TiCl4 and Grignard reagent-promoted ring-opening reactions of various epoxides: synthesis of γ-hydroxy-α,α-difluoromethylenephosphonates

Article abstract of DOI:10.1016/j.tetlet.2008.07.146

A straightforward methodology for the synthesis of diethyl γ-hydroxy-α,α-difluoromethylenephosphonates is reported. In the presence of titanium tetrachloride, epoxide ring-opening reactions occurred upon treatment with lithium diethyl difluoromethylenephosphonate. When diethyl 3,4-epoxy-1,1-difluorobutylphosphonate was reacted with TiCl₄ and Grignard reagents, the corresponding halohydrins were obtained in very good yields.

Full text of DOI:10.1016/j.tetlet.2008.07.146

Tetrahedron Letters 49 (2008) 6046–6049
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
TiCl4 and Grignard reagent-promoted ring-opening reactions of various
epoxides: synthesis of c-hydroxy-a,a-difluoromethylenephosphonates
Romana Pajkert a, Alexander A. Kolomeitsev b, Magdalena Milewska a,
a,
Gerd-Volker Röschenthaler b, , Henryk Koroniak
*
*
a Adam Mickiewicz University, Faculty of Chemistry, Grunwaldzka 6, 60-780 Poznan, Poland
b Institute for Inorganic and Physical Chemistry, University of Bremen, Leobener Strasse, D-28334 Bremen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
A straightforward methodology for the synthesis of diethyl c-hydroxy-a,a-difluoromethylenephospho-
Received 15 April 2008
Revised 24 July 2008
Accepted 25 July 2008
Available online 31 July 2008
nates is reported. In the presence of titanium tetrachloride, epoxide ring-opening reactions occurred
upon treatment with lithium diethyl difluoromethylenephosphonate. When diethyl 3,4-epoxy-1,1-
difluorobutylphosphonate was reacted with TiCl4 and Grignard reagents, the corresponding halohydrins
were obtained in very good yields.
Ó 2008 Published by Elsevier Ltd.
Keywords:
Difluoromethylenephosphonates
Lewis acids
Epoxide ring-opening reactions
Halohydrins
Lewis acid-mediated cleavage of epoxides with various nucleo-
philes is an important transformation in organic synthesis.1 The
popularity of oxiranes as versatile intermediates in organic prepa-
rations generally emanates from their accessibility, reactivity, and
the stereochemical predictability of their reactions. Despite, how-
ever, the numerous transformations already reviewed, little atten-
tion has been given to the opening of epoxides containing a
difluoromethylenephosphonate moiety.2 Moreover, very few
examples are known in the literature, where a phosphorus and
fluorine-containing nucleophile has been utilized in these
reactions.3
reactions of epoxides with lithium diethyl difluoromethylene-
phosphonate. The reactivity of diethyl 3,4-epoxy-1,1-difluorobutyl-
phosphonate toward Lewis acids and Grignard reagents is also
presented.
Our first attempt involved the reaction of propylene oxide with
a phosphonodifluoromethyl carbanion (Scheme 1).
In this reaction, (EtO)2P(O)CF2Li 1a was generated by deproto-
nation of 1 with LDA in dry THF at ꢀ78 °C.6 Propylene oxide was
added dropwise, followed by the careful addition of a stoichiome-
tric amount of titanium tetrachloride. An exothermic and vigorous
reaction occurred on addition of the first drop of catalyst, but the
temperature of the reaction mixture was controlled so as to not
exceed ꢀ78 °C, which is required for reactions of lithium
difluoromethylenephosphonate. After stirring for 2 h at ꢀ78 °C,
the reaction mixture was quenched with distilled water, and the
temperature was allowed to rise to ambient. The reaction was
successful and propylene oxide was opened regioselectively at
the unsubstituted carbon, affording diethyl 1,1-difluoro-3-hydroxy-
Interest in the synthesis of c-hydroxy-a,a-difluoromethylene-
phosphonates results mostly from their potential in the design of
non-hydrolyzable analogues of biologically active phosphate
esters. More recently, they have been described as substrates for
NADH-linked sn-glycerol 3-phosphate dehydrogenase or as key
intermediates in the synthesis of various phosphatase-resistant
phosphonolipids.4 In contrast to non-fluorinated hydroxyl-methyl-
enephosphonates, the preparation of which has been studied
largely by oxirane ring-opening reactions with phosphonomethyl
organometallic reagents,5 this methodology has rarely been
applied in the synthesis of fluorinated analogues.2 Hence, we
report our results on the Lewis acid-promoted ring-opening
butylphosphonate
2
in 72% yield. It should be noted that
R
1. LDA/THF, -78 °C
2. Epoxide
3. TiCl4
(EtO)2P(O)CF2H
(EtO)2(O)PF2C
OH
4. H2O
1
2
* Corresponding authors. Tel.: +49 421 218 2493; fax: +49 421 218 4267
(G.-V.R.); tel.: +48 61 829 1303; fax: +48 61 829 1507 (H.K.).
E-mail addresses: gvr@chemie.uni-bremen.de (G.-V. Röschenthaler), koroniak@
amu.edu.pl (H. Koroniak).
R=Me, Et, Ph, cyclohexyl
-hydroxy- -difluoromethylenephosphonates.
Scheme 1. Synthesis of
c
a,a
0040-4039/$ - see front matter Ó 2008 Published by Elsevier Ltd.
doi:10.1016/j.tetlet.2008.07.146

R. Pajkert et al. / Tetrahedron Letters 49 (2008) 6046–6049
6047
O
O
O
(EtO)2P
(EtO)2P
P(OEt)2
F
F
F
F
O
3
(EtO)2P(O)CF2
other products
Scheme 2. Formation of tetraethyl difluoromethylenebisphosphonate 3.
1a
compound 2 was accompanied by a small amount of the corre-
sponding chlorohydrin, which is due to the known ability of metal
halides to open simple epoxides.7 Another byproduct appearing in
this reaction was tetraethyl difluoromethylenebisphosphonate 3.
Its formation was probably the consequence of the subsequent
slow reaction that product 2 experiences at the P@O center with
lithium diethyl difluoromethylenephosphonate 1a (Scheme 2). All
byproducts and unreacted starting materials were removed
successfully during work-up, and flash column chromatography
afforded pure 2 in 62% yield.
In the course of our studies to optimize the reaction conditions,
we observed that 1a was not sufficiently reactive to open the oxi-
rane ring without the presence of the Lewis acid (Table 1). This
observation is not surprising since anion 1a is a weak nucleophile,
and apparently the formation of the complex: epoxide—Lewis acid
is mandatory for the reaction to occur. Furthermore, we found that
the use of a stoichiometric amount of TiCl4 gave the best yield. This
is probably due to the poor solubility of TiCl4 at the low tempera-
ture at which the reaction is carried out. We also found that neither
the use of excess starting (EtO)2P(O)CF2H (2 equiv) nor lengthening
the reaction time improved the conversion yield.
The epoxide-opening reaction with lithium diethyl dif-
luoromethylenephosphonate 1a was also examined in the pres-
ence of other Lewis acids such as Ti(O-iPr)4, Ti(OTf)4, BF3ꢁEt2O
and Cu(OTf)2 (Table 1). Generally, none of these catalysts afforded
the desired ring-opened products. Ti(O-iPr)4 and titanium and cop-
per triflates were ineffective, and starting material was recovered.
Also, no activation was observed in the presence of LiBr, the use of
which was mentioned in the case of epoxide opening with perfluo-
rohexyl lithium.8
Surprisingly, the use of boron trifluoride BF3ꢁEt2O, the most
common catalyst applied in ether cleavage reactions was unsuit-
able.9 BF3ꢁEt2O has been applied successfully for oxirane ring-
opening reactions with fluorinated vinyllithium compounds and
non-fluorinated phosphonates, however, in our case it gave only
a 5% yield of the expected product 2. Boron trifluoride under these
conditions seems to be less prone to coordinate at the boron
center with an epoxide, and to catalyze nucleophilic ring opening.
Preferentially, in the case of difluoromethylenephosphonate
anion 1a, BF3ꢁEt2O undergoes reaction with 1a to afford
ðEtOÞ2PðOÞCF2BF3ꢀLiþ as the main product (87% yield). We envis-
age this new and very interesting compound as an organoboron
reagent for cross-coupling reactions, since the analogous
perfluoroorganotrifluoroborate salts have been described recently
as highly efficient reagents for reactions of this type.10
In an attempt to extend the scope of the above-described reac-
tion, we decided to explore the possibility of opening other oxir-
anes. All the representative epoxides studied here, 1,2-butene
oxide, styrene oxide, and cyclohexene oxide were ring-opened
and gave the corresponding products. As originally expected, the
reaction is regioselective, which is undoubtedly due to nucleophilic
attack at the less hindered site of the oxirane ring. Except for 1,2-
butene oxide, in all cases, the formation of analogous chlorohyd-
rins was observed. All impurities were removed easily during work
up to afford alcohols 4–7 in satisfactory yields (Table 2).
It should be noted that no product of type 8 resulting from
cleavage of the oxolane ring with lithium salt 1a was observed
(Table 2, entry 5), which makes THF a potential solvent for this
type of reaction.
These results prompted us to examine the reactivity of diethyl
3,4-epoxy-1,1-difluorobutylphosphonate 10 toward various Lewis
acids and Grignard reagents. We considered that this transforma-
tion could represent another route to
methylenephosphonates.
c-hydroxy-a,a-difluoro-
The synthesis of diethyl 3,4-epoxy-1,1-difluorobutylphospho-
nate 1011 was achieved successfully as shown in Scheme 3 by
the reaction of methyl(trifluoromethyl)dioxirane generated in situ
with unsaturated phosphonate 9 in a water-acetonitrile mixture as
solvent.12 This new and simple procedure, which is an alternative
to previously reported syntheses of this compound, was very effi-
cient and led to the desired product in excellent yield (92%), even
when the reaction was conducted on a preparative scale.
Next, we treated oxirane 10 with titanium tetrachloride
(1.1 equiv) in dry THF and after two hours of stirring at room
temperature, we observed complete consumption of the starting
material (by 19F NMR). After typical work up and column
chromatography, the sole product, diethyl 4-chloro-1,1-difluoro-
3-hydroxybutylphosphonate, was isolated in 82% yield. Moreover,
we found that the reaction occurred with complete regioselectivity
via preferential attack on the terminal carbon of epoxyphospho-
nate 10. With milder Lewis acids, for example BF3ꢁEt2O and Me3Al,
no reaction occurred and only starting materials were recovered.
Interestingly, similar behavior was observed when Grignard
reagents were used as nucleophiles. In this case, the reaction
resulted in the regioselective formation of the corresponding
chloro- or bromohydrin (attack on the less hindered side of the
epoxide) and no other products were detected in the reaction mix-
tures (based on 19F NMR analysis). This unexpected behavior of
diethyl 3,4-epoxy-1,1-difluoromethylenephosphonate 10 with
organomagnesium reagents suggests that these could also act as
Lewis acids. Probably, due to the reaction conditions, the Schlenk
equilibrium, which occurs in solutions of Grignard reagents,13
tends to favor the formation of diorganyl magnesium and a
Table 1
Lewis acids used in ring-opening reactions with (EtO)2P(O)CF2H (1)
Entry
Catalyst
Reaction conditions
Yield of 2a (%)
1
2
3
4
5
6
7
TiCl4
None
LDA, THF, ꢀ78 °C
LDA, THF, ꢀ78 °C
LDA, THF, ꢀ78 °C
LDA, THF, ꢀ78 °C
LDA, THF, ꢀ78 °C
LDA, THF, ꢀ78 °C
t-BuLi, Et2O, ꢀ90 °C
72
—
5b
—
—
—
5
BF3ꢁEt2O
Ti(Oi-P)4
Ti(OTf)4
Cu(OTf)2
TiCl4
a
19F NMR yields from starting (EtO)2P(O)CF2H.
Main product: ðEtOÞ2PðOÞCF2BF3ꢀLiþ (87%).
b

6048
R. Pajkert et al. / Tetrahedron Letters 49 (2008) 6046–6049
Table 2
Diethyl c-hydroxy-a,a-difluoromethylenephosphonates produced as shown in Scheme 1
Entry
Epoxide/Oxolane
Product
Yielda (%)
62
O
CF2P(O)(OEt)2
1
OH
4
HO
O
CF2P(O)(OEt)2
2
34
5
O
CF2P(O)(OEt)2
3
4
47
26
0
OH
6
OH
O
CF2P(O)(OEt)2
7
O
HO
5
CF2P(O)(OEt)2
8
a
Isolated yields after purification by flash column chromatography.
O
CF3C(O)CH3, Oxone
CF2P(O)(OEt)2
CF2P(O)(OEt)2
10(92%)
AcCN-H2O, pH~7
9
Scheme 3.
Table 3
magnesium halide salt (a weak Lewis acid). Complexation of this
Lewis acid with epoxide 10 is then possibly preferred over the
reaction with the Grignard reagent. As a consequence, halohydrins
were obtained in very good yields as sole products, even when the
reactions were carried out under different reaction conditions
(Scheme 4, Table 3).
As has already been mentioned, all of these reactions occured at
the less hindered side of the oxirane ring. The predominant regi-
oselectivity can be explained by Coulombic repulsion between
the lone pairs of the difluoromethyl group and the negative charge
on the nucleophile (Scheme 5).14
Grignard reagents used in the formation of halohydrins produced as shown in
Scheme 4
Entry
Grignard reagent
Reaction conditions (product)
Yielda (%)
1
2
3
4
5
6
7
CH3MgCl
CH3MgCl
AllylMgCl
AllylMgCl
PhMgBr
ꢀ78 °C ? rt, THF, 24 h (11)
rt, 12 h (11)
97
97
78
78
79
73
78
ꢀ78 °C ? rt, THF, 24 h (11)
rt, 12 h (11)
ꢀ78 °C ? rt, THF, 24 h (12)
rt, 12 h (12)
rt, 12 h (11)
PhMgBr
PhCH2MgCl
a
Isolated yields after purification by flash column chromatography.
It is also worth mentioning that all the
c-hydroxy-a,a-di-
fluoromethylenephosphonates show complicated spectra, but par-
ticularly significant are the CF2 fluorine signals in the 19F NMR
spectra which exhibit a characteristic AB type pattern. In these
molecules, both the fluorine atoms as well as both protons at the
b carbon are not equivalent (Fa, Fb, and Ha, Hb), which is due to
the stereogenic carbon. The complicated multiplicity of each fluo-
rine atom is a consequence of coexisting couplings between fluo-
rine–fluorine, fluorine–phosphorus, and finally fluorine–proton.
What is more, the presence of hydroxyl, difluoromethylene, and
phosphonyl groups in the same molecule creates the possibility of
hydrogen bonding. Parameters such as state (solid or in solution),
concentration, temperature and polarity of the solvent will deter-
mine the preferred conformation (synclinal or antiperiplanar) and
will be crucial for the formation of either intramolecular or inter-
molecular hydrogen bonds. Simple NMR experiments at room tem-
perature in solvents of different polarity—CDCl3 and D2O, have
delivered much information already.15
OH
O
1. TiCl4 or RMgX
(EtO)2P(O)CF2
(EtO)2P(O)CF2
2. H2O
10
X
11 X=Cl
12
X=Br
In conclusion, a general access to diethyl
c-hydroxy- and d-
Scheme 4.
halo- -hydroxy- -difluoromethylenephosphonates via epoxide
c
a,a

R. Pajkert et al. / Tetrahedron Letters 49 (2008) 6046–6049
6049
LA
OH
X
-
O
(EtO)2P(O)CF2
X
X
(EtO)2P(O)CF
2
+
O
X
Lewis Acid
relatively stable
(EtO)2P(O)CF2
(EtO)2P(O)CF
2
LA
O
+
-
(EtO)2P(O)CF2
OH
Scheme 5.
3. (a) Gillet, J. P.; Sauvetre, R.; Normant, J. F. Synthesis 1986, 355; (b) Normant, J. F.
J. Organomet. Chem. 1990, 400, 19; (c) Li, Z.; Racha, S.; Dan, L.; El-Subbagh, H.;
Abushanab, E. J. Org. Chem. 1993, 58, 5779; (d) Azuhata, T.; Okamoto, Y.
Synthesis 1983, 916; (e) Azuhata, T.; Okamoto, Y. Synthesis 1984, 417; (f)
Tanaka, H.; Fukui, M.; Haraguchi, K.; Masaki, M.; Miyasaka, T. Tetrahedron Lett.
1989, 30, 2567.
4. (a) Nieschalk, J.; O’Hagan, D. Chem. Commun. 1995, 719; (b) Nieschalk, J.;
Batsanov, A. S.; O’Hagan, D.; Howard, J. A. K. Tetrahedron 1996, 52, 165; (c)
Chambers, R. D.; Jaouhari, R.; O’Hagan, D. Chem. Commun. 1988, 1988; (d)
Chambers, R. D.; Jaouhari, R.; O’Hagan, D. Tetrahedron 1989, 45, 5101; (e)
Moolenar, W. H. Exp. Cell Res. 1999, 253, 230; (f) Goetzl, E. J.; Dolezalova, H.;
Kong, Y.; Hu, Y. L.; Jaffe, R. B.; Kalli, K. R.; Conover, C. A. Cancer Res. 1999, 59,
5370; (g) Xu, Y.; Prestwich, G. D. Org. Lett. 2002, 4, 4021.
ring-opening reactions has been developed. Representative
oxiranes underwent the reaction with lithium difluoro-
methylenephosphonate in the presence of TiCl4 to give the corre-
sponding products in good yields. The successful opening of cyclo-
hexene oxide broadens the utility of the reaction described herein
to 1,2-disubstituted epoxides. Moreover, when diethyl 3,4-epoxy-
1,1-difluorobutylphosphonate was reacted with TiCl4 or Grignard
reagents, the corresponding halohydrins were obtained. Com-
pounds of this type are of synthetic potential and can serve as
useful precursors for a variety of derivatives.4
5. (a) Smith, J. G. Synthesis 1984, 629; (b) Tanaka, H.; Fukui, M.; Haraguchi, K.;
Masaki, M.; Miyasaka, T. Tetrahedron Lett. 1989, 30, 2457; (c) Montchamp, J. L.;
Migaud, M. E.; Frost, J. W. J. Org. Chem. 1993, 58, 7679 and references cited
therein.
Acknowledgment
6. Obayashi, M.; Ito, E.; Matsui, K.; Kondo, K. Tetrahedron Lett. 1982, 23,
2323.
7. Bonini, C.; Righi, G. Synthesis 1994, 225.
The Deutschen Forschungsgemeinschaft (Grant 436 POL 113/
120/0-1) is greatly acknowledged for the financial support of this
work.
8. Uno, H.; Suzuki, H. Synlett 1993, 91.
9. (a) Gillet, J. P.; Sauvetre, R.; Normant, J. F. Synthesis 1986, 355; (b) Normant, J. F.
J. Organomet. Chem. 1990, 400, 19; (c) Li, Z.; Racha, S.; Dan, L.; El-Subbagh, H.;
Abushanab, E. J. Org. Chem. 1993, 58, 5779; (d) Bakalarz-Jeziorna, A.; Helinski,
J.; Krawiecka, B. J. Chem. Soc., Perkin Trans. 1 2001, 1086.
Supplementary data
10. Kolomeitsev, A. A.; Kadyrov, A. A.; Szczepkowska-Sztolcman, J.; Milewska, M.;
Koroniak, H.; Bissky, G.; Barten, J. A.; Röschenthaler, G-V. Tetrahedron Lett.
2003, 44, 8273 and references cited therein.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2008.07.146.
11. Chambers, R.; O’Hagan, D.; Jauhari, R. Tetrahedron 1989, 45, 5101.
12. Yang, D.; Wonk, M. K.; Yip, Y. C. J. Org. Chem. 1995, 60, 3887.
13. Tammiku-Taul, J.; Burk, P.; Tuulmets, A. J. Phys. Chem. A 2004, 108, 133.
14. Katagiri, T.; Uneyama, K. J. Fluorine Chem. 2000, 105, 285.
15. See Supplementary data.
References and notes
1. March, J. Advanced Organic Chemistry, 3rd ed.; John Wiley: New York, 1985.
2. Ozouf, P.; Binot, G.; Pommelet, J.-C.; Lequeux, T. P. Org. Lett. 2004, 6, 3747.