Stereoselective Wittig olefination reactions employing a novel ortho-P-aryl alkoxide effect

Article abstract of DOI:10.1016/j.tetlet.2008.09.150

Non-stabilized ortho-P-alkoxy-substituted ylides react with aromatic and aliphatic aldehydes providing (E)-olefins with high stereocontrol, also allowing easy phosphine oxide removal in certain cases.

Full text of DOI:10.1016/j.tetlet.2008.09.150

Tetrahedron Letters 49 (2008) 7054–7057
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Stereoselective Wittig olefination reactions employing
a novel ortho-P-aryl alkoxide effect
*
James McNulty , Kunal Keskar
Department of Chemistry, McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada L8S 4M1
a r t i c l e i n f o
a b s t r a c t
Article history:
Non-stabilized ortho-P-alkoxy-substituted ylides react with aromatic and aliphatic aldehydes providing
(E)-olefins with high stereocontrol, also allowing easy phosphine oxide removal in certain cases.
Ó 2008 Elsevier Ltd. All rights reserved.
Received 18 September 2008
Accepted 25 September 2008
Available online 1 October 2008
1. Introduction
qualitatively outlined in Scheme 1, is now believed to involve
essentially irreversible, rate-determining addition of the ylide to
Since its discovery in 1953, the Wittig reaction1 has emerged as
a powerful strategic carbon–carbon alkene bond forming process
of wide applicability, Scheme 1. This highly reliable reaction allows
for olefination with complete positional selectivity, relatively high
chemoselectivity, and may be conducted in many cases with reli-
able and high stereocontrol.2 The reaction has been the subject of
extensive experimental2 and theoretical3a investigations, and has
been comprehensively reviewed.2,3b The reaction mechanism,
the carbonyl component yielding either a cis- or a trans-substi-
tuted oxaphosphetane (OPA) intermediate. An interplay of 1,2-
and 1,3-steric and electronic effects governs the relative transition
states leading to either, or both of these intermediates. In general,
it is believed that cis- to trans-OPA equilibration or Wittig reversal
does not occur. Stereospecific syn-elimination of the phosphine
oxide then leads to the (Z)- or (E)-olefins, respectively. Despite
its advantages, the Wittig reaction still suffers from some
planar-TS
Puckered-TS
R'
PR3
O
R"
cis-OPA
R'
PR3
O
R"
trans-OPA
R'
R''
(Z)-olefin
R
P O
R
R
R'
+
P
R
R
R'
R
H
(E)-olefin
+
O
R''
R"
H
Scheme 1. General mechanism of the Wittig olefination reaction.
* Corresponding author. Tel.: +1 905 525 9140x27393; fax: +1 905 522 2509.
E-mail address: jmcnult@mcmater.ca (J. McNulty).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.09.150

J. McNulty, K. Keskar / Tetrahedron Letters 49 (2008) 7054–7057
7055
stereochemical limitations, for example, the stereoselective forma-
tion of aliphatic (E)-olefins from non-stabilized ylides and aromatic
or aliphatic aldehydes,2,4 as well as from the very practical issue of
phosphine oxide side-product removal.5 The selective formation of
(E)-olefins from non-stabilized ylides has only been achieved in
limited cases, for example, using the Schlosser modification,4a,b
which employs excess of an alkyl lithium base, or through the
use of ylides containing bulky alkyl groups attached to
phosphorus.4c
In this regard, the use of ortho-substituted aryl groups on phos-
phorus has been shown to modulate (E)/(Z)-selectivity also,
increasing the level of (E)-selectivity slightly with ortho-methoxy
groups,6a but increasing (Z)-selectivity in the case of ortho-
methoxymethyl-substituted aryl groups.6b The precise role of the
ortho-alkoxy groups in these cases is ambiguous at present.
ortho-Halo substituents have also been shown to provide higher
levels of (E)-stereoselectivity in limited cases.6c We were attracted
to a recent paper that disclosed a fascinating solvent-quench effect
on the stereoselectivity of the Wittig reaction, in which protic sol-
vents were shown to increase levels of (E)-stereoselection signifi-
cantly in the reaction with Garner’s aldehyde.7 While an
intramolecular stereochemical effect of alkoxy and other nucleo-
philic groups on the alkylidene portion on ylides has been investi-
gated, also resulting in higher levels of (E)-olefin,8 to our
knowledge, the role of related substituents on the aryl portion of
the ylide has not been investigated. Although the effect was limited
in scope, mechanistically it was attributed to cis-OPA protonation
and alkoxide-mediated isomerization to the trans-OPA. Overall,
the combination of the ortho-substituent effects6 and this intrigu-
ing intermolecular alkoxide effect7 prompted us to consider intra-
molecular ortho-aryl alkoxy effects as
a
possible means of
modulating (E)-stereocontrol, and that might also allow easy phos-
phine oxide removal. Herein, we report our preliminary findings
that support both of these hypotheses.
2. Results and discussion
The ethyl phosphonium salts of the three ortho-substituted
phosphines A–C were prepared employing standard techniques
outlined in Scheme 2, as well as of the unsubstituted control deriv-
ative D.
The ylides were generated from phosphonium salts A–D
employing 1.1 or 2.1 equiv of sec-BuLi under kinetically controlled
conditions at ꢀ78 °C in the presence of HMPA, followed by the
addition of a range of aldehydes as shown in Scheme 3.10 The over-
all results of this study are reported in Table 1.
To begin with, the control ylide derived from phosphonium salt
D reacted with aromatic aldehydes giving alkenes with good (Z)-
selectivity as expected and with high selectivity in the case of
the aliphatic aldehyde undecanal. The ylide derived from the
ortho-methoxy derivative A was less selective, but delivered more
(E)-olefin, similar to the literature data,6a with the exception of the
reaction with the undecanal which provided very high (Z)-selectiv-
ity. This result appears to indicate an overriding steric effect of the
long alkyl chain, possibly due to 1,2-interactions with the ylide
substituent, favoring the early cis-puckered transition state. More
revealing however are the results with the ylides derived from
the ortho-hydroxyl and benzyl alcohol-substituted derivatives B
and C. In both these cases, the ylides were generated under
kinetically controlled conditions using 2.1 equiv of sec-BuLi in
Preparation of Phosphonium Salts A and B
CH3
O
H3C
I
1.n-BuLi/
THF
Acetonitrile
P+
OMe
P
2.(Ph)2PCl
-78 ºC to RT
OMe
HI
C2H5I, Reflux
~6 hrs
1
3
2
Phosphonium salt A
H
O
H3C
P+
Acetonitrile
C2H5I, Reflux
~6 hrs
P
OH
5
4
Phosphonium salt B
Preparation of Phosphonium Salt C
Br
1.nBuLi/
THF
Br
Br
O
O
NaBH4
OH
OTHP
OMe
P
2.(Ph)2PCl
DCM, PTSA
THF/MeOH
66 ºC, 6 hrs
OTHP
8
-78 ºC to RT
7
6
9
OH
H3C
Acetonitrile
Methanol
P+
P
C2H5I, Reflux
~6 hrs
PTSA
2hrs RT
OH
11
10
Phosphonium salt C
Scheme 2. Synthesis of phosphonium salts A–C.

7056
J. McNulty, K. Keskar / Tetrahedron Letters 49 (2008) 7054–7057
CH3
CH3
CH3
CH3
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
P
P
P
P
+
+
+
+
HO
MeO
A
I
I
I
I
OH
B
C
D
CH3
1) THF, HMPA, sec-BuLi, -78 ºC
2) RCHO, -78 ºC to RT
A, B, C or D
R
Scheme 3. ortho-P-Substituted phosphonium salts employed in this study.
Table 1
Wittig reactions of aromatic and aliphatic aldehydes (Z):(E)a,b
H
H3C
H3C
R
H
R
O
O
Aldehyde
Ylide derivative
C
O
P
O
P
A
B
D
(Ph)2
(Ph)2
O
cis-OPA
from B
cis-OPA
from C
H
39:61
ND
12:88
8:92
71:29
Cl
1a
Figure 1. Oxaphosphetane intermediates.
O
O
H
23:77
54:46
31:69
95:5
8:92
63:37
80:20
67:33
90:10
1b
OPA through a 6- or 7-member transition state, and leading to
the (E)-olefins with high selectivity. It is not clear if the isomeriza-
tion involves initial ionization of the OPA to a betaine, but we do
not believe that such ionization is a requirement for the cis to trans
isomerization. Given the dipole moment and polarizability of the
O–P bond in the OPA, we believe that the acidity of the alkyl proton
MeO
MeO
H
10:90
8:92
17:83
12:88
38:62
1c
O
a
to phosphorus in conjunction with the intramolecular alkoxide is
O
O
H
sufficient to allow epimerization. Our present view is therefore the
direct epimerization of the OPA intermediates as depicted in Fig-
ure 1. These results support the alkoxide-mediated isomerization
pathway postulated by Kim and co-workers,7 which were limited
to the Garner aldehyde. The intramolecular variant described here
expands the scope to aromatic aldehydes, allowing for a significant
increase in (E)-stereoselectivity in their Wittig reaction with unsta-
bilized ylides.
1d
O
38:62
H
8
1e
(Z):(E) ratios were determined by 1H NMR.
Reported ratios are the averaged values of (Z):(E) after carrying out individual
a
b
Finally, in the case of the reactions with the phenoxide reagent
B, in addition to the high (E)-olefination observed, we were able to
remove the ortho-hydroxyl-triphenylphosphine oxide side product
employing a simple base extraction protocol.11 The reaction mix-
ture was quenched with saturated NH4Cl, and the olefin parti-
tioned into dichloromethane. The combined organic fractions in
dichloromethane were simply shaken with 10% NaOH (aq) which
completely removed the phosphine oxide, followed by normal
work-up procedures.
reaction for at least 3 or 4 times. NMR data for all the olefins are in accordance with
the literature data.12 There was no significant change in the (Z):(E) ratios of the
crude and purified material.
order to assure the formation of the phenoxide and alkoxide inter-
mediates. A pronounced effect was observed with both of these
ylides in their reactions with aromatic aldehydes, yielding olefins
with high (E)-selectivity, while a moderate increase in (E)-selectiv-
ity was observed with the aliphatic undecanal.
In comparing the results of the methoxy derivative A to the alk-
oxide containing derivatives B and C, the stereochemical results are
not simply due to lone pair interactions or steric factors alone,6a,b
the involvement of an alkoxide effect is evident. The present intra-
molecular alkoxide effect does not require initial protonation of the
OPA intermediate, and therefore does not appear to involve a b-hy-
droxy phosphonium intermediate.2c,7 Furthermore, the similarity
of the stereochemical effect in comparing the ylides derived from
B and C strongly suggests that the result is not manifested via con-
jugative or inductive dipolar effects,9 but it involves a direct
through space interaction of the alkoxy oxygen.
3. Conclusion
In conclusion, we have demonstrated an intramolecular alkox-
ide effect in stereocontrolled Wittig olefination reactions that is
consistent with ready cis to trans isomerization of OPA intermedi-
ates. This intramolecular probe extends the known intermolecular
effect7 to a wider range of aldehydes. In particular, aromatic alde-
hydes are shown to react with alkoxy-ylides leading to (E)-olefins
with high selectivity. Phosphine oxide removal is readily accom-
plished in the case of phenolic derivative B. Further extension
and application of the method are currently under investigation.
Figure 1 depicts the expected kinetically produced cis-OPA
intermediates derived from salts B and C. These species are now
set up to participate in an intramolecular variation of the alkoxy
effect demonstrated by Kim and co-workers7 through rapid iso-
merization of the cis-OPA intermediate, isomerizing to the trans-
Acknowledgments
We thank NSERC, Cytec Canada Inc. and McMaster University
for financial support of this work.

J. McNulty, K. Keskar / Tetrahedron Letters 49 (2008) 7054–7057
7057
consumption of the aldehyde. The reaction mixture was quenched with
saturated NH4Cl solution. The clear solution was extracted with (2 ꢁ 20 mL)
dichloromethane. The combined organic fractions were washed with brine and
dried over Na2SO4. Dichloromethane was removed on a rotary evaporator
leaving an oily residue which was purified by column chromatography (10:90
ethyl acetate/hexane) to give the alkene as an oil.
References and notes
1. Wittig, G.; Geissler, G. Liebigs Ann. Chem. 1953, 580, 44–57.
2. (a) Takeda, T. Modern Carbonyl Olefination; Wiley-VCH: Weinheim, 2004; (b)
Vedejs, E.; Peterson, M. J. Top. Stereochem. 1994, 21, 1–157; (c) Maryanoff, B. E.;
Reitz, A. B. Chem. Rev. 1989, 89, 863–927.
11. General protocol for removal of phosphine oxide: After complete consumption
of the corresponding aldehyde in the Wittig reactions employing phosphonium
salt ‘B’, the reaction mixture was quenched with saturated NH4Cl. The reaction
mixture containing the required olefin and the phosphine oxide by product
was partitioned between dichloromethane, and washed successively with
(2 ꢁ 10 mL) 10% NaOH solution to remove the phosphine oxide from the
reaction mixture.
3. (a) Robiette, R.; Richardson, J.; Aggarwal, V. K.; Harvey, J. N. J. Am. Chem. Soc.
2006, 128, 2394–2409; (b) Edmonds, M.; Abell, A. In Modern Carbonyl
Olefination; Takeda, T., Ed.; Wiley-VCH: Weinheim, 2004; pp 1–17.
4. (a) Schlosser, M.; Christmann, K. F. Angew. Chem., Int. Ed. Engl. 1966, 5, 126; (b)
Schlosser, M.; Christmann, K. F. Liebigs Ann. Chem. 1967, 708, 1–35; (c) Meyers,
A. I.; Lawson, J. P.; Carver, D. R. J. Org. Chem. 1981, 46, 3119–3123.
5. For discussion of phosphine oxide removal see: (a) Fukumoto, T.; Yamamoto, A.
US patent 5,292,973, 1994 (Niigata, JP); (b) Wang, Q.; Khoury, M. E.; Schlosser,
M. Chem. Eur. J. 2000, 6, 420–426; (c) Griffin, S.; Heath, L.; Wyatt, P. Tetrahedron
Lett. 1998, 39, 4405–4406; (d) Bootle-Wilbraham, A.; Head, S.; Longstaff, J.;
Wyatt, P. Tetrahedron Lett. 1999, 40, 5267–5270.
6. (a) McEwen, W. E.; Cooney, J. V. J. Org. Chem. 1983, 48, 983–987; (b) Zhang, X.;
Schlosser, M. Tetrahedron Lett. 1993, 34, 1925–1928; (c) Dunne, E. C.; Coyne, E.
J.; Crowley, P. B.; Gilheany, D. G. Tetrahedron Lett. 2002, 43, 2449–2453.
7. Oh, J. S.; Kim, B. H.; Kim, Y. G. Tetrahedron Lett. 2004, 45, 3925–3928.
8. Maryanoff, B. E.; Reitz, A. B.; Duhl-Emswiler, B. A. J. Am. Chem. Soc. 1985, 107,
217–226.
12. NMR data for olefins (derived from (1a–e): 1H NMR (200 MHz, CDCl3); d (ppm):
as a mixture of (Z)- and (E)-isomers,
1-(40-Chlorophenyl)prop-1-ene (from 1a):13 d 7.29 (2H, d, J = 8.74 Hz, ArH), 7.21
(2H, d, J = 8.52 Hz, ArH), 6.30–6.43 (1Ha, m, J = 16.37 Hz, CHa@CHbMe), 6.19
(1Hb, dq, J = 6.12 Hz, J = 15.76 Hz, E-isomer), 5.80 (1Hb, dq, J = 7.18 Hz,
J = 11.53 Hz, Z-isomer), 1.87 (3H, d, J = 7.2 Hz, Z-isomer and J = 6.12 Hz, E-
isomer).
1-(40-Methoxyphenyl)prop-1-ene (from 1b):14–18 d 7.26 (2H, d, J = 8.7 Hz, ArH),
6.83 (2H, d, J = 8.7 Hz, ArH), 6.34 (1Ha, m, J = 15.86 Hz, CHa@CHbMe), 6.08 (1Hb,
dq, J = 6.44 Hz, J = 15.66 Hz, E-isomer), 5.70 (1Hb, dq, J = 7.04 Hz, J = 11.37 Hz,
Z-isomer), 3.81 (3H, s, OMe, Z-isomer), 3.79 (3H, s, OMe, E-isomer), 1.89 (3H, d,
J = 7.04 Hz, Z-isomer), 1.85 (3H, d, J = 6.46 Hz, E-isomer).
9. Robiette, R.; Richardson, J.; Aggarwal, V. K.; Harvey, J. N. J. Am. Chem. Soc. 2005,
127, 13468–13469.
1-(30-Methoxyphenyl)prop-1-ene (from 1c):14–17 d 7.16–7.29 (1H, m, ArH), 6.83–
6.96 (2H, m, ArH), 6.71–6.80 (1H, m, ArH), 6.33–6.45 (1Ha, m, J = 16.73 Hz,
CHa@CHbMe), 6.22 (1Hb, dq, J = 6.1 Hz, J = 15.82 Hz, E-isomer), 5.79 (1Ha, dq,
J = 7.21 Hz, J = 11.60 Hz, Z-isomer), 3.81 (3H, s, OMe, Z-isomer), 3.80 (3H, s,
OMe, E-isomer), 1.90 (3H, d, J = 7.21 Hz, Z-isomer), 1.87 (3H, d, J = 6.1 Hz, E-
isomer).
10. (a) Typical procedure for the Wittig reaction with phosphonium salts A and D:
Phosphonium salt A or D (0.50 mmol), in a 25 mL flame-dried flask equipped
with magnetic stir bar, rubber septum and an argon inlet, was suspended in
anhydrous THF (2.0 mL), and the suspension was cooled to ꢀ78 °C. sec-BuLi
(0.55 mmol, 1.4 M in pentane) was added to the reaction mixture to yield a red
solution. After 15 min, HMPA (0.55 mmol) was added to the reaction mixture
at ꢀ78 °C. After 2 h, the corresponding aldehyde (0.50 mmol) was slowly
added to the flask at ꢀ78 °C. The reaction mixture was stirred for 2 h at ꢀ78 °C,
and then gradually warmed to room temperature and stirred for an additional
3 h. TLC analysis of the reaction mixture showed complete consumption of the
aldehyde. The reaction mixture was quenched with saturated NH4Cl solution.
The clear solution was extracted with (2 ꢁ 20 mL) dichloromethane. The
combined organic fractions were washed with brine and dried over Na2SO4.
Dichloromethane was removed on a rotary evaporator leaving an oily residue
which was purified by column chromatography (10:90 ethyl acetate/hexane)
to give the alkene as an oil.
1-(30,40-Dioxymethylenephenyl)prop-1-ene (from 1d):15,17,19 d 6.69–6.90 (3H, m,
ArH), 6.31 (1Ha, m, J = 15.75 Hz, CHa@CHbMe), 6.05 (1Hb, dq, J = 6.47 Hz,
J = 15.65 Hz, E-isomer), 5.69 (1Hb, dq, J = 7.12 Hz, J = 11.54 Hz, Z-isomer), 5.94
(2H, s, –OCH2O–, Z-isomer), 5.92 (2H, s, –OCH2O–, E-isomer), 1.87 (3H, d,
J = 7.12 Hz, Z-isomer), 1.84 (3H, d, J = 6.47 Hz, E-isomer).
Tridec-2-ene (from 1e):20 1H NMR (600 MHz, CDCl3); d 5.37–5.45 (2H, m), 1.94–
1.97 (2H, m, E-isomer), 2.00–2.05 (2H, m, Z-isomer), 1.63–1.64 (3H, d,
J = 4.2 Hz, E-isomer), 1.59–1.60 (3H, d, J = 6.6 Hz, Z-isomer), 1.24–1.33 (16H,
m), 0.86 (3H, t, J = 7.2 Hz).
13. Ikeda, Y.; Nakamura, T.; Yorimitsu, H.; Oshima, K. J. Am. Chem. Soc. 2002, 124,
6514–6515.
(b) Typical procedure for the Wittig reaction with phosphonium salts B and C:
14. Yu, J.; Gaunt, M. J.; Spencer, J. B. J. Org. Chem. 2002, 67, 4627–4629.
15. Joshi, B. P.; Sharma, A.; Sinha, A. K. Tetrahedron 2005, 61, 3075–3080.
16. Roberts, J. C.; Pincock, J. A. J. Org. Chem. 2006, 71, 1480–1492.
17. Buss, A. D.; Warren, S. J. Chem. Soc., Perkin. Trans. 1 1985, 2307–2325.
18. Goez, M.; Eckert, G. Helv. Chim. Acta 2006, 89, 2183–2199.
19. Mohottalage, S.; Tabacchi, R.; Guerin, P. Flavour Fragr. J. 2007, 22, 130–138.
20. Hodgson, D. M.; Fleming, M. J.; Stanway, S. J. J. Org. Chem. 2007, 72, 4763–
4773.
Phosphonium salt
B or C (0.50 mmol) was suspended in anhydrous THF
(2.00 mL) in a 25 mL flame-dried flask equipped with magnetic stir bar, rubber
septum, and an argon inlet, and the suspension was cooled to ꢀ78 °C. sec-BuLi
(1.05 mmol, 1.4 M in pentane) was added to the reaction mixture to yield a red
solution. After 15 min, HMPA (1.05 mmol) was added to the reaction mixture
at ꢀ78 °C. After 2 h, the corresponding aldehyde (0.50 mmol) was slowly
added to the flask at ꢀ78 °C and the reaction mixture was stirred for 2 h at
ꢀ78 °C, and then gradually warmed to room temperature and stirred for an
additional 3 h. TLC analysis of the reaction mixture showed complete