Regioselectivity and stereoelectronic effects in the reactions of the dinitroaniline herbicides trifluralin and benefin with nucleophiles

Article abstract of DOI:10.1139/v98-089

The reactions of two members of the dinitroaniline class of herbicides, N,N-di-n-propyl-2,6-dinitro-4-(trifluoromethyl)aniline (trifluralin; 1) and N-ethyl-N-n-butyl-2,6-dinitro-4-(trifluoromethyl)aniline (benefin; 2), along with their analogue, N-phenyl-2,6-dinitro-4-(trifluoromethyl)aniline (3), with the nucleophiles, OD^- and SO₃^2-, have been investigated using 400 MHz ¹H NMR spectroscopy. The reactions of both 1 and 2 with OD^- result in formation of Meisenheimer anionic σ-complexes according to a K3T1 (kinetic preference for formation of the C-3 adduct with thermodynamic preference for formation of the C-1 adduct) reaction sequence while the reaction of 3 with OD^- and that of 1 with SO₃^2- follow a K3T3 (kinetic and thermodynamic preference for formation of the C-3 adduct) sequence. There was no observation of the C-1 adducts of 1 and 2 with OD^-, but the products of S(N)Ar displacement at C-1 were observed as the final thermodynamic products. Geometry optimization calculations support our hypothesis of n → σ* stabilization of the C-1 adduct leading to S(N)Ar displacement. In the reaction of 3 with OD^-, initial N-deprotonation to form the anion, 3a, is followed by σ-complex formation. The final thermodynamic product observed in this system is 3,5-dinitro-4-(N-phenylamino)benzoic acid formed through hydrolysis of the trifluoromethyl group on the anion, 3a. Aryl H-D exchange has been found for the systems of 1 and 2 with OD^-, but not for the SO₃^2- system and neither for the reaction of 3 with OD^-. Since dimethylpicramide showed significantly slower H-D exchange under identical conditions, it is argued that this discrepancy has as origin the ability of the amino N lone electron pair to interact with the π-system of the ring. With both 1 and 2 the larger size of the amino alkyl chains prevent the amino N lone pair from aligning with the π-system of the ring, thus hindering electron density donation to the electron-deficient ring carbons.

Full text of DOI:10.1139/v98-089

8
73
Regioselectivity and stereoelectronic effects in
the reactions of the dinitroaniline herbicides
trifluralin and benefin with nucleophiles
Michael T. Annandale, Gary W. vanLoon, and Erwin Buncel
Abstract: The reactions of two members of the dinitroaniline class of herbicides, N,N-di-n-propyl-2,6-dinitro-4-
(
trifluoromethyl)aniline (trifluralin; 1) and N-ethyl-N-n-butyl-2,6-dinitro-4-(trifluoromethyl)aniline (benefin; 2), along with
–
2−
their analogue, N-phenyl-2,6-dinitro-4-(trifluoromethyl)aniline (3), with the nucleophiles, OD and SO , have been
investigated using 400 MHz H NMR spectroscopy. The reactions of both 1 and 2 with OD result in formation of
3
1
–
Meisenheimer anionic σ-complexes according to a K3T1 (kinetic preference for formation of the C-3 adduct with
–
thermodynamic preference for formation of the C-1 adduct) reaction sequence while the reaction of 3 with OD and that of 1
2
−
with SO3 follow a K3T3 (kinetic and thermodynamic preference for formation of the C-3 adduct) sequence. There was no
–
observation of the C-1 adducts of 1 and 2 with OD , but the products of S Ar displacement at C-1 were observed as the final
N
thermodynamic products. Geometry optimization calculations support our hypothesis of n → σ* stabilization of the C-1
–
adduct leading to S Ar displacement. In the reaction of 3 with OD , initial N-deprotonation to form the anion, 3a, is followed
N
by σ-complex formation. The final thermodynamic product observed in this system is 3,5-dinitro-4-(N-phenylamino)benzoic
acid formed through hydrolysis of the trifluoromethyl group on the anion, 3a. Aryl H–D exchange has been found for the
–
2−
–
systems of 1 and 2 with OD , but not for the SO system and neither for the reaction of 3 with OD . Since
3
dimethylpicramide showed significantly slower H–D exchange under identical conditions, it is argued that this discrepancy
has as origin the ability of the amino N lone electron pair to interact with the π-system of the ring. With both 1 and 2 the larger
size of the amino alkyl chains prevent the amino N lone pair from aligning with the π-system of the ring, thus hindering
electron density donation to the electron-deficient ring carbons.
Key words: trifluralin, benefin, herbicides, Meisenheimer complexes, stereoelectronic effects.
1
Résumé : Faisant appel à la spectroscopie RMN du H à 400 MHz, on a étudié les réactions de deux herbicides membres de la
classe des dinitroanilines, soit la N,N-dipropyl-2,4-dinitro-4-(trifluorométhyl)aniline (trifluraline; 1) et la
N-éthyl-N-butyl-2,6-dinitro-4-(trifluorométhyl)aniline (bénéfine, 2), et de leur analogue, la N-phényl-2,6-dinitro-4-
–
2−
–
(
trifluorométhyl)aniline (3), avec les nucléophiles OD et SO3 . Les réactions des composés 1 et 2 avec OD conduisent à la
formation de complexes σ anioniques de Meisenheimer suivant une séquence réactionnelle K3T1 (préférence cinétique pour
la formation de l’adduit C-3 avec préférence thermodynamique pour la formation de l’adduit C-1); par ailleurs, la réaction de
–
2−
3
avec OD ainsi que celle de 1 avec SO d’après une séquence K3T3 (préférences cinétique et thermodynamique pour la
3
–
formation de l’adduit C-3). On n’a pas observé de formation d’adduits C-1 entre 1 et 2 et OD , mais les produits de
substitution S Ar en C-1 ont été observés comme produits thermodynamiques finals. Des calculs avec optimisation de la
N
géométrie supportent notre hypothèse d’une stabilisation n → σ* de l’adduit C-1 conduisant à une substitution S Ar. Dans la
N
–
réaction de 3 avec OD , la N-déprotonation initiale conduisant à la formation de l’anion 3a est suivie de la formation d’un
complexe σ. Le produit thermodynamique final observé dans ce système est l’acide 3,5-dinitro-4-(N-phenylamino)benzoïque
formé par hydrolyse du groupe trifluorométhyle de l’anion 3a. On a observé un échange H–D des protons aromatiques pour
–
2−
les systèmes de 1 et de 2 avec le OD ; on ne l’a toutefois pas observé ni pour le système avec SO ni pour la réaction de 3
3
–
avec OD . Puisque l’échange H–D de la diméthylpicramide est beaucoup plus lent dans des conditions identiques, il a été
suggéré que cette différence provient de l’habilité de la paire d’électrons non partagée de l’azote à interagir avec le système π
du noyau. Avec les composés 1 et 2, la taille plus grande des chaînes alkyles du groupe aminé empêche la paire d’électrons
non partagée de l’azote de l’amine de s’aligner avec le système π du noyau et ceci empêche d’accroître la densité électronique
sur les carbones électrodéficients du noyau.
Mots clés : trifluraline, bénéfine, herbicides, complexes de Meisenheimer, effets stéréoélectroniques.
[
Traduit par la rédaction]
Received December 8, 1997.
This paper is dedicated to Professor Erwin Buncel in recognition of his contributions to Canadian chemistry.
1
M.T. Annandale, G.W. vanLoon, and Erwin Buncel. Department of Chemistry, Queen’s University, Kingston, ON K7L 3N6, Canada.
1
Author to whom correspondence may be addressed. Telephone: (613) 545-2616. Fax: (613) 545-6669.
Can. J. Chem. 76: 873–883 (1998)
© 1998 NRC Canada

8
74
Can. J. Chem. Vol. 76, 1998
biological activity may arise from their ability to form stable
covalent anionic adducts (Meisenheimer complexes) with nu-
cleophiles (e.g., -NH of lysine protein residues, -SH of cyste-
ine residues, etc.) in plant cells (2–4).
The present study focuses on the reaction pathways of her-
bicides 1 and 2, and their aryl analogue, N-phenyl-2,6-dinitro-
Introduction
Trifluralin (N,N-di-n-propyl-2,6-dinitro-4-(trifluoromethyl)-
aniline; 1) and benefin (N-ethyl-N-n-butyl-2,6-dinitro-4-(tri-
fluoromethyl)aniline; 2) are members of the dinitroaniline
family of herbicides, widely used for the control of most
grasses and broad-leafed weeds. They are selective pre-emer-
gence soil-herbicides used on a variety of crops, including
beans, carrots, tomatoes, potatoes, and strawberries. They are
also used for the control of annual grasses and broad-leafed
weeds in grain and cereal crops as well as in orchards and
vineyards (1).
2
4
-(trifluoromethyl)aniline, 3, with hydroxide and sulfite ions.
1
The reactions were followed by H NMR spectroscopy in order
to eludicate the degradation products formed, and to identify
any Meisenheimer complex intermediates that may lie along
the reaction path.
1H nuclear magnetic resonance (NMR) spectroscopy has
proved to be the most useful tool in the determination of the
structures of σ-complexes in solution (5–11). For example, in
the reaction of 1,3,5-trinitrobenzene (TNB) with nucleophiles,
the ring carbon (C-2) to which the nucleophile becomes at-
2
3
tached, undergoes a change in hybridization from sp to sp ,
which affects the shielding felt by the attached proton. The
resonances of the C-4 and C-6 protons are also shifted upfield
relative to uncomplexed TNB, but to a lesser extent than the
C-2 proton, due to loss of aromaticity (reduction of ring cur-
rent) in the trinitrocyclohexadienate moiety and the increased
negative charge on the ring in the complex.
The mode of action of these herbicides is not fully under-
stood at present. However, since dinitroanilines possess an
electron-deficient aromatic ring, it has been postulated that the
©
1998 NRC Canada

Annandale et al.
875
a
Table 1. Proton NMR resonance shifts of substrates, σ-complexes, and products observed for the reactions of 1–3 with deuteroxide and
sulfite ion.
Compound
Chemical shift (ppm)
H3
H5
Ha
Hb
Hc
Hd
He
Hf
0.75 (t)
b
1
8.39 (s) 8.39 (s)
2.89 (t)
1.45 (m)
0.75 (t)
2.89 (t)
1.45 (m)
b
1
1
1
1
a
5.77 (s) 7.43 (s) 3.35 (m), 2.64 (m) 1.55 (m), 1.36 (m)
0.75 (bt)
3.35 (m), 2.64 (m) 1.55 (m), 1.36 (m) 0.75 (bt)
b
b
Not observed
b
c
Not obs. Not obs.
3.53 (m)
2.05 (bm)
Obscured
0.75 (bt)
Obscured
Obscured
2.90 (t)
3.53 (m)
2.05 (bm)
Obscured
b
d
5.50 (s) 7.28 (s) 3.39 (m), 2.62 (bm) 1.57 (m), 1.32 (bm)
3.39(m), 2.62 (bm) 1.57 (m), 1.32 (bm) 0.75 (bt)
b
cis-1e
trans-1e
5.01 (d) 7.15 (d)
5.01 (d) 7.68 (d)
8.42 (s) 8.42 (s)
Obscured
Obscured
3.02 (q)
Obscured
Obscured
1.02 (t)
Obscured
Obscured
1.40 (m)
1.46 (bm)
Obscured
Obscured
1.17 (m)
1.17 (m)
Obscured
Obscured
0.76 (t)
b
c
2
c
2
2
2
3
3
3
4
5
6
7
8
a
5.74 (s) 7.43 (s) 2.75 (m); 3.37 (m)
1.02 (t)
2.75 (m); 3.37 (m)
Not observed
Obscured
7.09 (bm)
6.69 (t)
0.76 (t)
c
b
c
c
Not obs. Not obs.
8.53 (s) 8.53 (s)
7.74 (s) 7.74 (s)
5.56 (s) 6.91 (s)
8.02 (s) 8.02 (s)
8.34 (s) 8.34 (s)
Obscured
7.09 (bm)
6.44 (d)
Obscured
N/A
2.06 (m)
7.26 (bm)
7.04 (t)
Obscured
N/A
Obscured
N/A
Obscured
N/A
Obscured
N/A
c
c
c
a
N/A
N/A
N/A
b
Obscured
N/A
N/A
N/A
N/A
b
N/A
N/A
N/A
b
b
c
c
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
2.35 (t)
2.44 (q)
6.62 (d)
1.35 (m)
0.94 (t)
7.00 (t)
1
0.79 (t)
2.35 (t)
1.35 (m)
N/A
1.35 (m)
1.35 (m)
N/A
0.79 (t)
0.81 (t)
N/A
2.39 (t)
8.06 (s) 8.06 (s)
6.60 (t)
a
Measurements obtained using a Bruker AM-400 spectrophotometer ( H: 400.1 MHz). Residual DMSO-d H
5
was used as a reference.
b
8
0% DMSO-d –20% D O (v/v) solvent.
6 2
Results
significant chemical shift differences from one another, but
have chemical shifts similar to the corresponding proton on the
other propyl group. Therefore, one can assign the four multi-
plets to the following four pairs of protons: H , H , H , and
H6,7 (using the numbering scheme portrayed in Fig. 1). Spe-
cific assignments identifying the separate N-bound methylene
proton multiplets or the C-bound methylene proton multiplets
cannot be made at this time.
The proton chemical shifts of the observed Meisenheimer
complexes, intermediates, and products for the reactions of the
three dinitroaniline compounds (1–3) with deuteroxide nu-
cleophile and for the reaction of 1 with sulfite ion are given in
Table 1. Proton-labelled structures are shown in Chart 1.
1,4
2,3
5,8
–
Reaction of trifluralin (1) and benefin (2) with OD
Upon addition of the OD solution to the substrate solutions of
–
The initial formation of the C-3 Meisenheimer adduct (1a,
2a) is followed by the appearance of aryloxide, 4, and the
1
and 2 in DMSO-d (final DMSO concentration 80% v/v),
6
amines, 6 and 7 (6 from 1; 7 from 2). In time, 4 becomes
hydrolyzed to 5 in the basic medium. Adducts 1c and 2c be-
come observable as the concentration of 6 and 7 increases in
their respective systems. However, as the reaction proceeds
and the available substrate concentration decreases, the peaks
assigned to 1c and 2c also decrease. The C-1 adducts (1b and
there was an immediate disappearance of the peak at 8.39 ppm
1) or 8.42 ppm (2) and, concurrently, the appearance of down-
field peaks at 5.77 and 7.43 ppm (1a) and 5.74 and 7.43 ppm
2a) characteristic of formation of Meisenheimer adducts
(
(
–
through addition of OD at C-3. Due to the asymmetry of 1a,
the N-propyl resonances appear to be in a pseudo-ABMNX3
spin system, similar to the ABX spin system described by
2
b) were not detected.
Reaction of both 1 and 2 with OD was accompanied by a
3
–
Servis (12) for the C-3 σ-adduct of N,N-diethyl-2,4,6-trini-
troaniline with MeO . The AB and MN methylene groups arise
–
notable decrease in the aromatic proton signals of 1 and 1a (2
and 2a) as a result of H–D exchange, which was confirmed
through product isolation and MS analysis.
–
from the asymmetric environment induced by the OD group
in a preferred conformation as shown in Fig. 1, where AB and
MN represent the two protons on the N-bound methylene
group and the two protons on the C-bound methylene group,
respectively. The methylene protons of 1a, H and H , are
Reaction of trifluralin (1) with SO3 −
2
a
d
found at 3.35 (2H) and 2.64 (2H) ppm; H and H at 1.55 (2H)
Addition of the nucleophile (Na SO –D O) to 1 in DMSO-d
b
e
and 1.36 (2H) ppm; and the methyl protons, H and H , at 0.76
2
3
2
6
(final DMSO concentration 80% v/v) resulted in the immedi-
ate reduction of the peak at 8.39 ppm (1). This was accompa-
nied by formation of the C-3 adduct (1d) with characteristic
ring proton resonances at 5.50 and 7.28 ppm; however, no C-1
adduct or S Ar displacement products were detected. With
N
time, new peaks assignable to the cis and trans diadducts of 1e
c
f
(
6H) ppm (Fig. 2).
Homonuclear decoupling (13) experiments, leading to
more specific assignments of the individual peaks, were per-
formed. The decoupling results indicated that each of the four
methylene protons on the individual propyl groups have
©
1998 NRC Canada

8
76
Can. J. Chem. Vol. 76, 1998
Chart 1.
were observed at 5.01, 7.15, and 7.68 ppm. No aromatic proton
exchange was observed for this system.
Discussion
Regioselectivity and stereoelectronic factors in reaction
–
Reaction of N-phenyl-2,6-dinitro-4-(trifluoromethyl)-
of trifluralin (1) and benefin (2) with OD
–
aniline (3) with OD
The reaction pathways for the reactions of trifluralin (1) and
benefin (2) with OD are shown in Scheme 1. The reactions of
both 1 and 2 with OD appear to follow a K3T1 sequence, i.e.,
–
Upon the addition of the base (KOH–D O) to the solution of 3
2
in DMSO-d (final DMSO concentration 85% v/v), there was
–
6
an immediate distinct colour change from orange to deep pur-
ple. Concurrently the peak at 8.53 ppm (3) disappeared and
was replaced by a singlet at 7.74 ppm assigned to the anion, 3a.
This deprotonation was followed by the appearance of a singlet
at 8.06 ppm assignable to 8, a hydrolysis product of 3a. At low
base:substrate ratios (ca. < 4:1 molar ratio) no σ-adduct forma-
tion was observed. Only with a large excess of base, specifi-
kinetic preference for formation of the C-3 adduct with ther-
modynamic preference for formation of the C-1 adduct. This
kind of behaviour has been observed with a number of oxygen
nucleophiles (14). General qualitative free energy diagrams for
common reaction sequences, including the K3T1 sequence,
are shown in Fig. 3.
Although the products of nucleophilic aromatic substitution
(S Ar) at C-1 are predicted to be the final thermodynamic
–
cally, 10:1 molar equivalents of OD :3, formation of the C-3
N
adduct (3b) was detected by the appearance of small peaks at
5.56 and 6.91 ppm. No aromatic proton exchange was ob-
served for this system.
products, the C-1 adduct was not spectroscopically observed
in the present system. Thus our assumption, that the C-1 adduct
is thermodynamically more stable than the C-3 adduct and that
©
1998 NRC Canada

Annandale et al.
877
Fig. 1. The preferred conformation of 1a showing the asymmetric environments yielding AB and MN methylene protons. (A) View with the
aromatic ring in the x–y plane. The second propyl group is not shown in this view. This perspective shows how H , H , H , and H on the same
1
2
5
6
propyl group are all in chemically different environments due to their relative positions to the OD substituent. (B) View with the aromatic ring
in the y–z plane showing the two propyl groups in different environments induced by the OD substituent. (C) View with down the
amine-N—aromatic-C bond with the aromatic ring in the x–z plane. This perspective shows the eight methylene protons in chemically
different environments arising from their relative position to the OD substituent.
1
Fig. 2. 400 MHz H NMR spectrum of 1a in 80% DMSO-d – 20% D O solution. Only the upfield aliphatic region is shown.
6
2
once formed the C-1 adduct rapidly decomposes with loss of
a secondary amine (6 from 1, 7 from 2) to yield the phenoxide
4, is supported by past findings that C-1 OH adducts have not
been observed in any system where the C-1 position is also
substituted by an acceptable leaving group (15).
been documented in 1,3,5-trinitrobenzene (TNB) and other
polynitroaromatics (16, 17), that the nitro group para to the
site of attack plays an important role in stabilizing adduct for-
mation through resonance. Since the adducts derived from 1
and 2 have a CF group rather than NO in the para-position,
–
3
2
Since there is no direct observation of the C-1 adduct in the
present system, one needs to consider the possibility that the
C-3 adduct is actually more stable than the C-1 adduct. It has
this kind of stabilization is reduced. However, there remains
the possibility of stereoelectronic stabilization of the C-1 ad-
duct (14, 18, 19) with lone pair electron density donated from
©
1998 NRC Canada

8
78
Can. J. Chem. Vol. 76, 1998
Scheme 1.
Fig. 3. Qualitative comparative energy – reaction coordinate
profiles for the four general patterns of regioselectivity. Barrier
heights and relative stabilities are exaggerated for clarity. K3T1
describes those systems in which formation of the C-3 adduct is the
product of kinetic control, while the the C-1 adduct is the thermo-
dynamic product. In the K1T1 profile the C-1 adduct is favoured by
both kinetics and thermodynamics. Conversely, K3T3 represents
the situation where the C-3 adduct is doubly preferred: by kinetics
and thermodynamics. The K1T3 profile describes the inverse
behaviour of that indicated by K3T1; now, the C-1 adduct is
favoured kinetically, but the C-3 adduct is the most stable product (14).
the oxygen of the OD moiety into the antibonding orbital of
the C-1—NR bond, and vice versa. Such n → σ* interactions
2
(20, 21) have previously been shown to be important in stabi-
lizing structurally similar C-1 adducts (14, 18, 19, 22). The
calculated (geometry optimized, PM3) structure supports the
supposition of n → σ* stabilization for the C-1 adduct 1b (vide
infra).
–
A further complexity in the reactions of 1 and 2 with OD
results from the displacement of secondary amines (6 from 1;
7 from 2); once liberated, these amines can act as nucleophiles
and compete with OD for reaction with the substrates. From
–
separate experiments in which 1 was reacted with 6 in the
–
presence of OD , it was noted that 6 formed the C-3 adduct,
1c, but no adduct resulting from attack at C-1 was seen, in
accord with a K3T3 sequence. However, an important note of
caution has been raised (14), namely, that if the rate constants
are of vastly different magnitudes and the kinetic conditions
k > k but k > > k–3 both apply (shown by the dotted lines
1
3
–1
in Fig. 3, K1T3 profile), then it may not be possible to observe
the C-1 adduct owing to the time scale of the reaction. In the
present system, it is unlikely that the C-1 adduct would be
kinetically favoured because of the greater steric interactions
associated with attack at the C-1 position (F-strain) (11, 18,
23–25).
Geometry optimization calculations on 1, 1a, and 1b yield
conformations that concur with the aforementioned hypothe-
ses. The geometry-optimized structure of 1 (Fig. 4) has both
nitro groups aligned perpendicular to the plane of the ring
©
1998 NRC Canada

Annandale et al.
879
Fig. 4. Geometry-optimized structure using SPARTAN for trifluralin (1) showing the alignment of the N,N-di-n-propyl groups such that the
amine nitrogen lone electron pair is parallel with the plane of the ring and cannot interact with the ring π-system. The ChemWindow3 structure
is shown in the centre.
Fig. 5. Geometry-optimized structure using SPARTAN for N,N-dimethylpicramide (DMP) showing the alignment of the N,N-dimethyl groups
such that the amine nitrogen lone electron pair is perpendicular to the plane of the ring and has maximum interaction with the ring π-system.
The ChemWindow3 structure is shown in the centre.
Fig. 6. Geometry-optimized structure using SPARTAN for 1a showing the alignment of the nitro group para to the site of nucleophilic attack
such that its π-system can interact with the π-system of the ring, allowing for greater charge stabilization of the anionic Meisenheimer adduct.
The ChemWindow3 structure is shown in the centre.
along with the two N-n-propyl groups. The orientation of the
nitro groups perpendicular to the plane of the ring probably
results from steric interactions with the two N-n-propyl groups.
This may be expected since the optimized geometry for N,N-
dimethylpicramide (DMP; Fig. 5) suggests significant ring
strain (indicated by the displacement of C-1 from the plane of
the ring) due to interaction between the N-methyl and the nitro
groups.
such that one of the oxygen lone pairs is approximately anti-
periplanar to the C-1—NR bond. This orientation is optimal
for n → σ* interactions where electron density is donated from
the oxygen lone electron pair into the antibonding orbital of
2
the C-1—NR bond.
2
The n → σ* stabilization indicated above provides support
for the assumption that the C-1 adduct is more thermodynami-
cally stable than the C-3 adduct in the reactions of both 1 and
3
–
The importance of the nitro group para to the sp hybridized
carbon (C-3) in 1a can be seen in the geometry-optimized
structure shown in Fig. 6; the para-NO has become aligned
2
2 with OD . This stereoelectronic stabilization does not appear
to occur from the amine-N lone pair to the C-1—OH bond
since the amine-N lone electron pair is not in an antiperiplanar
alignment to the C-1—OH bond.
with the ring for maximum π–π interaction while the ortho-
NO remains perpendicular to the plane of the ring. This con-
2
formation allows for greater delocalization of negative charge
Aryl H–D exchange in trifluralin (1) and benefin (2)
onto the para-NO group, and such stabilization is unavailable
2
to the C-1 adduct of 1. However, the geometry-optimized
structure for 1b (Fig. 7) shows the -OH substituent aligned
It was noted in the Results that in the reaction of 1 and 2 with
OD there is a significant decrease in the intensity of the reso-
–
nances due to H and H on 1 and 1a, as well as on 2 and 2a,
3
5
©
1998 NRC Canada

8
80
Can. J. Chem. Vol. 76, 1998
Fig. 7. Geometry-optimized structure using SPARTAN for 1b showing the alignment of the deuteroxide oxygen such that one of its lone
electron pairs is aligned antiperiplanar to the ring-C—NR bond allowing for n → σ* stabilization. The antiperiplanar lone electron pair is
2
shaded. The ChemWindow3 structure is shown on the right.
Scheme 2.
relative to their aliphatic signals, arising from aryl ring hydro-
gen–deuterium exchange. Exchange of polynitroaromatic ring
hydrogen by deuterium is well known to compete with σ-com-
plex formation (26–31). It has been documented that σ-com-
plexes do not themselves undergo exchange (24, 26, 32), and
neither do polynitroaryloxides (26, 33). Thus H–D exchange
occurs in the uncomplexed trifluralin and benefin molecules.
Although exchange of aryl hydrogen is well documented,
its observation in this system is unexpected for a number of
reasons. First, aryl ring H–D exchange generally occurs pref-
erentially at a position ortho to two nitro groups (26–31, 34),
whereas with 1 and 2 the aryl hydrogens are ortho to one nitro
group and a trifluoromethyl group. Equally important is the
slower rate of aryl H–D exchange in N,N-dimethylpicramide
has occurred within 24 h whereas with DMP under the same
conditions only ca. 50% H–D exchange was observed over a
48 h period. It is suggested that this reduced rate of exchange
has as origin the electron-donating properties of the C-1 N
substituent. The ability of the lone pair of electrons on the
amine-nitrogen in DMP to donate electron density by aligning
with the π-orbital system of the ring (π–π overlap) (21) will
lead to an increase in the electron density of the ring carbons.
In trifluralin and benefin, however, the increased steric inter-
actions associated with the larger N-alkyl groups prevent the
alignment of the amine-N lone electron pair for effective
π–π overlap and, therefore, aryl proton abstraction leading to
H–D exchange is facilitated.
Geometry optimization calculations (semi-empirical;
AM1) support this hypothesis. The optimized configuration
for 1 and 1a with the N-propyl groups aligned perpendicular
(DMP) under identical conditions. Preliminary observations
indicate that 100% H–D exchange of trifluralin aryl protons
©
1998 NRC Canada

Annandale et al.
881
Scheme 3.
to the plane of the ring are displayed in Figs. 4 and 6, respec-
tively. The same geometry optimization calculation performed
for DMP (Fig. 5) yielded an optimized configuration with the
N-methyl groups aligned with the plane of the ring, allowing
for maximum interaction of the amine-N lone pair of electrons
with the ring π-electron system. A comparison of partial
atomic charges (semi-empirical; AM1; Mulliken population
analysis) shows that the amine nitrogen of DMP is more posi-
tively charged than the amine nitrogen of 1. This supports our
arguement that there is significant overlap between the C-1
nitrogen lone pair of DMP with the ring π-system that is hin-
dered in the sterically constrained molecules of trifluralin (1)
and benefin (2).
–
However, another pathway is available for expulsion of R N
2
in the case of 1b and 2b, namely, a prior ionization of the
deuteroxide adduct followed by expulsion of R N , as shown
in Scheme 3.
–
2
Proton abstraction versus σ-complex formation in the
reaction of N-phenyl-2,6-dinitro-4-(trifluoro-
–
methyl)aniline (3) with OD
Since N-phenyl-2,6-dinitro-4-(trifluoromethyl)aniline (3) has
only one substituent on the C-1 nitrogen, the interaction with
base can, in principle, lead to N-deprotonation as well as σ-
complex formation. Such behaviour has been reported for the
reaction of N-methylpicramide (NMP) with MeO in
DMSO–MeOH solutions, which was reported to undergo in-
itial N-deprotonation upon the addition of 1 molar equivalent
of MeO , with further reaction to the C-3 adduct upon addition
of excess base (38). In fact this behaviour has also been found
in the present study.
Scheme 4 shows the reaction pathway for the reaction of 3
with OD . It was found that, initially, reaction results in N-de-
protonation to form the anion, 3a. This is followed by the
further reaction of 3a with OD to form the C-3 adduct, 3b,
but only when a large excess of base is present. Of notable
significance is the fact that with the NMP–MeO system, σ-
complexation was reported for solutions containing 1.1 molar
equivalents of MeO to NMP (38). However, in the present
system σ-complexation required ca. 10 molar equivalents of
OD . The origin of this contrasting behaviour is not clear at
–
2
−
Regioselectivity in reaction of trifluralin (1) with SO
3
2
−
As shown in Scheme 2, reaction of SO3 with 1 initially re-
sults in formation of the C-3 adduct. By analogy with the
–
2−
SO3 adducts of picramides (5, 6, 9, 35–37), bonding to sulfite
nucleophile is presumed to occur through the S-center. The
C-1 adduct is not observed in the present system, presumably
due to steric hindrance (F-strain) arising from the bulky N,N-
di-n-propylamine substituent at the C-1 position. Nevertheless,
further attack on 1d by SO3 occurs at the C-1 position, form-
ing the cis and trans isomers of diadduct 1e. Formation of the
diadduct, 1e, is thought to be stabilized by n → σ* stabilization
–
–
2
−
–
from the NR lone electron pairs and the forming C-
–
2
2−
1
—SO3 bond. Due to the alignment of the NR group in 1,
2
this stereoelectronic interaction is not present to help stabilize
–
2
3
−
the formation of the C-1 monoadduct with SO . Interest-
present, although a possible explanation is the decreased elec-
tron deficiency of 3 relative to the NMP system.
2
−
ingly, a diadduct of 1 with the two -SO3 groups at the C-3 and
C-5 positions was not observed, even though such diadducts
are seen with related picramides (35–37). This is probably due
2−
As in the reaction of 1 with SO3 , there was no observation
of a C-1 monoadduct, in accord with a K3T3 sequence. The
final thermodynamic product of this reaction is the hydrolysis
product of 3a, 3,5-dinitro-4-(N-phenylamino)benzoic acid, 8.
to the presence of the C-4 -CF group in 1, whereas in pi-
3
cramides with a -NO substituent at C-4 negative charge can
2
be delocalized onto the -NO group. Since no C-1 monoadduct
2
is observed in this system, we conclude that this reaction fol-
lows a K3T3 sequence.
Experimental
The fact that there is no overall nucleophilic displacement
of the dialkylamino group by sulfite, in contrast to the hydrox-
ide reaction, is worthy of comment. One factor is the ease of
formation and the stability of adduct 1d, attack at the 3-posi-
tion of 1 being favored over the 1-position for steric reasons
as discussed above. Moreover, there is clearly no expulsion of
the dialkylamino group from the cis and trans diadducts 1e,
N,N-Di-n-propyl-2,6-dinitro-4-(trifluoromethyl)aniline (tri-
fluralin; 1) was available from Chem Service, Inc. with a re-
ported purity of 99.0% and was used without further
purification. N-Ethyl-N-n-butyl-2,6-dinitro-4-(trifluoromethyl)-
aniline (benefin; 2) was prepared via a variation of method II
reported by Marshall et al. (39) by adding ca. 4 molar equiva-
lents of N-ethyl-n-butylamine to a solution of 2,6-dinitro-4-
(trifluoromethyl)chlorobenzene in diethyl ether and acidifying
the resulting solution with HCl (ca. 1 M). After removal of the
colourless aqueous layer, the clear orange-yellow diethyl ether
solution was washed with H O and dried over MgSO fol-
2−
–
i.e., in Scheme 2, 1e may eject SO3 but not R N . The much
2
greater nucleofugality of sulfite relative to the dialkylamino
group is readily evident on the basis of pK values but, as
a
shown in Scheme 1, expulsion of the dialkylamino group does
2
4
occur from 1b and 2b despite an unfavorable pK situation
(
lowed by evaporation of the ether under reduced pressure. The
resulting solid was further purified using flash chromatography
a
though clearly not as unfavorable as in the sulfite case).
©
1998 NRC Canada

8
82
Can. J. Chem. Vol. 76, 1998
Scheme 4.
(
(
CHCl eluent), yielding bright yellow crystals, mp 66–67°C
3
lit. (40) mp 67–68°C). N-Phenyl-2,6-dinitro-4-(tri-
final concentration of DMSO-d was 80% v/v. The reactions
1
6
were then monitored at various intervals by H NMR until
fluoromethyl)aniline (3) was prepared using the same proce-
dure as described for 2 with the addition of aniline in place of
N-ethyl-n-butylamine, yielding bright yellow crystals, mp
completion. Nuclear magnetic resonance (NMR) measure-
ments were obtained using a Bruker AM-400 spectrophotome-
1
1
ter ( H: 400.1 MHz) with DMSO-d H acting as reference ( H:
5
1
23°C (lit. (41) mp 119–121°C).
N,N-Di-n-propyl-2,6-dinitro-4-(trifluoromethyl)aniline-d2
2.50 ppm).
(
trifluralin-d ) was isolated by adding ca. 2 molar equivalents
2
Acknowledgement
of NaOD (D O) to a solution of 1 in DMSO-d . The solution
2
was let stand for ca. 24 h at which time ca. 3 molar equivalents
6
Financial support of this research by the Natural Sciences and
Engineering Research Council of Canada and valuable discus-
sions with Professor Julian Dust are gratefully acknowledged.
Some preliminary experiments by Louisa Wong are also ac-
knowledged with thanks.
of trifluoroacetic acid were added to quench the reaction. The
reaction mixture was then added to H O and orange crystals
2
were collected via vacuum filtration and dried under reduced
1
pressure (< 0.1 Torr; 1 Torr = 133.3 Pa); mp = 48–49°C. H
NMR (DMSO-d ) δ: 2.93 (4H, t, J
tq, JH–H = 7.3, Hb,b’), 0.80 (6H, t, JH–H = 7.3, Ha,a’). C NMR
= 7.3, Hc,c’), 1.50 (4H,
6
H–H
13
References
(
DMSO-d ) δ: 145.5 (1C, C ), 140.3 (2C, C ), 126.9 (2C, t,
6 1 2,6
1
. D. Hartley and H. Kidd (Editors). The agrochemicals handbook.
Royal Society of Chemistry, Nottingham, England. 1983.
. F.M. Ashton and A.I. Crafts. Mode of action of herbicides.
Wiley–Interscience, New York. 1973.
JC–D = 23, C ), 122.4 (1C, q, J
JC–F = 36, C ), 53.2 (2C, C8,11), 20.4 (2C, C9,12);,10.8 (6C,
C10,13). EI-MS: M = 337 m/z (calcd. for C H D F N O :
3
= 270, C ), 121.0 (1C, q,
3,5
C–F
7
4
2
13
14
2
3
3
4
37.0 g/mol).
3
4
. S.J. Parke and Q.F. Soper. Weed Sci. 25, 79 (1977).
. R.C. Hall and C.S. Giam. J. Agric. Food. Chem. 22, 461 (1974).
Sodium deuteroxide solutions were prepared by adding
freshly cut sodium metal to D O (99.9% D) under a stream of
2
N . Potassium deuteroxide solutions were prepared from a
5. M.J. Strauss. Chem. Rev. 70, 667 (1970).
2
6. E. Buncel, M.R. Crampton, M.J. Strauss, and F. Terrier. Electron
deficient aromatic– and heteroaromatic–base interactions: The
chemistry of anionic σ-complexes. Elsevier Science Publishing
Company, Inc., New York. 1984.
weighed amount of potassium hydroxide (AnalaR by BDH,
Inc.), which was added to D O under N . The alkaline solu-
tions were standardized against 0.108 M potassium hydrogen
phthalate (KHP). Sodium sulfite solutions were prepared by
dissolving a calculated amount of Na SO (Aldrich Chemical
2
2
7
8
9
. F. Terrier. Nucleophilic aromatic displacement: the influence of
the nitro group. Organic Nitro Chemistry Series, VCH Publishers
Inc., New York. 1991.
. E. Buncel. The chemistry of functional groups. Supplement F:
The chemistry of amino, nitro and nitroso compounds. Wiley,
London. 1982.
2
3
Company, Inc.) in D O under N .
2
2
Stock solutions of substrate were prepared in DMSO-d6
while stock solutions of nucleophile were prepared in D O.
2
Solutions used in the reaction studies were prepared in clean
dry NMR tubes by adding appropriate volumes of the stock
solutions to a known volume of DMSO-d –D O such that the
. F. Terrier. Chem. Rev. 82, 77 (1982).
10. E. Buncel, J.M. Dust, and R.A. Manderville. J. Am. Chem. Soc.
118, 2072 (1996).
6
2
©
1998 NRC Canada

Annandale et al.
883
1
1
1
1
1
1
1
1
1
2
1. E. Buncel, R.A. Manderville and J.M. Dust. J. Chem. Soc. Perkin
Trans. 2, 1019 (1997).
2. (a) K.L. Servis. J. Am. Chem. Soc. 87, 5495 (1965); (b) J. Am.
Chem. Soc. 89, 1508 (1967).
3. J.K.M. Sanders and B.K. Hunter. Modern NMR spectroscopy.
Oxford University Press, London. 1987. pp. 56–62.
4. E. Buncel, J.M. Dust, and F. Terrier. Chem. Rev. 95, 2261
24. M.R. Crampton and V. Gold. J. Chem Soc. B: 893 (1966).
25. A. Cooney and M.R. Crampton. J. Chem. Soc. Perkin Trans. 2,
1973 (1984).
26. J.M. Dust and M.D. Secord. J. Phys. Org. Chem. 8, 810 (1995).
27. R. Foster and R.K. Mackie. Tetrahedron, 19, 691 (1963).
28. E. Buncel and W.A. Zabel. J. Am. Chem. Soc. 89, 3082 (1967).
29. E. Buncel, J.A. Elvidge, J.R. Jones, and K.T. Walkin. J. Chem.
Res. 272 (1980).
30. E. Buncel and E.A. Symons. Can. J. Chem. 44, 771 (1966).
31. E. Buncel and E.A. Symons. J. Chem. Soc. Chem. Commun. 771
(1967).
32. E. Buncel and E.A. Symons. J. Org. Chem. 38, 1201 (1978).
33. R. Bacaloglu, C.A. Bunton, G. Cerichelli, and F. Ortega. J. Am.
Chem. Soc. 110, 3495 (1988).
34. R. Bacaloglu, A. Blasko, C. Bunton, E. Dorwin, F. Ortega, and
C. Zucco. J. Am. Chem. Soc. 113, 238 (1991).
35. N. Marendic and A.R. Norris. Can. J. Chem. 51, 3927 (1973).
36. M.R. Crampton. J. Chem. Soc. B: 1208 (1968).
37. E. Buncel, A.R. Norris, K.E. Russell, and P.J. Sheridan. Can. J.
Chem. 52, 25 (1974).
38. (a) E. Buncel, M. Hamaguchi, and A.R. Norris. Can. J. Chem.
58, 1615 (1980); (b) Can. J. Chem. 58, 1609 (1980).
39. F.J. Marshall, R.E. McMahon, and R.G. Jones. J. Agric. Food.
Chem. 14, 498 (1966).
(
1995).
5. C.F. Bernasconi, J. Fassberg, R.B. Killion, D.F.Schuck, and Z.
Rappaport. J. Am. Chem. Soc. 113, 4937 (1991).
6. R.A. Manderville. Ph.D Thesis. Queen’s University, Kingston,
Ont. 1992. p. 46.
7. R.A. Manderville, J.M. Dust, and E. Buncel. J. Phys. Org. Chem.
9, 515 (1996).
8. E. Buncel, J.M. Dust, A. Jo czyk, R.A. Manderville, and I. Ony-
ido. J. Am. Chem. Soc. 114, 5610 (1992).
9. E. Buncel, R.M. Tarkka, and J.M. Dust. Can. J. Chem. 72, 1709
(
1994).
0. (a) A.J. Kirby. The anomeric effect and related stereoelectronic
effects at oxygen. Springer Verlag, Berlin. 1983; (b) Stereoelec-
tronic effects. Oxford University Press Inc., New York. 1996.
1. W. Szarek and D. Horton. Anomeric effect. Origin and conse-
quences. ACS Symp. Ser. 87. American Chemical Society,
Washington, D.C. 1979.
2
2
2. R.A. Manderville and E. Buncel. J. Am. Chem. Soc. 115, 8985
(1993).
3. M.R. Crampton and V. Gold. J. Chem Soc. 4293 (1964).
40. R.C. Hall and C-S. Giam. J. Agric. Food. Chem. 20, 546, (1972).
41. B.F. Malichenko, E.M. Levchenko, and L.M. Yagupol’skii. Ukr.
Khim. Zh. (Ukr. Ed.) 33, 717, (1967).
2
©
1998 NRC Canada