McKenna reaction - Which oxygen attacks bromotrimethylsilane?

Article abstract of DOI:10.1021/jo4021612

The first experimental proof of the course of silylation in the McKenna reaction, one of the most widely used reactions for the synthesis of organophosphorus acids, is presented. The reaction (in acetonitrile) proceeds via an attack of the terminal oxygen from the dialkyl phosphonate on the silicon atom in bromotrimethylsilane. Isotopically enriched diethyl phenylphosphonates (Pi - ¹⁷O or Pi - ¹⁸O) were used as the model compounds. The location of the isotopic tracers was detected using ³¹P and ¹⁷O NMR spectroscopy.

Full text of DOI:10.1021/jo4021612

Note
pubs.acs.org/joc
McKenna ReactionWhich Oxygen Attacks Bromotrimethylsilane?
Katarzyna M. Błazewska*
̇
Institute of Organic Chemistry, Faculty of Chemistry, Lodz University of Technology, Zeromskiego 116, 90-924 Lodz, Poland
S
* Supporting Information
ABSTRACT: The first experimental proof of the course of
silylation in the McKenna reaction, one of the most widely used
reactions for the synthesis of organophosphorus acids, is presented.
The reaction (in acetonitrile) proceeds via an attack of the terminal
oxygen from the dialkyl phosphonate on the silicon atom in
bromotrimethylsilane. Isotopically enriched diethyl phenylphos-
17
18
phonates (P O or P O) were used as the model
compounds. The location of the isotopic tracers was detected
using 31P and 17O NMR spectroscopy.
he McKenna reaction1,2 is one of the most widely used
attacks silicon, leading to phosphonium-like intermediates 2a
and 4a (Scheme 2, mechanism A). Earlier, Rabinowitz had
proposed this mechanism for CTMS-assisted dealkylation of
dialkyl phosphonates.5 Chojnowski, Michalski, and co-workers
studied mechanistic aspects of the reaction of ITMS and BTMS
with thio- and selenoesters of phosphorus.7 However, in a
recent kinetics study demonstrating that the BTMS reaction is
second order, an alternate mechanism (Scheme 2, mechanism
B) was postulated, involving nucleophilic attack on silicon by a
bridging oxygen.11
In the mechanism proposed by McKenna (Scheme 2,
mechanism A),2 the terminal oxygen of compound 1 attacks
the silicon atom, expelling bromide anion from BTMS and
leading to the formation of a phosphonium-like intermediate
2a. Next, PO is recreated from one of the alkoxy groups
upon attack of the bromide anion on the carbon neighboring a
bridging oxygen, with formation of ethyl bromide. The new
PO then attacks another BTMS molecule to form 4a, which
upon an attack of bromide anion on the alkoxy carbon is
transformed into bis(trimethylsilyl) phosphonate 5a.
T
methods in the synthesis of organophosphorus acids. The
application of commercially available bromotrimethylsilane
(BTMS) ensures mild and efficient conditions for the
transformation of pentavalent organophosphorus alkyl esters
into trimethylsilyl esters, which are easily cleavable to free acids
by the action of protic solvents.2,3 Although it originally was
applied only to dialkyl phosphonates, the method is also useful
for dealkylation of phosphates and phosphinates. According to
the Reaxys database,4 BTMS and hydrochloric acid are the
most often used reagents for dealkylation of organophosphorus
esters. In the last two years (2012−2013), BTMS was used in at
least half of such cases. Recent applications of other
trimethylsilyl halides, chlorotrimethylsilane (CTMS)5 and
iodotrimethylsilane (ITMS),6,7 for dealkylation of organo-
phosphorus esters are less frequent.8
Two steps are involved in the BTMS-assisted synthesis of
phosphonic acids (Schemes 1 and 2). In the first one,
silyldealkylation takes place, converting dialkyl phosphonate
ester 1 into bis(trimethylsilyl) phosphonate 5 (Scheme 2). In
the second step, the silyl ester product 5 is subjected to
solvolysis with the use of a protic solvent (Scheme 1).9 The
solvolysis step proceeds via nucleophilic attack on silicon by the
oxygen atom of a protic solvent molecule, leaving a bridging
oxygen from the ester in the molecule of the product acid 7
(Scheme 1). This was confirmed by Rabinowitz, who
determined that the action of an alcohol on bis(trimethylsilyl)
phosphonate resulted in the formation of a mixed ether, 8.5
Later on, Orlov and co-workers applied labeled water-18O for
solvolysis, and by using IR the authors detected isotopically
labeled derivative 8 without formation of the labeled
phosphonic acid.10
A proposed alternative mechanism (Scheme 2, mechanism
B) involves the bridging oxygen of 1, which attacks the silicon
atom, resulting in the formation of intermediate 2b. The
bromide anion then attacks the carbon neighboring the
positively charged oxygen in 2b, forming ethyl bromide. The
second alkoxy group (via 4b) then undergoes an analogous
transformation, and the final product 5b is obtained.
Here, the first experimental proof of the mechanism of
silylation in the McKenna reaction is presented. 31P and 17O
NMR spectroscopy was used to determine which oxygen atom
in the phosphonate moiety attacks silicon by following the fate
17
18
of the labeled terminal oxygen (P O and P O). Readily
available diethyl phenylphosphonate (11) was selected as a
In contrast to the solvolysis step, the mechanism of the
silylation is relatively unexamined. In their first report, in which
BTMS silyldealkylation was carried out without added solvent,
McKenna et al.2 suggested that the terminal oxygen (PO)
Received: September 29, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jo4021612 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Note
Scheme 1. Mechanism of Solvolysis of Bis(trimethylsilyl) Phosphonate
a
Scheme 2. Proposed Courses of the Silylation Reaction
a
The terminal oxygen is marked as a black oval to differentiate the two plausible mechanisms.
model compound, since it is a typical representative of
phosphonates, the most abundantly studied class of organo-
phosphorus esters. 17O-enriched 11c and its bis(trimethylsilyl)
analogue 12c have been fully characterized previously by 17O
NMR spectroscopy, with the chemical shifts of both the
bridging and terminal oxygens being determined.12 Such data
make this ester an ideal reference compound for the studies
presented here.
Diethyl phenylphosphonite (10) was prepared from
dichloro(phenyl)phosphine (9).13 Labeled and unlabeled
diethyl phosphonates 11a−c were prepared from 10 in the
oxidation reaction with water and iodine.14 This was a vital
synthesis for these studies, since it enabled the introduction of
an oxygen atom in the terminal PO position.15 Depending
on the water used, phosphonates 11a−c (Scheme 3) were
obtained with high purity in 60−88% yield upon bulb-to-bulb
distillation.
To get good 31P NMR resolution of the reaction mixture, the
use of pyridine (as a proton scavenger) was necessary.
Otherwise, an unresolved broad singlet in the 31P NMR
spectrum was observed (Figure S8 in the Supporting
Information). This phenomenon might be associated with the
presence of residual HBr in BTMS, leading to chemical
exchange of a proton between the oxygen and proton
donor.16,17
18O is not magnetically active, but when it is directly linked
with 31P, it induces a small upfield shift in the 31P NMR
resonances18,19 whose magnitude depends on, for example, the
bond order.20,21 The effect of isotopic substitution on the
magnetic shielding of nuclei was predicted in 1952,22 and since
then it has been widely applied for studies of the mechanisms of
a broad spectrum of chemical reactions.20 Since in this case all
of the oxygen atoms are directly connected to the 31P nucleus,
the bond order between phosphorus and oxygen was used as a
probe, allowing determination of the reaction mechanism.
Depending on the reaction course, the phosphorus−oxygen
bond order would change with PO attack center (mechanism
A) or would remain the same if the bridging oxygen were
responsible for the attack on the silicon atom (mechanism B).
It was reported that the difference between the chemical shifts
a
Scheme 3. Synthesis of Isotopically Labeled Substrate 11
16
18
of pentavalent phosphorus P O and P O should be in
the range 0.043−0.052 ppm, whereas for P−16O−R/P−18O−R
the difference should be smaller (0.016−0.032 ppm).20
In the first experiment, a mixture of diethyl 16O-phosphonate
ester 11a and 18O-phosphonate ester 11b was monitored by 31P
NMR, giving a chemical shift difference of 0.045 ppm between
11a and 11b (Figure 1, on the left), corresponding to the value
observed for analogous phosphonates labeled at the PO
oxygen.20 Upon the action of BTMS, the chemical shift
difference dropped to 0.028 ppm (Figure 1, on the right). This
result confirms the phosphonium-like mechanism of the
McKenna reaction, where the terminal oxygen attacks the
silicon atom, changing the bond order from PO to P−O−
SiMe3 (Scheme 2, mechanism A).
a
Silylation conditions are shown in the red rectangle. The labeled
oxygen is marked as a black oval only for substrate 11.
Silylation of the generated phosphonates 11 was carried out
in an NMR tube in dry CD3CN at room temperature. Spectra
were recorded before and after addition of BTMS. The progress
of the reaction was monitored by 31P NMR spectroscopy.
Exchange of one ethyl ester group for one trimethylsilyl residue
shifts the phosphonate signal upfield by ∼9 ppm, which is a
characteristic marker for the progress of the McKenna reaction.
B
dx.doi.org/10.1021/jo4021612 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Note
Figure 1. Superimposed 31P NMR spectra of mixtures (molar ratio 1/0.8) of substrates 11a and 11b (blue, on the left) and products 12a and 12b
(red, on the right).
Among the three naturally occurring oxygen isotopes, only
5
17O possesses a nuclear spin (I = /2). It is quadrupolar, which
leads to rapid relaxation of the 31P nucleus directly linked with
17O and broadening of the line width.20 Still, 17O NMR analysis
of P−O-containing compounds is appropriate for probing their
structure, bonding, and dynamics23 and is a valuable tool in
some stereochemical analyses.24 17O NMR spectroscopy allows
differentiation between the terminal PO and bridging P−O−
R oxygens in organophosphorus derivatives12 on the basis of
their peak shapes (usually a well-resolved doublet for the PO
oxygen vs a broad peak for the P−O−R oxygen) or chemical
shifts.12 Application of 17O NMR in titration studies showed
that protonation of the phosphonate oxygen is accompanied by
an upfield shift ∼50 ppm per charge neutralized.25
17
In the experiment with P O-labeled phosphonate 11c,
al.,12 in the case of phosphonate esters, substituting an alkyl
17O NMR was used as the analytical tool. According to Dahn et
Figure 2. Superimposed 17O NMR spectra of 11c (in blue) and 12c
(in red).
group with a trimethylsilyl group shifts the signals of the
terminal and bridging oxygens downfield by the same
by 31P and 17O NMR spectroscopies. The methodology
presented here should be applicable to studies of other
trimethylsilyl halogenides (CTMS and ITMS) as well as to
other organophosphorus esters that are prone to BTMS-
magnitude of 20−26 ppm. The chemical shifts for the terminal
oxygen in 11c and the bridging oxygen in bis(trimethylsilyl)
ester 12c were almost the same under the conditions applied.12
Here, upon addition of BTMS, the signal of oxygen originally
assisted dealkylation. It could be also useful to evaluate how
17
present in P O shifted downfield by only ∼4 ppm,
different reaction conditions may influence the reaction
suggesting that the phosphorus−oxygen bond order changed
mechanism.
from PO into P−O. This conclusion is strongly supported by
the change in the peak shape from a well-resolved doublet for
EXPERIMENTAL DETAILS
■
11c into a broad “singlet” for 12c (Figure 2).
Diethyl phenylphosphonite (10),13 diethyl phenylphosphonate (11),26
To sum up, the first experimental proof of the mechanism of
and bis(trimethylsilyl) phenylphosphonate (12)27 were previously
the silylation in the McKenna reaction is presented. The results
show that it is the PO oxygen in diethyl phenylphosphonate
that is responsible for the attack on the silicon atom in BTMS.
Thanks to isotopic labeling of the terminal oxygen, the change
in the phosphorus−oxygen bond order was directly observed
synthesized. Here compounds 11 were characterized by 31P, 17O, and
1H NMR, and compounds 12 were characterized in the crude reaction
mixture by 31P, H, and 17O NMR. NMR spectra of substrates 11a−c
1
were acquired at 250.1 MHz for 1H NMR and 101 MHz for 31P NMR
in CDCl3, unless otherwise stated. NMR spectra of reaction mixtures
C
dx.doi.org/10.1021/jo4021612 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Note
before and after addition of BTMS were measured at 700.0 MHz for
1H NMR, 283.4 MHz for 31P NMR, and 94.9 MHz for 17O NMR.
Chemical shifts (δ) are reported in parts per million (ppm) relative to
ACKNOWLEDGMENTS
■
The author thanks Mr. Grzegorz Ciepielowski for recording
700 MHz spectra, Ms. Joanna Gmach for assistance with some
of the experiments, and Prof. Tadeusz Gajda and Prof. Stefan
Jankowski for helpful discussions. Financial support by the
Ministry of Science and Higher Education in Poland (IP2011
003771) is gratefully acknowledged.
1
(a) in H NMR, internal residual CH3CN in CD3CN (δ 1.96) or
internal residual CHCl3 in CDCl3 (δ 7.26); in 31P NMR, external 85%
H3PO4 (0 ppm); and in 17O NMR, external residual D217O in D2O (0
ppm). Experiments were performed in NMR tubes, dried with the heat
gun, and stored above P2O5 under vacuum. CD3CN was dried with
freshly activated 3 Å molecular sieves. Pyridine was distilled and dried
with 3 Å molecular sieves. BTMS was distilled under nitrogen and
stored in sealed ampules at −20 °C. Water-17O (20−24.9 atom % 17O)
was acquired from Sigma-Aldrich and was also enriched with 18O
(20.56 atom % 17O, 15.65 atom % 18O, 63.78 atom % 16O). Water-18O
(98%) was acquired from Sercon (0.2 atom % 17O, 98.6 atom % 18O,
1.2 atom % 16O). The instrumental settings were as follows. For 17O
NMR: FIDRES, 0.072385 Hz; AQ, 6.9 s; D1, 2s; LB, 15 Hz. For
determination of the 18O isotopic shift in 31P NMR: a spectral width of
11312.2 Hz and TD = 65536 gave high spectral resolution (FIDRES:
0.1726 Hz); AQ, 2.9 s; D1, 2 s; LB, 0.5 Hz. Compound 10 was
prepared according to the reported procedure in 75% yield.13
REFERENCES
■
(1) Thottathil, J. In Handbook of Organophosphorus Chemistry; Engel,
R., Ed.; Marcel Dekker: New York, 1992; pp 61−62.
(2) McKenna, C. E.; Higa, M. T.; Cheung, N. H.; McKenna, M. C.
Tetrahedron Lett. 1977, 18, 155−158.
(3) McKenna, C. E.; Schmidhauser, J. J. Chem. Soc., Chem. Commun.
1979, 739.
(4) Reaxys is a chemical database that is owned and protected by
Reed Elsevier Properties SA. The search profile (Sept 2013) is
presented in Scheme S1 in the Supporting Information.
(5) Rabinowitz, R. J. Org. Chem. 1963, 28, 2975−2978.
(6) Zygmunt, J.; Kafarski, P.; Mastalerz, P. Synthesis 1978, 609−612.
(7) Borecka, B.; Chojnowski, J.; Cypryk, M.; Michalski, J.; Zielinska,
J. J. Organomet. Chem. 1979, 171, 17−34.
(8) ITMS is a more vigorous reagent than BTMS, often requiring
lower temperatures of application (e.g., −20 °C) to avoid side
reactions. In some cases (e.g., thiophosphonates) it successfully
replaces BTMS in reactions with analogues requiring harsher
conditions for dealkylation.7
(9) The ease of deprotection of bis(trimethylsilyl) phosphonates by
solvolysis was shown earlier by Rabinowitz.5 In their later works,
McKenna and co-workers showed the importance of pH and
buffering29 and the possibility of controlled deprotection of other
groups by the action of the deprotected phosphonic acid.30
(10) D’yakov, V. M.; Voronkov, M. G.; Orlov, N. F. Izv. Akad. Nauk
SSSR, Ser. Khim. 1972, 2484−2488.
(11) Conibear, A. C.; Lobb, K. A.; Kaye, P. T. Tetrahedron 2010, 66,
8446−8449. Curiously, these authors attributed mechanism B to ref. 2
although it is not mentioned there.
Diethyl Phenylphosphonate (11). Diethyl phenylphosphonite
(0.3 g, 1.51 mmol) was dissolved in THF (6 mL). After addition of
pyridine (0.25 mL), the solution was cooled to −45 °C, and two
liquids, appropriately labeled water (0.09 mL, 3 equiv)28 and iodine
(0.45 g, 1.77 mmol, 1.17 equiv) dissolved in THF, were added
simultaneously via separate syringes until the yellow color persisted.
After addition, the temperature was maintained at −45 to −35 °C for
10 min and then at rt for 10 min. The reaction was then quenched
with a saturated solution of Na2S2O3, and THF was evaporated. The
residue was suspended in diethyl ether (50 mL), washed with 0.1 M
HCl (5 × 1 mL) and water (3 × 1 mL), and then dried over MgSO4.
The resultant crude product was distilled bulb-to-bulb at 135 °C/0.1
mmHg. The final products were obtained in good yields (11a, 60%;
11b, 83%; 11c, 88%). 1H NMR (250 MHz, CDCl3, on the example of
3
11c): δ 1.32 (t, JHH = 7.10 Hz, 2CH3CH2O, 6H); 3.99−4.23 (m,
2CH3CH2O, 4H); 7.42−7.59 (m, 3Har); 7.76−7.86 (m, 2Har). 31P
NMR (101 MHz, CDCl3): δ 19.78. 17O NMR (94.9 MHz, CD3CN): δ
96.78 (d, JOP = 158.83).26
1
Bis(trimethylsilyl) Phenylphosphonate (12). The appropriately
labeled diethyl phenylphosphonate (15 mg scale; 20% 17O for 17O
NMR experiments or 45% 18O for 31P NMR experiment) was
dissolved in dry acetonitrile-d3 (0.6 mL), placed in an NMR tube, and
subjected to the action of BTMS (5 equiv). After storage over the
weekend at rt without stirring, NMR spectra were recorded and
showed full conversion of the diethyl ester into the expected product
12. NMR spectra were recorded for the crude reaction mixture. Below
only signals confirming structure 12 are presented. For the spectra of
reaction mixtures (with signals from additional compounds described),
see Figures S13 and S19 in the Supporting Information. 1H NMR (700
MHz, CD3CN, on the example of 12c): δ 0.43 (bs, 2Si(CH3)3,
18H);27 7.50−7.53 (m, 2Har); 7.59−7.61 (m, 1Har); 7.73−7.77 (m,
2Har). 31P NMR (283.35 MHz, CD3CN): δ 1.12 (12a);27 1.09 (12b);
1.12 and 1.09 in a 3.7:1 molar ratio (12c). 17O NMR (94.91 MHz,
CD3CN): δ 100.8 (very broad singlet).
(12) Dahn, H.; Toan, V. V.; Ung-Truong, M.-N. Magn. Reson. Chem.
1992, 30, 1089−1096. It should be noted that the spectroscopic
conditions used were slightly different.
(13) Coskran, K. J.; Jenkins, J. M.; Verkade, J. G. J. Am. Chem. Soc.
1968, 90, 5437−5442.
(14) Bentrude, W. G.; Sopchik, A. E.; Gajda, T. J. Am. Chem. Soc.
1989, 111, 3981−3987.
(15) Letsinger, R. L.; Lunsford, W. B. J. Am. Chem. Soc. 1976, 98,
3655−3661.
(16) Other explanations may be possible.
(17) The use of pyridine before addition of BTMS resulted in N-
alkylation of pyridine with ethyl bromide, which forms upon the attack
of bromide anion on phosphonium intermediates 2 and 4 (Scheme 2).
This side reaction did not have any influence on the silylation reaction
but could be avoided by adding the pyridine after the silylation was
completed. According to the experience of the author and her co-
workers, such N-alkylation is sometimes encountered when the
McKenna reaction is applied to organophosphorus esters containing a
heterobase residue. However, it can be easily circumvented by running
the reaction in an atmosphere of flowing neutral gas, which removes
the relatively volatile alkyl bromide from the reaction mixture.31
(18) Cohn, M.; Hu, A. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 200−
203.
(19) Lowe, G.; Sproat, B. S. J. Chem. Soc., Chem. Commun. 1978,
565−566.
(20) Cullis, P. M. In Phosphorus-31 NMR Spectral Properties in
Compound Characterization and Structural Analysis; Quin, L. D.,
Verkade, J. G., Eds.; Wiley-VCH: Weinheim, Germany, 1994; pp 315−
332.
ASSOCIATED CONTENT
■
S
* Supporting Information
Copies of 31P NMR, 1H NMR, and 17O NMR spectra of diethyl
phenylphosphonates 11 and crude reaction mixtures before and
after addition of BTMS. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: katarzyna.blazewska@p.lodz.pl.
Notes
The author declares no competing financial interest.
D
dx.doi.org/10.1021/jo4021612 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Note
(21) Wolfsberg, M.; Van Hook, W. A.; Paneth, P.; Rebelo, L. P. N.
Isotope Effects in the Chemical, Geological, and Bio Sciences; Springer:
Dordrecht, The Netherlands, 2010; pp 203−244.
(22) Ramsey, N. F. Phys. Rev. 1952, 87, 1075−1079.
(23) Gerothanassis, I. P. Prog. Nucl. Magn. Reson. Spectrosc. 2010, 56,
95−197.
(24) Sammons, R. D.; Frey, P. A.; Bruzik, K.; Tsai, M.-D. J. Am.
Chem. Soc. 1983, 105, 5455−5461.
(25) Gerlt, J. A.; Reynolds, M. A.; Demou, P. C.; Kenyon, G. L. J. Am.
Chem. Soc. 1983, 105, 6469−6475.
(26) Kalek, M.; Ziadi, A.; Stawinski, J. Org. Lett. 2008, 10, 4637−
4640.
(27) Kiddle, J. J.; Gurley, A. F. Phosphorus, Sulfur Silicon Relat. Elem.
2000, 160, 195−205.
(28) In the case of water-16O, 14 equiv were used. In the case of
water-18O (98 atom % 18O) and water-17O (20−24.9 atom % 17O), 3
equiv were used.
(29) Gross, H.; Keitel, I.; Costisella; McKenna, C. E. Phosphorus,
Sulfur Silicon Relat. Elem. 1991, 61, 177−181.
(30) Marma, M. S.; Khawli, L. A.; Harutunian, V.; Kashemirov, B. A.;
McKenna, C. E. Bioorg. Med. Chem. Lett. 2004, 126, 1467−1475.
McKenna, C. E.; Levy, J. N. J. Chem. Soc., Chem. Comm. 1989, 246.
(31) Joachimiak, Ł.; Janczewski, Ł.; Błazewska, K. M. Unpublished
̇
results.
E
dx.doi.org/10.1021/jo4021612 | J. Org. Chem. XXXX, XXX, XXX−XXX