Substitution of a Fluorine Atom in Perfluorobenzonitrile by a Lithiated

Mar 30, 2017 - Vasily Romanov , Irina Bagryanskaya , Dmitry Gorbunov , Nina Gritsan ... Pavel V. Petunin , Ekaterina A. Martynko , Marina E. Trusova ,...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/joc

Substitution of a Fluorine Atom in Perfluorobenzonitrile by a Lithiated Nitronyl Nitroxide Evgeny V. Tretyakov,*,†,‡ Pavel A. Fedyushin,† Elena V. Panteleeva,†,‡ Dmitri V. Stass,‡,§ Irina Yu. Bagryanskaya,†,‡ Irina V. Beregovaya,† and Artem S. Bogomyakov∥ †

N. N. Vorozhtsov Institute of Organic Chemistry, 9 Ac. Lavrentiev Avenue, Novosibirsk 630090, Russia Novosibirsk State University, 2 Pirogova Str., Novosibirsk 630090, Russia § Institute of Chemical Kinetics and Combustion, 3 Institutskaya Str., Novosibirsk 630090, Russia ∥ International Tomography Center, 3a Institutskaya Str., Novosibirsk 630090, Russia ‡

S Supporting Information *

ABSTRACT: A 4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazole-3-oxide-1-oxyl (1) lithium derivative was found to react with perfluorobenzonitrile (2) substituting its para-fluorine atom to form 2-(4-cyanotetrafluorophenyl)4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazol-3-oxide-1-oxyl (3), a new nitronyl nitroxide containing a multifunctional framework of strong electronwithdrawing nature. This result shows the possibility of obtaining multifunctional nitronyl nitroxides via the interaction of paramagnetic lithium derivatives as C-nucleophiles with polyfluoroarenes activated for nucleophilic substitution. The reaction regioselectivity is supported by the data of quantum-chemical calculations, which also show that the reaction follows a concerted pathway without formation of an intermediate. Reduction of nitronyl nitroxide 3 in system NaNO2−AcOH yielded corresponding iminonitroxide 4. Characterization of persistent radicals 3 and 4 obtained by the SNF synthetic strategy includes X-ray crystal structures, electron spin resonance data, and static magneticsusceptibility measurements. X-ray diffraction analysis of both nitronyl nitroxide and iminonitroxide revealed a complete match of the parameters of their crystal lattices.

■

INTRODUCTION For many years, polyfunctional nitronyl nitroxides and iminonitroxides have been in demand in various fields of research, and this demand contributed to the development of synthetic chemistry related to this class of compounds.1−3 There has also been a feedback effect: the discovery of new approaches to nitronyl and iminonitroxides enabled investigators to create paramagnetics with a previously unknown combination of functional groups thereby stimulating new ideas in research.4 Thus, we recently developed a new method for producing nitronyl nitroxides that involves nucleophilic substitution of a hydrogen atom (SNH) in a series of azines and their N-oxides by the action of a paramagnetic lithium derivative (1-Li) of radical 1. This approach paved the way for the synthesis of previously inaccessible mono- and diradicals as well as their heterospin complexes with a new type of structural organization.5−10 In addition to these results, the SNAr substitution of fluorine11 in electron-deficient polyfluorinated aromatics (nitrobenzene, pyridine, and phthalo- and benzonitrile) by means of a paramagnetic O-nucleophilethe alkoxide anion generated from the 4-hydroxy-2,2,6,6-tetramethylpiperidine oxide (TEMPOL)has been reported as a facile route to the polynitroxides showing strong electron exchange between nitroxide groups.12 With the aim to develop new methods for nitronyl nitroxide modification based on SNAr-type substitu© 2017 American Chemical Society

tion, we tested whether it is possible to use 1-Li as a Cnucleophile to replace a fluorine atom in polyfluorinated benzonitriles. In general, it is known that C-nucleophiles can attack the nitrile group of fluorobenzonitriles and can substitute a fluorine atom in the aromatic ring. Accordingly, reactions of the cyanomethyl anion with mono- and pentafluorobenzonitrile produce the corresponding β-aminocinnamonitriles, with yields varying in a wide range depending on the conditions of the Canion generation and the number and positions of fluorine atoms in the benzonitrile. For example, interaction of 4-fluoroor pentafluorobenzonitrile with the anion generated by sodium amide from acetonitrile in liquid ammonia produces βaminocinnamonitriles with yields of 68% and 35%, respectively.13 When sodium tert-butoxide in diethyl ether is used to generate the cyanmethyl anion, the yields of β-aminocinnamonitriles are low: 19% (for 2-fluorobenzonitrile), 24% (for 3-fluorobenzonitrile), and 20% (for 4-fluorobenzonitrile).14 The highest yield of β-aminocinnamonitrile (98%) was obtained after the interaction of 4-fluorobenzonitrile with the anion generated by potassium tert-butoxide from benzene.15 It should be noted that none of the above studies mentioned Received: January 20, 2017 Published: March 30, 2017 4179

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185

Article

The Journal of Organic Chemistry products of fluorine substitution, formation of which was very likely in an activated aromatic ring. Nevertheless, the SNAr reaction of 2,3,4-trifluorobenzonitrile with the diethyl malonate anion generated by sodium hydride in THF was successfully used for quantitative synthesis of ethyl 2-(cyanodifluorophenyl)malonate (yield 98%, ratio of o-/p-regioisomers 1.35:1).16 Taken together, the available data allowed us to assume that an increase in the number of fluorine atoms in benzonitrile promotes fluorine substitution by the action of a C-nucleophile. Here we report the interaction of perfluorobenzonitrile 2 with nitronyl nitroxide lithium derivative 1-Li (in THF) producing 2-(4-cyano-2,3,5,6-tetrafluorophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazol-3-oxide-1-oxyle 3 with a moderate yield (26−34%). Because this transformation can serve as a clue to the development of the synthesis of nitronyl nitroxides and iminonitroxides bearing polyfluorinated aromatics substituted with electron-withdrawing groups, we found it necessary to describe the synthesis of not only nitroxide 3 but also of its deoxygenated derivative 4 as well as X-ray diffraction (XRD) structures, results of electron spin resonance (ESR) analysis, and temperature dependences of the magnetic susceptibility of polycrystalline samples of these compounds.

According to the calculations, there are two different types of prereaction complexes (Figure 1) resulting from the interaction of 1-Li with the cyano group of nitrile 2 (Li+ ion with N atom of the CN group) as well as with the para−F-C fragment of molecule 2. The most stable complex [RCortho+CN] is a starting point of two processes, namely the substitution of the fluorine atom at the ortho-position providing radical 3′ and 1-Li addition to the CN triple bond with adduct A formation. The second complex [RCpara] precedes the substitution of the fluorine atom at the para-position in relation to the CN group leading to target product 3, and it is higher in terms of energy than the first one (8.1 kcal/mol in the gas phase and 4.4 kcal/ mol in THF). The relative energy of the complexes and the calculated activation energies (Table 1) predict preferability for the 1-Li addition in comparison with fluorine atom substitution, but the addition is a reversible process. The major product of interaction of 2 and 1-Li is determined by the irreversible process of fluorine atom substitution. The calculated barrier heights and heat values of reactions comply with the experimental fact of exclusive substitution of the fluorine atom at the para-position. Furthermore, the calculations show that F atom substitution in nitrile 2 by 1-Li proceeds without formation of an anionic intermediate like a Meisenheimer complex. In other words, the nucleophilic attack and elimination take place simultaneously, crossing a single barrier through the transition state. The possibility of such a mechanism (called concerted or one-step mechanism) was first postulated by Bunnett.17 Later, in their seminal work, Glukhovtsev et al. conducted an ab Initio computational study of aromatic nucleophilic substitution of C6H5X with Nu = F−, Cl−, Br−, I− and (O2N)nC6H5−nCl with Nu = Cl−. It was revealed that the concerted mechanism is observed in cases when the formation of the second C−Nu bond does not significantly diminish the aromatic nature of the benzene moiety.18 Moreover, it was shown19 that the SNAr reaction of O2NC6H4X and (O2N)2C6H4F can change from a concerted mechanism in a gas phase to a stepwise pathway in liquid ammonia, and the halide elimination step is assisted by proton transfer from the reacting NH3 to leaving X−. On the basis of this knowledge, let us consider in detail the reaction of nitrile 2 with 1-Li via the Cpara pathway. The first step is formation of prereaction complex RCpara in which the Li ion is located almost in the plane of the irregular triangle with vertices occupied by the nitroxide oxygen atom and two fluorine atoms at para- and meta-positions. The distance between the lithium ion and oxygen atom is the shortest (1.86 Å), and notably, it is very close to the distance Li···O in solid LiOH or LiOMe (1.90−1.98 Å).20 On the way to TSpara, the Li···O distance shortens slightly (Δ = 0.08 Å) as compared with significant shortening of Li···Fpara and Li···Fmeta distances, respectively, from 2.41 to 1.96 Å and from 2.25 to 2.03 Å, which are smaller than those in the lithium fluoride ionic crystal (2.013 Å).20 Simultaneously, the distance CONCNO...Cpara also significantly decreases to 2.17 Å (Δ = 0.98 Å), while the F−Cpara bond increases from 1.34 to 1.40 Å and bends out of the mean aromatic ring plane with the angle of 27°. It is noteworthy that this angle is considerably smaller than the angle between the CONCNO−Cpara bond and aromatic ring (63°). This means that in TSpara, the CONCNO−Cpara bond is only partly formed, while the C−F bond is almost intact. Additionally, an estimation of the benzene ring aromaticity in TSpara using the A index18 gives a value of ∼0.980, which is only slightly less than that in 2 (A =

■

RESULTS AND DISCUSSION To determine the effect of the activation degree of the fluorinated aromatic system on the possibility of its involvement in a reaction with paramagnetic derivative 1-Li, we tested a set of benzonitriles differing in the number of fluorine atoms in the ring (2,4-difluoro-, 2,4,6-trifluoro-, and pentafluorobenzonitrile 2). The choice of a substrate was based on the location of its fluorine atoms at positions activated for SNAr. The experiments showed that only perfluorobenzonitrile 2 could react with 1-Li leading to para-fluorine substitution and producing nitronyl nitroxide 3 (Scheme 1). Subsequent deoxygenation of resulting 3 by means of the AcOH/NaNO2 system produced iminonitroxide 4 with a yield up to 93%. Scheme 1. Synthesis of Nitronyl Nitroxide 3 and Iminonitroxide 4

The regularities observed experimentally for the interaction of 2 with 1-Li including the process regioselectivity can be realized on the basis of quantum-chemical calculations of the potential energy surface (PES) sections along the expected reaction coordinates. We considered the substitution of orthoor para-fluorine atoms in the pentafluorobenzonitrile aromatic ring and addition of 1-Li to the carbon atom of the cyano group. These reaction pathways are schematically depicted in Figure 1; the calculation results are given in Table 1. 4180

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185

Article

The Journal of Organic Chemistry

Figure 1. Reaction paths considered for the interaction of lithiated nitronyl nitroxide 1-Li with perfluorobenzonitrile 2 at the CAM-B3LYP/6-31+G* level of calculations (RCortho+CN and RCpara denote prereaction complexes; TSpara, TSortho, and TSCN are abbreviations of the corresponding transition states). Hydrogen atoms of the methyl groups are omitted. The numbers are interatomic distances in Å.

Table 1. Total Energy Values for the Initial Reagent Complexes (ERC, arbitrary atomic units), the Heights of Energy Barriers for Nucleophilic Substitution or Addition (ΔEA, kcal/mol), and the Heat Values of the Reactions (ΔE, kcal/mol) gas reaction position Cortho− CCN Cpara−

ERC −1361.52249 −1361.50955

Scheme 2. Plausible Pathway of Diradical 5 Formation

PCM (THF)

ΔEAa

ΔEb

ERC

17.4 9.8 12.4

−15.2 −6.8 −24.9

−1361.53679 −1361.52982

ΔEA c

8.8 18.0

ΔEb −38.7 −12.1 −45.5

a

The heights of energy barriers were estimated as total energy differences between the TSes and the potential energy surface (PES) minima corresponding to the initial RCs. bThe heat values of reactions were estimated as total energy differences between the PES minimum corresponding to the initial RCs and the reaction products. cThe value was not calculated.

formation of 5 at the first step should involve 1-Li addition to the CN bond with formation of adduct A, which subsequently reacts with 2 thus producing substituted imine B. The latter can then attach another 1-Li molecule and yield diradical 5 of unusual structure. It would be relevant to mention that the approach to the construction of heterocyclic ring-fused systems21 implies fluorine atom substitution as well as the addition of nitrogen-centered nucleophiles at the cyano group of nitrile 2. Well-shaped single crystals were obtained after slow evaporation of the solutions of 3 and 4 in a mixture of CH2Cl2 with n-heptane at ∼5 °C. XRD analysis revealed that both paramagnetics crystallize isostructurally in the orthorhombic Ibca space group. In the crystals, the molecules are aligned along the C2 axis; therefore, only one-half of the molecules are independent. In compound 4, atom O1 has weight 50% and is disordered over two positions; this state of affairs does not allow us to determine the real values of iminonitroxide bond lengths (Figures 2 and 3). The X-ray crystal structure of diradical 5 is shown in Figure 4. By its structure, 5 belongs to an extremely scarce series of

0.994). All the data show that the SNAr reaction of 2 with 1-Li follows a concerted pathway, for the following probable reasons: (1) in early TSpara, the CONCNO···Cpara bond is loose and does not significantly disturb the aromatic nature of the benzene moiety, (2) the fluoride elimination is assisted by the lithium ion at an early stage when the CONCNO···Cpara bond is only partly formed. Quantum-chemical calculations showed that the reaction of nitrile 2 with 1-Li via the CCN pathway has the lowest kinetic barrier. This mechanism could effectively compete with the Cpara pathway if it were not reversible. Nevertheless, we succeeded in obtaining the evidence of CCN competing directions of the nucleophilic attacks of 1-Li on 2. A minor amount (several crystals) of 2,2′-([(4-cyano-2,3,5,6-tetrafluorophenyl)amino]-[perfluorophenyl]methylene)bis(4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazole 3-oxide 1-oxyl) (5) was isolated and structurally characterized (Scheme 2). Apparently, 4181

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185

Article

The Journal of Organic Chemistry

the sum of van der Waals radii of oxygen atoms (ca. 3.0 Å). There are intramolecular H-bonds between the NH group and nitroxide oxygen atoms O1 and O3. Regarding the magnetically important intermolecular contacts, there are two contacts between methyl hydrogen H9 and oxygen O1 atoms (2.685 Å). Electronic structures of radicals 3 and 4 were first studied by electronic absorption spectroscopy. The spectrum of 3 contains a very weak absorption maximum (λmax) in the visible range at 559 and 561 nm (solution and solid state), corresponding to n−π* transition in the ONCNO moiety although arylsubstituted nitronyl nitroxides usually show two bands in the 600−680 nm range.3,23 Therefore, a strong hypsochromic shift of the long-wave absorption band of 3 is probably caused by the violation of π-conjugation between the ONCNO moiety and aromatic ring because of their considerable twist relative to each other (α = 66.6°). A similar effect was observed previously in the 4,8-dimethoxy-5-formyl-naphth-1-yl-substituted nitronyl nitroxide, where angle α reached 81°, which also shifted the band of n−π* transition to 566.5 nm.24 Other absorption patterns include intense bands with maxima at 303 and 388 nm corresponding to the π−π* electronic transitions of radical 3. The UV−vis spectrum of 4 contains a very weak band in the visible range with λmax = 398 nm (both solution and solid state), corresponding to n−π* transition in the NCNO moiety, whereas iminonitroxides are commonly characterized by an absorption band in the region of 440−500 nm. The considerable hypsochromic shift of this band in the spectrum of 4 is also obviously due to a violation of π-coupling because of the large (in fact the largest so far25) angle (α = 66.6°) between the NCNO and aromatic planes. Other intense absorption peaks with maxima at 232 and 268 nm correspond to the π−π* electronic transitions of the iminonitroxide. ESR spectroscopy provided further information about the electronic structure of 3 and 4 as well as the intramolecular hyperfine interactions in these compounds. Figures 5 and 6 show the entire ESR spectra and zoomed second low-field lines for the radicals to demonstrate the finer structure of the main lines. The isotropic g-values were determined using solid 2,2diphenyl-1-picrylhydrazyl (DPPH) as a standard and were found to be typical for nitronyl and iminonitroxides at giso of 2.0066 and 2.0060 for 3 and 4, respectively. The spectra were simulated with the following sets of parameters: for 3, A(2N) = 0.703 mT, A(12H) = 0.017 mT, A(2F) = 0.074 mT, A(2F) = 0.020 mT, A(N) = 0.005 mT; for 4, A(N1) = 0.842 mT, A(N2) = 0.428 mT, A(12H) = 0.017 mT, A(2F) = 0.045 mT, A(2F) = 0.020 mT, A(N) = 0.005 mT. The accuracy of determining hyperfine coupling constants and g-values is estimated as 0.005 mT and 0.0001, respectively. The minor coupling of 0.005 mT at the nitrile nitrogen in both radicalsalthough not explicitly resolved in the spectrum and equal to the stated accuracyis necessary to obtain the correct overall envelopes of the lines, which are quite sensitive to its magnitude and are thus rather reliable. Characterization of the magnetic properties of nitroxides 3 (Figure 7) and 4 (Figure S12) indicated the absence of noticeable exchange interactions. The μeff values are 1.78 and 1.69 μB at 300 K for 3 and 4, respectively, and do not change when the temperature is lowered down to 20 K. Below 20 K, the μeff values slightly decrease to 1.41 and 1.53 μB at 2 K. The 1/χ(T) dependences obey the Curie−Weiss law, with the best fit parameters C and Θ equal to 0.395 K·cm3/mol and −1.7 K, respectively, for 3 and 0.355 K·cm3/mol and −0.6 K for 4. The Curie constant values C are in good agreement with the

Figure 2. Structure of molecule 3 in crystal and the numbering scheme (the dependent parts are indicated with the letter A). Selected bond lengths (Å) and bond angles (deg): O1−N1 1.276(3), N1−C1 1.339(3), N1−C2 1.495(4), C2−C2a 1.573(4), O1−N1−C1 125.6(3), O1−N−C2 122.4(2), C1−N1−C2 111.9(2), N1−C1− N1a 110.4(3), N1−C2−C2a 101.2(2).

Figure 3. Structure of molecule 4 in crystal and the numbering scheme (the dependent parts are indicated with the letter A). Selected bond lengths (Å) and bond angles (deg): {O1−N1 1.255(1) and N1−C1 1.319(2) are average values because of disordering}, N1−C2 1.488(2), C2−C2a 1.568(3); O1−N1−C1 123.7(2), O1−N−C2 126.9(2), C1− N1−C2 109.3(2), N1−C1−N1a 114.6(2), N1−C2−C2a 101.9(1).

Figure 4. Structure of molecule 5 in crystal and the numbering scheme. Selected bond lengths (Å) and bond angles (deg): O1−N1 1.279(3), N2−O2 1.272(3), O3−N3 1.267(3), O4−N4 1.281(3), N5−C7 1.362(3), N5−C6 1.461(3), N1−C13 1.340(3), N1−C14 1.498(4), N2−C13 1.340(3), N2−C15 1.509(4), N4−C23 1.335(3), N4−C24 1.500(4), N3−C23 1.343(3), C14−C15 1.563(4), C24− C25 1.533(5), C13−N1−C14 112.6(2), C13−N2−C15 111.5(2), C23−N4−C24 112.2(2), C23−N3−C25 110.7(2).

sp3-C spaced diradicals.22 The dihedral angle between the planes of paramagnetic moieties (ON−C13−NO and ON− C23−NO) is ca. 64°. The shortest distance between the two oxygen atoms (O1 and O3) is 3.068 Å, which is very close to 4182

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185

Article

The Journal of Organic Chemistry

Figure 7. Experimental μeff(T) (●) and 1/χ(T) (■) dependences for 3.

theoretical spin-only value 0.375 K·cm3/mol for a monoradical with spin S = 1/2 and g = 2. A small decrease in μeff below 20 K and negative values of Weiss constants Θ are caused by weak antiferromagnetic exchange interactions.

■

CONCLUSION We demonstrated the possibility of obtaining multifunctionalized nitronyl nitroxides (and related iminoxides) by substituting a fluorine atom in highly activated aromatic substrates by means of a C-nucleophile generated from a paramagnetic source. This strategy can be useful during planning of the retrosynthesis of nitroxides bearing strong electron-withdrawing groups. Such paramagnetics arouse much interest in the field of molecular design of magnets; they can be used for creation of paramagnetic chemical sensors and organic rechargeable batteries.

Figure 5. Experimental ESR spectra and zoomed second low-field line (black curves) of 3 recorded at room temperature in a degassed dilute toluene solution and its modeling performed (red curves) in Winsim v. 0.96, as described elsewhere.26

■

EXPERIMENTAL SECTION

Reagents and General Methods. 4,4,5,5-Tetramethyl-4,5dihydro-1H-imidazol-3-oxide-1-oxyl was synthesized as reported elsewhere.8 THF was freshly distilled over benzophenone sodium ketyl. Other chemicals were of the highest purity commercially available and were used as received. The reactions were monitored by thin-layer chromatography (TLC) on silica gel 60 F254 aluminum sheets. The chromatography was carried out in silica gel (0.063−0.200 mm) for column chromatography. Fourier transform infrared (FT-IR) spectra were acquired in KBr pellets on a Bruker Vector-22 spectrometer. Elemental analyses were performed using a Euro EA 3000 elemental analyzer. 2-(4-Cyano-2,3,5,6-tetrafluorophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazol-3-oxide-1-oxyle (3). A 1.0 M solution of LiN(SiMe3)2 (2.2 mL, 2.2 mmol) in THF was added at −90 °C into a vigorously stirred solution of 4,4,5,5-tetramethyl-4,5-dihydro-1Himidazol-3-oxide-1-oxyl (314 mg, 2.0 mmol) in THF (30 mL) under argon. The reaction mixture was stirred at −90 °C for 30 min. Then, the solution of pentafluorobenzonitrile 2 (425 mg, 2.2 mmol) in THF (5 mL) was added at −90 °C in an argon atmosphere, stirring was continued, and the reaction was monitored by TLC (Silufol F254, EtOAc as eluent). After 4 h, TLC changes ceased, the cooling was stopped, and the reaction mixture was allowed to warm up to room temperature and brought into contact with the atmosphere. Flash chromatography (SiO2, column 3 × 4 cm, EtOAc as eluent) yielded a dark purple solid mixture (876 mg) after solvent removal under reduced pressure at room temperature. The resulting solid mixture was separated by column chromatography (SiO2, column 3 × 20 cm, CH2Cl2 as eluent), which produced a blue-violet fraction of radical 3, which was concentrated under reduced pressure to a volume of ∼5 mL. n-Heptane (5 mL) was added, and the mixture was incubated for

Figure 6. Experimental ESR spectra and zoomed second low-field line (black curves) of 4 recorded at room temperature in a degassed dilute toluene solution and its modeling (red curves).

4183

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185

Article

The Journal of Organic Chemistry ∼60 h at 0−5 °C for slow crystallization of radical 3. Yield 225 mg (34%); dark purple crystals; mp 195.7−196.0 °C (uncorrected); IR (KBr) ṽmax, cm−1: 436, 476, 538, 661, 712, 872, 914, 968, 989, 1030, 1070, 1142, 1176, 1265, 1300, 1373, 1431, 1452, 1466, 1504, 2247 (CN), 2937, 2989, 3001, 2437. UV (C2H5OH) λmax, nm (lg ε): 559 (2.50), 388 (3.65), 303 (4.26), 227 (3.19). UV (KBr) λmax, nm (lg ε): 561, 402, 312, 225. Anal. Calcd for C14H12F4N3O2: C, 50.91; H, 3.66; F, 23.01; N, 12.72. Found: C, 50.75; H, 3.73; F, 23.02; N, 12.62. 4-(4,4,5,5-Tetramethyl-1-oxyl-4,5-dihydro-1H-imidazol-2-yl)2,3,5,6-tetrafluorobenzonitrile (4). A mixture of radical 3 (165 mg, 0.50 mmol), NaNO2 (42 mg, 0.60 mmol), CHCl3 (14 mL), AcOH (5.00 mL), and water (0.8 mL) was stirred until starting radical 1 was completely consumed (after 30 min according to TLC). The reaction mixture was neutralized with a saturated solution of NaHCO3, then MnO2 (10 mg, 0.12 mmol) was added. The organic layer was separated, dried over anhydrous MgSO4, filtered through a layer of Al2O3 (2 × 5 cm, EtOAc as eluent), and evaporated under reduced pressure. The residue was recrystallized from a mixture of n-heptane with CH2Cl2. Yield 146 mg (93%); orange crystals; mp 130.9−131.6 °C (uncorrected). IR (KBr) ṽmax, cm−1: 478, 575, 625, 658, 685, 733, 744, 789, 868, 908, 943, 989, 1061, 1146, 1211,1246, 1263, 1333, 1369, 1379, 1452, 1502, 1583, 1643, 1668, 2249, 2648, 2688, 2874, 2933, 2945, 2993, 3198, 3225, 3354, 3456. UV (C2H5OH) λmax, nm (lg ε): 398 (3.00), 268 (4.19), 232 (4.23). UV (KBr) λmax, nm (lg ε): 398, 297, 228. Anal. Calcd for C14H12F4N3O: C, 53.51; H, 3.85; F, 24.18; N, 13.37. Found: C, 53.49; H, 3.61; F, 24.15; N, 13.07. X-ray Structural Analysis. XRD data were obtained on a Bruker Kappa Apex II CCD diffractometer using φ, ω scans of narrow (0.5°) frames with Mo Kα radiation (λ = 0.71073 Å) and a graphite monochromator at 296 K. The structures of 3 and 4 were solved by direct methods and refined by the full-matrix least-squares method against all F2 in anisotropic approximation using the SHELX-97 software suite.27 Positions of the H atoms were calculated by means of the riding model. Absorption corrections were applied empirically using the SADABS software. 3. C14H12F4N3O2, FW = 330.27, T = 296 K, orthorhombic, Ibca, a = 10. 4901(1), b = 10. 6393(1), c = 26.130(4) Å, V = 2916.3(6) Å3, Z = 8, Dcalc = 1.504 g/cm3, μ(Mo Kα) = 0.137 mm−1, a total of 14912 reflections (θmax = 25.10°), 1202 unique reflections (Rint = 0.0841), 742 (F > 4σF), 109 parameters. Goof = 1.090, R1 = 0.0654, wR2 = 0.1415 (I > 2σI), R1 = 0.1087, wR2 = 0.1699 (all data), max/min diff. peak 0.421/−0.481 eÅ−3. 4. C14H12F4N3O, FW = 314.27, T = 296 K, orthorhombic, Ibca, a = 10.3528(3), b = 10.8047(3), c = 26.1688(6) Å, V = 2927.2(1) Å3, Z = 8, Dcalc = 1.426 g/cm3, μ(Mo Kα) = 0.127 mm−1, a total of 16298 reflections (θmax = 26.03°), 1452 unique reflections (Rint = 0.0500), 1167 (F > 4σF), 109 parameters, 1 restraint. Goof = 1.032, R1 = 0.0565, wR2 = 0.1364 (I > 2σI), R1 = 0.0773, wR2 = 0.1687 (all data), max/min diff. peak 0.341/−0.391 eÅ−3. 5. C28H24F9N6O4, FW = 679.53, T = 296 K, triclinic, P-1, a = 10.0863(6), b = 12.4846(7), c = 13.1297(7) Å, λ = 76.259(2), β = 78.474(2), γ = 87.054(2)°, V = 1573.6(2) Å3, Z = 2, Dcalc = 1.434 g/ cm3, μ(Mo Kα) = 0.133 mm−1, a total of 28664 reflections (θmax = 25.74°), 5991 unique reflections (Rint = 0.0356), 4112 (F > 4σF), 431 parameters. Goof = 1.235, R1 = 0.0665, wR2 = 0.1749 (I > 2σI), R1 = 0.1234, wR2 = 0.2495 (all data), max/min diff. peak 0.793/−0.899 eÅ−3. ESR Spectroscopy. Continuous-wave ESR spectra were acquired on a Bruker EMX spectrometer at room temperature in dilute (ca. 10−6 M) toluene solutions degassed by means of repeated freeze− pump−thaw cycles. Modulation 0.01 mT@100 kHz, MW power 2 mW, the spectra were recorded in single slow (ca. 3 h) scans. Magnetic Measurements. Magnetic susceptibility of the polycrystalline samples was measured with a Quantum Design MPMSXL SQUID magnetometer in the temperature range 2 to 300 K with magnetic field of up to 5 kOe. None of the radicals showed any field dependence of molar magnetization at low temperatures. Diamagnetic corrections were made using the Pascal constants. The effective magnetic moment was calculated as μeff(T) = [(3k/ NAμB2)χT]1/2 ≈ (8χT)1/2.

Computational Details. Calculations of the PES sections along the reaction coordinates for the nucleophilic substitution of ortho- or para-fluorine atoms in the pentafluorobenzonitrile aromatic ring and for the addition of 1-Li to the carbon atom of the cyano group in nitrile 2 were performed at the CAM-B3LYP/6-31+G* calculation level. Stationary points on the PESes were located, and their types were determined by the normal vibrations analysis. Interrelations between TSes and the minima corresponding to the initial complexes of reagents and to imino anion A were determined by the intrinsic reaction coordinate (IRC) method. Because of the high computational cost, we did not follow all the IRC calculations leading to the complexes of the nucleophilic substitution products. We assumed the pairs (sums) of isolated 3 or 2-(2-cyano-3,4,5,6-tetrafluorophenyl)4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazol-3-oxide-1-oxyle (3′) and LiF molecule to be these products (Figure 1). The solvent effect was accounted for within the polarizable continuum model (PCM) using built-in parameters for THF. The untabulated value for the van der Waals radius of Li was assigned 1.60 Å. All calculations were done in the GAMESS package.28 The structure images were generated by means of the MOLDEN software.29

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.joc.7b00144. Crystal packings, IR and UV spectra of nitroxides 3 and 4, experimental μeff(T) and 1/χ(T) dependences for 4, computational results, and Cartesian coordinates (PDF) X-ray crystallographic data of compounds 3, 4, and 5 (CIF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] ORCID

Evgeny V. Tretyakov: 0000-0003-1540-7033 Artem S. Bogomyakov: 0000-0002-6918-5459 Notes

The authors declare no competing financial interest. Additional crystallographic data (CCDCs 1486802 for nitronyl nitroxide 3, 1486803 for iminonitroxide 4, and 1510989 for diradical 5) for this article may be accessed at no charge at the Cambridge Crystallographic Data Centre at http://www.ccdc. cam.ac.uk/conts/retrieving.html.

■

ACKNOWLEDGMENTS The authors thank the Federal Agency for Scientific Organizations (FASO Russia), the Siberian Branch of the Russian Academy of Sciences (project 0302-2015-0001), the Multi-Access Chemical Service Center, SB RAS, for spectral and analytical measurements.

■

REFERENCES

(1) Suzuki, S.; Okada, K. In Organic Redox Systems Synthesis, Properties, and Applications; Nishinaga, T., Ed.; Wiley: NJ, 2016; pp 269−285. (2) Stable Radicals: Fundamentals and Applied Aspects of Odd-Electron Compounds; Hicks, R. G., Ed.; John Wiley and Sons: Chichester, UK, 2010. (3) Tretyakov, E. V.; Ovcharenko, V. I. Russ. Chem. Rev. 2009, 78, 971−1012. (4) Tanimoto, R.; Suzuki, S.; Kozaki, M.; Okada, K. Chem. Lett. 2014, 43, 678−680. 4184

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185

Article

The Journal of Organic Chemistry (5) Weiss, R.; Kraut, N.; Hampel, F. J. J. Organomet. Chem. 2001, 617−618, 473−482. (6) Ovcharenko, V. I.; Chupakhin, O. N.; Kovalev, I. S.; Tretyakov, E. V.; Romanenko, G. V.; Stass, D. V. Russ. Chem. Bull. 2008, 57, 2227− 2229. (7) Chupakhin, O. N.; Utepova, I. A.; Varaksin, M. V.; Tretyakov, E. V.; Romanenko, G. V.; Stass, D. V.; Ovcharenko, V. I. J. Org. Chem. 2009, 74, 2870−2872. (8) Tretyakov, E. V.; Utepova, I. A.; Varaksin, M. V.; Tolstikov, S. E.; Romanenko, G. V.; Bogomyakov, A. S.; Stass, D. V.; Ovcharenko, V. I.; Chupakhin, O. N. Arkivoc 2011, viii, 76−98. (9) Varaksin, M. V.; Tretyakov, E. V.; Utepova, I. A.; Romanenko, G. V.; Bogomyakov, A. S.; Stass, D. V.; Sagdeev, R. Z.; Ovcharenko, V. I.; Chupakhin, O. N. Russ. Chem. Bull. 2012, 61, 1469−1473. (10) Kovalev, I. S.; Kopchuk, D. S.; Zyryanov, G. V.; Rusinov, V. L.; Chupakhin, O. N.; Charushin, V. N. Russ. Chem. Rev. 2015, 84, 1191− 1225. (11) Reviews: (a) Chambers, R. D.; Hall, C. W.; Hutchinson, J.; Millar, R. W. J. Chem. Soc., Perkin Trans. 1 1998, 1705−1714. (b) Chambers, R. D.; Hassan, M. A.; Hoskin, P. R.; Kenwright, A.; Richmond, P.; Sandford, G. J. Fluorine Chem. 2001, 111, 135−146. (c) Revesz, L.; Di Padova, F. E.; Buhl, T.; Feifel, R.; Gram, H.; Hiestand, P.; Manning, U.; Wolf, R.; Zimmerlin, A. G. Bioorg. Med. Chem. Lett. 2002, 12, 2109−2112. (d) Sasaki, S.; Tanabe, Y.; Yoshifuji, M. Bull. Chem. Soc. Jpn. 1999, 72, 563−572. (e) Transformations of Fluorinated Compounds. Houben-Weyl Methods of organic Chemistry; Baasner, B., Hagemann, H., Tatlow, J.-C., Eds.; Thieme: Stuttgart, NY, 1999; Vol. E10/2. (f) Shelke, N. B.; Ghorpade, R.; Pratap, A.; Tak, V.; Acharya, B. N. RSC Adv. 2015, 5, 31226−31230. (g) Chambers, R. D.; Hassan, M. A.; Hoskin, P. R.; Kenwright, A.; Richmond, P.; Sandford, G. J. Fluorine Chem. 2001, 111, 135−146. (12) Zeika, O.; Li, Y.; Jockusch, S.; Parkin, G.; Sattler, A.; Sattler, W.; Turro, N. J. Org. Lett. 2010, 12, 3696−3699. (13) Lang, S. A.; Cohen, E., Jr. J. Med. Chem. 1975, 18, 441−443. (14) Ridge, D. N.; Hanifin, J. W.; Harten, L. A.; Johnson, B. D.; Menschik, J.; Nicolau, G.; Sloboda, A. E.; Watts, D. E. J. Med. Chem. 1979, 22, 1385−1389. (15) Yamaguchi, Y.; Katsuyama, I.; Funabiki, K.; Matsui, M.; Shibata, K. J. Heterocycl. Chem. 1998, 35, 805−810. (16) Kreutter, K. D.; Lu, T.; Lee, L.; Giardino, E. C.; Patel, S.; Huang, H.; Xu; Fitzgerald, G. M.; Haertlein, B. J.; Mohan, V.; Crysler, C.; Eisennagel, S.; Dasgupta, M.; McMillan, M.; Spurlino, J. C.; Huebert, N. D.; Maryanoff, B. E.; Tomczuk, B. E.; Damiano, B. P.; Player, M. R. Bioorg. Med. Chem. Lett. 2008, 18, 2865−2870. (17) Bunnett, J. F. Tetrahedron 1993, 49, 4477−4484. (18) Glukhovtsev, M. N.; Bach, R. D.; Laiter, S. J. Org. Chem. 1997, 62, 4036−4046. (19) Moors, S. L. C.; Brigou, B.; Hertsen, D.; Pinter, B.; Geerlings, P.; Van Speybroeck, V.; Catak, S.; De Proft, F. J. Org. Chem. 2016, 81, 1635−1644. (20) FIZ/NIST, Inorganic Crystal Data Base; National Institute of Standard and Technology: Gaitersburg, 2015. (21) Cargill, M. R.; Linton, K. E.; Sandford, G.; Yufit, D. S.; Howard, J. A. K. Tetrahedron 2010, 66, 2356−2362. (22) (a) Matsumoto, K.; Oda, M.; Kozaki, M.; Sato, K.; Takui, T.; Okada, K. Tetrahedron Lett. 1998, 39, 6307−6310. (b) Suzuki, S.; Itoh, N.; Furuichi, K.; Kozaki, M.; Shiomi, D.; et al. Chem. Lett. 2011, 40, 22−24. (23) Ullman, E. F.; Osiecki, J. H.; Boocock, D. G. B.; Darcy, R. J. Am. Chem. Soc. 1972, 94, 7049−7059. (24) Ding, L.; Zhang, D. Q.; Zhang, B.; Zhu, D. B. Chin. Chem. Lett. 2000, 11, 749−752. (25) Cambridge Crystallographic Data Centre, v1.1.2. (26) Sviridenko, F. B.; Stass, D. V.; Kobzeva, T. V.; Tretyakov, E. V.; Klyatskaya, S. V.; Mshvidobadze, E. V.; Vasilevsky, S. F.; Molin, Yu. N. J. Am. Chem. Soc. 2004, 126, 2807−2819. (27) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2008, 64, 112−122.

(28) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. J.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347−1363. (29) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000, 14, 123−134.

4185

DOI: 10.1021/acs.joc.7b00144 J. Org. Chem. 2017, 82, 4179−4185