Dramatic effect of the metal cation in dealkylation reactions of phosphinic esters promoted by complexes of polyether ligands with metal iodides

Article abstract of DOI:10.1021/jo0056388

Metal ion electrophilic catalysis has been revealed in dealkylation reactions of phosphinic esters 1-4 promoted by complexes of polyether ligands 5-7 with metal iodides MI_n (Mⁿ⁺ = Li⁺, Na⁺, K⁺, Rb⁺, Ca²⁺, Sr²⁺, Ba²⁺) in low polarity solvents (chlorobenzene, 1,2-dichlorobenzene, and toluene) at 60 °C. The catalytic effect increases with increasing the Lewis acid character of the cation, in the order Rb⁺ < K⁺ < Na⁺ < Li⁺ and Ba²⁺ < Sr²⁺ < Ca²⁺. The results are interpreted in terms of a transition state where the complexed cation (Mⁿ⁺ ? Lig) assists the departure of the leaving group Ph₂P(O)O^- and, at the same time, favors the attack at carbon of the nucleophile I^- ( push-pull mechanism). The rate sequence found for 1-4 (Me > Et ? i-Pr and t-Bu) shows that this reaction can be utilized for the selective dealkylation of these substrates.

Full text of DOI:10.1021/jo0056388

VOLUME 66, NUMBER 10
MAY 18, 2001
© Copyright 2001 by the American Chemical Society
Articles
Dr a m a tic Effect of th e Meta l Ca tion in Dea lk yla tion Rea ction s of
P h osp h in ic Ester s P r om oted by Com p lexes of P olyeth er Liga n d s
w ith Meta l Iod id es
Domenico Albanese, Dario Landini, and Angelamaria Maia*
Centro CNR and Dipartimento di Chimica Organica e Industriale,
Via Venezian 21, I-20133 Milano, Italy
angelamaria.maia@unimi.it
Received September 7, 2000
Metal ion electrophilic catalysis has been revealed in dealkylation reactions of phosphinic esters
1-4 promoted by complexes of polyether ligands 5-7 with metal iodides MIn (Mn+ ) Li+, Na+, K+,
Rb+, Ca2+, Sr2+, Ba2+) in low polarity solvents (chlorobenzene, 1,2-dichlorobenzene, and toluene)
at 60 °C. The catalytic effect increases with increasing the Lewis acid character of the cation, in
the order Rb+< K+< Na+< Li+ and Ba2+< Sr2+< Ca2+. The results are interpreted in terms of a
transition state where the complexed cation (Mn+ ⊂ Lig) assists the departure of the leaving group
Ph2P(O)O- and, at the same time, favors the attack at carbon of the nucleophile I- (“push-pull”
mechanism). The rate sequence found for 1-4 (Me > Et . i-Pr and t-Bu) shows that this reaction
can be utilized for the selective dealkylation of these substrates.
In tr od u ction
etc. preferentially attack the phosphorus, more polariz-
able nucleophiles, like bromides or iodides, yield products
mainly arising from nucleophilic attack at carbon.3 The
selectivity obtained in solution can be, however, com-
pletely reversed in the gas phase where, in the absence
of any solvation, it is possible to measure the intrinsic
reactivity of carbon versus phosphorus.3
In recent years, increasing interest has been focused
on the production of compounds, analogues of the nucleic
acids, with modified internucleosidic bonds such as alkyl
phosphotriesters and methylphosphinates of oligonucle-
otides.1 In the synthesis of oligonucleotides with the
phosphotriester method, the key step is often represented
by the final removal of the protective group, usually a
methyl group, from the phosphoric function of the
substrate.2 The selectivity of this reaction, expressed as
the ratio between the attack at phosphorus (phosphory-
lation) and at carbon (dealkylation), largely depends on
the nucleophile used. Whereas nucleophiles with a high
charge density such as hydroxides, alkoxides, enolates,
By contrast, the effect of the counterion was only
qualitatively reported in a few publications and just for
divalent cations.4,5 Later on, Buncel et al. revealed and
quantitatively evaluated for the first time the participa-
tion of the metal cation (“electrophilic catalysis”) in
reactions of nucleophilic substitution at phosphorus of
phosphoric esters by ethoxides6 and phenoxides7 of alkali
(1) Koole, L. H.; Moody, H. M.; Broeders, N. L.; Quaedflieg, P. J .;
Kuijpers, W. H.; van Genderen, M. H.; Coenen, A. J .; van der Wal, S.;
Buck, H. M. J . Org. Chem. 1989, 54, 1657.
(2) Smith, D. J .; Ogilvie, K. K.; Gillen, M. F. Tetrahedron Lett. 1980,
21, 861.
(3) Lum, R. C.; Grabowski, J . J . J . Am. Chem. Soc. 1992, 114, 8619.
(4) Steffens, J . J .; Sampson, E. J .; Siewers, I. J .; Benkovic, S. J . J .
Am. Chem. Soc. 1973, 95, 936.
(5) Wadsworth, W. S., J r. J . Org. Chem. 1981, 46, 4080.
(6) Dunn, E. J .; Buncel, E. Can. J . Chem. 1989, 67, 1440.
10.1021/jo0056388 CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/21/2001

3250 J . Org. Chem., Vol. 66, No. 10, 2001
Albanese et al.
metals in ethanol. The catalytic effect depends on the
metal ion, increasing in the order K+ < Na+ , Li+. More
recently, we found an analogous behavior in nucleophilic
substitution reactions of alkylsulfonates promoted by
complexes of crown ethers and polyethylenglycols with
metal salts in chlorobenzene.8,9
In the present work, a kinetic study of the effect of the
metal cation on the dealkylation rate of alkyl diphe-
nylphosphinates 1-4 by complexes of polyether ligands
5-7 with alkali and alkaline-earth metal iodides in low
polarity media (chlorobenzene,1,2-dichlorobenzene, and
toluene) is reported. A comparative study of the corre-
sponding uncatalyzed reactions with bulky quaternary
onium iodides has also been included.
F igu r e 1. Second-order plot for the reaction of methyl diphen-
ylphosphinate 1 (0.025 M) with (Na+ ⊂ PHDB18C6)I- (0.0095
M) in chlorobenzene, at 60 °C. Y ) [(Na+ ⊂ PHDB18C6)I-].
Resu lts a n d Discu ssion
Kinetics were performed by reacting comparable
amounts of alkyl diphenylphosphinates 1-4 (0.006-0.03
M) and preformed (Mn + ⊂ Lig)nI- complex (0.002-0.014
M) in the appropriate solvent (chlorobenzene, 1,2-dichlo-
robenzene, toluene) at 60 ( 0.1 °C (eq 1).
Ta ble 1. Secon d -Or d er Ra te Con sta n ts k (M-1 s-1) for
th e Dea lk yla tion of Alk yl Dip h en ylp h osp h in a tes
P h 2P (O)OR (R ) Me, Et, i-P r , t-Bu ) 1-4 by Com p lexes
(M+ ⊂ Lig)I- (Lig ) 5 a n d 7) in Ch lor oben zen e, a t 60 °Ca
c
F
103kb
-∆Gcat
R
M+ (Å-1
)
ligand
(M-1 s-1
)
krel
(kcal mol-1
)
1
1
1
1
Me
Me
Me
Me
Rb+ 0.67
K+
0.75
5
5
5
5
7
5
5
5
0.029
1
3.8
4.3
5.8
7.3
0.061(0.23)d
2.1
Na+ 1.02
0.63
5.6
21.7
193
Li+
Li+
1.28
0.084
0.064
e
2
3
4
Et
1.28
1.28
1.28
i-Pr Li+
t-Bu Li+
e
a
A chlorobenzene solution (30 mL) of substrates 1-4 (0.02-
0.03M) and (M+ ⊂ Lig)I- complex (0.009-0.014 M). Average of
b
at least two determinations. The error in these values is estimated
to be 5%. c ∆Gcat ) ∆Gq - ∆Gq
) - RT ln (kcat/kuncat); see
cat
uncat
d
Figure 2 for determination of kuncat
.
Value obtained by using
.
e No reaction after 2 weeks.
+
NH4
Ta ble 2. Effect of Solven t P ola r ity on th e Secon d -Or d er
Ra te Con sta n t of th e Dem eth yla tion of 1 w ith (M+
⊂
Rates have been measured by potentiometric titration
of the complexed iodide and/or by gaschromatographic
determination of the unreacted substrate 1-4 evaluated
with respect to an internal standard (tetradecane). Under
these conditions reactions follow a regular second-order
kinetic eq 2 up to at least 3 half-life times (Figure 1).
Results are reported in Tables 1-3.
Lig)I- Com p lexes, a t 60 °Ca
rate ) k[substrate][complexed I-]
(2)
a
A solution (30 mL) of substrate (0.015-0.03M) and (M+
⊂
Lig)I- complex (0.005-0.013Μ). Average of at least two deter-
b
In the case of K+, Rb+, and NH4+, the reaction 1 goes
to the equilibrium (3) already for low conversion values
(<20%). The equilibrium can be fully shifted toward the
formation of the dealkylation product by addition, in the
reaction medium, of equimolar quantities, with respect
to the (Mn+ ⊂ Lig)nI- complex, of triethylamine.
minations. The error is estimated to be 5%. c See ref 12. The
d
extent of complexation is very low (e2%).
The latter, in fact, reacts with the alkyl iodide formed
to give the corresponding quaternary ammonium salt
Et3N+RI- that quantitatively precipitates during the
reaction. Blank experiments performed in parallel have
shown that the solubility of Et3N+RI- is negligible in the
organic medium.
For having a comparison with the uncatalyzed reac-
tion, the dealkylation (1) has been carried out with the
lipophilic quaternary ammonium salt hexyl4N+I- in the
presence of an equimolar amount of triethylamine. As
expected, the reaction is very slow, and a conversion
<10% was found after about 1 month. On the other hand,
the use of complexable NH4+I- showed a reduced but still
detectable catalysis (Table 1).
(7) Dunn, E. J .; Moir, R. Y.; Buncel, E.; Purdon, J . G.; Bannard, R.
A. Can. J . Chem. 1990, 68, 1837.
(8) Gobbi, A.; Landini, D.; Maia, A.; Secci, D. J . Org. Chem. 1995,
60, 5954.
(9) Gobbi, A.; Landini, D.; Maia, A.; Penso, M. J . Chem. Soc., Perkin
Trans. 2 1996, 2505 and references therein.

Electrophilic Catalysis in Dealkylation of Phosphinic Esters
J . Org. Chem., Vol. 66, No. 10, 2001 3251
Ta ble 3. Secon d -Or d er Ra te Con sta n ts k (M-1 s-1) for
th e Dem eth yla tion of Meth yl Dip h en ylp h osp h in a te 1 by
Com p lexes (M2+ ⊂ P HDB18C6)2I- in Ch lor oben zen e a t
60 °Ca
Sch em e 1
b
M2+
F (Å-1
)
103 k (M-1 s-1
)
krel
Ba2+
Sr2+
Ca2+
1.40
1.57
1.89
0.46
6.24
61.7
1
13.6
134
a
A chlorobenzene solution (30 mL) of substrate (0.006-0.008M)
An analogous mechanism was proposed by Buncel in
reactions of alkali metal ethoxides with sulfonic and
phosphinic esters.10 The higher the interaction of the
complexed cation (M+ ⊂ Lig) with the leaving group, the
higher must be the reactivity of the iodide and hence the
higher the corresponding rate constant. As shown in
Table 1, in the demethylation reaction 1 with the series
of alkali iodides the second-order rate constants k (M-1
s-1) progressively increase, up to about 200 times, in the
b
and complexed iodide (0.002-0.005M). Average of at least two
determinations. The error is estimated to be e10%.
+
+
+
+
order kRb < kK < kNa , kLi . Such a sequence is in line
with the increasing interaction of the metal ion with the
-
leaving group Ph2PO2 on increasing the Lewis acidity
of the cation (Rb+ < K+ < Na+ < Li+).11
The electrophilic catalysis is also confirmed by the
excellent (r ) 0.9999) linear correlation found by plotting
the second-order rate constants k (M-1 s-1) of the reaction
1 as a function of the charge density F of the cation
(Figure 2).
In addition, the comparison with the extrapolated
value of the rate constant for the uncatalyzed reaction
kuncat (M-1 s-1) (Figure 2) allows a rough evaluation of
the entity of electrophilic catalysis observed in the
reaction 1. Interestingly, the catalytic effect ∆Gcat, ex-
pressed as decrease of the activation free energy ∆∆Gq,
is particularly remarkable in this case, passing from 3.8
kcal/mol for Rb+ up to more than 7 kcal/mol in the case
of Li+ cation where the interaction with the phosphinic
oxygens reaches its maximum value (Table 1).
An analogous behavior has been obtained when
the demethylation reaction 1 was performed with the
complexes of crown ether 5 with alkaline-earth metal
iodides MI2 (M2+ ) Ca2+, Sr2+, Ba2+). Also in this case,
the rate constant noticeably increases (up to more
than two powers of 10) by increasing the charge
density of the cation in the order Ba2+ < Sr2+ < Ca2+
(Table 3). The transition state for divalent cations is
substantially the same as that proposed for the mono-
valent ions M+ (Scheme 1) and involves only one anion
at a time, as confirmed by the reaction order found: first-
order both in substrate 1 and in complexed iodide
(M2+ ⊂ PHDB18C6)2I-.
F igu r e 2. Correlation between second-order rate constant k
(M-1 s-1) and charge density F (Å-1) of the cation in the
demethylation reaction of methyl diphenylphosphinate 1
with complexes (M+ ⊂ PHDB18C6)I-, where M+ ) Li+, Na+,
K+, Rb+, in chlorobenzene (r ) 0.9999, kuncat) 9.35 × 10-8
(M-1 s-1) extrapolated value for F ) 0).
In addition, the dealkylation rate strongly depends on
the alkyl substituent R (R ) Me, Et, i-Pr, t-Bu). As
reported in Table 1, the cation being the same, the
second-order rate constant k (M-1 s-1) decreases about 2
orders of magnitude on changing from the methyl 1 to
the corresponding ethyl 2 derivative, and no reaction at
all was observed in the case of the corresponding isopro-
pyl 3 and tert-butyl 4 phosphinates, even after prolonged
times. A very good linear correlation (r ) 0.9999) was
obtained in the demethylation of 1 by plotting the second-
order rate constants k (M-1 s-1) measured for the series
of alkali cations M+ (Li+, Na+, K+, Rb+) vs the corre-
sponding charge densities F, obtained by dividing the
charge z of the cation for its ionic radius r (Figure 2).
An analogous correlation was also obtained in the case
of the alkaline-earth cations Ca2+, Sr2+, and Ba2+. For
these divalent cations, however, the corresponding second-
order rate constants are lower than those expected from
their charge density values (Tables 1 and 3).
The lower reactivity of the latter complexes if compared
with that expected from their charge density values
(Tables 1 and 3) can likely be attributed to a higher
energy of interaction of the divalent cation with the anion
within the ion pair in the ground state. A very similar
trend was previously observed in nucleophilic substitu-
tion reactions promoted by cryptates of metal iodides in
low polarity media and explained in the same way.9
The variation of reactivity on changing the solvent does
not indicate any substantial modification in the transition
state with the polarity (Table 2). The cation being the
same, the sequence found, k1,2-dichlorobenzene < kchlorobenzene
Kinetic data as a whole evidence the participation of
the metal cation in the dealkylation reaction of alkyl
diphenylphosphinates 1 and 2 promoted by complexes of
polyethers PHDB18-crown-6 5 and PEG400Me2 6 with
metal iodides in low polarity media (eq 1).
Results can be explained in terms of a transition state
where the complexed cation (M+ ⊂ Lig) can interact with
both the ion-paired nucleophile I- as well as with the
-
leaving group Ph2PO2 (Scheme 1).
In the activation process, the ion-pair (M+ ⊂ Lig)I-
reacts following a concerted “push-pull” mechanism
where the complexed metal cation (M+ ⊂ Lig) stabilizes
the developing negative charge on the phosphinic oxygens
and at the same time favors the nucleophilic attack of
the iodide at the alkylic carbon (“electrophilic catalysis”).
(10) Pregel, M. J .; Buncel, E. J . Org. Chem. 1991, 56, 5583 and
references therein.
(11) Pearson, R. G. Coord. Chem. Rev. 1990, 100, 403.

3252 J . Org. Chem., Vol. 66, No. 10, 2001
Albanese et al.
Ma ter ia ls a n d Solven ts. Tetrahexylammonium iodide,
ligands PHDB18C6 5, PEG 400Me2 6 and [2.2.2,C10] 7 were
commercial products, utilized as purchased. Metal and am-
monium iodides were Analar grade commercial products used
as such after dehydration of the salt in an oven, at 110-120
°C, under vacuum for several hours. In all cases the water
content is e0.05 mol of H2O per mol of salt (Karl Fischer
titration). Dry (Fluka) chlorobenzene, 1,2-dichlorobenzene and
toluene (H2O e 15 ppm) were used. Alkyl diphenylphosphi-
nates Ph2P(O)OR 1-4 were prepared by reacting diphenyl-
phosphinyl chloride and the corresponding alcohol in the
presence of triethylamine in CH2Cl2 following a previously
reported procedure.16
< ktoluene, reflects the increasing nucleophilic reactivity
of the iodide on diminishing the polarity of the medium.12
Con clu sion s
The results as a whole reveal the fundamental role
played by the cation Mn+ (“electrophilic catalysis”) and
the alkyl group in determining the dealkylation rate of
alkyl diphenylphosphinates 1-4 by complexes of poly-
ethers 5 and 6 with alkali and alkaline-earth metal
iodides in low polarity solvents. The catalytic effect
increases with increasing the Lewis acid character of the
cation and is mainly related to the ability of the ligand
to effectively shield the metal ion charge.
It is worth noting that cyclic polyethers such as crown
ether PHDB18crown6 5 or open chain as PEG400Me2 6,
that are anion activators less efficient than macrocyclic
cryptands such as [2.2.2,C10] 7,13,14 play a major role in
cation-assisted reactions. Besides solubilizing inorganic
salts in low polar media, these ligands form stable
inclusion complexes where the metal cation, even if
partially shielded by the ethereal oxygens, still keeps a
remarkable density of charge and hence can interact with
the leaving group Ph2PO2- in the transition state (Scheme
1). Metal ion electrophilic catalysis largely overcomes,
particularly for high charge density cations, the lower
anion activation realized by ligands 5 and 6 as experi-
mentally confirmed by the comparison with cryptand
[2.2.2,C10] 7 (Table 1).
Data for 1 (R ) Me): white hygroscopic solid; mp 56-58 °C
[lit.16 bp 178 °C (2.4 mmHg)]; 1H NMR δ 3.75 (d, J ) 11.1 Hz,
3H), 7.30-7.80 (m, 10H); 31P NMR δ 33.7.
Data for 2 (R ) Et): colorless oil; bp 170-172 °C (1.5 mmHg)
1
[lit.17 bp 172 °C (1.5 mmHg)]; H NMR δ 1.37 (t, J ) 7.0 Hz,
3H), 4.11 (dt, J PH ) J HH ) 7.0 Hz, 2H), 7.30-7.96 (m, 10H);
31P NMR δ 31.9.
Data for 3 (R ) i-Pr): white solid; mp 97-99 °C (lit.16 mp
97-99 °C); 1H NMR δ 1.35 (d, J ) 6.2 Hz, 6H), 4.60-4.70 (m,
1H), 7.28-7.90 (m, 10H); 31P NMR δ 30.9.
Data for 4 (R ) t-Bu): white solid mp 110-112 °C (lit.16
1
mp 111-112 °C); H NMR δ 1.53 (s, 9H), 4.60-4.70 (m, 1H),
7.42-7.81 (m, 10H); 31P NMR δ 26.3.
Kin etic Mea su r em en ts. In a typical procedure, a stan-
dardized solution (15 mL) of substrate 1-4 (0.012-0.07 M)
was added to a standardized solution (15 mL) of complexed
MI or MI2 (0.004-0.03 M) in a 50 mL flask thermostated at
60 ( 0.1 °C. Samples (2-10 mL), withdrawn periodically, were
quenched in ice-cold MeOH (50 mL), and the unreacted
nucleophile I- was potentiometrically titrated with 0.01 N
AgNO3.
The variations of rate constant obtained with these
complexed cations (M+ ⊂ Lig) in chlorobenzene (k(K
/
+
⊂Lig)
+
+⊂Lig)
k(Na ⊂Lig)/k(Li
) 1:10:92) are even higher than those
found by other authors in analogous demethylation
reactions with metal ions M+ in acetone (kK /kNa /kLi
)
+
+
+
Alternatively, a standardized solution (10 mL) of substrate
1 (0.0030-0.04 M), tetradecane as internal standard (0.02 M),
and anhydrous triethylamine (0.02 M) was added to a stan-
dardized solution (10 mL) of preformed complex or hexyl4N+I-
(0.02 M). Samples (0.5-1 mL), withdrawn periodically, were
quenched with an aqueous saturated solution of AgNO3 (1 mL)
and analyzed by GC. The mass balance was g97% in all cases.
1:1.6:9).15 In dipolar aprotic (acetone) and protic (MeOH)
media, where metal salts are soluble as such, metal ions
M+ in equilibrium with ion pairs M+Y- can directly
interact with the substrate in the transition state. In
these systems, the addition of the polyether, which
selectively binds the cation15 or destroys the ion pair,7
inhibits electrophilic catalysis and is most likely the main
reason for the decelerating effect observed.
The second-order rate constants were evaluated using a
least-squares computer program from the eq 1/([B0] - [A0]) ln
([B][A0])/[A][B0]) ) kt, where A ) substrate and B ) complexed
iodide or hexyl4N+I- or vice versa. All rates involved at least
eight samplings and gave correlation coefficients of 0.995 or
better.
Finally, the sequence of reactivity obtained for alkyl
diphenylphosphinates 1-4 with the complexes of LiI
(Me > Et . i-Pri and t-Bu) (Table 1) clearly indicates
that, from a practical point of view, this reaction can be
utilized with success for the selective demethylation of
these substrates.
The solutions of preformed complexes of 5-7 were prepared
by magnetically stirring a standardized solution (30 mL) of
ligand in the organic solvent (chlorobenzene, 1,2-dichloroben-
zene or toluene) (0.01-0.03 M) with the appropriate quantity
of salt MI or MI2 (0.2-1.3 mol per mol of ligand) as a solid
phase, in a flask thermostated at 60 ( 0.1 °C. The heteroge-
neous system was stirred for 1-3 h and then kept without
stirring for an additional 10 min to allow good separation of
the two phases. Aliquots (5-8 mL) of the organic phase were
centrifuged and samples (2-5 mL) were withdrawn and
titrated with 0.01 N AgNO3.
Exp er im en ta l Section
Meth od s. 1H and 31P NMR spectra were recorded in CDCl3
on Bruker AC 300 and AMX 300 spectrometers using CHCl3
and aqueous 85% H3PO4 as external references. Potentiometric
titrations were performed with a Metrohm 670 titroprocessor
by using a combined silver electrode isolated with a potassium
nitrate bridge. Karl Fischer determinations were carried out
with a Metrohm 684 KF coulometer. GC data were obtained
with a Perkin-Elmer 8310 equipped with a 50 × 1/8 in
OV-101-5% on chromosorb WHP 100/120 mesh column.
Ack n ow led gm en t. We gratefully acknowledge the
support of this research by CNR and MURST.
J O0056388
(12) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry,
2nd ed.; Verlag-Chemie: Weinheim, 1988.
(13) Maia, A. Pure Appl. Chem. 1995, 67, 697.
(14) Landini, D.; Maia, A.; Penso, M. Comprehensive Supramolecular
Chemistry; Lehn, J . M., Ed.; Pergamon Press: Oxford, 1996; Vol. 1,
Chapter 11, pp 417-464.
(16) Berlin, K. D.; Austin, T. H.; Nagabhushanam, M. J . Org. Chem.
1965, 30, 1267.
(17) Allen, D. W.; Hutley, B. G.; Meller, M. T. J . Chem. Soc., Perkin
Trans. 2 1977, 1705.
(15) Modro, A. M.; Modro, T. A. Can. J . Chem. 1993, 71, 469.