Catalytic aldol-transfer reactions with Al-alkoxide trapping

Article abstract of DOI:10.1016/j.tetlet.2006.02.044

Al-BINOL catalyzed aldol-transfer reactions of aldehydes with diacetonealcohol conducted with Al-alkoxide trapping gave aldol adducts in good yields. The best yields were obtained with electron-rich aromatic aldehydes (e.g., 83% yield with 3,4,5-trimethoxybenzaldehyde).

Full text of DOI:10.1016/j.tetlet.2006.02.044

Tetrahedron Letters 47 (2006) 2561–2564
Catalytic aldol-transfer reactions with Al-alkoxide trapping
Bixia Xi and Vesa Nevalainen*
Department of Chemistry and Biochemistry, University of Massachusetts Dartmouth, 285 Old Westport Road,
North Dartmouth, MA 02747, USA
Received 20 January 2006; revised 7 February 2006; accepted 8 February 2006
Available online 24 February 2006
Abstract—Al-BINOL catalyzed aldol-transfer reactions of aldehydes with diacetonealcohol conducted with Al-alkoxide trapping
gave aldol adducts in good yields. The best yields were obtained with electron-rich aromatic aldehydes (e.g., 83% yield with
3,4,5-trimethoxybenzaldehyde).
ꢀ 2006 Elsevier Ltd. All rights reserved.
Nucleophile-transfer reactions have been shown to offer
great potential for organic transformations.1–9 They are
analogous to hydride-transfer reactions found by pio-
neers Meerwein, Ponndorf, Verley, and Oppennauer.1
During recent years, several publications considering
new transfer reactions of nucleophiles other than
hydride have appeared. Inoue et al. have described the
use of Al- and Ti-based promotors (stoichiometric)2
and La-alkoxides (catalytic)3 for cyanohydrin transfer
reactions. Maruoka et al. have described an alkynyl-
transfer reaction mediated by an Al-o,o0-biphenyldioxy
species4 and Zr-alkoxides.5 Allyl and homopropargyl-
transfer reactions have been reported by Nokami et al.
and by Loh et al.6 (including an enantioselective allyl-
transfer by Loh et al.6d–6g). Aldol-transfer reactions cata-
lyzed by Al-BINOL (Scheme 1) were discovered by
Nevalainen et al.7 In this reaction aldol 1 is converted
through 2 and 3 to a diolmonoester 4 or aldol 5. Later
Schneider et al. reported8 the related Zr-BINOL cata-
lyzed process (including enantioselective8c formation of
analogs of 4). Chandrasekhar et al. described the first
L-proline catalyzed enantioselective aldol-transfer reac-
tion for aldols 5.9
Tandem reactions leading to 4 (Scheme 1) give better
yields than those of 5. This indicates that intermediate
3 is formed. However, the newly formed 5 (via
3+1!2+5) may undergo side reactions, which lower
the isolated yield of 5 (several by-products detected by
TLC). In tandem reactions intermediate 3 is ‘trapped’
by the conversion of 3 to stable 4. This is posing a
question: Can the newly formed 5 be trapped so that
it survives until the isolation step? For example, if
0.33 equiv extra trimethylaluminum is used, all 5 would
O
Al
Al
Al
HO
O
O
O
O
R
O
O
BINOL (0.05 eq)
Me3Al (0.1 eq)
O
RCHO
RCHO
R
R
2
1
3
4
(60 - 80%)
HO
O
by-products
+
O
O
Me3Al (0.33 eq)
"Alkoxide trap"
HO
O
side-reactions
Al
R
(40 - 70%)
5
R
5
3
6
Scheme 1. Catalytic aldol-transfer reactions of diacetonealcohol 1 with RCHO.7
*
Corresponding author. Tel.: +1 508 999 8276; fax: +1 508 999 9167; e-mail: vnevalainen@umassd.edu
0040-4039/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.02.044

2562
B. Xi, V. Nevalainen / Tetrahedron Letters 47 (2006) 2561–2564
OH
1)
2)
BINOL (10 mol%)
Me3Al (40 mol%)
OH
O
OH
O
O
O
+
+
OH
R
R
R
3)
4)
RCHO (100 mol%)
CH2Cl2, r.t., 43 h
HCl (1M)
1
7
8
5
Scheme 2. Aldol-transfer reactions conducted with alkoxide trapping.
Table 1. Aldol-transfer reactions of aldehydes RCHO with diacetonealcohol 1a
Entry
R
Derivative
Yield of 5 (%)b
Yield of 7 (%)b
Yield of 8 (%)b
1
2
3
4
5
6
7
8
9
C6H5
p-Cl–C6H4
a
b
c
d
e
f
29
18
38
63
84
83
20
74
30
40
10
33
—
—
—
—
—
—
—
—
3
5
p-NO2–C6H4
p-CH3O–C6H4
3-CH3O,4-C2H5O–C6H3
3,4,5-(CH3O)3C6H2
1-Naphthyl
Pyridine-2
Furan-2
Thiophene-2
12
—
—
—
—
—
—
—
g
h
i
10
j
a See Scheme 2. The products were characterized by spectral data.
b Isolated yields (relative to RCHO) after flash chromatography.
be converted to 6 (Scheme 1). Converting BINOL
(10 mol %) to its aluminum chelate would require
2/3 · 10 mol % = 0.06 equiv trimethylaluminum. There-
fore, the total amount of trimethylaluminum needed is
0.4 equiv or 40 mol %. In order to test this trapping
concept 10 (hetero)aromatic aldehydes were reacted
with 1 and 40 mol % trimethylaluminum (Scheme 2) to
obtain 5. In three cases by-products 7 and 8 were also
isolated. The results are summarized in Table 1.
version of this aldol-transfer reaction gave 5d only in
40% yield (30 mol % L-proline with 2 equiv of 1 in
DMSO, 120 h at rt).9 With di- and trialkoxybenzalde-
hydes the new method (Scheme 2) gave still significantly
better results. Aldols 5e and 5f were obtained in 84%
and 83% yields (Table 1, entry 5 and 6), respectively.
Furthermore, in the crude of 5d–f much less by-products
were detected than in the case of 5a–c. The new method
was also checked with 1-naththaldehyde and o-chloro-
benzaldehyde to see whether o-substitution would have
any effect. However, aldol 5g was isolated only in 20%
yield and the corresponding reaction of o-chlorobenz-
aldehyde gave a bad mixture not worth purification.
Yields of reactions of chloro-, nitro-, and alkoxy-substi-
tuted benzaldehydes (Table 1, entries 1–6) indicate that,
in contrast to that observed earlier,7 the parent benzal-
dehyde is now a worse substrate. For p-chlorobenzalde-
hyde the aldol-transfer step occurs modestly but the
yield is low because of water elimination (enone 7b iso-
lated in 33% yield, Table 1, entry 2). A small amount
(10%) of enone 7a was obtained also with benzaldehyde.
With p-nitrobenzaldehyde the yield of 5c was lowered
because a second aldol-transfer reaction took place con-
verting 3c to a novel compound 8c (Scheme 2) character-
When 2-heteroarylcarbaldehydes were used as sub-
strates there was not a clear trend in the results. For
example, the reaction of 2-pyridylcarbaldehyde with
diacetonealcohol 1 under the standard conditions
(Scheme 2) rendered the desired aldol 5h in 74% yield
whereas 5i and 5j were isolated only in 30% and 40%
yields. For 5h the method of isolation required modifica-
tions. By working in accord to the isolation method of
Marvel and Stille11 it was possible to isolate 5h in 41%
yield, but 12 extractions with CHCl3 were necessary.
The method of Chimni and Mahajan12 (extracting a
water solution of 5h three times with CH2Cl2 to obtain
5h in 98% yield) did not give satisfactory results either.
For the isolation of 5h the crude product was neutral-
ized with aqueous NaOH (10%), saturated with NaCl
and extracted several times with EtOAc. Another good
method was the following: the neutralized crude reac-
tion mixture is evaporated to dryness and the resulting
semisolid is washed with CHCl3. Filtration followed
by evaporation of the filtrate to dryness renders crude
5h (for purification by flash chromatography).
1
ized by H NMR, 13C NMR, MS (FAB), and HRMS
methods.10 From the crude of 5a and 5b derivatives 8a
and 8b were also found but in very low abundance
(3% and 5%). The formation of 8 was expected, because
ketones are known to undergo aldol-transfer reactions,
although in very poor yields. For example, acetophen-
one reacted with 1 (instead of RCHO, conditions as
depicted in Scheme 1), to give 2-hydroxy-2-phenyl-4-
pentanone in 6% yield.7a Furthermore, the purification
of 5a–c was hampered by the formation of 8a–c. Namely
the Rf values of 5a–c and 8a–c were very similar.
With electron-rich benzaldehydes the new concept
worked much better. For example, the p-methoxy deriva-
tive 5d was obtained in 63% yield whereas the other
method (Scheme 1, with 10 mol % BINOL, 22 h) gave7a
5d only in 47% yield. Interestingly, an organocatalytic
Aldol-transfer reaction of 2-thiophenecarbaldehyde
gave 5j in 40% yield. Aldol 5j was simultaneously pre-

B. Xi, V. Nevalainen / Tetrahedron Letters 47 (2006) 2561–2564
2563
pared also by Chimni and Mahajan12 in 30% yield using
an organocatalytic direct aldol method (reactions con-
ducted in water with 30 mol % pyrrolidine added as a
catalyst). As regarding side reactions Chimni and Maha-
jan report extensive water elimination to accompany the
formation of 5j.12 The yield of 5j (Table 1, entry 10) was
lowered because of extensive polymerization (a broad
hump with numerous sharp signals in it was observed
in the area of aromatic and enone protons in the
1H spectrum of the crude). Very similar conclusions
were drawn in the case of aldol 5i, which was obtained
in 30% yield only (Table 1, entry 9). Interestingly, Chim-
ni and Mahajan12 report a slightly better yield (44%) for
5i.
NMR (300 MHz, CDCl3, 20 ꢁC, CHCl3, 7.26 ppm): d
6.58 (s, 2H), 5.07 (d, 1H, J = 7.2 Hz), 3.87 (m, 6H),
3.83 (m, 3H), 3.26 (s, 1H), 2.83–2.87 (m, 2H) 2.21 (s,
3H); 13C NMR (75 MHz, CDCl3, 77.0 ppm): d 208.75,
153.17, 138.28, 137.21, 102.39, 69.83, 60.60, 55.92,
51.90, 30.54. These spectral values match well with the
literature data.13
1
All compounds were characterized using H NMR, 13C
NMR, and IR spectroscopy. The spectral data of known
5a–d, 5f–j, and 7a,b were well consistent with the litera-
ture values.7,12,13 New aldol 5e14 and the novel by-prod-
uct 8c10 gave satisfactory analyses.
References and notes
For comparison the 2-heteroaryl aldols 5h–j were syn-
thesized using the conventional base-catalyzed meth-
od:11 the aldehydes were reacted with acetone in
aqueous NaOH (10%) at ꢀ20 ꢁC for 40 min. The result-
ing mixture, which was first neutralized and evaporated
to dryness, was washed with CHCl3 and filtered. Evapo-
ration of the filtrate to dryness rendered aldol 5h in 72%
yield after recrystallization from methanol. Aldols 5i
and 5j were prepared accordingly in 20% and 24% yields
(after flash chromatography with EtOAc:hexane from
1:8 to 1:3), respectively. Interestingly, the yields were
better with the new method (Table 1).
1. (a) Meerwein, H.; Schmidt, R. Liebigs Ann. Chem. 1925,
444, 221; (b) Verley, A. Bull. Soc. Chim. Fr. 1925, 37, 537;
(c) Ponndorf, W. Angew. Chem. 1926, 39, 138.
2. Mori, A.; Kinoshita, K.; Osaka, M.; Inoue, S. Chem. Lett.
1990, 1171.
3. Mori, A.; Osaka, M.; Inoue, S. Chem. Lett. 1993, 375.
4. Ooi, T.; Miura, T.; Maruoka, K. J. Am. Chem. Soc. 1998,
120, 10790.
5. (a) Ooi, T.; Takaya, K.; Miura, T.; Maruoka, K. Synlett
2000, 69; (b) Ooi, T.; Miura, T.; Takaya, K.; Ichikawa, H.;
Maruoka, K. Tetrahedron 2001, 57, 867–873.
6. Allyl-transfer: (a) Nokami, J.; Yoshizane, K.; Matsuura,
H.; Sumida, S. J. Am. Chem. Soc. 1998, 120, 6609; (b)
Sumida, S.; Ohga, M.; Mitani, J.; Nokami, J. J. Am.
Chem. Soc. 2000, 122, 1310; (c) Lee, C. H. A.; Loh, T.-P.
Tetrahedron Lett. 2006, 47, 809–812; Asymmetric allyl-
transfer: (d) Lee, C.-H. A.; Loh, T.-P. Tetrahedron Lett.
2004, 45, 5819–5822; (e) Lee, C.-L. K.; Lee, C.-H. A.; Tan,
K.-T.; Loh, T.-P. Org. Lett. 2004, 6, 1281–1283; (f) Loh,
T.-P.; Lee, C.-L. K.; Tan, K.-T. Org. Lett. 2002, 4, 2985–
2987; (g) Loh, T.-P.; Hu, Q.-Y.; Chok, Y.-K.; Tan, K.-T.
Tetrahedron Lett. 2001, 42, 9277–9280; Homopropargyl-
transfer: (h) Lee, K.-C. K.; Lin, M.-J.; Loh, T. P. Chem.
Commun. 2004, 2456.
In conclusion, using the new alkoxide trapping method
electron-rich alkoxy-substituted benzaldehydes gave
the best yields of desired aldols 5. In contrast, elec-
tron-poor aldehydes gave low yields of 5 and the forma-
tion of either enones 7 or tetrahydropyran diols 8, or
both, was observed. Therefore, the purification of aldols
derived from electron-poor aldehydes was tedious.
When comparing the aldol-transfer (Scheme 2) and the
base-catalyzed method similar yields were obtained with
2-pyridin-carbaldehyde whereas for 2-furan and 2-thio-
phenecarbaldehyde the new methods worked better.
Finally, the new method gave the best yields for
Al-catalyzed aldol-transfer reactions published so far.
7. (a) Simpura, I.; Nevalainen, V. Angew. Chem., Int. Ed.
2000, 39, 3422; (b) Simpura, I.; Nevalainen, V. Tetra-
hedron Lett. 2001, 42, 3905; (c) Simpura, I.; Nevalainen, V.
Tetrahedron 2003, 59, 7435.
Typical procedure for the aldol-transfer reaction of
aromatic aldehydes is as follows: At room temperature
under argon, trimethylaluminum (0.64 mmol, 0.32 mL,
2 M in toluene or heptane) was added to a suspension
of 1,10-bi-2-naphthol (0.32 mmol, 0.93 mg) in dry
CH2Cl2 (1 mL) and stirred for 20 min. Then 3,4,5-tri-
methoxybenzaldehyde (3.33 mmol, 654.7 mg) and 4-
hydroxy-4-methyl-2-pentanone (3.4 mmol, 399 mg) were
added simultaneously. After stirring for 1 h trimethyl-
aluminum (0.64 mmol, 0.32 mL) was again added. After
further stirring for 42 h the mixture was quenched with
aqueous HCl (1 M, 10 mL). After adding 10 mL EtOAc
the mixture was stirred until homogeneous. The organic
layer was separated. The aqueous layer was extracted
with EtOAc (2 · 10 mL). The combined extracts were
washed with aqueous NaHCO3, dried over MgSO4,
and filtered. Evaporation of the filtrate gave 973.3 mg
crude product. Flash chromatographic purification of
320 mg of the crude gave 4-hydroxy-4-(30,40,50-trimeth-
oxyphenyl)-2-butanone (227.6 mg, 0.895 mmol, 83%)
as light yellow solid. Rf = 0.17 (EtOAc/hexane); 1H
8. (a) Schneider, C.; Weide, T. Chem. Eur. J. 2005, 11, 3010;
(b) Schneider, C.; Klapa, K.; Hansch, M. Synlett 2005, 91;
(c) Schneider, C.; Hansch, M. Synlett 2003, 837; (d)
Schneider, C.; Hansch, M. Chem. Commun. 2001, 1218.
9. Chandrasekhar, S.; Narsihmulu, C.; Ramakrishna Reddy,
N.; Shameen Sultana, S. Chem. Commun. 2004, 2450.
1
10. Characterization of 8c: Rf = 0.40 (EtOAc/hexane 1:2); H
NMR (300 MHz, CDCl3, 20 ꢁC, CHCl3, 7.26 ppm): d 8.19
(d, 2H, J = 8.7 Hz), 7.54 (d, 2H, J = 8.8 Hz), 5.26 (dd, 1H,
J = 2.2 Hz, J0 = 11.9 Hz), 4.27 (s, 1H), 3.23 (s, 1H), 1.96
(m, 2H), 1.65 (m, 2H), 1.53 (s, 3H), 1.30 (s, 3H); 13C NMR
(75 MHz, CDCl3, 77.0 ppm): d 149.67, 126.79, 126.47,
123.42, 97.40, 69.44, 67.58, 45.67, 44.60, 30.44, 29.49; MS
(m/z) BP: 268.1 (M+1), 249.1 (MꢀH2O), 231.1, 216.0,
192.0, 149.0, 98.1, 77.0, 58.1; HRMS (FAB) calcd for
C13H15NO4 [M+ꢀH2O]: 249.1001, found: 249.1061.
11. Marvel, C. S.; Stille, J. K. J. Org. Chem. 1957, 22, 1451.
12. Chimni, S. S.; Mahajan, D. Tetrahedron 2005, 61, 5019.
13. Dambacher, J.; Zhao, W.; El-Batta, A.; Anness, R.; Jiang,
J.; Bergdahl, M. Tetrahedron Lett. 2005, 46, 4473.
1
14. Characterization of 5e: Rf = 0.15 (EtOAc/hexane 1:2); H
NMR (300 MHz, CDCl3, 20 ꢁC, CHCl3, 7.26 ppm): d 6.92
(s, 1H), 6.83 (s, 2H), 5.09 (d, 1H, J = 8.0 Hz), 4.09 (q, 2H,

2564
B. Xi, V. Nevalainen / Tetrahedron Letters 47 (2006) 2561–2564
J = 6.97 Hz), 3.88 (s, 3H), 3.30 (s, 1H), 2.90 (dd, 1H, J =
147.73, 135.37, 117.75, 112.52, 109.09, 69.69, 64.32, 55.86,
51.99, 30.73, 14.74; MS (m/z) BP: 238.2 (M), 221.1, 181.1,
57.4; HRMS (FAB) calcd for C13H18O4 [M+]: 238.1205,
found: 238.1216.
9.05 Hz, J0 = 17.42 Hz), 2.79 (dd, 1H, J = 8.48 Hz,
J0 = 17.42 Hz), 2.20 (s, 3H), 1.45 (t, 3H, J = 2.8 Hz); 13C
NMR (75 MHz, CDCl3, 77.0 ppm): d 209.07, 149.34,