A mechanistic study of the Hiyama-Nozaki allylation: Evidence for radical intermediates

Article abstract of DOI:10.1016/j.tetlet.2004.09.179

A novel mechanism is suggested for the Nozaki-Hiyama allylation of aldehydes. The key intermediate is a chromium complex containing an allylic radical and the carbonyl component, from which the allylic radical may escape and undergo separate reactions. In course of chromium(II) mediated additions of relatively complex allylic bromides to aldehydes (Hiyama-Nozaki allylation) radical intermediates have been observed. From these findings a new mechanism of the allylation is derived.

Full text of DOI:10.1016/j.tetlet.2004.09.179

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 8867–8870
A mechanistic study of the Hiyama–Nozaki allylation: evidence
for radical intermediates
Johann Mulzer,* Achim R. Strecker and Lars Kattner
Institut fu¨r Organische Chemie der Universita¨t Wien, Wa¨hringerstrasse 38, A-1090 Wien, Austria
Received 5 August 2004; revised 28 September 2004; accepted 29 September 2004
Abstract—In course of chromium(II) mediated additions of relatively complex allylic bromides to aldehydes (Hiyama–Nozaki allyl-
ation) radical intermediates have been observed. From these findings a new mechanism of the allylation is derived.
Ó 2004 Elsevier Ltd. All rights reserved.
The Hiyama–Nozaki allylation1 is a well documented
THF
H
Ii
THF
THF
reaction of which numerous applications have been re-
ported in natural product synthesis,2 however the mech-
anism of this reaction has never been studied in greater
detail.
THF
Cl
X
Cr
R1
Cl
2
(X = Cl, Br, I,
1
OP(O)(OR)2, OTf)
In the early literature1c the formation of a covalent Cr–
allyl intermediate 3 from the solvated chromium(II) salt
(1) and the allylic halide (or phosphate or triflate) 2 was
postulated, which adds to the aldehyde 4 via a Zimmer-
man–Traxler like complex 5 from which the homoallylic
alcohols 6 and 7 are formed with varying syn-stereose-
lectivity (from ca. 70:30 up to >95:5; in the original lit-
erature the syn- anti-designation is opposite to the one
we use here. We do this because it is the easier way to
describe linear carbon chains with multiple stereogenic
centers). It was also claimed,1b that the 6/7-ratio is inde-
pendent of the E/Z-geometry of 2, which indicated that
species 3 rapidly isomerizes to the thermodynamically
favored E-configuration. As these mechanistic conclu-
sions, which are summarized in Scheme 1, were derived
from very simple substrates we decided to validate them
by using more complex allylic halides and aldehydes.
Thus, as shown in Scheme 2, halide 83 and aldehyde
93 under the usual conditions (THF, CrCl2 prepared in
situ from CrCl3 and LiAlH4)1 formed the diastereomeric
adducts 10 and 11 in a ratio of 2.5:1.
H
H
R2CH=O
(4)
THF
H
R1
Cr
III
THF
THF
R1
Cr
Cl
X
Cl
O
R2
H
3, rapid E/Z-
5
isomerization
OH
OH
R1
R1
R2
R2
+
6 (syn)
7(anti)
Scheme 1. Standard mechanism of the Hiyama–Nozaki allylation.
In keeping with our earlier results,3 the addition pro-
ceeded with complete syn-stereoselectivity around the
3,4- and with (moderate) Felkin–Anh selectivity around
the 4,5-axis. Additionally olefin 12 and di-olefin 13 were
isolated. The formation of 12 and, in particular, 13 can
be interpreted in terms of an allylic radical intermediate
14, which is quenched either by hydrogen abstraction or
by dimerization.4 Remarkably, the hydrogen abstrac-
tion did not produce regioisomer 12a, which would
come from the most stable radical position in 14. It
Keywords: Allylic radicals; Organochromium compounds; SET mech-
anism; Stereoselective allylation of aldehydes; Homoallylic alcohols.
*
Corresponding author. Tel.: +43 1427752190; fax: +43
1427752189; e-mail: johann.mulzer@univie.ac.at
0040-4039/$ - see front matter Ó 2004Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.09.179

8868
J. Mulzer et al. /Tetrahedron Letters 45 (2004) 8867–8870
Et
Me
O
O
OTBS
OBn
O
O
(16)
OBn
Br
9
O
Br
Me
O
Ph
Me
15
CrCl2, THF
65 %
CrCl2, THF
OBn
8
OBn
Me
OH
OBn
5
O
O
OBn
OH
Me
3
O
O
Me
+
Ph
4
Ph
O
O
OTBS
Me OH Et
O
Me OH Et
O
Me
OTBS
11 (12 %)
10 (28 %)
17
18
17 : 18 = 9 : 1
OBn
OBn
Me
Me
O
Ph OBn
O
O
16
Me
O
Me
O
Br
O
BnO
Me
19
CrCl2, THF
13 (6 %)
12 (29 %)
61 %
Ph OBn
OH
OBn
OBn
Ph OBn
OH
Me
Me
+
O
O
Me
OTBS
Me
12a
O
Me
O
Me
OTBS
14
20
21
20 : 21 = 10 : 1
Scheme 2. Dimerization and hydrogen abstraction in the Hiyama–
Nozaki allylation.
Scheme 3. Stereoselective Hiyama–Nozaki allylation with halides 15
and 19.
appears, however, that the radical transfer favors the least
hindered, that is terminal position of the allylic radical.
Workup with D2O did not lead to deuterium incorpora-
tion in 12 or 13, which excluded the presence of anionic
precursors. Halide 8 was inert toward chromium(II) in
the absence of aldehyde 9 and could be re-isolated. In
a second set of experiments (Scheme 3), (E)- and (Z)-ha-
lides 153 and 193 were first shown to behave ÔnormallyÕ
by giving the Hiyama–Nozaki adducts 17/18 and 20/21
in high diastereoselectivity. However, when 15 (or its
(Z)-isomer 19) was treated with CrCl2 and 1mol equiv
of acetone or benzophenone, no Hiyama–Nozaki allyl-
ation was observed. Instead the crystalline dimer 22 was
isolated5 whose structure was elucidated via single crys-
tal diffraction.12 The formation of 22 can be rationalized
(Scheme 4) in terms of a 5-exo-trig-cyclization of allylic
radical 23 to benzyl radical 24 and subsequent radical
dimerization to form the least hindered dimer possible
(the two phenyl groups on the bridge in 22 are anti-peri-
planar). In the absence of the ketone, no reaction oc-
curred and 15 remained untouched. Similarly, acetone
and benzophenone were inert toward chromium(II)chlo-
ride and benzophenone could be re-isolated. To probe
the stereochemical result of a ÔrealÕ free radical cycliza-
tion, halide 15 was subjected to a standard 5-hexenyl-
cyclization6 in presence of n-Bu3SnH as a hydrogen do-
nor. In fact, cyclopentane 25 was formed with the same
relative configuration as the one found in dimer 22,
according to extensive NOE experiments.
To probe the amount of negative charge that is accumu-
lated in the allylic moiety during the Hiyama–Nozaki
allylation, halide 26 was tested, which has an oxygen
leaving group in the d-position (Scheme 5). On treatment
with reducing metals such as Zn or Mg such halides un-
dergo elimination to form olefins.7 Under Hiyama–Noz-
aki conditions, however, 26 reacted with a variety of
aldehydes 27 to give the homoallylic alcohols 28/29 with
varying syn-selectivity.8 This result renders the formation
of an organometallic species such as 3 with carbanion-
like properties unlikely. Similarly, all attempts to quench
Cr–allyl intermediates with D2O did not lead to deute-
rium incorporation (vide supra).
In conclusion, we suggest an alternative mechanistic
pathway for the Hiyama–Nozaki allylation (Scheme
6), which in contrast to the ÔstandardÕ mechanism shown
in Scheme 1 avoids the formation of 3 and focuses on a
primary chromium–carbonyl interaction instead.
Hence, in the first step the reversible formation of a ke-
tyl radical 30 in low concentration is postulated. An-
other reversible SET to the allylic bromide places
bromide into the ligand sphere of the chromium and
generates complex 31, in which the carbonyl species
and an allylic radical are coordinated with the metal.
It remains a matter of speculation whether the allylic
species is coordinated to the chromium in a g1- or g3-
fashion.9 The regiochemistry of the ensuing Hiyama–

J. Mulzer et al. /Tetrahedron Letters 45 (2004) 8867–8870
8869
Nozaki—
Hiyama
allylation
Ph
THF
Me2C=O
H
THF
X
H
OBn
2
R2R3C=O
SET
THF
Cl
or Ph2C=O
Me
R2
THF
Cl
Cr O
R1
O
R2
6, 7
1
Cr
Cl
15
Me
Cl
SET
CrCl2, THF
40 %
R3
30
BnO
R3
H
Ph
31
22
H
increasing
R1
bulkiness of
R2 and/or R3
/ CrCl2
X
13
H
Ph
H
dimerization
X
Ph
5-exo-trig cyclization
+ dimerization
THF
R1
R2
R3
Cr
Cl
22
12
O
32
Cl
33
OBn
H-abstraction
H
Me
CrCl2
Ph
Ph
BnO
x 2
15
22
Scheme 6. Alternative mechanism for the Hiyama–Nozaki allylation.
Me
sociation were reversible, the allylic species in 31 would
be symmetrized and the corresponding Hiyama–Nozaki
adducts 6/7 would necessarily show allylic transposition.
From 30, a pinacol dimerization of the ketyl radical with
unreacted ketone could occur. However, such pinacol
dimerizations are slow in the absence of TMSCl.1g,10
24
23
OBn
Me
Bu3SnH, AIBN
94 %
15
Ph
Allylic radicals such as 32 are known to have a relatively
high barrier toward E/Z-isomerization,11 which would
be in contradiction to the behavior of intermediate 3
(Scheme 1). Thus, the E- and Z-allylic halides 343 and
373 were treated with benzaldehyde in a standard Hiy-
ama–Nozaki allylation (Scheme 7). Although the syn-
diastereomer 35 was formed selectively in both experi-
ments, the diastereomeric ratios of 35:36 were clearly
different. This observation suggests that the allylic inter-
mediates, which have been formed from 34 and 37 are
different and do not equilibrate.
25
Scheme 4. Radical cyclization and dimerization of allylic halide 15.
RCHO
(27)
R
R
Br
O
O
O
CrCl2, THF
O
OH
O
OH
O
26
28
29
OBn
PhCH=O
Br
O
O
Me
R
28 : 29
total yield (%)
CrCl2, THF
75 %
a Ph
93 : 7
92 : 8
60 : 40
60 : 40
86
84
91
88
b naphthyl
c Me
34
d C13H27
OBn
OBn
Ph
Ph
OH
Scheme 5. Hiyama–Nozaki allylation with d-alkoxy-allylic halides.
O
O
+
O
Me
OH
O
Me
Nozaki allylation, however, suggests it is g1, which
means that the geometry of 31 strongly resembles the
one of complex 5. The reversibility of the formation of
31 is necessary to explain the Cl/Br-exchange in the ally-
lic bromides 8, 15, and 19, which was observed as a side
reaction. For small substituents R2 and R3 31 may col-
lapse to generate the homoallylic alcohols 6/7 after
hydrolysis. For bulky residues R2 and R3 31 dissociates
into the Cr–carbonyl complex 33 and radical 32, which
either undergoes dimerization to 13 and 22, or forms
12 via hydrogen abstraction. This dissociation most
likely is irreversible, as it is driven by the release of steric
repulsion between the residues R1, R2, and R3. If the dis-
(96 : 4)
36
35
OBn
PhCH=O
O
O
Me
Br
35 + 36
CrCl2, THF
73 %
(85 : 15)
37
Scheme 7. Hiyama–Nozaki allylation of (E) and (Z)-allylic halides 34
and 37.

8870
J. Mulzer et al. /Tetrahedron Letters 45 (2004) 8867–8870
4. Okude, Y.; Hiyama, T.; Nozaki, H. Tetrahedron Lett.
1. Experimental
1977, 18, 3829; Sustmann, R.; Altevogt, R. Tetrahedron
Lett. 1981, 22, 5167.
5. Colorless crystals (mp 95°C). H NMR (CDCl3): d 0.66
A suspension of anhydrous chromium(III)chloride
(27mmol) in THF (100mL) was treated under vigorous
stirring at 0°C with lithium aluminumhydride
(13.5mmol) in small portions. After the evolution of
hydrogen has ceased the mixture was stirred for
additional 20min at ambient temperature. Then the
mixture was cooled to 0°C and the aldehyde (15mmol)
and the allylic bromide (10mmol) in THF (20mL each)
were added dropwise in succession and the mixture
was stirred at room temperature for 36h. Then the
reaction was quenched with saturated aqueous sodium
hydroxide (15mL) and solid sodium sulfate (20g) was
added. The mixture was filtered over Celite, dried over
sodium sulfate, concentrated under reduced pressure,
and purified by column chromatography (silicagel,
hexane–ethylacetate) or HPLC (0.8% 1-propanol in
hexane).
1
and 0.71 (d, 6H, J = 7.5Hz), 0.92 and 0.95 (ddd, 2H,
J = 13Hz, J = 7.5Hz, 2H), 1.41 and 1.42 (dddq, 2H,
J = 8Hz, J = 7.5Hz), 1.50 and 1.56 (ddd, 2H, J = 13Hz,
J = 9Hz, J = 8Hz), 1.87 and 2.01 (mc, 1H), 1.99 and 2.16
(mc, 2H), 3.33 and 3.40 (dd, 2H, J = 8Hz, J = 7Hz), 3.45
and 3.84(dd, 2H,
J = 13Hz, J = 5Hz), AB-system
(dA = 4.21, dB = 4.33, JAB = 12Hz, 2H), AB-system
(dA = 4.35, dB = 4.44, JAB = 12Hz, 2H), 4.90 and 5.00
(dd, 2H, J = 10Hz, J = 2Hz), 5.00 (2dd, 2H, J = 17Hz,
J = 2Hz), 5.40 and 5.50 (ddd, 2H, J = 17Hz, J = 10Hz,
J = 9Hz), 7.31 (mc, 10H). 13C NMR (CDCI3): d 15.17,
15.28, 34.26, 34.70, 37.46, 39.84, 42.34, 45.01, 45.79, 48.88,
50.26, 50.62, 71.04, 71.33, 82.88, 84.89, 112.75, 114.23,
125.89, 126.03, 126.13, 127.13, 127.27, 127.40, 127.52,
127.70, 127.83, 128.00, 128.10, 128.18, 128.25, 128.33,
128.61, 133.75, 142.46, 144.98. IR (KBr): m 3080m, 3065m,
3025s, 3020w, 2955vs, 2920vs, 2865vs, 2855vs, 1945w,
1898w, 1710m, 1635m, 1491m, 1450s, 1421w, 1400m,
1372m, 1350br m, 1295w, 1259m, 1211m, 1176m, 1153m,
1103vs, 1082s,1070s, 1028m, 999m, 906s, 821w, 788w,
753s, 703vs, 682w, 652mcmÀ1. MS (EI, 80eV, 90°C): m/e
610 (1%, M), 519 (3%, MÀC7H7), 502 (3%, MÀC7H7O),
411 (1%, MÀC7H7ÀC7H7O), 394(3%, M À2C7H7O), 305
(13%, M/2), 287 (5%), 231 (2%), 215 (3%, M/2ÀC7H7), 199
(47%, M/2ÀC7H7O), 143 (10%), 107 (22%, C7H7O), 91
(100%, C7H7). Anal. Calcd for C44H50O2: C, 86.51; H,
8.25. Found: C, 86.66, H, 8.28.
Acknowledgements
We thank Professor Dr. Christian Lehmann for per-
forming the crystal structure analysis of compound 22,
Dr. Jurgen Buschmann and Professor Dr. Peter Luger,
¨
FU Berlin for performing the one of 28a, and Dr.
Burkhard Kirste, FU Berlin for the NOE experiments
with 25.
6. For a review, see: Curran, D. P.; Porter, N. A.; Giese,
B. Stereochemistry of Radical Reactions; VCH: Wein-
heim, 1995, p 31ff; For related radical cyclizations with
Cr(II), see: Takai, K.; Nitta, K.; Fujimura, O.; Uti-
moto, K. J. Org. Chem. 1989, 54, 4732; Hackmann, C.;
Supplementary data
Supplementary data associated with this article can be
found, in the online version at doi:10.1016/j.tet-
let.2004.09.179. Some analytical and spectroscopic data
of compounds 12, 13, 17, 18, 20, 21, 25, 28a, 28b, 28c,
29c, 28d, 29d.
Scha¨fer, H. J. Tetrahedron 1993, 49, 4559; Lubbers, T.;
Scha¨fer, H. J. Synlett 1992, 743.
¨
7. Mulzer, J.; Greifenberg, S.; Beckstett, A.; Gottwald, M.
Liebigs Ann. Chem. 1992, 1131.
8. This result is contrary to the behavior of SmI2, which in
related systems induces eliminations due to the formation
of transient carbanions.2b The structure of 28a was
elucidated via single crystal diffraction.12
References and notes
9. Review on chromium-g3 complexes: Whiteley, M. N.
Abel, E. W.; Stone, F. G. A.; Wilkinson, G., Eds. in chief
In Comprehensive Organometallic Chemistry II; Labinger,
J. A., Winter, M. J., Eds.; Pergamon: Oxford, 1995; Vol. 5,
pp 335–346.
1. (a) Okude, Y.; Hirano, S.; Hiyama, T.; Nozaki, H. J. Am.
Chem. Soc. 1977, 99, 3179; (b) Hiyama, T.; Okude, Y.;
Kimura, K.; Nozaki, H. Bull. Chem. Soc. Jpn. 1982, 55,
561; (c) Buse, C. T.; Heathcock, C. H. Tetrahedron Lett.
1978, 19, 1685; (d) Hiyama, T.; Kimura, K.; Nozaki, H.
Tetrahedron Lett. 1981, 22, 1037; (e) Nowotny, S.; Tucker,
Ch. E.; Jubert, C.; Knochel, P. J. Org. Chem. 1995, 60,
10. Svatos, A.; Boland, W. Synlett 1998, 549.
11. Franklin, J. L.; Field, F. H. J. Am. Chem. Soc. 1953, 75,
2819; Szwarz, M.; Sheon, A. H. J. Chem. Phys. 1950, 18,
237; Homer, J. B.; Lossing, F. P. Can. J. Chem. 1966, 44,
2211; Berson, J. A.; Walsh, E. J. J. Am. Chem. Soc. 1968,
90, 4730; Field, R. J.; Abell, P. J. J. Am. Chem. Soc. 1969,
91, 7226; Golden, D. M.; Gac, N. A.; Benson, S. W. J.
Am. Chem. Soc. 1969, 91, 2136; Doering, W. v. E.;
Beasley, G. H. Tetrahedron 1973, 29, 2231; Roth, W. R.;
Ruf, G.; Ford, Ph. W. Chem. Ber. 1974, 107, 48; Egger, K.
W.; Golden, D. M.; Benson, S. W. J. Am. Chem. Soc.
1964, 86, 5420; Dewar, M. J. S.; Schmeising, H. N.
Tetrahedron 1960, 11, 96.
12. The crystal data can be obtained free of charge via www
from the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge, CB21EZ, UK, fax: +44 1223
336 033 or deposit@ccdc.cam.uk under the registration
numbers CCDC 241179 (22) and 241180 (28a).
2762, and earlier work; (f) Furstner, A.; Shi, N. J. Am.
¨
Chem. Soc. 1996, 118, 12349; (g) Bandini, M.; Cozzi, P.
G.; Melchiorre, P.; Umani-Ronchi, A. Angew. Chem., Int.
Ed. 1999, 38, 3357; (h) Wipf, P.; Lim, S. J. Chem. Soc.,
Chem. Commun. 1993, 1654.
2. (a) Reviews: Cintas, P. Synthesis 1992, 248; Wessjohann,
L. A.; Scheid, G. Synthesis 1999, 1; Furstner, A. Chem.
¨
Rev. 1999, 99, 991; Semmelhack, M. F.. In Organometal-
lics in Organic Synthesis, 2nd ed.; Schlosser, M., Ed.,
Wiley: Chichester, 2000; pp 1012–1021; (b) For related
work on SmI2-reactions, see: Curran, D. P.; Fevig, T. L.;
Jasperse, C. P.; Totleben, M. J. Synlett 1992, 943.
3. Mulzer, J.; Kattner, L.; Strecker, A. R.; Schro¨der, C.;
Buschmann, J.; Lehmann, C.; Luger, P. J. Am. Chem. Soc.
1991, 113, 4218.