Synthesis and Spectroscopic Properties of 2,3-Diphenyl-1,3-thiaza-4-one Heterocycles

Synthetic and spectroscopic data (H NMR, C NMR, IR, UV/Vis) for a series of six 2,3-diphenyl-1,3-thiaza-4-one heterocycles which differ in ring size and substitution is reported. The results show that there are significant differences in spectroscopic signals common to all six compounds. Distinctions can be made among the compounds using the IR absorbance of the C4 carbonyl and the H NMR signal at C2, and to a lesser extent the C NMR signal at C4 and the UV/Vis spectrum.

Reasoning that the amide formation would likely be the rate-determining step, we looked to a paper published by Dunetz et al.
Without any further optimization of the reaction, we have successfully reacted a number of other thioacids with N-benzylideneaniline 2 to prepare six-and seven-membered 2,3-diphenyl-1,3-thiaza-4-one heterocycles (Figure 5, Table 1).We have previously reported the syntheses and x-ray crystal structures of all products (Yennawar & Silverberg, 2013, Yennawar & Silverberg, 2014, Yennawar, Bendinsky et al., 2014, Yennawar, Singh & Silverberg, 2014, Yennawar, Singh & Silverberg, 2015).The syntheses of 6, 8 and 12 have since been repeated and updated procedures and yields are reported here.Thioacid 3 reacted to give 4, the compound Surrey et al. (1958) had been unable to synthesize.A successful outcome was also achieved with 2-thionicotinic acid 7, which Dandia et al. (2004) reported as unreactive with an N-aryl imine, under a variety of conditions and catalysts, without microwave radiation.Also notable is N-acetyl-L-cysteine 9, which reacted to give the cis diastereomer 10 as the major product in reasonable yield.Thioacid 11 worked as well, giving the synthetically challenging seven-membered ring 12.In fact, every thioacid attempted to this point has worked in this reaction.Products in each case were purified by column chromatography and/or recrystallization.The reader may note that in most cases tetrahydrofuran and 2-methyltetrahydrofuran were both used as solvents (T3P came as a 50% solution in 2-methyltetrahydrofuran, and THF was additionally added to the reaction) while in the later preparation of 12 only 2-methyltetrahydrofuran was used.It does not appear to make any difference in the reaction, and in our current work we are using only 2-methyltetrahydrofuran.Yields and reaction times were not optimized.a) Yield of isolated material after chromatography of crude product and then recrystallization; b) Yield of isolated material after recrystallization from crude product; c) Yennawar, Bendinsky et al., 2014, Ponci et al., 1963, Kollenz & Ziegler, 1970, Oae & Numata, 1974 2013) have proposed a mechanism for T3P-promoted 1,3-benzothiazin-4-one formation in which the acid is rapidly activated by reaction with T3P, and then is attacked by the imine nitrogen to give an iminium ion.On the other hand, it has been shown that 1,3-thiazolidin-4-one formation without a promoter occurs by initial attack of sulfur on the imine carbon (Tierney, 1989, Surrey, 1947).Reaction of 2 with 5 in THF, without T3P/2-methyltetrahydrofuran and with or without pyridine, gave no apparent reaction (Figure 6).This supports the mechanism proposed by Unsworth et al.
The results are compiled in Table 2, with only key signals common to each product compared.Full spectral data is provided in the Experimental Section.

1 H NMR
The signal that is common to all of the structures in Table 2 is the proton at C2, a carbon which is also connected to the sulfur, the ring nitrogen, and a phenyl ring.In comparing the 5-, 6-, and 7-membered rings (13, 4, and 12), the most upfield resonance was at 5.91 ppm in 4, > 0.2 ppm less than in 13 and 12.The signal in compound 12 was 0.05 ppm more downfield than 13.Thus the C2 signal appears to be useful diagnostically for distinguishing between the thiazolidinone and thiazinone rings, and possibly between the thiazinone and thiazepanone rings, although this is less certain because of the presence of the cyclopropyl substituent at C6.
The fused benzene ring in 6 moved the C2 resonance downfield by 0.18 ppm compared to 4. The pyridine ring in 8 moved the signal another 0.1 ppm farther downfield than in 6.Thus 6, 4, and 8 are distinguishable according to the C2 resonance.
Having a cis N-acetyl group at C5 moved the C2 resonance in 10 downfield, as compared to 4, to 6.09.A 1 H-1 H-COSY NMR experiment in CD 2 Cl 2 showed that the protons at C5 and C6 were all coupled to each other (Figure 7).Coupling of the proton on C5 was also seen with the adjacent NH (6.75 ppm).Although the proton at C2 appeared as a singlet in the COSY, some coupling was seen with one of the protons (dd at 3.5 ppm at room temp.) at C6. Processing of the NMR data with resolution enhancement (Figure 8) did in fact show small couplings on C2 (dt, 6.1 ppm) and one of the C6 protons (ddd, 3.5 ppm).
Figure 8. 1 H NMR data of 10 processed with resolution enhancement.
In the crystal structure of 10 previously reported (Yennawar, Singh & Silverberg, 2014), one of the two conformations was a boat conformation, with the proton on C2 in a pseudo-axial positon, whereas the other was a half-chair, typical of the other structures, with the proton on C2 in a pseudo-equatorial position.The ring carbonyl at C4 is also common to all of the compounds.Here some dramatic differences were seen.Five-membered 13 gave a resonance at 171.2, whereas six-membered 4 had a signal at 169.7, and seven-membered 12 shifted downfield to 174.2 ppm.Compounds 6 and 8 had nearly the same chemical shift, which was upfield of 4 by more than six ppm.There are two carbonyls in 10, with the signals close together, so no firm conclusions can be drawn, but they were both close to the value for C4 in 4.
The signal at C4 thus appears to be diagnostic for distinguishing five-, six-, and seven-membered rings, and for distinguishing thiazinones (4, 10) from thiazinones with an aromatic ring fused at C5 and C6 (6, 8).

IR
The amide carbonyl C=O stretching vibration at C4 is common to all of the structures.
A five-membered lactam is expected to give this signal at ~1750-1700 cm -1 (Silverstein & Webster, 1998), but the experimental result for 13 was much lower at 1668.8 cm -1 .
A six-membered lactam is expected (Silverstein & Webster, 1998) to give a signal at ~1650 cm -1 , and the experimental result for 4 was 1633.2 cm -1 .Compound 10 has a second amide carbonyl on the sidechain at C5.Comparison of the spectra of 4 and 10 indicated that in 10 the ring carbonyl was at 1643.2 cm -1 and the amide sidechain at C5 was the absorbance at 1678.1 cm -1 .
A seven-membered lactam is also expected to absorb at ~1650 cm -1 (Silverstein & Webster, 1998).The experimental result for 12 was close at 1646.6 cm -1 .Thus, while the wavenumber decreased going from the five-( 13) to the six-membered ring (4), it increased going from the six-to the seven-membered ring ( 12).The differences in absorbances for 4, 12, and 13, suggest that the value should be diagnostic for distinguishing the thiazolidinones from the thiazinones and thiazepanones, and possibly distinguishing the latter two from each other.
Fusion of a lactam to another ring usually increases the absorbance by 20-50 cm -1 (Silverstein & Webster, 1998).
Benzothiazinone 6 gave the peak at 1682.3 cm -1 , an increase of 49.1 cm -1 compared to 4, whereas pyridothiazinone 8 had an absorbance at 1650.7 cm -1 , an increase of 17.5 compared to 4. The experimental difference of 31.6 cm -1 makes the benzothiazinone and pyridothiazinone readily distinguishable.

UV/Vis
Four (6, 4, 10, and 13) compounds displayed a  max at 272 nm, and 12 was at 268.The only major change was in 8. Two very strong peaks of nearly equal absorbance were observed at 284 and 308 nm.The absorbance at 308 is believed to be from the pyridine ring.Thus, UV/Vis may be diagnostic for establishing the presence of the pyridine ring.

Conclusions
Five different 2,3-diphenyl-1,3-thiaza-4-one rings were prepared and six were studied spectroscopically.The T3P synthetic method has thus far proven to be general and versatile and we are continuing to use it in ongoing studies.
Of greatest overall success in distinguishing the different compounds spectroscopically was the infrared absorbance of the C4 carbonyl.The 1 H NMR signal at C2 also displayed significant differences among the compounds.The 13 C signals at C2 were roughly similar, but the C4 signals distinguished between those that had a fused arene ring and those that didn't.The pyridothiazinone 8 gave a distinctive UV/Vis spectrum.The data collected here should be useful to researchers in identifying these types of compounds.

Experimental
General: Toluene, N-benzylideneaniline 2, THF, pyridine, thiosalicylic acid 5, and 3-mercaptopropionic acid 3 were purchased from Sigma-Aldrich (St. Louis, MO).N-acetyl-L-cysteine 9, 2-thionicotinic acid 7, and [1-(sulfanylmethyl)cyclopropyl]acetic acid 11 were obtained from Oakwood Products, Inc. T3P in 2-methyltetrahydrofuran (50 weight %) was obtained from Euticals, Inc. TLC plates (silica gel GF, 250 micron, 10 x 20 cm, catalog No. 21521) were purchased from Analtech (Newark, DE).TLC's were visualized under short wave UV, and then with I 2 and then by spraying with ceric ammonium nitrate/sulfuric acid and heating.Infrared spectra were run on a Perkin-Elmer Spectrum One using a diamond-ATR attachment for the direct powder analysis (Villanova University).Spectra were taken at a resolution of 4 cm -1 , 16 scans averaged. 1H and 13 C experiments (Penn State University Park) were carried out on a Bruker 600.07-MHzAvance-III instrument using a 5-mm cryoprobe TCI 1H-13C/15N/D Z-GRD, or a Bruker Avance-III-HD 500.20-MHzinstrument using a 5 mm CPPBBO BB-1H/19F/D Z-GRD probe, or a Bruker 850.24-MHzAvance-III also using a 5-mm cryoprobe TCI 1H-13C/15N/D Z-GRD.Samples were dissolved in CDCl 3 and analyzed at RT.Low temperature NMR experiments were carried out on a Bruker DPX-300 operating at 1 H frequency of 300.13 MHz using a RT BBO probe.Typical conditions for 1 H acquisition were 1 sec relaxation delay, acquisition time of 2.76 sec, spectral width of 12 kHz, 16 scans.Spectra were zero-filled to 128k points, and multiplied by exponential multiplication (EM with LB = 0.3 Hz) prior to FT.For 13 C experiments a 2 sec relaxation delay was employed, acquisition time of 0.9088 sec, spectral width of 36 kHz, 128 scans.Spectra were zero-filled once, and multiplied by EM with LB = 2 Hz prior to FT.An Applied Biosystems API 2000 Triple Quadrupole Mass Spectrometer was used to determine molecular masses by electrospray ionization (Villanova University).A 0.1% (v:v) formic acid methanol mixture containing the compound at 100 ppm was infused at 20 μL/min into the electrospray source.Source and compound dependent parameters for the MS/MS product ion analysis were as follows: curtain gas (CUR) = 20, nebulizer gas (GAS1) = 15, heater gas (GAS2) = 15, electrospray voltage (IS) = 5500 V, source temperature (TEM) = 398 K, declustering potential (DP) = 40 V, focusing potential (FP) = 400 V, entrance potential (EP) = 10 V, collision energy (CE) = 25 V, cell exit potential (CXP) = 4 V. Ultraviolet/Visible spectroscopy was performed on a Thermo Electron Corp. Genesys 10 UV (Penn State Schuylkill).Melting points (Penn State Schuylkill) were performed on a Thomas Hoover Capillary Melting Point Apparatus (Arthur H. Thomas Co., Philadelphia, PA).

Table 2 .
Comparison of key spectroscopic signals.