Novel cobalt(0)- or magnesium-mediated approaches to β-ketophosphonates

Article abstract of DOI:10.1055/s-2002-33653

Two novel approaches to β-ketophosphonates, based on cobalt(0)- or magnesium-mediated reactions of α-halophosphonates with esters are described.

Full text of DOI:10.1055/s-2002-33653

PAPER
1683
Novel Cobalt(0)- or Magnesium-Mediated Approaches to -Ketophosphonates
N
ovel Approa
u
che
s
to -Ke
l
tophosp
v
honates ia Orsini,* Emanuela Di Teodoro, Marinella Ferrari
Dipartimento di Chimica Organica e Industriale, Universita’ degli Studi di Milano, via Venezian, 21-20133 Milano, Italy
Fax +39(2)50314106; E-mail: Fulvia.Orsini@unimi.it
Received 4 April 2002; revised 18 June 2002
(as well as anhydrides or imidazolides) in the presence of
Abstract: Two novel approaches to -ketophosphonates, based on
cobalt(0)- or magnesium-mediated reactions of -halophosphonates
with esters are described.
magnesium chloride/triethylamine, followed by decar-
boxylation: this last step, however, may require heating
with strong aqueous acids, not recommended for sensitive
substrates.13 Furthermore reactions are slow and heating,
that speeds the substrate transformation, also entails side
products formation. A further, different approach is based
on the reaction of organometallic reagents (such as orga-
nocuprates or organolithiums) with 2-dialkylphospho-
noalkanoyl chlorides from -lithiumalkyl phosphonates.14
Other miscellaneous approaches include: base-induced
isomerization of enol phosphates; reaction of ketone eno-
lates with dialkyl phosphorochloridrites, followed by
oxidation;15 acylation of 1-(trimethylsilyl)vinyl phospho-
nates;16 and hydrolysis of vinylogous phosphoramides.17
Key words: -ketophosphonates, organomagnesium compounds,
organocobalt compounds
-Ketophosphonates, phosphonic analogues of -keto es-
ters, are one of the very important class of organic com-
pounds. Their complexing ability has been exploited for
selective metal extraction.1 Furthermore they have found
several applications as organic intermediates. Examples
are the homologation of carbonyl compounds to , -un-
saturated derivatives via the Horner–Emmons reaction;2
the synthesis of difficult-to-prepare complex phospho-
nates, such as phosphorylated heterocyclic compounds;3
the synthesis of chiral -hydroxyphosphonic acids, im-
portant surrogates for the corresponding carboxylic ac-
ids;4 and the synthesis of -amino phosphonic acids,
important antimetabolites of amino acids.5
In this paper we discuss two novel protocols for the syn-
thesis of -ketophosphonates based on the reaction of -
halophosphonates with esters: in one case the reactions
were promoted by a soluble Co(0) complex; in the other
case by magnesium metal (Scheme 1).
The above mentioned Co(0) complexes have been already
successfully used to promote Reformatsky-type18 and al-
dol-type19 reactions and are easily prepared from Co(II)
salts by reduction with magnesium metal in the presence
of a selected phosphine.20 They can be used either in sto-
ichiometric or in substoichiometric (catalytic) amounts,
provided, in the latter case, magnesium metal is present,
to carry out the reduction from Co(II) to Co(0).
Many synthetic approaches to -ketophosphonates, that
not general and suffering from limitations, have been de-
veloped. The reaction of -haloketones with trialkyl
phosphites6 (Michaelis–Arbuzov reaction) or with metal-
lated dialkyl phosphites7 (Michaelis–Becker reaction) is
restricted to highly reactive -haloketones and generally
gives a mixture of isomers ( -ketophosphonates, , -ep-
oxyphosphonates, and enol phosphates). The use of
masked carbonyls8, although successful, is limited by the
availability of starting reagents. The Claisen condensation
between -lithiumalkyl phosphonates and esters9 give a
mixture of -ketophosphonates and alkyl phosphonates,
due to the higher acidity of the former with respect to the
latter. Also attack of the -metallated phosphonate on the
carbonyl carbon of -ketophosphonate has been ob-
served.10,11 In any case, -lithiumalkyl phosphonates re-
quire low temperatures, strictly anhydrous media and
strong bases such as butyllithium for their preparation.
The modified version using organocopper reagents12
overcomes the acid-base exchange but requires acid chlo-
rides and a large amount of butyllithium, which makes the
procedure less convenient in terms of safety and costs.
Scheme 1
To define the role of the metal and to optimize the exper-
imental conditions and the yields of the reactions, three
series of experiments were carried out. Firstly a Co-medi-
ated, ‘stoichiometric’ procedure was considered: the halo-
phosphonate and the ester were simultaneously added to a
tetrahydrofuran solution of the Co(0)-complex at room
temperature. The optimum halophosphonate/Co(0)-com-
plex/ester ratio was 2.0:2.0:1.0: a lower halophosphonate/
Co(0)-complex/ester ratio gave lower yields. The best re-
sults were obtained with a one-pot procedure, adding the
In an approach related to malonic synthesis, -ketophos-
phonates have been obtained by reaction of acid chlorides
Synthesis 2002, No. 12, Print: 06 09 2002.
Art Id.1437-210X,E;2002,0,12,1683,1688,ftx,en;P01302SS.pdf.
© Georg Thieme Verlag Stuttgart · New York
ISSN 0039-7881

1684
F. Orsini et al.
PAPER
two organic reagents simultaneously to the Co(0) com- amounts depending on the ester. These results indicate
plex.
that in the absence of Co-complex, magnesium alone is
capable of giving -ketophosphonates from halophospho-
nates and ester with equal, or better efficiency in some
cases, albeit in a less predictable way.
In the second series of experiments, (‘substoichiometric’
procedure) a halophosphonate/Co(0)-complex/ester ratio
10:1:10 was used and the reaction was performed in the
presence of magnesium metal. In both series of experi- Whereas there is no doubt on the key role played by cobalt
ments diethyl -iodomethylphosphonate, already success- in the ‘stoichiometric’ procedure, the observation that
fully used in Reformatsky-type reactions, and the Co(0)- magnesium on its own mediates the reactions, could
complex with trimethylphosphine were first used; the sol- arouse doubts on the role of cobalt in the ‘substoichiomet-
id, easier-to-handle triphenylphosphine can be used as ric’ conditions. However, in addition to the different reac-
well as a ligand to the metal.
tion speed and reproducibility observed in the presence or
in the absence of the Co(0)-complex (higher in the former
case), one more piece of evidence that might prove the ac-
tual involvement of cobalt, also under ‘substoichiometric’
conditions, is the color of the solution which, as reported
above, continuously changes during the addition of the or-
ganic reagents. It is therefore reasonable to assume that, as
far as the reduction Co(II)/Co(0) is faster than the reaction
of magnesium with the -halophosphonate, the soluble
Co(0)-complex reacts with the -halophosphonate before
magnesium, that must work in heterogeneous phase.
In the third series of experiments, the -halophosphonate
and the ester were added to magnesium alone in tetrahy-
drofuran under the same experimental conditions used for
the Co-mediated reactions under ‘substoichiometric’ con-
ditions (speed of stirring, temperature, dilution, etc.).
The results are summarized in the Table and deserve few
comments. A variety of ester structural forms are tolerated
in this reaction. The desired -ketophosphonates are gen-
erally obtained in reasonable to good yields (45–90%),
depending on the substrate and on the protocol used. With
esters containing a nitrogen atom elsewhere, the reaction
was less efficient, as exemplified by 2-carbomethoxypyri-
dine which gave the lowest yields (34–39% at the best in
Co-mediated reactions). The Co-mediated reaction was
not practicable with , -unsaturated esters, which gave
mixtures of unidentified products. Concurrent attack of
the organometallic reagent on the carbonyl carbon of the
In conclusion we have presented two novel procedures for
the synthesis of -ketophosphonates via a Reformatsky-
type addition of -halophosphonates to esters. In one pro-
cedure the reactions are performed using a catalytic
amount of a soluble Co(0)-complex in the presence of
magnesium metal. In the other procedure magnesium met-
al alone is used.
formed -ketophosphonate was not observed, even in the In general, the reactions proceed with comparable effi-
presence of the reagent in excess. Besides the -ketophos- ciency with both procedures; in few cases better yields
phonate, the only other compounds observed in the reac- were obtained with magnesium metal alone, but this metal
tion mixture were unreacted starting ester and diethyl works in a less reproducible way due to the heterogeneous
methanephosphonate.21 The former was easily separated conditions and to the requirement of the metal activation.
by chromatography and recycled. In most cases a simple Advantages of these cobalt- and/or magnesium methodol-
filtration on a very short column of silica gel was enough; ogies with respect to known procedures are: mild reaction
a procedure easily applicable also if the reaction has to be conditions and easily available carbonyl substrates; no
scaled up.
strong bases, acids or heat required; lower costs and high-
er safety; reagents easier to be handled and stored; and no
side products deriving from the attack on the carbonyl car-
bon of the -ketophosphonate are formed. When cobalt is
used, an added advantage, which also makes the cobalt-
procedure more valuable, is the color of the solution,
which provides an efficient monitoring of the course of
the reaction and allows a timely visualization of its end
point with minimization of byproducts.
Comparing the Co-mediated reactions, the ‘stoichiome-
tric’ procedure can be the protocol of choice for substrates
with delicate, highly-magnesium sensitive, functional-
ities; but there is no doubt that, in general, the ‘substoichi-
ometric’ procedure is more convenient: it is not necessary
to filter-off the magnesium; the amounts of cobalt and
phosphines are minimized (thus offsetting the disadvan-
tage of their use); the workup of the reaction and the puri-
fication of the products are easier and, as a consequence,
a higher recovery of products may be observed. Further-
more the course and the end point of the reaction are easily
monitored by the alternation of the brown color of the
Co(0)-complex and the blue-violet color of the Co(II)-
complex during the addition of the -halophosphonate/es-
ter reagents to the Cobalt-complex.22
Anhyd THF from Aldrich was generally used. With Mg-mediated
reactions, reagent grade THF was refluxed over LiAlH4 and dis-
tilled before use. 1H NMR and 13C NMR spectra were recorded at
200 and 300 MHz on Bruker spectrometers. MS spectra were re-
corded on a VG7070 E9 spectrometer. Flash-column chromatogra-
phy was performed on silica gel (Merck Kieselgel 60, 230–400
mesh). TLC was performed on silica gel plates (60 F254, Merck),
spots were detected visually by ultraviolet irradiation (254 nm) or
by spraying with MeOH–H2SO4 (9:1), followed by heating at 100
°C. All reactions were performed under dry N2, using glassware
dried by flaming in a stream of dry N2. Petroleum ether used had bp
40–60°C.
The reactions performed with magnesium alone, in the ab-
sence of cobalt, were less reproducible and required a
variable period of induction. Once started, the reactions
gave the expected -ketophosphonates in variable
Synthesis 2002, No. 12, 1683–1688 ISSN 0039-7881 © Thieme Stuttgart · New York

PAPER
Novel Approaches to -Ketophosphonates
1685
Table Reactions Between Ethyl -Iodomethylphosphonate and Esters
Entry
1
Ester
Product
Yield (%)
Procedurea
51 (22)b
63 (11)b
37 (49)b
A
B
C
1a
1b
1c
2a
2b
2
3
46 (40)b
70 (22)b
45 (42)b
A
B
C
–
A
B
C
83
90c
2c
4
mixture of products
56
B
C
1d
1e
2d
2e
5
6
mixture of products
30
A
C
57
B
1f
2f
7
8
57
B
C
85c
2g
1g
1h
56
B
C
80c
2h
2i
9
53
47
B
C
1i
1j
10
34 (55)b
39
A
B
C
14 (15)b
2j
a Procedure A (stoichiometric): ester, ethyl -iodomethylphosphonate and [Co(PMe3)]4 were used in a 1:2.0:2.0 ratio. Procedure B (substoichi-
ometric): the reaction was performed using a substoichiometric quantity of Co(PMe3)4 (10–15%). Procedure C: the reaction was performed
using metallic Mg only.
b The amount given in parenthesis refers to recovered starting ester.
c Yields determined by 1H NMR spectroscopy.
Stoichiometric Reaction Mediated by (Me3P)4Co; General Pro-
cedure
Procedure A: A 1 M solution of Me3P in THF (10 mL) was added
to a mixture of activated Mg turnings23,24 (0.3 g) and anhyd CoCl2
(0.26 g, 2.0 mmol). The reaction mixture was stirred at r.t. till a dark
brown color developed.25 The excess of Mg was filtered, and to the
resulting solution were added diethyl -iodomethylphosphonate
(0.32 mL, 2.0 mmol) and the ester (1 mmol) in THF (2–3 mL) drop-
wise over a period of 1 h. The reaction was monitored by TLC (sil-
ica gel, eluent: EtOAc–petroleum ether, 7:3). At the end of the
Synthesis 2002, No. 12, 1683–1688 ISSN 0039-7881 © Thieme Stuttgart · New York

1686
F. Orsini et al.
PAPER
reaction, the mixture was stirred in the air till a blue color devel-
oped. Two different workup procedures were then used depending
on the water-solubility of the ester and/or of the product. For water-
insoluble compounds the reaction mixture was diluted with EtOAc
and poured into 5% aq HCl (2–3 mL). The aqueous phase was ex-
tracted with EtOAc (3 15 mL). The organic layers were collected,
dried (Na2SO4), and evaporated to dryness. Occasionally a light
blue or green color still persistent in the crude material was elimi-
nated by dissolving in EtOAc and by washing with a sat. EDTA
(bis-sodium salt) solution. For partially soluble-compounds, the re-
action mixture was filtered on silica gel eluting with EtOAc–petro-
leum ether to remove cobalt salts. The crude material was
chromatographed on silica gel, eluting with EtOAc–petroleum ether
mixtures of increasing polarity, to separate the product from the un-
reacted starting ester. Me3P can be substituted by Ph3P withouth af-
fecting the course of the reactions.
31P NMR (CDCl3): = 20.21.
MS: m/z = 270 (M+), 179 (M+ – CH3CH2O – CH3CH2OH), 151
[CH2P(O)(OEt)], 137 (151 – CH2), 91 (PhCH2).
HRMS: m/z Calcd for C13H19O4P: 270.1020977200. Found:
270.0995000.
2c
Colorless thick oil.
1H NMR (CDCl3): = 1.30 (t, 6 H, J = 7.5 Hz), 2.85–2.95 (m, 4 H),
3.05 (d, 2 H, JCP = 22 Hz), 4.1 (dq, JCP = 7.5 Hz), 7.15–7.35 (m, 5
H).
13C NMR (CDCl3): = 16.18 (dq, J = 5.18 Hz), 29.35 (t), 42.52 (dq,
JCP = 127.5 Hz), 45.29 (t), 62.40 (dt, JCP = 5.48 Hz), 2 126.03 (d),
3
128.33 (d), 140.57 (s), 200.94 (s).
31P NMR (CDCl3): = 20.46.
Substoichiometric Reaction Mediated by (Me3P)4Co; General
Procedure
MS: m/z = 284 (M+), 255 (M+ – CH3CH2), 179 (M+ – PhCH2CH2),
152 [CH3P(O)(OEt)2], 105 (PhCH2CH2), 91 (PhCH2).
Procedure B: A 1 M solution of Me3P in anhyd THF (1 mL) was
added to a mixture of activated Mg (0.2 g) and anhyd CoCl2 (0.019
g, 0.15 mmol). The mixture was stirred at r.t. till the brown color of
the cobalt(0) complex had developed. To the mixture, a THF solu-
tion (3 mL) of ester (1 mmol) and diethyl -iodomethylphosphonate
(0.334 g, 1.2 mmol) was added dropwise under stirring. The speed
of addition was modulated so as to reduce the time of persistence of
the blue color [Co(II) complex] developed during the addition to a
minimum. At the end of the reaction, indicated by the persistence of
the brown color of the original Co(0) complex, the Mg was filtered
and the resulting solution was stirred in air till a blue color devel-
oped. The reaction mixture was then treated as described above for
the stoichiometric procedure. Me3P can be substituted by Ph3P with-
out affecting the course of the reactions.
HRMS: m/z Calcd for C14H21O4P: 284.1177478000. Found:
284.1186000.
2d
Colorless thick oil.
1H NMR (CDCl3): = 1.27 (t, 6 H, J = 7.0 Hz), 1.88 (dd, 3 H,
J = 6.78, 1.8 Hz), 3.15 (d, 2 H, J = 22.6 Hz), 4.1 (dq, 4 H, J = 7.32
Hz), 6.2 (dq, 1 H, J = 14.09, 1.8 Hz), 6.91 (dq, 1 H, J = 15.7, 6.9
Hz).
13C NMR (CDCl3): = 16.11 (q), 39.8 (dq, JCP = 128.06 Hz), 62.32
(dq, JCP = 5.65 Hz), 131.46 (d), 145.26 (d), 140.75 (s), 190.8 (s).
31P NMR (CDCl3): = 20.91.
MS: m/z = 220 (M+), 191 (M+ – CH3CH2), 179 (M+ – CH3CH=CH),
152 [CH3P(O)(OEt], 69 [M+ – CH2P(O)(OEt)2], 41 (CH3CH=CH).
Reaction Mediated by Magnesium; General Procedure
Procedure C: To a suspension of activated Mg turnings (0.3 g) in
freshly distilled anhyd THF (2 mL) was added dropwise a THF so-
lution (3 mL) of ester (1 mmol) and diethyl -iodomethyl phospho-
nate (0.334 g, 1.2 mmol). The reaction was monitored by TLC
(silica gel, eluent: EtOAc–petroleum ether). The Mg was filtered
and the resulting solution was treated according to one of the two
procedures described in the previous section, depending on the wa-
ter solubility of the product.
HRMS: m/z Calcd for C9H17O4P: 220.0864476. Found:
220.0828000.
2e
Colorless thick oil.
1H NMR (CDCl3): = 1.32 (t, 6 H, J = 6.94 Hz), 3.32 (d, 2 H,
J = 22.68 Hz), 4.18 (dq, 4 H, J = 6.94 Hz), 6.9 (d, 1 H, J = 16.03
Hz), 7.21–7.6 (m, 5 H), 7.65 (d, 1 H, J = 16.03 Hz).
2a
13C NMR (CDCl3):
CP = 127.9 Hz), 62.55 (dt, JCP = 4.63 Hz), 125.69 (d), 2 128.52
(d), 2 128.92 (d), 130.80 (d), 134.12 (s), 144.65 (d), 191.01 (s).
= 16.24 (dq, J = 4.53 Hz), 40.98 (dq,
Colorless thick oil.
J
1H NMR (CDCl3): = 1.25 (t, 6 H, J = 7.8 Hz), 3.6 (d, 2 H,
JCP = 23.7 Hz), 4.1 (dq, 4 H, J = 7.8 Hz), 7.4–7.62 (m, 3 H), 7.97
(dd, 2 H).
31P NMR (CDCl3): = 20.85.
MS: m/z = 282 (M+), 253 (M+ – CH3CH2), 179 (M+ – PhCH=CH),
13C NMR (CDCl3): = 16.07 (dq, JCP = 16.0 Hz), 38.26 (dq,
152 [CH3P(O)(OEt)2], 131 [M+
(PhCH=CH).
–
CH2P(O)(OEt)2], 103
JCP = 129.55 Hz), 62.50 (dt, JCP = 5.35 Hz), 2 128.45 (d),
2
129.87 (d), 133.42 (d), 136.55 (s), 191.69 (s).
HRMS: m/z Calcd for C14H19O4P: 282.1020977. Found:
282.1039000.
31P NMR (CDCl3): = 20.64.
MS: m/z = 256 (M+), 227 (M+ – C2H5), 211 (M+ – C2H5O), 183 (M+
– C2H5O – CH2=CH2), 151 [CH2P(O)(O C2H5)2].
2f
Colorless thick oil.
HRMS: m/z Calcd for C12H17O4P: 256.0864476. Found:
1H NMR (CDCl3): = 1.22 (t, 6 H, J = 7.0 Hz), 1.10–1.45 (m, 5 H),
1.55–1.64 (m, 1 H), 1.64–1.77 (m, 2 H), 1.77–1.88 (m, 2 H), 2.48–
2.59 (m, 1 H), 3.02 (d, 2 H, J = 21.0 Hz), 4.06 (dq, 4 H, J = 7.3 Hz).
256.0906000.
2b
Colorless thick oil.
1H NMR (CDCl3): = 1.34 (t, 6 H, J = 6.5 Hz), 3.1 (d, 2 H, J = 23.3
Hz), 3.9 (s, 2 H), 4.15 (dd, 4 H, J = 6.5, 6.5 Hz), 7.15–7.35 (m, 5 H).
13C NMR (CDCl3): = 16.23 (dq, J = 5.45 Hz), 2 25.41 (t), 25.70
(t), 28.18 (t), 40.21 (dq, JCP = 117.05 Hz), 41.07 (t), 51.35 (d), 62.38
(dt, J = 5.55 Hz), 205.20 (s).
13C NMR (CDCl3):
CP = 127.61 Hz), 50.67 (t), 62.53 (dt, JCP = 7.29 Hz), 127.14 (d),
128.69 (d), 2 129.51 (d), 133.42 (s), 199.34 (s).
= 16.19 (dq, J = 5.28 Hz), 41.30 (dq,
31P NMR (CDCl3): = 20.99.
J
2
Synthesis 2002, No. 12, 1683–1688 ISSN 0039-7881 © Thieme Stuttgart · New York

PAPER
Novel Approaches to -Ketophosphonates
1687
MS: m/z = 262 (M+), 233 (M+ – CH2CH3), 217 (M+ – OCH2CH3),
MS: m/z = 257 (M+), 212 (M+ – OCH2CH3), 179 (M+ – C5H4N), 151
152 [CH3P(O)(OEt)2], 151 [CH2P(O)(OEt)2], 111 [M+
CH2P(O)(OEt)2], 83 (C6H11).
–
[CH2P(O)(OEt2], 106 [M+ – CH2P(O)(OEt)2], 78 (C5H4N).
HRMS: m/z Calcd for C11H16NO4P: 257.0816966080. Found:
Anal. Calcd for C11H23O4P: C, 54.96; H, 8.78. Found: C, 54.88; H,
8.68.
257.0822000.
Acknowledgement
2g
Colorless thick oil.
National Research Council (C.N.R) – Italy and Ministero della
Ricerca Scientifica e Tecnologica (MURST) are acknowledged for
financial support.
1H NMR (CDCl3): = 0.85 (t, 3 H, J = 7.0 Hz), 1.26 (t, 6 H, J = 7.0
Hz), 1.20–1.32 (m, 2 H), 1.5 (tt, 2 H, J = 7.0, 7.0 Hz), 2.56 (t, 2 H,
J = 7.0 Hz), 3.01 (d, 2 H, J = 23.0 Hz), 4.10 (dq, 4 H, J = 7.0 Hz).
13C NMR (CDCl3): = 13.75 (d), 16.24 (dq, J = 4.7 Hz), 22.41 (t),
25.47 (t), 42.31 (dq, JCP = 83.75 Hz), 2 43.73 (t), 62.49 (t), 202.18
References
(s).
(1) McCabe, D. J.; Duesler, E. N.; Paine, R. T. Inorg. Chem.
1985, 24, 4626; and references quoted therein.
(2) (a) Maryanoff, B. E.; Reitz, A. B. Chem. Rev. 1989, 89, 863.
(b) Corey, E. J.; Gorzynsky, S. J. J. Am. Chem. Soc. 1979,
101, 1038. (c) Bonjoch, J.; Casamitjana, N.; Quirante, J.;
Garriga, C.; Bosch, J. Tetrahedron 1992, 48, 3131.
(3) (a) Aboujaoude, E. E.; Collignon, N.; Savignac, P.
Tetrahedron 1985, 41, 427. (b) Mikolajczyk, H.;
Zurawinsky, R.; Kielbasinsky, P. Tetrahedron Lett. 1989,
30, 1143. (c) Miller, D. B.; Raychaudhuri, S. R.; Avasthi,
K.; Lal, K.; Levison, B.; Salomon, R. G. J. Org. Chem. 1990,
55, 3164.
(4) Gautier, I.; Ratovelomanana-Vidal, V.; Savignac, P.; Genet,
J. P. Tetrahedron Lett. 1996, 37, 7721.
(5) Ryglowsky, A.; Kafarski, P. Tetrahedron 1996, 52, 10685.
(6) Gaydou, E. M.; Bianchini, J. P. J. Chem. Soc., Chem.
Commun. 1975, 541.
(7) Sturtz, G. Bull. Soc. Chim. Fr. 1964, 2333.
(8) (a) Chatta, M. S.; Aguiar, A. M. J. Org. Chem. 1973, 38,
2908. (b) Gasteiger, J.; Herzig, C. Tetrahedron Lett. 1980,
21, 2687. (c) Sturtz, G.; Corbel, B.; Medinger, L.; Haelthers,
J. P.; Sturtz, G. Synthesis 1985, 1048. (d) Kim, D. Y.; Lee,
K.; Oh, D. Y. J. Chem. Soc., Perkin Trans.1 1992, 2451.
(e) Kim, D. Y.; Mang, J. Y.; Oh, D. Y. Synth. Commun.
1994, 24, 629. (f) Kim, D. Y.; Kong, M. S. J. Chem. Soc.,
Perkin Trans. 1 1994, 3359.
31P NMR (CDCl3): = 20.67.
MS: m/z = 236 (M+), 207 (M+ – CH2CH3), 179 [M+ – CH3(CH2)3],
152
[CH3P(O)(OEt)2],
151
[CH2P(O)(OEt)2],
137
[P(O)(OCH2CH3)], 85 [M+ – CH2P(O)(OEt)2], 57 [CH3(CH2)3].
Anal. Calcd for C10H21O4P: C, 50.85; H, 8.90. Found: C, 50.55; H,
8.85.
2h
Colorless thick oil.
1H NMR (CDCl3): = 0.95 (dt, 2 H, J = 4.2, 7.2 Hz), 1.1 (dt, 2 H,
J = 8.0, 4.4 Hz), 1.3 (t, 6 H, J = 7.0 Hz), 2.18 (tt, 1 H, J = 7.2, 7.2,
4.0, 4.0 Hz), 3.18 (d, 2 H, J = 23.25 Hz), 4.12 (dq, 4 H, J = 7.0 Hz).
13C NMR (CDCl3): = 2 11.96 (t), 16.46 (dq, JCP = 5.13 Hz),
21.68 (d), 43.54 (dq, JCP = 127.86 Hz), 62.66 (dt, JCP = 5.38 Hz),
202.02 (s).
31P NMR (CDCl3): = 21.03.
MS: m/z = 220 (M+), 191 (M+ – CH2CH3), 175 (M+ – OCH2CH3),
151 [CH2P(O)(OEt)2], 69 [M+ – CH2P(O)(OEt)2].
HRMS: m/z Calcd for C9H17O4P: 220.0864476. Found:
220.0853000.
2i
(9) (a) Corey, E. J.; Kwiatkowski, G. T. J. Am. Chem. Soc. 1966,
88, 5653. (b) Corey, E. J.; Kwiatkowski, G. T. J. Am. Chem.
Soc. 1968, 90, 6816. (c) Mikolajczyk, M.; Balczewski, P.
Synthesis 1984, 691.
(10) Savignac, P.; Mathey, F. Tetrahedron Lett. 1976, 2829.
(11) Exceptionally, these side reactions were not observed with
diethyl -chloromethylphosphonate, due to a complete
enolization of the resulting -chloro- -ketophosphonate in
the reaction medium: Teulade, M. P.; Savignac, P.;
Aboujaoude, E. E.; Collignon, N. J. Organomet. Chem.
1985, 287, 145.
Colorless thick oil.
1H NMR (CDCl3): = 1.1 (s, 9 H), 1.25 (t, 6 H, J = 6 Hz), 3.08 (d,
2 H, J = 22 Hz), 4.08 (dq, 4 H, J = 6 Hz).
13C NMR (CDCl3): = 16.18 (dq, JCP = 5.56 Hz), 3 25.93 (q),
35.69 (dq, JCP = 134.77 Hz), 45.00 (s), 62.16 (dt, JCP = 5.59 Hz),
207.00 (s).
31P NMR (CDCl3): = 21.88.
MS: m/z = 236 (M+), 191 (M+ – OCH2CH3), 179 (M+ – t-C4H9), 152
[CH3P(O)(OEt)2], 151 [CH2P(O)(OEt)2].
Anal. Calcd for C10H21O4P: C, 50.85; H, 8.90. Found: C, 51.01; H,
8.98.
(12) Mathey, F.; Savignac, P. Tetrahedron 1978, 34, 649.
(13) (a) Corbel, B.; L’Hostis-Kervella, I.; Haelter, J. P. Synth.
Commun. 1996, 26, 2561. (b) Kim, D. Y.; Kong, M. S.;
Kim, T. H. Synth. Commun. 1996, 26, 2487. (c) Corbel, B.;
L’Hostis-Kervella, I.; Haelthers, J. P. Synth. Commun. 2000,
30, 609.
(14) Coutrot, P.; Grison, C.; Lachgar, M.; Ghribi, A. Bull. Soc.
Chim. Fr. 1995, 132, 925.
(15) (a) Lee, K.; Wiemer, D. F. J. Org. Chem. 1991, 56, 5556.
(b) An, J. G.; Wiemer, D. F. J. Org. Chem. 1992, 57, 317.
(c) An, Y. Z.; An, J. G.; Wiemer, D. F. J. Org. Chem. 1994,
59, 8197.
2j
Colorless thick oil.
1H NMR (CDCl3): = 1.28 (t, 6 H, J = 5.9 Hz), 4.0 (d, 2 H, J = 23.5
Hz), 4.15 (dq, 4 H, J = 5.9 Hz), 7.48 (dd, 1 H, J = 1.15, 5.75 Hz),
7.83 (d, 1 H, J = 2.3, 8.05 Hz), 8.06 (d, 1 H, J = 9.2 Hz), 8.69 (d, 1
H, J = 4.6 Hz).
13C NMR (CDCl3):
= 16.40 (dq, J = 5.48 Hz), 36.18 (dq,
JCP = 130.53 Hz), 62.60 (dt, J = 5.43 Hz), 122.44 (d), 127.59 (d),
137.16 (d), 149.15 (d), 153.07 (s), 193.89 (s).
31P NMR (CDCl3): = 21.35.
(16) Hong, S.; Chang, K.; Ku, B.; Oh, D. Y. Tetrahedron Lett.
1989, 30, 3307.
(17) Boeckman, R. K.; Walters, M. A.; Koyano, H. Tetrahedron
Lett. 1989, 30, 4787.
Synthesis 2002, No. 12, 1683–1688 ISSN 0039-7881 © Thieme Stuttgart · New York

1688
F. Orsini et al.
PAPER
substitution of a phosphine ligand by a carbonyl donor
group. Reaction between the two organic moieties bound to
the cobalt center would then give the product together with
a Co(II) species from which the Co(0) precursor would be
regenerated by reduction with Mg metal.
(18) (a) Orsini, F.; Pelizzoni, F.; Pulici, M.; Vallarino, L. M. J.
Org. Chem. 1994, 59, 1. (b) Orsini, F. Tetrahedron Lett.
1998, 39, 1425.
(19) Orsini, F. J. Org. Chem. 1997, 62, 1159.
(20) Orsini, F.; Pulici, M.; Vallarino, L. M. J. Organomet. Chem.
1995, 495, C1.
(21) Diethyl methanephosphonate may form both from acid-base
exchange between the -ketophosphonate and the -
metallated phosphonate (as observed for lithium
phosphonate) or from hydrolysis, most likely during
workup, from an unreacted organometallic reagent (as
observed in Reformatsky-type reactions).
(22) This behaviour would suggest a mechanism involving the
oxidative addition of the -halophosphonate to the cobalt(0)-
complex to give a cobalt(II) species, as well as the
(23) Mg metal is used in excess and can be recycled after use.
(24) Mg was activated by adding a few drops of 1,2-
dichloroethane to the metal turnings in THF so that a
vigorous reaction ensued. After few minutes the mixture was
cooled in an ice-bath, the solvent was removed and the Mg
was washed with THF (3 2 mL).
(25) The solution can be also warmed with an external bath at
about 50 °C to speed the reduction of Co(II) to Co(0).
Warming may be useful when Ph3P is used instead of Me3P.
Synthesis 2002, No. 12, 1683–1688 ISSN 0039-7881 © Thieme Stuttgart · New York