Study of reactions of pentafluorophenylhydrazine with activated enol ethers. Synthesis of N -pentafluorophenylpyrazoles

Activated enol ethers derived from methyl or ethyl acetoacetate/cyanoacetates or nitriles and pentane-2,4-dione react with pentafluorophenylhydrazine through the primary amino-group to afford pyrazoles bearing a preferred 5-amino-over 5-methyl-or 5-hydroxy-substituent in the resulting 4-substituted pyrazoles


Introduction
Enol ethers are a wide group of very reactive organic systems 1 which, with the introduction of one or two electron withdrawing groups 1 to the alkoxy group, produce push-pull systems.[6] 2: Ar = C6H5, 3 -9: Ar = C6F5 X, Y, X′, Y′ for 1, 4 -9 (OH or NH2 can exist also in their tautomeric forms 7,8 ) Recently we studied the reaction of pentafluorophenylhydrazine with 2-ethoxymethylene-3oxobutanenitrile and isolated an intermediate enehydrazine (2-pentafluorophenylhydrazinylmethylene-3oxobutanenitrile) due to the lower nucleophilicity of the second nitrogen in comparison with phenylhydrazine 5 (route a).In continuation we have now studied the reaction of the monosubstituted hydrazine (pentafluorophenylhydrazine 3) with enolethers 1a-d,f-i and the chemoselectivity of the cyclisation reaction in the case of 'unsymmetrical' enol ethers 1f-i.Aminonitrile 6a has been previously reported, 9 but no data were given.

Quantum-chemical calculations
Standard geometry optimizations of the compounds under study at DFT level of theory using hybrid B3LYP functional 10 and cc-pVDZ basis sets 11,12 were performed using the Gaussian03 program package. 13Their stability was confirmed by vibrational analysis (no imaginary vibrations).Frontier electron densities (FED) of individual atoms in the highest occupied molecular orbital (HOMO) and in the lowest unoccupied molecular orbital (LUMO) were evaluated as the corresponding net electron populations and used as reactivity indices of these atoms in nucleophilic and electrophilic reactions, respectively.1) which is equal for both compounds.The nucleophilic character of the N(terminal) atom is significantly lower and ca.twice higher in the case of the fluorinated aromatic ring.Higher nucleophilic character of N(central) atoms in comparison with the N(terminal) ones is also supported by their higher negative charges.The electrophilic character of both N atoms in phenylhydrazine based on FED LUMO reactivity indices is very small, unlike its fluorinated analogue where both N atoms exhibit nearly equal reactivity indices, which are ca.three times higher than for phenylhydrazine.The low HOMO-LUMO energy separation of both compounds (ca 0.2 eV) indicates their high reactivity.

Reactivity of pentafluorophenylhydrazine 3
According to quantum-chemical calculations, the arylhydrazines 2, 3 are highly reactive and there is no large difference between their reactivities.However, cyclisation in the case of phenylhydrazine 2 occurs to the cyano group (thus producing a 5-amino-4-acetylpyrazole) while with pentafluorophenylhydrazine 3 the cyclisation is to the acetyl group (producing the 5-methyl-4-cyanopyrazole derivative 7e) 5 .Another interesting fact is that, in the case of the reaction of 1e with 3 we were able to isolate the intermediate enehydrazine 4e and on the basis of 1 H NMR analysis (two different doublets) to establish that the first nucleophilic attack was through the terminal (primary) amino group of the pentafluorophenylhydrazine 3 (route a) and therefore the succeeding cyclisation should produce 6 or 7 and not 8 or 9.These facts motivated us to study the reactivity of pentafluorophenylhydrazine with various enol ethers 1 displaying moderate reactivity. 14nder the chosen reaction conditions (reflux in methanol or ethanol for the appropriate time) we were never able to detect any other product than that of nucleophilic substitution through the terminal (primary) amino group (no route b and subsequent routes c,d in Scheme 1).In some cases we were able to isolate intermediate 4 (4d and 4c) confirming nucleophilic substitution at the terminal amino group (4d was obtained in almost quantitative yield, 4c in only 64% yield).The lower yield of 4c is perhaps due to the lower boiling temperature of methanol (Table 2).
The acyclic structural pattern of 4c,d was easily recognized from the 1 H NMR spectra -the methine proton signal at position 3 was split by a 3 JHH coupling with a neighboring proton from one of the present N-H pairs.The peak shapes excluded the possibility that the reactions proceeded via route b yielding analogues 5c,d (see Scheme 1).The narrow peaks of the N-H proton signals (related to their sufficiently slow exchange rate and ordinary T2* relaxation times) enabled the observation of heteronuclear correlations with nearby 13 C nuclei ( 2 JCH and 3 JCH), thus explicitly proving the linkage between the perfluorinated and non-fluorinated part of the molecules.The methyl ester and ethyl ester moieties in 4c and 4d, respectively, were also proved by the analysis of HMBC correlations (Table 5).
Cyclisation of (un)isolated intermediate in the reaction mixture can take place through the same functional group in the case of symmetrically-substituted derivatives 4a-d (X=Y) and thus only a single product can arise.In a successful cyclisation giving products 6c, 6d we tried to enhance the nucleophilicity of the secondary amino group by addition of sodium ethoxide in ethanol under reflux, but we isolated only traces of the desired products.If we used potassium carbonate in boiling water the product was isolated in 22-26% yields.Heating in toluene, xylene or DMSO led to decomposition of the starting compound and no product was detected.In the case of differently substituted derivatives 4e-i enhydrazine intermediates can exist as geometrical isomers which could not be isolated due to the low energy barrier of their isomerization. 6Upon thermal cyclisation, these enhydrazines gave a single product: in the case of cyano-substituted enehydrazines (4a,f,g) only the corresponding 5-aminopyrazoles were formed.If no cyano group is present, cyclisation to the acetyl group is observed and 5-methylpyrazoles were obtained (Scheme 2).6a,f,g 4 6b,h,i 6a (R=CN), 6f (R=COOMe), 6g (R=COOEt) 6b (R=Ac), 6h (R=COOMe), 6i (R=COOEt) Scheme 2. Cyclisation of arylaminoenhydrazines 4.
Non-cyclised intermediates (4c,d) have slightly lower melting points in comparison with their cyclized products (6c,d).In the compounds 4c,d the melting point of the corresponding methyl ester has higher melting point in comparison with ethyl ester.This tendency is reversed for cyclic pyrazoles 6h/i and 6f/g.HRMS spectra were used instead of elemental analysis (Table 2).Typically, electrospray ionization registered the protonated cation of the molecule.Differences between measured m/z and calculated values are stated in ppm units.In the IR spectra (Table 3) only selected characteristic vibrations are presented: for 6a of the cyano group (2228 cm -1 ), for 6a,f,g of the amino groups the presence of which confirms cyclisation in these cases through cyano groups and having intramolecular hydrogen bond between hydrogen of the amino group and oxygen atom of the carbonyl of the alkoxycarbonyl group for 6f,g evident from the frequency of the carbonyl group (comparison 6f,g with 6h,i) (Figure 1).Ester groups without this hydrogen bond (derivatives 6h,i) had peaks about 5 cm -1 higher frequency due to weak hydrogen bond of the previous compounds 6f,g.Methyl esters had these frequencies higher than ethyl esters.Absorption bands corresponding to C-F were found between 1000 -1400 cm -1 .Amino tautomers of 6a,e,f,g were indicated with no evidence for the tautomeric imino form. 6UV spectra showed absorption bands between 220 -240 nm belonging to pentafluorophenyl and pyrazole units which are not coplanar but twisted out of plane.Initially the identification of 6a-h was hampered by the lack of evidence about the precise order of the pyrazole ring substituents, as well as the unavailability of heteronuclear correlations between the pyrazole and the pentafluorophenyl group.Thanks to the good solubility of 3-methylpyrazoles 6b,h,i in chloroform it was possible to prepare concentrated samples for measurements of uncommon spectra.The measurement and analysis of an INADEQUATE spectrum from a sample of 6b offered an unambiguous proof of the reaction path c.Generally the presence of five fluorine atoms on the pentafluorophenyl group resulted in the formation of complicated multiplet patterns in the standard proton-decoupled 13 C NMR spectra (Tables 4, 5).The 1 JCF and n JCF couplings causing these multiplets were exploited in order to obtain 19 F-heterocorrelated 2D NMR spectra -these served for the exact assignment of the atoms in the perfluorinated group and for the precise extraction of chemical shifts in cases when 19 F-decoupled 13 C NMR spectra were not measured.It was noted that the pyrazoles 6a-h provide 19 F NMR spectra with a characteristic motif, in which the fluorine in the paraposition resonates at a higher frequency than the fluorine in the meta-position.The cyclisation has a dramatic effect on the 19 F NMR shifts as the acyclic derivatives 4c,d and similar acyclic derivatives yield 19 F NMR spectra with a shift of the para-fluorine signal towards lower resonance frequencies.This detailed analysis revealed that 19 F NMR is an accessible (especially when a NMR probe with a double-tuned high frequency channel is available) and seemingly reliable indicator in the reaction pathway of pentafluorophenyl derivatives described in Scheme 1.

Conclusions
Quantum chemical calculations confirm that the reactivity the two amine groups in pentafluorophenylhydrazine is not reversed in comparison with those in phenylhydrazine and there is no difference in principle between them.In some cases we were able to isolate an intermediate thus confirming this.

Experimental Section
General.All NMR spectra were obtained using a Varian VNMRS 600 MHz spectrometer (operating frequencies 599.76 MHz ( 1 H), 150.83 MHz ( 13 C) and 564.25 MHz ( 19 F)) equipped by an inverse triple resonance probe and a standard tuneable X/H probe with the possibility to tune the high frequency channel to the resonance frequency of 19 F. These spectra include standard 1 H, 13 C, 19 F, 19 F-decoupled 13 C spectra (bridged WALTZ16 decoupling), non-uniform sampled HSQC spectra with gradient coherence selection, HMBC with gradient coherence selection, 19 F- 13 C HSQC ( 1 JCF set to 250 Hz) spectra and 19 F- 13 C HMBC spectra with gradient coherence selection ( n JCF set to 15 Hz) and one 13 C-13 C INADEQUATE spectrum.Tetramethylsilane was used for the calculation of the 1 H and 13 C chemical shift scales and correctly referenced using the (residual) solvent signals (2.50 and 39.52 ppm for DMSO and 7.26 and 77.00 ppm for chloroform).CFCl3 was used for the calculation of the 19 F chemical shift scale; in order to correctly reference the 19 F chemical shift scale an automatic referencing mechanism exploiting the 2 H signal of the deuterated solvent was used.ATR IR spectra were recorded on Perkin-Elmer FT-IR spectrometer Spectrum Two UTa (ZnSe).UV-vis spectra were measured on two-beam UV-vis spectrometer Specord® 250 Plus (Analytik Jena).HRMS data were recorded on high resolution mass spectrometer Orbitrap Elite (Thermo Scientific) with resolution 240 000 for m/z 200.For ionization of samples we used electrospray with voltage set at 4.0kV and sheath gas flow at 5 unit.Melting points were measured on Boetius micro hot stage and are uncorrected.CC were performed using 60 μm silicagel, TLC by UV 256 plates from Merck.

General procedure for reaction of pentafluorophenylhydrazine with enol ethers
The pentafluorophenylhydrazine 3 and corresponding enol ether 1a-i were mixed (1:1) in ethanol or methanol (corresponding to the alkoxy group of the enol ether (R in Scheme 1), 10 mL per 10 mmol) and heated under reflux 2-6 hour (TLC control).The reaction mixture was cooled, solvent evaporated and an appropriate method used for separating the product.Products 6b, 6h, 4d were separated immediately after cooling of the reaction mixture by filtration followed by recrystallization from an appropriate solvent to obtain analytically pure compounds.Analytical data for all obtained compounds are summarized in Table 2 and IR and UV spectra in Table 3, NMR data in Tables 4, 5 .Preparation of 4-methoxycarbonylpyrazol-3-one (6c) and 4-ethoxycarbonylpyrazol-3-one (6d).The pentafluorophenylhydrazine 3 and corresponding enol ether 1c/d were mixed (1:1) and heated in water (50 mL) with an equivalent of potassium carbonate at reflux for two hours (TLC).The mixture was cooled, washed with ethyl acetate and then acidified to pH 2 (5N hydrochloric acid).The resultant precipitate was filtered off, washed with water and dried under vacuum.The obtained sample was pure enough for analysis.

Table 3 .
IR and UV spectra of the compounds 4c,4d and 6a-d,f-i