Regioselectivity of Meisenheimer complexation in reaction of oxygen-centred nucleophiles with picryl aryl ethers: Polar vs. SET mechanisms

Article abstract of DOI:10.1139/v03-014

Picryl alkyl ethers react with hydroxide and methoxide ions to give regioisomeric Meisenheimer (anionic σ-) adducts; the C-3 adduct is kinetically favoured and the C-1 adduct is thermodynamically favoured (K3T1 behaviour). In the current 400 MHz NMR spectroscopic study of the reactions of two picryl aryl ethers, picryl phenyl ether (PicOPh, 1) and picryl mesityl ether (PicOMes, 2), the charge localized nucleophiles OH^- and MeO^- displayed the same K3 regioselectivity as found with picryl alkyl ethers; attachment at C-1 leads to S_NAr displacement of the aryloxide. In contrast, phenoxide (PhO^-) and the sterically demanding 2,4,6-trimethylphenoxide (mesitoxide, MesO^-) react with 1 and 2 to form the C-1 O-adduct as the product of kinetic control (i.e., K1 behaviour). These reactions were studied at low temperature (-40°C in acetonitrile-d₃:dimethoxyethane-d₁₀ 1:1) and as a function of increasing temperature (-40°C to ambient). On the thermodynamic side, the C-1 PhO^- O-adduct of 1 is also the more stable of the possible phenoxide O-adducts; it shows T1 regioselectivity within the manifold of O-adducts (K1T1), but the C-3 C-adduct (via para-attack of PhO^-) is the ultimate thermodynamic product. The C-1 O-adducts formed by MesO ^- with 1 or 2 give way with time (or temperature increase) in favour of their C-3 regioisomers or a C-1,3-O-diadduct. Mesitoxide, therefore, displays K1T3 regioselectivity. Stereoelectronic stabilization is discussed as a factor influencing T1 regioselectivity in O-adduct formation. Frontier molecular orbital (FMO) interactions between the HOMO of the nucleophile and the LUMO of the picryl ether may play a role in the K1 preference of aryloxides. An alternative argument is presented based on a single electron (radical) transfer (SET) pathway for the aryloxide nucleophiles rather than the polar (S_NAr) pathway for hydroxide and methoxide. The SET pathway also predicts a kinetic preference for C-1, as the C-1 position is of higher spin density than C-3 in the radical anion of the picryl ether and thus should be the preferred site for coupling by the aryloxide radical.

Full text of DOI:10.1139/v03-014

443
Regioselectivity of Meisenheimer complexation in
reaction of oxygen-centred nucleophiles with
picryl aryl ethers: Polar vs. SET mechanisms
Erwin Buncel, Julian M. Dust, Richard A. Manderville, and Richard M. Tarkka
Abstract: Picryl alkyl ethers react with hydroxide and methoxide ions to give regioisomeric Meisenheimer (anionic σ-)
adducts; the C-3 adduct is kinetically favoured and the C-1 adduct is thermodynamically favoured (K3T1 behaviour). In
the current 400 MHz NMR spectroscopic study of the reactions of two picryl aryl ethers, picryl phenyl ether (PicOPh,
1) and picryl mesityl ether (PicOMes, 2), the charge localized nucleophiles OH– and MeO– displayed the same K3
regioselectivity as found with picryl alkyl ethers; attachment at C-1 leads to SNAr displacement of the aryloxide. In
contrast, phenoxide (PhO–) and the sterically demanding 2,4,6-trimethylphenoxide (mesitoxide, MesO–) react with 1
and 2 to form the C-1 O-adduct as the product of kinetic control (i.e., K1 behaviour). These reactions were studied at
low temperature (–40°C in acetonitrile-d3:dimethoxyethane-d10 1:1) and as a function of increasing temperature (–40°C
to ambient). On the thermodynamic side, the C-1 PhO– O-adduct of 1 is also the more stable of the possible phenoxide
O-adducts; it shows T1 regioselectivity within the manifold of O-adducts (K1T1), but the C-3 C-adduct (via para-at-
tack of PhO–) is the ultimate thermodynamic product. The C-1 O-adducts formed by MesO– with 1 or 2 give way with
time (or temperature increase) in favour of their C-3 regioisomers or a C-1,3-O-diadduct. Mesitoxide, therefore, dis-
plays K1T3 regioselectivity. Stereoelectronic stabilization is discussed as a factor influencing T1 regioselectivity in O-
adduct formation. Frontier molecular orbital (FMO) interactions between the HOMO of the nucleophile and the LUMO
of the picryl ether may play a role in the K1 preference of aryloxides. An alternative argument is presented based on a
single electron (radical) transfer (SET) pathway for the aryloxide nucleophiles rather than the polar (SNAr) pathway for
hydroxide and methoxide. The SET pathway also predicts a kinetic preference for C-1, as the C-1 position is of higher
spin density than C-3 in the radical anion of the picryl ether and thus should be the preferred site for coupling by the
aryloxide radical.
Key words: anionic Meisenheimer adducts, regioselectivity, kinetic–thermodynamic control, FMO, stereoelectronic
stabilization, single electron transfer (SET).
Résumé : Les oxydes de picryle et d’alkyle ré4a5g6issent avec les ions hydroxydes et méthylates avec formation d’adduits
(anioniques σ-) de Meisenheimer régioisomères; l’adduit en C-3 est favorisé d’un point de vue cinétique alors que
l’adduit en C-1 est favorisé d’un point de vue thermodynamique (comportement K3T1). Dans l’étude courante, par
spectroscopie RMN à 400 MHz, des réactions de deux oxydes de picryles et d’aryles, l’oxyde de picryle et de phényle
(PicOPh, 1) et l’oxyde de picryle et de mésityle (PicOMes, 2), les nucléophiles à charge localisée, OH– et MeO–, pré-
sentent la même régiosélectivité K3 que celle observée avec les oxydes de picryle et d’alkyle; la fixation en C-1
conduit à un déplacement SNAr de l’aryloxyde. Par opposition, le phénolate (PhO–) et le 2,4,6-triméthylphénolate (mé-
sitylate, MesO–) qui imposent des demandes stériques beaucoup plus importantes réagissent tous les deux avec les
composés 1 et 2 pour former, comme produit de contrôle cinétique, un O-adduit en C-1 (comportement K1). On a étu-
dié ces réactions à basse température (–40 °C, dans un mélange acétonitrile-d3:diméthoxyéthane-d10 1:1) et en fonction
d’une augmentation de la température (–40 °C à la température ambiante). D’un point de vue thermodynamique, le O-
adduit du PhO– en C-1 du produit 1 est aussi le plus stable des O-adduits possibles pour le phénolate; il présente du
régiosélectivé T1 parmi les plusieurs O-adduits (K1T1); toutefois, le C-adduit C-3 (obtenu par une attaque en para du
Received 5 November 2002. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 17 April 2003.
Dedicated to Prof. Don Arnold for his contributions to chemistry in Canada.
E. Buncel.1 Department of Chemistry, Queen’s University, Kingston, ON K7L 3N6, Canada
J.M. Dust.2 Departments of Chemistry and Environmental Science, Sir Wilfred Grenfell College, Corner Brook, NL A2H 6P9,
Canada.
R.A. Manderville.3 Department of Chemistry, Wake Forest University, Winston-Salem, NC 27109-7486, U.S.A.
R.M. Tarkka. Department of Chemistry, University of Central Arkansas, Conway, AR 72035, U.S.A.
1Corresponding author (e-mail: buncele@chem.queensu.ca).
2Corresponding author (e-mail: jdust@swgc.mun.ca).
3Corresponding author (e-mail: manderra@wfu.edu).
Can. J. Chem. 81: 443–456 (2003)
doi: 10.1139/V03-014
© 2003 NRC Canada

444
Can. J. Chem. Vol. 81, 2003
PhO–) est le produit thermodynamique ultime. Les O-adduits C-1 qui se forment initialement entre le MeSO– et les
produits 1 et 2 laissent la place, avec le temps ou une augmentation de la température, à leurs régioisomères C-3 ou à
un C-1,3-O-diadduit. Le mésitylate présente donc une régiosélectivité K1T3. On discute de la stabilisation électronique
comme facteur influençant la régiosélectivité T1 dans la formation des O-adduits. Les interactions des orbitales molé-
culaires frontières (OMF) entre l’OM haute occupée et la basse vacante de l’oxyde de picryle pourrait jouer un rôle
dans la préférence K1 des aryloxydes. On présente un argument alternatif, basé sur une voie réactionnelle impliquant le
transfert d’un seul (radical) électron (SET) pour les nucléophiles aryloxydes plutôt qu’une voie polaire (SNAr) pour
l’hydroxyde et le méthanolate. La voie SET permet aussi de prédire une préférence cinétique pour C-1 parce que cette
position possède une densité de spin plus élevée que celle en C-3 dans l’anion radical de l’oxyde de picryle et qu’elle
devrait donc être le site favorisé pour le couplage avec le radical aryloxyde.
Mots clés : adduits anioniques de Meisenheimer, régiosélectivité, contrôles cinétique et thermodynamique, OMF, stabili-
sation stéréoélectronique, transfert d’un seul électron (SET).
[Traduit par la Rédaction] Buncel et al.
plicit in the synthetic results found in vicarious nucleophilic
substitution (VNS) reactions, as documented by Makosza
and co-workers (10). Finally, we have reported the results of
the reaction of TNA with the bulky aryloxide nucleophile
2,4,6-trimethylphenoxide (mesitoxide, MesO–) (6c). In this
Introduction
Reaction of a base (nucleophile) with an electron-deficient
aromatic may give rise to diverse products, namely π-com-
plexes (1), radical anions (2), aryl carbanions (3), and an-
ionic σ-adducts (termed Meisenheimer complexes) (1–4).
With 1-X-2,4,6-trinitrobenzenes, nucleophiles react to form
Meisenheimer complexes according to an accepted general
pattern of the initial formation of a C-3 adduct that gives
way over time to a thermodynamically favoured C-1 adduct
(5). This regioselectivity has been elaborated in a number of
1-X-2,4,6-trinitrobenzene-nucleophile reaction systems, no-
tably with alkoxides and hydroxide as nucleophiles (5), and
has been labelled K3T1 (6, 7) (cf. Fig. 1).
system, attack at C-1 to yield
a
C-1 TNA·OMes–
Meisenheimer complex is kinetically favoured at –40°C, but
as the NMR probe temperature was raised, the C-1 adduct
rapidly declined in concentration and was supplanted by the
C-3 TNA·OMes– adduct, in accord with K1T3 behaviour
(Fig. 1). This reactivity pattern is the inverse of the “nor-
mal” K3T1 isomerization pathway displayed by alkoxides
and hydroxide.
Although a wide range of kinetic and thermodynamic fac-
tors have been advanced (4, 5, 11, 12) to account for the
regioselectivity found in picryl ether – base Meisenheimer
complexation, our analysis has focused primarily on thermo-
dynamics and has highlighted the contribution made by
stereoelectronic stabilization of the relevant C-1 adducts
(6b–d, 7, 9a). In this approach, C-1 adducts, as acetal ana-
logues, can be stabilized by n → σ* donation from an oxy-
gen lone pair of one RO group into the anti-bonding orbital
of the C—OR bond (and vice versa); the interaction is maxi-
mized if the relevant lone pairs and C—OR fragment bonds
can be arranged antiperiplanar to one another. Where stereo-
electronic stabilization is negligible, the C-3 adduct may be-
come the thermodynamic product (7).
The arguments concerning kinetic factors have either re-
lied on assumptions made about the position of the transition
state for adduct formation (5d, 6b) or have consisted of as-
sessments of steric hindrance to attack at C-1 (F-strain) (11,
12d, 12e, 13, 14). In light of the K1T3 behaviour exhibited
in the reaction of TNA with mesitoxide, a bulky nucleophile,
the importance of F-strain in determining the regioselectivity
of Meisenheimer complexation warrants re-examination.
The present article extends the study of picryl ether –
nucleophile interactions to the picryl aryl ethers picryl phenyl
ether (PicOPh, 1) and picryl mesityl ether (PicOMes, 2). The
reactions of these ethers, 1 and 2, with the aryloxide nucleo-
philes, phenoxide ion (PhO–) and mesitoxide (MesO–), were
monitored in acetonitrile–dimethoxyethane (MeCN-d3:DME-
d10 1:1 v/v) as a function of temperature (–40° to ambient),
and the PicOPh–PhO– system was studied in dimethyl sulfox-
ide (DMSO-d6) at room temperature. Thus, the current study
In a series of articles (6, 7), we have delineated a full
range of regioselectivities in Meisenheimer complexation.
Thus, while 2,4,6-trinitroanisole (TNA) reacts with meth-
oxide according to the K3T1 pattern (cf. Fig. 1) (5a–d, 6b),
TNA reacts with the oxygen centre of a phenoxide ion
(PhO–) with K1T1 regioselectivity. In the latter case, a C-1
TNA·OPh– adduct was the first (and only) phenoxide O-
1
adduct detected by 400 MHz H NMR spectroscopy at –40°C
in acetonitrile–dimethoxyethane (MeCN-d3:DME-d10 1:1)
solvent (6b). The regioisomeric C-3 TNA·OPh– adduct was
not observed, in accord with a previous stopped-flow UV–
vis kinetic study in which the initially observed adduct was
identified as the C-1 species (8). Furthermore, the structur-
ally similar phenoxide O-adduct of 1,3,5-trinitrobenzene
1
(i.e., TNB·OPh–) can be observed by H NMR spectroscopy
under the same conditions used in the TNA·PhO– study (9a).
On this basis, the regioselectivity exhibited by the TNA–
PhO– system was classified as K1T1, i.e., a system in which
the C-1 adduct is favoured by both kinetics and thermody-
namics (Fig. 1). However, the ultimate phenoxide product
was the C-3 TNA·PhO(H)– para C-bonded adduct, consis-
tent with the ambident (O- and C-) nucleophilic nature of
PhO– (6b) and in agreement with results gleaned from re-
lated systems (9). The behaviour of phenoxide as a C-
nucleophile corresponds to K3T3 regioselectivity wherein
the C-3 C-adduct is the product of both kinetic and thermo-
dynamic control. Our AM1 calculations on the regioisomeric
–
adducts formed by TNA with OH– and CH3 as prototypical
O- and C-nucleophiles, respectively, confirm the thermody-
namic preference for C-3 attack by carbon nucleophiles (6d).
This K3T3 behaviour (Fig. 1) of C-nucleophiles is also im-
© 2003 NRC Canada

Buncel et al.
445
Fig. 1. 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 ki-
netic control, but the C-1 adduct is the thermodynamic product. In the K1T1 profile the C-1 adduct is favoured by both kinetics and
thermodynamics. The profile designated K3T3 represents the situation where the C-3 adduct is doubly preferred; that is by kinetics and
by thermodynamics. The K1T3 profile describes the inverse behaviour from that indicated by K3T1; now the C-1 adduct is favoured
kinetically but the C-3 adduct is the most stable product.
further probes the effect of steric hindrance at C-1 on the kinet-
ics of these systems, particularly in the case of the highly hin-
dered PicOMes reacting with the bulky MesO– anion.
in the MeCN-d3:DME-d10 medium than in DMSO-d6. 13C
NMR peak positions for some relevant σ-adducts measured
in MeCN–DME and substrates measured in DMSO-d6 are
given in Table 2.
The results are compared with those obtained in the re-
lated TNA–Nu– systems and are discussed with regard to
possible frontier molecular orbital (FMO) interactions be-
tween the highest occupied molecular orbital (HOMO) of
the aryloxide nucleophiles and the lowest unoccupied molec-
ular orbital (LUMO) of the polynitroaromatic substrate. To
model a 1-X-2,4,6-trinitrobenzene, we have carried out a
semi-empirical (AM1) (15) molecular orbital calculation on
TNA, and these results are included in this article. Finally,
more recent suggestions that related reactions proceed
through transition states having varying degrees of radical
character (16) or via single electron transfer (SET) to give a
radical–radical anion pair (2) are considered.
As in previous studies of the reaction of aryloxides with
picryl systems (6b, 6c, 9), C-3 hydroxide adducts of 1 (i.e.,
6) and 2 (i.e., 15) were observed also. The assignment of the
signals of these adducts (6 and 15) was confirmed by control
experiments that involved the electron-deficient substrates
and tetramethylammonium hydroxide (Me4NOH) in DMSO.
In general, these C-3 hydroxide adducts gave way over time
to picrate anion (PicO–, 8) following a pattern seen previ-
ously (6b, 6c, 9). Methoxide reacted with 1 and 2 to yield
the respective C-3 adducts 9 and 17, as the products of ki-
netic control; the corresponding C-1 adducts 10 and 18 were
not observed. Relevant spectroscopic data for the observed
adducts are given in Table 1.
Results
Reaction of 1 with excess PhOK in MeCN–DME
Reactions of the aryloxides PhO– and MesO– with the
electrophiles PicOPh, 1, and PicOMes 2, were monitored by
400 MHz 1H NMR in acetonitrile–dimethoxyethane (MeCN-
d3-DME-d10 1:1 v/v), a solvent system that has proven use-
ful in NMR studies down to temperatures of –50°C (6b, 6c,
9a). Reactions of 1 and 2 with OH– and MeO– were con-
ducted in DMSO-d6 at ambient temperature. Spectroscopic
characteristics of the species (including coupling constants,
J, in Hz) shown in Schemes 1 and 2 are listed in Table 1. In
general, resonances are located at positions farther downfield
To an NMR tube that contained a solution of PicOPh (1)
in MeCN-d3:DME-d10 (1:1, v/v), cooled to –50°C, was in-
jected 1.5 equiv of a similarly cooled MeCN-d3:DME-d10
solution of phenoxide ion (PhOK; final concentrations of
1:PhOK were 0.06:0.09 M). The first H NMR spectrum re-
corded at –40°C contained a singlet at δ 8.63 that is consis-
1
tent with the resonance for two equivalent ring protons (H3,5
)
of a C-1 O-adduct. Other signals in the spectrum are simi-
larly attributable to the C-1 PicOPh·OPh– O-adduct, 3
(Scheme 1). The initial low-temperature spectrum was re-
© 2003 NRC Canada

446
Can. J. Chem. Vol. 81, 2003
Scheme 1.
markably free from signals that would arise from formation
of other adducts at this temperature. Therefore, a full assign-
ment could be made for 3: 8.63 (2H, s, H3,5), 7.15 (4H, m,
Hm), 6.95 (2H, m, Hp), and 6.69 (2H, m, Ho).4 It is pertinent
to note that the signals assigned to the ring protons of the at-
tached phenoxyl group are equivalent, indicating that the
adduct is symmetrical. The 13C NMR spectrum of 3 was also
recorded (Table 2).
As the temperature was gradually raised, resonances of 3
began to broaden, and after ca. 1 h at 10°C a new set of dou-
blets could be seen at 8.57 (1H, J = 1.9) and 6.42 (1H, J =
1.9). These peaks are assignable to H5 and H3, respectively,
of the C-3 PicOPh·OH– adduct, 6 (Scheme 1), that arises
from equilibration of PhO– and adventitious water present in
the solvent (cf. ref. 6b, 6c, 17). The OH resonance of 6 was
not observed and its state of ionization is, therefore, uncer-
tain.
As the temperature was further raised to ambient, a singlet
was noted at 8.65. Comparison with related systems (6)
shows that this singlet represents the two equivalent ring
protons of picrate anion, i.e., PicO–, 8. Eventually (>5 h)
peaks appear that are ascribable to the C-3 para-bonded
adduct, 5 (Scheme 1); resonances belonging to 5 are given in
Table 1. The OH of the attached phenoxyl group was not ob-
served, so the state of ionization of this OH is uncertain.
Moreover, the para proton (Hp) of the C-1 phenoxyl moiety
of the C-3 PicOPh·PhO(H)– adduct, 5, was apparently ob-
scured by resonances of free PhOH (i.e., Hm′, Ho′, Hp′).
In summary, as an O-nucleophile, PhO– reacts at C-1 of
PicOPh, 1, to yield the C-1 O-adduct, 3, as the first phen-
oxide adduct. More significantly, at no time were peaks ob-
served that could be attributed to a C-3 phenoxide O-adduct
(i.e., 4, Scheme 1). We have previously shown (9a) that the
chemical shift of the diagnostic proton bonded to the sp3-hy-
bridized ring carbon in the phenoxide O-adduct of 1,3,5-
trinitrobenzene, TNB·OPh–, is located 0.5–0.8 ppm down-
field from the signal for the comparable proton in analogous
TNB·OR– adducts. On this basis, the signal for the similar
sp3-bound proton in the putative C-3 PicOPh·OPh– adduct,
4, should appear in a region of the spectrum well separated
from the signals of the C-3 hydroxide adduct, 6, and so
should be readily identifiable. In fact, no such signal was
seen before the appearance of the peaks assigned to the C-1
O-adduct, 3, nor did it appear later in response to increasing
temperature. As the temperature was raised, 6 and PicO–, 8,
were observed in the spectrum. Slowly, 6 gave way to 8 and
free phenol, presumably through the intermediacy of a tran-
sient C-1 PicOPh·OH– adduct, 7, that is not observed. Such
adducts have been postulated in analogous reaction systems
(6b–d). In accord with the ambident nature of phenoxide
ion, the eventual product of phenoxide attack is the C-3 para
C-bonded adduct, 5.
Reaction of 1 with equimolar MesOK in MeCN–DME
Upon addition of 1 equiv of potassium 2,4,6-trimethyl-
phenoxide ion (potassium mesitoxide, MesOK) in MeCN-
d3:DME-d10 to the NMR tube that contained a cooled (–50°C)
solution of 1 (final concentration 0.06 M), the sample turned
a deep orange colour. Interestingly, the initial 1H NMR spec-
4 Signals assigned to the para, meta, and ortho (Hp, Hm, and Ho) protons typically appear as somewhat broadened triplets, and triplets and
doublets, respectively. The broadening found is indicative of further unresolved coupling and so throughout this article the signals ascribed
to attached phenoxyl groups and to phenoxide ion or phenol will be listed as multiplets.
© 2003 NRC Canada

Buncel et al.
447
Scheme 2.
trum, recorded at –40°C, showed the presence of two σ-com-
plexes, the C-1 PicOPh·OMes– O-adduct, 11 (Scheme 2),
and the C-1 PicOPh·OPh– adduct, 3 (Scheme 1). At this
stage, peaks of 11 were found at δ 8.69 (2H, s, H3,5), 7.08
(2H, m, Hm), 6.80 (1H, m, Hp), 6.68 (2H, s, H3′,5′ mesitoxyl),
6.60 (2H, m, Ho), 2.16 (3H, s, p-Me, mesitoxyl), and 1.97
(6H, s, o-Me) and were clearly distinguishable from the sig-
nals assigned to 3 (Table 1). The resonances of 11 were pre-
dominant in this initial spectrum: 11:3 = 2:1.
Subsequent monitoring of the reaction as a function of in-
creasing temperature (from –10 to 0°C in ca. 30 min)
showed that the peaks due to 11 vanished while signals at-
tributed to 3 and the C-3 aryloxide O-adducts remained in
the spectrum. Furthermore, peaks assigned to the
PicOPh·OH– adduct, 6, were joined by those of the C-3
PicOMes·OH– adduct, 15 (Table 1, Scheme 2). At ambient
temperature the signals attributed to 3 and the C-3 aryloxide
O-adducts were no longer present; those of the OH– adducts,
6 and 11, dominated the spectrum. Peaks for 2, MesOH, and
8 were also found and a spectrum acquired after 2 days at
ambient still contained peaks for 8 and mesitol as well as
PhOH.
Thus, the results suggest that in the reaction of MesO–
with 1, initial attack gives the C-1 PicOPh·OMes– O-adduct,
11. Breakdown of 11 via an SNAr pathway (18–20) leads to
formation of PicOMes and PhO– and, therefore, to a system
that contains two substrates (1 and 2) as well as two nucleo-
philes (PhO– and MesO–). The liberated PhO– ion reacts at
C-1 of 1 to give 3. As the reaction proceeds, peaks due to 11
decrease in intensity in tandem with growth of the reso-
nances of 2, MesOH and 3. At no time are peaks of 1 and
free PhOH observed at this early stage of the reaction.
The next observable σ-adducts appeared at –10°C and
were identifiable as C-3 aryloxide O-adducts (6, 9a). Al-
though the exact assignment of these adducts is tentative, the
spectroscopic evidence supports the assignment of the sig-
nals to C-3-type aryloxide O-adducts. Moreoever, the C-3
adducts arise from attack of MesO– on 2 and are not C-3 ad-
ducts formed by attack of either PhO– or MesO– on 1. First,
the previous experiment (vide supra) confirmed PhO– attack
at C-1 of 1 to give 3 as the only process that involves 1 and
phenoxide as an O-nucleophile. Thus, none of the C-3 arylo-
As the temperature of the system was slowly raised to
−20°C, broad resonances appeared, a set at ca. 8.60 and 6.80
and a signal at ca. 6.67; these resonances were ascribable to
PicOMes, 2, and mesitol (MesOH), respectively. Coincident
with the appearance of the new signals, the resonances as-
signed to 11 declined in favour of those of 3. With the in-
crease to –10°C and the passage of time (ca. 2 h), new
resonances appeared that were assigned to C-3 aryloxide O-
adducts. The sp3-bound protons (H3) of the O-adducts were
doublets at 6.90 (J = 2.2) and 6.87 (J = 2.2), whereas the
sp2-bound protons (H5) appeared as a single triplet that
arises from overlapping doublets centred at 8.43. The down-
field shift of the H3 protons (at 6.90 and 6.87) of these spe-
cies relative to the shift of the corresponding proton
established for the C-3 PicOPh·OH– adduct, 6, was consis-
tent with attachment of the O-centre of an aryloxide (6, 9a).
Although additional signals that would be expected for these
C-3 aryloxide O-adducts were obscured by the peaks of 2, 3,
11, and MesOH, the signals at 6.90, 6.87, and 8.43 (as well
as those for 3, 6, and 11) do not survive acidification
(trifluoroacetic acid, TFA; 5 µL); it is a characteristic of O-
adducts that they are acid labile (4b, 6, 9). Tentative assign-
ment of the structures of the C-3 O-adducts corresponding to
the 6.90, 6.87, and 8.43 set of signals will be made below.
© 2003 NRC Canada

448
Can. J. Chem. Vol. 81, 2003
xide O-adducts can arise from interaction of 1 and PhO–. Sec-
ondly, decomposition of the initially formed C-1 PicO-
Ph·OMes– adduct, 11, releases PicOMes, 2, which could
undergo further attack by mesitoxide. Note that 2 and MesOH
are both observed in spectra acquired at this temperature. Fur-
ther justification for the assignments was obtained from study
of the PicOMes–MesO– system (vide infra). By the reasoning
outlined, the peaks of the C-3 aryloxide O-adducts in the
present system are ascribed to the C-3 PicOMes·OMes–
2–
adduct, 12, and the C-1,3 PicOMes·(OMes)2 diadduct, 13.
The H3 resonance at 6.90 (d, J = 2.2) is attributed to 12; that
of 13 appears at 6.87 (d, J = 2.2).
The final σ-adducts found in the system are the C-3 hy-
droxide complexes 6 and 15 that result from equilibration of
PhO– and MesO– with residual water in the medium that, in
turn, generates OH–. Ultimately, these too give way to
picrate ion, 8, PhOH, and MesOH.
Reaction of 2 with equimolar PhOK in MeCN–DME
The interaction of PicOMes, 2, with PhOK in MeCN-
d3:DME-d10 (1:1 v/v) was found to give results similar to
those for the PicOPh–MesO– reaction system described
above. Consequently, after the addition of equimolar PhOK
in MeCN-d3:DME-d10 to a solution of 2 in the same medium
(final concentrations: 0.06 M), cooled to –50°C, the first
spectrum (recorded at –40°C) includes resonances of the C-1
PicOMes·OPh– O-adduct, 11 (Scheme 2) as well as the C-1
PicOPh·OPh– O-adduct, 3 (Scheme 1). As the temperature
is raised, peaks reappear for 2 and appear for MesOH, but
at –10°C, signals are observed that correspond to the C-3
mesitoxide adducts, 12 and 13. Observation of the peaks for
these species validates their observation and assignment in
the PicOPh–MesO– reaction system.
Peaks due to 11 disappear from the spectrum at 0°C and
resonances attributable to the C-3 hydroxide adducts, 6 and
15, appear. At ambient temperature the signals for these hy-
droxide adducts, 6 and 15, and for PicO– were dominant,
whereas the peaks of 3, 12, and 13 diminish. Eventually
(>12 h), the spectrum obtained at room temperature consists
mainly of 8, PhOH, and MesOH, although small extraneous,
unidentified signals were also present.
Reaction of 2 with equimolar MesOK in MeCN–DME
To a solution of 2, cooled to –50°C, was added a similarly
cooled solution of MesOK in MeCN-d3:DME-d10 (final con-
1
centration: 0.06 M). The major peaks in the initial H NMR
spectrum measured at –40°C (acquired within 5 min of mix-
ing) were identified as belonging to unmodified 2 and
MesOK. However, two equivalent broad singlets were also
noted at δ 8.62 and 7.17. The 8.62 resonance could corre-
spond to the H3,5 protons of a C-1 MesO– adduct; compari-
son of the chemical shift with those for the H3,5 protons of
the related C-1 TNA·OMes– adduct, 18, (i.e., 8.62 (6c)) and
the C-1 PicOPh·OPh– adduct, 3 (i.e., 8.63) favours this as-
signment. However, in the present system the relative
integrals link the 8.62 peak to the broad signal at 7.17 (i.e.,
1:1 integral ratio). Moreover, peaks, albeit poorly resolved
ones, are observed in the 6.5–6.8 and 1.9–2.2 ppm regions
(other than those for 2 and MesOK) and are taken to repre-
sent the mesitoxyl protons of this initially formed species.
© 2003 NRC Canada

Buncel et al.
449
Table 2. 13C NMR spectral characteristicsa of the C-1 PicOPh phenoxide adduct, 3, and the C-1 TNA phenoxide adduct 10,b in
MeCN–glyme,c as well as the substrates, PicOMe (TNA), PicOPh (1), and PicOMes (2) in DMSO.d
Species
C-1
C-2,6
C-3,5
C-4
C-7
C-8,12
C-9,11
C-10
3c
104.4
104.5
151.5
145.2
145.5
130.8
129.5
144.3
144.0
141.0
130.5
131.1
125.0
125.5
124.4
119.9
119.7
141.7
143.3
140.9
155.1
155.4
65.2
156.4
148.0
120.8
120.5
130.4
130.4
124.0
124.0
10b,c
TNAd
1d
115.8
127.9
130.1
130.9
124.6
135.7
2d,e
aChemical shifts are given in ppm (δ) measured at 100 MHz.
bGenerated from reaction of TNA with 1 equiv PhO–.
cCD3CN-DME-d10 (1:1, v/v); obtained at –40°C.
dDMSO-d6 at ambient temperature.
eMesitoxyl o-Me δ 15.9; p-Me 20.2.
Combination of these observations with the kinetic
preference shown by aryloxides for attack at the C-1 posi-
tion of picryl alkyl ethers (6, 7) leads to assignment of the
resonances at 8.62 and 7.17 to the ring protons of the C-1
PicOMes·OMes– adduct, 14 (Scheme 2). In this case, the
H3 and H5 protons are non-equivalent as a result of steric
interactions between the C-1 mesitoxy substituents and
the flanking C-2,6 nitro groups. For example, if one of the
C-2,6 NO2 groups in 14 is twisted from the plane of the
cyclohexadienate ring to relieve the proposed steric strain,
then the negative charge of 14 could only be delocalized
to the C-4 and one of the C-2,6 NO2 groups. Thus, assum-
ing that the C-4 and C-6 nitro groups form part of the con-
jugated anionic system, the C-2 nitro group and the
moiety to which it is bonded would resemble an isolated
nitroalkene, and H3, in this example, would approximate a
Discussion
Reaction pathways and classification of regioselectivity
The interactions of the series of O-nucleophiles, phenoxide
(PhO–), 2,4,6-trimethylphenoxide (mesitoxide, MesO–), hy-
droxide (OH–), and methoxide (MeO–) ions with the picryl
aryl ethers PicOPh, 1, and PicOMes, 2, provide interesting
insights into the variable regioselectivity of Meisenheimer
complex formation. On the basis of comparison of the pKa
values of the parent phenols (23) in water and with the as-
sumption that nucleophilicity follows basicity (24), MesO–
as the more basic aryloxide of the two would be expected to
be a more reactive nucleophile than PhO– (i.e., pKa
(PhOH) = 9.95; pKa (MesOH) = 10.88, see also ref. 25).
However, PhO– may act as an ambident (C- and O-)
nucleophile, whereas MesO– is restricted to O-attack. The
ortho-Me groups in MesO– make it a sterically hindered
nucleophile. This steric bulk would introduce further F-
strain to attack at C-1, which would partly offset the higher
nucleophilicity of MesO– compared with PhO–. Therefore,
the two aryloxides could display similar kinetic activity
overall. Conversely, the steric factor may also render the re-
sultant C-1 MesO– adduct significantly less stable than its C-
1 PhO– oxygen-centred counterpart. Consequently, these
nucleophiles would show different thermodynamic prefer-
ences while displaying the same kinetic preference for C-1
attachment.
The ambident nature of PhO– also results in the formation
and observation of C-centred adducts as the final, stable
Meisenheimer complexes observed in the reaction systems
involving phenoxide as nucleophile. In each case, C-adducts
were identified as the C-3 regioisomers bonded via the para
site of phenoxide.
In all of the aryloxide systems examined, hydroxide ad-
ducts also formed during the period of study as a result of
reaction with adventitious H2O in the solvent systems. The
assignments of these adducts was confirmed in separate con-
trol experiments that included reaction of 1 and 2 with tetra-
methylammonium hydroxide in DMSO and with potassium
methoxide (in MeOH) in DMSO. As in other studies of
picryl ether – alkoxide and picryl ether – hydroxide reac-
tions (5), these localized O-nucleophiles reacted according
to K3 classification in which attack at C-3 is kinetically pre-
proton at the 2-position of
a 1-nitroalkene (14a,
Scheme 2). In this regard, it is noteworthy that the reso-
nance for the olefinic proton at the 2-position of 1-
nitrocyclohexene appears at 7.20 in DMSO-d6 (21).
Twisting of NO2 groups out of the aromatic plane has
been noted in numerous X-ray crystallographic studies of
neutral nitroaromatic compounds that possess an alkoxy
group adjacent to the nitro group (22).
As the temperature was gradually raised, the spectra ob-
tained showed a sequential decline in the signals that represent
14 (Scheme 2). The spectrum acquired at –30°C (recorded ca.
30 min after the initial spectrum acquired at –40°C) lacked
peaks of 14; signals of 2 were notably broad. Upon further
warming, peaks sharpened, and at –10°C the spectrum con-
tained signals assigned to two C-3 MesO– adducts as fol-
lows: 12 at 8.44 (1H, d, J = 2.2, H5) and 6.90 (1H, d, J =
2.2, H3) and 13 at 8.44 (1H, d, J = 2.2, H5 overlapped with
H5 of 12) and 6.87 (1H, d, J = 2.2, H3). Observation of sig-
nals of these Meisenheimer complexes in this system was in
accord with their previous identification in the PicOPh–
MesO– and PicOMes–PhO– reactions and confirmed their
assignment as adducts arising from attack of MesO– on
PicOMes in all three studies.
At ambient temperature the peaks of 12 and 13 are re-
placed by those due to the C-3 PicOMes·OH– adduct, 15
(Scheme 2). After monitoring the reaction for 12 h at room
temperature, the singlet for 8 at 8.65 is also seen.
© 2003 NRC Canada

450
Can. J. Chem. Vol. 81, 2003
Fig. 2. Qualitative reaction coordinate diagram for reaction of
PicOPh (1) with MeO–.
ferred. On the other hand, the C-1 O-adducts 7 and (or) 10
(Scheme 1) and 16 and (or) 18 (Scheme 2) are not detected
and represent metastable intermediates in the SNAr displace-
ment reaction to yield picrate (8), TNA, and phenol or
mesitol.
The situation in the picryl aryl ether – methoxide systems
is particularly informative. Although the C-1 adducts
PicOPh·OMe– (10) and PicOMes·OMe– (18) are observed in
the TNA–PhO– (6b) and TNA–MesO– (6c) systems, respec-
tively, these species are not observed in the present studies
where the corresponding picryl aryl ethers react with meth-
oxide to yield the C-3 adducts 9 and 17 as the only observ-
able Meisenheimer complexes. It is important to note that in
the TNA systems, 10 and 18 are the first and only adducts
detected at –40°C in 1:1 MeCN–DME. Both adducts decom-
pose to yield the C-3 hydroxide adduct of TNA via the ther-
modynamically unfavourable equilibrium between the
aryloxide and adventitious water in the MeCN–DME me-
dium (6b–6d, 9a). This observation suggests that the C-3
TNA·OH– adduct is more stable than 10 and 18 and by ex-
tension the C-3 TNA·OMe– adduct should also be more sta-
ble. These arguments suggest strongly that the inability to
detect 10 and 18 in reactions of 1 and 2 with MeO– stems
from the fact that they are not as stable as the C-3 adducts 9
and 17, respectively. The very slow rate of conversion of the
C-3 PicOPh·OMe– adduct 9 and C-3 PicOMes·OMe– adduct
17 into the SNAr products of TNA, PhO–, MesO–, and
picrate (8) is consistent with this hypothesis. These argu-
ments are summarized in the reaction coordinate diagram
depicted in Fig. 2 for the reaction of PicOPh (1) with MeO–.
Here the C-1 O-adduct 10 is shown to be less stable than the
C-3 adduct 9. However, the SNAr product from decomposi-
tion of 10 (TNA, PhO–, and eventually picrate (8)) are the fi-
nal products formed following irreversible processes. Thus,
even though the products from C-1 attack are thermodynami-
cally favoured, the reactions of 1 and 2 with methoxide (and
by extension hydroxide) can be designated as K3T3 on the
basis of the kinetic and thermodynamic preferences in the
initial reactions of 1 and 2 with the alkoxide (hydroxide)
nucleophiles.
Focusing on the aryloxide systems, PicOPh(1)–PhO– and
PicOMes(2)–MesO– are regarded as symmetrical systems, as
the nucleophile and leaving group in the SNAr process are
the same; conversely, PicOPh(1)–MesO– and PicOMes(2)–
PhO– are regarded as nonsymmetric systems. In the symmet-
rical PicOPh(1)–PhO– system, it is clear that reaction of
PhO– with 1 yields the C-1 PicOPh·OPh– adduct, 3, as the
first observable phenoxide O-adduct (at –40°C). No C-3
phenoxide O-adduct (i.e., 4, Scheme 1) is detected either
prior to observation of 3 or later in the reaction. In our previ-
ous study of the TNA–PhO– system we were able to demon-
strate that a similar observation at low temperature indicated
K1T1 regioselectivity (Fig. 1) in which formation of the cor-
responding C-1 TNA·OPh– Meisenheimer complex was fa-
voured by both kinetics and thermodynamics (6b). However,
the TNA–PhO– system had been previously examined by
fast kinetic techniques that supported the assignment of the
first formed O-adduct to the C-1 TNA·OPh– species (8). Fur-
ther, it had been established that the related TNB·OPh–
adduct could be identified and fully characterized by 1H
NMR spectroscopy under the same experimental conditions
(9a). Since the TNB·OPh– O-adduct is a reasonable model
for a hypothetical C-3 TNA·OPh– O-complex, it follows that
any C-3 phenoxide O-adduct would have been observed if it
had formed. It is, therefore, important in following this train
of logic to note that the TNB·OPh– adduct is also a structural
analogue of the C-3 PicOPh·OPh– Meisenheimer complex,
4, and that 4 would be expected to be observed if it formed
during the course of the reaction. Thus, in the symmetrical
PicOPh–PhO– system, PhO– (as an O-nucleophile) follows
K1T1 regioselectivity.
The next observable species, as a function of increasing
temperature, is 6, the C-3 PicOPh·OH– complex. The neces-
sary hydroxide is formed in low concentration via the ther-
modynamically unfavourable equilibrium between PhO– and
adventitious water in the MeCN–DME medium (6b–d, 9a).
The timing of the appearance of the hydroxide adduct in
these systems does not parallel the nucleophilicity of OH–
relative to PhO– because of the expected significant differ-
ence in concentration of the two nucleophiles in these reac-
tion systems. As the reaction proceeds, PicO– (picrate anion,
8) and PhOH appear, presumably as decomposition products
that arise from the transient C-1 PicOPh·OH– adduct, 7
(Scheme 1), whose existence has been postulated in a num-
ber of related systems (6b–6d).
The ultimate product of thermodynamic control is the C-3
para-bonded C-adduct, 5. Formation of this C-3 PicO-
Ph·PhO(H)– Meisenheimer complex further illustrates the
ambident (O- and C-) nucleophilic nature of phenoxide;
compound 5 is formed with effective irreversibility. It is ap-
parent that this C-3 C-adduct is the product of kinetic prefer-
ence for C-attack at the 3-position. AM1 calculations (6d)
suggest that the C-3 C-adduct is also the product of thermo-
dynamic control. The thermodynamic preference for C-3 at-
tack by carbon nucleophiles is also implicit in the reaction
scheme of the VNS reaction (10). Thus, in the PicOPh–
PhO– system, PhO– (as a C-nucleophile) follows K3T3
regioselectivity.
In the nonsymmetric systems (PicOPh(1)–MesO– and
PicOMes(2)–PhO–), reaction of 1 with MesO– forms the C-1
PicOPh·OMes– Meisenheimer complex, 11, as the
kinetically favoured species (Scheme 2). Although 11 is the
1
dominant adduct observed in the first H NMR spectrum ac-
quired at –40°C, significant amounts of the C-1 PicO-
Ph·OPh– adduct, 3, were also present. This observation im-
© 2003 NRC Canada

Buncel et al.
451
Fig. 3. (a) Illustration of the stereoelectronic stabilization of the
plied that 11 broke down to yield low concentrations of 2
and free PhO–, which could then attack 1 to yield 3. In this
sense, two nucleophiles (MesO– and PhO–) and two sub-
strates (1 and 2) must be simultaneously present here and,
by extension, in the other nonsymmetric PicOMes(2)–PhO–
reaction system.
Importantly, the initially formed C-1 O-adducts of 1 and 2
eventually give way to the C-3 mesitoxide O-centred adducts
C-3 PicOMes·OMes–, 12, and the C-3 diadduct PicO-
Mes·(OMes)22–, 13. The assignment of the structures of 12
and 13 as progeny of 2 was confirmed by the separate study
of the symmetrical PicOMes(2)–MesO– system. Here, the
initially formed C-1 PicOMes·OMes– adduct, 14, decom-
poses in favour of the C-3 species, 12 and 13. Thus, the
complexity of the nonsymmetric systems (PicOPh–MesO–
and PicOMes–PhO–) can render only a tentative classifica-
tion as K1T3 (Fig. 1); however, as noted above, the simpler
symmetrical PicOMes–MesO– system can be definitively
classified as following K1T3 regioselectivity.
C-1 TNA·OPh– adduct 10 through antiperiplanar interaction be-
tween the lone pair on the methoxy oxygen and the C— OPh
bond; (b) Illustration of the TNA·OMes– adduct, where stereo-
electronic stabilization as in (a) is not possible; (c) Illustration of
the C-1 PicOPh·OPh– adduct 3 where, as in (a), stereoelectronic
stabilization is possible; (d) Illustration of the C-1 PicOMes·OMes–
adduct 14 where, as in (b), stereoelectronic stabilization is not
possible.
Stereoelectronic stabilization of the C-1 adducts
Of the factors that stabilize the respective C-1 picryl ether
O-adducts, stereoelectronic n → σ* interactions have
emerged from our analysis of the TNA–PhO– and TNA–
MesO– systems as an important mechanism of stabilization
of these adducts (6b–d, 7). In brief, full n → σ* stabilization
is possible only when one O—R sigma bond is aligned
antiperiplanar to one of the nonbonding lone pairs of the
other C(1)-OR acetal-like group and vice versa (26), i.e., a
“doubly antiperiplanar” conformation. We previously argued
that in the C-1 TNA·OPh– adduct (i.e., 10), rotameric forms
that would permit single n → σ* interactions would be sig-
nificantly populated in solution. Thus, as depicted in Fig. 3,
Fig. 3a, which represents the most stable rotamer of 10, will
partake in this stereoelectronic stabilization through the anti-
periplanar configuration of orbitals and consequently, adduct
10 is more stable than its C-3 counterpart (6b). However,
from inspection of molecular models (Fieser or Darling) and
on the basis of downfield 13C chemical shifts (6c) that sug-
gested puckering of the cyclohexadienate ring of the TNA–
moiety, it became apparent that the rotamer shown in Fig. 3b
was preferred for the mesitoxide C-1 adduct of TNA, even
though it does not permit stereoelectronic stabilization (6c,
7). Consequently, the C-1 TNA·OMes– adduct is not as sta-
ble as its C-3 analogue (6c).
A further factor impinges on the degree of n → σ* stabili-
zation provided to a given C-1 adduct that is geminally di-
substituted by electronegative groups. Even if suitable dou-
bly antiperiplanar conformers are readily accessible through
rotation and even if such forms would be expected to be
populated at a given temperature, the efficacy of the n → σ*
stabilization will depend on the relative energies of the two
σ* fragment orbitals. Thus, n donation from OPh in adduct
10 to the σ* orbital of the OCH3 fragment may be more or
less effective than n donation from OCH3 to the σ* orbital of
OPh. In general then, unsymmetrical C-1 adducts should ex-
hibit less stereoelectronic stabilization than their symmetri-
cal counterparts regardless of whether their most favourable
rotameric forms are accessible or not (7, 26, 27). Conse-
quently, the C-1 PicOPh·OMe– adduct, 10, the C-1 PicO-
Mes·OMe– adduct, 18, the C-1 PicOMes·OPh– adduct, 11,
and the C-1 hydroxide adducts, 7 and 16, would all be ex-
pected to partake of less stereoelectronic stabilization than
their symmetrical analogues. This factor, no doubt, partly
accounts for the inability to observe the hydroxide adducts,
for example.
Focusing on the symmetrical aryloxide systems
(PicOPh(1)–PhO– and PicOMes(2)–MesO–), whose regio-
selectivity could be clearly defined, it would be expected
that the C-1 PicOPh·OPh– adduct, 3, if it could achieve suit-
able rotameric forms that would permit n → σ* donation,
would be stabilized relative to its C-3 counterpart, 4. There-
fore, the PicOPh–PhO– system would be expected to display
thermodynamic preference for formation of the C-1 adduct,
i.e., T1 in the observed K1T1 regioselectivity, and 4 would
not be detected. Relating the C-1 PicOPh·OPh– adduct, 3, to
its TNA analogue (10, Fig. 3a), inspection of molecular
models (Fieser or Darling) predicts that the rotamer repre-
sented in Fig. 3c would be the most stable for 3. This con-
figuration would be stabilized by n → σ* donation, and
consequently 3 would be more stable than its C-3 analogue
4. These expectations are borne out by the observation of 3
as the sole O-adduct in this system.
For the PicOMes(2)–MesO– system, the rotamer repre-
sented in Fig. 3d is favoured for the C-1 PicOMes·OMes–
adduct, 14, on inspection of molecular models. This adduct
is even more sterically congested than its TNA analogue
(Fig. 3b), and 14 will be unable to derive any stabilization
from n → σ* donation. In fact, the non-equivalence of the
2,4,6-trinitrocyclohexadienate ring protons of 14 showed
that to accommodate the bulky mesitoxyl groups at C-1, the
adduct must sacrifice delocalization of negative charge into
one of the ortho nitro groups as indicated in structure 14a.
© 2003 NRC Canada

452
Can. J. Chem. Vol. 81, 2003
The C-1 PicOMes·OMes– adduct (14) has a relatively short
lifetime even at low temperatures (i.e., 14 is no longer pres-
ent in the spectrum taken at –30°C, a temperature at which
the analogous C-1 PicOPh·OPh– adduct (3) is the only
adduct seen in the PicOPh–PhO– system). The lack of
stereoelectronic stabilization and the consequent instability
of 14 distinguishes the PicOMes–MesO– system that follows
K1T3 regioselectivity from the PicOPh–PhO– system that
exhibits K1T1 behaviour.
On the basis of the energetics, phenoxide (and, by exten-
sion, mesitoxide and other nucleophiles with relatively low-
lying HOMO energies) may interact most strongly with the
LUMO of TNA and, consequently, attack C-1, the site of
highest orbital coefficient, preferentially. Conversely, hy-
droxide and methoxide may interact most strongly with the
SUMO of TNA, which has its highest orbital coefficients at
C-3 and C-5. Hence, aryloxides would show K1 behaviour
while alkoxides and (or) hydroxide display K3 behaviour.
However, such an analysis attributes significant changes in
regioselectivity to relatively small energy differences; the
energy gap between LUMO and SUMO is not large, and in
fact, the HOMO of aryloxide nucleophiles would be ex-
pected to interact both with the LUMO and SUMO.
Molecular orbital approaches to the kinetic preferences
Frontier molecular orbital analysis
An important question arises from the above consider-
ations. If the PicOMes·OMes– adduct, 14, is inherently so
unstable, why does mesitoxide attack at C-1 occur at all?
Moreover, given the steric bulk of the mesitoxide nucleo-
phile, why is C-1 attack by MesO–, particularly on the con-
gested C-1 centre in PicOMes, 2, favoured kinetically?
Molecular orbital theory, and, notably, the frontier molec-
ular orbital (FMO) approach (28, 29), have been invoked to
rationalize the regioselectivity in aromatic substitution reac-
tions. As applied to the current systems, the FMO treatment
holds that the preferred sites of attack will be those that per-
mit maximum orbital overlap between the incoming (donor)
nucleophile and the nitroaromatic (acceptor) substrate. In
turn, this implies that the regioselectivity will be determined
by the magnitude of the lowest unoccupied molecular orbital
(LUMO) lobe at each ring carbon, as indicated by the orbital
coefficients at each site, for a given nitroaromatic substrate.
Given the similarity in regioselectivity found for PicOPh,
1, and PicOMes, 2, in the current study and the behaviour of
TNA with the same nucleophiles, namely PhO– (6b), MesO–
(6c), and MeO– and OH– (6b, 6c, 5a, 5b), TNA would ap-
pear to be an acceptable general model for all picryl ethers.
In this regard, we have undertaken AM1 calculations on
TNA (and some related compounds (6d)). The two degener-
ate LUMOs of 1,3,5-trinitrobenzene become split upon in-
troduction of the C-1 methoxyl group to form TNA. The
AM1 calculated LUMO and superjacent unoccupied molecu-
lar orbital (SUMO) are illustrated in Fig. 4. The energy of
the LUMO is –2.50 eV according to the AM1 calculation,
while the SUMO is higher in energy at –2.39 eV (a differ-
ence of 0.11 eV or 2.5 kcal mol–1).
Examination of the LUMO (Fig. 4) shows that C-1 is the
site with the largest orbital coefficient. Interestingly, C-4,
which bears a nitro group, is the site with the second highest
orbital coefficient. The regioselectivity found should, in the
FMO treatment, depend also on the energy of the highest oc-
cupied molecular orbital (HOMO) of the attacking nucleo-
phile. The most favourable interaction will then be between
nucleophile HOMO and substrate LUMO, when these are
similar in energy.
Estimates of the HOMO energies of some of the anions
have been made by Pearson (30) using the approximation
that the electron affinity of the corresponding radicals (i.e.,
PhO•, OH•) represents the ionization potential (IP) of the an-
ions that, according to Koopmans’ theorem (31), is taken to
be equal to the negative value of the HOMO energies.
Therefore, the HOMO energy for PhO– and for OH– is esti-
mated to be –2.35 and ca. –1.83 eV, respectively.
An alternative to the FMO approach is the configuration
mixing model that is considered in the next section.
Configuration mixing model: Polar vs. SET pathways
The configuration mixing model advanced by Shaik,
Pross, and Hoz in a series of articles (32) and reviews (16)
defines the full reaction profile for an organic reaction by as-
suming, in general, that transition states possess varying pro-
portions of covalent-bonding and radical character. This
model provides a mechanistic spectrum between a single
electron transfer (SET) pathway and a polar one, where the
polar process involves synchronous electron shift and bond
formation through a radical coupling pathway. In the SET
route, SET precedes bond formation. Thus, which particular
pathway is followed in any given reaction depends on the
feasibility of coupling of the two spin-paired electrons (bond
formation) following the electron shift. Factors governing
SET vs. polar pathways include effects of donor–acceptor
ability, steric interactions, the donor–acceptor bond strength,
and radical delocalization (16b). Taking these factors into
consideration allows assessment of which donor–acceptor
pair is likely to follow the SET pathway.
In the present system, single electron transfer from the
nucleophile to the picryl aryl ether would produce the radi-
cal anion of the picryl aryl ether and the requisite radical of
the nucleophile. Change in nucleophile from an alkoxide-
HO– to an aryloxide may cause a shift from a polar to a SET
pathway. Electrochemical measurements in acetonitrile show
that the one-electron half-peak oxidation potential (E1/2) for
the conversion of phenoxide into the phenoxyl radical is
~0.30 V vs. NHE. The corresponding value for HO– is
~0.6 V (33). The lower oxidation potential of phenoxide in-
dicates that the phenoxyl radical is more stable than the
hydroxyl radical as a consequence of the stabilizing effect
exerted by the neighbouring aromatic ring on the radical
centre. Further, since methyl substituents on a benzene ring
are known to impart further stability to a benzylic (and pre-
sumably to an analogous aryloxide) radical, mesitoxyl radi-
cal would be expected to be more stable than phenoxyl
radical and, therefore, to form even more readily (34, 35).
Clearly the SET donor ability of the aryloxide is superior to
that of alkoxides-HO–. In terms of steric interactions, it is
equally clear that mesitoxide is a more sterically hindered
nucleophile than methoxide-HO–. Steric hindrance favours
the SET pathway since the polar pathway is energetically fa-
vourable when the reacting species can approach each other
to within bonding distance. Steric repulsions will cancel en-
© 2003 NRC Canada

Buncel et al.
453
Fig. 4. Orbital coefficients (pz), according to a semi-empirical calculation (AM1), for the lowest unoccupied molecular orbital (LUMO)
and for the superjascent unoccupied molecular orbital (SUMO) of 2,4,6-trinitroanisole (TNA), a model for picryl ethers, generally. (The px
and py coefficients are negligibly small). The LUMO, which is calculated to be lower in energy (–2.50 eV) than the SUMO (–2.39 eV),
has its largest ring orbital coefficient at C-1 (i.e., –0.600), whereas the largest ring orbital coefficients in the SUMO are located at C-3
and C-5 (0.512 and –0.529, respectively).
ergy lowering because of bonding changes in the transition
state, thus favouring the SET pathway, as SET processes
take place at distances significantly greater than those at
which incipient bonding takes place (16b). Furthermore,
alkoxides and HO– form strong bonds to give stable
Meisenheimer complexes and generate localized radicals.
These factors favour the polar pathway. In contrast, the
aryloxides generate relatively unstable O-adducts that are
transient in DMSO at room temperature (8, 9a) and
delocalization of the aryloxide radical inhibits bond forma-
tion through radical coupling. These factors favour the SET
pathway.
steric interactions at C-1. However, if the electron shift pre-
cedes bond formation, then the spin density of the picryl
ether radical anion dictates C-1 attachment even though C-1
is clearly the more sterically hindered site.
In summary, on the basis of the above argument it can be
concluded that hydroxide and methoxide follow the polar
(SNAr) pathway; this will favour kinetically C-3 attachment
due to F-strain at C-1. However, phenoxide and mesitoxide
follow the SET pathway. In these cases, C-1 attachment is
favoured kinetically due to the higher C-1 spin density in the
radical anion of the picryl ether.
Shown in Fig. 5 is a schematic illustrating the relationship
between the polar and SET reaction pathways (16b) for the
present picryl aryl ether – nucleophile systems. The reac-
tants are in the lower left corner where phenoxide and
mesitoxide are represented as ArO–. The polar process is in-
dicated by the diagonal arrow, where both electron transfer
and bond formation are synchronous.
Conclusions
The present results on the course of the reactions between
picryl phenyl ether (PicOPh, 1) and picryl mesityl ether
(PicOMes, 2) with the oxygen-nucleophiles, phenoxide,
mesitoxide, hydroxide, and methoxide allow us to make the
following conclusions.
The reactions of HO– and MeO– with 1 and 2 carried out
in DMSO-d6 at ambient temperature show K3 behaviour, as
documented in numerous picryl ether – alkoxide systems
(5a–d). In contrast, phenoxide and mesitoxide display ki-
netic preference for C-1 attachment, i.e., K1 behaviour. This
observation is analogous to the reactivity of these aryloxide
nucleophiles towards 2,4,6-trinitroanisole (6b, 6c), and now
appears to be a general trend for the reactions of aryloxide
nucleophiles with picryl ethers. For the reaction of phen-
oxide with 1, the C-1 O-adduct is also thermodynamically
favoured for phenoxide acting as an O-nucleophile. Stereo-
electronic stabilization of the C-1 PicOPh·OPh– adduct
through n → σ* donation confers thermodynamic preference
on this C-1 adduct relative to its C-3 oxygen-centred regio-
isomer. With mesitoxide ion, the C-1 adducts of both sub-
strates are less stable than their C-3 counterparts. In these
cases the steric congestion in the C-1 adducts associated
with the presence of the ortho methyl groups of mesitoxyl
moiety precludes significant stereoelectronic stabilization.
Now the C-3 adducts of mesitoxide are more stable and the
The SET pathway depicted in Fig. 5 involves first electron
transfer to generate the picryl aryl ether radical anion and
the aryloxide radical (ArO–) shown in the upper left corner.
The second step involves bond formation. Now the regio-
selectivity (C-1 or C-3) of Meisenheimer complex formation
would be controlled by the spin density at the C-1 or C-3
positions of the radical anion of the picryl aryl ether, an in-
termediate on the reaction pathway. The position of higher
spin density would be expected to be the site where radical
coupling will be most favoured. Calculation of the spin den-
sity of the radical anion of TNA, the model for picryl ethers,
using the AM1 method shows that C-1 is the site of highest
spin density. Thus, subsequent radical combination to form the
Meisenheimer complex would be expected to occur at C-1.
The configuration mixing model provides a rationale for
the regioselectivity observed in reactions of picryl ethers
with O-centred nucleophiles. The regioselectivity depends
on the degree of concertedness of electron shift and bond
formation. When electron shift and bond formation are con-
certed, C-3 attachment is kinetically favoured because of
© 2003 NRC Canada

454
Can. J. Chem. Vol. 81, 2003
Fig. 5. Schematic of a potential energy surface diagram illustrating the relationship between polar and SET reaction pathways for reac-
tion of 1 and (or) 2 with aryoxide (ArO–) nucleophiles to give the C-1 Meisenheimer adducts.
initially formed C-1 adducts decline in favour of their C-3
regioisomers.
spective phenol and standard MeOK–MeOH in an N2-filled
glovebox as described previously (6b, 6c). 2,4,6-Trimethyl-
phenol (mesitol; Aldrich) was recrystallized from petroleum
ether and dried in vacuo before use. Trifluoroacetic acid
(TFA; Aldrich) and tetramethylammonium hydroxide
(Me4NOH, 25 wt % solution in water; Aldrich) were used
without further purification. Melting points were measured
on a Thomas Hoover capillary apparatus and are not cor-
rected.
Although steric inhibition of stereoelectronic stabilization
can account for the changeover in adduct stability (i.e., C-1
PhO– adduct of 1 are more stable than its C-3 counterpart,
whereas the C-1 MesO– adducts of 1 and 2 are less stable
than their C-3 isomers), there is still kinetic preference to
formation of the C-1 adducts of 1 and 2 even when the at-
tacking nucleophile is the bulky mesitoxide ion. Frontier
molecular orbital considerations suggest that all nucleophiles
would preferentially attack the C-1 site of picryl ethers,
based on AM1 calculations of 2,4,6-trinitroanisole (TNA),
as a general model for picryl ethers. An alternative explana-
tion, which we propose, is that the reactions involving the
aryloxides do not follow the polar SNAr pathway, but rather
proceed according to the SET pathway outlined in Fig 5.
NMR experiments: General
NMR experiments were carried out using a Bruker AM-
400 specrometer (1H: 400.1 MHz, 13C: 100.0 MHz) in
MeCN-d3:DME-d10 (1:1 v/v) or in dried DMSO-d6. In the
mixed solvent system, CD2HCN served as chemical shift
reference (1H: δ = 1.93 ppm) and lock signal, whereas reso-
nances found in spectra determined in DMSO were refer-
enced to the peak for residual CD3SOCD2-H in the solvent
(1H: δ = 2.50 ppm). Chemical shifts are given in parts per
million (ppm) and coupling constants (J) are reported in
hertz. Wilmad PP-507 NMR tubes (5 mm) were used in all
experiments. All stock solutions were prepared in the appro-
priate solvents under a nitrogen atmosphere. NMR tubes
were capped with rubber septa and swept out with dry N2
prior to injection of the reactants into the NMR tube with a
gas-tight syringe.
Experimental
Materials and methods
2,4,6-Trinitrophenyl phenyl ether (PicOPh,1) was pre-
pared from picryl chloride and potassium phenoxide in etha-
nol as described by Dyall (36), mp 154–155°C (lit. (36)
value mp 153–154°C). 2,4,6-Trinitrophenyl-2′,4′,6′-trimethyl-
phenyl ether (PicOMes, 2) was prepared from picryl chlo-
ride and ethanolic potassium mesitoxide, mp. 157–158°C
(lit. (37) value mp 154–155°C). Picryl chloride used in these
preparations resulted from the reaction of pyridinium picrate
with POCl3 (38), mp. 80–81°C, after recrystallization from
CCl4 (lit. (39) value mp 81°C). Acetonitrile-d3 and dimeth-
oxyethane-d10 (Merck) were dried by sequential treatment
with 4 Å molecular sieves as advocated by Burfield et al.
(17). 1,4-Dibromobenzene (DBB; Eastman) was recrys-
tallized from ethanol and dried in vacuo prior to use as an
internal integration standard in the NMR experiments. Potas-
sium phenoxide and mesitoxide were prepared from the re-
Low temperature NMR experiments in MeCN–DME
(1:1)
Typically, a weighed quantity of the aryloxide was dis-
solved in a 1:1 (v/v) mixture of MeCN-d3:DME-d10 under a
nitrogen atmosphere such that 300 µL of this stock solution
would yield 1–1.5 equiv of the aryloxide relative to the sub-
strate. An aliquot of the stock solution (300 µL) was injected
into the NMR tube, and the solution frozen by immersion in
liquid N2. To the frozen solution, 1 equiv of PicOPh or
© 2003 NRC Canada

Buncel et al.
455
3. (a) E.A. Symons and E. Buncel. Can. J. Chem. 50, 1729
(1972); (b) E. Buncel and E.A. Symons. J. Org. Chem. 38,
1201 (1973); (c) E. Buncel and A.W. Zabel. Can. J. Chem. 59,
3168; 3177 (1981).
PicOMes was injected by a gas-tight syringe (200 µL; final
concentrations 0.06–0.09 M). The final mixture was placed in
a dry ice – acetone bath that had been maintained at –50°C.
The contents of the tube were allowed to mingle at this tem-
perature. The tube was inverted several times to promote
mixing and was then immersed again in liquid N2. The tube
was transferred to the spectrometer probe (at –40°C), the in-
strument was tuned as previously described (6b), and spectra
were recorded at various intervals. A standard collection of
FID would be made at 3, 5, 7, and 9 min and then as war-
ranted by observed changes in the spectrum. Simulta-
neously, the temperature of the probe was gradually raised.
4. (a) E. Buncel, J.M. Dust, K.T. Park, R.A. Renfrow, and M.J.
Strauss. In Nucleophilicity. Edited by J.M. Harris and S.P.
McManus. ACS Adv. in Chem. Ser. 215. American Chemical
Society, Washington, D.C. 1987. pp 369–383; (b) E. Buncel,
M.R. Crampton, M.J. Strauss, and F. Terrier. Electron deficient
aromatic – and heteroaromatic – base interactions. The chem-
istry of anionic sigma complexes. Elsevier, Amsterdam. 1984;
(c) E. Buncel. In The chemistry of functional groups. Supple-
ment F. The chemistry of amino, nitro, and nitroso com-
pounds. Edited by S. Patai. Wiley, London. 1982; (d) F.
Terrier. Chem. Rev. 82, 77 (1982); (e) F.A. Artamkina, M.P.
Egorov, and I.P. Beletskaya. Chem. Rev. 82, 427 (1982);
(f) C.F. Bernasconi. MTP Int. Rev. Sci. Org. Chem. Ser. One,
3, 33 (1973); (g) M.J. Strauss. Chem. Rev. 70, 667 (1970); (h)
E. Buncel, A.R. Norris, and K.E. Russell. Quart. Rev. Chem.
Soc. London, 22, 123 (1968).
5. (a) K.L. Servis. J. Am. Chem. Soc. 87, 5495 (1965); (b) K.L.
Servis. J. Am. Chem. Soc. 89, 1508 (1967); (c) R.A. Foster,
C.A. Fyfe, P.H. Emslie, and M.I. Foreman. Tetrahedron, 23,
227 (1967); (d) C.F. Bernasconi. J. Am. Chem. Soc. 92, 4682
(1970); (e) G. Baldini, G. Doddi, G. Illuminati, and F. Stegel.
J. Org. Chem. 41, 2153 (1976); (f) M.P. Simonnin, M.J. Pouet,
and F. Terrier. J. Org. Chem. 43, 855 (1978); (g) A.D.A.
Alaruri and M.R. Crampton. J. Chem. Res. Synop. 140 (1980);
J. Chem. Res. Miniprint, 2157 (1980); (h) M.P. Simonnin, J.C.
Halle, F. Terrier, and M.J. Pouet. Can. J. Chem. 63, 866
(1985).
Room temperature experiments in DMSO
A solution of the substrate was prepared in DMSO-d6 un-
der an N2 atmosphere. An aliquot of this solution was trans-
ferred via syringe into an NMR tube and an initial spectrum
run to ascertain the purity of the substrate. To this solution
1 equiv of nucleophile was injected through the septum and
spectra were recorded as rapidly as possible initially and
then at longer times as the reaction proceeded. 1,4-Dibromo-
benzene was present in the solution of the substrate and in
the NMR tube as an internal integration standard.
Molecular orbital calculations
Semi-empirical calculations were performed using
AMPAC version 2.10 of AM1 (15a) on an IBM 3081 com-
puter. The structure of 2,4,6-trinitroanisole (TNA) was cal-
culated using the PRECISE option within the restricted
Hartree–Fock (RHF; closed shell) scheme. The TNA struc-
ture was fully optimized following a Fletcher–Powell
minimization (40) to a self-consistent field (SCF) that satis-
fied Herbert’s test. The final structure was characterized as a
local minimum when no negative force constants were ob-
tained using the FORCE option.
6. (a) E. Buncel, S.K. Murarka, and A.R. Norris. Can. J. Chem.
62, 534 (1984); (b) E. Buncel, J.M. Dust, A. Jonczyk, R.A.
Manderville, and I. Onyido. J. Am. Chem. Soc. 114, 5610
(1992); (c) R.A. Manderville and E. Buncel. J. Am. Chem.
Soc. 115, 8985 (1993); (d) E. Buncel, R.M. Tarkka, and J.M.
Dust. Can. J. Chem. 72, 1709 (1994).
7. E. Buncel, J.M. Dust, and F. Terrier. Chem. Rev. 95, 2261
(1995).
8. C.F. Bernasconi and M.C. Muller. J. Am. Chem. Soc. 100,
5530 (1978).
9. (a) E. Buncel and R.A. Manderville. J. Phys. Org. Chem. 6, 71
(1993); (b) E. Buncel and J.M. Dust. Can. J. Chem. 66, 1712
(1988); (c) E. Buncel and W. Eggimann. J. Am. Chem. Soc.
99, 5958 (1977); (d) E. Buncel, J.G.K. Webb, and J.F.
Wiltshire. J. Am. Chem. Soc. 99, 4429 (1977); (e) E. Buncel
and W. Eggimann. Can. J. Chem. 54, 2436 (1976); (f) E.
Buncel, A. Jonczyk, and J.G.K. Webb. Can. J. Chem. 53, 3761
(1975); (g) E. Buncel and J.G.K. Webb. J. Am. Chem. Soc. 95,
328 (1973).
Acknowledgments
This research has been financed by a number of agencies.
The support of the Natural Sciences and Engineering Re-
search Council of Canada (NSERC) (to E.B.) and Wake For-
est University (to R.A.M.) as well as the award of an R.T.
Mohan Fellowship (to R.M.T.) is gratefully acknowledged.
Discussions with Professors S. Hoz and F. Terrier were help-
ful, as always, and are appreciated.
References
10. (a) M. Makosza. Pol. J. Chem. 66, 3 (1992); (b) M.
Makosza,W. Daniekiewicz, and E. Wojciechowski. Phosphorus
Sulfur Silicon Relat. Elem. 53, 457 (1990); (c) M. Makosza
and A. Kinowski. Bull. Pol. Acad. Sci. 37, 127 (1989); (d) M.
Makosza. Russ. Chem. Rev. 58, 747 (1989); (e) M. Makosza
and J. Winiarski. Acc. Chem. Res. 20, 282 (1987).
11. (a) M.R. Crampton and V. Gold. J. Chem. Soc. 4293 (1964);
(b) M.R. Crampton and V. Gold. J. Chem. Soc. B, 893 (1966).
12. (a) C.F. Bernasconi. J. Am. Chem. Soc. 93, 6975 (1971);
(b) M.R. Crampton and H.A. Khan. J. Chem. Soc. Perkin
Trans. 2, 1173 (1972); (c) M.R. Crampton, B. Gibson, and
F.W. Gilmore. J. Chem. Soc. Perkin Trans. 2, 91 (1979); (d) R.
Destro, C. Grammicioli, and M. Simonetta. Acta Crystallogr.
Sect. B Struct. Crystallogr. Cryst. Chem. 24, 1369 (1968);
1. (a) R.A. Foster. J. Phys. Chem. 84, 2135 (1980); (b) M.A.
Slifkin. Mol. Interact. 2, 271 (1981).
2. (a) R. Bacaloglu, C.A. Bunton, and G. Cerichelli. J. Am.
Chem. Soc. 109, 621(1987); (b) R. Bacaloglu, C.A. Bunton,
G. Cerichelli, and F. Ortega. J. Am. Chem. Soc. 110, 3495
(1988); (c) R. Bacaloglu, C.A. Bunton, and F. Ortega. Int. J.
Chem. Kinet. 20, 195 (1988); (d) R. Bacaloglu, A. Blasko,
C.A. Bunton, F. Ortega, and C.Zucco. In Atual. Fis-Quim.
Org. Conf. Latinoam. Fis-Quim. Org. 1st, Edited by E.
Humeres. 1991. p. 165; (e) R. Bacaloglu, A. Blasko, C.A.
Bunton, E. Dorwin, F. Ortega, and C. Zucco. J. Am. Chem.
Soc. 113, 238 (1991); (f) R. Bacaloglu, A. Blasko, C.A.
Bunton, F. Ortega, and C. Zucco. J. Am. Chem. Soc. 114,
7708 (1992).
© 2003 NRC Canada

456
Can. J. Chem. Vol. 81, 2003
(e) H. Ueda, N. Sakabe, J. Tanaka, and A. Furusaki. Bull.
Chem. Soc. Jpn, 41, 2866 (1968).
215. American Chemical Society, Washington, D.C. 1987. pp
137–155.
13. M.R. Crampton. Adv. Phys. Org. Chem. 7, 211 (1969).
14. A. Cooney and M.R. Crampton. J. Chem. Soc. Perkin Trans. 2,
1973 (1984).
15. (a) M.J.S. Dewar, E.G. Zoeblisch, E.F. Healy, and J.J.P. Stew-
art. J. Am. Chem. Soc. 107, 3902 (1985); (b) M.J.S. Dewar
and K.M. Dieter. J. Am. Chem. Soc. 108, 8075 (1986).
16. (a) A. Pross and S.S. Shaik. Acc. Chem. Res. 16, 363 (1983);
(b) A. Pross. Acc. Chem. Res. 18, 212 (1985); (c) S. Hoz. Acc.
Chem. Res. 26, 69 (1993).
17. D.R. Burfield and R.H. Smithers. J. Org. Chem. 43, 3966
(1978).
18. (a) B. Gibson and M.R. Crampton. J. Chem. Soc. Perkin
Trans. 2, 648, (1979); (b) C.F. Bernasconi, J. Fassberg, R.B.
Killiom, Jr., D.F. Schuck, and Z. Rappoport. J. Am. Chem.
Soc. 113, 4937 (1991).
25. (a) E.H. Baughman and M.M. Krevoy. J. Phys. Chem. 4, 421
(1974); (b) F.G. Bordwell. Acc. Chem. Res. 21, 456 (1988).
26. (a) A. J. Kirby. The anomeric effect and related stereo-
electronic effects at oxygen. Springer-Verlag, Berlin. 1983. pp
9–11, 82, 90, 135–136; (b) P. Deslongchamps. Stereoelectronic
effects in organic chemistry. Pergamon Press, New York. 1983.
pp 5–20.
27. M.L. Sinnott. Adv. Phys. Org. Chem. 24, 114 (1988).
28. K. Fukui, T. Yonezawa, C. Nagata, and H. Shingu. J. Chem.
Phys. 22, 1433 (1954).
29. I. Fleming. Frontier orbitals and organic reactions. Wiley. New
York. 1976.
30. R.G. Pearson. Inorg. Chem. 27, 734 (1988).
31. T. Koopmans. Physica Utrecht, 1, 34 (1934).
32. (a) S.S. Shaik. Prog. Phys. Org. Chem. 15, 197 (1985);
(b) S.S. Shaik. J. Am. Chem. Soc. 103, 3692 (1981); (c) A.
Pross. Isr. J. Chem. 26, 360 (1985).
19. (a) J.F. Bunnett and R.E. Zahler. Chem. Rev. 49, 273 (1951);
(b) J.F. Bunnett. Quart. Rev. Chem. Soc. London, 12, 1 (1958);
(c) J.A. Orvik and J.F. Bunnett. J. Am. Chem. Soc. 92, 2417
(1970).
33. P.K.S. Tsang, P. Cofre, and D.T. Sawyer. Inorg. Chem. 26,
3604 (1987).
20. (a) J. Miller. Nucleophilic aromatic substitution. Elsevier, Am-
sterdam. 1968; (b) F. Terrier. Nucleophilic aromatic displace-
ment. The influence of the nitro group. VCH, New York. 1991.
21. C.J. Pouchert. The Aldrich library of NMR spectra. Vol. 1.
Aldrich, Milwaukee, WI. 1986. p. 355
34. (a) J.M. Dust and D.R. Arnold. J. Am. Chem. Soc. 105, 1221
(1983); (b) D.D.M. Wayner and D.R. Arnold. Can. J. Chem.
62, 1164 (1984).
35. R.A. Jackson and M. Sharifi. J. Chem. Soc. Perkin Trans. 2,
775 (1996).
22. (a) J. Trotter. Acta Crystallogr. 12, 605 (1959); (b) P. Coppens.
Acta Crystallogr. 17, 573 (1964); (c) F. Terrier, L. Xaio, P.G.
Farrell, and D. Moskowitz. J. Chem. Soc. Perkin Trans. 2,
1259 (1992); and refs. cited therein.
23. G.R. Sprengling and C.W. Lewis. J. Am. Chem. Soc. 75, 5709
(1953).
36. L.K. Dyall. J. Chem. Soc. 5160 (1960).
37. L.B. Clapp, H. Lacey, G.G. Beckwith, R.M. Srivastava, and N.
Muhammad. J. Org. Chem. 33, 4262 (1968).
38. R. Boyer, E.Y. Spencer, and G.F. Wright. Can. J. Res. 24B,
200 (1946).
39. P.F. Frankland and F.H. Garner. Chem. Ind. London, 39, 257
(1921).
40. R. Fletcher and M.J.D. Powell. Comput. J. 6, 163 (1963).
24. F. Bordwell, T.A. Cripe, and D.L. Hughes. In Nucleophilicity.
Edited by J.M. Harris and S.P. McManus. Adv. Chem. Ser.
© 2003 NRC Canada