Mono-N-methylation of functionalized anilines with alkyl methyl carbonates over NaY faujasites. 4. Kinetics and selectivity

Article abstract of DOI:10.1021/jo048076r

(Chemical Equation Presented) In the presence of NaY faujasite as the catalyst, the reaction of bifunctional anilines (1-4: XC₆H ₄-NH₂; X = OH, CO₂H, CH₂OH, and CONH₂) with methyl alkyl carbonates [MeOCO₂R′: R′ = Me or MeO(CH₂)₂O(CH₂)₂] proceeds with a very high mono-N-methyl selectivity (XC₆H ₄NHMe up to 99%), and chemoselectivity as well, with other nucleophilic functions (OH, CO₂H, CH₂OH, CONH₂) fully preserved from alkylation and/or transesterification reactions. Aromatic substituents, however, modify the relative reactivity of amines 1-4: good evidence suggests that, not only steric and electronic effects, but, importantly, direct acid-base interactions between substituents and the catalyst are involved. Weakly acidic groups (OH, CH₂OH, CONH₂, pK_a ≥ 10) may help the reaction, while aminobenzoic acids (pK _a of 4-5) are the least reactive substrates. The solvent polarity also affects the reaction, which is faster in xylene than in the more polar diglyme. The mono-N-methyl selectivity is explained by the adsorption pattern of reagents within the zeolite pores: a B_Al2 displacement of the amine on methyl alkyl carbonate should occur aided by the geometric features of the NaY supercavities. Different factors account for the reaction chemoselectivity. Evidence proves that the polarizability of the two nucleophilic terms (NH ₂ and X groups) of anilines is relevant, although adsorption and confinement phenomena of reagents promoted by the zeolite should also be considered.

Full text of DOI:10.1021/jo048076r

Mono-N-methylation of Functionalized Anilines with Alkyl Methyl
Carbonates over NaY Faujasites. 4. Kinetics and Selectivity
Maurizio Selva,* Pietro Tundo, and Tommaso Foccardi
Dipartimento di Scienze Ambientali dell’Universita` Ca’ Foscari, Calle Larga S. Marta 2137,
30123, Venezia, Italy
selva@unive.it
Received October 29, 2004
In the presence of NaY faujasite as the catalyst, the reaction of bifunctional anilines (1-4: XC6H4-
NH2; X ) OH, CO2H, CH2OH, and CONH2) with methyl alkyl carbonates [MeOCO2R′: R′ ) Me or
MeO(CH2)2O(CH2)2] proceeds with a very high mono-N-methyl selectivity (XC6H4NHMe up to 99%),
and chemoselectivity as well, with other nucleophilic functions (OH, CO2H, CH2OH, CONH2) fully
preserved from alkylation and/or transesterification reactions. Aromatic substituents, however,
modify the relative reactivity of amines 1-4: good evidence suggests that, not only steric and
electronic effects, but, importantly, direct acid-base interactions between substituents and the
catalyst are involved. Weakly acidic groups (OH, CH2OH, CONH2, pKa g 10) may help the reaction,
while aminobenzoic acids (pKa of 4-5) are the least reactive substrates. The solvent polarity also
affects the reaction, which is faster in xylene than in the more polar diglyme. The mono-N-methyl
selectivity is explained by the adsorption pattern of reagents within the zeolite pores: a BAl2
displacement of the amine on methyl alkyl carbonate should occur aided by the geometric features
of the NaY supercavities. Different factors account for the reaction chemoselectivity. Evidence proves
that the polarizability of the two nucleophilic terms (NH2 and X groups) of anilines is relevant,
although adsorption and confinement phenomena of reagents promoted by the zeolite should also
be considered.
SCHEME 1. M ) K, Na; X ) CN, CO2Me; Y Fau:
Alkali Metal-Exchanged Y Faujasites
Introduction
In the past decade, the need for safer and more
selective processes has fueled a growing interest for
dimethyl carbonate (MeOCO2Me, DMC), as a methylat-
ing agent.1 DMC, in fact, is a nontoxic compound2 that
allows an unprecedented high selectivity (up to 99%, at
complete conversion) in the monomethylation of both
CH2-active compounds and primary aromatic amines
(Scheme 1).3,4
In particular, we recently reported that the combined
use of DMC and sodium-exchanged Y-zeolite (NaY) is also
a valuable protocol for the N-methylation of functional-
ized anilines such as amino-phenols (1), -benzoic acids
(2), -benzyl alcohols (3), and -benzamides (4) (Scheme 2).5
(1) (a) Shaik, A.-A. G.; Sivaram, S. Chem. Rev. 1996, 96, 951-976.
(b) Tundo, P.; Selva, M. Acc. Chem. Res. 2002, 35, 706-716.
(2) (a) Romano, U.; Rivetti, F.; Di Muzio, N. U.S. Pat. 4,318,862,
1981, C.A. 80141, 1979. (b) Rivetti, F.; Romano, U.; Delledonne, D. In
Green Chemistry: Designing Chemistry for the Environment; Anastas,
P., Williamson, T. C., Eds.; ACS Symposium Series; American Chemi-
cal Society: Washington, DC, 1996; Vol. 626, pp 70-80.
(3) (a) Selva, M.; Marques, C. A.; Tundo, P. J. Chem. Soc., Perkin
Trans. I 1994, 1323-1328. (b) Fu, Z.-H.; Ono, Y. J. Catal. 1994, 145,
166-70. (c) Tundo, P.; Selva, M.; Bomben, A. Org. Synth. 1999, 76,
169-177. (d) Selva, M.; Tundo, P.; Perosa, A.; Memoli, S. J. Org. Chem.
2002, 67, 1071-1077.
10.1021/jo048076r CCC: $30.25 © 2005 American Chemical Society
Published on Web 03/10/2005
2476
J. Org. Chem. 2005, 70, 2476-2485

Mono-N-methylation of Functionalized Anilines
SCHEME 2. X ) OH, CO2H, CH2OH, CONH2
This reaction not only shows a very high mono-N-
methyl selectivity (up to 99%), but it proceeds with a
complete chemoselectivity toward the amino group, with
the other nucleophilic functionalities (OH, CO2H, CH2-
OH, CONH2) fully preserved from alkylation and/or
transesterification reactions. Moreover, with respect to
conventional methylation procedures, which require sto-
ichiometric bases and harmful reagents (methyl halides
or dimethyl sulfate), the reaction of Scheme 2 takes place
with catalytic amounts of NaY and neither organic nor
inorganic wastes need to be disposed of.6
With the aim of further exploring the features of this
process, we wish to report here a kinetic analysis of the
methylation of compounds 1-4 (Scheme 3) with either
DMC or 2-(2-methoxyethoxy)methyl ethyl carbonate
[MeO(CH2)2O(CH2)2OCO2Me, MEC] as methylating
agents, and NaY faujasite as a catalyst.
FIGURE 1. The mono-N-methylation of o-aminophenol with
DMC, at 90 °C (Q ) 2): disappearance of 1a (f) and formation
of its mono-N-methyl derivative (9).
Results
This study shows how the presence of aromatic sub-
stituents OH, CH2OH, CO2H and CONH2 modifies the
relative methylation rate of amines 1-4, plausibly by
altering the adsorption of these substrates within the
pores of the zeolite used. Some general conclusions on
the role of steric hindrance and of interactions of aniline
substituents with the zeolite cages are also drawn from
the reaction of some O-protected bifunctional amines
(XC6H4NH2: X ) OMe, 6a,b; CO2Me, 7a,b; CO2Et, 8a,b;
Scheme 4) with DMC and MEC, in the presence of NaY.
The overall investigation has been developed through
the following lines: (i) at first, the order and the regime
of the reaction were evaluated to validate the kinetic
examination; (ii) then, pseudo-first-order rate constants
were obtained for the reaction of DMC with amines 1 and
3,4, and for the reaction of MEC with amines 1-4
(including aminobenzoic acids 2a,b), and with amino-
thiophenols (ortho- and para-isomers, respectively); (iii)
finally, pseudo-first-order rate constants were obtained
for the reaction of MEC with anisidines and esters of
aminobenzoic acids (6-8).
1. Mono-N-methylation of o-Aminophenol (1a).
Order and Regime. o-Aminophenol (1a) was initially
chosen as a model compound. Solutions of 1a in DMC
(6.5 × 10-2 M, 42 mL; DMC serving both as the methy-
lating agent and as the solvent) were made to react at
90 °C, under a N2 atmosphere, in the presence of different
amounts of the faujasite NaY as the catalyst [weight ratio
NaY:1a (Q) in the range of 0.5-3], and of triglyme
[CH3O(CH2O)2(CH2)2OCH3, 0.56 mmol] as the internal
standard. Both the disappearance of o-aminophenol and
the formation of its N-methyl derivative (o-HOC6H4-
NHMe) were monitored by GLC and GC/MS. In all cases,
no other products were observed, except for traces (<1%)
of o-(N,N-dimethylamino)phenol at conversions >90%. As
an example, Figure 1 reports the reaction run with Q )
2.
From these data, in accordance with the treatment
already described by us for the methylation of aniline
over NaY zeolite,4d pseudo-first-order rate constants (kobs
)
for the decay of 1a could be evaluated at different Q
ratios.7 The results are reported in Table 1.
SCHEME 3
SCHEME 4
J. Org. Chem, Vol. 70, No. 7, 2005 2477

Selva et al.
FIGURE 2. Mono-N-methylation of 1a with DMC and NaY (Q ) 2): (A) the effect of the stirring rate on kobs; (B) apparent
activation energy from the Arrhenius equation.
TABLE 1. Pseudo-First-Order Rate Constants for the
Mono-N-methylation of o-Aminophenol with DMC in the
Presence of Different Amounts of NaY (Q)
(kobs) tended to an asymptotic value provided that the
stirring rate was above 970 rpm (Figure 2A), and (iii)
the apparent activation energy calculated from the
Arrhenius equation (ln kobs ) ln A - Eatt/RT; Figure 2B)
was 93.5 kJ mol-1. The A factor was of 2.68 × 10-2 (see
Experimental Section for details).
2. a. Mono-N-methylation of ortho- and para-
Isomers of Aminophenols (1a,b), Aminobenzyl Al-
cohols (3a,b), and Aminobenzamides (4a,b) with
DMC. Two sets of experiments (a and b) were performed.
(a) In the first, the reactivity of ortho-substituted
compounds (3a and 4a) was explored under the condi-
tions above-described for o-aminophenol.10 Solutions of
3a and 4a in DMC (6.5 × 10-2 M, 42 mL) were made to
react at 90 °C in the presence of the faujasite NaY [for
convenience, the weight ratio NaY:substrate (Q) was set
at 2, see Experimental Section]. Triglyme [CH3O(CH2O)2-
(CH2)2OCH3, 0.63 and 0.47 mmol for 3a and 4a, respec-
tively] was the internal standard.
(b) In the second, to compare rate constants for both
para- and ortho-derivatives, reaction conditions were
slightly modified: because p-aminophenol (1b) and p-
aminobenzamide (4a) were sparingly soluble in DMC,
1,2-dimethoxyethane (DME) was used as a cosolvent.11
Accordingly, solutions of compounds 1a,b, 3a,b, and 4a,b
in a mixture of DMC/DME (4.2 × 10-2 M, 42 mL,
DMC:DME ) 4 v:v), were made to react at the reflux of
DME (84 °C), under a N2 atmosphere, in the presence of
NaY (Q ) 2) and of an internal standard (triglyme, 1a,b,
3a, and 4a,b; hexadecane, 3b).
The disappearance of the reagents and the formation
of the corresponding N-methyl derivatives (XC6H4NHMe;
X ) OH, CH2OH, CONH2) were monitored by GLC and
GC/MS. In all cases, pseudo-first-order rate constants
(kobs) for the decay of the substrates were obtained as
described for 1a in Table 1.12
entry
Qa
kobs × 10-3 (min-1
)
1
2
3
4
5
0.5
1.0
1.6
2.0
3.0
6.7
13.6
20.7
24.5
34.0
a Q: weight ratio NaY:1a.
General criteria recommended for heterogeneously
catalyzed processes8 were then applied to evaluate the
regime under which the reaction occurred: both the effect
of the catalyst weight and of the stirring rate on the
kinetic constant (kobs), and the activation energy of the
reaction were considered.
In particular, it was noted that: (i) a linear relation-
ship held between kobs of Table 1 and the Q ratio,9 and,
under the conditions of Figure 1, (ii) the reaction rate
(4) (a) Fu, Z.-H.; Ono, Y. Catal. Lett. 1993, 22, 277-81. (b) Selva,
M.; Bomben, A.; Tundo, P. J. Chem. Soc., Perkin Trans. 1 1997, 1041-
1045. (c) Selva, M.; Tundo, P.; Perosa, A. J. Org. Chem. 2001, 66, 677-
680. (d) Selva, M.; Tundo, P.; Perosa, A. J. Org. Chem. 2002, 67, 9238-
9247.
(5) Selva, M.; Tundo, P.; Perosa, A. J. Org. Chem. 2003, 68, 7374-
7378.
(6) The coproduct MeOH (Scheme 2) can, in principle, be recycled
to the preparation of the carbonate itself.
(7) If one considers that the concentration of the DMC (reagent/
solvent) is a constant during the reaction, a simplified rate law
equation for the disappearance of 1a can be used: v ) -(d[1a]/dt) )
k‚[1a]p‚[DMC]q ) k′[1a]p, where k′ ) k[DMC]q. We observed that,
throughout the range of Q, the concentration profiles of 1a (determined
by GC) fitted the integrated rate law with p ) 1 (ln[1a]o/[1a] ) k′t,
where [1a]o and [1a] were the concentrations of 1a at t ) 0 and at a
later time t, respectively). The resulting plots gave straight lines (r >
0.99), whose slopes were the pseudo-first-order rate constants kobs, for
the disappearance of the substrate. According to the above equation,
the evaluation of the order with respect to DMC is not possible: this
would require a large excess of the amine and, consequently, the
presence of a solvent. Under such conditions, compound 1a as well as
all the other amines 1-4 would dissolve only in polar solvents (i.e.,
alcohols, DMF, DMSO, sulfolane, etc.), which strongly adsorb on the
zeolite and deactivate it (see ref 4d and references therein, as well as
the discussion section).
(8) (a) Boudart, M.; Diega-Mariadasson, G. In Kinetics of Hetero-
geneous Catalytic Reactions; Prausnitz, J. M., Brewer, L., Eds.;
Princeton University Press: Princeton, NJ, 1984. (b) Alcorn, W. R.;
Sullivan, T. J. In Catalysis of Organic Reactions; Kosak, J. R., Ed.;
Marcel Dekker Inc.: New York, 1984. (c) Satterfield, C. N. Mass
Transfer in Heterogeneous Catalysis; MIT Press: Cambridge, MA,
1970.
(9) See Figure 1 in the Supporting Information. It should be noted,
however, that saturation effects of the catalyst by both aromatic amines
and dialkyl carbonates should be expected. We already observed these
effects in (i) the reaction of aniline with MEC (ref 4d), and (ii) the
reaction of aniline with a large excess of DMC (ref 4b).
(10) At 90 °C, both anthranilic and p-aminobenzoic acids (2a,b) did
not react with DMC in the presence of NaY (ref 5). Therefore, they
were not considered at this stage.
(11) The choice of DME was based on results previously reported
by us (refs 4d and 5).
(12) The Supporting Information reports the example of the mono-
N-methylation of p-amino benzamide (4b).
2478 J. Org. Chem., Vol. 70, No. 7, 2005

Mono-N-methylation of Functionalized Anilines
TABLE 2. Mono-N-methylation of o- and p-Aminophenols (1a,1b), o- and p-Aminobenzyl Alcohols (3a,3b), and o- and
p-Aminobenzamides (4a,4b)
a Entries 1-3: experiments were carried out at 90 °C using DMC as a solvent. Entries 4-9: experiments were carried out at 84 °C
c
using a mixture of DMC/DME (4:1 v/v) as a solvent. b Reaction conversion, % by GC. SM/D% ) XC6H4NHCH3/(XC6H4NHCH3
+
XC6H4N(CH3)2) × 100 is the mono-N- to N,N-dimethyl selectivity.
SCHEME 5. 2-(2-Methoxyethoxy)methylethyl
Carbonate, MEC
however, reacted with DMC only at T g 130 °C, in
autoclaves with internal pressures up to 6-7 bar.5,10
Because these conditions were not suitable for a kinetic
study, an asymmetrical alkyl methyl carbonate, 2-(2-
methoxyethoxy)methylethyl carbonate (MEC, Scheme 5),
was chosen as the methylating agent.
MEC in fact, was easily prepared from DMC, and it
allowed us to run both O- and N-methylation reactions
with 99% chemoselectivity, at temperatures up to 160
°C, and at atmospheric pressure.4c,d,13
Two sets of experiments (c and d) were performed.
In (c), solutions of ortho-substituted compounds 1a-
4a in xylene14 (7 × 10-2 M, 30 mL) were made to react
at 135 °C, in the presence of MEC (in a 10 molar excess
with respect to the reactant amine), of the faujasite NaY
[weight ratio NaY:substrate (Q) of 2], and of octadecane
as the internal standard (Scheme 6a).
In (d), both ortho- and para-substituted compounds
(1a,b, 2a,b, and 4a,b, respectively) were made to react
with MEC under the same conditions described for set
c, except for the solvent used: because p-aminobenzoic
acid (2b) and p-aminobenzamide (4b) were not soluble
in xylene, a glycol-derived dimethyl ether (DME-like)
such as diglyme [CH3O(CH2)2O(CH2)2OCH3, bp 162 °C]
was chosen (Scheme 6b).
FIGURE 3. Competitive mono-N-methylation of o- and p-
aminophenols with DMC/DME.
Table 2 reports kobs for the investigated compounds.
The table also shows the mono-N- to N,N-dimethylselec-
tivity (SM/D%), which is always very high (from 89% to
99%) even at substantially quantitative conversions.
SM/D% defines also the overall reaction selectivity because
no other products were observed except XC6H4NHMe and
XC6H4NMe2.
To further compare the reactivity between ortho- and
para-isomers, also a competitive reaction between o- and
p-aminophenol (1a,b) was carried out using a solution
of equimolar amounts of the two isomers (2.1 × 10-2
M
in each substrate) in DMC/DME (42 mL; DMC/DME )
4 v/v), in the presence of NaY [NaY/(1a + 1b) ) 2 weight
ratio]. The result is reported in Figure 3.
For both sets c and d, in accordance to the procedure
described for Tables 1 and 2,7,15 pseudo-first-order rate
2. b. Mono-N-methylation of ortho- and para-
Isomers of Aminophenols (1a,b), Aminobenzoic
Acids (2a,b), Aminobenzyl Alcohols (3a,b), and
Aminobenzamides (4a,b) with 2-(2-Methoxyethoxy)-
methylethyl Carbonate (MEC). To complete the ki-
netic analysis for amines 1-4, the mono-N-methylation
of acids 2a,b had to be explored. Compounds 2a,b,
(13) Perosa, A.; Selva, M.; Tundo, P.; Zordan, F. Synlett 2000, 1,
272-274.
(14) The adsorption of hydrocarbons over zeolites takes place to a
much lesser extent with respect to more polar compounds. Thus, when
possible, apolar media (i.e., xylene) are excellent solvents for liquid-
phase reactions catalysed by faujasites. See refs 4d and: (a) Espeel,
P. H. J.; Vercruysse, K. A.; Debaerdemaker, M.; Jacobs, P. A. Stud.
Surf. Sci. Catal. 1994, 84, 1457. (b) Jayat, F.; Sabater Picot, M. J.;
Guisnet, M. Catal. Lett. 1996, 41, 181.
J. Org. Chem, Vol. 70, No. 7, 2005 2479

Selva et al.
TABLE 3. The Reactions of Amines 1-4 with MEC in the Presence of NaY
a Entries 1-4: experiments were carried out at 135 °C using xylene as a solvent. Entries 5-10: experiments were carried out at 135
°C using diglyme as a solvent. b Conversions, SM/D% (as defined in Table 2), and kobs were averaged over two runs. c The overall selectivity
toward the wanted product (o-CH2OHC6H4NHMe) was of only 62% due to the occurrence of side-reactions.
SCHEME 6. The Methylation of Amines 1-4 with
MEC
SCHEME 7. The Reaction of Aminothiophenols
with MEC: Distribution (%) of Products
Determined by GC/MS
constants (kobs) were evaluated for the disappearance of
reagents 1-4. Results are reported in Table 3.
It should be noted that in xylene as a solvent, kobs
values related to o-aminophenol and o-aminobenzyl
alcohol (1a and 3a, respectively) are missing (entries 1
and 3). In these cases, the kinetic investigation was
precluded because of two different reasons: (i) the
reaction of 1a took place too rapidly (10 min, entry 1),
and (ii) the concentration profile of 3a was altered with
respect to a first-order decay, by the onset of side-
reactions. In particular, the formation of three high-
boiling byproducts in a total amount of 35% was observed
at complete conversion of the substrate: two of them
reasonably derived from the N-alkylation of 3a [HO-
CH2C6H4NHR and HOCH2C6H4N(CH3)R, R ) CH3O-
(CH2)2O(CH2)2: molecular ions M+ of 225 and 239,
respectively], while the structure of the third compound
(M+ ) 281) was not assigned. This lack of selectivity
suggested to us that we not investigate the reactivity of
aminobenzyl alcohols (3a,b) in diglyme.
2. c. Methylation of ortho- and para-Isomers of
Aminothiophenols (5a,b) with MEC. Under the con-
dition of Table 3 (entries 1-4), also the reaction of other
ambident nucleophiles such as aminothiophenols (HSC6H4-
NH2; ortho and para, compounds 5a and 5b, respectively)
with MEC was explored. The results are reported in
Scheme 7.
3. The Reaction of MEC with Anisidines and
Methyl and Ethyl Aminobenzoates. Finally, to fur-
ther examine the role of aromatic substituents, the
methylation of some O-protected bifunctional aromatic
amines (6a,b, 7a,b, and 8a,b) with MEC was investi-
gated (Scheme 8). All reactions were carried out under
the conditions of entries 1-4 of Table 3. At 135 °C,
solutions of compounds 6-8 in xylene (7 × 10-2 M, 30
mL) were made to react with MEC (in a 10 molar excess
(15) The concentration of MEC (molar ratio MEC:substrate ) 10)
was assumed to be constant during the reactions.
2480 J. Org. Chem., Vol. 70, No. 7, 2005

Mono-N-methylation of Functionalized Anilines
SCHEME 8. Methylation of Protected Anilines 6-8 with MEC over NaY
with respect to the reactant amine), in the presence of
the faujasite NaY [weight ratio NaY:substrate (Q) of 2],
and of an internal standard (0.35 mmol: dodecane for
6a,b, tetradecane for 7a,b, hexadecane for 8a,b).
The course of reactions was monitored by GLC-GC/
MS: only mono-N- and di-N,N-methyl derivatives of
compounds 6-8 were observed. According to the proce-
dure described for Tables 1-3, pseudo-first-order rate
constants (kobs) were evaluated for the disappearance of
reagents. The results are reported in Table 4.
(1) are the more reactive substrates, and a general
reactivity scale can be recognized (Scheme 9).
SCHEME 9
The same trend holds at 135 °C with MEC as the
methylating agent (Table 3). For instance, in either
xylene or diglyme as solvents, the reaction of 1a proceeds
6-8 times faster with respect to amide 4a (entries 1 and
4, 5 and 9).17 Phenols 1a,b are also more reactive than
acids 2a,b (entries 1 and 4, 5 and 7, 6 and 8).
Reasonably, the electron-donating and steric properties
of the OH substituent may account for both an increase
in the nucleophilicity and a reduced hindrance around
the NH2 group, of 1 with respect to compounds 2, 3, and
4 as well.
TABLE 4. The Reaction of Compounds 6-8 with MEC
over NaY
The role of steric hindrance seems readily appreciable
also from data of Table 4: for both anisidines (entries 1
and 2) and methyl and ethyl esters of aminobenzoic acids
(entries 4,5 and 8,9, respectively), para-isomers are by
far more reactive than the ortho ones; also, kobs values of
methyl p-aminobenzoate and methyl anthranilate (en-
tries 4 and 5) are higher than those of the corresponding
ethyl esters (entries 8 and 9).
The Role of H-Bonding. Steric and electronic effects,
however, do not explain other relevant aspects of the
reactivity of amines 1-4. In particular, (i) although the
CONH2 group shows a higher electron-withdrawing
capacity,18 and a moderately larger bulkiness with re-
spect to the CH2OH substituent (Table 3 of SI), the
reaction of DMC with amides is faster than with alcohols,
b
a Reaction conversion, % by GC. SM/D (%) as defined in Tables
1-3. c Already reported in Table 3. d nd: not determined (see entry
1 of Table 4). e After 33 h, a conversion of 73% was obtained with
a SM/D of 94%.
especially in the case of o-substituted compounds (k4a
≈
For a direct comparison, Table 4 also indicates the rate
constants for the methylation of o-aminophenol, anthra-
nilic acid, and anthranilamide (compounds 1a, 2a, and
4a, respectively) obtained under the same reaction condi-
tions (entries 1, 2, and 4, Table 3).
2k3a, entries 2 and 3, Table 2); and (ii) more surprisingly,
a striking difference of reactivity is evident between acids
2 and amides 4: at 135 °C with MEC, the ratio of rate
constants for ortho-derivatives (k4a/k2a) is 16 (entries 2
and 4, Table 3), and it is even higher for para-reagents
2b and 4b (k4b/k2b ) 24; entries 8 and 10, Table 3). This,
despite CO2H and CONH2 substituents possess medium-
to-strong electron-withdrawing effects and a similar
steric hindrance (Table 3 of SI).
Discussion
The Regime of the Reaction. Data of Table 1 and
Figure 2A, and the apparent activation energy of 93.5
kJ mol-1 (Figure 2B) well above the threshold of 25 kJ
mol-1 accepted for chemically controlled reactions,16 allow
us to conclude that the model reaction of compound 1a
with DMC is not diffusion-limited under the examined
conditions. Therefore, the kinetic investigation is vali-
dated.
To discuss this kinetic evidence, the knowledge of
adsorption phenomena of reagents over the catalyst
becomes mandatory. Model cases of aniline and phenol
are suitable to this scope.
According to Czjiek et al.,19 the steric requisites of
aniline allow it to diffuse into the pores of NaY (super-
Steric and Electronic Effects. At 90 °C, in either
DMC [set a: entries 1-3, Table 2] or DMC and DME as
a cosolvent [set b: entries 4-9, Table 2], aminophenols
(17) The relative reactivity for compounds 1a and 4a (entries 1 and
4) is estimated from the time required for both reactions to be
completed.
(18) (a) March, J. Advanced Organic Chemistry, 4th ed.; Wiley: New
York, 1991. (b) Hansh, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91,
165.
(16) Bond, G. C. Heterogeneous catalysis principles and applications,
2nd ed.; Clarendon Press: Oxford, 1987.
(19) Czjiek, M.; Vogt, T.; Fuess, H. Zeolites 1991, 11, 832.
J. Org. Chem, Vol. 70, No. 7, 2005 2481

Selva et al.
FIGURE 4. Two adsorption sites for PhND2 in the NaY faujasite. Site II: Na+ is the blue dot.
cages), where two sites of adsorption have been identified
through diffraction techniques (Figure 4): on the first
(site I), the amine is held through H-bonds between
amine protons and basic oxygen atoms of the zeolite
framework; on the second (site II), the amine is pinned
to the lattice, forming a π complex between the aryl ring
and a Lewis-acid cation (e.g., Na+).
through H-bonds of intensity comparable to that of
alcohols 3, while a strong acid-base interaction between
carboxylic functions of acids 2a,b (pKa ) 4-5) and oxygen
atoms of the zeolite may plausibly block the reagents over
the catalytic sites (slowing the reaction).25 The relative
rates of methylation of compounds 2 and 4 with both
DMC and MEC (Table 4) may be so explained.
Different authors reported that also the hydroxyl group
of a phenol adsorbed into an alkali metal-exchanged
faujasite (i.e., NaX) is predominantly H-bonded to basic
oxygen atoms of the cavity of the zeolite.20,21
Overall, the balance between steric/electronic effects
and the acidity associated with the aromatic substituents
of amines 1-4 appears as one of the key factors account-
ing for the kinetic results of Tables 2 and 3. This seems
confirmed also from the reactions of O-protected anilines
(6-8) with MEC (Table 4). Data show that: (i) the
reaction of o-aminophenol (1a, entry 3) is much faster
with respect to o-anisidine and even more surprisingly,
with respect to p-anisidine too (entries 1 and 2); and (ii)
anthranilammide (4a, entry 7) is more reactive than
methyl anthranilate (entry 4) and, despite the ortho
substitution, also than methyl p-aminobenzoate (entry
5). These remarkable differences reflect the positive effect
of H-bonding interactions between the catalyst and OH
or CONH2 groups of 1a and 4a, which is missing for
O-protected anilines (6a,b and 7b). On the other hand,
when acid-base interactions between the zeolite and the
substrate become too strong, then reactivity drops: the
reaction of anthranilic acid (2a, entry 6) is only slightly
faster with respect to that of the corresponding methyl
ester (entry 5). Like compounds 1-4, also amines 6-8
should be hosted by the zeolite supercavity (Table 3 of
SI).
The Reaction Mechanism: Mono-N-methyl Selec-
tivity. To formulate a mechanistic hypothesis for the
investigated reaction, the methylation of aniline with
DMC over NaY is a suitable model. Again, the role of
the adsorption of reagents over the catalyst should be
considered, and having already described the behavior
of aniline (see above), only the case of DMC is detailed.
The adsorption of DMC on faujasites has been inves-
tigated via IR experiments carried out by us over NaY26
and by others over NaX zeolites:21 two modes of interac-
H-bond interactions should be expected to play an
important role also in the adsorption of amines 1-4
within NaY. In fact, as supported by the comparison of
the molecular size of compounds 1-422 with the dimen-
sions of the zeolite channels and cavities (diameters of
7.4 and 11.8 Å, respectively23), also amines (1-4) may
have a relatively free access to the faujasite supercages.
If so, the kinetic outcome of Tables 2 and 3 has to be
affected by specific interactions of aromatic substituents
such as OH, CO2H, CH2OH, and CONH2, with the
catalyst itself.
Aminophenols (1a,b). The OH group of aminophenols
can exert a 2-fold favorable effect: (a) because of its
moderate steric hindrance and its electron-donating
properties, it can help the interaction of the amino group
with the catalyst (particularly on site I, Figure 4); (b) it
may stabilize the adsorption of reagents 1a,b through a
direct H-bond with the zeolite (ArO-H‚‚‚‚O-zeolite).
Aminobenzyl Alcohols (3a,b). Like phenols, also the
OH group of benzyl alcohols undergoes H-bonding with
oxygen atoms of alkali metal-exchanged faujasite.24
However, a poorer interaction with respect to phenols
(pKa of 10) is expected because of the poorer acidity of
alcohols (pKa of 16). Along with the steric hindrance and
the moderate electron-withdrawing properties of the CH2-
OH functionality, this fact offers a rationale for the
relative reactivity of compounds 1 and 3.
Aminobenzoic Acids (2a,b) and Aminobenzamides
(4a,b). Weakly acid amidic protons (pKa ) 17) should go
(20) Fu, Z. H.; Ono, Y. Catal. Lett. 1993, 21, 43.
(21) Beutel, T. J. Chem. Soc., Faraday Trans. 1998, 94, 985.
(22) The molecular size of amines 1-4 was estimated with both
semiempirical and ab initio methods: see Table 3 of SI.
(23) Schwochow, F.; Puppe, L. Angew. Chem., Int. Engl. Ed. 1975,
14, 620.
(25) Although the proposed interactions between substrates 2 and
4 with NaY are plausible, to our knowledge, investigations of adsorp-
tion modes of carboxylic acids and/or amides over Y zeolites have not
been reported so far.
(26) Damin, A.; Bonino, F.; Bordiga, S.; Tundo, P.; Selva, M.;
Zecchina, A. EuroConference on Guest-Functionalized Molecular Sieve
Systems, Hattingen (Germany), March 20-25, 2004.
(24) Bezoukhanova, C. P.; Kalvachev, Y. A. Catal. Rev.-Sci. Eng.
1994, 36, 125.
2482 J. Org. Chem., Vol. 70, No. 7, 2005

Mono-N-methylation of Functionalized Anilines
tions can be envisaged where DMC acts as a base to form
an acid-base complex with the Lewis acidic sites of the
catalyst (Scheme 10).
According to Scheme 11, the key requisite for selectiv-
ity is that the reactant amine can migrate through the
zeolite channels, to reach the catalytic supercavity. We
already noticed that this condition was crucial also in
the methylation of p-alkylanilines with MEC (Scheme
12).4d
SCHEME 10. Modes of Adsorption (III and IV) of
DMC over Na-Exchanged Faujasites
SCHEME 12. Mono-N-methylation of p-Alkyl
Anilines with MEC
It should be noted that the formation of both complexes
(III and IV) implies a lengthening of the O-CH3 bond of
DMC, meaning that DMC undergoes an electrophilic
activation within the pores of the solid.
As far as the size of p-substituents allowed the diffu-
sion of reagents into the catalyst, good mono-N-methyl
selectivities (up to 99%) and conversions (45-100%) were
observed, while the bulky 3,5-di-terbutylamine gave a
modest selectivity of 82%, even at a low conversion of
9%.
In line with this observation, the mechanism of Scheme
10 is plausibly validated also for amines 1-8: in fact, the
molecular size estimation of these substrates indicates
that they are all expected to be hosted by the catalytic
supercavity of NaY (see Table 3, SI). For compounds 6-8,
however, the mono-N-methyl selectivity (SM/D) is some-
what unsatisfactory: for instance, in the case of p-
anisidine, SM/D is of only 77% at a conversion of 88%
(entry 2, Table 4). The alteration of the adsorption
geometry, induced by OR groups, into the catalytic cages
seems the most probable reason for this result.
Based on Figure 4 and Scheme 10, the mono-N-methyl
selectivity of the reaction of aniline and DMC can be
discussed through a mechanism which involves the
geometric features of the catalyst (shape selectivity) and
its amphoteric properties as well. In particular, once the
amine and DMC diffuse into the supercages of NaY, they
may approach each other only according to the steric
requisites of their adsorption patterns. Scheme 11 gives
a pictorial description where, for simplicity, only one
adsorption mode is considered for both reagents (sites I
and IV, respectively).27
SCHEME 11. BAl2 Displacement of Aniline on
DMC within the NaY Supercage
The Chemoselectivity. According to the principle of
hard and soft acids and bases,30 in a SN2-type reaction
involving a soft electrophilic center (i.e., a carbon atom),
an ambident nucleophile becomes more likely to attack
with its less electronegative atom.31 This statement fits
the investigated reaction: the methylation takes place
exclusively at the aminic N-atom of anilines 1-4, with
both DMC and MEC.
The role played by the nature of the substrate clearly
emerges also from results of Scheme 7: for aminothio-
phenols (5a,b), the higher polarizability of the thiol group
(with respect to OH) does not allow one to discriminate
between the N- and S-nucleophilic terms of reagents, and
both N- and S-methylations are observed.
In addition to the hard-soft character of the nucleo-
phile, “solvation effects” induced by the zeolite should also
be considered. Zeolites, in fact, may behave as solid
solvents, which, beyond the adsorption of organic mol-
ecules, allow other noncovalent host-guest interactions
(of the van der Waals type) able to confine preferably
some molecules versus others, within their intracrystal-
line volume.32 These adsorption and confinement phe-
nomena should contribute to the reaction chemoselectiv-
ity observed for amines 1-4. In fact, when other catalysts
are used, the N,O-chemoselectivity is elusive even with
DMC as a methylating agent. Scheme 13 exemplifies the
Contrary to our previous hypothesis,4b the reaction
then proceeds via a BAl2 displacement of aniline on DMC.
The product, mono-N-methyl aniline (PhNHMe), plausi-
bly adsorbs into the zeolite in a different way with respect
to aniline, because different H-bonds (N-H‚‚‚‚O-zeolite)
take place with the solid. Moreover, as recently reported
by Su et al.,28 N-methyl amines may also interact with
NaY by H-bonding between the protons of the methyl
group and the oxygen atoms of the zeolite. It may be so
imagined that the NHMe group forces the molecule a bit
far from the catalytic surface in a fashion less suitable
to meet DMC and react with it. This behavior can account
for the mono-N-methyl selectivity observed, which is
peculiar to the use of DMC: in the presence of alkali
metal-exchanged faujasites, in fact, the bis-N-methyla-
tion of primary aromatic amines occurs easily with
conventional methylating agents (i.e., dimethyl sulfate).29
(27) According to Beutel (ref 21), in the coadsorption of phenol and
DMC over a NaX zeolite, DMC would be forced into a preferential
adsorption state corresponding to complex IV of Scheme 10.
(28) Docquir, F.; Toufar, H.; Su, B. L. Langmuir 2001, 17, 6282-
6288.
(30) Pearson, R. G.; Songstand, J. J. Am. Chem. Soc. 1967, 89, 1827.
(31) March, J. Advanced Organic Chemistry, Reactions, Mechanisms,
and Structures, 4th ed.; Wiley: New York, 1992; Chapter 10, pp 367-
368.
(29) Onaka, M.; Ishikawa, K.; Izumi, Y. Chem Lett. 1982, 1783.
(32) Derouane, E. G. J. Mol. Catal. A 1998, 134, 29.
J. Org. Chem, Vol. 70, No. 7, 2005 2483

Selva et al.
SCHEME 13
behavior of reactions of DMC with o- and p-amino phenol
1a,b, both catalyzed by K2CO3.5,33
and 6; k2b/k2a ) 4.0, entries 7 and 8; k4b/k4a ) 6.3, entries
9 and 10). Overall, neither the temperature (from 90 to
135 °C) nor the use of two different carbonates (DMC and
MEC) show appreciable effects on the mono-N-methyl
selectivity (except for the alcohol 3a, SM/ D% ranges from
86% to 100% also for the experiments of Table 3).
The lower reactivity of ortho- with respect to para-
isomers in Table 3 may partly be ascribed to the steric
hindrance and the occurrence of intramolecular H-bonds
on ortho-compounds; if so, however, the role of the same
effects is not appreciable in the reactions with DMC
(Table 2), which show the opposite behavior.
In the case of 1a, N-methylbenzoxazolinon-2-one forms
through a O-methoxycarbonylation/N-methylation se-
quence followed by a cyclization reaction; for 1b, a
plethora of products are obtained from simultaneous N-/
O-methylation and N-methoxycarbonylation reactions.
The Influence of Solvent Polarity. The activity of
faujasites is rather affected by the solvent polarity: in
particular, the competitive adsorption between solvent
and substrates over the catalyst is more and more
important as the media polarity increases, and, as a
consequence, reactions are slowed.4d,14 This behavior
plausibly accounts for some of the kinetic results reported
in Tables 1-4.
In Table 3, for instance, methylations with MEC are
much faster in xylene (ꢀ of 2.4) than in diglyme (ꢀ of 7.2;
see also ref 14) (compare entries 1 and 5, 2 and 7, 4 and
9), at the point that kobs can be evaluated for amino-
phenols 1a,b only in diglyme (entries 5 and 6).
Also, the relative differences of kobs change with the
solvent.
Conclusions
This study demonstrates that several factors can
account for the trend of reactivity/selectivity shown by
amines 1-4 in their mono-N-methylation with DMC and
MEC.
The selectivity of mono-N-methylation is explained
through a BAl2 mechanism assisted by the steric requites
of the supercavities of the NaY catalyst: accordingly, the
estimate of the molecular size of the reactant amines is
in good agreement with the general premise that all of
these compounds can be hosted into the zeolite pores.
The kinetic analysis shows that the reactivity of
compounds 1-4 toward N-methylation is largely modified
by the properties of aromatic substituents (X ) OH,
CO2H, CH2OH, CONH2). Both steric and electronic effects
on the reaction site (NH2 group) and the establishment
of direct acid-base interactions between substituents X
and the catalyst must be simultaneously considered to
explain the observed scheme of rate constants. Weakly
acid groups (CH2OH, CONH2, up to OH: pKa of 16-17
up to 10), may help the adsorption of amines 1 and 3,4
(particularly of 1) over the NaY surface, and so favor the
reaction. In fact, O-protected anilines (anisidines and
esters of aminobenzoic acids, 6-8), where RO-substitu-
ents cannot undergo H-bond interactions with the zeolite,
show lower N-methylation rates with respect to ami-
nophenols and aminobenzamides.
In Table 2, for example, the comparison between
phenols and alcohols shows that in DMC alone, the ratio
k1a/k3a (∼12, entries 1 and 2) doubles that achievable in
DMC/DME (∼5.5, entries 4 and 6).
The most striking evidence, however, seems the alter-
ation of the ortho/para reactivity.
In the presence of DME as a cosolvent (Table 2), the
reactions of DMC with ortho-isomers of phenols and
amides proceed faster than with para-ones (entries 4,5
and 8,9), while the reverse is true for alcohols 3a,b
(entries 6 and 7). This is also confirmed by the competi-
tive reaction between o- and p-aminophenol (1a,b) with
DMC/DME (Figure 3),34 which shows that the ortho-
substituted compound (1a) react more rapidly than the
para-isomer 1b. In all cases, the selectivity (SM/D%) is
moderately better for the more reactive isomers: also in
Figure 3, the N,N-dimethyl derivative is observed only
for 1b (SM/D% for 1b of 93% at a conversion of 91%).
In diglyme solvent instead (Table 3), less hindered
p-substituted compounds exhibit a reactivity higher than
the corresponding o-isomers within all three couples of
ortho/para isomers investigated (k1b/k1a ) 1.4, entries 5
Acids 2a,b (pKa of 4-5), however, are the least reactive
substrates among compounds 1-4, presumably because
carboxylic groups go through strong acid-base interac-
tions with the catalyst.
The solvent polarity also plays a role: methylations of
amines 1-4 with MEC are faster in xylene rather than
in the more polar diglyme, and, particularly, the relative
reactivity of ortho/para isomers changes when reactions
(33) Selva, M. Synthesis 2003, 18, 2872-76.
(34) In Figure 3, both isomers 1a and 1b react more rapidly with
respect to entries 4 and 5 of Table 2; this should be ascribed to the
different reaction conditions used.
2484 J. Org. Chem., Vol. 70, No. 7, 2005

Mono-N-methylation of Functionalized Anilines
apparent activation energy of 93.5 kJ mol-1 was then calcu-
lated from the Arrhenius equation (ln kobs ) ln A - Eatt/RT;
Figure 2B).
Table 2, Entries 1-3: Amines 1a, 3a, and 4a. The above-
described procedure (a) was followed also for the experiments
of entries 1-3 in Table 2. The Q ratio (weight ratio NaY:
substrate) was set equal to 2 to allow most reactions to proceed
near to (or at) completion in a working day.
in DMC/DME or diglyme solvents are compared. The
competitive adsorption between solvent and substrates
over the catalyst can account for this behavior.
Finally, the reaction chemoselectivity is explained
through the different polarizability of the two nucleophilic
terms (NH2 and X groups) of amines 1-4 and solvation
effects promoted by the zeolite.
Table 2, Entries 4-9: Amines 1a,b, 3a,b, and 4a,b. The
above-described procedure a was followed for the experiments
of entries 4-9 in Table 2, with the following modifications:
(i) solutions of compounds 1a,b, 3a,b, and 4a,b (4.2 × 10-2 M,
42 mL) in a mixture of DMC/DME (DMC:DME ) 4 v:v) were
made to react at the reflux of DME (84 °C); (ii) a constant
weight ratio NaY:amine of 2 was always used; and (iii) the
internal standards were triglyme (amines 1a,b, 3a, and 4a,b)
and hexadecane (compound 3b).
(b) Table 3, Entries 1-4: Amines 1a-4a. A three-necked,
jacketed, 50 mL round-bottomed flask fitted with a stopcock,
an adapter for the withdrawal of samples, a reflux condenser
capped with a N2-containing rubber reservoir, and a magnetic
bar was loaded with a solution of the given amine (7.0 × 10-2
M, 30 mL) in xylene as the solvent, and n-octadecane as the
internal standard.
2-(2-Methoxyethoxy)methylethyl carbonate [MeOCO2(CH2)2O-
(CH2)2OMe, MEC] and the faujasite NaY were added to the
mixture (MEC:amine ) 10 molar ratio, NaY:amine ) 2 weight
ratio), which was then degassed under vacuum (10 mmHg)
and left under a N2 atmosphere. The flask was heated at the
desired temperature (135 °C), while the mixture was vigor-
ously stirred. At intervals, samples (0.1 mL) were withdrawn
and were analyzed by both GC and GC/MS.
Table 3, Entries 5-10: Amines 1a,b, 2a,b, and 4a,b. The
above-described procedure b was followed for the experiments
of entries 5-10 in Table 3, with the following modifications:
(i) solutions of compounds 1a,b, 2a,b, and 4a,b (7.0 × 10-2 M,
30 mL) in diglyme [CH3O(CH2)2O(CH2)2OCH3) as the solvent
were made to react at 135 °C. In the case of 4b, the reaction
mixture was analyzed by HPLC.
Scheme 7 and Table 4. The above-described procedure (b)
was followed for the reactions of amines 1a, 2a, 4a, 5a,b, 6a,b,
7a,b, and 8a,b.
Except for the acid 2b, calibration curves were built for each
substrate and were used to follow the reaction by GC. The
amount and the type of internal standards used for any given
substrate in the different solvents are specified in the Sup-
porting Information.
The structures of mono-N- and di-N,N-methyl derivatives
of the investigated amines were assigned through GC/MS
analyses. A synoptic table of GC/MS spectra is available in
the Supporting Information.
In the case of amines 1-4, the structure of the correspond-
ing mono-N-methyl derivatives was confirmed also by com-
parison to authentic samples previously prepared by us.5
Experimental Section
Compounds 1-4, anisidines 6a,b, methyl anthranilate (8a),
and DMC were ACS grade and were employed without further
purification. The zeolite NaY was from Aldrich (art. 33,444-
8), and, before each reaction, it was dried under vacuum (65
°C; 8 mbar) overnight.
2-(2-Methoxyethoxy)methylethyl carbonate [MeO(CH2)2O-
(CH2)2OCO2Me, MEC], methyl and ethyl p-aminobenzoates (7b
and 8b), and ethyl anthranilate (8a) were prepared according
to already described procedures:4c,35 their physical and spec-
troscopic data were in agreement with those reported in the
literature (7b,36 8a,37 8b,38 MEC4c). MEC was also compared
to an authentic sample.
GLC and GC/MS (70 eV) analyses were run using CPSil24,
FFAP, and HP5/MS capillary columns (30 m), respectively.
HPLC analyses were carried out with an inverse-phase C18
column (“Aqua”, 250 × 4.5 mm, length × thickness; size
particles, 5 µm): p-aminobenzoic acid (2b) and its mono-N-
and di-N,N-methyl derivatives (whose structures were as-
signed by comparison to authentic samples) were eluted with
a binary mixture of MeCN/H2O (gradient: 20-100% MeCN,
20 min) buffered to pH ) 3, and then revealed by an UV
detector.
1H NMR spectra were recorded on a 300 MHz spectrometer,
using CDCl3 with TMS as the internal standard.
The kinetic results of Tables 1-4 were gathered through
procedures a and b, described below. To check for repeatability,
each data point should be considered as an average value of
at least two subsequent runs whose corresponding kobs values
deviated e10% from each other.
(a) The Reaction of o-Aminophenol with DMC. Figure
1 and Table 1. A three-necked, jacketed, 50 mL round-
bottomed flask fitted with a stopcock, an adapter for the
withdrawal of samples, a reflux condenser capped with a
N2-containing rubber reservoir, and a magnetic bar was loaded
with a solution of o-aminophenol (1a, 6.5 × 10-2 M, 42 mL) in
dimethyl carbonate. Triglyme (0.56 mmol) as the internal
standard and the faujasite NaY (weight ratio NaY:1a in the
range of 0.5-3) were added to the mixture, which was then
degassed under vacuum (10 mmHg) and kept under a N2
atmosphere. The flask was heated at the reflux temperature
(90 °C), while the mixture was vigorously stirred. At intervals,
samples (0.1 mL) were withdrawn and were analyzed by both
GC and GC/MS.
Figure 2A. Under the conditions of Figure 1, kinetic
constants were measured at different stirring rates of 500, 970,
and 1350 rpm.
Figure 2B. Under the conditions of Figure 1, kinetic
constants (kobs) were also measured at different temperatures
(T) of 343, 348, 353, and 363 K (70, 75, 80, and 90 °C,
respectively), and a plot of ln kobs versus 1/T was built up. An
Acknowledgment. MIUR (Italian Ministry of Uni-
versity and Research) and INCA (Interuniversity Con-
sortium Chemistry for the Environment) are gratefully
acknowledged for financial support.
Supporting Information Available: GC/MS spectra of
mono-N- and di-N,N-dimethyl derivatives of amines 1-8,
details of some kinetic results, and a synoptic table of internal
standards. The estimation of molecular size of amines 1-8 is
also given. This material is available free of charge via the
Internet at http://pubs.acs.org.
(35) Hosangadi, B.; Dave, R. Tetrahedron Lett. 1996, 37, 6375-6378.
(36) Ramesha, A. R.; Bhat, S.; Chandrasekaran, S. J. Org. Chem.
1995, 60, 768.
(37) Herweh, J. E.; Hoyle, C. E. J. Org. Chem. 1980, 45, 2195.
(38) Zafar, A.; Melendez, R.; Geib, S. J.; Hamilton, A. D. Tetrahedron
2002, 58, 683.
JO048076R
J. Org. Chem, Vol. 70, No. 7, 2005 2485