Reactions of polyfluoropyridines with bidentate nucleophiles: Attempts to prepare deazapurine analogues

Article abstract of DOI:10.1016/S0022-1139(00)00335-3

Reactions of pentafluoropyridine with guanidine, thiourea and urea with the hope of preparing derivatives which may be cyclised to polyfluorodeazapurines are described The reactions with urea and thiourea produced, somewhat unexpectedly, bis-(2,3,5,6-tetrafluoro-4-pyridyl)amine and bis-(2,3,5,6-tetrafluoro-4-pyridyl)sulfide, respectively, the structure of the amine was confirmed by X-ray crystallography. The product from the guanidine reaction was the expected 4-substituted tetrafluoropyridine derivative. Attempts to cyclise the latter are described and a rationalisation to explain the unexpected melamine derivatives which were produced is given.

Full text of DOI:10.1016/S0022-1139(00)00335-3

Journal of Fluorine Chemistry 107 (2001) 13±22
Reactions of poly¯uoropyridines with bidentate nucleophiles:
attempts to prepare deazapurine analogues
Paul L. Coe*, Anthony J. Rees, Jane Whittaker
School of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
Received 5 April 2000; accepted 7 August 2000
Abstract
Reactions of penta¯uoropyridine with guanidine, thiourea and urea with the hope of preparing derivatives which may be cyclised to
poly¯uorodeazapurines are described The reactions with urea and thiourea produced, somewhat unexpectedly, bis-(2,3,5,6-tetra¯uoro-4-
pyridyl)amine and bis-(2,3,5,6-tetra¯uoro-4-pyridyl)sul®de, respectively, the structure of the amine was con®rmed by X-ray
crystallography. The product from the guanidine reaction was the expected 4-substituted tetra¯uoropyridine derivative. Attempts to
cyclise the latter are described and a rationalisation to explain the unexpected melamine derivatives which were produced is given.
# 2001 Elsevier Science B.V. All rights reserved.
Keywords: Poly¯uoropyridines; Bidentate nucleophiles; Cyclisation; Deazapurines
1. Introduction
functional groups which could either spontaneously or in a
two-step process cyclise to the desired ring systems. This
kind of process has been exempli®ed by the reaction of
dimethyl-2-(S-(tetra¯uoro-4-pyridyl)fumarate (2) leading to
3 [4] and by the work of Herkes [5] who showed that
poly¯uorobenzothiazoles (4) could be prepared by cyclisa-
tion of poly¯uorothioanilides which in turn can be prepared
by thioacylation of the corresponding anilines as shown in
Scheme 1. There are relatively few examples of reactions of
¯uorinated aromatic compounds with bidentate nucleophiles
as indicated in a recent comprehensive review of the area
[6]. There is, to our knowledge, only one example in the
benzenoid series in which the ®rst attack by nitrogen
is followed by a spontaneous cyclisation, this is the forma-
tion of 5,6,7,8-tetra¯uoro-1,2,3,4-tetrahydroquinoxaline (5)
reported by Burdon et al. [7] who reacted hexa¯uorobenzene
with ethylene diamine at high temperature over a prolonged
period The reason for this is almost certainly the very
high deactivating effect of nitrogenous functions on nucleo-
philic substitutions of ¯uoroaromatics [8]. Most other nitro-
gen heterocycles described in this ®eld are prepared by
reactions of suitable vicinal diamines in standard hetero-
cyclic syntheses.
In a previous paper [1], we have described work directed
towards the formation of ¯uorodeazapurine derivatives as
potential precursors for anti-viral agents and more impor-
tantly as precursors in the preparation of anti-sense nucleo-
sides. Part of the rationale for this work was the discovery
that the replacement of nitrogen by the isoelectronic CF
group in ribovirin analogues albeit in a ®ve-membered ring
afforded compounds with signi®cant anti-in¯uenza activity
[2]. As part our program in the study of antiviral agents it
was of interest to see if similar replacement of a purine
nitrogen both in the imidazole ring akin to the work of
Townsend and coworkers [3] and also in the pyrimidine ring
would lead to activity. We were able to prepare 3-deaza-
2,3,6-tri¯uoro-8-hydroxypurine (1) (see Scheme 1) that
could after reduction and suitable nucleophilic substitution
lead to 3-¯uoro-3-deazaguanine or the corresponding 2,3-
di¯uoro-3-deaza-adenine analogues. Unfortunately, we
could not deoxygenate 1 to the required purine using a
variety of methods and hence we were unable to study the
elaboration of the molecule to the desired derivatives.
Thus, in a different approach, we decided to investigate
the possibility of preparing 4-substituted poly¯uoropyridine
derivatives using bidentate nucleophiles containing suitable
A search of the literature reveals only one report of a
successful reaction between guanidine and a ¯uoroaromatic
by Popova [9] who showed that 2,4,6-tri¯uoropyrimidine, a
highly reactive species, afforded 2-guanidino-4,6-di¯uoro-
pyrimidine. We found no reports of reactions between urea
* Corresponding author. Fax: ‡44-121-414 4403.
E-mail address: plcoe@chemistry.bham.ac.uk (P.L. Coe).
0022-1139/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 0 0 ) 0 0 3 3 5 - 3

14
P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
Scheme 1.
and any ¯uoroaromatics. We showed earlier [10] that reac-
tion between thiourea and its N-substituted derivatives and
poly¯uorobenzenes yielded poly¯uorodiarylsulphides. We
therefore, decided to investigate the reactions of these three
reagents with poly¯uoropyridines, in particular penta¯uor-
opyridine (6) and 3-chlorotetra¯uoropyridine (7), in the
hope of obtaining the corresponding 4-substituted com-
pounds, which is the expected orientation for nucleophilic
substitution in these pyridines. The possibility of cyclisation
of these products to the desired deazapurine analogues, as
shown in Schemes 2 and 3, would then be investigated,
hopefully leading to the desired deazapurines.
2. Results and discussion
The ®rst reaction we studied was that between guanidine
(as its hydrochloride) and penta¯uoropyridine (6) in the
presence of excess sodium hydride ®rstly to release the free
base and secondly to generate the anion from it. The
Scheme 2.

P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
15
Scheme 3.
reaction, monitored by TLC, proceeded rather slowly but
yielded a single compound in 63% yield after reaction for 2
days in re¯uxing ether. The product was identi®ed by a
combination of physical techniques. The 1H NMR spectrum
showed only a broad singlet which disappeared on shaking
the sample with D2O, indicating the presence of exchange-
able protons. The 19F NMR spectrum showed the character-
istic bands at d 97.2 and 154.9 for 4-substituted
tetra¯uoropyridines [11] and the 13C NMR spectrum
was consistent with this pattern showing four bands attri-
butable to the 2 and 6, 3 and 5, and 4 ring and guanidino
carbon atoms, respectively. The mass spectral and elemental
analysis data con®rmed the structure as N-(2,3,5,6-tetra-
¯uoropyridyl)guanidine (8a). In the same way 3-chlorote-
tra¯uoropyridine (7) afforded N-(3-chloro-2,5,6-tri¯uoro-4-
pyridyl)guanidine (8b) identi®ed as above for 8a by similar
physical methods (see Section 3 for details). We found using
different solvents (ethanol, THF or acetonitrile) with or
without added base were of no advantage, indeed without
the addition of two equivalents of base very little reaction
occurred. Attempts to cyclise 8a or 8b to the desired
deazapurine derivatives 9a and 9b are described below. This
result is interesting in that a similar reaction with hexa-
¯uorobenzene gave no reaction at low temperature and at
higher temperature in a sealed tube gave only polymeric
material [12]. This presumably re¯ects the greater reactivity
of poly¯uoro¯uoropyridines towards nucleophiles vis-a-vis
hexa¯uorobenzene and is in keeping with the results
obtained by Popova with 2,4,6-tri¯uoropyrimidine.
[13]. We found that substitution did not occur under the
conditions described above for the reaction with guanidine.
We changed the solvent from ether to the higher boiling
dimethoxyethane and after heating the reaction for 18 h
obtained a single compound in 92% yield as a white crystal-
1
line solid. The H NMR spectrum showed a broad peak
exchangeable with D2O, the 19F NMR spectrum again
showed the characteristic bands for a 4-substituted tetra-
¯uoropyridine at d 92.7 and 153.9. The 13C spectrum
was at ®rst somewhat puzzling showing bands for only the
ring carbon atoms again characteristic of a 4-substituted
tetra¯uoropyridine. The mass spectrum and elemental ana-
lysis indicated that the product had the molecular formula
C10HF8N3. This latter, together with the above data sug-
gested that the compound was bis-(2,3,5,6-tetra¯uoro-4-
pyridyl)amine (10a). This compound has been reported
previously by a different route [14±16]. The structure was
con®rmed by X-ray analysis1 on a single crystal as shown in
Fig. 1 (Tables 1 and 2). A similar reaction occurred with 3-
chlorotetra¯uropyridine (6) to yield the corresponding
amine 10b. Caution! We subsequently found that both
of these compounds have very high biological activity as
1 The X-ray data was collected using a Rigaku R-axis II area detector
refracrometer using graphite monochoromated Mo Ka radiation. The
structure was determined by direct methods (molecular structure
corp. TEXSAN single crystal structure analysis software version 1.6 and
G.M. Sheldrick SHELXL-93 program for crystal structure refinement. The
data and analysis was carried out by Dr. T.A. Hamor of the School of
Chemistry, University of Birmingham. The formula of the compound is
The next reaction we tried was that of urea with 6. The
rationale for this choice was that urea is frequently used as a
bidentate nucleophile in purine and deazapurine synthesis
C10HN3F8 and its Found mass is 315.004239 and the required mass is
315.004273, the error is 0.1 ppm.

16
P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
Fig. 1.
insecticides but also may well have very high mammalian
toxicities and should be treated with extreme caution
[17]. Compound 10a has been reported in a list of com-
pounds in a patent, but with no experimental details, as being
bioactive [16]. Although these reactions were in themselves
unusual and provided some biologically interesting com-
pounds they did not enable us to proceed further towards our
original goal. We, therefore, investigated the reaction of
penta¯uoropyridine (6) with the more reactive thiourea. We
found that reaction in ethanol in the absence of added base
NMR data indicated that the product contained 4-substituted
tetra¯uoropyridyl residues and no other carbon atoms and
that there were no protons present in the molecule. Mass
spectral and elemental analysis data con®rmed the mole-
cular formula as C10N2F8S and all the data together con®rms
the compound to be bis-(2,3,5,6-tetra¯uoro-4-pyridyl)sul-
®de (11a). This compound has been reported as a byproduct
in the formation of tetra¯uoro-4-mercaptopyridine [18]. In a
similar way the pyridine (7) and thiourea gave the sul®de
(11b). These results particularly in the urea reactions
were somewhat surprising although in the thiourea case
afforded a crystalline product in 71% yield. 1H, 19F and 13
C

P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
17
Table 1
Table 1 (Continued )
Data for X-ray crystallography
F(6)±C(8)±C(7)
119.5 (5)
116.4 (5)
125.3 (5)
118.3 (5)
120.7 (4)
120.5 (4)
118.8 (4)
Temperature (K)
Ê
Wavelength (A)
293
N(2)±C(9)±F(7)
N(2)±C(9)±C(10)
F(7)±C(9)±C(10)
F(8)±C(10)±C(9)
F(8)±C(10)±C(6)
C(9)±C(10)±C(6)
0.71069
Orthorhombic
Pnma
Crystal system
Space group
Unit cell dimensions
Ê
a (A)
13.027
15.018
11.435
2237.2
8
Ê
b (A)
Ê
c (A)
Ê 3
Volume (A )
Z
the formation of the sul®de may have been expected based
on the results we obtained from a similar reaction with
hexa¯uorobenzene [10].
Density (Mg m3)
Absorption coefficient (mm
R indices
1.871
0.210
0.0577
1
)
A rationalisation of these observations is shown in
Scheme 4. We believe the ®rst step in these reactions is
the expected nucleophilic attack by nitrogen and sulfur,
respectively, to give the ®rst formed products A and B
which under any conditions reacted rapidly and could not
be isolated. In the case of the urea reaction there are now two
possible reactions, either direct attack of A on another
molecule of penta¯uoropyridine, or attack of the base
(NaH) on A. Loss of isocyanic acid or carbodiimide leaves
the corresponding anion C attack of which on a further
molecule of the starting pyridine yields 10a or 10b. In the
case of the thiourea reaction the same pathways are both
possible via B. However, for the anion D to form it is
necessary to postulate ethanol acting as the base since the
reaction proceeds without the addition of base. Iso-thiour-
onium salts of the type B normally require strong bases for
Ê
Bond length (A)
F(1)±C(2)
F(2)±C(3)
F(3)±C(4)
F(4)±C(5)
F(5)±C(7)
F(6)±C(8)
F(7)±C(9)
F(8)±C(10)
N(1)±C(3)
N(1)±C(4)
N(2)±C(9)
N(2)±C(8)
N(3)±C(1)
N(3)±C(6)
C(1)±C(2)
C(1)±C(5)
C(2)±C(3)
C(4)±C(5)
C(6)±C(7)
C(6)±C(10)
C(7)±C(8)
C(9)±C(10)
1.349 (5)
1.348 (5)
1.346 (5)
1.351 (4)
1.353 (5)
41.33 (5)
1.357 (5)
1.345 (5)
1.315 (6)
1.318 (5)
1.300 (6)
1.318 (6)
1.375 (5)
1.397 (5)
1.392 (6)
1.396 (6)
1.364 (6)
1.354 (6)
1.393 (6)
1.396 (6)
1.365 (7)
1.378 (7)
Table 2
Atomic coordinates (Â104) and equivalent isotropic displacement para-
2
ꢀ
meters (A Â 103)a
Bond angles (8)
C(3)±N(1)±C(4)
C(9)±N(2)±C(8)
C(1)±N(3)±C(6)
N(3)±C(1)±C(2)
N(3)±C(1)±C(5)
C(2)±C(1)±C(5)
F(1)±C(2)±C(3)
F(1)±C(2)±C(1)
C(3)±C(2)±C(1)
N(1)±C(3)±F(2)
N(1)±C(3)±C(2)
F(2)±C(3)±C(2)
N(1)±C(4)±F(3)
N(1)±C(4)±C(5)
F(3)±C(4)±C(5)
F(4)±C(5)±C(4)
F(4)±C(5)±C(1)
C(4)±C(5)±C(1)
C(7)±C(6)±C(10)
C(7)±C(6)±N(3)
C(10)±C(6)±N(3)
F(5)±C(7)±C(8)
F(5)±C(7)±C(6)
C(8)±C(7)±C(6)
N(2)±C(8)±F(6)
N(2)±C(8)±C(7)
x
y
z
U (eq)
114.8 (4)
116.0 (5)
128.6 (4)
125.9 (4)
119.0 (4)
115.0 (4)
120.6 (4)
120.3 (4)
119.0 (4)
115.9 (4)
125.9 (4)
118.2 (4)
115.7 (4)
124.9 (4)
119.4 (4)
120.9 (4)
118.8 (4)
120.3 (4)
115.4 (4)
119.3 (4)
125.1 (4)
121.0 (4)
119.0 (4)
119.9 (5)
116.1 (5)
124.4 (5)
F(1)
F(2)
F(3)
F(4)
F(5)
F(6)
F(7)
F(8)
N(1)
N(2)
N(3)
C(1)
C(2)
C(3)
C(4)
C(5)
C(6)
C(7)
C(8)
C(9)
C(10)
H(1)
4282 (2)
6098 (2)
4795 (2)
2885 (2)
1134 (2)
258 (2)
618 (2)
896 (2)
437 (2)
1492 (3)
3860 (3)
2967 (2)
438 (2)
64 (1)
86 (1)
88 (1)
76 (1)
74 (1)
90 (1)
98 (1)
75 (1)
62 (1)
70 (1)
57 (1)
46 (1)
50 (1)
58 (1)
57 (1)
52 (1)
50 (1)
56 (1)
65 (1)
66 (1)
56 (1)
79 (17)
3016 (2)
2801 (2)
1931 (2)
580 (2)
1718 (2)
310 (3)
2000 (3)
2893 (2)
5446 (3)
1152 (3)
2515 (3)
3491 (3)
4355 (3)
5275 (3)
4632 (3)
3677 (3)
2099 (3)
1385 (3)
949 (4)
1511 (2)
273 (2)
1700 (2)
2688 (3)
714 (4)
1948 (3)
456 (3)
1588 (3)
1670 (3)
1208 (3)
1364 (3)
2406 (3)
2303 (3)
887 (3)
1259 (3)
1704 (4)
1323 (4)
1855 (4)
3018 (4)
2572 (4)
610 (4)
1073 (3)
394 (4)
264 (4)
882 (4)
1803 (4)
2283 (3)
2091 (39)
636 (3)
15 (3)
115 (5)
813 (4)
1982 (32)
1493 (43)
a U (eq) is defined as one-third of the traces of the orthogonolised Uij
tensor.

18
P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
Scheme 4.
their hydrolysis/decomposition and it seems that proton
abstraction by ethanol is unlikely in this case. We may,
therefore, have to postulate that the iso-thiouronium salts
react as shown in a different way, e.g. ethanol acting as a
nucleophile leading to formation of the anion which then
reacts with more starting material.
Thus, only in the reaction with guanidine we were able to
isolate a product with the potential for cyclisation to the
desired deazapurine analogues. In the light of the literature
evidence, we would expect the cyclisation process to be
dif®cult for two reasons; ®rstly the lone pair on the nitrogen
atom when attached to a poly¯uoroaryl ring has a powerful
deactivating effect on the positions ortho and para to it [7].
Secondly, the 3-position in poly¯uoropyridines has been
shown to be unreactive towards most nucleophiles [19]. We,
therefore, decided to use forcing conditions to attempt the
cyclisation reaction. To achieve cyclisations of the type we
required the previous studies have indicated that the use of
relatively high boiling dipolar aprotic solvents or poly ethers
such as diglyme is required [20], frequently base, usually a
tertiary amine or sodium hydride, is added. We tried a series
of reactions using a wide range of conditions and obtained
some interesting and potentially useful results. In all cases in
the absence of base we recovered only starting materials
when 8a or 8b were heated below 1008C. Heating of 8a in
DMF, NMP, sulpholane or DMSO at temperatures between
80 and 2008C in the presence of sodium hydride or triethy-
lamine afforded only 4-amino-tetra¯uoropyridine (12) as the
sole isolable product. Thus, our attempts to prepare the
desired deazapurine structure by this method failed. How-
ever, when 8a was heated in re¯uxing anisole (1548C)
without base a different reaction occurred and we obtained
a mixture of four products, the major one of which
was unchanged starting material. Separation of the mixture
by column chromatography afforded starting material 65%,
4-amino-tetra¯uoropyridine (12) (10%) identical to an
authentic sample, and crystalline solids 13 and 14. The ®rst
1
of these 13 was identi®ed by physical methods. The H

P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
19
1
NMR spectrum showed a broad single band which was
exchangeable with D2O at d 9.5 indicative of an amino
function, whilst the 19F spectrum showed a set of bands with
very similar chemical shifts to those found for 4-amino-
tetra¯uoropyridine. An accurate mass measurement gave a
mass peak corresponding to a molecular formula of
C18H3F12N9. These data together are compatible with the
structure of 13 as tris-(N-[tetra¯uoro-4-pyridyl])melamine.
Compound 14 was, in a similar manner indicated, to be bis-
(N-[tetra¯uoro-4-pydridyl])melamine (see Section 3 for
details). The origin of these latter compounds is interesting
and possible mechanisms which account for the products are
shown in Scheme 5. The loss of carbodiimide from the
guanidine derivative 8a under the in¯uence of base leads
directly to the formation of anion of 4-amino-tetra¯uoro-
pyridine (12) which would readily protonate. We believe, as
shown, that the thermal reaction leading to 12, 13 and 14
takes a different course. The rationale for this comes ®rstly
from the early work of Davis and Underwood [20] who
showed that heating of guanidine carbonate in an open tube
led to the formation of cyanamide and ammonia whereas
heating the carbonate with a condensing system in a sub-
limation apparatus led to the formation of melamine. Sub-
sequent work, notably by Pankratov [21], has shown that
heating of N-arylcyanamides leads to the formation of N-
aryl-melamine derivatives possibly via dicyanamide and the
isomelamine derivative. The latter has been shown to rear-
range irreversibly to the melamine at temperatures in the
range 150±2008C [22]. We believe that the ®rst step in the
reaction follows the literature precedent to give 4-cyana-
mido-tetra¯uoropyridine. This latter then has two possible
reaction pathways either direct thermal trimerisation to give
the observed melamine, a known process [21] or the sequen-
tial nucleophilic addition process as shown. This has the
merit that it more satisfactorily explains the formation of the
dimeric product. Our preferred route is via the stepwise
build up, but we have no evidence to support one mechanism
against the other.
unless stated otherwise. H NMR spectra (400 MHz) were
measured on a Bruker AMX 400 NMR spectrometer. 19F
NMR spectra were carried out either on a Jeol NMR
spectrometer, type FX 90 Q (84.26 MHz) or on a Bruker
AC 300 NMR spectrometer (282.4 MHz); tetramethylsilane
and ¯uorotrichloromethane were used as internal references.
For the characterisation of the signals the following abbre-
viations are used: s: singlet; d: doublet, t: triplet, q: quartet,
quin: quintet, sext: sextet, m: multiplet, dd: doublet of
doublets, dt: doublet of triplets, quin: pseudo-quintet, etc.
J values are given in Hz. The mass spectra (CI-MS/EI-MS)
were measured on a VG-Prospec-triple focusing mass spec-
trometer. For GC-MS analysis, a Carlo Erba, 8000 series GC
was used with a 50 m column, BPX 5 (helium carrier gas,
70 eV, electron impact).
Thin layer chromatography was performed on TLC plas-
tic sheets silica gel 60 F254, pre-coated with a layer thickness
of 0.2 mm from Merck, Art. 5735. Gas chromatographic
analysis was carried out using a Philips PYE Unicam, Series
304 chromatograph with a 50 m CD-SIL-CB 19 column.
The data were registered by a JCL 600 chromatography data
system.
3.1. Materials
Penta¯uoropyridine and 3-chlorotetra¯uoropyridine were
prepared as previously described [1] although they are
commercially available at relatively high cost (Aldrich).
All other reagents were obtained from Aldrich.
3.1.1. Preparation of N-(2,3,5,6-tetrafluoro-4-pyridyl)
guanidine (8a)
Sodium hydride (0.236 g of 60% dispersion in oil, washed
three times with dry ether) was suspended in dry ether
(40 cm3). The suspension was cooled in ice and dry guani-
dine hydrochloride (0.283 g, 2.96 mmol) was added and the
mixture stirred for 30 min after which penta¯uoropyridine
(6) (0.5 g, 2.95 mmol) was added dropwise. The mixture
was heated at re¯ux for 2 days when TLC and GLC of the
liquid showed all the pyridine had reacted. Saturated ammo-
nium chloride solution (30 cm3) was carefully added, all the
solids dissolved in the aqueous phase. The ether layer and
the combined extracts of the aqueous phase (3 Â 20 cm3)
were dried (MgSO4) and the solvent evaporated to yield a
small amount of an oil which contained no ¯uorinated
products. The aqueous phase was made basic (saturated
Na2CO3) and re-extracted with ether (3 Â 50 cm3). After
drying (MgSO4) the ether layer was evaporated to afford N-
(2,3,5,6-tetra¯uoro-4-pyridyl)guanidine (8a) (nc) (0.4 g,
64%) mp 134±1358C (from hexane/ethyl acetate); (Found:
C, 34.4; H, 2.1; N, 26.6% C6H4N4F4 requires C, 34.6; H, 1.9;
N, 26.9%). dH (DMSO-d6) 6.15 (bs, equivalent NH), dF
(DMSO-d6) 97.2 (m, 2F, 2F and 6F), 154.9 (m, 2F, 3F
Thus, we conclude that although we have been unable to
prepare the desired deazapurine precursors by this approach
we have discovered some unexpected and potentially useful
reactions. The melamines and their method of synthesis
could lead to the production of some unusual and interesting
ligands since by careful choice of pyridine derivatives we
could in principle prepare a range of nine co-ordinate
ligands which may have interesting properties. As indicated
above some of the compounds we have made have shown
extremely potent biological activities and work is continuing
to study structure activity relationships with a range of their
analogues.
3. Experimental
1
and 5F), dC (DMSO-d6) 156.3 (s, C7), 143.6 (dm, JFC
1H NMR (300 MHz) and 13C NMR spectra (75 MHz)
were measured on a Bruker AC 300 NMR spectrometer
233.4, C2), 143.5 (m, C4), 136.1 (dm, 1JFC 247.4, 3C); m/z
‡
‡
‡
208 [M] , 192 [M NH2] , 166 [M NHCNH] .

20
P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
Scheme 5.

P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
21
‡
3
4
3.1.2. Preparation of N-(3-chloro-2,5,6-trifluoro-4-
pyridyl)guanidine (8b)
2JFC 37.5, JFC or JFC 6.2, C3); m/z 347/349/351 [M] .
Found mass, 346.94517. Required mass (C10HCl2F6N3),
346.94517.
In the same manner as above, the chloropyridine (7)
(0.24 g, 2.96 mmol) afforded N-(3-chloro-2,5,6-tri¯uoro-
4-pyridyl) guanidine (8b) (0.4 g) mp 131±1348C (from
hexane/ethyl acetate); (Found: C, 32.2; H, 1.5; N, 24.8%
C6H4ClN4F3 requires C, 32.1; H, 1.8; N, 24.9%); dH
(DMSO-d6) 6.2 (bs, equivalent NH), dF (DMSO-d6)
3.1.5. Preparation of bis-(tetrafluoro-4-pyridyl)sulfide
(11a)
Penta¯uoropyridine (6) (1.50 g, 8.88 mmol) and thiourea
(7.5 g, 98.7 mmol) were heated together in ethanol (50 cm3)
at re¯ux for 20 h. The mixture was cooled and the unreacted
thiourea was ®ltered off. The ®ltrate and the ether washings
of the precipitate were combined and the solvents evapo-
rated to leave a residual solid which was re-extracted with
cold ether (2 Â 20 cm3) to leave a further quantity of
thiourea. The ether layer was washed successively with
2M HCl (20 cm3) and water (20 cm3), dried (MgSO4) and
the solvent evaporated to leave a pale yellow solid which
was freed from any residual thiourea by column chromato-
graphy (hexane:ethyl acetate 1:1 as eluant) to yield bis-
(tetra¯uoro-4-pyridyl)sul®de (11a) (1.05 g, 71%) mp 43±
468C (cited 52±53 [18]); (Found: C, 36.0; N, 8.4% C10F8N2S
requires C, 36.1; N, 8.4%); dF (CDCl3) 90.4 (m, F2 and
5
4
78.4 (dd, 1F, JFF 21.4, JFF 15.3, F2) 96.2 (dd, 1F,
4
5
5JFF 21.4, JFF 15.3, 6F), 155.4 (t, 1F, JFF ˆ 3JFF
21.4,5F), dC (DMSO-d6) 156. (s, C7), 151.4 (m, C4),
3
1
151.1 (dd, 1JCFC 232.6, JFC 17.7, C2),147.1 (dt, JFC
2
1
235.9, JFC ˆ 3JFC 17.8, C6), 135.7 (ddd, JFC 246.6,
3JFC 24.9, JFC 5.4, C5), 105.7 (m, C3); m/z 224/226
[M] , 208/210 [M NH2] , 182/184 [M NHCNH] .
4
‡
‡
‡
3.1.3. Preparation of bis-(tetrafluoro-4-pyridyl)amine
(10a)
Sodium hydride (0.237 g as a suspension 60% in mineral
oil) which had been washed with dry hexane (2 Â 10 cm3)
then with dry ether (2 Â 10 cm3) and ®nally with diglyme
(5 cm3) was suspended in diglyme (20 cm3). A solution of
urea (0.178 g, 2.97 mmol) in diglyme (20 cm3) was added
dropwise; when the initial evolution of hydrogen has sub-
sided penta¯uoropyridine 6 (0.5 g, 2.96 mmol) was added
and the reaction heated at re¯ux for 18 h. The mixture was
cooled and the solvent evaporated to leave a gum to which a
mixture of ether (40 cm3), water (10 cm3) and 3 M HCl
(20 cm3) was added. The solution was separated and the
ether layer dried (MgSO4) and the solvent evaporated to a
leave a crystalline mass, which was recrystallised from ether
to yield bis-(tetra¯uoro-4-pyridyl)amine (10a) (0.43 g,
93%) mp 144±1468C (cited 149±1518C [14,15]); (Found:
C, 38.1; H, 0.4; N, 13.2% C10HF8N3 requires C, 38.1; H,
0.3;N, 13.3%); dH (acetone-d6) 9.6 (bs, D2O exchangeable
NH); dF (acetone-d6) 92.7 (m, 2F, F2 and F6), 153.8 (m,
1
F6), 135.1 (m, F3 and F5); dC (CDCl3) 144.3 (dm, JFC
245.3, 2C and 6C), 143.2 (dm, 1JFC 259.8, C3 and C5), 125.2
‡
(m, C4); m/z 332 [M] .
3.1.6. Preparation of bis-(3-chloro-trifluoro-4-
pyridyl)sulfide (11b)
In the same manner as above 3-chloro-tetra¯uoropyridine
(7) (1 g, 5.4 mmol) and thiourea (5.0 g, 65.8 mmol) afforded
bis-(3-chlorotri¯uro-4-pyridyl)sul®de (11b) (nc) (0.25 g,
25%) mp 33±368C; (Found: C, 33.0; N, 7.7% C10Cl2F6N2S
requires C, 32.9; N, 7.7%); dF (CDCl3) 73.0 (dd, 2F, 5JFF,
27.5, 4JFF 12.2, F2), 88.5 (dd, 2F, 3JFF 21.4, 4JFF 12.2, F6),
133.7 (dd, 2F, 3JFF 21.4, 5JFF 27.5, F5); dC (CDCl3) 151.9
(dd, 1JFC 242.6, 3JFC 13.4, C2), 148.1 (ddd, 1JFC 246.9, 2JFC
3
1
2
17.1, JFC 15.1, C6), 143.7 (ddd, JFC 258.7, JFC 28.3,
4JFC6.1, C5) 136.5 (dm, 2JFC 15.3, C4), 116.8 (m, C3); MS
1
2F, F3 and F5); dC (acetone-d6) 144.3 (ddt, JFC 238.6,
‡
2JFC ˆ 3JFC 16.6 4JFC is seen as an outer band of triplets but
m/z 364/366/368[M] .
is too small to measure, C2), 135.9 (dm, 1JFC 256, C3); m/z
‡
‡
315 [M] 296 [M F] . Found mass, 315.00423. Required
mass (C10HF8N3), 315.00427.
3.1.7. Reaction of N-(tetrafluoro-4-pyridyl)guanidine (8a)
with sodium hydride
The guanidine derivative 8a (0.5 g, 2.4 mmol) in NMP
(50 cm3) was treated with sodium hydride (0.1 g, 2.4 mmol,
60% dispersion in oil washed from the oil as described
above) at 1008C for 20 h. The mixture was cooled and ether
(100 cm3) and 2 M HCl (100 cm3) were added and the
mixture was continuously extracted for 24 h. The ether layer
was dried (MgSO4) and the solvent evaporated to leave a
solid (0.15 g) which was identi®ed as 4-amino-tetra¯uor-
opyridine (12) [23] by comparison with an authentic sam-
ple.Similar reactions in other solvents, e.g. diglyme, DMSO,
DMF and THF all gave essentially the same results.
3.1.4. Preparation of bis-(3-chloro-trifluoro-4-
pyridyl)amine (10b)
In the same manner as above reaction of 3-chloro-tetra-
¯uoropyridine (7) (2.0 g, 10.8 mmol) afforded, after puri®-
cation by column chromatography (hexane: ethyl acetate 1:1
as eluant), bis-(3-chloro-tri¯uoro-4-pyridyl) amine 10b (nc)
(1.8 g, 96%) as a gum; (Found: C, 34.6; H, 0.4; N, 12.0%
C10HCl2F6N3 requires C, 34.5; H, 0.3; N, 12.1%); dH
(CDCl3) 7.0 (bs, D2O, exchangeable NH); dF (CDCl3),
4
5
4
73.7 (dd, 1F, JFF 12.2, JFF 24.4, F2), 88.9 (dd, JFF
5
3
3
12.2, JFF 21.4, F5), 153.8 (dd, 1F, JFF 21.4, JFF 24.4,
1
3
4
F5); dC (CDCl3) 151.3 (ddd, JFC 241, JFC 15, JFC2, C2)
147.5 (dt, 1JFC 244.2, 2JFC ˆ 3JFC 15.3, C6), 138.9 (m, C4),
135.3 (ddd, 1JFC 258.3, 2JFC 29.4, 4JFC 6.8, C5), 104.9 (dd,
3.1.8. Thermal decomposition of 8a
Heating in re¯uxing anisole for 2 days. The guanidine
derivative 8a (1.0 g) was heated in anisole (15 cm3) at the

22
P.L. Coe et al. / Journal of Fluorine Chemistry 107 (2001) 13±22
re¯ux temperature for 2 days. The reaction mixture was
cooled, water (15 cm3), saturated sodium bicarbonate solu-
tion (20 cm3) and ether (30 cm3) were added. The ether layer
was separated, dried (MgSO4) and the solvents evaporated to
leave a pasty solid. TlC showed the presence of four
components the major of which had an identical Rf value
to starting material. The mixture was separated by column
chromatography with ether:hexane (1:1) to yield (i) 4-
amino-tetra¯uoropyridine (12) (0.08 g, 10%) mp 84±858C
(cited [23] 85±868C) identi®ed by comparison with an
authentic sample, (ii) Tris-(N-[tetra¯uoro-4-pyridy])mela-
mine (13) (nc) (0.04 g, 4.4%) mp 142±1438C; (Found: C,
37.4; H, 0.4; N, 21.8% C18H3F12N9 requires C, 37.7; H, 0.5;
N, 22%); dH (acetone-d6) 9.5 (bs, exchangeable with D2O,
NH); dF (acetone-d6) 97.2 (m, 2F, F2 and F6), 145.3
References
[1] P.L. Coe, A.J. Rees, J. Fluorine Chem. 101 (2000) 45.
[2] R. Storer, C.J. Ashton, A.D. Baxter, M.M. Hann, C.L.P. Marr, A.M.
Mason, C.-L. Mo, P.L. Myers, S.A. Noble, C.R. Penn, N.O. Weir,
J.M. Woods, P.L. Coe, Nucleosides and Nucleotides 18 (1999)
203.
[3] J.S. Pudlo, M.R. Massiri, E.R. Kern, L.L. Wotring, J.C. Drach, L.B.
Townsend, J. Med. Chem. 33 (1990) 1984.
[4] A.M. Sipyagin, I.V. Efremova, I.A. Ponytkin, N.N. Alunikova, S.A.
Kashtanov, Khim. Geterotsikl. Soedin 9 (1994) 21.
[5] F.E. Herkes, J. Fluorine Chem. 13 (1979) 1.
[6] G.M. Brooke, J. Fluorine Chem. 86 (1997) 1.
[7] J. Burdon, V.A. Damodaran, J.C. Tatlow, J. Chem. Soc. (1964)
763.
[8] R.D. Chambers, J. Hutchinson, W.K.R. Musgrave, J. Chem. Soc.
(1964) 3736.
‡
‡
[9] L.M. Popova, E. Studentsov, Russ. J. Org. Chem (Eng. Transl.) 33
(1996) 1370.
(m, 2F, F3 and F5); MS m/z 573 [M] , 554 [M F] .
Found mass, 573.032980. Required mass (C18H3F12N9),
573.032922, (iii) bis-(N-[tetra¯uoro-4-pyridyl])melamine
(14) (nc) (0.07 g, 6.9%) mp 206±2078C; (Found: C, 37.1;
H, 1.1; N, 26.2% C13H4F8N8 C, 36.8; H, 1.0; N, 26.4%); dH
(acetone-d6) 9.1 (bs,2H, D2O exchangeable, 2 Â NH), 0.6
(bs, 2H, D2O exchangeable, NH2), dF (acetone-d6) 93.6
(m, 2F, F2 and F6) 145.7 (m, 2F, F3 and F5); dC (acetone-
d6) 168.9 (s, CNH2), 165.6 (s, CNH), 144.3 (dt, 1JFC 239.6,
2JFC ˆ 3JFC 16.4, 4JFC toosmall tomeasure, C2), 138.6(ddm,
[10] P.L. Coe, N.E. Milner, J.C. Tatlow, R.T. Wragg, Tetrahedron 28
(1972) 105.
[11] J. Lee, K.G. Orrell, J. Chem. Soc. (1965) 582.
[12] J. Burdon, G.M. Brooke, private communication.
[13] J. Davoll, J. Chem. Soc. (1960) 131.
[14] G.G. Furin, A.O. Miller, G.G. Yakobson, Izv. Sib. Akad. Nauk.
SSSR. Ser Khim. 1 (1985) 127.
[15] A.O. Miller, G.G. Furin, J. Fluorine Chem. 75 (1995) 169.
[16] Ger. Offen. (1972), 2139042, CA 76 126795a.
[17] G.P. Lahm, Du Pont Agricultural Products, personal communication.
[18] L.S. Kobrina, G.G. Furin, G.G. Yakoson, J. Gen. Chem. Russ. (Eng.
Transl.) 38 (1968) 505.
‡
1JFC 258, 2JFC 35.3, C3), 131.8 (m, C4); m/z 424 [M] , 405
‡
[M F] . Found mass, 424.043118. Required mass
[19] R.D. Chambers, J.S. Waterhouse, D.L.H. Williams, J. Chem. Soc.,
Perkin Trans. 1 (1977) 585.
(C13H4F8N8), 424.0425145. and (iv) starting material (0.4 g).
In a repeat of the above reaction but with heating for 10
days under re¯ux we obtained complete reaction of the start-
ing material to afford a mixture (0.6 g) which was separated
as above to give(i) 4-amino-tetra¯uoropyridine 46%, (ii)
tris-(N-[tetra¯uoro-4-pyridyl])melamine 20% and (iii) bis-
(N-[tetra¯uoro-4-pyridyl])melamine 34%, these ratios are
approximately the same as we found in the 2-day experiment.
[20] T.L. Davis, H.M. Underwood, J. Am. Chem. Soc. 22 (1922) 259.
[21] V.A. Pankratov, A.E. Chesnokova, Russ. Chem. Rev. (Engl. Trans.)
(1989) 879.
[22] V.V. Korshak, D.F. Kutepov, V.A. Pankratov, N.P. Antsiferova,
S.V. Vinogradova, Isz. Akad. Nauk. SSSR. Ser. Khim. 6 (1973)
1408.
[23] R.E. Banks, J.E. Burgess, W.M. Cheng, R.N. Haszeldine, J. Chem.
Soc. (1965) 575.