1,2-Diaryl(3-pyridyl)ethanone Oximes. Intermolecular Hydrogen Bonding Networks Revealed by X-ray Diffraction

The synthesis of a set of 1-aryl-2-aryl(3-pyridyl)ethanones 1-5 and the corresponding ketoximes 6-9 is reported. Structural studies of oximes 6, 7 and 9 were performed in solution using 1H-NMR and in the solid state by X-ray crystallography, providing evidence of H-bonding networks. The crystal packing was controlled by homomeric intermolecular oxime···oxime H-bond interactions for 6 and cooperative oxime···N(pyridyl) and CH/π interactions for 7 and 9.

In parallel, hydrogen bonds [6,7] are the main driving forces for oxime non-covalent interactions and oxime functionality can be stereoelectronically adjusted for precise directed assembly of homomeric intermolecular oxime···oxime and oxime···N-heterocycle molecular motifs [8][9][10]. Moreover, the CH/π hydrogen bonds [11] are weaker intermolecular hydrogen bond interactions that modulate the crystal packing of molecules and their self-assembly.  Distinct methodologies for the preparation of diarylethanones II could lead to compounds 1 and 2, whereas for 1-aryl-2-(3-pyridyl)ethanones 3-5, the synthetic protocols were limited by the presence of a 3-pyridyl moiety. By coupling of acid chlorides with organozinc bromides in the presence of CuCN/LiBr, compounds 1 and 2 were directly obtained, whereas ethanones 3-5 were prepared following a modified two-step Horner-Wittig reaction protocol.
The structural properties of oximes 6-9 are discussed with the data obtained in solution by 1 H-NMR and in the solid state by X-ray crystallography of compounds 6, 7 and 9. Crystal packing was mainly governed by hydrogen bond networks based on homomeric intermolecular oxime···oxime interactions for 6. The oxime and pyridyl structural motifs present in oximes 7 and 9 driven their selfassembly through intermolecular oxime···N-heterocycle and CH/π interactions.

Results and Discussion
Synthesis Different synthetic protocols have been developed for syntheses of 1,2-diarylethanones [1,4,12] and for 1-substituted-2-azinylethanones [1b] each of which has its own area of application. Moreover, both the 2-aryl-1-(3-pyridyl)ethanones [Ar-CH 2 -CO-(3-Py)] [3,5] and their isomers [Ar-CO-CH 2 -(3-Py)] [13] have received far less attention, probably due to the presence of the 3-pyridyl group which may limit their synthetic accessibility. A widely used pathway for the preparation of ketones is the coupling of acid halides with organometallic reagents and two alternative procedures were then examined for the synthesis of ethanones 1 and 2, the best results were obtained with Cu(I)-mediated coupling of organozinc reagents [14] and acid halides. As shown in Scheme 2, reaction of either acid chloride 11 or 13 with the organozinc bromide 10 or 14 in the presence of CuCN/LiBr gave ethanones 1 and 2 in 59% and 68% yield, respectively. At the same time, diarylethanone 2 has been prepared by a conventional Friedel-Crafts acylation reaction of thioanisole using phenylacetyl chloride in 75% yield [12a].
Among several alternative protocols for 1,2-diaryl/heteroaryl ethanone synthesis, two reasonable approaches to new ethanones 3-5 were examined, given that other methods appeared to be inefficient or might be useless, due to the presence of the 3-pyridyl moiety (Scheme 3). Method A refers to the reaction of Weinreb amides [15] with Grignard reagents and method B involves the Horner-Wittig reaction [16], a general procedure for the synthesis of arylacetylpyridines [3,16b]. Since under standard conditions [15], the preparation of the N-methoxy-N-methylamide 15 either from ester 16 or acid 17 gave in low yield (33%), this alternative was then abandoned. In consequence, the Horner-Wittig process appeared to be the method of choice to prepare the pyridylacetylarenes 3-5, and the reaction was first tested for preparation of ethanone 3 because pyridylcarbaldehyde 20 was commercially available. Reaction of arylcarbaldehyde 18 with aniline and diphenylphosphite gave the corresponding N,P-acetal 19 in 98% yield and the coupling of 19 and pyridylcarbaldehyde 20 in the presence of Cs 2 CO 3 in THF/i-PrOH produced enamine followed by HCl 3N hydrolysis to give ethanone 3 in 82% yield (Scheme 3). The best reaction conditions for method B were then applied to the preparation of targeted pyridylacetylarenes 4 and 5, giving overall yields of 46% and 69%, respectively. In accordance with standard protocol, oximes 6-9 were prepared from ethanones 2-5 in ≥ 81% yield and standard crystallization in 96% EtOH produced crystals of 6, 7 and 9 suitable for X-ray diffraction analysis (see below).

Structural Studies
The molecular structures of the new compounds were identified from their spectroscopic data and the structural properties of ketoximes 6-9 were examined in solution by 1 H-NMR and in the solid state by X-ray crystallography (only ketoximes 6, 7 and 9).
The IR spectra (KBr) of ethanones 1-5 showed the characteristic carbonyl band in the range of 1694-1681 cm -1 and the ν C=N absorption of ketoximes 6-9 in the 1597-1588 cm -1 range. Broad absorption at 2703 cm -1 (ν OH ) was shown by ketoxime 6, while the ketoximes 7-9 showed broad absorption bands in the range of 2805-2700 cm -1 (ν OH ). Both the 1 H-and 13 C-NMR data permitted unambiguous assignments with the aid of HMQC. In the 1 H-NMR spectra of oximes 6, 7 and 9, the characteristic δH chemical shift of the oxime proton was similar to those observed for several oximesubstituted pyridines in DMSO-d 6 [9,17] and the δ NOH values appeared in the range of 12.02 to 11.38 ppm in DMSO-d 6 and 9.65 to 9.07 ppm in CD 3 CN (Table 1).  Figure 2 shows the molecular structures of ketoximes 6, 7 and 9 and the molecular shape was similar for oximes 6 and 9. In all cases, the oxime functionality [8-10] bears the common antiperiplanar conformation of the thermodynamically preferred E-isomer. Particularly relevant is their crystal packing, which is governed by a robust hydrogen bonding network. The nature of the ring fragments modulated the intermolecular H-bond interactions. Each of oximes 6, 7 and 9 adopted a crystal structure [18] in which the molecular units were linked by either H-bond intermolecular oxime···oxime for 6 or oxime···N(py) molecular motifs for 7 and 9.
Oximes 7 and 9 showed a different pattern of C-H/π hydrogen bond, consistent with the acidity of the hydrogen bond donor, the soft acid CH group. Thus, the p-(methylsulfonyl)aryl CH of oxime 7 interacted with the face of the π-deficient heteroaromatic moiety (3-pyridyl) of the adjacent molecule and conversely, the α-CH of pyridyl moiety of oxime 9 pointed towards the p-(methylthio)aryl πexcessive aromatic ring (Figure 4). For oxime 7, the C-H···centroid(py) distance is 2.66 Å with an angle of 164.86° and oxime 9 has the C-H/π contact the C-H···centroid(aryl) distance of 2.62 Å and an angle of 176.20°. The present crystallographic results show that the homomeric O-H···N(pyridyl) hydrogen bond [8a,b,f] is the driving force that modulates and controls the assembly of compounds containing both oxime and pyridyl structural motifs [8,9]. In crystal structures of oximes 7 and 9, O-H···N(py) and CH/π hydrogen bonds cooperated in bringing the whole molecule into infinite layers for 7 and nonclassical packing of dimers for 9.

Conclusions
Depending on the aryl/3-pyridyl moieties present in 1,2-diaryl/heteroarylethanones 1-5, two synthetic approaches were used for their preparation. For cyclohexy/arylethanone 1 and diarylethanone 2, the Cu(I)-mediated coupling reaction between the corresponding organozinc reagent and acid chloride was performed in ≥ 59% yield. The selected route for pyridylacetylarenes 3-5 was the Horner-Wittig reaction. We identified the optimal conditions for the preparation of ethanone 3 and the best experiment gave 80% overall yield, whereas the yield was reduced to 45% for 4 and 69% for 5. Moreover, ethanones 2-5 were transformed to their ketoximes 6-9 in ≥ 81% yield. The structural properties of oximes 6-9 were examined in solution by 1 H-NMR and in the solid-state by X-ray crystallography. The crystal structures of oximes 6, 7 and 9 showed that the dominating hydrogen bond network was head-to-head OH···N(oxime) for 6 and, that both intermolecular head-to-tail O-H···N(py) and CH/π interactions modulated and controlled the self-assembly of molecules of oximes 7 and 9.

Method B:
A solution of pyridin-3-ylacetic acid hydrochloride (17, 1 g, 5.76 mmol) in 96% ethanol (50 mL) was passed through a column packed with a strongly basic anion-exchange (Amberlite IRA 401, hydroxide form). The eluates were concentrated to dryness to afford pyridin-3-ylacetic acid as a white solid. To a solution of the former solid in dry DMF/CH 3 CN (1:19, 50 mL) was added CDI (1.12 g, 6.91 mmol) and the solution was stirred 1.5 h, and Me(MeO)NH·HCl (0.68 g, 6.91 mmol) was added. After 12 h at room temperature, the reaction mixture was concentrated. The resulting residue was dissolved with EtOAc (200 mL), washed with brine (100 mL) and water (2 × 100 mL), dried over anhydrous Na 2 SO 4 and evaporated to dryness to give the amide 15 as a brown residue (yield 33%). (19) 4-(Methylsulfonyl)benzaldehyde (18, 4 g, 21.71 mmol) was dissolved in isopropyl acetate (100 mL) and stirred at room temperature. Aniline (2.38 mL, 26.05 mmol) was added in one portion followed by the addition of diphenylphosphite (6.68 mL, 34.74 mmol) in one portion. The resulting slurry was stirred 12 h at room temperature, filtered and washed with i-PrOH to give N,P-acetal 19 as a off-white solid (yield 98%).
Data collections for compounds 6, 7 and 9 were performed at 120(2) K on a Nonius KappaCCD single crystal diffractometer, using Cu-Kα radiation (λ= 1.5418 Å). Images were collected at a 29 mm fixed crystal-detector distance, using the oscillation method, with 2º oscillation and 40 s exposure time per image. Data collection strategy was calculated with the program Collect [21]. Data reduction and cell refinement were performed with the programs HKL Denzo and Scalepack [22]. A semi-empirical absorption correction was applied on each data set using the program SORTAV [23]. The structures were solved using direct methods, with the program SIR-92 [24]. Anisotropic least-squares refinement was carried out with SHELXL-97 [25]. All non-hydrogen atoms were anisotropically refined. All hydrogen atoms, except those of the methyl group, were located on a difference Fourier map and their isotropic displacement parameters were freely refined. Hydrogen atoms at the methyl groups were geometrically placed and refined as riding on their parent atoms with isotropic displacement parameters set to 1.5 times the U eq of the atoms to which they are attached.

Supplementary Material
The Supplementary Material for this paper is available via the Internet at http://www.mdpi.org/molecules/papers/13020301sm.pdf, 12 pages.