Chromium(II)-Mediated Reformatsky Reactions of Carboxylic Esters with Aldehydes

Full text of DOI:10.1021/jo961910v

3772
J . Org. Chem. 1997, 62, 3772-3774
Ta ble 1. Rea ction of Sym m etr ica l
Ch r om iu m (II)-Med ia ted Refor m a tsk y
Rea ction s of Ca r boxylic Ester s w ith
Ald eh yd es
r-R1/2-r-br om oa ceta tes w ith Ch r om iu m Dich lor id ea /
Ca ta lytic Lith iu m Iod id e a n d Ald eh yd es (R4-CHO) in
THF (R1-4: cf. Sch em e 1)
entry
no.
T
(°C)
t
yield
(%)
R1 ) R2
R3
R4
(h)
Ludger Wessjohann* and Tobias Gabriel
1
2
3
4
5
6
7
H
H
H
Me
Me
Me
Me
Me
Et
Ph
Ph
Ph
Et
i-Pr
Ph
55
55
55
20
55
55
20
1.0
1.0
1.0
0.5
4.0
4.0
1.0
89b
72b,c
63b,c
81c,d
93
Institut fu¨r Organische Chemie,
Ludwig-Maximilians-Universita¨t Mu¨nchen, Karlstrasse 23,
D-80333 Mu¨nchen, Germany
t-Bu
Me
Me
Me
Me
Received October 9, 1996
90
98
Ph-CH2-
In tr od u ction
a
CrCl2, 99.9% Strem Chemicals Inc. if not stated otherwise.
CrCl2, ca. 90%, Merck-Schuchardt. c Reaction time and workup
are not optimized. In DMF.
b
The Reformatsky reaction,1,2 like other aldol-type reac-
tions, may be regarded as a two-step process with (ester)
enolate formation as the first step followed by the actual
aldol reaction with a keto component. The value of the
Reformatsky variation is threefold: The enolate can be
formed at an invariably predetermined site, under neu-
tral conditions and in the presence of many functional
groups.3 The position thus activated may differ from
those accessible by thermodynamically or kinetically
controlled base-induced enolizations.
d
because the related Nozaki-Hiyama reaction of allyl
compounds6,7 gives excellent chemo- and simple diaste-
reoselectivity toward aldehydes and anti products,
respectively.6,8-10 In addition only kinetic products would
be expected for reasons discussed elsewhere.11-13
Resu lts a n d Discu ssion
The major disadvantage of the classical procedure with
zinc was low reactivity, low reproducibility, and access
to thermodynamic products only. The more recent
introduction of highly activated metals4 led to a renais-
sance of the Reformatsky reaction but requires at least
one additional step for the preparation of the reagents.
Many problems, however, could not be addressed by these
modifcations. Microscale preparations are difficult to
perform with the often pyrophoric materials, and selec-
tivity is sacrificed for reactivity in many instances. Other
limitations originate from the still heterogeneous reaction
conditions, e.g. in applications with polymer supported
substrates. Problems associated with the nature of the
metal, usually zinc, include low stereo-, regio-, and
chemoselectivity. Thus with propionate and most alde-
hydes a <2:1 ratio in favor of the more common syn
isomer is achieved thermodynamically, whereas low-
temperature methods with activated metals give 1:1
mixtures kinetically.
The chromium dichloride mediated reaction of R-bromo
esters is run in a Barbier-type fashion.2 The reaction
proved to be highly reproducible with easy handling even
on a micromolar scale. Unexpectedly it appeared to be
somewhat slow at room temperature in most solvents in
comparison to the much more reactive ketones or vin-
ylogous esters.3,11-15 However, addition of lithium iodide,
especially in tetrahydrofuran, and slightly elevated tem-
peratures result in clean formation of the aldol products
in a few hours in good to excellent yields (Tables 1 and
2, Scheme 1). The effect of lithium iodide certainly
includes general lewis acid16 and nucleophilic iodide
catalysis: in some cases traces of R-iodo esters could be
detected as intermediates. The main influence, however,
seems to be the salts visible ability to solubilize and
modify chromium dichloride. Whether Li2[CrX4L2] or a
similar complex species is involved in this process is
speculation at this time.
We envisaged that a metal salt with a suitable reduc-
tion potential for selective reactivity should be a better
reagent. As such, it might be (partly) soluble, allowing
for homogeneous reaction conditions, which in addition
might be controllable by the ligands (solvent). Chro-
mium(II) lent itself toward this purpose. Apart from
excellent chemoselectivity in the enolate forming step,5
we expected further advantages in the C-C coupling,
(6) Hiyama, T.; Okude, Y.; Kimura, K.; Nozaki, H. Bull. Chem. Soc.
J pn. 1982, 55, 561.
(7) (a) Hoppe, D. In Houben-Weyl: Methods of Organic Chemistry.
Stereoselective Synthesis, 4th ed.; Helmchen, G., Ed.; Thieme: Stut-
tgart, 1995; Vol. E 21b, p 1584. (b) Hodgson, D. M. J . Organomet. Chem.
1994, 476, 1. Cintas, P. Synthesis 1992, 248. Saccomano, N. A. In
Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon Press: Oxford, 1991; Vol. 1, p 173.
(8) Nowotny, S.; Tucker, C. E.; J ubert, C.; Knochel, P. J . Org. Chem.
1995, 60, 2762.
(9) Maguire, R. J .; Mulzer, J .; Bats, J . W. Tetrahedron Lett. 1996,
37, 5487. Yamamoto, Y.; Asao, N. Chem. Rev. 1993, 93, 2207. J ubert,
C.; Nowotny, S.; Kornemann, D.; Antes, I.; Tucker, C. E.; Knochel, P.
J . Org. Chem. 1992, 57, 6384.
(10) Buse, C. T.; Heathcock, C. H. Tetrahedron Lett. 1978, 1685.
(11) Gabriel, T.; Wessjohann, L. In 11th International Conference
On Organic Synthesis (ICOS-11); Stichting Chemische Congressen V
(KNCV and IUPAC), Ed.; H. Hiemstra (Amsterdam Institute of
Molecular Studies): Amsterdam, 1996; p 317.
* To whom correspondence should be addressed. Phone: (Germany
+89) 5902-654. FAX: (+89) 5902-483. E-mail: law@org.chemie.uni-
muenchen.de.
(1) Rathke, M. W.; Weipert, P. In Comprehensive Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 2,
p 277. Heathcock, C. H. In Asymmetric Synthesis; Morrison, J . D., Ed.;
Academic Press: Orlando, 1984; Vol. 3, p 144. Fu¨rstner, A. Synthesis
1989, 571.
(2) Blomberg, C. The Barbier Reaction and Related One-Step
Processes; Springer: Berlin, 1993.
(3) Wessjohann, L.; Wild, H. Synlett, in press.
(12) Wessjohann, L.; Wild, H. Synthesis, in press.
(13) Gabriel, T.; Wessjohann, L. Tetrahedron Lett. 1997, 38, 1363.
(14) Dubois, J .-E.; Axiotis, G.; Bertounesque, E. Tetrahedron Lett.
1985, 26, 4371.
(4) Fu¨rstner, A. Angew. Chem. 1993, 105, 171. Cintas, P. Activated
Metals in Organic Synthesis; CRC Press: Boca Raton, 1993. Csuk, R.
Nachr. Chem., Tech. Lab. 1987, 35, 828. Csuk, R.; Fu¨rstner, A.;
Weidmann, H. J . Chem. Soc., Chem. Commun. 1986, 775. Boldrini,
G. P.; Savoia, D.; Tagliavini, E.; Trombini, C.; Umani-Ronchi, A. J .
Org. Chem. 1983, 48, 4108.
(5) In contrast to more powerful reducing agents like samarium-
(II), chromium(II) does not reduce carbonyl groups or simple halides
normally.
(15) Wessjohann, L.; Mu¨hlbauer, A. In 24. GDCh-Hauptversam-
mlung Hamburg; GDCh, Ed.; VCH: Weinheim, 1993; p 502.
(16) Chromium dichloride of >99.9% purity purchased from Strem
Chemicals was used to avoid catalytically active contaminations. Less
reactive esters sometimes give increased yields if either freshly
prepared CrCl2 × 2THF complex26 (which may contain residual
aluminum or lithium chloride) or a preformed mixture of 99.9% purity
CrCl2 with 10 mol % AlCl3 is used.
S0022-3263(96)01910-X CCC: $14.00 © 1997 American Chemical Society

Notes
J . Org. Chem., Vol. 62, No. 11, 1997 3773
Ta ble 2. Rea ction of Asym m etr ica l
r-R1/2-r-br om oa ceta tes a n d Ald eh yd es (R4-CHO) w ith
Ch r om iu m Dich lor id ea /Ca ta lytic Lith iu m Iod id e in THF
entry
no.
T
t
yield
R1
R2
Me
Me
Me
Me
Me
Me
Et
R3
Me
R4
(°C) (h) anti:synb (%)
8
9
H
H
H
H
H
H
H
H
H
H
i-Pr
i-Pr
i-Pr
Ph
Ph
Ph
Ph
Ph
Ph
55 3.0
20 1.0
55 1.0
20 2.0
20 2.0
55 2.0
55 3.0
55 2.0
55 2.0
71:29
77:23
60:40
70:30
78:22
60:40
74:26
72c,d,e
84
86
95
81
Me
Me
t-Bu
Me
Me
Me
Me
Me
10
11
12
13
14
15
16
17
18
19
F igu r e 1.
88
82f
30:70g 85c
time. This is exemplified by the successful reaction of
phenylacetaldehyde as electrophile (entry 7) which oth-
erwise dimerizes easily. In contrast to the literature on
allyl chromium reagents,19 we found aldiminessactivated
or notsand iminium salts to be unsuitable electrophiles.
Anti products are formed preferentially (Table 2),
which is in sharp contrast to the syn selectivity obtained
with ester enolates of other metal ions (Li, Zn, etc.)1,20 or
those reported for ketones.14 Following the assignments
made in the literature, increased bulk of the R-substitu-
ent leads to a reversal of stereochemistry in favor of syn
products (entry 15 and footnote g), again in contrast to
the anti preference induced by other counterions. Inter-
estingly R-bromo lactones (entry 17), which must form
the E-enolate, also give the anti (threo) product prefer-
entially.
i-Pr
Ph
79:21
84d
-CH2-CH2- Ph
20 2.0 >95:<5g 81f,c
20 1.0
PhCH(CH3) 20 3.0
Me Ph
Me Me
Me
Me
i-Pr
68:32g 86
76:24g 84e
a
CrCl2, 99.9%, Strem Chemicals Inc. if not stated otherwise.
b
Determined from crude product (Celite- or silica-filtered when
necessary). c CrCl2, ca. 90%, Merck-Schuchardt. X ) Cl. e The
d
reaction time is not optimized. f CrCl2 × 2THF.28 Assignment
g
uncertain, based on coupling constants and models.
Sch em e 1
These observations are in accordance with the reaction
of an E-enolate and a Zimmermann-Traxler transition
state model as presented in Figure 1. The related
crotylchromium reagents are well-known for their ability
to give anti products from both (E)- and (Z)-crotyl
halides.6,8 The extraordinary tendency of crotylchromium
intermediates to react exclusively as the E-isomer via a
similar transition state is widely accepted8,10,21 and thus
may be adapted to the heteroequivalent discussed here
as well. An alternative boat transition state should favor
(E)-enolates even more in order to avoid 1,4-interaction
of the axial chromium ligand and R1.22 The anti prefer-
ence, nevertheless, seems to be less strong quantitatively
in the Reformatsky -aldol. This may be caused by allylic
strain from the additional axial alkoxy substituent
(OR3)21 or by a heteroatom effect. Independent of enolate
stereochemistry, the small ionic radius of chromium(III)17
combined with an octahedral coordination and its en-
hanced steric influence through the superaxial ligand
should force the aldehyde into the proper relative orien-
tation with the R4 group equatorial. In this respect
chromium-centered transition states may best be com-
pared to those of titanium(IV).22
R-Chloro esters give similar results, but react more
slowly (cf. entry 8). With aromatic aldehydes, diol
coupling can become a competing side reaction with
chlorides. Nonactivated halogenides in the reactants or
in the solvent (e.g. methylene chloride) are not effected.
Two substituent effects can be observed: Larger ester
groups can give lower yields (R3: Me > Et > t-Bu, entries
1-3) or require longer reaction times, probably due to
increased hindrance around the sterically sensitive chro-
mium(III) ion.17 Ester groups even larger then t-Bu still
seem to allow the reaction with chromium(II) but not the
subsequent attack of the electrophile. Accordingly, in-
creased amounts of dehalogenated starting material from
proton quenching during workup are observed.18 In
contrast to these observations, increased substitution at
the R-carbon results in equal or even better yields in most
cases. Quarternary centers are generated with ease
(entries 4-7, 18, and 19).12 Sterically this contradicts
the first observation. The obvious argument of increased
stability of either an enolate or especially a radical
intermediate (as proposed for the Hiyama-Nozaki reac-
tion) is not necessarily conclusive considering the similar
yields observed with R,R-disubstituted acetates and phen-
ylacetates (entries 16 and 18).
Accordingly the diastereomeric excess in most cases is
slightly better with chromium(III) than with other metal
counterions in identical systems (Zn, Li, etc.),1,20 but is
still unsatisfactory. Anti:syn ratios are usually in the
range of 60:40 to 80:20, depending mainly on the reaction
temperature (cf. entries 9 + 10, 12 + 13). However,
future improvements can be expected. These may be
achieved either by lower temperatures paired with more
reactive chromium salts or rather by solvent effects
A significant influence on yields through the steric
effect of aldehydes was not observed (cf. entries 4-7).
However, in contrast to the Hiyama-Nozaki reaction,
aliphatic aldehydes often give slightly better results than
aromatic benzaldehyde which tends to form stronger
chromium complexes in its products. Transenolization,
a frequent problem in reactions under basic conditions
or in those prone to retro-aldolization, was never observed
despite temperatures of 55 °C and several hours reaction
(19) Giammaruco, M.; Taddai, M.; Ulivi, P. Tetrahedron Lett. 1993,
34, 3635.
(20) (a) Heathcock, C. H.; Buse, C. T.; Kleschick, W. A.; Pirrung,
M. C.; Sohn, J . E.; Lampe, J . J . Org. Chem. 1980, 45, 1066. (b)
Heathcock, C. H.; Pirrung, M. C.; Sohn, J . E. J . Org. Chem. 1979, 44,
4294. (c) Rieke, R. D.; Uhm, S. J . Synthesis 1975, 452.
(21) Mulzer, J .; Kattner, L.; Strecker, A. R.; Schro¨der, C.; Bus-
chmann, J .; Lehmann, C.; Luger, P. J . Am. Chem. Soc. 1991, 113, 4218.
(22) Cf. Nerz-Stormes, M.; Thornton, E. R. J . Org. Chem. 1991, 56,
2489.
(17) Kauffmann, T. In Organometallics in Organic Synthesis 2;
Werner, H., Erker, G., Eds.; Springer: Berlin, 1989; p 162. Kauffmann,
T.; Mo¨ller, T.; Wilde, H.-W. Chem. Ber. 1994, 127, 2277.
(18) This implies differences to a suggested model for the reactive
species in the Nozaki-Hiyama reaction: cf. Mulzer in ref 7a.

3774 J . Org. Chem., Vol. 62, No. 11, 1997
Notes
of chromium(III) residues, dried over MgSO4, filtered, and
concentrated in vacuo by rotary evaporation. In single cases
filtration through Celite or silica to remove residual chromium
complex is necessary. Diastereomers can be separated by
column chromatography on silica preferentially with petroleum
ether/ethyl acetate.
(ligand effects). Both approaches are under current
investigation, and the latter has already proven partly
successful for vinylogous systems.3,15
In the reaction of chromium Reformatsky reagents
with 2-phenylpropanal the simple diastereofacial selec-
tivity is quantitatively similar to that of most other
enolates23 giving a 3:1 ratio (at room temperature!, entry
19). Again the less usual â,γ anti product (anti-Cram)
seems to form preferentially, in contrast to reactions of
other enolates and allyl chromium reagents.21,23,24 Chemi-
cal yields with chromium(II) are much better than those
reported with zinc (84% vs 35%, respectively).25
In summary we presented a variation of the Refor-
matsky reaction which offers excellent reproducibility
even on a microscale, convenient handling without
activation, excellent chemo- and improved simple dia-
stereoselectivity, the latter being inverse to that of more
common approaches, i.e. anti selective.
Typ ica l P r oced u r es: Meth yl 2,2-Dim eth yl-3-h yd r oxy-4-
p h en ylbu tyr a te (en tr y 7). To a suspension of CrCl2 (251 mg,
2.05 mmol) and dry LiI (11 mg, 0.08 mmol) in dry THF (3.2 mL)
were added via syringe phenylacetaldehyde (86 µL, 0.738 mmol)
and methyl 2-bromoisobutyrate (102 µL, 0.820 mmol). The
resulting suspension was stirred for 60 min at room tempera-
ture. After usual workup the resulting residue was purified by
flash chromatography on silica with 4:1 petroleum ether:ethyl
acetate to afford 160 mg (98%) of butyrate (Rf ) 0.42): 1H NMR
3
(300 MHz, CDCl3/TMS) δ 1.27 (s, 3 H), 1.28 (s, 3 H), 2.24 (d, J
2
3
) 5.3 Hz, 1 H), 2.54 (dd, J ) 13.4 Hz, J ) 10.3 Hz, 1 H), 2.81
2
3
3
(dd, J ) 13.4 Hz, J ) 2.1 Hz, 1 H), 3.69 (s, 3 H), 3.92 (ddd, J
3
3
) 10.3 Hz, J ) 5.3 Hz, J ) 2.1 Hz, 1 H), 7.19-7.33 (m, 5 H);
13C NMR (75.5 MHz, CDCl3/TMS) δ 20.7, 21.8, 38.4, 47.2, 52.0,
126.4, 128.5, 129.3, 129.7, 139.1, 177.7.
Meth yl 2,4-Dim eth yl-3-h yd r oxyp en ta n oa tes (en tr y 18).
To a suspension of CrCl2 (843 mg, 6.86 mmol) and dry LiI (36
mg, 0.27 mmol) in dry THF (11 mL) were added via syringe
2-methylpropanal (274 µL, 3.02 mmol) and methyl 2-bromo-2-
phenylpropionate (660 mg, 2.74 mmol, as solution in dry THF).
The resulting suspension was stirred for 60 min at room
temperature. The usual workup afforded 640 mg of crude
product. A 322 mg portion was subjected to flash chromatog-
raphy with 10:1 hexanes:ethyl acetate to afford a total of 278
mg (86%) methyl 2,4-dimethyl-2-phenyl-3-hydroxypentanoates
separated in two diastereomers of 90 mg (18-I: 32 relative %, Rf
) 0.40) and 188 mg (18-II: 68 relative %, Rf ) 0.18), presumably
the syn and anti isomers, respectively.
Exp er im en ta l Section
Gen er a l. Anhydrous chromium dichloride (99.9%) was pur-
chased from Strem Chemicals Inc. or from Merck-Schuchardt
(90%) or prepared according to the literature.26 Aldehydes were
freshly distilled (Kugelrohr). THF was dried over K/benzophe-
none; dry DMF was purchased from Aldrich. Petroleum ether
(<70 °C) and ethyl acetate were distilled. For chromatography
silica 60 F254 (TLC) and 40-63 µm at ca. 1.3 bar (LC) from E.
Merck was used. Detection was achieved by iodine vapor followed
by ethanolic molybdato phosphate solution and/or UV fluores-
cence.
18-I (syn ?): 1H NMR (300 MHz, CDCl3/TMS) δ 0.77 (d, 3J )
NMR spectra were recorded in CDCl3/TMS on a Bruker ARX
300, mass spectra on a Finnigan MAT 90 and 95Q, and IR
spectra on a Perkin-Elmer 1420.
3
6.6 Hz, 3 H), 0.92 (d, J ) 6.6 Hz, 3H), 1.62 (s, 3 H), 1.73-1.80
3
(m, 1 H), 2.02 (d, J ) 4.3 Hz, 1 H, OH), 3.65 (s, 3 H), 4.06 (dd,
3J ) 4.9 Hz, 3J ) 4.3 Hz, 1 H), 7.25-7.49 (m, 5 H); 13C NMR
(75.5 MHz, CDCl3/TMS) δ 16.9, 18.5, 21.6, 30.4, 52.1, 54.8, 80.3,
Syn:anti ratios were determined by 1H-NMR from crude
products, filtered as described below if residual paramagnetic
chromium(III) had to be removed. Yields are of isolated
products. If diastereomers were not separated, their distribution
may slightly deviate from the ratio in crude material.
Gen er a l P r oced u r e. Approximately 2.5 equiv of anhydrous
chromium dichloride and 0.1 equiv of dry lithium iodide are
suspended in dry THF (ca. 1.5 mL/mmol of CrCl2) under an
argon atmosphere. If more dissolved chromium dichloride or a
faster reaction is required, THF may be substituted by DMF
(entry 4) or DMA. To the light gray-green suspension are added
via syringe 1.0 equiv of aldehyde and 1.1 equiv of R-halo ester
(the ratio maybe inversed with valuable R-halo esters). After
completion of the reaction, usually 1-6 h at room temperature
or up to 60 °C, it is quenched with brine and vigorously stirred
for 15 min. The organic layer is separated, and the aqueous
phase is extracted three times with ether. The combined organic
layers are washed with ion-exchanged water to remove traces
126.6, 126.8, 127.2, 127.3, 128.5, 140.4, 176.3; IR (film) ν [cm-1
]
3600-3300, 3030, 2940, 2860, 1740, 1595, 1575, 1490, 1460,
1439, 1425; MS (m/z, CI: i-BuH) 237 (0.4, [M + H]+), 219 (8.6
[237 - H2O]), 164 (12.3, [237 - C4H9O]+). Anal. Calcd for
C
14H20O3 (236.3): C, 71.2; H, 8.5. Found: C, 71.2; H, 8.4.
3
18-II (a n ti ?): 1H NMR (300 MHz, CDCl3/TMS) δ 0.69 (d, J
3
) 6.9 Hz, 3 H), 0.90 (d, J ) 6.9 Hz, 3 H), 1.64 (s, 3 H), 3.14 (d,
3J ) 6.0 Hz, 1 H, OH), 3.66 (s, 3 H), 4.17 (dd, J ) 6.0 Hz, J )
4.0 Hz, 1 H), 7.24-7.38 (m, 5 H); 13C NMR (75.5 MHz, CDCl3/
TMS) δ 19.3, 20.0, 25.0, 31.5, 54.8, 58.1, 82.3, 128.9, 129.5, 130.9,
143.6, 179.8; IR (film) ν [cm-1] 3600-3300, 3030, 2940, 2860,
1740, 1595, 1575, 1490, 1460, 1439, 1425; MS (m/z, CI: i-BuH)
237 (3.35, [M + H]+), 220 (7.36, [237 - OH]+), 219 (53.78, [237
- H2O]), 164 (16.94, [237 - C4H9O]+). Anal. Calcd for C14H20O3
(236.3): C, 71.2; H, 8.5. Found: C, 71.3; H, 8.3.
3
3
Ack n ow led gm en t. We would like to thank Prof. Dr.
W. Steglich, patent attorneys Maiwald & Partner, the
BASF AG, and the Fonds der chemischen Industrie.
Su p p or tin g In for m a tion Ava ila ble: Tables of 1H NMR,
13C NMR, IR, and MS data and elementary analyses (or HRMS
spectra) for new compounds; literature references for known
compounds; and selected additional spectral data (5 pages).
This material is contained in libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
(23) Yamamoto, Y.; Maruyama, K.; Matsumoto, K. Tetrahedron Lett.
1984, 25, 1075.
(24) The assignment is based on NMR data and the model of ref
20b. In the more complex reaction between 2-phenylpropanal and
methyl 2-bromopropionate (no. 20)10 the product ratio is approximately
4:2:1 (79% yield). Although this is an improvement to existing
Reformatsky approaches, it appears not to be synthetically useful at
this stage. Cf.: Matsumoto, T.; Hosoda, Y.; Mori, K.; Fukui, K. Bull.
Chem. Soc. J pn. 1972, 45, 3156.
(25) Overberger, C. G.; Bonsignore, P. V. J . Am. Chem. Soc. 1958,
80, 5427.
(26) Handl´ır, K.; Holocek, J .; Klikorka, J . Z. Chem. 1979, 19, 265.
Kern, R. J . J . Inorg. Nucl. Chem. 1962, 24, 1105.
J O961910V