A highly efficient asymmetric organocatalytic aldol reaction in a ball mill

Article abstract of DOI:10.1002/chem.200700188

Anti-aldol products with up to >99% enantiomeric excess (ee) have been obtained by proline catalysis in excellent yields under experimentally simple solvent-free conditions. Efficient mixing of all the components is accomplished by applying a mechanochemi

Full text of DOI:10.1002/chem.200700188

FULL PAPER
DOI: 10.1002/chem.200700188
4710
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722

FULL PAPER
A Highly Efficient Asymmetric Organocatalytic Aldol Reaction in a Ball Mill
BelØn Rodríguez, Angelika Bruckmann, and Carsten Bolm*[a]
Abstract: Anti-aldol products with up
to >99% enantiomeric excess (ee)
have been obtainedby proline catalysis
in excellent yields under experimental-
ly simple solvent-free conditions. Effi-
cient mixing of all the components is
accomplishedby applying a mechano-
chemical technique (ball milling). The
catalysis is air andmoisture tolerant
andcan be performedwith non-puri-
fiedstarting materials. Even mixtures
of solely solidcompounds react, giving
(mostly solid) products through a par-
tially homogeneous (honey-like) inter-
mediate melt. Since the reactant ratio
is almost 1:1 (avoiding the common
excess of ketone), the product isolation
is easy leading to high aldol product
yields.
Keywords: aldol reaction
·
ball
milling · organocatalysis · proline ·
solvent-free
Introduction
most of these studies have been performed in organic sol-
vents andcrucial factors such as appropriate solvation and
reaction temperature are often not taken into account, this
discussion appears somehow superficial.
The direct catalytic aldol reaction is a well-studied and
broadly applicable C C bond-forming reaction, which pro-
ꢀ
vides enantiomerically enriched b-hydroxy carbonyl com-
pounds. Among other asymmetric catalytic methods, the
proline-catalyzedversion, which proceeds via in situ gen-
eratedenamine intermediates, has attractedmuch attention,
since it affords aldol products with high chemo- and stereo-
selectivity under very mild reaction conditions (Scheme 1).[1]
Performing organic reactions in aqueous media has re-
[4]
cently attractedconsiderable attention,
since water is non-
toxic, environmentally friendly, nonflammable and inexpen-
sive. Independent studies reported by Hayashi[5] and
Barbas[6] revealedthat high diastereoselectivities andenan-
tioselectivities can be obtained, when organocatalyzed aldol
[7–9]
reactions are performedin aqueous solution.
Generally,
apolar organic molecules are only sparingly soluble in water,
andpresumably hydrophobic effects, which increase the ef-
fective concentration of the reagents by hydrophobic hydra-
tion,[10,11] leadto the observeddistinctive rate accelerations
and/or selectivity improvements.[8a,12,13] In this scenario, the
reaction wouldoccur either in the organic phase of a bipha-
sic system[5a] or in an emulsion.[5b,6] Noteworthy, in the
above-mentionedreports, unmoidfiedproline, which is
highly soluble in water, does not catalyze under those condi-
tions, and only with hydrophobic proline derivatives signifi-
cant reactivities have been achieved.[8a] As a whole, these re-
sults indicate that the main role of water is to bring the reac-
tants closer together and not to act as a solvent. On that
Scheme 1. Proline-catalyzedaldol reaction.
Besides the natural amino acid itself, various proline de-
rivatives andmany other small molecules have been success-
[2]
fully appliedas catalysts. Noticeably, several authors have
referredto the “prebiotic” relevance of this transformation,
because amino acid-catalyzed processes may have played a
role in the evolution of homochirality.[3] However, since
basis, performing the reaction in the absence of any solvent
[14,15]
appearedpromising.
By intense mixing of all reactants,
[a] Dr. B. Rodríguez, A. Bruckmann, Prof. Dr. C. Bolm
Institute of Organic Chemistry, RWTH Aachen University
Landoltweg 1, 52056 Aachen (Germany)
Fax : (+49)241-809-2391
phase separations couldbe overcome, andwe hypothesized
that such modified conditions would allow inexpensive and
readily available proline to act as active catalyst. Hence, the
use of lengthy procedures to prepare hydrophobic proline
derivatives would be unnecessary and an environmentally
benign process requiring a minimum amount of solvent (for
E-mail: Carsten.Bolm@oc.rwth-aachen.de
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2007, 13, 4710 – 4722
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4711

C. Bolm et al.
[16]
the work-up) wouldresult.
Although a few isolatedexam-
shaped).[30] Then, using these establishedconditions, several
experiments were performed in order to determine the ef-
fects of concentration, water andimpurities on the reaction
outcome.
ples of organocatalytic aldol reactions performed in the ab-
sence of a solvent can be foundin the literature,
[17]
to the
best of our knowledge such approach has not been the focus
of an intense study. Furthermore, in those previous exam-
ples high yields and acceptable stereoselectivities could only
be achievedwith an excess of liquidketone ( ꢁ5 equiva-
lents), which appears to act as both reactant andreaction
A solvent-free proline-catalyzed aldol reaction
should avoid this disadvantage and provide good results
with stoichiometric amounts of reactants.
Effect of concentration: In a solvent-free reaction the com-
position of the medium, in which the reaction occurs, is only
determined by the ratio of reactants. Therefore, and with
the goal to findthe optimal reactant ratio, this crucial factor
was focussed on first. The study was carried out under the
conditions described above using various ratios of ketone 1a
to aldehyde 2a, ranging from 1:1 to 5:1 (Table 1).
media.
In order to achieve a high conversion in solvent-free cata-
lytic systems, an efficient mixing of all reagents andthe cata-
lyst is critical. In this respect, conventional magnetic stirring
methods may prove insufficient. The mechanochemical tech-
nique of ball milling offers an alternative. Ball milling has
widely been applied for the grinding of minerals into fine
particles andfor the preparation andmodification of inor-
ganic solids.[18,19] In synthetic organic chemistry it has sel-
domly been used, and the few reported examples include
the functionalization of fullerenes,[20] the reductive benzyla-
Table 1. Effect of the ratio of reactants on the solvent-free aldol reaction
between cyclohexanone (1a) and p-nitrobenzaldehyde (2a) catalyzedby
(S)-proline.[a]
[b]
Entry
Product
1a/2a
Yield[%]
anti/syn[c]
ee [%][d]
tion of malonitrile,[21] the protection of amines,[22a] Knoeve-
nagel[22b] andaldol
condensations, Michael additions,[22b]
[23]
1
2
3
4
5
6
3a
3a
3a
3a
3a
3a
1.0:1.0
1.1:1.0
1.5:1.0
2.0:1.0
3.0:1.0
5.0:1.0
77 (18)
90 (8)
93
94
91
82:18
82:18
80:20
77:23
72:28
62:38
95
95
94
94
93
87
the preparation of phosphorus ylides,[24] the oxidative cou-
pling of 2-naphthol,[25] andHeck-type cross-coupling reac-
tions.[26]
On that basis, a study was initiated focussing on the inves-
tigation of asymmetric organocatalytic reactions under (sol-
vent-free) ball-milling conditions.[27] Besides enantioselective
anhydride openings mediated by cinchona alkaloids,[28] pro-
line-catalyzedasymmetric aldol reactions became research
targets. For the latter, 10 mol% of (S)-proline were com-
monly appliedas catalyst, andfor comparison the reactions
were performedunedr conventional magnetic stirring or
ball-milling conditions. Here, the details of the study are dis-
closed.
55 (38)
[a] Reaction conditions: under argon atmosphere, cyclohexanone (1a), p-
nitrobenzaldehyde (2a, 2.0 mmol) and( S)-proline (0.2 mmol) were
stirredfor 24 h at room temperature using conventional magnetic stirring.
[b] Combined yield of the isolated diastereomers. The data in parenthe-
ses indicate the amounts of recovered aldehyde. [c] Determined by
1H NMR spectroscopic analysis of the crude product. [d] Determined by
HPLC analysis of the anti isomer using a chiral stationary phase.
The first gratifying observation was that inexpensive and
readily available (S)-proline was a highly efficient catalyst
of the aldol reaction under solvent-free conditions. When
stoichiometric amounts of reactants were used, the aldol
product 3a was obtainedin 77% yield(along with 18% of
recovered aldehyde), with an anti/syn ratio of 82:18 and
with 95% ee (Table 1, entry 1). Improvedconversions could
be achievedby using 1.1 equivalents of ketone, leading to
the desired product in 90% yield, without loss of selectivity
(Table 1, entry 2).[31] As anticipated, the ratio of reactants
was a crucial factor. It not only determined the rate of the
reaction, but also the selectivity of the process. It is particu-
larly noteworthy that both reactivity and selectivity were
higher in close to equimolar mixtures of reactants. The best
results were obtainedwhen smaller excesses of cyclohexa-
none (between 0 and 50 mol%) were employed, affording
products in high yields with both excellent diastereoselectiv-
ities andenantioselectivities ( ꢁ80:20 anti/syn, ꢁ94% ee;
Table 1, entries 1–3). The diastereoselectivity was lower
(ꢂ77:23 anti/syn) when more than two equivalents of cyclo-
hexanone were employed(Table 1, entries 4–6). Moreover,
the use of a large excess of the liquidketone, which was ex-
Results and Discussion
In order to investigate the effect of the solvent-free ball
milling on the proline-catalyzedaldol reaction, the reaction
conditions were first optimized. Subsequently, comparative
studies using either conventional magnetic stirring or ball
milling were carriedout for each substrate. As moedl
system the reaction between cyclohexanone (1a) and p-ni-
trobenzaldehyde (2a) in the presence of 10 mol% of (S)-
proline was chosen, andinitially, conventional magnetic stir-
ring was applied. Freshly purified starting materials were
used under inert and dry conditions in order to avoid effects
causedby water, air, andimpurities. For the sake of consis-
tency and to ensure optimal mixing, solid aldehyde 2a was
finely ground[29] andthe resulting powder stirredwith the
ketone andthe catalyst at a speedof 70–100 rpm at room
temperature for 24 h. Furthermore, all reactions were con-
ducted on a 2 mmol scale using the same round-bottom
flask (10 mL) andmagnetic stirring bar (prolate ellipsoid-
4712
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722

Asymmetric Organocatalytic Aldol Reaction
FULL PAPER
pectedto favor the formation of a homogeneous mixture
where the components can react more easily, not only de-
creasedthe selectivity of the process, but also sloweddown
the reaction. For example, when five equivalents of cyclo-
hexanone were used (Table 1, entry 6), the desired adduct
3a was obtainedin only 55% yield(along with 38% of re-
covered aldehyde) and, additionally, the selectivity of the
process was drastically diminished (62:38 anti/syn, 87% ee).
Basedon these results, subsequent experiments were carried
out using a ketone to aldehyde ratio of 1.1:1.0.
dry conditions (Table 2, entry 6 and Table 1, entry 2, respec-
tively). Thus, this pleasing fact of “condition tolerance”
allowedus to perform subsequent reactions without rigor-
ously excluding air and moisture.
Direct use of commercial reagents—effect of acid: The es-
tablished tolerance towards water indicated that neither dry
reaction conditions nor strictly dried starting materials were
required. However, the presence of other impurities in the
commonly usedcommercially available reagents couldalso
affect the reaction outcome. Therefore, a comparative ex-
periment was carriedout using all reagents as receivedfrom
commercial suppliers, without further purification (Table 3,
Effect of water: In the proline-catalyzedaldol reaction the
presence of water in an organic solvent was demonstrated to
have positive effects increasing both reaction rate andenan-
tioselectivity.[12] However, under solvent-free conditions
Table 3. Effect of p-nitrobenzoic acid( 4) as additive on the solvent-free
[32]
water causeda deceleration of the reaction progress.
In
aldol reaction between cyclohexanone (1a) and p-nitrobenzaldehyde
(2a) catalyzedby ( S)-proline.[a]
order to determine the impact of water in our reaction
system, catalyses were performedfollowing the above-
described methodology along with controlled additions of
up to 5.0 equivalents of water (Table 2). It was observed
[b]
Entry
4 [equiv]
Yield[%]
anti/syn[c]
ee [%][d]
1[e]
2
3
4
5
–
95
89:11
85:15
87:13
88:12
86:14
94
90
87
82
82
0.005
0.01
0.02
0.04
92 (5)
90 (6)
92 (4)
93 (4)
Table 2. Effect of water as additive on the solvent-free aldol reaction be-
tween cyclohexanone (1a) and p-nitrobenzaldehyde (2a) catalyzedby
(S)-proline.[a]
[a] Reaction conditions: under argon atmosphere, cyclohexanone (1a,
2.2 mmol), p-nitrobenzaldehyde (2a, 2.0 mmol), (S)-proline (0.2 mmol),
andacid 4 (if used) were stirred for 24 h at room temperature using con-
ventional magnetic stirring. [b]–[d] As in Table 1. [e] All reagents were
utilizedas receivedfrom commercial suppliers, without further purifica-
tion.
[b]
Entry
H2O [equiv]
Yield[%]
anti/syn[c]
ee [%][d]
1
2
3
4
0.15
0.5
1.0
2.0
5.0
–
78 (17)
63 (30)
17 (70)
<7
traces
90 (8)
90:10
93:7
88:12
n.d.
n.d.
82:18
97
98
90
n.d.
n.d.
95
entry 1). Unlike in all previous cases, the initial heteroge-
neous reaction mixture rapidly evolved to a honey-like
paste andafter complete conversion a 95% yieldof the
aldol product with an anti/syn ratio of 89:11 andan ee of
94% resulted.
5
6[e]
[a] Reaction conditions: under argon atmosphere, cyclohexanone (1a,
2.2 mmol), p-nitrobenzaldehyde (2a, 2.0 mmol), (S)-proline (0.2 mmol),
andH 2O (if used) were stirred for 24 h at room temperature using con-
ventional magnetic stirring. [b]–[d] As in Table 1. [e] Reaction run under
air.
Taking into account the potential for aldehydes to be oxi-
dized in air, the effect of small amounts of the correspond-
ing acid, which could be present in commercial samples of
aldehydes, was studied. In order to assure that only the
impact of the acid(i.e., without other impurities or moisture
present) was observed, the reactions were performed under
argon using purifiedstarting materials (ketone 1a andalde-
hyde 2a), and0.5–4 mol% of p-nitrobenzoic acid( 4) was
added (Table 3, entries 2–5). Once again, almost complete
conversions were achieved, and the products were obtained
in high yields (ꢁ90%). For example, when 0.5 mol% of
acid 4 was added, aldol product 3a was obtainedin 92%
yield. With respect to the stereochemistry of the product,
the acidhadalmost no effect on the anti/syn ratio (85:15),
but the enantioselectivity decreased to only 90% ee
(Table 3, entry 2). The latter was even reduced further by
adding 1.0 mol% of acid 4 (87% ee; Table 3, entry 3). Upon
use of ꢁ2 mol% of acid 4, a dramatic decrease in the enan-
tioselectivity was observed, leading to anti-aldol product 3a
with only 82% ee (Table 3, entries 4–5). Presumably, at this
stage the acidcatalysis of the aldol reaction became more
dominate, promoting the (non-asymmetric) background re-
that even substoichiometric quantities of water (0.15 or 0.5
equivalents) ledto reducedyields (78 and63%, respective-
ly) as a consequence of lower conversions (17 and30% of
recovered aldehyde, respectively; Table 2, entries 1 and 2).
The stereoselectivities, however, improvedgiving anti-aldol
products with up to 93:7 anti/syn ratio and98% ee (Table 2,
entry 2). In contrast, when stoichiometric amounts of water
were added, the reaction was decelerated. Thus, by adding 1
equivalent of water, aldol product 3a was obtainedin only
17% yield, with an anti/syn ratio of 88:12 andwith 90% ee
(Table 2, entry 3). Besides, by adding ꢁ2 equivalents of
water almost no reaction progress was detected (Table 2, en-
tries 4 and 5). Probably, under those conditions proline is
dissolved in the aqueous phase and thereby removed from
the other reactants.
Gratifyingly, no difference in conversion, yield, or selec-
tivity was observedbetween reactions run in air or under
Chem. Eur. J. 2007, 13, 4710 – 4722
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4713

C. Bolm et al.
action. As a consequence, a larger fraction of racemic prod-
uct was formed leading to a reduction of the final product
ee. Fortunately, using reagents as receivedfrom commercial
suppliers (without further purification) did not show this
effect, andwith them both high yields andexcellent enantio-
selectivities were achieved.[33] Even more so, in all reactions
with non-purifiedstarting materials a more rapidformation
of a melt favoring the stirring was observed, and perhaps
the improvedmixing of this partially homogenous inter-
mediate phase favored higher yields after shorter times. The
presence of small quantities of impurities (such as the corre-
sponding acid) could promote this effect.
nitrobenzaldehyde (2c), are shown in Table 4. In these cases
the optimum milling cycle for the formation of aldol prod-
ucts 3b and 3c consistedof a 15 min milling periodat a ro-
Table 4. Optimization of the milling cycle for the aldol reaction between
ketone 1a and different aldehydes 2 catalyzedby ( S)-proline performed
in a ball mill.[a]
Based on these optimization studies, it can be concluded
that: 1) using an excess of ketone is not advisable, since it
causes a decrease in both yields and stereoselectivities;
2) neither dry conditions, nor inert atmosphere are required;
3) there is no advantage in using purified starting materials.
From these studies it became apparent that an efficient
mixing methodwas vitally important in order to achieve
high yields in this solvent-free reaction. In this respect, con-
ventional magnetic stirring appearedinsufficient, andto
overcome this limitation, the mechanochemical technique of
ball milling was considered.[18,19,27] With the aim to evaluate
the effect of ball milling, comparative studies of the solvent-
free aldol reaction using conventional magnetic stirring and
ball milling were carriedout for each substrate. Unless oth-
erwise mentioned, the reactions were performed in an aero-
bic atmosphere, using all the reagents directly as received
from commercial suppliers and a ratio of ketone to aldehyde
of 1.1:1.
Entry Product Milling cycle
t
Yield
anti/
ee
[h] [%][b]
syn[c]
[%][d]
1
2
3
4
5
6
7
8
3b
3b
3b
3b
3c
3c
3c
3c
25 min at 700 rpm +
5 min pause
5 min at 500 rpm +
5 min pause
15 min at 400 rpm +
5 min pause
30 min at 300 rpm +
15 min pause
25 min at 700 rpm +
5 min pause
5 min at 500 rpm +
5 min pause
15 min at 400 rpm +
5 min pause
30 min at 300 rpm +
15 min pause
6.5 65
79:21 88
5.5 87
82:18 97
7
94
88:12
> 99
16 92
84:16 95
6.5 51 (16) 86:14 91
5.5 76 (14) 89:11 94
7
97
93:7
97
16 83
90:10 96
[a] Reaction conditions: cyclohexanone (1a, 2.2 mmol), aldehyde
(2.0 mmol), and( S)-proline (0.2 mmol) were stirredin the ball mill. [b]–
[d] As in Table 1.
2
Studies in the ball mill: Ball milling is a mechanochemical
technique, which can be appliedas an alternative methodof
stirring. It involves milling (or grinding) of the reaction mix-
tational speedof 400 rpm, followedby a 5 min pause. Com-
plete conversions were achievedafter 7 h (Table 4, entries 3
and7), and anti-aldol products 3b and 3c were formedin
high yields (94 and97%, respectively) andwith excellent
stereoselectivities (88:12 and93:7 anti/syn ratio; >99 and
97% ee, respectively). It was observedthat higher rotational
speeds (ꢁ500 rpm) resultedin lower ee values, together
with decreased yields mainly due to the formation of by-
products, both presumably as a result of the observed tem-
perature increase (Table 4, entries 1, 2, 5 and6). To mini-
mize evolution of heat, shorter milling times, slower speeds
and/or longer pause periods were programmed. However, in
these systems lower rotational speed(300 rpm) ledto lower
substrate conversion as a result of inefficient mixing. Thus,
long reaction times were requiredin order to achieve high
yields (Table 4, entries 4 and 8).
The generality of the ball-milling process was examined
by varying the substrate combinations. Furthermore, a full
comparative study revealed the beneficial effect of the ball
milling with respect to conventional magnetic stirring
(Table 5). Under solvent-free conditions the aldol products
were generally obtainedin high yields andwith excellent
stereoselectivities. By means of the ball mill the aldol reac-
tion proceeded much faster, indicating the advantage of
ture with balls in a rotating vessel, both made of chemically
[34,35]
inert andnon-abrasive zirconium oxide.
The number of
revolutions per minute (andtherefore rate, temperature and
degree of mixing) can be specified in a grinding cycle that
includes a pause period to avoid overheating. Thereby, a se-
lected cycle defined by the rotational speed, grinding time
andpause periodcan be repeateduntil the reaction is com-
pleted.
First, the reaction between cyclohexanone (1a) and p-ni-
trobenzaldehyde (2a), in the presence of 10 mol% of (S)-
proline as catalyst, was carriedout in the ball mill. A milling
cycle consisting of a 15 min milling periodat a rotational
speedof 400 rpm, followedby a 5 min pause, was used. This
led to the desired aldol product 3a in quantitative yield,
with an anti/syn ratio of 89:11 andwith 94% ee, after only
5.5 h. It is important to note that under typical ball-milling
conditions, the mechanical agitation process led to a consid-
erable heating of the reaction vessel (andas an extent, of
the contents), in conjunction with increasedpressure inside
the vessel (detected when it was opened).[36]
As examples of the optimization of the milling cycle, the
results of some selectedexperiments, in which cyclohexa-
none (1a) was treatedwith m-nitrobenzaldehyde (2b) or o-
4714
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722

Asymmetric Organocatalytic Aldol Reaction
FULL PAPER
Table 5. Scope of the solvent-free enantioselective aldol reaction cata-
lyzedby ( S)-proline under ball-milling conditions and under conventional
magnetic stirring.[a]
loweredboth the reactivity andstereoselectivity (Table 5,
entries 4–6). Other liquidketones such as cyclopentanone
(1b) andacetone ( 1c) were also employed(Table 5, en-
tries 7–9). In the former case, poor diastereoselectivities
were observed, whereas the enantioselectivities were high
(Table 5, 90–95% ee). In contrast, acetone afforded a prod-
uct with only 56% ee (Table 5, entry 9). Notably, the above-
mentioned trends of the electronic effects and the depend-
ence of the stereoselectivity on the substrate structure are
analogous to those observed in aldol reactions under alter-
native reaction conditions (e.g., in organic or aqueous sol-
Entry Product
Method[b]
t
Yield
anti/
ee
[h] [%][c]
syn[d]
[%][e]
A
B
5.5 99
24 95
89:11
89:11
94
94
1
[1,2,5a,6,12]
vents andsolvent-free conditions).
The excellent results obtainedin the solvent-free aldol re-
action, promptedus to further examine the advantages of
the ball-milling technology by applying the methodto a re-
action between solely solid components, which additionally
provided the opportunity to increase the range of cyclic ke-
tones testedas donors. As such, 4- tert-butylcyclohexanone
(1d) andtetrahydro-4 H-thiopyran-4-one (1e) were treated
with a variety of solid aromatic aldehydes. The results of
this study, in conjunction with a comparative study using
conventional magnetic stirring, are summarizedin Table 6.
We were pleasedto observe that the reaction also proceed-
edefficiently between solely solidreagents under solvent-
free conditions affording anti-aldol products 3j–q with both
excellent diastereoselectivities (up to 99:1 anti/syn ratio,
Table 6, entry 8) andenantioselectivities (up to 98% ee,
Table 6, entry 7). In the case of 4-tert-butylcyclohexanone
(1d), the reaction produced mainly two diastereoisomers,[38]
anti,trans-3j–m and syn,trans-3j–m, with high diastereocon-
trol (77:23 to 93:7 anti/syn ratio), obtaining the major anti,-
trans-(ꢀ)-3j–m in up to 95% ee (Table 6, entries 1–4). The
relative stereochemistry of products 3j–m was deduced
A
B
7
94
88:12
82:18
>99
98
2
3
4
5
6
16 89 (10)
A
B
7
97
93:7
91:9
97
97
36 89
A
B
20 87 (10)
72 85 (9)
74:26
78:22
75
67
A
B
36 65 (25)
96 64 (26)
66:34
71:29
63
67
A
B
11 53
22 35
81:19
81:19
45
47
A
B
5
53
57:43
52:48
95
87
7
24 90
1
from H and 13C NMR spectroscopic data: i) The character-
istic 1H NMR coupling constant between the protons
COCHCH(OH) andCOC HCH(OH) of the products
(3JH,H =7.2 Hz) revealedthe anti-aldol stereochemistry; ii)
The 2,4-trans relationship was basedon a comparison of the
13C NMR data with those of the known cis- and trans-4-tert-
butyl-2-hydroxymethylcyclohexanones.[39] Finally, the abso-
lute configuration was assignedon the assumption of a simi-
lar reaction pathway.[40]
As expected, the reactions in the solid–solid system pro-
ceeded much slower than in the liquid–solid one (Table 5),
presumably due to a more difficult mixing. In all these
solid–solid reactions the application of a mechanochemical
technique was particularly useful, andby means of ball mill-
ing high conversions were achievedin relatively short times
(ꢂ1.6 days), when compared with the use of conventional
magnetic stirring (5–7 days). For example, while under con-
ventional stirring the reaction between 4-tert-butylcyclohex-
anone (1d) and 4-nitrobenzaldehyde (2a) afforded aldol
product 3j in 58% yieldafter five days, ball milling ledto
the same product in 85% yield after only 1.4 days (Table 6,
entry 1). In a similar way, reactions between ketone 1d and
nitrobenzaldehydes 2b–c were faster under ball-milling con-
ditions (Table 6, entries 2 and 3). As mentioned above, the
A
B
5
93
50:50
46:54
90
75
8
9
36 93
A
19 73
36 69
–
–
56
54
B[f]
[a] Reaction conditions: ketone 1 (2.2 mmol), aldehyde 2 (2.0 mmol), (S)-
proline (0.2 mmol). [b] MethodA: the reaction mixture was stirredin the
ball mill using a rotation speedof 250–400 rpm (see Supporting Informa-
tion for details). Method B: the reaction mixture was stirred at room
temperature using conventional magnetic stirring. [c] Combinedyieldof
isolated diastereomers. The data in parentheses indicate the amounts of
recovered aldehyde. [d] Determined by 1H NMR spectroscopy of the
crude product. [e] Determined by HPLC analysis of the anti isomer using
a chiral stationary phase. [f] Acetone (4.0 mmol, 2.0 equiv) was em-
ployed.
using this mechanochemical technique.[37] As expected,
while aldehydes with electron-withdrawing substituents af-
forded products with high stereoselectivities in short reac-
tion times (Table 5, entries 1–3), more electron-rich groups
Chem. Eur. J. 2007, 13, 4710 – 4722
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4715

C. Bolm et al.
Table 6. Solvent-free enantioselective aldol reaction between solely solid
milling conditions led to products in shorter reaction times
without significant loss of stereoselectivities (Table 6, en-
tries 5–8).
reagents catalyzedby ( S)-proline under ball-milling conditions.[a]
Entry Product
Method[b]
t
Yield
anti/
ee
[d] [%][c]
syn[d]
[%][e]
In the case of aldol reactions with purely solid reactants,
it is of particular advantage to conduct the catalysis without
using an excess of ketone (as requiredin solution). Besides
the economical factor, practical issues are important, in par-
ticular, because purification of the products is simplified.
In solid–solid reactions two macroscopic solid particles in-
teract directly and form a product (which can also be a
solid) without the intervention of a liquid or vapor phase.
Some proceedquantitatively without formation of waste,
which makes them easily scalable. Furthermore, no solvent
A
B
1.4 85
91:9
93:7
91
89
1
5
58 (24)
A
B
1
5
80 (9)
82
78:22
92:8
92
95
2
3
4
can be lost in an isolation step, making processes of this
[15,19]
type sustainable andenvironmentally benign.
Although
some reactions between solely solidcompounds have been
reportedto proceed“in the solidphase”, in most cases,
mixing the solidreactants results in the formation of a
liquidphase. Thus, some authors idfferentiate between
purely solid-state reactions (also called solid–solid reac-
A
B
1
6
66 (6)
76 (6)
81:19
77:23
88
92
[41]
tions)[15] andthose that proceedvia an intermediate melt.
A
B
1.6 75 (11)
85 (10)
92:8
87:13
93
93
7
When this liquefaction occurs at room temperature, it im-
plies the existence of a eutectic mixture with Tfusion below
ambient temperature (even when all reagents have higher
than ambient melting points).[41] In our experiments per-
formedat room temperature using conventional magnetic
stirring, reactions between solely solidcompounds only pro-
ceeded when a phase change to a melt occurred.[42] In a sim-
ilar way, when solid aldol products 3a–i were formedfrom
liquidketones 1a–c, it was possible to observe that the ini-
tial (paste-likes) heterogeneous mixture of solid aldehyde 2
A
B
1.4 79
96:4
98:2
90
96
5
6
7
7
47 (32)
A
B
1.5 75 (10)
88:12
93:7
89
96
7
59 (18)
andproline dispersedin the liquidketone
1 evolvedto a
partially homogeneous (honey-like) intermediate melt,[43]
from which the aldol products crystallized as they were
formed(Table 1 andTable 5). As a consequence of the con-
tinuous removal of the product from the reaction mixture,
the aldol reaction is inherently irreversible under solvent-
free conditions (unlike similar reactions in solution). In the
ball-milling process the increasedefficiency in mixing, the
slightly elevatedtemperature andthe pressure exhibitedby
the balls upon stirring are beneficial for a more rapidmelt
formation, which corresponds to the observed rate enhance-
ment.
A
B
1.5 59
73
82:18
93:7
96
98
6
A
B
1.4 72
77
95:5
99:1
85
82
8
7
[a]–[e] As in Table 5.
Additional evidence for the hypothesis that the existence
of a liquidphase is a prerequisite for the reaction to occur is
the observation that some reactions requiredheating
(Table 7). When less reactive, solidketones 1-Boc-4-piperi-
done (1 f) and4-phenylcyclohexanone ( 1g) were usedas
donors, no reaction or phase change was observed after stir-
ring of the reaction mixture for eight days at room tempera-
ture (Table 7, entries 1 and2). While the reaction compo-
nents remainedsolids at room temperature, melting was ob-
servedat 60 8C, which paralleledwith some reaction prog-
ress. At this temperature, however, the enantioselectivity
eroded (Table 7) and in some cases even racemates were ob-
tained. Noteworthy, by means of ball milling these aldol re-
actions proceeded much faster (Table 7, entries 1 and 2), in-
reactions were slower when aldehydes with more electron-
rich substituents were utilized. Thus, after seven days under
conventional stirring, the reaction between ketone 1d and
p-chlorobenzaldehyde (2d) afforded the aldol product 3m
in 85% yield. Remarkably, only 1.6 days under ball-milling
conditions were required to obtain 3m in 75% yieldwith
excellent stereoselectivities (92:8 anti/syn and93% ee;
Table 6, entry 4). Due to the poor reactivity observedin
these solid–solid systems, other aldehydes with more elec-
tron-rich substituents were not examined. Again in those re-
actions where solid4-thiopyranone 1e was employed, ball-
4716
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722

Asymmetric Organocatalytic Aldol Reaction
FULL PAPER
Table 7. Solvent-free enantioselective aldol reaction catalyzed by (S)-
andreactions between achiral organic molecules in the pres-
ence of enantiomerically enrichedamino acids in a presuma-
bly hot, geologically active prebiotic world. Such reactions
are a possible source of complex molecules containing mul-
tiple stereocenters starting from a catalytic amount of enan-
tiopure amino acidandsimilar processes couldbe impli-
catedas steps in the evolution of homochirality.
proline.[a]
Entry Product
Method[b]
t
Yield
anti/
ee
[d] [%][c]
syn[d]
[%][e]
A
1.5 50 (21)
86:14
–
41:59
55
–
23
1
B
8
8
–
B[f]
52 (6)
Conclusion
A[g]
B
2
8
8
42 (29)
–
42 (4)
69:31
–
45:55
84
–
rac
2
3
In this study, proline-catalyzed aldol reactions have been in-
vestigated under solvent-free conditions. The effects of con-
centration, water andimpurities on the reaction outcome
have been establishedrevealing that: 1) a reagent ratio
close to 1:1 is optimal, andusing an excess of ketone can
cause a decrease in both yield and stereoselectivity; 2) nei-
ther dry conditions nor inert atmosphere are required;
3) there is no needto use purifiedstarting materials.
For achieving optimal results in terms of yieldandstereo-
selectivity an intense mixing of all reagents provedto be in-
dispensable. Along these lines, ball milling has been used in
asymmetric organocatalysis for the first time. The advan-
tages of using this mechanochemical technique are: 1) Orga-
nocatalyzedasymmetric aldol reactions proceedfaster than
under commonly applied mechanical stirring leading to high
yieldof anti-aldol products 3 with excellent diastereoselec-
tivities andenantioselectivities. 2) After optimization of
milling cycles, the reactions are remarkably clean yielding
mostly highly crystalline solids, which are easy to isolate.
3) Catalytic amounts of inexpensive proline can be used,
anddifficult to prepare catalysts are not required. 4) Almost
equimolar amounts of the starting materials react well,
avoiding the common uneconomical and impractical use of
a large excess of ketone. As additional benefit, the isolation
of the product from the reaction mixture becomes easier,
since at that stage the entire starting materials are con-
sumed.
B[f,g]
A
3
8
8
traces
traces
26 (60)
n.d.
n.d.
40:60
n.d.
n.d.
rac
B
B[f]
[a]–[e] As in Table 5. [f] Reaction performedat 60 8C. [g] Depicted anti,-
trans-3s is the dominant isomer under ball-milling conditions. Conven-
tional stirring at 608C leads predominantly to syn,cis-3s.
dicating once more the advantage of using this mechano-
chemical technique.[37]
Use of liquidcycloheptanone ( 1h) as donor showed that
the existence of a liquidphase does not necessarily imply re-
activity either at ambient temperature or in the ball mill
(Table 7, entry 3). Indeed, this system was already known to
[6]
be poorly reactive andunselective, andhere only low con-
version was achievedwhen the mixture was stirredat 60 8C
for eight days leading to product 3t in 26% yieldwithout
any stereoselectivity.
Only a few recently reportedexamples can be foundin
the literature where the solidketones 1d–f were usedas
donors in an organocatalyzed direct aldol reaction, all of
them performedin organic solvents.
[12,44]
Among them, the
study published by Pihko and co-workers attracted our
attention, since there the proline-catalyzedenantioselective
aldol reaction was also performed with stoichiometric
amount of reactants.[12] This allowedus to directly compare
their results obtainedin DMF solution (with up to five
equivalents of water, which was foundto accelerate the ca-
talysis) with ours achieved under solvent-free conditions. In
their study, product 3n was formedin 45% yieldafter five
days, and product 3q in 54% yieldafter three days. By
means of a ball milling we obtainedproducts 3n and 3q in
79 and 72% yield, respectively, after only 1.4 days. The dia-
stereoselectivities were similar to those described by Pihko
andco-workers, while the enantioselectvities were somewhat
Noteworthy, the preparation of 3a can be performedin a
10 g scale in the ball mill without significant change in yield
[45]
andwith the same stereoselectivity.
The results reportedin this article may serve as useful
guidelines for the development of other enantioselective or-
ganocatalytic carbon carbon bond-forming reactions under
ꢀ
solvent-free conditions.
Experimental Section
lower (90 and85% ee, comparedwith 99 and98%
ee,
General information: All reagents were purchasedfrom commercial sup-
pliers. When purifiedreagents were used, cyclohexanone was distilled
and p-nitrobenzaldehyde was sublimated before use. Reactions in the
ball mill were conducted using a Fritsch Planetary Micro Mill model
“Pulverisette 7”. The milling instrument consists of a main disk which
can rotate at a speedof 100–800 rpm andaccommodates two grinding
bowls. Both bowls (45 mL) and balls (5 mm diameter) are made of chem-
ically inert and non-abrasive zirconium oxide. Grinding cycles (including
a pause period) are programmed and repeated. Flash chromatography
respectively). At least in terms of reactivity these results
also highlight advantages of the ball-milling methodology.
It is fascinating to speculate on the correspondence be-
tween catalytic enantioselective reactions in a ball mill
(which allow for the creation of conditions of high pressure
andtemperature, and, in the present case for crystallization
of a highly enantiomerically enrichedproduct from a melt)
Chem. Eur. J. 2007, 13, 4710 – 4722
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4717

C. Bolm et al.
was carriedout with Merck silica gel 60 (63–200 mesh). For thin-layer
chromatography (TLC), Merck 60 F254 silica gel plates were usedand
compounds were visualized by means of irradiation with UV light and/or
by treatment with a solution of p-anisaldehyde (9.2 mL), concentrated
H2SO4 (12.5 mL), acetic acid(3.8 mL), andethanol (340 mL) followedby
heating. 1H NMR and 13C NMR (300 or 400 MHz and75 or 100 MHz, re-
spectively) spectra were recorded in CDCl3. Chemical shifts (d) are
reported in ppm using TMS as internal standard, and spin–spin coupling
constants (J) are given in Hz. IR spectra were recorded on a Perkin-
Elmer FT/IR 1760 as KBr pellets, neat or in CHCl3 andare given in
cmꢀ1. Melting points were measuredin open glass capillaries andare un-
corrected. Optical rotation measurements were determined at 228C using
HPLC-grade solvents. Analytical HPLC measurements were carried out
on a Gynkotek (Dionex) machine with autosampler GINA 50, UV/VIS
detector UVD 170S, gradient pump M480G and degasser DG 503.
H), 2.32–2.41 [m, 1H; CHHC(O)], 2.47–2.53 [m, 1H; CHHC(O)], 2.55–
2.62 (m, 1H; CHCHOH), 4.07 (brs, 1H; CHOH), 4.90 (d, J=8.4 Hz,
1H; CHOH), 7.50–7.52 (m, 2H; ArH), 8.19–8.23 ppm (m, 2H; ArH);
13C NMR (75 MHz, CDCl3): d=24.7 (CH2), 27.6 (CH2), 30.7 (CH2), 42.6
(CH2), 57.2 (CH), 74.0 (CH), 123.5 (CH, 2C), 127.8 (CH, 2C), 147.5 (C),
148.3 (C), 214.7 ppm (C).
A
[(2S,1’R)-3b]:[27] The enantiomeric excess was determined by HPLC
(Daicel Chiralpak AD, heptane/iPrOH 92:8, flow rate 0.5 mLminꢀ1, l=
210 nm): tR =49.3 min (major), tR =63.1 min (minor). Data for the major
isomer (2S,1’R)-3b: 1H NMR (400 MHz, CDCl3): d=1.33–1.73 (m, 4H;
cyclohex-H), 1.79–1.86 (m, 1H; cyclohex-H), 2.08–2.15 (m, 1H; cyclohex-
H), 2.33–2.41 [m, 1H; CHHC(O)], 2.47–2.53 [m, 1H; CHHC(O)], 2.58–
2.65 (m, 1H; CHCHOH), 4.12 (brs, 1H; CHOH), 4.89 (d, J=8.5 Hz,
1H; CHOH), 7.52 (t, J=7.9 Hz, 1H; ArH), 7.65–7.68 (m, 1H; ArH),
8.16 (ddd, J=8.2, 2.3, 1.1 Hz, 1H; ArH), 8.20–8.21 ppm (m, 1H; ArH);
13C NMR (75 MHz, CDCl3): d=24.6 (CH2), 27.6 (CH2), 30.6 (CH2), 42.6
(CH2), 57.0 (CH), 73.9 (CH), 121.9 (CH), 122.8 (CH), 129.2 (CH), 133.2
(CH), 143.2 (C), 148.2 (C), 214.8ppm (C).
General experimental procedure for the solvent-free aldol reaction cata-
lyzed by (S)-proline under inert and dry conditions: Under argon, a dry
10 mL round-bottomed flask charged with a magnetic stirring bar was
filledwith cyclohexanone ( 1a, 230 mL, 2.2 mmol, 1.1 equiv), p-nitroben-
zaldehyde (2a, 302 mg, 2.0 mmol, 1.0 equiv), (S)-proline (23 mg,
0.2 mmol, 0.1 equiv), and an additive (if used). Stirring was started using
a conventional magnetic stirring bar. After 24 h, the vessel was washed
with Et2O or CH2Cl2 (6”7 mL), andthe combinedorganic fractions were
filteredandconcentratedin vacuo. Purification by flash chromatography
on silica gel (pentane/EtOAc 100:0 ! 80:20) afforded the pure aldol
product 3a in the yields and enantioselectivities reported in Tables 1–3
(see main Text). The identity and purity of the aldol product 3a was con-
firmedby 1H and 13C NMR spectroscopic analysis. The diastereoselec-
tivity of the reaction was determined by 1H NMR spectroscopy of the
crude product.
AHCTREUNG
,
l=254 nm): tR =32.0 min (major), tR =38.7 min (minor). Data for the
major isomer (2S,1’R)-3c: 1H NMR (300 MHz, CDCl3): d=1.55–1.87 (m,
5H; cyclohex-H), 2.05–2.14 (m, 1H; cyclohex-H), 2.28–2.39 [m, 1H;
CHHC(O)], 2.41–2.49 [m, 1H; CHHC(O)], 2.71–2.80 (m, 1H;
CHCHOH), 3.98 (brs, 1H; CHOH), 5.44 (d, J=7.1 Hz, 1H; CHOH),
7.39–7.45 (m, 1H; ArH), 7.60–7.66 (m, 1H; ArH), 7.75–7.78 (m, 1H;
ArH), 7.83–7.86 ppm (m, 1H; ArH); 13C NMR (100 MHz, CDCl3): d=
25.0 (CH2), 27.8 (CH2), 31.1 (CH2), 42.8 (CH2), 57.3 (CH), 69.7 (CH),
123.9 (CH), 128.3 (CH), 128.9 (CH), 132.9 (CH), 136.4 (C), 148.5 (C),
214.7 ppm (C).
Typical experimental procedure for the solvent-free aldol reaction cata-
lyzed by (S)-proline
Reaction performed in the ball mill (Method A): A ball-mill vessel was
chargedwith ketone 1 (2.2 mmol, 1.1 equiv), aldehyde 2 (2.0 mmol,
1.0 equiv), (S)-proline (23 mg, 0.2 mmol, 0.1 equiv), and60 zirconium
oxide balls. Stirring was started in a grinding bowl using the ball mill with
a rotation speedof 250–400 rpm (see Supporting Information for details).
After an appropriate reaction time, the crude product was washed off the
reaction vessel with Et2O or CH2Cl2 (4”40 mL), andthe combinedor-
ganic fractions were filteredandconcentratedin vacuo. Purification by
flash chromatography on silica gel (pentane/EtOAc 100:0 ! 80:20)
afforded the pure aldol product 3 in the yields and enantioselectivities re-
ported in Tables 4–7 (see main Text). The identity and purity of the prod-
uct was confirmedby 1H and 13C NMR spectroscopic analysis. Aldol
products anti-3a–h, 3i, anti-3n–r and anti-3t are known com-
pounds.[6,12b,27,44b] The diastereoselectivity of the reaction was determined
by 1H NMR spectroscopy of the crude product. The enantiomeric excess
of anti-aldol products was determined by chiral-phase HPLC analysis.
The absolute configuration of aldol products was determined by compari-
son with publishedHPLC retention times.
C
,
Reaction performed in a round-bottom flask (Method B): A 10 mL
round-bottomed flask was charged with ketone 1 (2.2 mmol, 1.1 equiv),
aldehyde
2 (2.0 mmol, 1.0 equiv), and( S)-proline (23 mg, 0.2 mmol,
0.1 equiv). Stirring was startedusing a conventional magnetic stirring bar.
After an appropriate reaction time, the vessel was washedwith Et 2O or
CH2Cl2 (6”7 mL), andthe combinedorganic fractions were filteredand
concentrated in vacuo. The diastereoselectivity of the reaction was deter-
minedby 1H NMR spectroscopy of the crude product. Purification by
flash chromatography on silica gel (pentane/EtOAc 100:0 ! 80:20) af-
forded the pure aldol product 3 in the yields and enantioselectivities re-
ported in Tables 5–7 (see main Text). The identity and purity of the prod-
uct was confirmedby 1H and 13C NMR spectroscopic analysis.
(2S,1’R)-2-[1’-Hydroxy-1’-(4-nitrophenyl)methyl]cyclohexan-1-one
[(2S,1’R)-3a]:[27] The enantiomeric excess was determined by HPLC
(Daicel Chiralpak AD, heptane/iPrOH 90:10, flow rate 0.5 mLminꢀ1, l=
254 nm): tR =51.6 min (minor), tR =69.6 min (major). Data for the major
isomer (2S,1’R)-3a: 1H NMR (400 MHz, CDCl3): d=1.32–1.72 (m, 4H;
cyclohex-H), 1.80–1.88 (m, 1H; cyclohex-H), 2.08–2.15 (m, 1H; cyclohex-
4718
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722

Asymmetric Organocatalytic Aldol Reaction
FULL PAPER
A
analysis calcd(%) for C 17H23NO4 (305.4): C 66.86, H 7.59, N 4.59; found:
C 66.64, H 7.69, N 4.63.
[(2S,1’R)-3g]:[27] The enantiomeric excess was determined by HPLC
(Daicel Chiralpak AD, heptane/iPrOH 97:3, flow rate 0.5 mLminꢀ1, l=
254 nm): tR =140.1 min (minor), tR =145.9 min (major). Data for the
major isomer (2S,1’R)-3g: 1H NMR (400 MHz, CDCl3): d=1.45–1.73 (m,
3H; cyclopent-H), 1.91–2.00 (m, 1H; cyclopent-H), 2.16–2.44 [m, 3H;
CH2C(O) + CHCHOH], 4.69 (d, J=0.9 Hz, 1H; CHOH), 4.78 (d, J=
9.2 Hz, 1H; CHOH), 7.45–7.49 (m, 2H; ArH), 8.13–8.17 ppm (m, 2H;
ArH); 13C NMR (100 MHz, CDCl3): d=20.4 (CH2), 26.9 (CH2), 38.6
(CH2), 55.1 (CH), 74.4 (CH), 123.6 (CH, 2C), 127.2 (CH, 2C), 147.5 (C),
148.5 (C), 222.0 ppm (C).
(2S,4S,1’R)-4-tert-Butyl-2-[1’-hydroxy-1’-(2-nitrophenyl)methyl]cyclohex-
an-1-one [(2S,4S,1’R)-3l]: The enantiomeric excess was determined by
HPLC (Daicel Chiralpak AD, heptane/iPrOH 85:15, flow rate
0.4 mLminꢀ1, l=210 nm): tR =22.7 min (minor), tR =25.8 min (major).
Data for the major isomer (2S,4S,1’R)-3l: M.p. 83–848C; [a]2D2 =ꢀ4.5 (c=
1.00 in CHCl3, 90% ee); 1H NMR (400 MHz, CDCl3): d=0.78 (s, 9H;
CH3), 1.50–1.70 (m, 4H; cyclohex-H), 1.88–1.93 (m, 1H; cyclohex-H),
2.37–2.46 [m, 2H; CH2C(O)], 2.80–2.86 (m, 1H; CHCHOH), 3.55 (br,
1H; CHOH), 5.46 (d, J=7.2 Hz, 1H; CHOH), 7.43 (dt, J=7.8, 1.4 Hz,
1H; ArH), 7.63 (dt, J=8.4, 1.4 Hz, 1H; ArH), 7.73 (dd, J=8.0, 1.4 Hz,
1H; ArH), 7.84 ppm (dd, J=8.0, 1.3 Hz, 1H; ArH); 13C NMR (100 MHz,
CDCl3): d=24.3 (CH2), 27.0 (CH3, 3C), 27.6 (CH2), 32.8 (C), 39.7 (CH2),
42.4 (CH), 53.7 (CH), 69.7 (CH), 124.3 (CH), 128.7 (CH), 129.1 (CH),
133.3 (CH), 136.9 (C), 148.7 (C), 216.1 ppm (C); IR (KBr): n˜ =3455,
2870, 1702, 1527, 1446, 1353, 1093, 1032, 861 cmꢀ1; MS (EI): m/z (%): 306
(53) [(M+1)+], 288 (13) [(M+1)+ꢀH2O], 155 (100), 152 (86); elemental
analysis calcd(%) for C 17H23NO4 (305.4): C 66.86, H 7.59, N 4.59; found:
C 66.48, H 7.76, N 4.63.
A
[(2S,1’R)-3h]:[27] The enantiomeric excess was determined by HPLC
(Daicel Chiralpak AD, heptane/iPrOH 96:4, flow rate 1.0 mLminꢀ1, l=
254 nm): tR =43.7 min (major), tR =67.4 min (minor). Data for the major
isomer (2S,1’R)-3h: 1H NMR (300 MHz, CDCl3): d=1.50–1.55 (m, 1H;
cyclopent-H), 1.61–1.78 (m, 2H; cyclopent-H), 1.92–2.01 (m, 1H; cyclo-
pent-H), 2.16–2.47 [m, 3H; CH2C(O)
CHOH), 4.77 (d, J=9.3 Hz, 1H; CHOH), 7.47 (t, J=7.9, Hz, 1H; ArH),
7.62–7.65 (m, 1H; ArH), 8.10 (ddd, J=8.1, 2.3, 1.1 Hz, 1H; ArH), 8.17–
8.18 ppm (m, 1H; ArH); 13C NMR (75 MHz, CDCl3): d=20.3 (CH2),
26.8 (CH2), 38.6 (CH2), 55.0 (CH), 74.3 (CH), 121.5 (CH), 122.9 (CH),
129.4 (CH), 132.6 (CH), 143.6 (C), 148.3 (C), 222.3 ppm (C).
(4R)-4-Hydroxy-4-nitrophenylbutan-1-one (3i):[27] The enantiomeric
excess was determined by HPLC (Daicel Chiralpak AS, heptane/iPrOH
70:30, flow rate 0.5 mLminꢀ1, l=254 nm): tR =26.4 min (major), tR =
35.4 min (minor). 1H NMR (400 MHz, CDCl3): d=2.22 (s, 3H; CH3),
2.84–2.86 (m, 2H; CH2CHOH), 3.59 (s, 1H; CHOH), 5.26 (dd, J=8.0,
4.3 Hz, 1H; CHOH), 7.52–7.55 (m, 2H; ArH), 8.19–8.22 ppm (m, 2H;
ArH); 13C NMR (100 MHz, CDCl3): d=30.7 (CH3), 51.5 (CH2), 68.8
(CH), 123.5 (CH, 2C), 126.3 (CH, 2C), 147.0 (C), 149.9 (C), 208.2 ppm
(C).
(2S,4S,1’R)-4-tert-Butyl-2-[1’-hydroxy-1’-(4-chlorophenyl)methyl]cyclo-
hexan-1-one [(2S,4S,1’R)-3m]: The enantiomeric excess was determined
by HPLC (Daicel Chiralpak AD, heptane/iPrOH 95:5, flow rate
0.5 mLminꢀ1, l=230 nm): tR =33.3 min (minor), tR =63.8 min (major).
Data for the major isomer (2S,4S,1’R)-3m: M.p. 150–1518C; [a]2D2 =ꢀ6.3
(c=1.00 in CHCl3, 93% ee); 1H NMR (400 MHz, CDCl3): d=0.79 (s,
9H; CH3), 1.38–1.59 (m, 4H; cyclohex-H), 1.96–2.02 (m, 1H; cyclohex-
H), 2.41–2.54 [m, 2H; CH2C(O)], 2.62 (dt, J=9.6, 6.6 Hz, 1H;
CHCHOH), 3.28 (br, 1H; CHOH), 4.86 (dd, J=9.6, 2.8 Hz, 1H;
CHOH), 7.27–7.36 ppm (m, 4H; ArH); 13C NMR (100 MHz, CDCl3): d=
25.2 (CH2), 27.2 (CH3, 3C), 27.6 (CH2), 32.9 (C), 39.3 (CH2), 42.3 (CH),
55.2 (CH), 74.2 (CH), 128.2 (CH, 2C), 128.7 (CH, 2C), 133.9 (C), 139.8
(C), 215.7 ppm (C); IR (KBr): n˜ =3472, 2958, 2871, 1699, 1486, 1091,
1041, 830 cmꢀ1; MS (EI): m/z (%): 294 (1) [M +], 276 (7) [M +ꢀH2O],
154 (66), 140 (32), 139 (100), 70 (51), 57 (35); elemental analysis calcd
(%) for C17H23O2Cl (294.8): C 69.26, H 7.86; found: C 69.19, H 8.07.
(2S,4S,1’R)-4-tert-Butyl-2-[1’-hydroxy-1’-(4-nitrophenyl)methyl]cyclohex-
an-1-one [(2S,4S,1’R)-3j]: The enantiomeric excess was determined by
HPLC (Daicel Chiralcel OD-H, heptane/iPrOH 92:8, flow rate
0.5 mLminꢀ1, l=280 nm): tR =53.4 min (major), tR =66.6 min (minor).
Data for the major isomer (2S,4S,1’R)-3j, which was obtainedin enantio-
merically pure form after recrystallization: M.p. 184–1858C; [a]2D2 =ꢀ11.4
(c=1.00 in CHCl3, ꢁ99% ee); 1H NMR (400 MHz, CDCl3): d=0.79 (s,
9H; CH3), 1.34–1.68 (m, 4H; cyclohex-H), 1.94–2.00 (m, 1H; cyclohex-
H), 2.38–2.47 (m, 1H; CHHC(O)), 2.51–2.57 [m, 1H; CHHC(O)], 2.63–
2.69 (m, 1H; CHCHOH), 3.62 (br, 1H; CHCHOH), 4.97 (dd, J=9.4,
2.9 Hz, 1H; CHCHOH), 7.53–7.55 (m, 2H; ArH), 8.23–8.25 ppm (m,
2H; ArH); 13C NMR (100 MHz, CDCl3): d=24.3 (CH2), 27.0 (CH2), 27.1
(CH3, 3C), 33.0 (C), 39.4 (CH2), 42.5 (CH), 54.4 (CH), 74.0 (CH), 123.7
(CH, 2C), 127.7 (CH, 2C), 147.7 (C), 148.5 (C), 215.7 ppm (C); IR
(KBr): n˜ =3468, 2956, 1702, 1600, 1520, 1452, 1392, 1343, 1267, 1042,
856 cmꢀ1; MS (EI): m/z (%): 305 (2) [M +], 287 (13) [M +ꢀH2O], 154
(98), 151 (72), 139 (97), 70 (78), 57 (100); elemental analysis calcd(%)
for C17H23NO4 (305.4): C 66.86, H 7.59, N 4.59; found: C 67.12, H 7.33, N
4.59.
A
one [(3S,1’R)-3n]:[12b,44b] The enantiomeric excess was determined by
HPLC (Daicel Chiralcel OD-H, heptane/iPrOH 92:8, flow rate
0.5 mLminꢀ1, l=254 nm): tR =79.7 min (major), tR =127.2 min (minor).
Data for the major isomer (3S,1’R)-3n: 1H NMR (400 MHz, CDCl3): d=
2.51 [ddd, J=13.7, 4.7, 2.2 Hz, 1H; SCHHCHC(O)], 2.66 [dd, J=13.7,
11.0 Hz, 1H; SCHHCHC(O)], 2.74–2.83 (m, 2H; SCH2), 2.94–3.04 [m,
3H; CH2C(O) + CHCHOH], 3.64 (d, J=4.1 Hz, 1H; CHOH), 5.05 (dd,
J=8.2, 3.9 Hz, 1H; CHOH), 7.54 (d, J=8.8 Hz, 2H; ArH), 8.23 ppm (d,
J=8.8 Hz, 2H; ArH); 13C NMR (100 MHz, CDCl3): d=30.8 (CH2), 32.8
(CH2), 44.8 (CH2), 59.5 (CH), 73.2 (CH), 123.8 (CH, 2C), 127.8 (CH,
2C), 147.7 (C), 147.8 (C), 211.2 ppm (C).
A
one [(3S,1’R)-3o]:[44b] The enantiomeric excess was determined by HPLC
(Daicel Chiralpak AD, heptane/iPrOH 92:8, flow rate 0.7 mLminꢀ1, l=
254 nm): tR =67.9 min (major), tR =85.0 min (minor). Data for the major
isomer (3S,1’R)-3o: 1H NMR (400 MHz, CDCl3): d=2.52 [ddd, J=13.7,
4.7, 2.2 Hz, 1H; SCHHCHC(O)], 2.68 [dd, J=13.7, 11.0 Hz, 1H;
SCHHCHC(O)], 2.76–2.85 (m, 2H; SCH2), 2.95–3.07 [m, 3H; CH2C(O)
+ CHCHOH], 3.69 (d, J=4.0 Hz, 1H; CHOH), 5.04 (dd, J=8.2, 4.0 Hz,
1H; CHOH), 7.55 (t, J=7.9 Hz, 1H, ArH), 7.70 (d, J=7.7 Hz, 1H,
ArH), 8.16–8.24 ppm (m, 2H, ArH); 13C NMR (100 MHz, CDCl3): d=
30.8 (CH2), 32.8 (CH2), 44.8 (CH2), 59.4 (CH), 73.23 (CH), 122.0 (CH),
123.2 (CH), 129.6 (CH), 133.0 (CH), 142.6 (C), 148.4 (C), 211.4 ppm (C).
(2S,4S,1’R)-4-tert-Butyl-2-[1’-hydroxy-1’-(3-nitrophenyl)methyl]cyclohex-
an-1-one [(2S,4S,1’R)-3k]: The enantiomeric excess was determined by
HPLC (Daicel Chiralcel OG, heptane/iPrOH 85:15, flow rate
0.7 mLminꢀ1, l=210 nm): tR =30.0 min (minor), tR =43.3 min (major).
Data for the major isomer (2S,4S,1’R)-3k, which was obtainedin enantio-
merically pure form after recrystallization: M.p. 154–1558C; [a]2D2 =ꢀ5.3
(c=1.00 in CHCl3, ꢁ99% ee); 1H NMR (400 MHz, CDCl3): d=0.79 (s,
9H; CH3), 1.37–1.68 (m, 4H; cyclohex-H), 1.97–2.02 (m, 1H; cyclohex-
H), 2.41–2.56 [m, 2H; CH2C(O)], 2.65–2.71 (m, 1H; CHCHOH), 3.69 (d,
J=3.3 Hz, 1H; CHOH), 4.99 (dd, J=9.3, 3.3 Hz, 1H; CHOH), 7.56 (t,
J=7.8 Hz, 1H; ArH), 7.72 (d, J=7.7 Hz, 1H; ArH), 8.18–8.21 (m, 1H;
ArH), 8.24–8.26 ppm (m, 1H; ArH); 13C NMR (100 MHz, CDCl3): d=
24.6 (CH2), 27.1 (CH3, 3C), 27.3 (CH2), 32.9 (C), 39.4 (CH2), 42.4 (CH),
54.7 (CH), 74.0 (CH), 121.9 (CH), 123.1 (CH), 129.5 (CH), 132.9 (CH),
143.5 (C), 148.2 (C), 215.7 ppm (C); IR (KBr): n˜ =3328, 2962, 1702, 1526,
1444, 1345, 1218, 1060, 823 cmꢀ1; MS (EI): m/z (%): 305 (2) [M +], 287
(13) [M +ꢀH2O], 154 (99), 151 (45), 139 (100), 70 (73), 57 (67); elemental
A
one [(3S,1’R)-3p]:[44b] The enantiomeric excess was determined by HPLC
(Daicel Chiralpak AD, heptane/iPrOH 90:10, flow rate 0.5 mLminꢀ1, l=
210 nm): tR =64.6 min (minor), tR =82.9 min (major). Data for the major
isomer (3S,1’R)-3p: 1H NMR (400 MHz, CDCl3): d=2.61 [ddd, J=13.6,
4.5, 2.4 Hz, 1H; SCHHCHC(O)], 2.73–2.84 (m, 2H; SCH2), 2.88–3.02
[m, 3H; SCHHCHC(O), CH2C(O)], 3.11–3.19 (m, 1H; CHCHOH), 5.54
(d, J=6.7 Hz, 1H; CHOH), 7.45 (dt, J=7.2, 1.5 Hz, 1H; ArH), 7.66 (dt,
Chem. Eur. J. 2007, 13, 4710 – 4722
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4719

C. Bolm et al.
J=7.7, 1.5 Hz, 1H; ArH), 7.77 (dd, J=7.9, 1.5 Hz, 1H; ArH), 7.89 ppm
(dd, J=8.2, 1.5 Hz, 1H; ArH); 13C NMR (100 MHz, CDCl3): d=30.8
(CH2), 33.4 (CH2), 45.2 (CH2), 59.6 (CH), 69.4 (CH), 124.4 (CH), 128.9
(CH), 129.0 (CH), 133.5 (CH), 136.0 (C), 148.6 (C), 211.5 ppm (C).
A
one [(3S,1’R)-3q]:[12b,44b] The enantiomeric excess was determined by
HPLC (Daicel Chiralcel OD-H, heptane/iPrOH 92:8, flow rate
0.5 mLminꢀ1, l=210 nm): tR =35.6 min (major), tR =53.3 min (minor).
Data for the major isomer (3S,1’R)-3q: 1H NMR (400 MHz, CDCl3): d=
2.49 [ddd, J=13.7, 4.9, 1.9 Hz, 1H; SCHHCHC(O)], 2.57 [dd, J=13.7,
10.1 Hz, 1H; SCHHCHC(O)], 2.72–2.87 (m, 2H; SCH2), 2.91–2.99 [m,
3H; CH2C(O) + CHCHOH], 3.46 (d, J=3.6 Hz, 1H; CHOH), 4.93 (dd,
J=8.5, 3.3 Hz, 1H; CHOH), 7.27–7.35 ppm (m, 4H; ArH); 13C NMR
(100 MHz, CDCl3): d=31.0 (CH2), 32.9 (CH2), 44.7 (CH2), 59.7 (CH),
73.3 (CH), 128.3 (CH, 2C), 128.8 (CH, 2C), 134.0 (C), 138.7 (C),
211.5 ppm (C).
A
boxylic acid tert-butyl ester [(3S,1’R)-3r]:[44b] When the reaction was per-
formedin the ball mill, the major isomer obtainedwas (3 S,1’R)-3r. The
enantiomeric excess was determined by HPLC (Daicel Chiralpak AD-H,
heptane/iPrOH 85:15, flow rate 0.5 mLminꢀ1, l=210 nm): tR =26.0 min
(major), tR =28.0 min (minor). Data for the isomer (3S,1’R)-3r: 1H NMR
(400 MHz, CDCl3): d=1.40 (brs, 9H; CH3), 2.48–2.59 (m, 2H;
CH2C(O)), 2.76 (br, 1H; CHCHOH), 2.93 (brt, J=11.3 Hz, 1H, piperi-
dinyl-H), 3.27 (br, 1H, piperidinyl-H), 3.70–3.88 (m, 1H, piperidinyl-H),
3.88 (d, J=3.6 Hz, 1H, CHOH), 4.11–4.17 (m, 1H, piperidinyl-H), 4.99
(dd, J=8.0, 2.7 Hz, 1H; CHOH), 7.56 (d, J=8.9 Hz, 2H; ArH),
8.24 ppm (d, J=8.9 Hz, 2H; ArH); 13C NMR (100 MHz, CDCl3): d=28.3
(CH3, 3C), 41.5 (CH2), 42.8 (CH2), 43.7 (CH2), 56.4 (CH), 71.7 (CH),
81.0 (C), 123.6 (CH, 2C), 127.7 (CH, 2C), 147.1 (C), 147.6 (C), 154.2 (C),
210.0 ppm (C).
flow rate 0.5 mLminꢀ1, l=254 nm): tR =39.0 min (major), tR =92.6 min
(minor). Data for the anti isomer (2S,1’R)-3t: 1H NMR (400 MHz,
CDCl3): d=1.25–1.42 (m, 3H; cyclohept-H), 1.56–1.92 (m, 5H; cyclo-
hept-H), 2.43–2.58 [m, 2H; CH2C(O)], 2.95–3.00 (m, 1H; CHCHOH),
3.72 (brs, 1H; CHOH), 4.92 (d, J=7.4 Hz, 1H; CHOH), 7.51–7.54 (m,
2H; ArH), 8.20–8.23 ppm (m, 2H; ArH); 13C NMR (75 MHz, CDCl3):
d=23.4 (CH2), 28.2 (CH2), 28.6 (CH2), 28.7 (CH2), 44.1 (CH2), 57.8
(CH), 74.8 (CH), 123.6 (CH, 2C), 127.7 (CH, 2C), 147.5 (C), 149.2 (C),
216.9 ppm (C).
(2S,4S,1’R)-2-[1’-Hydroxy-1’-(4-nitrophenyl)methyl]-4-phenylcyclohexan-
1-one [(2S,4S,1’R)-3s]: When the reaction was performedin the ball mill,
the major isomer obtainedwas (2 S,4S,1’R)-3s. The enantiomeric excess
was determined by HPLC (Daicel Chiralpak AD-H, heptane/iPrOH
90:10, flow rate 0.5 mLminꢀ1, l=210 nm): tR =78.5 min (minor), tR =
119.4 min (major). Data for the isomer (2S,4S,1’R)-3s: M.p. 142–1438C;
[a]2D2 =+33.6 (c=1.00 in CHCl3, 84% ee); 1H NMR (300 MHz, CDCl3):
d=1.76–1.86 (m, 1H; cyclohex-H), 1.93–2.04 (m, 1H; cyclohex-H), 2.15–
2.24 (m, 1H; cyclohex-H), 2.36–2.43 (m, 1H; cyclohex-H), 2.54–2.61 [m,
2H; CH2C(O)], 2.75–2.83 (m, 1H; CHCHOH), 3.16–3.19 (m, 1H;
CHPh), 3.67 (d, J=3.2 Hz, 1H; CHOH), 5.04 (dd, J=8.7, 3.1 Hz, 1H;
CHOH), 7.13–7.33 (m, 5H; ArH), 7.53 (d, J=8.7 Hz, 2H; ArH),
8.23 ppm (d, J=8.4 Hz, 2H; ArH); 13C NMR (75 MHz, CDCl3): d=31.2
(CH2), 34.2 (CH2), 36.9 (CH), 39.2 (CH2), 54.1 (CH), 74.0 (CH), 123.7
(CH, 2C), 126.4 (CH, 2C), 126.7 (CH), 127.8 (CH, 2C), 128.8 (CH, 2C),
142.2 (C), 147.7 (C), 148.2 (C), 213.8 ppm (C); IR (KBr): n˜ =3345, 2919,
2856, 1702, 1597, 1508, 1340, 1055, 851 cmꢀ1; MS (EI): m/z (%): 325 (2)
[M +], 307 (14) [M +ꢀH2O], 174 (68), 151 (36), 104 (100); elemental analy-
sis calcd(%) for C 19H19NO4 (325.4): C 70.14, H 5.89, N 4.31; found: C
69.74, H 6.27, N 4.25.
Acknowledgements
We are grateful to the Fonds der Chemischen Industrie and to the
Deutsche Forschungsgemeinschaft (DFG) within the SPP 1179 (“Organo-
catalysis”) for financial support. A.B. thanks DFG for a doctoral fellow-
ship (GRK 440).
[1] For some selectedexamples, see: a) B. List, R. A. Lerner, C. F.
Barbas III, J. Am. Chem. Soc. 2000, 122, 2395–2396; b) D. Enders,
C. Grondal, Angew. Chem. 2005, 117, 1235–1238; Angew. Chem.
Int. Ed. 2005, 44, 1210–1212.
[2] For some recent reviews, see: a) B. List, Acc. Chem. Res. 2004, 37,
548–557; b) A. Berkessel, H. Grçger, Asymmetric Organocatalysis,
Wiley-VCH, Weinheim, 2005, Chapter 6; c) B. List, Chem.
Commun. 2006, 819–824.
[3] For very interesting contributions on the possible role of amino
acids as catalysts in the evolution of prebiotic homochirality, see:
a) S. Pizzarello, A. L. Weber, Science 2004, 303, 1151; b) M. Kluss-
mann, H. Iwamura, S. P. Mathew, D. H. Wells Jr, U. Pandya, A.
Armstrong, D. G. Blackmond, Nature 2006, 441, 621–623, andrefer-
ences therein; c) J. Crusats, S. Veintemillas-Verdaguer, J. M. Ribó,
Chem. Eur. J. 2006, 12, 7776–7781; d) see also: R. M. Kellogg,
Angew. Chem. 2007, 119, 498–502; Angew. Chem. Int. Ed. 2007, 46,
494–497.
[4] For recent reviews on organic reactions in aqueous media, see:
a) C.-J. Li, Chem. Rev. 2005, 105, 3095–3165; b) C.-J. Li, L. Chen,
Chem. Soc. Rev. 2006, 35, 68–82; c) H. C. Hailes, Org. Process Res.
Dev. 2007, 11, 114–120.
When the reaction was performedin a flask, after 8 dat 60 8C the major
anti isomer was (2S,4R,1’R)-3s. The enantiomeric excess was determined
by HPLC (Daicel Chiralpak AD-H, heptane/iPrOH 90:10, flow rate
0.5 mLminꢀ1, l=210 nm): tR =83.6 min (minor), tR =86.6 min (major).
Data for the isomer (2S,4R,1’R)-3s: 1H NMR (400 MHz, CDCl3): d=
1.65–1.71 (m, 2H; cyclohex-H), 1.89–2.04 (m, 1H; cyclohex-H), 2.22–2.25
(m, 1H; cyclohex-H), 2.58–2.62 [m, 2H; CH2C(O)], 2.81–2.88 (m, 1H;
CHCHOH), 2.98–3.05 (m, 1H; CHPh), 4.17 (d, J=3.0 Hz, 1H; CHOH),
4.96 (dd, J=8.8, 3.3 Hz, 1H; CHOH), 7.11–7.28 (m, 5H; ArH), 7.50 (d,
J=8.8 Hz, 2H; ArH), 8.15 ppm (d, J=8.8 Hz, 2H; ArH); 13C NMR
(100 MHz, CDCl3): d=34.8 (CH2), 37.7 (CH2), 42.3 (CH2), 42.9 (CH),
56.5 (CH), 73.8 (CH), 123.6 (CH, 2C), 126.5 (CH, 2C), 126.9 (CH), 127.9
(CH, 2C), 128.7 (CH, 2C), 143.6 (C), 147.5 (C), 148.0 (C), 213.8 ppm
(C).
[5] a) Y. Hayashi, T. Sumiya, J. Takahashi, H. Gotoh, T. Urushima, M.
Shoji, Angew. Chem. 2006, 118, 972–975; Angew. Chem. Int. Ed.
2006, 45, 958–961; b) Y. Hayashi, S. Aratake, T. Okano, J. Takaha-
shi, T. Sumiya, M. Shoji, Angew. Chem. 2006, 118, 5653–5655;
Angew. Chem. Int. Ed. 2006, 45, 5527–5529.
A
[(2S,1’R)-3t]:[6] The enantiomeric excess was determined by HPLC analy-
sis of the anti isomer (Daicel Chiralpak AD-H, heptane/iPrOH 90:10,
[6] N. Mase, Y. Nakai, N. Ohara, H. Yoda, K. Takabe, F. Tanaka, C. F.
Barbas III, J. Am. Chem. Soc. 2006, 128, 734–735.
4720
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722

Asymmetric Organocatalytic Aldol Reaction
FULL PAPER
[7] Additional examples of organocatalyzed aldol reactions performed
in water can be foundin: a) D. Font, C. Jimeno, M. A. Pericàs, Org.
Lett. 2006, 8, 4653–4655; b) Y. Wu, Y. Zhang, M. Yu, G. Zhao, S.
Wang, Org. Lett. 2006, 8, 4417–4420; c) Z. Jiang, Z. Liang, X. Wu,
Y. Lu, Chem. Commun. 2006, 2801–2803; d) P. Dziedzic, W. Zou, J.
Hµfren, A. Córdova, Org. Biomol. Chem. 2006, 4, 38–40; e) S. S.
Chimni, D. Mahajan, Tetrahedron: Asymmetry 2006, 17, 2108–2119;
f) for other examples of aldol reactions performed in aqueous sys-
tems, see A. Córdova, W. Notz, C. F. Barbas III, Chem. Commun.
2002, 3024–3025; g) Y.-Y. Peng, Q.-P. Ding, Z. Li, P. G. Wang, J.-P.
Cheng, Tetrahedron Lett. 2003, 44, 3871–3875.
T. Itoh, T. Okano, T. Sumiya, M. Shoji, Chem. Commun. 2007, 957–
959.
[18] Reviews: a) G. Kaupp, M. R. Naimi-Jamal, H. Ren, H. Zoz in Ad-
vanced Technologies based on Self-Propagating and Mechanochemi-
cal Reactions for Environmental Protection (Eds.: G. Cao, F.
Delogu, R. Orrꢂ), Research Signpost, Kerala, 2003, pp. 83–100;
b) S. Kipp, V. Sepelµk, K. D. Becker, Chem. Unserer Zeit 2005, 39,
384–392.
[19] For an interesting review on waste-free, large-scale transformations
using the ball-milling technique, see: G. Kaupp, CrystEngComm
2006, 8, 794–804.
[8] The insolubility of organic compounds in water can also cause them
to react on its surface, which implies that hydrogen-bond-type inter-
actions are responsible for rate accelerations andgreater selectivi-
ties observedin aqueous reaction meida comparedwith organic
ones. Sharpless andco-workers introudcedthe term “on-water”
chemistry to denote the reaction between insoluble reactants sus-
pended on water. For the descriptions of the concept, see: a) S. Nar-
ayan, J. Muldoon, M. G. Finn, V. V. Fokin, H. C. Kolb, K. B. Sharp-
less, Angew. Chem. 2005, 117, 3339–3343; Angew. Chem. Int. Ed.
2005, 44, 3275–3279; b) some additional details on the discussion
concerning this terminology can be foundin: A. P. Brogan, T. J.
Dickerson, K. D. Janda, Angew. Chem. 2006, 118, 8278–8280;
Angew. Chem. Int. Ed. 2006, 45, 8100–8102; c) Y. Hayashi, Angew.
Chem. 2006, 118, 8281–8282; Angew. Chem. Int. Ed. 2006, 45, 8103–
8104.
[20] K. Komatsu, Top. Curr. Chem. 2005, 254, 185–206.
[21] Z. Zhang, J. Gao, J.-J. Xia, G.-W. Wang, Org. Biomol. Chem. 2005,
3, 1617–1619.
[22] a) G. Kaupp, M. R. Naimi-Jamal, V. Stepanenko, Chem. Eur. J. 2003,
9, 4156–4160; b) G. Kaupp, M. R. Naimi-Jamal, J. Schmeyers, Tetra-
hedron 2003, 59, 3753–3760.
[23] C. L. Raston, J. L. Scott, Green Chem. 2000, 2, 49–52.
[24] V. P. Balema, J. W. Wiench, M. Pruski, V. K. Pecharsky, J. Am.
Chem. Soc. 2002, 124, 6244–6245.
[25] M. O. Rasmussen, O. Axelsson, D. Tanner, Synth. Commun. 1997,
27, 4027–4030.
[26] a) E. Tullberg, D. Peters, T. Frejd, J. Organomet. Chem. 2004, 689,
3778–3781; b) E. Tullberg, F. Schacher, D. Peters, T. Frejd, Synthesis
2006, 1183–1189.
[27] B. Rodríguez, T. Rantanen, C. Bolm, Angew. Chem. 2006, 118,
7078–7080; Angew. Chem. Int. Ed. 2006, 45, 6924–6926.
[28] T. Rantanen, I. Schiffers, C. Bolm, Org. Process Res. Dev. ASAP;
DOI: 10.1021/op6002743.
[9] Here, we wouldlike to refer to these reactions as performedeither
in water, in the presence of water or in an aqueous environment,
without participating in any controversy on that issue.
[10] For a general concept article, see M. C. Pirrung, Chem. Eur. J. 2006,
12, 1312–1317, andreferences therein.
[11] For a recently publishedapplication of this concept in the study of
organocatalyzedring opening of oxiranes, see: C. M. Kleiner, P. R.
Schreiner, Chem. Commun. 2006, 4315–4317.
[12] For an interesting study on the water-accelerated aldol reaction
catalyzedby proline in DMF using a stoichiometric amount of solid
ketones, see: a) A. I. Nyberg, A. Usano, P. M. Pihko, Synlett 2004,
1891–1896; b) P. M. Pihko, K. M. Laurikainen, A. Usano, A. I.
Nyberg, J. A. Kaavi, Tetrahedron 2006, 62, 317–328, andreferences
therein.
[29] When the solid aldehyde 2a was not usedas finely groundpowder,
only low conversion was observedandunreactedstarting material
was recoveredas a crystalline aggregate.
[30] The prolate ellipsoid-shaped magnetic stirring bar, which adjusts to
the shape of a round-bottom vessel, was employed because higher
conversions were observedwith this stirring bar comparedto others.
[31] When the same experiment was performedusing a cylindrical mag-
netic stirring bar, the aldol product was isolated in only 76% yield
(along with 20% of recovered aldehyde), with no differences in se-
lectivities. This experiment illustrates the importance of a suitable
mixing method.
[13] Some examples of metal-catalyzedreactions acceleratedby the
presence of water can be foundin ref. [8a] andreferences therein.
See also: M. Carril, R. SanMartin, I. Tellitu, E. Domínguez, Org.
Lett. 2006, 8, 1467–1470.
[32] Some isolatedexamples can be foundin refs. [5a] and[6].
[33] An obvious advantage of this is that the solvent-free protocol can
be more easily appliedon industrial scale, where a reagent purifica-
tion is difficult.
[14] For an overview on solvent-free organic reactions, see: a) G. W.
Cave, C. L. Raston, J. L. Scott, Chem. Commun. 2001, 2159–2169;
b) K. Tanaka, Solvent-Free Organic Synthesis, Wiley-VCH, Wein-
heim, 2003.
[15] For a recent review on highly efficient organic solid-state reactions,
see: a) G. Kaupp, Top. Curr. Chem. 2005, 254, 95–183; b) for infor-
mation on the mechanism of solid-state reactions, see: G. Kaupp,
CrystEngComm 2003, 5, 117–133.
[16] a) Green Chemistry: Theory and Practice (Eds.: P. T. Anastas, J. C.
Warner), OxfordUniversity Press, Oxford, 1998; b) see also: P. A.
Wender, S. T. Handy, D. L. Wright, Chem. Ind. 1997, 765, 767–769;
for some recent reviews on the use of green solvents in organic syn-
thesis, see: c) C. K. Z. Andrade, L. M. Alves, Curr. Org. Chem. 2005,
9, 195–218; d) R. A. Sheldon, Green Chem. 2005, 7, 267–278; e) M.
Avalos, R. Babiano, P. Cintas, J. L. Jimꢁnez, J. C. Palacios, Angew.
Chem. 2006, 118, 4008–4012; Angew. Chem. Int. Ed. 2006, 45, 3904–
3908.
[17] Some isolatedexamples of solvent-free organocatalytic aldol reac-
tions can be foundin ref. [5], see also: a) Y. Sekiguchi, A. Sasaoka,
A. Shimomoto, S. Fujioka, H. Kotsuki, Synlett 2003, 1655–1658;
b) Z. Tang, Z.-H. Yang, X.-H. Chen, L.-F. Cun, A.-Q. Mi, Y.-Z.
Jiang, L.-Z. Gong, J. Am. Chem. Soc. 2005, 127, 9285–9289; c) M.
Jiang, S.-F. Zhu, Y. Yang, L.-Z. Gong, X.-G. Zhou, Q.-L. Zhou, Tet-
rahedron: Asymmetry 2006, 17, 384–387; d) Y. Hayashi, S. Aratake,
[34] Although ZrO2 presents very goodresistance to abrasion andacid
treatment, it is not entirely non-abrasive. In order to ensure that the
ZrO2 is chemically inert anddoes not affect the reaction outcome, a
comparative experiment between cyclohexanone and p-nitrobenzal-
dehyde as substrates, in the presence of one equivalent of ZrO2 salt
as additive was performed. In this experiment, freshly purified start-
ing materials were used under inert and dry conditions. The reaction
mixture was stirredfor 24 h in a flask at room temperature using
conventional magnetic stirring. In this case aldol product 3a was ob-
tainedin 67% yield, with an anti/syn ratio of 80:20 andwith 94% ee
(compare with the result showedin Table 1, entry 2). The decrease
in the reaction rate was not surprising, since the stirring mode was
different and the addition of an inert salt resulted in a decrease in
the concentration of the components in the reaction mixture.
[35] In accordance with the application and in order to prevent attrition,
a few measures shouldbe taken into account. First, overheating due
to friction should be avoided. Thus, long grinding times and/or high
rotational speeds should only be used in combination of pause pe-
riods. Second, depending on the sample amount an adequate
number of balls shouldbe placedin the grinding bowls. A larger
number of balls will reduce the grinding time, but also cause higher
abrasion, which should be avoided. In the 2 mmol scale experiments
reportedhere 60 balls were usedin order to mix andhomogenize
the reaction mixture.
Chem. Eur. J. 2007, 13, 4710 – 4722
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4721

C. Bolm et al.
[36] The temperature andpressure increase appearedto be higher when
higher rotational speeds were programmed. Unfortunately, both pa-
rameters couldnot precisely be measuredwhile performing the re-
action. When the bowls were opened, the crude reaction mixture ap-
pearedlike a thick slightly inhomogeneous paste.
[37] To some extend, this rate enhancement might be due to the elevated
temperature in the ball-milling experiment. In proline-catalyzed
Mannich reactions, a significant temperature tolerance has been ob-
served. B. Rodríguez, C. Bolm, J. Org. Chem. 2006, 71, 2888–2891.
[38] Only traces (< 5% each) of the corresponding cis-isomers were
formed.
[42] Since the existence of an intermediate melt is requisite for the reac-
tion to occur, these aldol reactions should not be called solid-state
reactions.
[43] In some cases this intermediate melt phase rapidly solidified, and
the appearance of the mixture was not completely homogeneous.
However, we assume that in the case of the liquid–solid systems the
melt intermediate phase (peritectic mixture) was also a requisite for
the reaction to occur.
[44] For aldol reactions using solid ketones 1d, 1e or 1 f as donors, see:
a) D. E. Ward, V. Jheengut, Tetrahedron Lett. 2004, 45, 8347–8350;
b) J.-R. Chen, X.-Y. Li, X.-N. Xing, W.-J. Xiao, J. Org. Chem. 2006,
71, 8198–8202; c) P. Dinꢁr, M. Amedjkouh, Org. Biomol. Chem.
2006, 4, 2091–2096.
[45] When the reaction was performedon a 10 g scale (40 mmol), about
150 balls were placedin the grinding bowl. The full vessel was ro-
tatedat 400 rpm (15 min milling + 5 min pause) over 6 h. An in-
creasedpressure was detectedwhen the vessel was opened. The
product was obtained quantitatively (yield determined by 1H NMR)
with an anti/syn ratio of 84:16 and92% ee (for the anti isomer).
After recrystallization in AcOEt/pentane, the anti-aldol product was
isolatedas white solidin 76% yield.
[39] For the 13C NMR spectroscopic data of cis- and trans-2-hydroxy-
methyl-4-tert-butylcyclohexanones, see: a) R. A. Wanat, D. B.
Collum, J. Am. Chem. Soc. 1985, 107, 2078–2082. Additionally, the
1H and 13C NMR spectroscopic data of the major products 3j–m cor-
respondwell with the data reportedfor other trans-4-tert-butylcyclo-
hexanones substitutedwith a hydroxymethyl-group at C2. For exam-
ples, see in ref. [12] andin b) D. E. Ward, W.-L. Lu, J. Am. Chem.
Soc. 1998, 120, 1098–1099.
[40] The signs of the specific rotations of the major products 3j–m corre-
spondwith those reportedin ref. [12].
[41] For a general concept article, see: G. Rothenberg, A. P. Downie,
C. L. Raston, J. L. Scott, J. Am. Chem. Soc. 2001, 123, 8701–8708.
Received: February 2, 2007
4722
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4710 – 4722