The influence of nucleophile substituents on the orientation in the reaction between 2,4-difluoronitrobenzene and lithium phenoxides in liquid ammonia

Article abstract of DOI:10.1002/1099-0690(200101)2001:2<405::AID-EJOC405>3.0.CO;2-6

The dependence of the orientation of aryloxydefluorination of 2,4-difluoronitrobenzene (1) (o/p ratio) by the action of X-substituted lithium phenoxides 2 (X = p-OMe, p-Me, p-Et, p-iPr, p-tBu, m-Me, H, p-F) in liquid ammonia in the temperature range from -55 to -35 °C has been investigated. The enthalpic preference for ortho-fluorine substitution decreases with weakening substituent electron-donating capability in the order: p-OMe > p-Me ≈ p-Et > m-Me > H ≥ p-F. The predominant fluorine substitution at the ortho position for X = p-Me, p-Et turns into a preference for substitution at the para position when X = p-iPr, and this increases further on going to X = p-tBu, PM3, AM1 and MNDO MO calculations showed greater stability of the intermediate anionic σ-complexes formed on nucleophile addition at the para position, thus suggesting that the predominant ortho substitution manifested for X = p-OMe, m-Me, H, p-F and p-Alk = Me, Et is due to control over orientation by the charge distribution in the substrate. The substrate electronic structure, as a controlling factor, is probably changed by the relative stability of intermediate anionic σ-complexes on going to p-Alk = iPr, tBu, as a consequence of an enhancement of the substituent's electron-withdrawing nature with the increase in alkyl group polarizability in the order: p-Me ≈ p-Et < p-iPr < p-tBu.

Full text of DOI:10.1002/1099-0690(200101)2001:2<405::AID-EJOC405>3.0.CO;2-6

FULL PAPER
The Influence of Nucleophile Substituents on the Orientation in the Reaction
between 2,4-Difluoronitrobenzene and Lithium Phenoxides in Liquid Ammonia
Larisa Politanskaya,[a] Evgenij Malykhin,*[a] and Vitalij Shteingarts*[a]
Keywords: Aromatic nucleophilic substitutions / 2,4-Difluoronitrobenzene / Liquid ammonia / Lithium phenoxides /
Substituent effects in nucleophile
The dependence of the orientation of aryloxydefluorination
of 2,4-difluoronitrobenzene (1) (o/p ratio) by the action of X-
substituted lithium phenoxides 2 (X = p-OMe, p-Me, p-Et, p-
showed greater stability of the intermediate anionic σ-com-
plexes formed on nucleophile addition at the para position,
thus suggesting that the predominant ortho substitution
iPr, p-tBu, m-Me, H, p-F) in liquid ammonia in the temper- manifested for X = p-OMe, m-Me, H, p-F and p-Alk = Me,
ature range from −55 to −35 °C has been investigated. The
enthalpic preference for ortho-fluorine substitution decreases
with weakening substituent electron-donating capability in
the order: p-OMe Ͼ p-Me ഠ p-Et Ͼ m-Me Ͼ H Ն p-F. The
predominant fluorine substitution at the ortho position for X =
p-Me, p-Et turns into a preference for substitution at the para
position when X = p-iPr, and this increases further on going
to X = p-tBu. PM3, AM1 and MNDO MO calculations
Et is due to control over orientation by the charge distribution
in the substrate. The substrate electronic structure, as a con-
trolling factor, is probably changed by the relative stability
of intermediate anionic σ-complexes on going to p-Alk = iPr,
tBu, as a consequence of an enhancement of the substituent’s
electron-withdrawing nature with the increase in alkyl group
polarizability in the order: p-Me ഠ p-Et Ͻ p-iPr Ͻ p-tBu.
Liquid ammonia has long been known as a perfect solv- benzenes in AlkONa/AlkOH, the ratios of ko/kp Ͼ 1 ob-
ent in which to carry out reactions involving strong bases served for Alk ϭ iPr and Et and ko/kp Ͻ 1 for Alk ϭ Me,
and nucleophiles, due to its endowing anionic species, such based in both cases under entropic control,[5,7,18] are
as amides and carbanions, with very high reactivity.[1] How- changed to ko/kp Ͼ 1 under enthalpic control when chan-
ever, its unprecedented suitability for aromatic nucleophilic ging to MeONa in liquid ammonia[3] (the isokinetic temper-
substitution reactions, conventionally considered as pro- ature is raised from Ϫ130 °C to ca. 0 °C in going from
ceeding via the intermediate formation of anionic Mei- MeOH[19] to liquid ammonia[3]). These findings have been
senheimer-type σ-complexes (SNAr mechanism), has only interpreted[3,15] as an indication that, at temperatures below
been demonstrated as recently as in the last few years.[2,3] 0 °C, solvation as a factor determining the relative stabilit-
Thus, the rate constant for the reaction of p-chloronitroben- ies of ortho and para substitution transition states for Alk ϭ
[8]
zene with sodium phenylthiolate has been found to be k Ն iPr and Et in their respective alcohols is replaced in the
3·10Ϫ2 L/mol·s in liquid ammonia at Ϫ45 °C,[2] compared case of MeONa in MeOH by the intrinsic structural charac-
with 4.8·10Ϫ4 L/mol·s in 60% aqueous dioxane at 25 °C.[4] teristics of reactants, and the operation of structural factors,
The reaction of o- and p-fluoronitrobenzenes with alkoxides as the governing ko/kp, under entropic control in the last
are 106Ϫ108 times faster in liquid ammonia[3] than in the case, is changed by their operation under enthalpic control
respective alcoholic media.[5Ϫ8] Obviously, the reason for on going to liquid ammonia. Enthalpically controlled
such an enhanced reactivity of nucleophiles in liquid values of ko/kp Ͼ 1 have also been found[15] for the reactions
ammonia lies in the combination of strong metal cation sol- of 2,4-difluoronitrobenzene (1) with ROM (R ϭ Me, Et,
vation[9,10] and weak anion solvation,[11] so that the situ- iPr, Ph; M ϭ Na, K) in liquid ammonia in the temperature
ation is similar to that in bipolar aprotic solvents.[12]
range from Ϫ70 to Ϫ33 °C, becoming substituted by en-
As well as the reaction rates, one aspect of the nature of tropically controlled values of ko/kp Ͻ 1 at still higher tem-
liquid ammonia as a solvent is manifested in the factors peratures (80Ϫ100 °C).
governing the correlation of substitution rates at positions
ortho and para to the nitro group (ko/kp): customarily a test-
This makes it worthwhile to study systematically the ef-
fects of diverse factors on the value and temperature de-
pendence of ko/kp, so as to improve understanding of the
influence of liquid ammonia as a solvent on the detailed
ing parameter for determining delicate details of nucleo-
philic substitution mechanisms in nitroarenes.[3,5Ϫ8,13Ϫ17] In
the alkoxydefluorination reactions of o- and p-fluoronitro-
mechanism of the reactions in question. Currently available
experimental data do not allow one to judge in detail the
relationship between the o/p ratio and the structural peculi-
arities of a reagent that can be used as a test case for the
investigation of the effect of nucleophile structure on the
character of transition states. For this purpose, the depend-
ence of the o/p ratio on nucleophile steric characteristics
[a]
N. N. Vorozhtsov Institute of Organic Chemistry, Siberian
Division of the Russian Academy of Sciences, Novosibirsk
State University,
Lavrentiev Ave. 9, Pirogova St. 2, Novosibirsk 630090, Russia
Fax: (internat.) ϩ 7-383/234-4752
E-mail: shtein@nioch.nsc.ru
malykhin@nioch.nsc.ru
Eur. J. Org. Chem. 2001, 405Ϫ411
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/0102Ϫ0405 $ 17.50ϩ.50/0
405

L. Politanskaya, E. Malykhin, V. Shteingarts
FULL PAPER
and ion association should be minimised. As for the steric ponding aryl 5-hydroxy-2-nitrophenyl ethers 6fϪh and aryl
factor, this requirement can be met by employing phenox- 3-hydroxy-4-nitrophenyl ethers 7fϪh having been isolated
ides containing varying substituents in positions meta and and characterized (Scheme 2).
para to the nucleophilic centre. As regards ion association,
Liϩ seems to be appropriate as a counter cation, since,
reasoning by the increase in solvation of alkali metal cat-
ions by liquid ammonia in the order Kϩ Ͻ Naϩ ϽϽ
Liϩ,[9,10] the regioselectivity observed for lithium phenolates
is not believed to be dominated by ion association,[14] par-
ticularly by chelate formation in the ortho substitution
Scheme 2
transition state (TS) (cf. ref.[5Ϫ8,13]). In line with this as-
sumption, the addition of bicyclohexyl-18-crown-6 resulted
In Table 1, Table 2 and Table 3, the spectral and analyt-
in a decreasing o/p ratio in the reaction of 1 with PhOK, ical data for 3, 6 and 7 are summarized. In the IR spectra of
but no change in this ratio in the case of PhOLi. Moreover, 7fϪh, the hydroxy group absorptions exhibit low frequency
the dependencies of o/p ratio on a substituent X in the shifts of ca. 100 cmϪ1 from their positions in the IR spectra
reactions of 1 with XC6H4OK and with XC6H4OLi are of p-nitrophenols,[21] including 6h; this is typical for com-
opposite.[14]
pounds with an intramolecular hydrogen bond.[22] This pe-
The goal of this work was to determine the temperature culiarity denotes mutual ortho location of nitro and hydroxy
dependence of the regioselectivity of phenoxydefluorination groups in 7fϪh, and in turn justifies the assignment of the
of 2,4-difluoronitrobenzene (1) with lithium phenoxides aforementioned signals at δ ഠ 50, disappearing from the
XC6H4OLi (2aϪh) [X ϭ p-OMe (a), m-Me (b), H (c), p- 19F NMR spectra after hydroxydefluorination, to 4fϪh.
F (d); p-Alk ϭ Me (e), Et (f), iPr (g), tBu (h)] in liquid
ammonia (Scheme 1).
The regioselectivity of phenoxydefluorination of 1 was
determined from the ratio of integral intensities of the 19F
NMR signals at δ ഠ 62 and 50, belonging to isomers 3 and
4, respectively. The data listed in Table 4 and 5 characterize
the dependence of o/p ratio on the nature of a substituent
in the nucleophile, over the temperature range of Ϫ33 to
Ϫ55 °C. It can be seen that the proportion of the ortho
substitution product increases with decreasing electron-do-
nating capability of the substituent in the nucleophile (i.e.,
on passing along the substituent series from p-OMe to p-F)
and decreasing temperature (Table 4). The latter finding, in
conjunction with the o/p ratio of Ͼ 1, suggests that the re-
gioselectivity is controlled by enthalpy over the temperature
range from Ϫ55 to Ϫ33 °C. For p-methyl- and p-ethyl-
Scheme 1
Results and Discussion
The reactions were performed with 1 and 2, taken in the phenoxides (2e,f), the proportion of ortho substitution also
ratio of 2:1, at Ϫ55 to Ϫ33 °C and quenched after reaching increases with decreasing temperature, which, together with
30Ϫ50% conversion of phenoxide (10Ϫ30 min) by adding the o/p value of Ͼ 1, indicates enthalpic control of regiose-
an excess of ammonium chloride. It was established that lectivity over the temperature range used (Table 5). For p-
one fluorine atom only is replaced, to give the respective Me and p-Et, ortho substitution is enthalpically more ad-
aryl 5-fluoro-2-nitrophenyl (3aϪh) and aryl 3-fluoro-4- vantageous than para substitution, the isokinetic temper-
nitrophenyl (4aϪh) ethers. In a parallel pathway, fluorine ature lying above Ϫ33 °C. In contrast, when p-isopropyl-
substitution by the amino group occurs in 1 at the position and p-tert-butylphenoxides (2g and 2h) are used as nucleo-
ortho to the nitro group, to form 3-fluoro-6-nitroaniline (5) philes, the isokinetic temperature shifts into the range from
in a maximum yield not exceeding 10% (according to the Ϫ33 to Ϫ55 °C. This fact is illustrated by the alteration of
19F NMR spectroscopic data) after 30 min reaction time.
o/p ratio from Ͼ 1 at Ϫ33 °C to Ͻ 1 at Ϫ55 °C, and corre-
The 19F NMR spectra of the reaction mixtures (C6F6 as sponds to the change from entropically controlled regiose-
internal reference) contained signals at δ ഠ 62 and 50 (be- lectivity to that controlled by enthalpy upon lowering the
longing to aryloxydefluorination products), at δ ϭ 60.3 (be- temperature, with para substitution of fluorine becoming
longing to 5),[20] and at δ ϭ 50.9 and 64.5 (belonging to the more preferred (Figure 1).
starting compound 1). The positions of the product signals
The differences between the activation parameters of
were practically invariant over the series of substituents phenoxydefluorination of 1 at the positions ortho and para
(aϪh) and had been anticipated for compounds 3 and 4, to the nitro group were calculated by the least-squares
respectively. The structures of the products 3aϪe and 4aϪe method, from the temperature dependencies of the o/p ratio
have been established previously.[17] Attempts to separate (more than 6 values at various temperatures). The activa-
3fϪh and 4fϪh by chromatography failed, so that the struc- tion parameters are given in Table 6 and 7. The differences
tures of these compounds were assigned by treatment of the in the enthalpies and entropies of activation for fluorine
product mixtures with alkali in liquid ammonia, the corres- substitution by the action of lithium phenoxide at the ortho
406
Eur. J. Org. Chem. 2001, 405Ϫ411

Reaction between 2,4-Difluoronitrobenzene and Lithium Phenoxides in Liquid Ammonia
FULL PAPER
1
Table 1. H and 19F NMR spectroscopic data for substituted aryl nitrophenyl ethers 3f, 3g, 6h, and 7f؊h
Table 2. IR data
Compound
IR spectra, ν˜ [cmϪ1
]
OH
CH (Ar)
CH (Alk)
CϪOϪC
NO2
Other bonds
3f
3030Ϫ3100
3030Ϫ3110
3030Ϫ3100
3030Ϫ3120
3030Ϫ3110
3050Ϫ3110
2930, 2870
2940, 2870
2960, 2865
2970, 2870
2970, 2865
2960, 2860
1270
1270
1270
1270
1270
1270
1520, 1335, 840, 740
1520, 1340, 830, 745
1510, 1340, 840, 745
1510, 1335, 840, 755
1520, 1330, 840, 750
1520, 1320, 830, 740
1075 (CϪF)
1080 (CϪF)
3g
6h
7f
3350 s.
3235 w.
3220 m.
3200 m.
7g
7h
and para positions of
1
(∆∆H϶
ϭ
Ϫ0.9 kJ/mol, Ϫ11.9 J ϫ molϪ1KϪ1).[15] This is in line with the previous
o/p
∆∆S϶ ϭ Ϫ1.1 J ϫ molϪ1KϪ1) are considerably smaller conclusion[14] of a minimal contribution to regioselectivity
o/p
than the corresponding values for the reaction of 1 with by chelation observed for lithium phenoxides as compared
sodium phenoxide (∆∆H϶ ϭ Ϫ4.4 kJ/mol, ∆∆S϶
ϭ
with other alkali metal cations.
o/p
o/p
Eur. J. Org. Chem. 2001, 405Ϫ411
407

L. Politanskaya, E. Malykhin, V. Shteingarts
FULL PAPER
Table 3. Exact values of molecular ion masses
Table 7. The differences between the ortho and para activation
parameters (∆∆H϶o/p and ∆∆S϶o/p) for the reaction of 2,4-difluoro-
nitrobenzene (1) with p-AlkC6H4OLi in liquid ammonia
Compound
Formula
Mϩ (found)
Mϩ (calcd.)
Alk
Me
Et
iPr
tBu
3f
C14H12FNO3
C15H14FNO3
C16H17NO4
C14H13NO4
C15H15NO4
C16H17NO4
261.0799
275.0954
287.1132
259.0840
273.1003
287.1150
261.0801
275.0958
287.1157
259.0844
273.1001
287.1157
3g
6h
7f
∆∆H϶
Ϫ1.7Ϯ0.2
Ϫ1.7Ϯ0.3
1.7Ϯ0.1
2.7Ϯ0.2
[a]
o/p
7g
7h
[kJ/mol]
∆∆S϶
Ϫ5.3Ϯ0.1
Ϫ5.2Ϯ0.1
7.4Ϯ0.1
12.0Ϯ0.1
[a]
[J/molo/ϫp K]
[a]
ln(o/p) ഠ ln(ko/kp) ϭ ln(Ao/Ap) Ϫ [(∆Eo Ϫ ∆Ep)/R] ϫ 1/T;
∆∆H϶ ϭ ∆∆E϶o/p; ∆∆S϶ ϭ Rln(Ao/Ap).
Table 4. o/p ratios in the reaction of 2,4-difluoronitrobenzene (1)
with XC6H4OLi (2a؊d) in liquid ammonia
o/p
o/p
The TS structure for the reaction under study may more
closely resemble either an anionic cyclohexadienyl interme-
diate (σ-complex) or the starting reactants. In the former
case, the ratio of ortho to para substitution rates in 1 under
enthalpic control should correlate with the relative stabilit-
ies of σ-complexes A and B (Figure 2). In the second case,
it is reasonable to apply reactivity indices associated with
the electronic structure of 1. Let us consider both possibil-
ities.
T [°C]
p-OMe
m-Me
H
p-F
Ϫ33
Ϫ55
1.12[a]
1.20
1.28
1.37
1.36
1.45
1.49
1.58
[a]
The deviations of o/p ratios do not exceed 0.02.
Table 5. o/p ratios in the reaction of 2,4-difluoronitrobenzene (1)
with p-AlkC6H4OLi (2e؊h) in liquid ammonia
T [°C]
Me
Et
iPr
tBu
Ϫ33
Ϫ55
1.23[a]
1.30
1.24
1.40
1.03
0.93
1.14
0.96
[a]
The deviations of o/p ratios do not exceed 0.02.
Figure 2. Structures of the intermediate anionic σ-complexes
Calorimetric measurements[23] indicate a higher stability
for anionic σ-complexes formed by 1-methoxy-2- and -4-
nitronaphthalenes and 2,4- and 2,6-dinitroanisoles with
MeONa in methanol when the nitro group is located para
rather than ortho to the geminal methoxy groups (in both
cases, ∆∆H0 ϭ 21 kcal/mol). Quantum chemical calcula-
o/p
tion[24] of the enthalpies of formation of the anionic inter-
mediates of SNAr reactions also indicates a slightly higher
stability for the structures with a nitro group located para
Figure 1. Temperature dependence of the regioselectivity in the re-
action of 1 with XC6H4OLi (p-Alk C6H4OLi)
It follows from these data that, firstly, the substitution of rather than ortho to the sp3-hybridized carbon atom. Thus,
the ortho-fluorine atom is more preferable in terms of en- according to ab initio calculations with the STO-3G basis
thalpy than that of the para-fluorine atom and, secondly, set, the difference between the energies of cyclohexadienyl
that this tendency weakens in the order: p-OMe Ͼ m-Me Ͼ anions with their nitro groups situated ortho and para to the
p-Me ഠ p-Et Ͼ H ഠ p-F (Table 6). As well as this, the sp3-hybridized carbon atom is as small as 2.09 cal/mol.[25]
enthalpic advantage moves from ortho to para substitution Semiempirical calculation of the enthalpies of formation of
on going along the alkyl substituent series from Me and Et the anionic σ-complexes derived from o- and p-fluoronitro-
to iPr and tBu (Table 7). Thus, the absolute values of benzenes and an amide ion gives the following values:
∆∆H϶o/p and ∆∆S϶o/p increase in the order: Me ഠ Et Ͻ iPr ∆∆H0 ϭ Ϫ1 kcal/mol (CNDO), Ϫ3 kcal/mol (INDO),
o/p
Ͻ tBu.
and 5 kcal/mol (MNDO).[24] According to the MNDO re-
Table 6. The differences between the ortho and para activation parameters (∆∆H϶o/p and ∆∆S϶o/p) for the reaction of 2,4-difluoronitroben-
zene (1) with XC6H4OLi in liquid ammonia
X
p-OMe
p-Me
p-Et
m-Me
H
p-F
[a]
∆∆H϶
Ϫ2.1Ϯ0.2
Ϫ1.7Ϯ0.2
Ϫ1.7Ϯ0.3
Ϫ1.4Ϯ0.3
Ϫ0.9Ϯ0.2
Ϫ0.9Ϯ0.1
o/p
[kJ/mol]
[J/molo/ϫp K]
[a]
∆∆S϶
Ϫ8.3Ϯ0.1
Ϫ5.3Ϯ0.1
Ϫ5.2Ϯ0.1
Ϫ3.8Ϯ0.1
Ϫ1.1Ϯ0.1
Ϫ0.5Ϯ0.1
[a]
ln(o/p) ഠ ln(ko/kp) ϭ ln(Ao/Ap) Ϫ [(∆Eo Ϫ ∆Ep)/R] ϫ 1/T; ∆∆H϶ ϭ ∆∆E϶o/p; ∆∆S϶ ϭ Rln(Ao/Ap).
o/p
o/p
408
Eur. J. Org. Chem. 2001, 405Ϫ411

Reaction between 2,4-Difluoronitrobenzene and Lithium Phenoxides in Liquid Ammonia
FULL PAPER
sults, which are considered the most reliable,[24] the σ-com- of the nitro group due to local interaction with solvent mo-
plex formed by nucleophile addition to the position para to lecules should not be considerable, the contribution of this
the nitro group is more stable.
factor is probably insignificant. On the other hand, mutual
In order to estimate the relative stabilities of isomeric σ- electrostatic repulsion between nitro group and fluorine
complexes forming only in the course of aryloxydefluorin- atom, as structural units bearing partial negative charges,
ation of 1, we calculated the formation enthalpies for the may be important. In any case, both their steric and electro-
structures A and B, with Ar ϭ Ph, using semiempirical static interactions favor structure A. However, on the whole,
PM3, AM1, and MNDO methods.[26] The initial para- we cannot judge with certainty the contribution of a par-
meters for the cyclohexadienyl fragments were obtained by ticular structural factor to the relative stability of the struc-
the geometric optimization of σ-complexes, formed by p- tures A and B, and the real stability difference between iso-
and o-fluoronitrobenzenes with an amide ion, in terms of meric σ-complexes seems not great, so their relative stability
MNDO approximation.[24] According to these calculations, might change on variation of structure. Nonetheless, in gen-
the most stable structures are those in which the mutual eral, the above data suggest the σ-complex-like TS approxi-
arrangement of bonds between the four-coordinated carbon mation to be more probably in favor of para substitution
and the fluorine atoms, and between the oxygen atom and than ortho substitution.
the phenyl ring, corresponds to an antiperiplanar con-
That is why one must also take into account the fact that
formation (cf. ref.[27]). The data presented in Table 8 show the experimentally observed enthalpic preference for ortho
the structure B to be preferred by a formation enthalpy substitution for X ϭ p-OMe, m-Me, H, p-F; p-Me, p-Et,
value of 2.11 to 2.91 kcal/mol, depending on the calculation and the dependence of ∆∆H϶ and ∆∆S϶ values on a
o/p
o/p
procedure. To estimate the effects exerted by the substitu- substituent in the nucleophile may originate from a struc-
ents in the phenyl ring on ∆∆H0o/p, we calculated the forma- tural similarity of the aryloxydefluorination TS to the start-
tion enthalpies for the structures A and B, using the PM3 ing compounds 1 and 2 rather than to the intermediate σ-
method. The data presented in Table 9 show no regular cor- complexes. In the context of this approximation, it is reas-
relation between the identity of the substituent and the onable to consider reactivity indices relating to the elec-
∆∆H0 value, the structure B being preferred in all cases.
tronic structure of the initial compound 1. Table 10 gives
the calculated charges on the carbon atoms that are elec-
trophilic centers in 1. As expected, the positive charge on
the carbon atom ortho to the nitro group exceeds that for
the para carbon atom that would provide the enthalpic pref-
erence for the nucleophilic attack, proceeding via the react-
ant-like TS, at the ortho position of 1. The occurrence of
such a TS is favored by the high reactivity of anionic nucle-
ophiles in liquid ammonia,[2,3] thanks to their lack of signi-
ficant specific solvation and association with a counter cat-
ion. It seems reasonable to assume that a fluorine atom at-
tached to the carbon atom attacked by a nucleophile, be-
cause of its strong electron-withdrawing inductive effect,
also acts in favor of the ‘‘early’’ TS. This is supported by
o/p
Table 8. Calculated enthalpies of formation (∆H0) of structure A
and B (Ar ϭ Ph)
Parameter
PM3
AM1
MNDO
∆H0 (A) [kcal/mol]
∆H0 (B) [kcal/mol]
Ϫ139.16
Ϫ142.07
2.91
Ϫ120.33
Ϫ123.10
2.77
Ϫ111.99
Ϫ114.10
2.11
∆∆H0 [kcal/mol]
o/p
Table 9. PM3 enthalpies of formation (∆H0) of structures A and B
with various substituents in the aryl fragment
Parameter
p-OMe p-Me
m-Me
H
p-F
the inversion of the signs of both ∆∆H϶ and ∆∆S϶
o/p
∆H0 (A) [kcal/mol] Ϫ176.67 Ϫ148.24 Ϫ148.27 Ϫ139.16 Ϫ184.86
∆H0 (B) [kcal/mol] Ϫ179.61 Ϫ151.16 Ϫ151.27 Ϫ142.07 Ϫ187.51
o/p
(vide supra), as well as by the substantial increase in ∆H϶
value for halogen substitution reactions of p- and o-hal-
ogenonitrobenzenes with charged nucleophiles, on going
from fluorine to other halogens, while the ∆S϶ values re-
main almost unchanged.[3,5Ϫ7]
∆∆H0 [kcal/mol] 2.94
2.82
3.00
2.91
2.65
o/p
Steric hindrance, arising from the phenoxy group, to a
coplanar arrangement of the nitro group and pentadienyl
fragment in the structure A could be a factor responsible
for its smaller relative stability. However, according to crys-
tallographic data,[28] the nitro groups in the σ-complex of
2,4,6-trinitrophenetole with potassium ethoxide are copla-
Table 10. Calculated charges [a. u.] on the carbon atoms bearing
fluorine atoms in positions ortho and para to the nitro group in
compound 1
Atom
PM3
AM1
ortho
ϩ0.191
ϩ0.151
0.040
ϩ0.196
ϩ0.167
0.029
nar with the ring. Therefore, the above hindrance can para
scarcely be significant. Another factor of the same nature
∆qo/p
lies in the fact that a fluorine atom, situated adjacent to the
nitro group in structure B, can hinder coplanarity of the
Owing to the above-mentioned weak solvation of anions
nitro group with the pentadienyl moiety. However, taking by liquid ammonia, it is not unreasonable to expect the nu-
into account that the effective radius of fluorine is small cleophilicity of substituted phenoxides in liquid ammonia
and the solvating ability of liquid ammonia toward anions to change in parallel with their gas phase basicity: i.e., in
is weak,[11] and hence that an increase in the effective size the order p-F Ͻ H Ͻ m-Me Ͻ p-Me Ͻ p-OMe.[29] The
Eur. J. Org. Chem. 2001, 405Ϫ411
409

L. Politanskaya, E. Malykhin, V. Shteingarts
FULL PAPER
increase in nucleophile reactivity makes the TS more react- in distinctly protic, particularly hydroxylic, solvents on the
ant-like, so that the charge density distribution in 1 would other.[5Ϫ8] Specifically, as a result of conferring charged nu-
be an increasingly significant factor in the difference be- cleophiles with considerably higher activity, TSs are ex-
tween the activation parameters for ortho- and para-arylox- pected to be generally more reactant-like in solvents of the
ydefluorination, thus enhancing the enthalpic preference for first type than in solvents of the second. As a consequence,
ortho substitution. The change of ∆∆H϶ from Ϫ0.9 to any decrease in nucleophile activity due to structural
o/p
Ϫ2.1 kJ/mol on going from X ϭ p-F to X ϭ p-OMe is reasons may change the correlation of activation para-
in line with this prediction. Conversely, the increase in the meters of competing reactions and regioselectivity, in com-
electron-accepting capability of a substituent in the phenox- pliance with the strengthening intermediate-like character
ide is expected to diminish the enthalpic advantage of ortho of the TS.
substitution, even as far as reversing the o/p ratio, as indeed
occurs, for example, on going from sodium methoxide to
Experimental Section
sodium phenoxide in their respective reactions with o- and
p-fluoronitrobenzenes in methyl alcohol (∆∆H϶ ϭ Ϫ0.5
General Methods and Materials: The 19F NMR spectra were re-
corded with a Bruker WP-200 SY instrument in 50% Et2O and 5%
[D6]acetone solutions, using C6F6 as an internal standard. The H
o/p
kcal/mol for MeOϪ and ϩ3.3 kcal/mol for PhOϪ).[5]
It is known[29] that p-alkylphenoxides display a variation
1
of gas phase acidity in the order Me Ͻ Et Ͻ iPr Ͻ H Ͻ
tBu,[29] thus illustrating the weakening of the electron-do-
nating influence of the alkyl substituent and its trend to-
wards a change to an electron-withdrawing one with the
branching at the alkyl α-carbon atom; i.e. the increase in
the alkyl polarizability is unfavorable for phenoxide basi-
city. Accordingly, if one assumes for the above reasons that
alkylphenoxide nucleophilicity in liquid ammonia should
follow the same sequence, the inversion of ∆∆H϶o/p, con-
comitant with passing from p-Me and p-Et to p-iPr and p-
tBu (Table 7), would suggest the growth of the interme-
diate-like character of TS, to the detriment of its reactant-
like character, as caused by weakening phenoxide nucleo-
NMR spectra were recorded with a Bruker WP-200 SY instrument
in 5% [D6]acetone and CD2Cl2 solutions, using [(CH3)3Si]2O as an
internal standard. Ϫ IR spectra were recorded with a UR-20 instru-
ment in KBr tablets (0.25% by weight). Ϫ Exact values of molecu-
lar ion masses were measured by high-resolution mass spectrometry
with a Finnigan MAT-8200 machine. 2,4-Difluoronitrobenzene (1)
was prepared according to the literature procedure.[30] Phenol, m-
cresol, p-cresol, p-ethylphenol, p-fluorophenol, p-methoxyphenol,
p-isopropylphenol, and p-tert-butylphenol were commercially avail-
able. Liquid ammonia was purified by dissolving metallic sodium
in it, with subsequent distillation into the reaction flask at Ϫ40 °C.
Metallic lithium and potassium were purified from their oxide films
and weighed directly before loading into the reaction flask.
General Procedure for the Preparation of Lithium Phenoxides in Li-
quid Ammonia: To liquid ammonia (75 mL) at Ϫ40 °C was added
lithium metal (0.02 g), with stirring. FeCl3·6H2O (ca. 0.005 g) was
added to the obtained dark blue solution, which was stirred until
lithium amide formation was complete, concomitant with decolor-
ization. To generate the lithium phenoxide, one equivalent (relative
to lithium amide) of phenol was added (concentration 0.04 mol/l).
philicity. Such a view is consistent with the ∆∆H϶ values
o/p
for the reactions of 1 with PhOLi and p-FC6H4OLi (ca.
Ϫ0.9 kJ/mol), p-iPrC6H4OLi (ϩ1.7 kJ/mol), and p-tBu-
C6H4OLi (ϩ2.7 kJ/mol). Thus, the p-iPr and p-tBu sub-
stituents manifest as electron-withdrawing ones, dimin-
ishing phenoxide nucleophilicity in the reaction in question:
this is similar to their influence on the phenol dissociation
equilibrium in the gas phase. If so, for the reactions between
1 and 2aϪf, the o/p ratios are apparently determined by
factors inherent in the ‘‘early’’ TS, first of all by the charge
density distribution in 1, which is characterized by the
greater proportion of positive charge in the position ortho
to the nitro group, as compared to the para position. De-
creasing nucleophilicity on passing to 2g and 2h makes the
TS more closely similar to the intermediate σ-complexes,
and heightens the influence of those factors inherent in the
‘‘late’’ TS. It is suggested that this is the reason for the in-
version of the correlation of TS energies in favor of para
substitution.
General Procedure for the Interaction of 2,4-Difluoronitrobenzene
(1) with Lithium Phenoxide in Liquid Ammonia: To a thermostatted
(Ϯ0.5 °C) solution of lithium phenoxide, prepared as described
above, was added 1 (1 g). It was stirred for 10 and 20 min for X ϭ
p-OMe, m-Me, p-Me, and p-Et; and for 20 and 30 min for X ϭ p-F,
H, p-iPr, and p-tBu, at Ϫ33 and Ϫ55 °C, respectively. To determine
temperature dependencies, not less than 6 experiments were carried
out for each compound at fixed temperatures (error Ϯ0.5 °C), with
a step of ca. 5 °C. The reaction mixture was poured into a stirred
suspension of ammonium chloride (ca. 2 g) in diethyl ether
(100 mL) cooled to Ϫ50 °C. After the ammonia had evaporated,
the mixture was diluted with water (50 mL) and extracted with di-
ethyl ether (2 ϫ 50 mL). The combined ether extracts were washed
with 5% aqueous NaOH solution (2 ϫ 50 mL), then with water
(50 mL), and dried with MgSO4. The mixture of reaction products
obtained after solvent evaporation (1.1Ϫ1.2 g for X ϭ H, p-F, m-
Me, p-Me; and 1.2Ϫ1.3 g for X ϭ p-OMe, p-Et, p-iPr, p-tBu), was
analyzed by 19F NMR.
Conclusion
The results given in this paper, together with other General Procedure for the Separation of 3 and 4 from other Reaction
data,[3,15,17] are believed to explain the significant difference
in the detailed mechanisms of fluorine nucleophilic substi-
tution in positions ortho and para to the nitro group by the
Products: A mixture of 1, 3, and 4 was purified by column chroma-
tography (silica gel 40Ϫ140 µ, hexane), followed by evaporation of
1 in vacuum (5 Torr with 70 °C bath), to give a mixture of 3 and 4.
action of anionic nucleophiles in liquid ammonia, as well
as, probably, in aprotic bipolar solvents on one hand, and Potassium Hydroxide in Liquid Ammonia: To a stirred solution of
General Procedure for the Interaction of a Mixture of 3 and 4 with
410
Eur. J. Org. Chem. 2001, 405Ϫ411

Reaction between 2,4-Difluoronitrobenzene and Lithium Phenoxides in Liquid Ammonia
FULL PAPER
[9]
K. R. Plowman, J. J. Lagowski, J. Phys. Chem. 1974, 78,
potassium metal in liquid ammonia (concentration 0.2 mol/l) at
Ϫ40 °C, containing ferric chloride (ca. 0.005 g), was added drop-
wise an equimolar amount of water. After formation of a potas-
sium hydroxide suspension, a mixture of 3 and 4 (concentration 0.1
mol/l) was added, and the reaction mixture was stirred at Ϫ35 °C
for 4 h. The reaction was quenched by pouring the mixture into a
stirred suspension of ammonium chloride (ca. 5 g) in diethyl ether
(150 mL) cooled to Ϫ50 °C. After the ammonia had evaporated,
the mixture was washed with 20% aqueous KOH solution (2 ϫ
100 mL), then with water (50 mL), and dried with MgSO4. Accord-
ing to 19F NMR spectroscopic data, the corresponding mixtures of
3 and 4 contained mainly isomers 3f and 3g, which were isolated
by TLC [Silufol UV-254 plates, hexane/diethyl ether (9:1), Rf ϭ
0.9Ϫ0.7]. The combined alkaline, aqueous solution was neutralized
with sulfuric acid (to pH ϭ 6) and extracted with diethyl ether (2
ϫ 100 mL). The ether extract was washed with water and then
dried with MgSO4. After solvent evaporation, 6 and 7 were isolated
by TLC (Silufol UV-254 plates, benzene, Rf ϭ 0.9Ϫ0.7 and
0.5Ϫ0.15, respectively).
143Ϫ148.
[10]
[11]
A. T. Lemley, J. J. Lagowski, J. Phys. Chem. 1974, 78, 708Ϫ713.
N. A. Izmailov, Dokl. Akad. Nauk SSSR 1963, 149,
1364Ϫ1367.
[12] [12a]
M. Herlem, A. I. Popov, J. Am. Chem. Soc. 1972, 94,
[12b]
1431Ϫ1434. Ϫ
W. R. Davidson, P. Kebarle, J. Am. Chem.
[12c]
Soc. 1976, 98, 6133Ϫ6138. Ϫ
B. J. Herold, M. Celina, R.
L. R. Lazana, H. M. Novais, Tetrahedron 1977, 33, 517Ϫ521.
M. Kimura, S. Sekiguchi, K. Matsui, Koguo Kagaku Zasshi
1970, 73, 513Ϫ516.
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
L. V. Politanskaya, E. V. Ryabizkaya, E. V. Malykhin, V. D.
Shteingarts, Zh. Org. Khim. 2000, 36, 835Ϫ841.
T. A. Kizner, V. D. Shteingarts, Zh. Org. Khim. 1985, 21,
2376Ϫ2382.
L. V. Politanskaya, E. V. Malykhin, V. D. Shteingarts, Zh. Org.
Khim. 1997, 33, 703Ϫ710.
L. V. Politanskaya, E. V. Malykhin, V. D. Shteingarts, Zh. Org.
Khim. 1996, 32, 1174Ϫ1185.
R. Bolton, J. P. B. Sandall, J. Chem. Soc., Perkin Trans. 2
1978, 141Ϫ144.
G. V. L. Bevan, A. F. Foley, J. Hirst, W. O. Uwamu, J. Chem.
Soc. B. 1970, 794Ϫ797.
G. A. Selivanova, T. V. Chuikova, A. A. Shtark, V. D. Shtein-
garts, Zh. Org. Khim. 1988, 24, 2513Ϫ2518.
The Sadtler Standard Spectra. Infrared Grating Spectra, Sadtler
Research Laboratories, Philadelphia, 1968؊1978.
R. Silverstein, Spectroscopic identification of organic com-
pounds, Mir, Moscow, 1977, pp. 590.
The spectral characteristics of new compounds 3f, 3g, 6h, and 7fϪh
are presented in Table 1, Table 2 and Table 3.
[23] [23a] Y. H. Fendler, W. L. Hinze, L.-J. Liu, J. Chem. Soc., Perkin
Trans. 2 1975, 1768Ϫ1775. Ϫ [23b] R. M. Murphy, C. A. Wulff,
M. J. Strauss, J. Am. Chem. Soc. 1974, 96, 2678Ϫ2682.
[1]
H. Smith, Organic Reactions in Liquid Ammonia, John Wiley &
Sons, New York, 1963, pp. 343.
[2]
A. A. Shtark, T. A. Kizner, V. D. Shteingarts, Zh. Org. Khim.
1982, 18, 2321Ϫ2326.
T. A. Kizner, V. D. Shteingarts, Zh. Org. Khim. 1984, 20,
1089Ϫ1096.
J. F. Bunnet, R. F. Snipes, J. Am. Chem. Soc. 1955, 77,
5422Ϫ5433.
T. O. Bamkole, J. Hirst, E. I. Udoessien, J. Chem. Soc., Perkin
Trans. 2 1973, 2114Ϫ2119.
B. A. Bolton, J. Miller, V. A. Williams, J. Chem. Soc. 1955,
2926Ϫ2929.
T. O. Bamkole, J. Hirst, E. I. Udoessien, J. Chem. Soc., Perkin
Trans. 2 1973, 110Ϫ114.
[24]
N. S. Nudelman, P. MacCormack, Tetrahedron 1984, 40,
4227Ϫ4235.
[3]
[25]
A. J. Birch, A. L. Hinde, L. Radom, J. Am. Chem. Soc. 1980,
102, 6430Ϫ6437.
[4]
[26]
A. A. Bliznyk, A. A. Voityk, Zh. Str. Khim. 1986, 27, 190Ϫ191.
[27]
E. Buncel, R. M. Tarkka, J. M. Dust, Can. J. Chem. 1994,
[5]
72, 1709Ϫ1721.
[28]
R. Destro, C. M. Gramaccioli, M. Simonetta, Acta Crys-
[6]
tallogr., B 1968, 24, 1369Ϫ1386.
[29]
R. W. Taft, R. D. Topsom, Prog. Phys. Org. Chem. 1987, 16,
[7]
1Ϫ83.
[30]
F. Swarts, Recl. Trav. Chim. Pays-Bas 1915, 35, 154Ϫ163.
[8]
F. Del Cima, G. Biggi, F. Pietra, J. Chem. Soc., Perkin Trans.
Received July 3, 2000
[O00327]
2 1973, 55Ϫ58.
Eur. J. Org. Chem. 2001, 405Ϫ411
411