Mechanisms of reactions of halogenated compounds. Part 7. Effects of fluorine and other groups as substituents on nucleophilic aromatic substitution

Article abstract of DOI:10.1016/j.jfluchem.2008.04.009

The effect of fluorine as a substituent group on nucleophilic aromatic substitution is discussed, where a fluorine atom located ortho to the point of substitution may be of variable activating influence, whereas fluorine located para is slightly deactivat

Full text of DOI:10.1016/j.jfluchem.2008.04.009

Journal of Fluorine Chemistry 129 (2008) 998–1002
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Mechanisms of reactions of halogenated compounds
Part 7. Effects of fluorine and other groups as substituents
on nucleophilic aromatic substitution
*
**
Richard D. Chambers , Peter A. Martin, Graham Sandford , D. Lyn H. Williams
University of Durham, Department of Chemistry, South Road, Durham DH1 3LE, UK
A R T I C L E I N F O
A B S T R A C T
Article history:
The effect of fluorine as a substituent group on nucleophilic aromatic substitution is discussed, where a
fluorine atom located ortho to the point of substitution may be of variable activating influence, whereas
fluorine located para is slightly deactivating and meta is activating. A rationale of these effects is
presented and evidence to support polar influences by ortho fluorine is advanced. The influence of CN,
CF3, CF2H and CFH2 is also established by comparison of appropriate measured rate constants and
compared with the activation effects of ring nitrogen.
Received 21 March 2008
Received in revised form 25 April 2008
Accepted 25 April 2008
Available online 2 May 2008
Dedicated to the memory of our good friend
and colleague Professor Lyn Williams.
ß 2008 Elsevier B.V. All rights reserved.
Keywords:
Nucleophilic aromatic substitution
Perfluoroheteroaromatic
Perfluoroaromatic
Activation effect
1. Introduction
to the point of nucleophilic attack, on a variety of perfluorinated
systems, including benzenoid [1,6], pyridine [5], and bicyclic [7,8]
Polyfluorinated aromatic systems have an interesting ‘mirror-
image’ relationship with the corresponding hydrocarbon com-
pounds, in that the former generally react with nucleophiles, via
Meisenheimer complexes [2–5], in contrast to the classical
chemistry of aromatic hydrocarbons, that undergo electrophilic
substitution, via Wheland intermediates as shown in Fig. 1.
To develop the chemistry of these systems, it is important to
understand at least the gross factors that determine the reactivity
and orientation of nucleophilic attack in various fluorinated
derivatives. We have noted a recent publication [9], relating to
calculations that are directed to the same problem but has
concentrated solely on the thermodynamic stability of the
corresponding Meisenheimer complexes. This assumes that the
Meisenheimer complex is a good model for the transition state in
each case, and that coulombic effects are not important. In this
paper, we wish to further demonstrate that neither of these
assumptions is valid.
We have previously established [1,2,6] that a fluorine atom that
is situated para- to the point of nucleophilic attack (Fig. 2) is slightly
deactivating, whereas fluorine atoms that are situated at corre-
sponding meta- and ortho- positions are activating at both sites. In
order to rationalize these findings we used our knowledge of the
effect of fluorine as a substituent on carbanion stabilities [5,10].
As indicated in Fig. 3, the stabilizing inductive effect of fluorine
on the stability of an attached carbanion is offset by electron-pair
repulsions and the overall effect is destabilizing for a planar
carbanion, where the repulsion is maximized. Of course, this is the
situation for a developing carbanionic transition state during
nucleophilic aromatic substitution and it is understandable,
therefore, that a para-F substituent is slightly deactivating. In an
analogous way, a fluorine atom attached to carbon that is adjacent
to a carbanionic site in the transition state is strongly stabilizing
aromatic compounds. In this connection,
emerged early in the study of these compounds; nucleophilic
attack occurs predominantly para- to the substituent in
a curious feature
X
pentafluorobenzene derivatives, Fig. 1, regardless of whether the
substituent is electron-withdrawing (activating towards nucleo-
philic aromatic substitution), hydrogen, or electron-donating
(deactivating). Consequently, we endeavoured to determine, by
experiment, the separate activating influences of fluorine, as a
substituent in a ring, at sites that are ortho-, meta- and para-, Fig. 2,
* Corresponding author. Tel.: +44 191 334 2020; fax: +44 191 384 4737.
** Corresponding author. Tel.: +44 191 334 2039; fax: +44 191 384 4737.
E-mail addresses: R.D.Chambers@durham.ac.uk (R.D. Chambers), Graham.Sand-
ford@durham.ac.uk (G. Sandford).
0022-1139/$ – see front matter ß 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2008.04.009

R.D. Chambers et al. / Journal of Fluorine Chemistry 129 (2008) 998–1002
999
Fig. 4. Ortho-activation.
us a short step to conclude that the activating effect of an ortho-
fluorine is due to enhancing the electrophilic character of
the carbon atom that is under attack. Of course, such effects fall
off rapidly with distance and it is unlikely that this is a
significant factor contributing to activation by a meta-fluorine
substituent.
Fig. 1. ‘Mirror-image’ chemistry.
This model is open to test, and we have previously done so
[12]: according to the Hammond postulate, as the transition
state moves closer to the initial-state, that is, if we use a more
reactive system, then the ortho- (coulombic) effect should
become more important, which it does and, conversely, as we
make the system less reactive, then the transition state becomes
more like the Meisenheimer complex and the ortho-activation
should become less important. This is indeed the situation that
is illustrated in Table 1, where we have compared reactions of
methoxide in methanol with systems containing F or H at the
corresponding positions. As the system is made much more
reactive, comparing benzenoid to the corresponding pyridine
systems, then the effect of an ortho-F becomes greater than that
of a meta-F in the pyridine series. In contrast, however, when the
much less reactive aniline is used as the nucleophile, then the
Fig. 2. Activating influences of fluorine.
because electron-pair repulsions are negligible; this is the
situation for the effect of a meta-F substituent on the developing
negative charge in the transition state for nucleophilic aromatic
substitution and it is understandable that a fluorine atom at this
position is strongly activating.
Intuitively, we might anticipate that a fluorine substituent
located at a site that is ortho- to the position of nucleophilic
attack would have an effect that is similar to that of a para-F; but
this is not the case. Fluorine substituents at the ortho- position
are usually activating and may be more activating than at the
corresponding meta- site, depending on the system. Our
Table 1
Relative rates of reaction (kF/kH) of systems bearing fluorine or hydrogen at
rationale of this situation is outlined in Fig. 4:
a factor
corresponding positions [5,6]
contributing to the reactivity of carbon–fluorine bonds in
unsaturated systems is the polarity of the bond. That is, the
approach of a nucleophile is aided by ion–dipole interaction, i.e.
a coulombic effect, and such effects have been shown previously
to change completely the orientation of nucleophilic substitu-
tion in some perfluorinated aromatic systems [11]. It seems to
Nucleophile
Substrates
kF/kH
ortho
meta
para
MeONa, MeOH, 58 8C
MeONa, MeOH, 58 8C
PhNH2, dioxane/H2O, 58 8C
Benzene systems
Pyridine systems
Pyridine systems
57
79
1.8
106
30
0.43
0.33
–
Fig. 3. Effect of fluorine as a substituent meta and para to the site of nucleophilic
34
aromatic substitution.

1000
R.D. Chambers et al. / Journal of Fluorine Chemistry 129 (2008) 998–1002
Table 2
Effects of ring nitrogen and trifluoromethyl
Effect of ring nitrogen: at various sites (derived from the ratios of k values for reactions with ammonia in dioxane) (60:40 v:v, at 25 8C) [13]
ortho-
meta-
para- (relative to the site of nucleophilic attack)
0.62 ꢀ 105
8.5 ꢀ 102
2.3 ꢀ 105 (relative to C–H)
[72
1
266] (relative to the meta- position)
Effect of trifluoromethyl: at various sites determined from pyridine derivatives [13] (derived from the ratios of k values for reactions with ammonia in dioxane) (60:40 v:v, at
25 8C)
ortho-
2.2 ꢀ 103
meta-
para- (relative to the site of nucleophilic attack)
4.5 ꢀ 103 (relative to C–H)
–
activating effect of an ortho-F virtually disappears. Thus, our
rationale, which incorporates coulombic effects, is quite con-
sistent with our findings, rather than the use of Meisenheimer
models only [9] for the transition state which, as we have
illustrated, will not account for the variable ortho-activation by
fluorine.
Development of these ideas has led us to determine the
separate activating influences of substituents (e.g. CF3, CN, etc.) at
sites that are ortho-, meta-, and para- to the site of nucleophilic
attack, to further compare other significant factors that govern the
orientation of nucleophilic aromatic substitution in these systems.
In this paper, we report our studies that explore the effects of CF3,
CF2H, CFH2, Cl and CN substituents on nucleophilic aromatic
substitution processes.
adjacent planar carbanion, is conformationally dependant. In
this way, introduction of the first fluorine would have
a
significant effect, and then the second would increase the
probability of attaining the favourable conformation. However,
introducing the third fluorine would not only increase the
probability further, but more importantly reduces the possibi-
lities for alternative conformations. A possible scenario would
be that the favourable conformation is as shown in Fig. 5, with a
C–F bond in plane with the
p-bond containing the developing
charge in the transition state, and an obvious way of accounting
for this stabilizing effect would be to invoke negative
hyperconjugation [5] as illustrated in Fig. 5. The possibilities
for attaining this favourable conformation are then maximized
in the system, X = CF3.
2. Results and discussion
2.2. Effect of cyano substituents
2.1. Effect of CF3, CF2H and CFH2 substituents
We have also explored the activating influence of the CN group
at sites that are ortho-, meta- and para- to the site of nucleophilic
attack, by comparing appropriate measured rate constants that
then enable us to make the comparisons with the activating effect
of the trifluoromethyl group shown in Table 4.
We have previously established the activating effects of
trifluoromethyl in perfluoroaromatic systems and compared these
effects with those of ring nitrogen (Table 2) [13].
Thus, trifluoromethyl is strongly activating but not as activating
as ring nitrogen at the corresponding position. Moreover,
trifluoromethyl, like ring nitrogen, activates both ortho- and para-
positions to a similar degree. Subsequently, we have explored
further the nature of the activating influence of fluorine in the side-
chain by comparing the relative effects of CF3, CF2H, and CFH2, in
reactions of C6F5–X (where X is the substituent group) with
ammonia in dioxane/water (60/40), the product in each case being
the corresponding 4-amino derivative as determined by 19F NMR.
The results are shown in Table 3 and they are very surprising,
because the difference in reactivity between CF2H and CFH2 are
smaller than we would have anticipated.
Clearly, CN in the para- position is significantly more activating
than trifluoromethyl, while introduction of a second CN group
into the meta- position has a relatively small effect. In contrast,
two CN groups para- to each other are little more activating than
two trifluoromethyl groups that are para- to each other. This
suggests that CN is significantly more activating in the para- than
in the ortho- positions and that CN is much less activating in the
meta- positions. The comparisons are also consistent with the
knowledge that the CN group is activating via conjugative
stabilization of the developing negative charge in the transition
state.
In further reactions to illustrate these conclusions, using
aniline as the reagent (Fig. 6), we have established that 4-
chlorotetrafluoropyridine 1 gave product 2 arising from sub-
stitution at the 2-position exclusively, replacing fluorine, rather
These results suggest that at least part of the stabilizing
influence of fluorine in the substituent groups X, on the forming
Table 3
Rate constants for reactions of polyfluoroalkylbenzene derivatives, C6F5–X, with
ammonia in dioxane/water, 60:40 v:v, at 80 8C
X
Rate constant, k (molꢁ1 sꢁ1
)
CF3
71 ꢀ 10ꢁ1
5.0 ꢀ 10ꢁ5
1.1 ꢀ 10ꢁ5
CF2H
CFH2
Fig. 5. Conformational dependence of C–F as a stabilizing influence on an adjacent
developing carbanion allowing stabilization by negative hyperconjugation.

R.D. Chambers et al. / Journal of Fluorine Chemistry 129 (2008) 998–1002
1001
Table 4
Rate constants (1 molꢁ1 sꢁ1) for reactions of cyano- and trifluoromethyl-benzene
derivatives, with ammonia in dioxane/water, 60:40 v:v, at 80 8C
Fig. 6. Orientation of substitution of reactions with aniline.
because this maximizes the number of activating fluorine
substituents (two ortho-fluorine and two meta-fluorine atoms
to the site of attack).
3. Conclusions
than chlorine which is attached to the electronically favoured 4-
position. 3,5-Dichloro-trifluoropyridine 3 is unusual in that
some attack at the 2-position has been reported previously and
attributed to steric effects arising from the chlorine atoms which
retard attack at the 4-position. In the case of attack by aniline
attack at the 4-position occurred preferentially to give 4a (92%),
whereas in 3-chloro-5-cyano-trifluoropyridine 5, products 6a
and 6b were formed in a 3:1 ratio. 6b arises from attack at the 2-
position and we may attribute this observation to the superior
activating influence of the 5-cyano group, rather than steric
effects.
The gross features of nucleophilic aromatic substitution
processes in perfluoroaromatic compounds are now reasonably
well understood and the activating effects of various ring
substituents are illustrated in the discussion above. This reaffirms
the conclusion that a consideration of the relative stabilities of
Meisenheimer intermediates alone is insufficient to explain
experimentally determined results.
4. Experimental
Furthermore, in the case of pentafluoropyridine 9 (Fig. 6),
the ring nitrogen has the major influence in determining
overall reactivity but is not significantly discriminating between
2- or 4-activation. This is clear from the fact that 2,3,4,6-
tetrafluoropyridine 11 gives a mixture of 2- and 4-substitution
products because the number of activating fluorine atoms is the
same for attack at both sites in this case. However, attack occurs
exclusively at the 4-position in pentafluoropyridine simply
4.1. General
Except where indicated, all of the compounds described in this
paper have been described previously and were synthesised by
known methods [3]. Also, reactions of these systems with
ammonia have been studied extensively [3] and, therefore, product
isomer analysis was carried out by NMR and by gas chromato-
graphy, using a gas-density balance detector.

1002
R.D. Chambers et al. / Journal of Fluorine Chemistry 129 (2008) 998–1002
4.2. Kinetic experiments
ꢁ65.8 (m, F-2); 5-chloro-2,4-difluoro-6-(phenylamino)nicotinoni-
trile 6b; dF ꢁ58.7 (m, F-2), ꢁ94.5 (m, F-4).
Rate measurements for reactions with ammonia were carried
out at 25 8C using dioxane:water, 60:40 (v:v). Samples were
withdrawn at suitable intervals and quenched in a large excess of
water, and the remaining base was titrated against standard
hydrochloric acid. The infinity value agreed well with that
calculated from the weight of material used. Generally, reactions
were followed for at least two half-lives and each run was carried
out in duplicate. Second-order rate constants were calculated [6]
from Eq. (i), where a and b are the initial concentrations of reagent
(e.g. ammonia) and substrate respectively, since the reagent
becomes protonated by the acid liberated in the reaction.
4.3.4. 3-Cyanotetrafluoropyridine 7
A mixture containing aniline (0.93 g, 10 mmol) and 3-cyanote-
trafluoropyridine 7 (0.88 g, 5 mmol) in dioxane (6 mL) was heated
in a sealed tube at 80 8C for 8 days. The NMR spectrum of the crude
mixture showed the presence of both the 4-anilino 8a and the 6-
anilino 8b derivatives in the ratio 7:3. No further purification was
carried out.
4.3.5. Pentafluoropyridine 9
A mixture containing aniline (1.03 g, 11 mmol) and penta-
fluoropyridine 9 (0.93 g, 5.5 mmol) (2:1 molar ratio) in dioxane
(6 mL) was heated in a sealed tube at 80 8C for 22 h. The tube was
cooled, opened to air, and the contents were extracted by dimethyl
ether; the combined extracts were washed with dilute HCl (1 M),
dried (MgSO4) and filtered. Removal of the dimethyl ether gave a
brown oil and HPLC showed one product, which was purified by
column chromatography (silica), using pentane/chloroform (50:50
v:v), to give 4-anilino-2,3,5,6-tetrafluoropyridine 10 (1.17 g, 88%) as
a white solid; m.p. 87 8C (Found: C, 54.8; H, 2.4; N, 11.9. C11H6F4N2
requires: C, 54.6; H, 2.5; N, 11.6%); dF ꢁ95.3 (2 F, m, F-2), ꢁ155.75
(2 F, m, F-3). The experiment was repeated, using a reaction
temperature of 25 8C, giving the same product in essentially
quantitative yield.
ln½bða ꢁ 2xÞ=aðb ꢁ xÞꢂ
k ¼
(i)
ða ꢁ 2bÞt
The standard error of the mean in any individual run was ꢃ1%
and duplicate runs agreed within ꢃ1%.
Rate measurements for reactions with aniline were carried out
by measuring fluoride ion concentrations using a fluoride ion
specific electrode, following calibration.
4.3. Reactions with aniline
4.3.1. 4-Chloro-2,3,5,6-tetrafluoropyridine 1
A mixture consisting of aniline (0.93 g, 10 mmol), 4-chloro-
2,3,5,6-tetrafluoropyridine 1 (0.93 g, 5 mmol) and dioxane (6 mL)
was heated in a sealed tube for 3 days at 80 8C. Work up and
purification as above gave 4-chloro-3,5,6-trifluoro-N-phenylpyridin-
2-amine 2 (1.02 g, 79%) as a white solid; (Found: C, 50.9; H, 2.2; N,
10.7. C11H6ClF3N2 requires: C, 50.98; H, 2.34; N, 10.83%); dF ꢁ94.2
(1 F, m, F-6), ꢁ143.7 (1 F, m, F-3), ꢁ158.5 (1 F, m, F-5).
4.3.6. 2,3,4,6-Tetrafluoropyridine 11
A mixture consisting of aniline (0.93 g, 10 mmol), 2,3,4,6-
tetrafluoropyridine 11 (0.76 g, 5 mmol) and dioxane (6 mL) was
stirred for 3 days at 25 8C. The mixture was worked up as described
above, giving a brown solid which was shown by 19F NMR and
mass spectrometry to contain 2,3,6-trifluoro-N-phenylpyridin-2-
amine 12a; dF ꢁ75.5 (1 F, m, F-2), ꢁ96.0 (1 F, m, F-6), ꢁ171.6 (1 F,
m, F-3); and, 3,4,6-trifluoro-N-phenylpyridin-2-amine 12b; dF
ꢁ71.6 (1 F, m, F-6), ꢁ126.3 (1 F, m, F-4), 170.9 (1 F, m, F-3); and, in a
3:1 ratio by 19F NMR.
4.3.2. 3,5-Dichloro-2,4,6-trifluoropyridine 3
A mixture consisting of aniline (0.93 g, 10 mmol), 3,5-dichloro-
2,4,6-trifluoropyridine 3 (1.1 g, 5 mmol) and dioxane (6 mL) was
heated in a sealed tube for 8 days at 80 8C. Work up gave a mixture
of 4-anilino- (4a, 92%) and 2-anilino-3,5- (4b, 8%) dichlorodi-
fluoropyridine which, after purification by column chromatogra-
phy (pentane:chloroform, 1:1 v:v), gave 3,5-dichloro-2,6-difluoro-
N-phenylpyridin-4-amine 4a (1.3 g, 92%) as a white solid; m.p.
73 8C, (Found: C, 48.3; H, 2.1; N, 10.2. C11H6Cl2F2N2 requires: C,
48.03; H, 2.20; N, 10.18%); dF ꢁ75.2 (m, F-2).
References
[1] For Part 6, see R.D. Chambers, M.J. Seabury, D.L.H. Williams, N. Hughes, J. Chem.
Soc., Perkin Trans. I (1988) 255–257.
[2] R.D. Chambers, Dyes and Pigments 3 (1982) 183–190.
[3] G.M. Brooke, J. Fluorine Chem. 86 (1997) 1–76.
[4] R.D. Chambers, C.R. Sargent, Adv. Heterocycl. Chem. 28 (1981) 1–71.
[5] R.D. Chambers, Fluorine in Organic Chemistry, Blackwell, Oxford, 2004.
[6] R.D. Chambers, J.S. Waterhouse, D.L.H. Williams, J. Chem. Soc., Perkin Trans. I
(1977) 585–588.
4.3.3. 3-Chloro-5-cyano-2,4,6-trifluoropyridine 5
[7] G.M. Brooke, R.D. Chambers, C.J. Drury, M.J. Bower, J. Chem. Soc., Perkin Trans. I
(1993) 2201–2209.
[8] R.D. Chambers, M.J. Seabury, D.L.H. Williams, N. Hughes, J. Chem. Soc., Perkin
Trans. I (1988) 251–254.
[9] M. Muir, J. Baker, J. Fluorine Chem. 126 (2005) 727–738.
[10] R.D. Chambers, M.R. Bryce, in: E. Buncel, T. Durst (Eds.), Comprehensive Carbanion
Chemistry, Part C, Elsevier, New York, 1987, pp. 271.
[11] R.D. Chambers, J.A.H. MacBride, W.K.R. Musgrave, J. Chem. Soc., Perkin Trans. I
(1968) 2989–2994.
[12] R.D. Chambers, D. Close, D.L.H. Williams, J. Chem. Soc., Perkin Trans. I (1980) 778–
780.
[13] R.D. Chambers, P.A. Martin, J.S. Waterhouse, D.H.L. Williams, B. Anderson, J.
Fluorine Chem. 20 (1982) 507–514.
A mixture consisting of aniline (0.93 g, 10 mmol), 3-chloro-5-
cyano-2,4,6-trifluoropyridine 5 (0.8 g, 4 mmol) and dioxane (6 mL)
was heated in a sealed tube for 8 days at 80 8C. Work up gave a
mixture (1.2 g, 66%), containing the corresponding 4- (6a, 75%) and
2- (6b, 25%) anilino derivatives, which after purification by column
chromatography (dichloromethane:hexane, 60:40 v:v), gave a
mixture of isomers; (Found: C, 53.9; H, 2.4; N, 15.7. C12H6ClF2N3
requires: C, 54.24; H, 2.26; N, 15.82%). Purification by analytical
scale HPLC gave small quantities of each isomer; 5-chloro-2,6-
difluoro-4-(phenylamino)nicotinonitrile 6a; dF ꢁ61.2 (m, F-6),