nickel on alumina in diox-
ane. In the case of the aliphatic substrate 1c, water was used
as an additive.
All the reactions from these two screenings were per-
formed on the “Quick Catalyst Screening 96” platform.
This equipment has a maximum pressure limit of 20 bar
and no individual temperature control for the reactors. Fur-
ther optimisation of pressure and temperature was for that
reason performed in a conventional autoclave.
In preparation for this, a test was conducted to determine
whether the results obtained with the nickel catalysts could
be improved by activation with H2 prior to the catalytic
test. After activation of the catalysts at 140 °C for 12 hours
Third Design
A temperature range of 80 to 160 °C and a pressure
at 40 bar H2, nickel on alumina gave similar yields to those range from 5 to 40 bar were tested. Only minor differences
of Raney nickel and nickel on silica and it was found conve- in the yield ( 7% for the aliphatic, 5% for the aromatic)
1668
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 1664–1671

A Novel Approach to N-Acylated β-Amino Alcohols
FULL PAPER
Table 4. Degrees of conversion and yields from the hydrogenation of acylated cyanohydrins 1a–f.
Substrate
Conversion of 1 [%]
NMR yield of 2 [%]
Isolated yields of 2 [%]
1a
1b
1c
1d
1e
1f
100
100
100
100
100
100
n.d.
n.d.
74
91
Ϸ 75
83
49[a]
50[a]
57[b]
72[a]
58[a]
30[b]
[a] Isolated by column chromatography. [b] Isolated by recrystallisation from ethyl acetate, yield not optimised.
were observed, except for reaction temperatures below nia released in the formation of the secondary amine side
90 °C, at which hardly any reaction occurred. Despite the product, or the secondary amine itself.
small differences in the observed yields an optimum of
10 bar H2 at 140 °C was found for the aliphatic substrate
1c, whilst 20 bar H2 at 120 °C was found for the benzylic
substrate 1a.
Conclusion
The catalytic hydrogenation of acylated cyanohydrins (1)
with subsequent intramolecular acyl group migration con-
stitutes a valuable one-pot route to the pharmaceutically
important N-acyl β-amino alcohols (2). Nickel on alumina
as catalyst in dioxane as solvent proved to be preferable to
the traditional catalysts (Pd/C and PtO2) used under acidic
conditions;[2e] both the hydrogenation and the migration
proceeded smoothly and the desired products could be ob-
tained in yields of up to 90% for the aliphatic substrates
and up to 50% for the more sensitive benzylic substrates,
as determined by GC. Application to a range of aliphatic
and aromatic substrates with different acyl groups was dem-
onstrated. Aliphatic substrates not only gave the highest
yields but in the case of an enantiopure aliphatic cyano-
hydrin acetate the stereocentre remained unaltered during
the reaction. On the other hand, a small amount of racemi-
sation could be observed for a sensitive enantiopure benzylic
substrate. Given the straightforward access to the (chiral)
starting materials and the mild, catalytic reaction condi-
tions this one-pot sequence represents a valuable addition
to the arsenal of organic chemists.
A multi-step DoE approach proved an efficient method
for the optimization of the reaction. Out of more than 2000
possible combinations of the parameters needing to be
studied, it proved possible to effect the optimization
through the use of only 70 experiments for each substrate.
This shows the great advantage of the DoE approach for
the optimisation of a new reaction, enabling a large param-
eter space to be investigated and the most interesting range
within the parameter space to be identified.
Other Substrates
In order to establish the versatility of the reaction, these
optimised conditions were applied to a number of other
acylated cyanohydrins: a substituted benzylic ester (1b), an
aliphatic substrate with an aromatic side chain (1d) and ali-
phatic substrates with a variety of acyl groups (1e, 1f)
(Scheme 2, Table 4). All these substrates were successfully
hydrogenated to yield the desired N-acyl β-amino alcohols
2a–f. Substrate conversion was 100% in all cases. The ben-
zylic substrates 1a and 1b gave more side products than the
aliphatic 1c–f. This difference between the substrates is in
accordance with the unstable benzylic C–O bond. In the
cases of 1c and 1f, the products were isolated from the reac-
tion mixtures by crystallisation.
The hydrogenation was also performed on the optically
active substrates (S)-1a (95% ee) and (S)-1c (94% ee). As
expected, the chiral centre of (S)-1c was found to remain
unchanged during both the hydrogenation and the intra-
molecular migration. This was not the case with (S)-1a,
however, and the isolated (S)-2a had an ee of only 75%
(see Scheme 3). This decrease in ee might be explained by a
combination of elevated temperatures and a base-catalysed
racemisation of the substrate, the base being either ammo-
Experimental Section
General: 1H and 13C NMR spectra were recorded on Varian VXR-
400S (400 and 100 MHz, respectively) or Varian Unity Inova 300
(300 MHz and 75 MHz, respectively) instruments. Chemical shifts
are expressed in parts per million (δ) relative to tetramethylsilane.
Abbreviations are as follows: s (singlet), d (doublet), t (triplet), q
(quadruplet) and m (multiplet).
Mass spectra were determined on a VG 70 SE spectrometer op-
erating at 70 eV. GC-MS was measured with a VG 250 SE instru-
Scheme 3. Catalytic hydrogenation of enantiopure acylated cya-
nohydrins.
Eur. J. Org. Chem. 2006, 1664–1671
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1669

L. Veum, S. R. M. Pereira, J. C. van der Waal, U. Hanefeld
FULL PAPER
ment fitted with a CP Sil 8 CB column of 25 m×0.25 mm and
0.4 µm DF. A Varian Star 3600 GC fitted with a CP Sil 5CB col-
umn with 50 m×0.55 mm and 1 µm DF was used to determine the
levels of conversion in the crude reaction mixtures. Optical rota-
tions were obtained with a Perkin–Elmer 241 polarimeter. Melting
points are uncorrected. Column chromatography was carried out
with silica gel packing of 0.060–0.200 mm, pore diameter ca. 6 nm,
with mixtures of petroleum ether (PE), methanol (MeOH) and
ethyl acetate (EtOAc) as solvents. TLC was performed on 0.20 mm
silica gel. The nickel catalysts were all activated at 140 °C for
12 hours at 40 bar H2 before use in General Procedures B, C and
D described below. All other catalysts, and the solvents employed,
were used as received from commercial sources. For all the sup-
ported catalysts the metal loading was 5%, except for Rh-silica
(1%), Ni-alumina (50%) and Ni-silica (66%). Racemic[3] and
enantiopure cyanohydrin acetates[4b,13] were synthesised by litera-
ture procedures. The optical purity of 2a was determined by HPLC
with use of a Waters 510 pump, a 4.6×250 mm 10 µ Chiracel OJ
column and a Waters 486 UV detector. The eluent was a mixture
of hexane and 2-propanol (90:10) with a flow of 0.8 mL min–1. The
optical purities of 1a, 1c and 2c were determined by chiral GC
with a Shimadzu Gas Chromatograph GC-17A fitted with a β-
cyclodextrin column (CP-Chirasil-Dex CB 25 m×0.25 mm). A Shi-
madzu Auto-injector AOC-20i and FID detector were employed,
and He with a linear gas velocity of 75 cm s–1 was used as the
carrier gas. The Avantium “Quick Catalyst Screening 96” platform
was used to perform the reactions in the first and second experi-
mental designs. This equipment has a maximum pressure limit of
20 bar and the temperature is controlled for all reactors simulta-
neously. Otherwise, a 100 mL Parr autoclave was used. The elemen-
tal analysis was performed on a Elementar Vario EL III analyser.
Commercially available NemrodW 2000 software from LPRAI
(France) was used for the statistical calculations of the experimen-
tal design.
umn chromatography (silica, EtOAc/MeOH, 95:5, Rf = 0.27). Yield
of (S)-2a: 503 mg (49%) as a white solid; m.p. 125–126 °C; 1H
NMR (300 MHz, CDCl3, 25 °C, TMS): δ = 2.01 (s, 3 H, CH3–
C=O), 3.32 (ddd, J = 5.0, 7.9, 14.1 Hz, 1 H, CH2–N), 3.70 (ddd, J
= 3.3, 7.0, 14.1, 1 H, CH2–N), 4.85 (dd, J = 3.3, 7.9, 14.1 Hz, 1 H,
CH–O), 5.92 (s, 1 H, NH), 7.28–7.38 (m, 5 H, aromatic) ppm. 13C
NMR (75 MHz, CDCl3, 25 °C, TMS): δ = 23.1 (CH3), 47.6 (CH2–
N), 73.6 (CH–O), 125.8, 128.8, 128.5 and 141.8 (aromatic), 171.6
(C=O) ppm. IR (KBr): ν = 3300, 3080, 1648, 1547, 1295 cm–1. MS
˜
(70 eV, EI): m/z (%) = 179 (1) [M]+, 161 (3) [M – H2O]+, 120 (14),
107 (21), 79 (31), 77 (31), 73 (100). Elemental analysis calculated
(%) for C10H13NO2 (179.22): C 67.02, H 7.31, N 7.82; found C
67.00, H 7.49, N 7.81.
(S)-N-(2-Hydroxy-2-phenylethyl)acetamide ((S)-2a): The solid pre-
pared from (S)-1a (95% ee) as described in General Procedure C
was purified by column chromatography (silica, EtOAc/MeOH,
95:5, Rf = 0.27). Yield of (S)-2a: 0.454 mg (45.4%) as a white solid;
ee = 75%, [α]2D0 = +8.1 (c = 1.0 in MeOH); other spectroscopic
data as for 2a.
N-[2-Hydroxy-2-(3-methoxyphenyl)ethyl]acetamide (2b): The solid
prepared from rac–1b as described in General Procedure C was
purified by column chromatography (silica, EtOAc/MeOH, 95:5,
Rf = 0.25). Yield of 2b: 0.570 mg (57%) as a white solid; m.p. 123–
1
124 °C; H NMR (300 MHz, CD3OD, 25 °C, TMS): δ = 1.93 (s, 3
H, CH3–C=O), 3.28 (dd, J = 7.9, 13.7 Hz, 1 H, CH2–N), 3.45 (dd,
J = 4.6, 13.5 Hz, 1 H, CH2–N), 3.78 (s, 3 H, OCH3), 4.71 (dd, J =
4.6, 7.9 Hz, 1 H, CH–O), 6.81 (ddd, J = 0.9, 2.6, 8.2 Hz, 1 H, C4–
H), 6.95 (m, 2 H, C2–H, C6–H), 7.24 (apparent t, J = 7.9 Hz, 1 H,
C5–H) ppm. 13C NMR (75 MHz, CD3OD, 25 °C, TMS): δ = 22.5
(CH3–CO), 48.3 (CH2–N), 55.6 (OCH3), 73.5 (CH–O), 112.6 (C2),
114.1 (C4), 119.4 (C6), 130.3 (C5), 145.5 (C1), 161.2 (C3), 173.6
(C=O) ppm. IR (KBr):
ν = 3290, 1634, 1596, 1552, 1259,
˜
1066 cm–1. MS (70 eV, EI): m/z (%) = 209 (7) [M]+, 191 (3) [M –
H2O]+, 150 (31), 109 (25), 73 (87), 62 (46), 45 (100). Elemental
analysis calcd. (%) for C11H15NO3 (209.24): C 63.14, H 7.23, N
6.69; found C 61.41, H 7.57, N 6.50.
General Procedure A – Screening on the Avantium “Quick Catalyst
Screening 96” Platform: The various supported metal catalysts
(5 mg) were weighed into the autoclaves and added to a 1.7  solu-
tion of the substrate in the desired solvent (1.5 mL). When water
was used as an additive, 10 µL was added. In the cases in which
ammonia was used as additive, the concentration of ammonia in
the reaction mixture was 0.5 . After the reaction mixture had been
stirred at 90/120 °C and 20 bar H2 for 3 or 24 hours, it was centri-
fuged and the supernatant liquid was analysed by GC and GC-
MS.
General Procedure D. Reductions in the Parr Autoclave under Op-
timised Conditions for Substrates Prepared from Aliphatic Alde-
hydes: Activated Ni on alumina (50%, 100 mg) was added to a
solution of the substrate (5.7 mmol) in dioxane (30 mL) and water
(0.2 mL). After the reaction mixture had been stirred at 140 °C and
10 bar H2 for 2 h it was filtered, a sample (2 mL) was taken from
the filtrate, the solvents from this sample were removed under vac-
uum, and the sample was then analysed by 1H NMR spectroscopy.
The combined filtrate and NMR sample were then evaporated to
dryness to yield the oil or solid products.
General Procedure B – Screening for Temperature and Pressure in
the Parr Autoclave: Preactivated Ni on alumina (50%, 100 mg) was
added to a solution of 1a or 1c (5.7 mmol) in dioxane (30 mL). In
the case of 1c, water (0.2 mL) was also added. After the reaction
mixture had been stirred at 80, 100, 120, 140, or 160 °C and 5, 10,
20, 30 or 40 bar H2 for 2 hours it was filtered and the filtrate was
analysed by GC.
N-(2-Hydroxyheptyl)acetamide (2c): The oil prepared from 1c as
described in General Procedure D was purified by recrystallisation
from EtOAc. Yield of 2c: 454 mg (56%) as a white solid; m.p. 75–
1
General Procedure C – Reductions in the Parr Autoclave under Opti-
mized Conditions for Substrates Prepared from Aromatic Aldehydes:
Activated Ni on alumina (50%, 100 mg) was added to a solution
of the substrate (5.7 mmol) in dioxane (30 mL). After the reaction
mixture had been stirred at 120 °C and 20 bar H2 for 3 h it was
filtered, a sample (2 mL) was taken from the filtrate, and the sol-
vents from this sample were removed under vacuum. The sample
was then analysed by 1H NMR spectroscopy. The combined filtrate
and NMR sample were then evaporated to dryness to yield the oil
or solid products.
76 °C; H NMR (300 MHz, CDCl3, 25 °C, TMS): δ = 0.89 (m, 3
H, CH3–CH2), 1.29–1.47 (m, 8 H, CH3–CH2–CH2–CH2–CH2),
2.00 (s, 3 H, CH3–C=O), 3.08 (ddd, J = 5.0, 7.9, 13.7 Hz, 1 H,
CH2–N), 3.45 (ddd, J = 2.9, 6.6, 13.9 Hz, 1 H, CH2–N), 3.45 (s, 1
H, OH), 3.69 (m, 1 H, CH–O), 6.49 (s, 1 H, NH) ppm. 13C NMR
(75 MHz, CDCl3, 25 °C, TMS): δ = 14.0 (CH3–CH2), 22.6 (CH3–
CH2), 23.2 (CH3–CO), 25.2 (CH3–CH2–CH2), 31.8 (CH2–CH2–
CH), 35.0 (CH2–CH), 45.9 (CH2–N), 71.2 (CH–O), 171.4
(C=O) ppm. IR (KBr): ν = 3425, 3279, 1661, 1627, 1586, 1569,
˜
1136 cm–1. MS (70 eV, EI): m/z (%) = 174 (3) [M + 1]+, 102 (10),
73 (100). Elemental analysis calcd. (%) for C9H19NO2 (173.25): C
62.39, H 11.05, N 8.08; found C 62.01, H 11.67, N 8.04.
N-(2-Hydroxy-2-phenylethyl)acetamide (2a): The solid prepared
from 1a as described in General Procedure C was purified by col-
1670
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 1664–1671

A Novel Approach to N-Acylated β-Amino Alcohols
FULL PAPER
(S)-N-(2-Hydroxyheptyl)acetamide ((S)-2c): The oil prepared from
(S)-1c (5.4 mmol, 94% ee) as described in General Procedure D
was purified by recrystallisation from EtOAc. Yield of (S)-2c:
Acknowledgments
The authors thank Dr. J. A. Peters, Dr. D. T. Schühle and Dr. R. S.
Downing for fruitful discussions and suggestions. Avantium Tech-
nologies are kindly acknowledged for the use of their “Quick Cata-
lyst Screening 96” platform at their laboratories. U. H. thanks the
Royal Netherlands Academy of Arts and Sciences (KNAW) for a
fellowship. L. V. and S. P. gratefully acknowledge NWO/ACTS for
financial support.
533 mg (57%) as a white solid; ee = 95%; m.p. 75–76 °C; [α]2D0
+14.1 (c = 1.0 in MeOH); other spectroscopic data as for 2c.
=
N-(2-Hydroxy-3-phenoxypropyl)acetamide (2d): The oil prepared
from rac-1d as described in General Procedure D was purified by
column chromatography (silica, EtOAc/MeOH, 95:5, Rf = 0.29).
Yield of 2d: 967 mg (72%) as a white solid; m.p. 49–50 °C; 1H
NMR (300 MHz, CDCl3, 25 °C, TMS): δ = 1.99 (s, 3 H, CH3–
C=O), 3.36 (ddd, J = 5.5, 6.8, 6.8 Hz, 1 H, CH2–N), 3.59 (ddd, J
= 3.3, 6.1, 14.0 Hz, 1 H, CH2–N), 3.92 (d, J = 5.5 Hz, 2 H, CH2–
O), 4.09 (m, 1 H, CH–O), 4.18 (s, 1 H, OH), 6.58 (s, 1 H, NH),
6.87 (m, 2 H, aromatic), 6.95 (m, 1 H, aromatic), 7.26 (m, 2 H,
aromatic) ppm. 13C NMR (75 MHz, CDCl3, 25 °C, TMS): δ = 23.0
(CH3–C=O), 43.0 (CH2–N), 69.5 (CH–OH and CH2–O), 114.5,
[1] a) M. North, Tetrahedron: Asymmetry 2003, 14, 147–176; b)
H. Griengl, H. Schwab, M. Fechter, Trends Biotechnol. 2000,
18, 252–256; c) J. Brussee, A. van der Gen, in: Stereoselective
Biocatalysis (Ed.: P. N. Ramesh), Marcel Dekker, Inc., New
York, 2000, pp. 289–320.
[2] a) L. T. Kanerva, Acta Chem. Scand. 1996, 50, 234–242; b) K.
Tanaka, A. Mori, S. Inoue, J. Org. Chem. 1990, 55, 181–185;
c) T. Ziegler, B. Hörsch, F. Effenberger, Synthesis 1990, 575–
578; d) T. Ooi, M. Kameda, J. Fujii, K. Maruoka, Org. Lett.
2004, 6, 2397–2399; e) W. J. Greenlee, J. P. Springer, A. A. Pat-
chett, J. Med. Chem. 1989, 32, 165–170.
121.2, 129.6, and 158.4 (aromatic), 171.9 (C=O) ppm. IR (KBr): ν
˜
= 3384, 3299, 1630, 1601, 1571, 1284, 1118, 751 cm–1. MS (70 eV,
EI): m/z (%) = 209 (3) [M]+, 191 (32) [M – H2O]+, 148 (7), 116
(100). Elemental analysis calcd. (%) for C11H15NO3 (209.24): C
63.14, H 7.23, N 6.69; found C 61.99, H 7.22, N 6.46.
[3] A. Fishman, M. Zviely, Tetrahedron: Asymmetry 1998, 9, 107–
118.
[4] a) M. Inagaki, J. Hiratake, T. Nishioka, J. Oda, J. Org. Chem.
1992, 57, 5643–5649; b) L. Veum, L. T. Kanerva, P. J. Halling,
T. Maschmeyer, U. Hanefeld, Adv. Synth. Catal. 2005, 347,
N-(2-Hydroxy-3-methylbutyl)benzamide (2e): The solid prepared
from rac-1e as described in General Procedure D was purified by
column chromatography (silica, EtOAc/PE, 45:55, Rf = 0.30). Yield
1015–1021; c) C. Paizs, P. Tähtinen, M. Tosa, C. Majdik, F. D.
¸
1
of 2e: 681 mg (58%) as a white solid; m.p. 116–117 °C; H NMR
Irimie, L. T. Kanerva, Tetrahedron 2004, 60, 10533–10540; d)
S. Lundgren, E. Wingstrand, M. Penhoat, C. Moberg, J. Am.
Chem. Soc. 2005, 127, 11592 –11593; e) Y. N. Belokon, P.
Carta, A. V. Gutnov, V. Maleev, M. A. Moskalenko, L. V.
Yashkina, N. S. Ikonnikov, N. V. Voskoboev, V. N. Khrustalev,
M. North, Helv. Chim. Acta 2002, 85, 3301–3312; f) Y. N. Belo-
kon, P. Carta, M. North, Lett. Org. Chem. 2004, 1, 81–83.
[5] M. Ikezaki, N. Umino, M. Gaino, K. Aoe, T. Iwakuma, T. Oh-
Ishi, Yakugaku Zasshi 1986, 106, 80–89.
[6] a) W. H. Hartung, J. Am. Chem. Soc. 1928, 50, 3370–3374; b)
J. S. Buck, J. Am. Chem. Soc. 1933, 55, 2593–2597.
[7] a) D. L. Massart, B. G. M. Vandeginste, L. M. C. Buydens, S.
de Jong, P. J. Lewi, J. Smeyers-Verbeke, Handbook of Chemo-
metrics and Qualimetrics, Part A, Elsevier, Amsterdam, 1997,
pp. 643–658, 701–737; b) E. W. Kirchhoff, D. R. Anderson, S.
Zhang, C. S. Cassidy, M. T. Flavin, Org. Process Res. Dev.
2001, 5, 50–53; c) R. Marchetti, M. E. Guerzoni, Cerevisia Bio-
technol. 1991, 16, 24–33.
[8] P. D. Haaland, Biotechnology Experimental Design in: Statisti-
cal Design and Analysis of Industrial Experiments (Ed.: S.
Ghosh), Marcel Dekker, New York, 1990, pp. 73–108.
[9] S. Gomez, J. A. Peters, T. Maschmeyer, Adv. Synth. Catal.
2002, 344, 1037–1057.
[10] a) J. Volf, J. Pasek, in Catalytic Hydrogenation, vol. 27 (Ed.: L.
Cerveny), Elsevier, Amsterdam 1986, pp. 105–144; b) H.
Greenfield, Ind. Eng. Chem. Prod. Res. Dev. 1976, 15, 156–158.
[11] Y. Huang, V. Adeeva, W. M. H. Sachtler, Appl. Catal., A 2000,
196, 73–85.
(300 MHz, CDCl3, 25 °C, TMS): δ = 0.97 (dd, J = 6.8, 9.0 Hz, 6
H, 2×CH3), 1.73 (m, 1 H, CH–(CH3)2), 3.06 (s, 1 H, OH), 3.30
(ddd, J = 4.6, 8.61, 13.7 Hz, 1 H, CH2–N), 3.50 (m, 1 H, CH–O),
3.72 (ddd, J = 2.8, 6.8, 13.7 Hz, 1 H, CH2–N), 6.86 (s, 1 H, NH),
7.38 (m, 2 H, aromatic), 7.46 (m, 1 H, aromatic), 7.77 (m, 2 H,
aromatic) ppm. 13C NMR (75 MHz, CDCl3, 25 °C, TMS): δ = 17.9
(CH3), 18.6 (CH3), 32.3 (CH–(CH3)2), 44.1 (CH2–N), 76.3 (CH–
O), 127.0, 128.5, 131.5, and 134.3 (aromatic), 168.5 (C=O) ppm.
IR (KBr): ν = 3398, 3319, 1633, 1578, 1541, 1057, 697 cm–1. MS
˜
(70 eV, EI): m/z (%) = 207 (1) [M]+, 189 (3) [M – H2O]+, 164 (16),
134 (89), 122 (29), 105 (100). Elemental analysis calcd. (%) for
C12H17NO2 (207.27): C 69.54, H 8.27, N 6.76; found C 68.75, H
8.61, N 6.68.
N-(2-Hydroxyhepyl)butanamide (2f): The oil prepared from rac-1f
as described in General Procedure D was purified by recrystalli-
sation from EtOAc. Yield of 2f: 345 mg (30%) as a white solid;
m.p. 62–63 °C; 1H NMR (300 MHz, CDCl3, 25 °C, TMS): δ = 0.89
(m, 3 H, pentyl-CH3), 0.95 (t, J = 7.5 Hz, 3 H, propyl-CH3), 1.29–
1.44 (m, 8 H, CH3–CH2–CH2–CH2–CH2), 1.67 (sext, J = 7.4 Hz,
2 H, CH2–CH2–C=O), 2.18 (t, J = 7.4 Hz CH2–CH2–C=O), 2.94
(s, 1 H, OH), 3.11 (ddd, J = 4.9, 7.7, 13.0 Hz, 1 H, CH2–N), 3.47
(ddd, J = 2.7, 6.2, 13.7 Hz, 1 H, CH2–N), 3.70 (m, 1 H, CH–O),
6.15 (s, 1 H, NH) ppm. 13C NMR (75 MHz, CDCl3, 25 °C, TMS):
δ = 13.8 (propyl-CH3), 14.0 (pentyl-CH3), 19.2 (CH2–CH2–C=O),
22.6 (CH3–CH2–CH2–CH2), 25.2 (CH3–CH2–CH2–CH2), 31.8
(CH2–CH2–CH), 35.0 (CH2–CH), 38.6 (CH2–C=O), 45.7 (CH2–
[12] P. F. Aguiar, B. Bourguignon, M. S. Khots, D. L. Massart, R.
Phan-Than-Luu, Chemometrics and Intelligent Laboratory Sys-
tems 1995, 30, 199–210.
[13] H. Griengl, N. Klempier, P. Pöchlauer, M. Schmidt, N. Shi,
A. A. Zabelinskaja-Mackova, Tetrahedron 1998, 54, 14477–
14486.
N), 71.5 (CH–O), 174.2 (C=O) ppm. IR (KBr): ν = 3418, 3283,
˜
2964, 2919, 1657, 1624, 1566 cm–1. MS (70 eV, EI): m/z (%) = 130
(17), 101 (100), the [M]+ peak could not be identified. Elemental
analysis calcd. (%) for C11H23NO2 (173.25): C 65.63, H 11.52, N
6.96; found C 64.78, H 12.03, N 6.83.
Received: November 4, 2005
Published Online: February 3, 2006
Eur. J. Org. Chem. 2006, 1664–1671
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1671