Formation and nucleophilic capture of N-nitrosiminium ions

Article abstract of DOI:10.1139/v99-108

A study of the solvolysis of a series of (N-nitrosomethylamino)arylmethyl esters and azides and the products of nucleophilic trapping of the corresponding N-nitrosiminium ion intermediates in aqueous media, 25°C, ionic strength 1 M is reported. Structure-reactivity data for the forward and reverse reactions have been obtained. In three cases, the rate constants for reactions of the cations with nucleophiles have been measured directly by laser flash photolysis. The data allow a comparison of the degree to which the N-methyl-N-nitroso functionality enhances cation stability from a thermodynamic and kinetic perspective. It has been possible to deduce that the carbon basicity of azide ion is less than 1 kcal/mol greater than that of acetate ion.

Full text of DOI:10.1139/v99-108

1
148
Formation and nucleophilic capture of
N-nitrosiminium ions
Latifa Chahoua, Alain Vigroux, Yvonne Chiang, and James C. Fishbein
Abstract: A study of the solvolysis of a series of (N-nitrosomethylamino)arylmethyl esters and azides and the products
of nucleophilic trapping of the corresponding N-nitrosiminium ion intermediates in aqueous media, 25°C, ionic strength
1
M is reported. Structure–reactivity data for the forward and reverse reactions have been obtained. In three cases, the
rate constants for reactions of the cations with nucleophiles have been measured directly by laser flash photolysis. The
data allow a comparison of the degree to which the N-methyl-N-nitroso functionality enhances cation stability from a
thermodynamic and kinetic perspective. It has been possible to deduce that the carbon basicity of azide ion is less than
1
kcal/mol greater than that of acetate ion.
Key words: nitrosiminium ions, α-acetoxynitrosamines, carbocations, iminium ions, nucleophilicity.
Résumé : On a effectué une étude, en milieu aqueux, à 25°C et à une force ionique de 1 M, sur la solvolyse d’une
série d’esters (N-nitrosométhylamino)arylméthyles et d’azotures apparentés et sur la nature des produits du piégeage
nucléophile des ions intermédiaires N-nitrosiminium correspondants. On a obtenu des données relatives de structure–
réactivité pour les réactions normales et inverses. Dans trois cas, faisant appel à la photolyse éclair au laser, on a pu
mesurer directement les constantes de vitesse des réactions des cations avec les nucléophiles. Les données permettent
de comparer à quel degré la fonction N-méthyl-N-nitroso accroît la stabilité du cation de points de vue tant
thermodynamique que cinétique. On a pu déduire que la basicité du carbone de l’ion azoture n’est plus grande que
celle de l’ion acétate que par moins de 1 kcal/mol.
Mots clés : ions nitrosiminium, α-acétoxynitrosamines, carbocations, ions iminium, nucléophilicité.
[
Traduit par la Rédaction] Chahoua et al. 1161
Introduction
ion that is subsequently hydrated by solvent water, as in
eq. [2] (3–6). Some characterization of the kinetic stability
of certain of these intermediates has been reported (7), in-
cluding the photochemical generation and direct determina-
tion of reactivity in one case (8). α-Substituted benzylic
cations have been relatively well characterized with regard
to the effects of substituents on cation stability and selectiv-
ity, and it is of interest to understand how the N-
nitrosomethylamino group compares to others as a cation
stabilizing substituent. The present study was undertaken to
quantitatively elucidate this cation stabilizing ability in a
α-Acetoxydialkylnitrosamines are important in under-
standing the mechanisms of mutagenesis and carcinogenesis
of dialkylnitrosamines (1, 2). They serve as precursors to
α-hydroxydialkynitrosmaines that are formed in the ac-
tivation of dialkylnitrosamines by P-450 enzymes, as il-
lustrated in eq. [1]. Evidence has been summarized that
indicates
that
in
the
case
of
most
α-
acetoxydialkylnitrosamines, formation of the α-hydroxy
intermediate occurs by formation of an N-nitrosiminium
NO
N
OAc
R
R'
NO
N
R N N OH
+
NO
N
P450
+
[
1]
DNA
OH
R
R N N
R-DNA
R
O
R'
R'
R'
Received November 17, 1998.
This paper is dedicated to Jerry Kresge, an outstanding chemist, collaborator, and friend, in recognition of his many achievements
in chemistry.
1
L. Chahoua, A. Vigroux, Y. Chiang, and J.C. Fishbein. Dept. of Chemistry, Wake Forest University, Winston–Salem, NC
2
7109, U.S.A.
1Author to whom correspondence may be addressed. Telephone: (336) 758-5328. Fax: (336) 758-4656. e-mail: fishbein@wfu.edu
Can. J. Chem. 77: 1148–1161 (1999)
© 1999 NRC Canada

Chahoua et al.
1149
NO
NO
N
NO
N
N
R
OAc
rls
+
AcO -
OH
[
2]
+
R
R
R'
R'
The pH values were obtained after the kinetic run using
an Orion model SA 720 pH meter equipped with a Corning
combination electrode. Two point calibrations were done
prior to recording pH values. Calibration of the meter was
performed using commercially available standards or those
prescribed by the Merck Index (9).
R'
benzylic system. In summary, the N-nitrosomethylamino
group stabilizes the adjacent benzyl cation both kinetically
and thermodynamically to an extent that is similar to the sta-
bilization that is provided by an α-azido group. In addition,
it has been possible to compare the cation generating reac-
tions involving the acetate and azide precursors and the cat-
ion capture reactions by acetate and azide ions, two
nucleophiles of very similar proton basicity. An upper limit,
of 1 kcal/mol, on the difference in carbon basicity of these
two nucleophiles has been derived.
Product analysis
The products of decomposition under various conditions,
vide infra, were quantitated after 10 halflives of reaction of
the starting materials. Product quantitation involved either
1
H-NMR or HPLC. In the former case, the reactant concen-
Experimental
trations were on the order of 0.01 M. In the latter case, con-
–
4
–5
centrations were typically in the range of 10 –10 M and
utilized a Waters chromatography system with Axxiom data
system and Waters Wisp-Ultra autoinjector attached. Separa-
tions were carried out with 6:4 acetonitrile/water as eluent
over a 15 cm waters C18 column. Quantitation of the prod-
ucts was accomplished using standard curves containing at
least two points that were generated from products that were
synthesized independently.
Warning! Most α-acyloxynitrosamines are powerful direct
acting carcinogens. Manipulations were carried out using
frequently changed double pairs of disposable gloves and in
the well-ventilated hood. Contaminated, and potentially con-
taminated, materials were treated with 50% aqueous sulfuric
acid containing the commercially available oxidant “No
Chromix” (Aldrich Chemical).
A series of experiments were carried out to examine the
partitioning of the nitrosiminium ion intermediates between
capture by water and alcohols. Reactions were carried out by
dissolution of the starting ester in buffer solution that was,
by volume, 1% ethanol, 9% other added alcohol, and 90%
H O, ionic strength 1 M with NaClO .
Materials
Organic solvents were dried and purified by distillation
before use. Organic and inorganic chemicals for synthesis
were ACS grade or better. Inorganic chemicals were used
without further purification. Organic chemicals were typi-
cally purified prior to use in synthesis or kinetic experi-
ments. Water was distilled in glass.
2
4
Synthesis
Esters, azide ion adducts, and ethers, products of
nitrosiminium ion trapping by alcohols, were synthesized
and are listed and correspondingly numbered in Table 1.
The α-acyloxynitrosamines 1a–7a and 1b–8b were syn-
thesized using previously published methods (3, 4). Physical
Methods
Kinetics
The kinetics of decomposition were observed using either
a Hewlett Packard 8452A diode array spectrophotometer, or
an Applied Photophysics DX 17MV stopped-flow spectro-
photometer. For compounds 3k and 4k, rate constants for
the addition of water and azide ion to the nitrosiminium ions
generated by KrF laser flash photolysis were measured as
described for the unsubstituted compound earlier (8). In all
cases, temperature was maintained at 25°C by means of at-
tached recirculating water baths.
1
data for 2a and 2b were reported previously (6). H NMR
(CDCl ) δ (ppm): 1a: (E) isomer (98%): 8.26 (1H, s), 7.32
3
(2H, s), 7.01 (2H, s), 3.83 (3H, s), 2.82 (3H, s), 2.24 (3H, s);
(Z) isomer (2%): 3.57 (3H, s); 2.18 (3H, s); Anal. calcd.: C
55.46, H 5.88, N 11.76; found: C 56.96, H 5.82, N 10.86;
3a:(E) isomer (86%): 8.28 (1H, s), 7.37 (4H, m), 2.8 (3H, s),
2.25 (3H, s); (Z) isomer (14%): 3.57 (3H, s), 2.18 (3H, s);
4a: (E) isomer (96%): 8.28 (1H, s), 7.38 (4H, m), 2.80
(3H, s), 2.26 (3H,s); (Z) isomer (4%): 3.57 (3H, s), 2.2
(3H, s); 5a: (E) isomer (96%): 8.34 (1H, s), 7.76 (2H, d),
7.53 (2H, d), 2.81 (3H, s), 2.28 (3H, s); 6a: (E) isomer
(96%): 8.38 (1H, s), 8.32 (2H, d), 7.61 (2H, d), 2.83 (3H, s),
2.30 (3H, s); (Z) isomer (4%): 3.60 (3H, s), 2.24 (3H, s).
Anal. calcd.: C 47.43, H 4.35, N 16.60; found: C 47.50, H
4.35, N 16.52; 7a: (E) isomer (96%): 8.38 (1H, s), 7.96
(1H, s), 7.86 (2H, s), 2.85 (3H, s), 2.31 (3H, s); (Z) isomer
(4%): 3.62 (3H, s), 2.24 (3H, s); 1b: (E) isomer (96%): 8.32
(1H, s), 7.34 (2H, d), 6.96 (2H, d), 4.21 (2H, s), 3.83
(3H, s), 2.83 (3H, s); 3b: (E) isomer (96%): 8.35 (1H, s), 7.4
(5H, m), 4.22 (2H, d), 2.82 (3H, s); (Z) isomer (4%): 3.59
(3H, s). Anal. calcd.: C 43.32, H 3.61, N 10.11; found: C
43.47, H 3.65, N 10.09; 4b: δ (E) isomer (94%): 8.35
(1H, s), 7.43 (5H, m), 4.24 (2H, d), 2.84 (3H, s); (Z) isomer
The buffer solution in each reaction cell was made up by
diluting a concentrated stock buffer solution. Ionic strength
of reaction solutions was maintained at 1 M with NaClO . A
4
stock solution of NaClO was added to the cuvette to main-
4
tain the ionic strength of 1 M. The buffers that were used in-
clude phosphate, MOPS, MES, CHES, acetic acid,
cyanoacetic acid, methoxyacetic acid, cacodylic acid, and
morpholine.
The stock solution of the nitrosamine was made up in
freshly distilled dry CH CN. Reactions performed in the
3
stopped-flow spectrophotometer were carried out by mixing
2
5 parts of the aqueous buffer with 1 part of the
nitrosoamine in CH CN solution using 2.5 and 0.1 mL sy-
3
ringes, respectively. Final concentration of the nitrosamine
in the reactions was -10 M.
–
4
©
1999 NRC Canada

1
150
Table 1. α-Substituted nitrosamines studied in the present work.
NO
Can. J. Chem. Vol. 77, 1999
R
N
ArX
No.
X
R
No.
X
R
No.
X
R
–
1
2
a
a
p-MeO
H
CH COO–
2c
3c
H
p-Cl
N3–
N3
3i
3j
p-Cl
p-Cl
Cl CHCH O
3
2
2
–
–
–
CH COO
F CCH O
3
3
2
3
k
p-Cl
4-CN-PhO-
–
3
4
5
6
7
1
2
a
a
a
a
a
b
b
p-Cl
m-Cl
CH COO–
4c
5c
2d
2e
2f
m-Cl
p-CF3
H
H
H
N3–
N3
4d
4e
4f
4g
4h
4i
m-Cl
m-Cl
m-Cl
m-Cl
m-Cl
m-Cl
m-Cl
m-Cl
p-CF3
p-CF3
p-CF3
p-CF3
p-CF3
p-CF3
p-CF3
CH CH O
3
3
2
–
–
–
CH COO
CH O
3
3
–
–
–
p-CN
p-NO2
3,5-(CF3)2
p-MeO
H
CH COO
CH CH O
CH OC H O
3 2 4
ClCH CH O
2 2
NCCH CH O
2 2
3
3
2
–
–
–
CH COO
CH O
3
3
–
–
–
CH COO
CH OC H O
3
3
2
4
–
–
–
–
ClCH COO
2g
2h
H
H
ClCH CH O
NCCH CH O
Cl CHCH O
2
2
2
2
2
–
–
ClCH COO
4j
4
F CCH O
2
2
2
3
2
k
4-CN-PhO-
–
–
3
4
5
6
7
8
1
b
b
b
b
b
b
c
p-Cl
m-Cl
p-CN
p-NO2
3,5-(CF3)2
p-CF3
p-MeO
ClCH COO–
2i
2j
H
H
Cl CHCH O
5d
5e
5f
5g
5h
5i
CH CH O
2
2
2
–
3
2
–
–
ClCH COO
F CCH O
CH O
2
3
2
3
–
–
–
ClCH COO
3d
3e
3f
3g
3h
p-Cl
p-Cl
p-Cl
p-Cl
p-Cl
CH CH O
CH OC H O
3 2 4
ClCH CH O
2 2
NCCH CH O
2 2
2
3
2
–
–
–
ClCH COO
CH O
2
3
–
–
–
ClCH COO
CH OC H O
2
3
2
4
–
–
–
ClCH COO
ClCH CH O
NCCH CH O
Cl CHCH O
2
2
2
2
2
–
–
–
N3
5j
F CCH O
2
2
3
2
(
6%): 3.63 (3H, s). Anal. calcd.: C 43.32, H 3.61, N 10.11;
found: C 43.60, H 3.85, N 9.84; 5b: δ (E) isomer (96%):
.45 (1H, s), 8.35 (2H, d), 7.65 (2H, d), 4.28 (2H, s), 2.84
3H, s); (Z) isomer (4%): 3.62 (3H, s); 6b: (E) isomer
96%): 8.45 (1H, s), 8.35 (2H, d), 7.65 (2H, d), 4.28
32H, s), 2.84 (3H, s); (Z) isomer (4%): 3.62 (3H, s); 8b: (E)
The synthesis of 4c and 5c was accomplish using
Wiessler’s procedure (10), since the method above was not
successful. The imines (11) (6.5 mmol) were dissolved in
10 mL of dry methylene chloride and added slowly to NOCl
(6.5 mmol) in 20 mL of dry methylene chloride at –30°C.
Sodium azide (6.5 mmol), dissolved in 5 mL of DMSO, was
added dropwise to the previous mixture. After 5 min, the
product was washed with water and extracted with methy-
8
(
(
(
isomer (94%): 8.41 (1H, s), 7.8 (2H, d), 7.6 (2H, d), 4.25
(
2H, d), 2.84 (3H, s); (Z) isomer (6%): 3.63 (3H, s).
lene chloride. The organic layer was dried over MgSO and
4
was filtered. The methylenechloride was then evaporated off
(
N-Methyl-N-nitroso)arylmethyl azides 1c–3c
α-Chloroacetoxy-N-nitrosamine ester (2.4 mmol) was dis-
using a flow of Argon. The product was purified by prepara-
1
tive TLC using 0.5:9.5 ether/pentane. H NMR (CDCl ) δ
solved in a solution (35 mL H O/25 mL methanol) contain-
3
2
(
(
7
3
ppm): 4c: (E) isomer (91%): 7.4 (5H, m), 7.27 (1H, s), 2.82
3H, s); (Z) isomer (9%): 3.53 (3H, s); 5c: (E) isomer (87%):
ing 1.0 M NaN3. The solution was stirred at room
temperature overnight (for 3c, we stirred at –10°C for 2 h).
A large amount of water was added to precipitate the prod-
uct. The product was extracted with methylene chloride. The
.63 (5H, 2d), 7.5 (1H, s), 2.82 (3H, s); (Z) isomer (13%):
1
3
.53 (3H, s). C NMR (CDCl ) δ: 4c: 135.5–124.6 (6 pics,
3
Ar), 79.0 (CH-N), 27.18 (CH -N).
organic layer was dried over MgSO and was filtered. Meth-
3
4
ylene chloride was then evaporated off using a flow of Ar-
gon and a warm water bath. Compound 1c was purified by
preparative TLC using 1:7 ether:pentane as the eluting sol-
vent. The crude material of 2c was washed with pentane to
get rid of aldehyde. The purification of 3c was hampered by
the presence of aldehyde, the mixture of crude material was
treated with bisulfite in methanol and subsequently purified
The ethers were synthesized by decomposition of an ester
in the neat alcohol. Thus, in a typical procedure, the acetoxy
or chloroacetoxy compound (1.0 mmol) was dissolved in dry
alcohol (4 mL) and allowed to reflux for 16 h. A large
amount of water was then added, the product was extracted
with methylene chloride, the organic layer was dried over
MgSO , and the solvent was removed under a stream of ar-
4
1
by preparative TLC using 1:9 ethyl acetate/pentane. H NMR
gon. The crude material was purified by preparative TLC us-
1
(
(
CDCl ) δ (ppm): 1c: (E) isomer (100%): 7.42 (1H, s), 7.33
2H, d), 6.93 (2H, d), 3.83 (3H, s), 2.82 (3H, s). Anal.
ing 1:20 ether/petroleum ether. H NMR (CDCl ) δ (ppm):
3
3
2d: (E) isomer (96%): 7.34 (5H, m), 6.78 (1H, s), 3.57
(2H, m), 2.68 (3H, s), 1.25 (3H, t); (Z) isomer (4%): 3.4
(3H, s). Anal. calcd.: C 61.85, H 7.2, N 14.4; found: C
61.95, H 7.26, N 14.31; 2e: (E) isomer (98%): 7.41 (5H, s),
6.75 (1H, s), 3.47 (3H, s), 2.75 (3H, s); (Z) isomer (2%):
3.38 (3H, s). Anal. calcd.: C 60.00, H 6.67, N 15.55; found:
C 60.21, H 6.75, N 15.52; 2f: 7.41 (5H, s), 6.91 (1H, s), 3.71
calcd.: C 48.84, H 5.01, N 31.7; found: C 49.1, H 5.05, N
1.43; 2c: (E) isomer (94%): 7.44 (1H, s), 7.42 (5H, s), 2.82
3H, s); (Z) isomer (6%): 3.51 (3H, s); 3c: (E) isomer (93%):
.4 (1H, s), 7.37 (5H, 2d), 2.8 (3H, s); (Z) isomer (7%): 3.5
3H, s). Anal. calcd.: C 42.58, H 3.57, N 31.04; found: C
2.64, H 3.60, N 30.92.
3
(
7
(
4
©
1999 NRC Canada

Chahoua et al.
1151
(
7
4H, m), 3.4 (3H, s), 2.76 (3H, s); 2g: (E) isomer (94%):
.42 (5H, s), 6.93 (1H, s), 3.84 (4H, m), 2.78 (3H, s); (Z)
isomer (6%): 3.5 (3H, s). Anal. calcd.: C 52.52, H 5.69, N
2.25; found: C 52.26, H 5.61, N 11.97: 2h: (E) isomer
97%): 7.43 (5H, s), 6.95 (1H, s), 3.8 (2H, m), 2.78 (3H, s),
.74 (2H, m); (Z) isomer (3%): 3.41 (3H, s). Anal. calcd.: C
0.26, H 5.98, N 19.16; found: C 59.84, H 6.91, N 18.48; 2i:
7.65 (4H, 2d), 7.00 (1H, s), 3.85 (2H, m), 2.78 (5H, s); (Z)
isomer (2%): 3.51 (3H, s); 5i: (E) isomer (97%): 7.66 (4H,
2d), 7.09 (1H, s), 5.9 (1H, t), 4.00 (2H, m), 2.78 (3H, s); (Z)
isomer (3%): 3.61 (3H, s); 5j: (E) isomer (97%): 7.65
(5H, s), 7.07 (1H, s), 4.03 (2H, m), 2.76 (3H, s); (Z) isomer
(3%): 3.5 (3H, s).
1
(
2
6
(
E) isomer (93.3%): 7.58 (5H, s), 7.16 (1H, s), 6.00 (1H, t),
.19 (2H, m), 2.92 (3H, s); (Z) isomer (6.7%): 3.61 (3H, s).
Anal. calcd.: C 45.65, H 4.6, N 10.65; found: C 46.08, H
(N-Methyl-N-nitroso)arylmethyl p-cyanophenyl ethers 3k, 4k
4
These were made by the method reported for the
1
unsubstituted compound. H NMR (CDCl3) δ (ppm): 3k: (E)
4
4
.63, N 9.3; 2j: (E) isomer (98%): 7.43 (5H, s), 7.00 (1H, s),
.00 (2H, m), 2.77 (3H, s); (Z) isomer (2%): 3.49 (3H, s).
isomer (95%): 7.88 (1H, s), 7.62 (2H, m), 7.44 (4H, s), 7.20
(2H, m), 2.77 (3H, s); (Z) isomer (5%): 3.51 (3H, s); 4k: (E)
isomer (96%): 7.78 (1H, s), 7.62 (3H, m), 7.40 (2H, m), 7.22
Anal. calcd.: C 48.4, H 4.44, N 11.29; found: C 48.56, H
1
3
4
2
1
6
5
7
.41, N 11.09; 3d: 7.38 (4H, s), 6.83 (1H, s), 3.63 (2H, m),
.73 (3H, s), 1.33 (3H, t). Anal. calcd.: C 52.52, H 5.73, N
2.25; found: C 53.74, H 5.92, N 11.66; 3e: 7.38 (4H, s),
.73 (1H, s), 3.47 (3H, s), 2.73 (3H, s). Anal. calcd.: C
0.36, H 5.16, N 13.05; found: C 50.45, H 5.19, N 12.96; 3f:
.38 (4H, s), 6.89 (1H, s), 3.69 (4H, m), 3.39 (3H, s), 2.74
(3H, m), 2.79 (3H, s); (Z) isomer (4%): 3.63 (3H, s).
C
NMR (CDCl3) δ: 3k: 158.61, 135.82, 134.27, 132.47,
129.37, 127.40, 118.37, 117.25, 106.78, 88.70, 26.78; 4k:
158.58, 136.04, 135.33, 134.33, 130.49, 129.99, 126.29,
124.17, 118.39, 117.28, 106.91, 88.52, 26.87.
(
(
3H, s); 3g: (E) isomer (94%): 7.4 (4H, s), 6.9 (1H, s), 3.81
4H, m), 2.76 (3H, s); (Z) isomer (6%): 3.5 (3H, s). Anal.
Results
calcd.: C 45.65, H 4.6, N 10.65; found: C 45.6, H 4.59, N
0.48; 3h: (E) isomer (98%): 7.4 (4H, s), 6.92 (1H, s), 3.82
The decay of (N-methyl-N-nitroso)arylmethyl acetates,
chloroacetates, and azides exhibited good first-order kinetics
of absorbance decay in the range of conditions studied for
4–6 halflives of reaction. In the case of the azides, the reac-
tions were carried out only in acidic solutions containing
1
(
(
5
7
2H, m), 2.77 (3H, s), 2.76 (2H, m); (Z) isomer (2%): 3.51
3H, s). Anal. calcd.: C 52.08, H 4.77, N 16.56; found: C
1.94, H 4.81, N 16.37; 3i: (E) isomer (98%): 7.4 (4H, s),
.00 (1H, s), 5.87 (1H, t), 4.00 (2H, m), 2.77 (3H, s); (Z)
0.01–0.1 M HClO in order to prevent nonsimple kinetics of
4
isomer (2%): 3.5 (3H, s). Anal. calcd.: C 40.36, H 3.73, N
decay due to the formation of azide ion product that could
inhibit decay of the starting material due to common ion in-
hibition. The rate constants kobsd in these cases were inde-
pendent of the acid concentration, agreeing within
experimental error for at least two different acid concentra-
tions that differed by at least a factor of two. In the cases of
the esters, the values of kobsd were determined typically at
three buffer concentrations for a given buffer ratio and a
concentration range between 0.05 and 0.3 M. The values of
kobsd in all cases increased less than 10% from low to high
buffer concentration, and extrapolation of the plot of kobsd
against buffer concentration to the y-axis yielded to value of
k1, the buffer-independent, pH-independent rate constant.
9
7
.41; found: C 40.16, H 3.67, N 9.28; 3j: (E) isomer (98%):
.41 (4H, s), 6.99 (1H, s), 4.00 (2H, m), 2.75 (3H, s); (Z)
isomer (2%): 3.49 (3H, s). Anal. calcd.: C 42.49, H 3.56, N
.91; found: C 42.63, H 3.57, N 9.97; 4d: (E) isomer
9
(
(
(
5
7
97.5%): 7.5 (1H, s), 7.33 (3H, m), 6.83 (1H, s), 3.64
2H, m), 2.75 (3H, s), 1.33 (3H, t); (Z) isomer (2.5%): 3.47
3H, s). Anal. calcd.: C 52.52, H 5.73, N 12.25; found: C
2.66, H 5.71, N 12.21; 4e: (E) isomer (98%): 7.49(1H, s),
.32 (3H, m), 6.73 (1H, s), 3.47 (3H, s), 2.75 (3H, s); (Z)
isomer (2%): 3.38 (3H, s). Anal. calcd.: C 50.36, H 5.16, N
1
7
3.05; found: C 50.47, H 5.17, N 12.97; 4f: 7.51 (1H, s),
.33 (3H, m), 6.89 (1H, s), 3.71 (4H, m), 3.4 (3H, s), 2.76
(
5
7
3H, s). Anal. calcd.: C 51.07, H 5.84, N 10.84; found: C
1.19, H 5.83, N 10.69; 4g: (E) isomer (98%): 7.52 (1H, s),
.44 (3H, m), 7.02 (1H, s), 4.06 (4H, m), 2.9 (3H, s); (Z)
Determinations of k were carried out in a minimum of two
1
different buffers, typically acetic and cacodylic acids, and
the values generally agreed within 10%. The mean values of
k1 for all compounds are reported in Table 2.
For a number of α-acetoxydialkylnitrosamines, the effect
on kobsd of increasing acetate ion concentration, while main-
isomer (3%): 3.41 (3H, s). Anal. calcd.: C 45.65, H 4.6, N
0.65; found: C 45.39, H 4.58, N 10.4; 4h: (E) isomer
97%): 7.47 (1H, s), 7.34 (3H, m), 6.92 (1H, s), 3.8 (2H, m),
.78 (5H, m); (Z) isomer (2%): 3.5 (3H, s); 4i: (E) isomer
97%): 7.49 (1H, s), 7.33 (3H, m), 6.99 (1H, s), 5.87 (1H, t),
.00 (2H, m), 2.78 (3H, s); (Z) isomer (3%): 3.55 (3H, s).
Anal. calcd.: C 40.36, H 3.73, N 9.41; found: C 40.21, H
.71, N 9.48; 4j: (E) isomer (95.6%): 7.47 (1H, s), 7.34
1
(
2
(
taining constant ionic strength (1 M) with NaClO , was de-
4
termined. A plot illustrating the effect for a number of
compounds is indicated in Fig. 1. The value of kobsd de-
creases, in all cases measured, to less than 25% the value of
kobsd in the absence of added acetate ion. Specific salt ef-
fects, due to replacement of sodium perchlorate by sodium
trifluoroacetate, have been shown to decrease the value of
kobsd by less than 10% at comparable concentrations (6).
To clarify the nature of the reaction that appeared to be
acetate independent at very high acetate ion concentrations,
a single experiment was run with compound 2a in which the
effect of acetate concentrations up to 1.35 M was deter-
mined at an ionic strength of 3 M (NaClO ) in D O. The
4
3
(
(
3H, m), 6.98 (1H, s), 4.00 (2H, m), 2.77 (3H, s); (Z) isomer
4.4%): 3.49 (3H, s). Anal. calcd.: C 42.49, H 3.57, N 9.91;
found: C 43.68, H 3.66, N 9.68; 5d: (E) isomer (97%):
.62(4H, 2d), 6.91 (1H, s), 3.66 (2H, m), 2.74 (3H, s), 1.34
3H, t); (Z) isomer (3%): 3.45 (3H, s); 5e: (E) isomer (98%):
.62 (4H, 2d), 6.80 (1H, s), 3.50 (3H, s), 2.74 (3H, s); (Z)
isomer (2%): 3.39 (3H, s); 5f: 7.64 (4H, 2d), 6.981 (1H, s),
.8 (4H, m), 3.41 (3H, s), 2.79 (3H, s); 5g: (E) isomer
7
(
7
3
(
(
4
2
1
93%): 7.65 (4H, 2d), 6.99 (1H, s), 3.90 (4H, m), 2.77
3H, s); (Z) isomer (7%): 3.5 (3H, s); 5h: (E) isomer (98%):
data are plotted in Fig. 2. Product analysis by H-NMR of a
reaction at completion that originally contained 1.35 M ace-
©
1999 NRC Canada

1
152
Can. J. Chem. Vol. 77, 1999
Table 2. Rate constants for the solvolysis of α-substituted (N-nitrosomethylamino)arylmethylamines in cacodylic acid buffer, ionic
a
strength µ = 1 (NaClO ) at 25°C.
4
Leaving group
Acetate
Chloroacetate
Azide
3
–1
–1
3
–1
1
0 × k (s )
k (s )
10 × k (s )
1
1
1
Ring substituent
<1% CH CN
<1% CH CN
<1% CH CN
3
3
3
4
4
H
4
3
4
4
4
3
-CH O
-F
6.9
0.60
1.17
3
0.057b
0.053
0.49
0.23
0.059c
0.049
0.022
0.002
-CL
-CL
-CF3
-CN
-NO2
,5-(CF3)2
a
0.027b
0.0095
0.0053
0.0037
0.0022
0.00107
d
0.023
0.0128
0.0083
Average of at least two values of intecepts of plots of kobsd against buffer concentration containing 3–4 points, typically 0.05–0.3 M. Standard error of
<
±5% except where otherwise indicated.
b
Standard error < ±20%.
c
From a single value of the intecept of a plot of kobsd against buffer concentration containing 4 points, typically 0.05–0.2 M.
d
Standard error < ±10%.
Fig. 1. Plot of kobsd against acetate ion concentration for the
solvolysis of acetates 2a (circles), 3a (squares), and 4a
Fig. 2. Plot of kobsd against acetate ion concentration for the
solvolysis of acetate 2a at 25°C, ionic strength 3 M (NaClO4).
(
diamonds) at 25°C, ionic stength 1 M (NaClO4).
0.0004
4
0
0
0
.00035
.0003
.00025
.0002
.00015
.0001
3
2
1
0
.5
3
0
.5
2
0
.5
1
0
-
5
5
10
.5
0
0
0
0.2 0.4 0.6 0.8
1
1.2 1.4 1.6
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
[
Acetate Buffer], (90% anion) M
[
Acetate lon] M
tate ion in D O showed a 12% yield of methylammonium
ion. The analysis of these experiments are discussed in more
detail later.
The rate constants for reactions of two N-nitrosiminium
ions with water and azide ion were measured by laser flash
photolysis using the p-cyanophenyl ethers, as described ear-
lier (8). The decay of absorbance of the cations exhibited
clean first-order kinetics for five halflives of decay. Rate
constants for decay of the cation were measured in 0.01 and
2
tively. The rate constants for the addition of azide ion are
9
–1 –1
9
–1 –1
2.4 (± 0.2) × 10 M
s
and 2.8 (± 0.2) × 10 M s , re-
spectively.
Characterization of products and yields were initally per-
formed under various conditions to check the generality of
what has previously been reported — these are summarized
in Table 3. Decay of the esters gives quantitative yields of
1
the aryl moiety. Analysis by H-NMR gives less than quanti-
0
.02 M cacodylic acid buffer (50% anion), and the two val-
ues of kobsd agreed within 5%. The rate constant for decay of
k in aqueous media containing 0.01 M cacodylic acid
tative yields of the methyl fragment, likely due to proton ex-
change of the methyl diazonium ion intermediate that gives
rise to the alcohol product (12). An experiment was carried
out to quantitate the extent of acetate ion stimulated
denitrosation of a nitrosiminium ion; this is discussed later.
A series of experiments was carried out to determine the
partitioning of the nitrosiminium ions, derived from chloro-
acetate esters, between capture by water, ethanol, and a
second alcohol, as indicated in eq. [3]. The observed prod-
uct ratios were corrected for the differences in nucleophile
3
buffer (50% anion) and 10% by volume trifluoroethanol was
identical with that in buffered water within experimental er-
ror. The second-order rate constants for reaction of azide ion
were taken as the slopes of plots of kobsd against azide ion
concentration that contained three points. The rate constants
for addition of water to the cations derived from 3k and 4k
6
–1
6
–1
are 1.15 (± 0.1) × 10 s and 1.35 (± 0.1) × 10 s , respec-
©
1999 NRC Canada

Chahoua et al.
1153
Table 3. Product yields for the decomposition of N-methyl-N-nitroso-N-arylmethyl
acetates at 25°C, ionic strength 1 M, NaClO4.
O
Ar
kH2O[H2O]
NO
NO
NO
+
k
EtOH[EtOH]
[
3]
N
OAcCl
N
N
O
Et
Ar
-
Ar
Ar
OAcCl
NO
N
kROH[ROH]
OR
Ar
indicate that the value of kobsd at 0.01 M azide ion is less than
0% larger than the intercept extrapolated to 0 M azide ion.
concentrations to give ratios of second-order rate constants,
and these are summarized in Table 4.
1
Studies were carried out to trap the similarly generated
nitrosiminium ions by azide ion. A plot of the fraction of
azide ion adduct as a function of azide ion concentration is il-
lustrated in Fig. 3. Control experiments establish that there is
no measurable effect on the rate constant, kobsd, for the decay
of the esters with increasing concentrations of azide ion up to
Discussion
Cation formation
Rate constants for the pH independent solvolysis of α-
acetoxy, α-chloracetoxy, and α-azido (N-nitrosomethyl-
amino)arylmethane amines were measured at 25°C, in aque-
0.01 M. Plots of kobsd against azide ion concentration contain-
ing 4–5 azide ion concentrations between 0.001 and 0.01 M
ous media of ionic strength 1 M (NaClO ) that contained 4%
4
©
1999 NRC Canada

1
154
Can. J. Chem. Vol. 77, 1999
Table 4. Ratios of second-order rate constants for the capture of N-nitrosiminium ions formed in the decay of (N-nitrosomethyl-
amino)arylmethyl chloroacetates in 0.05 M morpholino ethane sulfonic acid (MES) buffer (pH = 6.5), ionic strength 1 M (NaClO ),
4
a
in 90:1:9 H O:EtOH:ROH (v:v) at 25 C.
2
Alcohol
Substituent
H
p-Cl
m-Cl
p-CF3
kROH/kH2O
CH -OH
7.9
7.8
5.6
6.5
3
(
7.1
0.1)
(0.1)
6.1
(0.2)
3.9
(0.6)
5.2
CH -CH -OH
3
2
(
2.4
(
0.2)
(0.1)
2.11
(0.08)
0.85
(0.1)
1.63
(0.02)
0.61
(0.3)
1.54
(0.02)
CH O-(CH ) OH
3
2 2
0.05)
ClCH -CH -OH
1.09
2
2
(
0.01)
0.58
0.04)
0.15
0.01)
0.075
0.04)
(0.03)
0.45
(0.01)
0.21
(0.01)
0.040
(0.003)
(0.03)
0.49
(0.01)
0.18
(0.01)
0.078
(0.005)
CNCH -CH -OH
0.41
2
2
(
(0.02)
0.14
Cl CH -CH -OH
2
2
2
(
(0.02)
0.037
(0.001)
CF -CH OH
3
2
(
a
Rate constants as in eq. [3].
+
Fig. 3. Plot of the fractional yield of azide ion adduct against
azide ion concentration for the reaction of 1b (squares), 2b
Fig. 4. Plot of log k1 against σ for the solvolysis of acetates,
chloroacetates, and azides at 25°C, ionic strength 1 M (NaClO4).
(
triangles), 3b (diamonds), 4b (circles), and 8b (inverted
triangles) at 25°C, ionic strength 1 M (NaClO4).
1
0
0
0
0
.8
.6
.4
.2
0
0
0.002 0.004 0.006 0.008 0.01 0.012
Azide Ion] M
[
The negative values of the slopes are consistent with posi-
tive charge buildup in the transition state for formation of
nitrosiminium ion intermediates, as in eq. [2]. These values
by volume acetonitrile. Figure 4 is a plot of the log k
1
+
2
against σ . The slopes of the plots are –2.23 (r = 0.949),
2
2
+
–
1.76 (r = 0.993), and –1.61 (r = 0.994) for the azides, ac-
are considerably smaller than the values of ρ -–5 for the
etates, and chloroacetates, respectively. The uncertainty in
the value for the azides is greatest due to the fewer data
points, but the slope is likely greater than for the other sys-
tems. Inspection of the plot shows that the deviations of the
four points in the azide correlation are similar in the esters.
solvolysis of simple benzylic and phenethyl substrates (13).
This is expected due to the substantial resonance stabilizing
effect of the N-nitrosomethylamino substituent that can do-
nate electrons to the developing electron deficient center in
+
the transition state. Similar smaller values of ρ and ρ have
A plot of the log k values for the azide (y-axis) against
been reported for systems in which there are additional
substituents in the α-position that are capable of stablization
by resonance electron donation (14).
1
those for the same substituents in the acetate series (x-axis)
2
is linear with a slope of 1.31 (r = 0.998). Plots of log k
1
2
against σ give slightly larger slopes of –3.53 (r = 0.834),
1.98 (r = 0.978), and –1.90 (r = 0.988) for the azides, ac-
2
2
–
Cation capture
etates, and chloroacetates, respectively, when the point for
the para-methoxy substituent is excluded. This point devi-
ates positively from all three such correlations, by 0.3, 0.56,
and 0.46 log units for the azides, acetates, and chloro-
acetates, respectively.
There have been numerous reports indicating that the
dominant reaction of nitrosiminium ions with alcohol,
carboxylic acid, and water solvents involves carbon–oxygen
bond formation to give α-oxy substituted nitrosamines (3–8,
15–17). That this is the dominant reaction here, at neutral
©
1999 NRC Canada

Chahoua et al.
1155
pH, is confirmed by the data in Table 3. It was recently reported that nitrosiminium ions can undergo denitrosation (7), in par-
ticular in the presence of azide ion and a thiolate ion. The last two experiments in Table 3 were carried out as a check on the
extent of this reaction with solvent water and dilute (0.05 M) buffer. The yields of methylammonium ion of 7% confirm that
there is denitrosation of the nitrosiminium ion in weakly buffered solvent, although it is clearly a minor pathway. The domi-
nant pathway involves the production of methanol by formation of the α-hydroxynitrosamine, which decomposes to the
methanediazonium ion that yields methanol. The less than quantitative yield is almost certainly accounted for by proton ex-
change involving the diazonium ion that occurs to an appreciable extent at this pH in the presence of low buffer concentra-
tions (12). It must be noted that the literature does contain a single report of what appears to extensive denitrosation of a
nitrosiminium ion intermediate by solvent (18). Decomposition of the cyclic α-acetoxynitrosamine, eq. [4], in aqueous media
gives a 54% yield of myosmine that presumably arises by denitrosation that is competitive with α-carbon–oxygen bond for-
mation.
AcO
N
NO
N
[
4]
N
N
myosmine
The common ion rate depression illustrated in Fig. 1 indicates that acetate ion acts predominantly as a nucleophile at the
α-carbon to yield the starting α-acetoxynitrosamine. The effect of acetate appears to diminish at high acetate concentra-
tions such that there is remnant of reaction that is not further inhibited by the addition of acetate ion. This has been re-
ported previously (6) and was analyzed according to the mechanism of eq. [5] in which acetate can react to either
NO
kb[AcO-]
N
OH
OH
NO
N
NO
N
R
k1
+
OAc
[
5]
R'
R
R
R'
R'
AcO -
k-1ac [AcO-]
NO
N
kH2O
+
R
R'
regenerate starting material, by k–1ac, or assist formation of the α-hydroxynitrosamine by general base catalyzed addition of
water, k . An alternative to this scheme would be either nucleophilic attack, or general base catalyzed attack of water, by ace-
b
tate ion, on the nitrogen of the nitrosiminium ion to yield the imine which subsequently hydrolyzes. This is illustrated in the
mechanism of eq. [6].
NO
kb[AcO-]
N
OH
NO
N
NO
N
R
k1
+
OAc
R'
[
6]
R
R
R'
R'
AcO -
kcac[AcO-]
NO
N
kH2O
+
OH
R
kn[AcO-]
R'
+ R'CHO
N
+
R
RNH (H )
2
R'
An experiment comparing common ion rate depression in D O with the yield of methylammonium ion at high acetate ion
2
concentration indicates that denitrosation of the nitrosiminium ion is a minor pathway. The common ion rate depression of 2a
in D O was determined at 3 M ionic strength (NaClO ) at acetate concentrations up to 1.35 M, the data are shown in Fig. 2.
2
4
The most general mechanism, eq. [6], includes both general base catalyzed addition of water attack at carbon, k , and
b
©
1999 NRC Canada

1
156
Can. J. Chem. Vol. 77, 1999
Table 5. Partioning rate constant ratios for the reactions of (N-nitrosomethylamino)arylmethyl acetates in aqueous acetate buffer at
5°C (µ = 1, NaClO4).
2
k–1 /kH2Oa
ac
kb/kH2Oa
kACO–/kH2Ob
Ring substituent
–
1
–1
–1
(
M )
(M )
(M )
p-MeO
H
27.48 (0.57)
17.97 (0.34)
16.79 (2.94)
16.06 (2.23)
0.71 (0.06)
0.99 (0.05)
0.71 (0.48)
0.68 (0.42)
28.19
19.08
17.49
16.68
p-Cl
m-Cl
a
Values determined from common ion inhibition experiments, constants defined in eq. [5]. Standard errors in parentheses.
b
ac
kACO–/kH2O = k–1 /kH2O + kb/kH2O
.
denitrosation via k , by whatever mechanism. The rate equation for this mechanism is shown in eq. [7]. The best fit of the
n
data to this equation is shown as the solid line in Fig. 2 for which the following values are used for the given parameters:
k1
[
7]
kobsd =






1
 +1
k
H O
k + k 
2
n
b
+

–

k [AcO ]
kc

c

k1 = 0.00051 (± 0.00002), kH2O/k = 0.054 (± 0.006), and (k +k )/k = 0.077 (± 0.008). At 1.35 M acetate ion, based on these
c
n
b
–1
parameters, 68% of the remaining reaction entails acetate assisted product formation, by k + k . The yield of
b
n
1
methylammonium ion at 1.35 M acetate ion was determined by H-NMR to be 12%. This represents an upper limit of the part
of the acetate ion assisted product formation that involves denitrosation. This 12% is an upper limit because it is assumed that
no methylammonium ion is formed in the remaining 38% of the reaction that involves product formation through pathways
that are not assisted by acetate ion. Based on this analysis then, less than 18% (= 100 × 12%/68%) of the acetate assisted
product formation involves denitrosation, and the remaining 82% must involve general base catalyzed attack of water on the
cation, consistent with our original notion eq. [6].
The data on common ion inhibition at 1 M ionic strength in Fig. 1, and that not shown for compound 4a, were treated ac-
cording to the mechanism of eq. [6], which ignores the minor denitrosation pathway. The appropriate expression for kobsd is
indicated in eq. [8] and best fits to the data in Fig. 1, solid lines were drawn using the values for the parameters indicated in
Table 5.
k1
[
8]
kobsd =






1
H O
 +1
k
k 
b
2
+

–

k [AcO ]
k

c
c 
The reactivity of azide ion with several cations was also investigated. One method involved the azide ion trapping studies, for
which the data are included in Fig. 3. Azide ion proves to be quite effective, with limiting amounts of the azide ion product be-
ing reached in the millimolar concentration range. Figure 3 indicates that in two cases the upper limit of the trappable fraction of
cation falls well below 1.0. This has been observed previously and ascribed to azide ion stimulated denitrosation that competes
with N—C bond formation, as illustrated in the mechanism of eq. [9] (7). There is no clear correlation of substituents with the
NO
kcaz[N3-]
N
N3
NO
N
NO
N
R
+
k1
OAc
R'
[
9]
R
R
R'
R'
NO
N
-
kH2O
+
AcO
OH
R
knaz[N3-]
R'
N
+
+ R'CHO
RNH (H )
2
R
R'
©
1999 NRC Canada

Chahoua et al.
1157
Table 6. Partioning rate constant ratios for the reactions of N-(nitrosomethylamino)arylmethyl chloroacetate with azide ion in cacodylic
acid buffer at 25°C, ionic strength 1 M (NaClO4).a
p-MeO
H
p-Cl
m-Cl
p-CF3
kN3/kH2Ob
knaz/kcaz
a
6300
3000
2700
(100)
0.44
2100
1600
(100)
0.21
(
200)
0.047
0.027)
(300)
0.055
(0.005)
(100)
0.058
(0.011)
(
(0.02)
(0.04)
Rate constants defined in eq. [9], standard errors in parentheses.
b
az
az
kN3/kH2O = (kn + kc )/kH2O
.
Table 7. Rate and equilibrium constants for reactions of esters, azides, and N-nitrosiminium ions.a
p-MeO–
H
p-Cl
m-Cl
9
az
–1 –1
0.83b
0.14d
0.38
180
c
c
2.6c
1.35
2.2
0.27
0.077
1
1
1
1
1
0 k (M s )
1.8
1.9
c
6
–1
0.76e
1.4
1.15e
1.9
e
0 k
(s )
H2O
ac
7
–1 –1 f
0 k
(M s )
–
1
0 K (M)g
1
1
ac
3.5
2.7
1.2
1.1
1
1
4
az
h
0 K (M)
140
1
a
Rate and equilibrium constants are for processes defined in eqs. [10] and [11].
b
az
+
Calculated by extrapolation of the plot of log k against σ determined from the data for the three other compounds for which the values were
c
az
+
2
directly determined (log k = 0.416σ + 9.25, R = 0.989).
c
c
az
az
Directly measured values of k combined with kn /kc from the trapping studied, Fig. 2 and Table 2.
az
d
e
az
az
az
az
Calculated from kc /kH2O based on the measured values of kN3/kH2O and kn /kc and the calculated value of kc
.
Determined by laser flash photolysis in the present report and as previously published (8).
f
ac
Calculated from k–1 /kH2O (Table 5) and the determined or calculated value of kH2O.
g
h
ac
From the values of k (Table 2) and k
.
.
1
–1
–1
az
From the values of k (Table 2) and k
1
limiting fraction of azide adduct. The two substituents with
NO
N
NO
N
k1
the smallest limiting fractions, the p-Cl and p-CF , are elec-
-
3
[
[
11]
12]
OAc
+ AcO
tron withdrawing, but the m-Cl substituent of intermediate
character between the two gives azide adduct in excess of
R
kcac
Ar
R'
9
0%. The appropriate expression for the fraction of azide
adduct (F ) as a function of azide ion concentration is given
NO
N
az
NO
k1
+
in eq. [10]. Best fits to the data are indicated by the solid
-
N3
N
+ N3
kcaz
R
Ar
'
R
1
[
10]
F =
az
az
az
k
photolysis and the ratio knaz/kcaz in Table 6 that were deter-
mined from the azide ion trapping analysis. No correction
for the reaction of water at nitrogen was made, as this ap-
pears to be minimal (vide supra). For the p-methoxyphenyl
cation, the value for kc was calculated from linear extrapo-
lation of the plot of log kc against σ for the unsubstituted,
p-Cl and m-Cl compounds. The value of k was calculated
kn
H O
2
1
+
+
az
kc
kc [N3 ]
–
lines in Fig. 3 using values for the parameters that are in-
az
az
az
cluded in Table 6. The data for k /k
(= (kn + kc )/kH2O
)
N3 H2O
az
+
are in reasonable agreement with the ratios obtained from
the rate constants that were directly measured for three of
the compounds using laser flash photolysis (see Results).
H2O
az
from the value kc and the appropriate combination of ratios
The values of k /k
ing electron-withdrawing character of the substituent, the
show a small decrease with increas-
in Table 6. Values for the reaction at carbon by acetate ion,
N3 H2O
ac
kc , were then determined from the values of k
and the
H2O
+
2
ac
az
ac
plot of log k /k
against σ has a slope of –0.44 (r =
values of kc /kH2O in Table 5. The values of kc and kc can
N3 H2O
0
.961, maximum deviation in y = 0.1 log unit).
Based on the quantitative relationships established in Ta-
then be combined with values of k (Table 2) for the decom-
1
position of the azide and acetate adducts, respectively, to
give the equilibrium constants for cation formation.
The equilibrium constants for cation formation from the
acetates exceed those for formation from the azide, for a
given substituent, by a factor of -1000.
Comparison of some of the values in Table 7 with some
previously reported similar numbers for other α-substituted
arylmethyl cations is instructive relative to the cation stabi-
lizing ability of the N-methyl-N-nitroso group. Scheme 1 il-
lustrates the relevant cations and their equilibrium and rate
constants for reaction with solvent water (19). As can be
bles 5 and 6 and the directly measured rate constants for re-
actions of azide and water with the cations, microscopic
constants can be calculated that allow a clearer view of the
cation stabilizing characteristics of the N-methyl-N-nitroso
moiety. The equilibria of interest appear in eqs. [11] and
[
12]. A summary of the relevant rate and equilibrium con-
az
stants appears in Table 7. Rate constants, kc , for reaction of
azide ion at the α-carbon of the cations with H, p-Cl, and m-
Cl substituents were determined from the rate constants for
disappearance of the cations measured by laser flash
©
1999 NRC Canada

1
158
Can. J. Chem. Vol. 77, 1999
Scheme 1.
MeO
N3
MeO
N3
N3
N3
+
+
-
-
+
N3
+
N3
K1az
4.4 x 10-12
M
=
K1az
= M
8 x 10-14
kH2O
=
2 x 109s-1
6 x 107s-1
kH2O
=
MeO
N3
MeO
N3
N3
N3
+
+
-
-
+
N3
+
N3
OMe
OMe
OMe
OMe
K1az
kH2O
=
1.4 x 10-10
3 x 107s-1
M
K1az
kH2O
9 x 10-12
=
M
=
3 x 105s-1
=
seen for both the benzylic and p-methoxybenzyl systems, the
equilibrium constants for formation of the α-nitrosomethyl-
amino cations are slightly smaller than those for formation
of the analogous α-azido cations, right pair in Scheme 1, by
less than 1.1 kcal/mol. The equilibrium constants for α-
nitrosomethylamino cations are smaller by 3 and 2.7 kcal/mol
than the corresponding α-methoxy-benzyl and p-methoxy-
benzyl cations, respectively. Thus, regarding equilibrium
stablization, the α-nitrosomethylamino group is the least ef-
fective. In contrast, the α-nitrosomethylamino cations are
the most kinetically stable of the cations. They react with
water with rate constants that are a factor of 80 and 2 slower
than the corresponding α-azido-benzyl and p-methoxybenzyl
ditionally, there is likely the contribution of re-establishment
of resonance delocalization in the product cation-hydrates
that may contribute to the larger intrinsic barriers for capture
of the α-azido and α-nitrosomethylamino cations. There is
no resonance delocalization involving the α-methoxy group
in the product of water attack on the α-methoxy cation.
It is surprising that despite the slightly greater stability of
the -azido cations compared to the -nitrosomethylamino
α
α
cations, the ρeq = –3.8 for formation of the α-azido cations is
more negative than ρ = –2.7 for the α-nitrosomethylamino
eq
2
cations. These values suggest greater delocalization into the
benzene ring in the case of the α-azido system than for the
α-nitrosomethylamino system. It is possible that some of the
carbocation stabilization in the case of the α-azido cations
may arise from the relatively strong NϵN triple bond in the
imino-diazonium ion character of one of the resonance
3
cations, respectively; and factors of 10 and 40 slower than
the corresponding α-methoxybenzyl and p-methoxybenzyl
cations, respectively.
3
Amyes and Richard (19) rationalized the contrasting ki-
netic and thermodynamic effects in Scheme 1 — the smaller
rate constants for cation capture of the thermodynamically
less stable α-azido cations — as being due to a larger intrin-
sic barrier for cation capture in the case of the α-azido cat-
ions due to loss of more extensive resonance delocalization
compared to the α-methoxy cations. The more extensive loss
of resonance contributes to a larger intrinsic barrier (20a).
This explanation appears applicable to the present case. Res-
onance in the α-nitrosomethylamino cations is similarly more
extensive than in the case of the α-methoxy cations, and so,
the intrisic barrier for capture may be larger despite the
lesser thermodynamic stability of the former cations. It
should be noted that not only is there a more extensive loss
of resonance stablization in the case of the α-azido and α-
nitrosmethylamino compared to α-methoxy cations, but ad-
forms.
It is notable that the values of kH2O are similar for the
comparable α-azido and α-nitroamino cations, given the dif-
ferences in the functionalities involved. A range of Lewis
structures that need to be considered is illustrated in
Scheme 2 and highlights important uncertainties suggested
by the present results. It is the shift, in going from the cation
to the transition state, in the dominance of the various Lewis
forms, and the consequent changes in bond lengths and an-
gles, that contributes to the larger intrinsic barrier for cap-
ture of the α-azido and α-nitrosmethylamino cations
compared to the α-methoxy cations. The structure of the
transition state is a weighted average of reactant and prod-
uct, and probably other, structures. The larger values of kH2O
for the α-azido cations than for the α-nitrosmethylamino cat-
ions may be due to less resonance delocalization in the α-
2
We had not initially noted this but it was kindly pointed out by a referee (see ref. 20b).
Prof. J. Richard (SUNY, Buffalo) suggested this possibility.
3
©
1999 NRC Canada

Chahoua et al.
1159
Scheme 2.
2
azido cation than in the α-nitsosomethylamino cation. This
would be consistent with what is expected on the basis of
the electron-withdrawing abilities of the diazonium ion and
(r = 0.948) for the acetates and azides, respectively. This is
+
considerably smaller than the value of ρ = –9 for the equi-
librium formation of 1-phenethyl cations from the chlorides
–
–
+
nitroso groups. The value of σ is σ = 1.46 for the nitroso
(13, 22). The smaller values of ρ are consistent with consid-
–
group, while σ = 3–3.4 for the diazonium ion (21). Thus,
erable resonance donation by the nitrosomethylamino group
in the cation that lessens the demand for stabilization of the
cation by substituents in the benzene ring. Similar smaller
structure a (Scheme 2) should be relatively less important to
the total structure of the α-azido cation than structure b is to
the nitrosomethylamino cation. The smaller resonance stabi-
lization by the α-azido group should give rise to a smaller
intrinsic barrier — consistent with the larger rate constants
kH2O for capture of these cations by water.
+
values of ρ are seen in the acid catalyzed formation of
oxocarbocations from substituted benzaldehyde diethyl
+
acetals for which ρ = –3.3 (14).
Comparison of the equilibrium constants for the azide and
acetate adducts with identical substituents yields information
on the differences in carbon basicity between acetate and
azide ions. This is because the reaction of azide ion, as well
as acetate, with the unsubstituted nitrosomethylamino benzyl
cation appears to be significantly activation limited, not sim-
ply diffusion controlled. The rate constant for the azide ion
reaction was measured directly by laser flash photolysis (8)
and the increase in the rate constant for azide ion capture of
the m-chloro compound shows that the reaction with the
unsubstituted compound cannot be diffusion limited. In ad-
dition, it was shown (8) that a thiolate nucleophile reacts
with the nitrosomethylamino benzyl cation with a rate con-
Of the product structures, the contribution of d to the total
structure of α-azido compound is likely significantly greater
than that of c; and this difference is likely larger than the
difference in contribution of f compared to e in the α-
nitrosamino compound. This deduction is based on the rela-
tively energetically unfavorable unpaired electrons on the
anionic nitrogen in c compared to those on the amino nitro-
–
gen in e and on the differences in σ , mentioned above. All
things being equal, this would suggest more bond length and
angle reorganization in going to the transition state for cap-
ture of the α-azido cation. This would lead to the expecta-
tion of larger barriers for capture of the α-azido cations —
opposite the experimental observation. But Scheme 2 shows
that the comparison is not simple: the relevant resonance
forms have different bond multiplicities and orbital struc-
tures in the α-azido compared to the α-nitrosaminomethyl
system, and these differences likely impact on the extent of
structural reorganization in ground and transition states and
in the degree to which this is manifest in the intrinsic barri-
ers. An alternative explanation is that there is relatively little
contribution to the transition state of the product-like reso-
nance forms, but this seems unlikely. It is the change from
the dominant resonance forms of the cations to the different
dominant product-like resonance forms in the transition state
that contributes to the larger intrinsic barrier for capture of
the α-azido cations compared to the α-methoxycarbocations.
9
–1 –1
stant of 2.8 × 10 M
ion. Further, the rate constant for reaction with azide ion is
s
that is larger than that for azide
9
–1 –1
significantly smaller than the values of 5–7 × 10 M
s
that have been directly measured for the reaction of azide
ion with diaryl- and triarylmethyl cations that do not appear
to be appreciably less hindered than the nitrosiminium ions
studied here (23). The smaller rate constant for the reaction
8
–1
of acetate ion is also smaller than the value of -5 × 10 s ,
which represents the upper limit for encounter-controlled
capture of cations by acetate ion (24). Thus, the differences,
between the acetate and azide adducts, in the equilibrium
constants for cation formation reflect the balance of differ-
ences in intrinsic carbon basicity and (or) other properties
between the two nucleophiles.
The equilibrium constants for cation formation show only
small changes as a function of the substituent on the ben-
zene ring. The values of ρ are –2.41 (r = 0.993) and 2.67
The observed differences in equilibrium constants can al-
most entirely be accounted for by differences in solvation of
the acetate and azide anions. This indicates, surprisingly,
+
2
©
1999 NRC Canada

1
160
Can. J. Chem. Vol. 77, 1999
Fig. 5. Plot of log (ROH/H2O) against pKa ROH for the
alcohol–water selectivity of cations from chloroacetates 2b
(circles, C = 1.5), 3b (triangles, C = 1), 4b (diamonds, C = 0.5),
8b (triangles, C=0) at 25°C, ionic strength 1 M (NaClO4),
that there is little difference between the carbon basicity of
acetate and azide ions. This is most easily understood by
combination of the two equilibria for cation formation to
give eq. [13] below.
solvent H O/EtOH/ROH = 90/1/9.
2
NO
N
NO
N
3
2
1
0
Ktx
-
-
[
13]
OAc + N3
N3
+ AcO
Ar
Ar
Appropriate combination of the two equilibrium constants
gives the value of K = 1300 for transfer of the carbon frag-
tx
ment between the two nucleophiles, 4.3 kcal/mol in favor of
the azide adduct. This preference is a combination of the dif-
ferences in intrinsic carbon affinity of the two nucleophiles
as well as the differences in the strength of solvation of the
two anions, and the two neutral species as well. The free en-
ergy of transfer from the gas phase to water is larger for ace-
tate ion than azide ion by 6.6 kcal/mol (25). The free energy
of transfer from the gas phase to water is larger for acetic
acid than hydrazoic acid by 3.3 kcal/mol (25). If is is as-
sumed that there is a similar difference for the acetate adduct
compared to the azide adduct in eq. [13], then 3.3 (= 6.6 –
-
-
1
2
1
2
13
14
15
16
17
pKa ROH
3
.3) kcal/mol of the 4.3 kcal/mol for the equilibrium in
lutions, ionic strength 1 M (NaClO ), and show that there is
little change in selectivity as a function of the aryl group
substituent. This is illustrated in Fig. 5, which plots the alco-
4
eq. [13] is due to the stronger solvation of acetate ion than
azide ion. This means that the difference between azide ion
and acetate ion in intrinsic carbon basicity is on the order of
hol water selectivity as a function of alcohol pK . For the p-
a
1
kcal/mol in favor of azide ion. The assumption that the
H, p-Cl, and p-CF substituents, the slopes of the least-
3
difference in solvation of the neutral species is the same as
that for the corresponding neutral acids may be incorrect if
some of the difference in the case of the acids is due to
stronger hydrogen bond donation by acetic acid than by
2
2
squares correlations are 0.62 (r = 0.983), 0.66 (r = 0.989),
.60 (r = 0.989), respectively. The m-Cl substituent, of in-
2
0
termediate electron-withdrawing power, gives the smallest
2
slope, β = 0.52 (r = 0.989). The reason for this apparent
HN . This would make the difference in free energy of trans-
fer of the neutral species in eq. [13] less than 3.3 kcal/mol,
which would indicate that a larger fraction of the observed
3
anomaly is not understood. The intermediate values of β sug-
gest there is considerable charge buildup on the attacking al-
cohol, and significant bond formation in the transition state
for cation capture by these weak nucleophiles.
4
.3 kcal/mol for the equilibrium in eq. [13] would be ac-
counted for simply by differences in the solvation of the an-
ions. This makes the calculated difference of 1 kcal/mol
between the carbon basicities of azide and acetate ions an
upper limit.
Notably, Fig. 5 indicates that trifluoroethanol (pK = 12.4)
a
behaves as a normal nucleophile for its pK , and this con-
a
trasts with previous reports. In the capture of 1-phenylethyl
cations (24) and oxocarbocations (27), trifluoroethanol is a
significantly poorer nucleophile than expected for its pKa.
The reason for the different, “normal” behavior in compari-
son with other alcohols in the present case is unclear.
The recognition of the important contribution of solvation
to the equilibrium in eq. [13] may also account for the unex-
pected observation that the difference, between acetate and
azide ion, in the rate constants for the highly exoergic cap-
ture of the cations is larger than the difference, between
azide and acetate adducts, in the rate constants for the highly
endoergic formation of the cations. Table 7 shows that the
Acknowledgements
We are grateful to Prof. J. Bartmess (Tennessee) for pro-
viding the information on free energies of transfer from the
gas phase to water. This work was supported by grants RO1-
CA 52881 and KO4-CA 62124 from the U.S. National Insti-
tutes of Health (NCI).
az
ac
ratio, kc /kc , is a factor of -100, whereas Table 2 indicates
that the values of k for the azides is only a factor of 10
1
smaller than the acetates. The requirement for extensive
desolvation prior to the small amount of bond formation that
likely exists in the transition state for the highly exothermic
cation capture reaction could explain the larger difference in
reactivity between the two nucleophiles, given the stronger
solvation of the acetate ion that exists (vide supra). More
facile desolvation of azide ion has been invoked previously
to account for the larger limiting rate constants for reaction
of azide ion compared to carboxylate ions and amines with
carbocations (24, 26).
References
1
. P.D. Lawley. In Chemical carcinogens. ACS monograph 182.
Edited by C.D. Searle. American Chemical Society, Washing-
ton, D.C. 1984.
2
. W. Lijinsky. Chemistry and biology of N-nitroso compounds.
Cambridge University Press, Cambridge. 1992.
The reactivities of a series of alcohols with four of the
3. J.E. Baldwin, A. Scott, S.E. Branz, S.R. Tannenbaum, and L.
cations was determined in 90% H O/1% EtOH/9% ROH so-
Green. J. Org. Chem. 43, 2427 (1978).
2
©
1999 NRC Canada

Chahoua et al.
1161
4
5
. C. Revis, M. Rajamaki, and J.C. Fishbein. J. Org. Chem.
0 7733 (1995).
. L. Chahoua, M. Mesic’, A. Vigroux, and J.C. Fishbein. J. Org.
16. M. Mochizuki, T. Anjo, N. Sekiguchi, A. Ikarashi, A. Suzuki,
Y. Wakabayashi, and M. Okada. Chem. Pharm. Bull. 34, 3956
(1986).
6
Chem. 62, 2500 (1997);H. Cai and J.C. Fishbein. Tetrahedron,
17. M. Wiessler. N-Nitrosamines. Edited by J.-P. Anselme. ACS,
Washington, D.C. 1979. p. 57.
1
0 671 (1997).
6
7
. H. Cai and J.C. Fishbein. J. Am. Chem. Soc. In press.
. M. Rajamaki, A. Vigroux, L. Chahoua, and J.C. Fishbein. J.
Org. Chem. 60, 2324 (1995).
18. S.S. Hecht and C.B. Chen. J. Org. Chem. 44,1563 (1979).
19. T.L. Amyes and J.P. Richard. J. Am. Chem. Soc. 113, 1867
(1991).
8
9
. A. Vigroux, A.J. Kresge, and J.C. Fishbein. J. Am. Chem. Soc.
20. (a) C.F. Bernaconi. Tetrahedron, 41, 3219 (1985); J.P. Richard.
Tetrahedron, 51, 1535 (1995); (b) J.P. Richard, T.L. Amyes, V.
Jagannadham, Y.G. Lee, and D.J. Rice. J. Am. Chem. Soc.
117, 5198 (1995)
117, 4433 (1995).
. P.G. Stecher (Editor). The Merck index. 8th ed. Merck & Co.,
Rahway, N.J. 1968.
1
0. M. Wiessler. Angew. Chem. Int. Ed. 13, 743 (1974).
21. O. Exner. In Correlation analysis in chemistry, recent ad-
vances. Edited by N.B. Chapman and J. Shorter. Plenum, New
York. 1978.
22. N.C. Deno and W.L. Evans. J. Am. Chem. Soc. 79, 5804
(1957).
11. J. Graymore. J. Chem. Soc. 1353 (1932).
1
2. J. Hovinen. J.I. Finneman, S.N. Satapathy, J. Ho, and J.C.
Fishbein. J. Am. Chem. Soc. 114, 10 321 (1992), and
refs. within.
1
1
3. J.P. Richard, M.E. Rothenberg, and W.P. Jencks. J. Am. Chem.
Soc. 106, 1361 (1984).
4. R.B. Dunlap, G.A. Ghanim, and E.H. Cordes. J. Phys. Chem.
23. R.A. McClelland, V.M. Kanagasabapathy, N.S. Banait, and S.
Steenken. J. Am. Chem. Soc. 113, 1009 (1991).
24. J.P. Richard and W.P. Jencks. J. Am. Chem. Soc. 106, 1373
(1984).
7
3, 1898 (1970); P.R. Young and W.P. Jencks. J. Am. Chem.
Soc. 99, 8238 (1977).
25. B. Wilson and J.E. Bartmess. J. Am. Chem. Soc. 113, 1762
(1991).
26. R.A. McClelland, V.M. Kanagsabapathy, N.S. Banait, and S.
Steenken. J. Am. Chem. Soc. 114, 1816 (1992).
27. M.L. Sinnott and W.P. Jencks. J. Am. Chem. Soc. 102, 2026
(1980).
1
5. M. Wiessler, G. Rossnagel, and B. Rugewitz-Blackholm. N-
Nitroso compounds: occurrence, biological effects and
relevence to human cancer. Vol. 52. Edited by I.K. O’Neill,
R.C. von Borstel, R.C. Miller, J. Long, and H. Bartsch. IARC
Sci. Pub., Lyon. 1977. p. 465; N. Frank and M. Wiessler. Car-
cinogen, 7, 365 (1986).
©
1999 NRC Canada