2,3-Dihydrobenzo[e][1,3]oxazin-4-one

Abstract

The title compound and its hydroxymethyl precursor have been fully characterised for the first time. The IR spectra, fully assigned ¹H and ¹³C NMR spectra, and X-ray structures are presented for both compounds. Both compounds form hydrogen-bonded dimers in the crystal structures.

1. Introduction

The fundamental heterocyclic compound 2,3-dihydrobenzo[e][1,3]oxazin-4-one (or 2,3-dihydro-1,3-benzoxazin-4-one) 3 has only been mentioned in a single literature paper [1] over 30 years ago where it was characterised by ¹H NMR, elemental analysis and a melting point. The compound was prepared as a reference in a study of the conformational preference of fully saturated analogues [1]. We recently had cause to prepare this compound using the convenient two-step synthesis from salicylamide 1 (Scheme 1), and present here for the first time the full characterisation of both it and its N-hydroxymethyl precursor 2, including fully assigned ¹H and ¹³C NMR spectra, UV and IR spectra and X-ray crystal structures.

2. Results

The reaction of salicylamide 1, conveniently obtained by heating methyl salicylate with aqueous ammonia [2] with a 1:1 mixture of 37% aqueous formaldehyde and formic acid gave the hydroxymethyl compound 2 and heating a solution of this under reflux in toluene for 1 h led to loss of formaldehyde and formation of the target compound 1 [1].

The ¹H NMR spectra obtained for both 2 and 3 (

Table 1. ¹H and ¹³C chemical shifts (ppm) for 3 and 2.

Position	3		2
	δC	δH	δC	δH
2	73.8	5.25	78.0	5.34
4	164.1	–	163.2	–
4a	118.6	–	118.2	–
5	128.2	7.94	128.4	7.95
6	122.6	7.13	122.6	7.12
7	134.6	7.48	134.7	7.46
8	116.7	7.01	116.5	6.97
8a	158.3	–	158.2	–
9	–	–	69.2	5.06
NH/OH		7.58		3.72

) were broadly in agreement with the data already reported [1], and the aromatic ring protons in the spectrum of 2 gave particularly well-resolved signals (see Supplementary Materials), allowing the determination of all H–H coupling constants (

Table 2. ¹H–¹H coupling constants (Hz) in 3 and 2.

Compound	3JH5–H6	4JH5–H7	5JH5–H8	3JH6–H7	4JH6–H8	3JH7–H8	3JH9–OH
3	8.0	~0	~0	7.6	~0	8.0	—
2	8.0	1.6	0.4	7.2	1.2	8.4	7.6

). In order to achieve full NMR assignments for compounds 2 and 3, a range of additional spectra were obtained, including DEPTQ ¹³C, COSY, HSQC and HMBC (see Supplementary Materials). The HSQC data allowed the association of H and C shifts, as shown in Table 1, but the HMBC results were key to arriving at an unambiguous assignment. Thus, for 3, the high-carbon chemical shift of 164.1 ppm was clearly C=O, and this correlated on HMBC with the proton signal at 7.94 ppm, while the carbon signal at 158.3 ppm corresponding to C–O was correlated on HMBC with the proton signals at 7.48 and 7.94 but not 7.13. Similarly, for 2, the quaternary carbon signal at 163.2 ppm was correlated to the proton signal at 7.95 ppm, which was also correlated to the carbons at 134.7 and 158.2 ppm. Combining this information with the coupling information derived from the proton spectra (Table 2) led to an unambiguous assignment of all the signals, as summarised in Figure 1. There is a high degree of consistency between the data for the two compounds, with the major difference being the replacement of the broad ¹H signal at 7.58 ppm for NH in 3 by a doublet at 5.06 ppm and broad triplet at 3.72 ppm for the CH₂OH group in 2.

The infrared spectra showed the expected features with prominent peaks for O–H at 3275 cm^–1 and C=O at 1643 cm^–1 for 2, and for N–H at 3163, 3092 and 3053 cm^–1 and C=O at 1661 cm^–1 for 3.

Since both compounds 3 and 2 formed suitable crystals after recrystallisation, we have been able to determine their structure using X-ray diffraction (Figure 2). The molecular structures both have a planar benzene ring; however, the heterocyclic ring takes up a twisted conformation, with C=O slightly below the benzene ring plane and the 2-CH₂ significantly above it (distances from benzene plane for O4 0.4812(18) and 0.5181(19) Å, and for C2 0.601(2) and 0.647(2) Å, for 3 and 2, respectively). The lactam nitrogen is quite accurately planar in both cases: angle sum at N(3) 358.8° (3) and 359.9° (2). The dimensions of the heterocyclic rings are summarised in

Table 3. Heterocyclic ring dimensions for 3 and 2 (Å, °).

Compound	3	2
O(1)–C(2)	1.4272(14)	1.4187(13)
C(2)–N(3)	1.4412(14)	1.4542(14)
N(3)–C(4)	1.3423(13)	1.3531(13)
C(4)–O(4)	1.2380(12)	1.2379(13)
C(4)–C(5)	1.4837(14)	1.4801(15)
C(5)–C(10)	1.3995(15)	1.3961(14)
C(10)–O(1)	1.3725(13)	1.3725(12
C(10)–O(1)–C(2)	111.43(8)	111.68(8)
O(1)–C(2)–N(3)	111.07(9)	111.27(9)
C(2)–N(3)–C(4)	119.34(9)	119.15(9)
N(3)–C(4)–C(5)	114.05(9)	114.42(9)
C(4)–C(5)–C(10)	119.11(9)	119.31(9)
C(5)–C(10)–O(1)	120.77(9)	119.88(10)
Flap angle *	48.85(11)	49.32(10)

* Defined as angle between planes defined by C(5)…C(10) and O(1)–C(2)–N(3).

.

As expected, both structures exhibited hydrogen bonding, with 3 forming symmetrical N–H to O=C $R_2^2$(8) dimers and 2 forming O–H to O=C $R_2^2$(12) dimers in the crystal (Figure 3). The hydrogen bonding parameters (

Table 4. Hydrogen bonding parameters for 3 and 2 (Å, °).

	D—H···A	D—H	H···A	D···A	D—H···A
3	N(3)–H(3)···O(4)	0.925(13)	1.936(13)	2.8609(12)	178.0(13)
2	O(11)–H(11)···O(4)	0.921(14)	1.812(15)	2.7309(12)	175.4(17)

) are conventional. The structure of 2 also showed the formation of offset π-stacked chains involving the benzene ring, running along the crystallographic b-axis (centroid···centroid distance 3.7923(4) Å).

A search of the Cambridge Crystallographic Database (CSD) showed that only a few comparable structures have been determined before (Figure 4). The enantiomerically pure dimethyl compound 4 [3], as well as the 2,2-dimethyl compounds 5 [4] and 6 [5], all show similar molecular structures to 1 and 3, with C(2) significantly out of the plane defined by the benzene ring. The removal of the carbonyl group, as in 7 [6], results in a half-chair conformation with the pyramidal amine nitrogen and C(2) on opposite sides of the mean molecular plane, while the introduction of a second carbonyl group results in a completely planar geometry for molecules such as 8 [7], 9 [8] and 10 [9]. The fully saturated analogue 11 [1] makes an interesting comparison with the structure of 2.

The hydrogen bonding patterns in these structures also make an interesting comparison. Compound 4 has an unusual structure with three independent molecules, two forming a hydrogen-bonded dimer with the third bonded only by N–H to one O=C of the dimer [3]. In contrast, compound 5 forms hydrogen-bonded dimers that are almost identical to 3 [4], as also does the dione 8 [7]. The fluorinated analogue 10 [9] forms a ribbon structure, with each molecule linked to two others via N–H to carbamate (as opposed to lactam) O=C hydrogen bonding. Finally, the N-hydroxy compound 6 [5] consists of two independent molecules forming an O–H to O=C hydrogen-bonded dimer somewhat analogous to 2. In contrast to 2, the saturated analogue 11 forms O–H to O=C hydrogen-bonded chains along the c axis [1].

In summary, we have been able to fully assign the ¹H and ¹³C NMR spectra for compounds 3 and 2 for the first time. The X-ray structures of 3 and 2 feature an out of plane 2-CH₂, with the molecules forming hydrogen-bonded dimers in the crystal.

3. Experimental

3.1. General Experimental Details

The melting points were recorded on a Reichert hot-stage microscope (Reichert, Vienna, Austria) and are uncorrected. NMR spectra were obtained using a Bruker AV300 instrument (Bruker, Billerica, MA, USA). Spectra were run with internal Me₄Si as the reference, and chemical shifts are reported in ppm to high frequency of the reference. NMR spectra were processed, and simulations were produced using iNMR reader, version 6.3.3 (Mestrelab Research, Santiago de Compostela, Spain). Salicylamide 1 was prepared by heating a mixture of methyl salicylate and aqueous ammonia under reflux for 18 h and filtering off the crystalline product obtained upon cooling [2].

3.2. Synthesis of 2-Hydroxymethyl-2,3-dihydrobenzo[e][1,3]oxazin-4-one 2

A solution of salicylamide 1 (3.63 g, 26.5 mmol) in aqueous formaldehyde (37%, 25 mL, 308 mmol) and formic acid (25 mL) was heated under reflux for 1 h. After cooling, this was added to ice (200 g) and the mixture was neutralised by the addition of solid sodium bicarbonate, then extracted with CH₂Cl₂ (2 × 40 mL). Drying and evaporation, followed by chromatography of the residue (SiO₂, Et₂O), gave first a material containing formaldehyde oligomers, followed by the desired product 2 (2.29 g, 49%) as colourless crystals, mp 93–94 °C (lit. [1] 92–94 °C). IR (ATR) ν_max/cm^–1 3275 (OH), 1643, 1612, 1481, 1462, 1362, 1315, 1188, 1038, 1005; ¹H NMR (400 MHz, CDCl₃) δ 7.95 (1H, ddd, J = 8.0, 1.6, 0.4 Hz), 7.46 (1H, ddd, J = 8.4, 7.2, 1.6 Hz), 7.12 (1H, ddd, J = 8.0, 7.2, 1.2 Hz), 6.97 (1H, ddd, J = 8.4, 1.2, 0.4 Hz), 5.34 (2H, s), 5.06 (2H, d, J = 7.6 Hz), 3.72 (1H, t, J = 7.6 Hz); ¹³C NMR (100 MHz, CDCl₃) δ 163.2 (C=O), 158.2 (C–O), 134.7 (CH), 128.4 (CH), 122.6 (CH), 118.2 (C), 116.5 (CH), 78.0 (CH₂), 69.2 (CH₂).

3.3. Synthesis of 2,3-Dihydrobenzo[e][1,3]oxazin-4-one 3

A solution of compound 2 (2.29 g, 12.8 mmol) in toluene (70 mL) was heated under reflux for 4 h. The solution was decanted off from the solid paraformaldehyde byproduct and evaporated to afford the product 3 (1.37 g, 71%) as colourless crystals, mp 129–132 °C (lit. [1] 129–131 °C). UV-Vis (CH₂Cl₂): λ_max (log ε) 300 (3.47), 238 (3.92); IR (ATR) ν_max/cm^–1 3163, 3091, 3053 (NH), 1661, 1609, 1479, 1352, 1215, 1153, 1029, 993, 827; ¹H NMR (400 MHz, CDCl₃) δ 7.94 (1H, d, J = 8.0 Hz), 7.58 (1H, br s, NH), 7.48 (1H, dd, J = 8.0, 7.6 Hz), 7.13 (1H, dd, J = 8.0, 7.6 Hz), 7.01 (1H, d, J = 8.0 Hz), 5.25 (2H, s); ¹³C NMR (100 MHz, CDCl₃) δ 164.1 (C=O), 158.3 (C–O), 134.6 (CH), 128.2 (CH), 122.6 (CH), 118.6 (C), 116.7 (CH), 73.8 (CH₂).

3.4. X-ray Structure Determination of 3 and 2

X-ray diffraction data for compounds 3 and 2 were collected at 173 K using a Rigaku FR-X Ultrahigh Brilliance Microfocus RA generator/confocal optics with XtaLAB P200 diffractometer [Mo Kα radiation (λ = 0.71073 Å)]. Data for both compounds were collected (using a calculated strategy) and processed (including correction for Lorentz, polarisation and absorption) using CrysAlisPro [10]. The structures were solved by dual-space methods (SHELXT) [11] and refined by full-matrix least-squares against F² (SHELXL-2019/3) [12]. Non-hydrogen atoms were refined anisotropically, and hydrogen atoms were refined using a riding model except for the hydrogen atoms on N3 and O11, in 3 and 2, respectively, which were located from the difference Fourier map and refined isotropically subject to a distance restraint. All calculations were performed using the Olex2 [13] interface.

3.4.1. 2,3-Dihydrobenzo[e][1,3]oxazin-4-one 3

Crystal data for C₈H₇NO₂, M = 149.15 g mol^–1, colourless block, crystal dimensions 0.10 × 0.065 × 0.04 mm, monoclinic, space group P2₁/n (No. 14), a = 6.4500(3), b = 5.05307(19), c = 21.1492(8) Å, β = 94.081(3)°, V = 687.55(5) ų, Z = 4, D_calc = 1.441 g cm^–3, T = 173 K, goodness of fit on F² 1.066, 14408 reflections measured, 1701 unique (R_int = 0.0276) which were used in all calculations. The final R₁ [I > 2σ(I)] was 0.0338, and wR₂ (all data) was 0.0965. The data have been deposited at the Cambridge Crystallographic Data Centre as CCDC 2373268. The data can be obtained free of charge from the Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/structures.

3.4.2. 2-Hydroxymethyl-2,3-dihydrobenzo[e][1,3]oxazin-4-one 2

Crystal data for C₉H₉NO₃, M = 179.17 g mol^–1, colourless block, crystal dimensions 0.11 × 0.09 × 0.02 mm, orthorhombic, space group Pbca (No. 61), a = 13.7423(3), b = 6.86948(16), c = 17.5448(4) Å, V = 1656.28(7) ų, Z = 8, D_calc = 1.437 g cm^–3, T = 173 K, goodness of fit on F² 1.039, 32521 reflections measured, 2113 unique (R_int = 0.0283) which were used in all calculations. The final R₁ [I > 2σ(I)] was 0.0345, and wR₂ (all data) was 0.0946. The data have been deposited at the Cambridge Crystallographic Data Centre as CCDC 2373269. The data can be obtained free of charge from the Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/structures.