The proton sponge-triethylamine tris(hydrogen fluoride) system as a selective nucleophilic fluorinating reagent for chlorodiazines

Article abstract of DOI:10.1016/S0040-4039(00)01054-6

The proton sponge-triethylamine tris(hydrogen fluoride) mixture provides a mild and efficient fluorinating - reagent for the selective introduction of a fluorine atom, by halogen exchange, into dichlorodiazines. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01054-6

Tetrahedron Letters 41 (2000) 6763±6767
The proton sponge±triethylamine tris(hydrogen ¯uoride)
system as a selective nucleophilic ¯uorinating reagent for
chlorodiazines
Mircea Darabantu, Thierry Lequeux, * Jean-Claude Pommelet, * Nelly Ple
Â
,b
a
a,
a,
b
c
Alain Turck and Loõc Toupet
È
aLaboratoire de Chimie MoleÂculaire et Thio-organique, CNRS, Universite de Caen-ISMRA,
Boulevard du Mar e chal Juin, 14050 Caen, France
Universite de Rouen, IRCOF, 76131 Mont Saint-Aignan, France
6
b
c
Universit e de Rennes I, Groupe Mati eÁ re Condens e e et Mat e riaux, CNRS 35042 Rennes, France
Received 25 May 2000; accepted 21 June 2000
Abstract
The proton sponge±triethylamine tris(hydrogen ¯uoride) mixture provides a mild and ecient
uorinating reagent for the selective introduction of a ¯uorine atom, by halogen exchange, into
¯
dichlorodiazines. # 2000 Elsevier Science Ltd. All rights reserved.
Keywords: ¯uorine and compounds; regioselection; pyrimidines; pyridazines; phthalazines; quinoxalines.
Both triethylamine tris(hydrogen ¯uoride) (Et N.3HF) and 1,8-bis(dimethylamino)naphtha-
3
1
lene (`proton sponge', PS) as its hydrogen ¯uoride salt (PS.1HF) have recently been reviewed
and reported2 to be ¯uoride ion donors in their ability to form carbon±¯uorine bonds. The
latter is easily prepared and has been successfully shown to generate 2,4,6-tri¯uoropyrimidine
from its trichloro analogue (Scheme 1).
�4
Scheme 1.
*
Corresponding authors. Fax: 2 31 45 28 53; e-mail: pommelet@ismra.fr
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01054-6

6
764
We now report our preliminary results focused on this subject since we considered it of interest
to broaden the use of PS.1HF to various (poly)chlorodiazines in connection to our previous work
in the ®eld.6 To our knowledge, the selective ¯uorination of these derivatives has never been
attempted and the use of PS.1HF, which is active under mild conditions, appeared to us to be
appropriate for this new approach.
,7
Preliminary attention was paid to the synthesis of the ¯uorinating reagent. Thus, in order to
overcome the hazards involved in handling anhydrous hydrogen ¯uoride (even as an ethereal
4
solution), our option was to react the commercially available Et N.3HF with PS (1:3 molar
3
ꢀ
ratio) and then completely remove the triethylamine under vacuum (100 C, NMR monitoring,
Eq. (1)):
ꢁ1†
The resulting solid residue (98±99% yield) exhibited the same IR and NMR spectra previously
,5,8a
reported for PS.1HF and gave satisfactory elemental analysis.4
We normally assigned this
compound as PS.1HF. However, treatment of this product with ether (or pentane), at room
temperature, a€orded a crystalline solid,8b corresponding to the PS.3HF stoichiometry. Work-up
of the mother liquor yielded the theoretical amount of the starting free base, PS (NMR
9
monitoring). Although this behaviour is similar to other `onium' salts of hydrogen ¯uoride we
nevertheless examined whether our product PS.1HF is speci®cally a unitary structure or just a
trivial mixture isolated in a 2PS:1(PS.3HF) molar ratio.
The evidence provided by spectra con®rmed the existence of two distinct compounds (at least
in the solid state) of di€erent stoichiometry, PS.1HF and PS.3HF.
The IR spectra (KBr) were in agreement with the previously reported assignments (e.g. the
broad peak at 600 cm^ consistent with the N±H stretching vibration in protonated forms,4,5 and
the appearance of a Bohlmann band region, typical for PS salts).5
1
The NMR spectra revealed little di€erence regarding the relevant ꢀ values of the two forms
(
e.g. the resonance of ¯uorine at ^166.2 ppm in PS.1HF and at ^167.6 ppm in PS.3HF) but
8b
showed a strong dependence on the acid versus base stoichiometry. Thus, PS.3HF exhibited
almost all the expected sharp splitting in the aromatic region whereas PS.1HF8a revealed only
broad singlets for all four types of `rigid' protons, assigned to a slow change of the environment
(Eq. (2)).
ꢁ2†
+
^
Finally, the X-ray analysis of PS.3HF was consistent with the structure of (PSH ) (HF ) (HF)
2
(Fig. 1).
However, for the use of this reagent in synthesis, we considered it more convenient to generate
in situ the PS.1HF by using mixtures of the type xPS+y(Et N.3HF) (the coecients x, and y are
3
referred to as the number of mols of each component used to 1 mol of the chlorodiazine; see Table
1). In this way, the mono¯uorination of a selection of dichlorodiazines (1±4) was straightforward.
The results are collected in Table 1 (Scheme 2).

6
765
Figure 1. The ORTEP diagram of `proton sponge' tris(hydrogen ¯uoride) PS.3HF
The identity of all reaction products was fully con®rmed by combined spectroscopic data
0,11
(
NMR and HR-MS).1
In none of the selected examples did the use of Et N.3HF alone a€ord
3
the desired products other than in small traces (NMR detection). Thus, the addition of PS not
only improved the ¯uorination ability of this reagent but it also led either to regioselective (entry
1
) or chemioselective chlorine replacement (entries 2±4). In turn, in the absence of triethylamine,
shown by the use of PS.3HF alone, seemed to us unpromising, according to the poor yields
obtained.
4
Surprisingly, the use of a solvent was not mandatory to ensure selectivity. In turn, the rather
long reaction time was found to be essential for a reasonable compromise between selectivity and
conversion. Thus, the method presented allows, for the ®rst time, access and isolation of
mono¯uorinated chlorodiazines with satisfactory yields. In some cases, if desired, complete
substitution of the chlorine atoms becomes the dominant process by using excess of the reagent
(e.g. compound 3b, entry 3).
Liquid compounds (entries 1, 2) were isolated by Kugelrohr distillation. The solids (entries 3, 4)
were separated by ¯ash column chromatography.
Finally, we mention that the recycling of the free base PS, which is quite an expensive reagent,
was regularly undertaken. Thus, simple protocols in the work-up of the crude reaction mixtures
a€orded the pure recovered PS (> 80% for entries 1, 3 and >60% for entries 2, 4). We veri®ed
that its reactivity was not a€ected and that it can be routinely reused in this new method, to
generate the same ¯uorinating reagent.

6
766
Table 1
Results of selective ¯uorination of chlorodiazines 1±4
Scheme 2.
Acknowledgements
This work was supported by the Re
Â
RINCOF) (Contrat de Plan Etat Bassin Parisien-Re
seau Interre
Â
gional Normand de Chimie Organique Fine
Â
(
gions Haute Normandie, Basse Normandie).

6
767
References
1
2
3
4
5
6
7
8
. Mc. Clinton, A. M. Aldrichimica Acta 1995, 28, 31±34.
. Alder, W. R. Chem. Rev. 1989, 89, 1215±1223.
. Chambers, D. R.; Holmes, F. T.; Korn, R. S.; Sandford, G. J. Chem. Soc., Chem. Commun. 1993, 855±856.
. Chambers, D. R.; Korn, R. S.; Sandford, G. J. Fluorine Chem. 1994, 69, 103±108.
. Brzezinsky, B.; Grech, E.; Malarsky, Z.; Sobczyk, L. J. Chem. Soc., Faraday Trans. 1990, 86, 1777±1780.
. Ple
Â
, N.; Turck, A.; Heynderickx, A.; Que
Â
guiner, G. J. Heterocyclic Chem. 1994, 31, 1311±1315.
guiner, G. Tetrahedron 1998, 54, 4899±4912.
. Ple, N.; Turck, A.; Heynderickx, A.; Que
Â
. (a) 1,8-Bis(dimethylamino)naphtalene (hydrogen ¯uoride), sticky crystals; thermal dissociation as shown by Eq.
Â
ꢀ
1
(
8
7
2) between 50±126 C. Anal. calcd for C14
H
19FN
2
: C, 71.76%; H, 8.17%; N, 11.95%; found: C, 72.41%; H,
, ꢀ ppm, J Hz): 2.91 (bs, 12H, -CH ), 7.17 (bs, 2H), 7.38 (bs, 2H),
.48 (bs, 2H); 13C NMR (62.5 MHz, ꢀ ppm, CDCl ): 45.5 (4C, -CH ); only two distinct broad signals were found
.50%; N, 12.35%. H NMR (250 MHz, CDCl
3
3
3
3
in the aromatic region at 120.5 ppm and 126.3 ppm as suggested by the equilibrium depicted by Eq. (2); 19F NMR
(
crystalline powder; mp=125±126 C (ether or pentane). Anal. calcd for C14
235 MHz, CDCl
3
, ꢀ ppm, J Hz): ^166.2 (s). (b) 1,8-Bis(dimethylamino)naphtalene tris(hydrogen ¯uoride), white
ꢀ
H
21
F
3
2
N : C, 61.29%; H, 7.71%; N,
1
1
0.21%; found: C, 61.38%; H, 7.79%; N, 10.23% H NMR (CDCl , ꢀ ppm, J Hz): 3.25 (s, 12H, -CH ), 7.68 (dd
3 3
1
3
as t, 2H, 7.9), 7.93 (dd as t, 4H, J=7.8); C NMR (CDCl
2C), 129.5 (2C), 135.8 (1C), 144.9 (2C); 19F NMR (CDCl
. Yoneda, N. Tetrahedron 1991, 47, 5329±5365.
3
, ꢀ ppm): 46.8 (4C, -CH
3
), 119.3 (2C), 121.6 (1C), 127.5
(
3
, ꢀ ppm, J Hz): ^167.6 (s).
9
1
0. (a) In the case of compounds 1a,b the mild conditions already reported (Tullock, C. W.; Carboni, R. A.; Harder,
ꢀ
R. J.; Smith, W. C.; Co€man, D. D. J. Am. Chem. Soc. 1960, 82, 5107±5110), SF
4
/50±150 C, yielded almost the
di¯uoro derivative 1b. (b) For the compounds 2a,b, Fukuhara and Yoneda reported a 92% yield (as 56% di¯uoro
ꢀ
2
b versus 44% mono ¯uoro derivative 2a, Table 1) by using anhydrous HF in an autoclave (2 h at 100 C); no
structural assignment accompanied these data (Fukuhara, T.; Yoneda, N. Chem. Lett. 1993, 3, 509±512);
combined hard conditions are also reported by: Ishikawa, N.; Kuroda, K.; Onodera, N. Kogyo Kagaku Zasshi
1
971, 74, 1490±1491 (Chem. Abstr. 1971, 75, P76709g). (c) For the di¯uoro derivative 3b, the required hard
ꢀ
conditions (KF/DMSO, 200 C, 70±120 h) were reported as inappropriate for its isolation (Homer, J.
J. Heterocyclic Chem. 1966, 3, 244).
1. Typical procedure for the synthesis of compounds 3: 2,3-Dichloroquinoxaline (1.00g, 5.02 mmol) was dissolved in
1
ꢀ
acetonitrile (5 mL); to this solution, heated at 80 C, `proton sponge' (1.08 g, 5.02 mmol) and triethylamine
tris(hydrogen ¯uoride) (0.300 mL, 1.80 mmol) were added over 96 h, as three equal portions. After an additional
2
4 h the reaction mixture was cooled at room temperature, taken up in 15 mL dichloromethane and the organic
layer was washed several times with water (15 mL). After drying and removing the solvent, ¯ash column
chromatography (silica gel, eluent pentane:chloroform, 5:1) was performed to yield the desired products. 2-
ꢀ
1
Chloro-3-¯uoroquinoxaline 3a (white crystalline powder, mp=88±89 C); H NMR (CDCl , 250 MHz, ꢀ ppm, J
3
Hz): 8.11±8.05 (1H, m), 8.03±7.92 (1H, m), 7.86±7.53 (2H, m); 13C NMR (CDCl
3
, 62.5 MHz, ꢀ ppm, J Hz): 152.1
(
1C, d, 256.3), 141.2 (1C, d, 2.5), 139.0 (1C, d, 8.8), 137.1 (1C, d, 40.6), 131.7 (1C, s), 130.6 (1C, d, 3.1), 128.6 (1C,
s), 128.2 (1C, d, 1.9); 19F NMR (CDCl
and 184.0019; found: 182.0052 and 184.0007. Anal. calcd for C H ClFN : C, 52.63%; H, 2.20%; N, 15.33%;
found: C, 52.96%; H, 2.29%; N, 15.06%. 2,3-Di¯uoroquinoxaline 3b (a yellowish crystalline powder, mp=92±
3 8 4 2
, 235 MHz, ꢀ ppm, J Hz): ^82.7 (1F, s). HR-MS for C H ClFN : 182.0047
8
4
2
ꢀ
1
, 250 MHz, ꢀ ppm, J Hz): 7.88 (2H, dd, 6.3, 3.5), 7.69 (2H, dd, 6.3, 3.5); 13C NMR (CDCl
3
C); H NMR (CDCl
2.5 MHz, ꢀ ppm, J Hz): 146.5 (2C, dd, 261.3, 39.4), 138.8 (2C, dd as t, 5.6), 130.7 (1C, s), 128.2 (1C, s); F NMR
, 235.35 MHz, ꢀ ppm, J Hz): ^82.7 (1F, s). HR-MS for C 20: 166.0342; found: 166.0347. Anal.
calcd. for C : C, 57.83%; H, 2.42%; N, 16.86%; found: C, 57.93%; H, 2.54%; N, 16.61%.
3
3
,
19
6
(
CDCl
3
8 4 2
H N F
8 4 2 2
H F N