Reactions of carbonyl compounds with phosphorus ylide generated from tribromofluoromethane and tris(dimethylamino)phosphine

Article abstract of DOI:10.1246/cl.150523

The reactions of fluorinated ylide, generated from tris(dimethylamino) phosphine and tribromofluoromethane, with simple aldehydes and reactive ketones gave the expected Wittig reaction products. However, a ketone having a galactose skeleton afforded an acid fluoride, probably through an unprecedented CoreyChaykovski- type epoxide formation reaction, followed by spontaneous Meinwald rearrangement.

Full text of DOI:10.1246/cl.150523

Received: May 30, 2015 | Accepted: July 9, 2015 | Web Released: July 18, 2015
CL-150523
Reactions of Carbonyl Compounds with Phosphorus Ylide Generated
from Tribromofluoromethane and Tris(dimethylamino)phosphine
Go Hirai,*1,2,3 Eri Nishizawa,1,4 Daiki Kakumoto,1,4,5 Masaki Morita,1,3 Mitsuaki Okada,1,4
Daisuke Hashizume,1 Sayoko Nagashima,5 and Mikiko Sodeoka*1,2,3,4,5
1RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198
2RIKEN Center for Sustainable Resource Science, 2-1 Hirosawa, Wako, Saitama 351-0198
3AMED-CREST, 2-1 Hirosawa, Wako, Saitama 351-0198
4Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510
5Graduate School of Science and Engineering, Saitama University, 255 Shimo-okubo, Sakura-ku, Saitama 338-8570
(E-mail: gohirai@riken.jp; sodeoka@riken.jp)
The reactions of fluorinated ylide, generated from tris(di-
methylamino)phosphine and tribromofluoromethane, with sim-
ple aldehydes and reactive ketones gave the expected Wittig
reaction products. However, a ketone having a galactose
skeleton afforded an acid fluoride, probably through an
unprecedented Corey-Chaykovski-type epoxide formation re-
action, followed by spontaneous Meinwald rearrangement.
A) Reported bromofluoromethylenation of carbonyl compounds
O
heat,
Zn,
F
Br
R2
F
R1
R2
Ph3P + CFBr3
Ph3P
Et2Zn,
or (RO)3P
R1
Br
Putative Ylide 1
B) Difluoromethylenation and chlorofluoromethylenation using HMPT
O
F
X
(Me2N)3P (2 equiv)
R1
R2
F
Fluorinated ylides are widely used for fluoro-, difluoro-,
chlorofluoro-, and bromofluoromethylenations of carbonyl
compounds in the fields of synthetic organic chemistry and
medicinal chemistry.1 However, relatively few methods are
available for bromofluoromethylenation reactions, and the
triphenylphosphonium ylide 1 has been mainly utilized. It is
usually prepared in situ, without isolation, from tribromofluoro-
methane (CFBr3) and triphenylphosphine (PPh3) under heating2
or with the use of zinc metal,1,3,4 diethylzinc,5 or phosphite6 as
an additive (Figure 1A).1,7
+
(Me2N)3P
R1
R2
CF2Br2 (X = F)
or
CFCl3 (X = Cl)
X
Putative Ylide
2a (X = F); 2b (X = Cl)
Ph
Ph
CF2Br2
(Me2N)3P
O
O
O
O
MP = 4-methoxyphenyl
OMP
O
O
THF
OMP
F
O
OTIPS
OTIPS
76%
F
3
4
C) Reaction of 2c with carbonyl compounds (This Work)
On the other hand, tris(dimethylamino)phosphine (HMPT)
has been used for difluoromethylenation1,8 and chlorofluorome-
thylenation9 reactions of carbonyl compounds (Figure 1B).
Mixing of HMPT (2 mol) with CF2Br2 or CFCl3 (1 mol) would
afford the corresponding ylides 2a or 2b, which should react
with reactive carbonyl compounds to give the corresponding
fluoro-olefins. Indeed, we employed the HMPT-CF2Br2 system
with the relatively reactive ketone 3 to obtain difluoro-olefin 4
(Figure 1B).10 These methods are convenient owing to their
simple procedures, but, as far as we know, no example of
bromofluoromethylenation reaction using HMPT has yet been
reported.11 Herein we report unique reactivity of the putative
ylide 2c (Figure 1C).
We first investigated the reaction of carbonyl compounds
with 2c. Most of the aldehydes examined were converted to the
corresponding bromofluoro-olefin, and the scope of this reaction
is summarized in Table 1. The reaction was conducted with
CFBr3 (X equiv) and HMPT (2X equiv) at room temperature in
THF (0.1 M concentration of the substrate), unless otherwise
noted.12 The optimum amounts of reagents were different for
each substrate. For example, use of four equivalents of CFBr3
(X = 4) gave the best result for 6a (yield of 7a: 92%), whereas in
the case of 6e, a smaller amount of reagents (X = 3) gave a
better result (yield of 7e: 72%). In all cases, the product 7 was a
mixture of E- and Z-isomers, in accordance with previous
reports on the reactions using PPh3.5 Several electron-donating
substituents on the aromatic ring were tolerated, but electron-
O
F
Br
R1
R2
Ph
R1
R2
(Me2N)3P (2 equiv)
F
+
(Me2N)3P
CFBr3
Br
O
O
3
Putative Ylide 2c
F
O
OMP
OTIPS
O
Br
5
Figure 1. A) Bromofluoromethylenation of carbonyl compounds
with putative ylide 1; B) Fluoromethylenation of carbonyl compounds
with putative ylide 2a or 2b and its application to ketone 3 (ref 10);
C) Two different reactions of putative ylide 2c generated from HMPT
and CFBr3 with carbonyl compounds.
withdrawing nitro or fluoro substituents decreased the chemical
yield of 7. The reaction of benzaldehydes with -NHAc or
-NHBoc groups gave messy mixtures and no bromofluoro-olefin
was isolated. Aliphatic aldehydes 6p and 6q were converted to
7p and 7q in moderate yields.
Although ketones such as acetophenones or cyclohexanones
showed poor reactivity with 2c (Table S1),12 a few ketones such
as 6r or the α-ketoester 6s were converted to the corresponding
bromofluoro-olefin in acceptable yield. Thus, 2c appeared to
be a moderate nucleophile.13 However, ketones that were not
Chem. Lett. 2015, 44, 1389–1391 | doi:10.1246/cl.150523
© 2015 The Chemical Society of Japan | 1389

Table 1. Bromofluoromethylenation of various aldehydes and ke-
tones 6
Table 2. Formation of acid fluoride 5
Ph
(Me2N)3P (2X equiv)
CFBr3 (X equiv)
Crystal structure
of 8
F
Br
O
CFBr3
(Me2N)3P
O
O
R1
R2
3
R1
R2
THF (0.1 M), rt
X
O
R2 = H
6
7
THF, 24 h
OMP
OTIPS
O
R1
X
Time
Yield/%
E:Z ratiob
Br
p-MeOC6H4 (6a)
m-MeOC6H4 (6b)
o-MeOC6H4 (6c)
p-MeC6H4 (6d)
o-MeC6H4 (6e)
p-BrC6H4 (6f)
o-BrC6H4 (6g)
p-FC6H4 (6h)
o-FC6H4 (6i)
p-O2NC6H4 (6j)
o-O2NC6H4 (6k)
p-HC¸CC6H4 (6l)
o-HC¸CC6H4 (6m)
p-MeO2CC6H4 (6n)
4
4
4
4
3
4
3
3
3
2
2
2
4
3
22 h
2 h
1 h
1 h
24 h
40 min
2 h
30 min
1.5 h
30 min
15 min
4 h
92a
81
84
66
72
58
75
44
44
46
81
62
71
63
1.2:1
1:1.2
1:1.3
1:1.0
1:1.0
1:1.3
1:1.1
1:1.2
1:1.3
1:1.7
1:1.0
1:1.0
1:1.2
1:1.4
5
(X = F)
NaOMe
MeOH, 45%
Br
8
(X = OMe)
CFBr3
/equiv
(Me2N)3P
/equiv
Temp.
/°C
Yield of
Recoverd
3/%
Entry
5a/%
1b
2
3
4
5
1.5
3.0
3.0
3.0
3.0
3.0
3.0
2.0
6.0
5.0
4.0
3.1
3.1
3.1
rt
rt
rt
rt
rt
27
0
0
31
24
32
69
51
NDc
33
32
45
6
7d
0
¹40
21
NDc
1.5 h
20 min
aTrace amounts of impurities were detected. bReaction time: 1.5 h.
d
cND: Not detected. Reaction time: 13 h.
a6% of 6a was recovered. bThe E/Z ratio was determined by 1H NMR
of crude material. cThe carbonyl group indicated by bold face in 6s and
6t was converted to the corresponding bromofluoro-olefin.
2c
O
3
5
O
O
O
O
O
H
R3P
PR3
F
Br
PR3
F
Br
F
Br
10
H
H
R = NMe2
O
11
9
O
6o
6p
6q
F
X = 4, 1 h, 78%
E:Z = 1:1.0
X = 3, 14 h, 68%
E:Z = 1:1.9
X = 4, 1 h, 50%
E:Z = 1:1.8
7
Br
(19% recovery)
O
O
O
Scheme 1. Plausible mechanism for the formation of acid fluoride 5.
H
Ot-Bu
O
analysis revealed that, surprisingly, it was an α-bromoacid
fluoride derivative 5. The acid fluoride functionality was
confirmed by 13C NMR (¤ 156.5 ppm, J = 365 Hz) and
19F NMR (¤ 24.8 ppm), while mass spectrometry showed the
presence of a bromine atom. Treatment of 5 with sodium
methoxide gave ester 8, whose structure was confirmed by X-ray
crystallographic analysis (Table 2).14
6tc
6r
6sc
O
X = 4, 0.5 h, 44%
E:Z = 1:1.3
X = 3, -40 °C, 24 h, 56% X = 2, 45 min, 63%
E:Z = 1:1.3
(30% recovery)
E:Z = 1:1.4
converted to 7, were not completely inert, and were partially
degraded (Table S1).12 In fact, chemoselective olefination of
aldehyde in 6t in the presence of a ketone functionality was
realized, though in only moderate yield without recovery of 6t.
These results suggested that 2c might react with ketones, but the
formation of bromofluoro-olefin was unfavored. We anticipated
that other reaction pathways might exist after addition of 2c to
a carbonyl group. During the course of these investigations, we
found a very interesting reaction process.
We examined bromofluoromethylenation of ketone 3, which
is quite electrophilic due to the ring-strain of the sugar skeleton
(Figure 1C). We assumed that bromofluoromethylenation would
proceed, because a similar difluoromethylenation reaction was
successful.10 However, treatment of ketone 3 with CFBr3
(1.5 equiv) and HMPT (3 equiv) gave a complex mixture.
Careful TLC analysis revealed a new spot on TLC upon partial
addition of HMPT, but this later disappeared. Hence, the amount
of HMPT was reduced to 2 equiv (CFBr3: 1.5 equiv, Table 2,
Entry 1). As a result, the new product could be isolated, but
it was not the corresponding bromofluoro-olefin. Structural
Formation of the α-bromoacid fluoride 5 was investigated
with an increased amount of CFBr3 (3 equiv, Table 2). It was
found that one of the key factors determining the chemical yield
of 5 was the amount of HMPT. When 6 equiv of HMPT (2 equiv
with respect to CFBr3) was used, a messy mixture was obtained
again, probably due to decomposition of 5 (Entry 2). In contrast,
when the amount of HMPT was decreased from 6 to 3.1 equiv,
the yield of 5 increased (Entries 2-5). 5 was obtained when less
than 5 equiv of HMPT was used (Entry 4), and 3.1 equiv of
HMPT was enough to produce 5 (Entry 5). Another key factor
was the reaction temperature. Lowering the reaction temperature
improved the conversion of 3 and the chemical yield of 5, and
eventually 5 was obtained in 69% yield at ¹40 °C (Entry 7).
A possible reaction mechanism leading to 5 is shown in
Scheme 1. We speculated that α-bromoacid fluoride function-
ality could be formed from α-bromo-α-fluoroepoxide 11 via
tautomerization, corresponding to a spontaneous Meinwald-
type rearrangement. A similar transformation was reported by
Saloutin and co-workers.15 Although α-bromo-α-fluoroepoxide
1390 | Chem. Lett. 2015, 44, 1389–1391 | doi:10.1246/cl.150523
© 2015 The Chemical Society of Japan

with the perfluoroalkyl group has been isolated,16 11 lacking the
perfluoroalkyl group would be destabilized and spontaneously
converted into the α-bromoacid fluoride. Electron-donating alkyl
groups and the negative hyperconjugation effect of a fluorine
atom might enhance the rearrangement. If this is the case,
Corey-Chaykovsky-type epoxidation (CC-type reaction) should
be implicated in the formation of 11 from 3 with 2c. The
observed experimental fact that a slight excess of HMPT with
respect to CFBr3 was sufficient for the formation of 5 supports
the proposed mechanism.17 In the mechanism shown in
Scheme 1, HMPT would be regenerated after CC-type reaction,
while two HMPT molecules are required to generate 2c. CC-
type reaction is well-known in sulfur ylide-mediated epoxida-
tion.18 Based on the proposed mechanism of the sulfur ylide-
mediated CC-type reaction, an intramolecular SN2 reaction of
betaine 10, which might be in equilibrium with the usual
oxaphosphetane intermediate 9 in Wittig reaction, would give
the labile 11. Although selenium,19 tellurium,20 arsine,21 and
bismuth22 ylide-mediated CC-type reactions have been already
reported, as far as we know, this is the first example of CC-type
reaction of a phosphorus ylide.23
On the other hand, application to simple aldehydes, ketones,
or the corresponding glucose and mannose derivative of 3 did
not result in the isolation of α-bromoacid fluorides or their
hydrolyzed products (Scheme S2).12 It is possible that this new
process occurred with simple substrates because the α-bromo-
acid fluorides or epoxide intermediates would be readily
decomposed under the reaction conditions. In contrast, com-
pound 5 is sufficiently stable to be isolated, and may be
kinetically stabilized due to the neighboring bulky TIPS group.
In conclusion, we found two different reactions of putative
ylide 2c, although the structural requirements for each process
remain unclear. At least for more reactive carbonyl compounds,
such as aldehydes or α-ketoesters, the reaction with 2c provided
the corresponding bromofluoro-olefin derivatives, which should
be precursors of fluoro-olefin derivatives. For the specific
substrate 3, another reaction process from carbonyl compound
into α-bromoacid fluoride derivative via a novel phosphorus
ylide-mediate CC-type reaction, followed by spontaneous
Meinwald rearrangement reaction, proceeded in a similar
reaction system. Isolation of 5 suggested that this unique
multi-sequence reaction might be more generally available as
a highly functionalized one-carbon elongation reaction. Further
investigations are in progress to examine this possibility.
Zemmouri, M. Kajjout, Y. Castanet, S. Eddarir, C. Rolando, J. Org.
Chem. 2011, 76, 7691.
4
5
Bromofluoromethylenation reactions appeared to proceed for sev-
eral aldehydes even at room temperature. D. Alloatti, G. Giannini,
W. Cabri, I. Lustrati, M. Marzi, A. Ciacci, G. Gallo, M. O. Tinti, M.
Marcellini, T. Riccioni, M. B. Guglielmi, P. Carminati, C. Pisano,
J. Med. Chem. 2008, 51, 2708.
a) X. Lei, G. Dutheuil, X. Pannecoucke, J.-C. Quirion, Org. Lett.
2004, 6, 2101. b) H. Zhang, Y. Xu, Z. Zhang, E. R. Liman, G. D.
Prestwich, J. Am. Chem. Soc. 2006, 128, 5642. c) L. Zoute, G.
Dutheuil, J.-C. Quirion, P. Jubault, X. Pannecoucke, Synthesis 2006,
3409. d) G. Dutheuil, C. Pierry, E. Villiers, S. Couve-Bonnaire, X.
Pannecoucke, New J. Chem. 2013, 37, 1320.
R. M. Flynn, D. J. Burton, D. M. Wiemers, J. Fluorine Chem. 2008,
129, 583.
A stereoselective method to synthesize 1-bromo-1-fluorostyrenes
from benzaldehydes via hydrazones has been reported: A. V.
Shastin, V. M. Muzalevsky, E. S. Balenkova, V. G. Nenajdenko,
Mendeleev Commun. 2006, 16, 179.
a) D. G. Naae, D. J. Burton, Synth. Commun. 1973, 3, 197.
b) J. M. J. Tronchet, A. Bonenfant, F. Barbalat-Rey, Carbohydr. Res.
1978, 67, 564. c) W. A. Vinson, K. S. Prickett, B. Spahic, P. R. O.
de Montellano, J. Org. Chem. 1983, 48, 4661. d) W. B. Motherwell,
M. J. Tozer, B. C. Ross, J. Chem. Soc., Chem. Commun. 1989, 1437.
e) J. S. Houlton, W. B. Motherwell, B. C. Ross, M. J. Tozer, D. J.
Williams, A. M. Z. Slawin, Tetrahedron 1993, 49, 8087. f) J. Fried,
S. Kittisopikul, E. A. Hallinan, Tetrahedron Lett. 1984, 25, 4329.
a) M. J. Van Hamme, D. J. Burton, J. Fluorine Chem. 1979, 13, 407.
b) M. J. Van Hamme, D. J. Burton, J. Organomet. Chem. 1979, 169,
123.
6
7
8
9
10 a) G. Hirai, T. Watanabe, K. Yamaguchi, T. Miyagi, M. Sodeoka,
J. Am. Chem. Soc. 2007, 129, 15420. b) M. Sodeoka, G. Hirai, T.
Watanabe, T. Miyagi, Pure Appl. Chem. 2009, 81, 205.
11 Formation and stability in water of dibromofluoromethyltris(di-
¹
methylamino)phosphonium bromide [(Me2N)3PCFBr2]+Br have
been reported. D. J. Burton, R. M. Flynn, R. G. Manning, R. M.
Kessler, J. Fluorine Chem. 1982, 21, 371.
12 See Supporting Information.
13 To compare the reactivities of the ylides 1 and 2c, reactions of
some substrates with 1 at room temperature were examined
(Scheme S1).12 Better results were obtained for 2c.
14 Crystallographic data reported in this manuscript have been
deposited with Cambridge Crystallographic Data Centre as supple-
mentary publication no. CCDC-1402667.
15 Oxidation of 1-bromo-1,2,3,3,4,4,4-heptafluorobut-1-ene provided
the corresponding epoxide along with a 2-bromo-2,3,3,4,4,4-hexa-
fluorobutanoyl fluoride: T. I. Filyakova, A. Y. Zapevalov, M. I.
Kodess, V. I. Saloutin, Russ. J. Org. Chem. 2006, 42, 1696.
16 D. Seyferth, S. P. Hopper, J. Organomet. Chem. 1973, 51, 77.
17 Low reproducibility of the reaction was observed when equimolar
amounts of HMPT and CFBr3 were used.
18 a) E. J. Corey, M. Chaykovsky, J. Am. Chem. Soc. 1962, 84, 867.
b) E. J. Corey, M. Chaykovsky, J. Am. Chem. Soc. 1965, 87, 1353.
c) A.-H. Li, L.-X. Dai, V. K. Aggarwal, Chem. Rev. 1997, 97, 2341.
19 a) W. Dumont, P. Bayet, A. Krief, Angew. Chem., Int. Ed. Engl.
1974, 13, 274. b) K. Takaki, M. Yasumura, K. Negoro, Angew.
Chem., Int. Ed. Engl. 1981, 20, 671.
20 A. Osuka, H. Suzuki, Tetrahedron Lett. 1983, 24, 5109.
21 a) M. C. Henry, G. Wittig, J. Am. Chem. Soc. 1960, 82, 563. b) P. S.
Kendurkar, R. S. Tewari, J. Organomet. Chem. 1973, 60, 247.
c) W. C. Still, V. J. Novack, J. Am. Chem. Soc. 1981, 103, 1283.
22 Y. Matano, J. Chem. Soc., Perkin Trans. 1 1994, 2703.
23 Corey-Chaykovsky reaction of a phosphorus ylide was predicted to
be a disfavored process by means of theoretical studies. F. Volatron,
O. Eisenstein, J. Am. Chem. Soc. 1987, 109, 1.
We thank Prof. Teiji Chihara for helpful discussions. This
work was partially supported by Grants-in-Aid for Scientific
Research (No. 26102744).
Supporting Information is available electronically on J-STAGE.
References and Notes
1
2
D. J. Burton, Z.-Y. Yang, W. Qiu, Chem. Rev. 1996, 96, 1641.
a) R. W. Vanderhaar, D. J. Burton, D. G. Naae, J. Fluorine Chem.
1972, 1, 381. b) J. Xu, Y. Wang, D. J. Burton, Org. Lett. 2006, 8,
2555. c) Y. Wang, D. J. Burton, Tetrahedron Lett. 2006, 47, 9279.
d) J. Xu, D. J. Burton, J. Org. Chem. 2006, 71, 3743.
3
a) S. Eddarir, Z. Abdelhadi, C. Rolando, Tetrahedron Lett. 2001, 42,
9127. b) P. S. Bhadury, M. Palit, M. Sharma, S. K. Raza, D. K.
Jaiswal, J. Fluorine Chem. 2002, 116, 75. c) A. Saito, M.
Nakagawa, T. Taguchi, J. Fluorine Chem. 2005, 126, 1166. d) R.
Chem. Lett. 2015, 44, 1389–1391 | doi:10.1246/cl.150523
© 2015 The Chemical Society of Japan | 1391