Choline hydroxide, an efficient, green, and recyclable base catalyst, promoted the synthesis of 3-aroylflavones via Baker–Venkataraman rearrangement

ABSTRACT

3-Aroylflavones, a new synthesized group of flavones, have been recently studied via Baker–Venkataraman rearrangement under ‘one-pot’ or sequential ‘one-pot’ reaction conditions. Choline hydroxide, an inexpensive, easily prepared, and green homogeneous catalyst has been found to be the most efficient catalyst among several base catalysts such as KOH/pyridine, KOH/triethylamine, and CaO-ES/triethylamine in the sequential ‘one-pot’ synthesis of 3-aroylflavones. Moreover, the catalyst’s efficiency, and high recyclability to six times without considerable loss of 3-aroylflavones have enhanced the importance of the procedure. Furthermore, 1D-NMR, 2D-NMR, and HRMS have also been used to elucidate the structures of synthesized compounds.

GRAPHICAL ABSTRACT

KEYWORDS:

• 3-aroylflavone
• chromen-4-one
• choline hydroxide
• sequential one-pot synthesis
• Baker–Venkataraman rearrangement

Introduction

Flavones have been classified in the largest subgroups of flavonoids which are widespread in the plant kingdom as specialized metabolites (1). Flavones possessed several physiological, biochemical, and ecological properties that allowed plants to thrive in harsh environments (2), besides they also showed important anti-inflammatory (3), anti-microbial (4), and anti-cancer activities in vitro (5). With the usefulness of these compounds, several synthetic pathways of flavones were performed via Baker–Venkataraman rearrangement (6), cyclodehydration of aromatic chalcones possessing hydroxy group at ortho position obtained from Claisen–Schmidt condensation (7), cyclocarbonylation (8), intramolecular Wittig reaction (9), annulation reactions of phenols with 5-alkylidene Meldrum's acids (10), Suzuki–Miyaura coupling (11), and Kostanecki–Robinson cyclization (12).

Baker–Venkataraman rearrangement was a traditional synthetic method for flavones via several reaction steps such as the rearrangement of 2-acetylphenolic esters promoted by a base catalyst to produce 1,3-diketones and subsequent treatment of 1,3-diketone intermediate compounds under the acidic condition to afford flavone derivatives (12). In the first step, several basic catalysts such as NaH or KOH in pyridine (12), LiHMDS (13), DBU, and pyridine were utilized (14), besides acidic catalysts such as H₂SO₄/AcOH (15), NaHSO₄/SiO₂ (16), FeCl₃ (17), and InCl₃ were involved in the second step (18). In these above procedures, 3-aroylflavones were found as minor products; however, several syntheses of 3-aroylflavones as main products via Baker–Venkataraman rearrangement were recently performed by using various catalysts such as DBU/pyridine (19), K₂CO₃/pyridine (20), K₂CO₃/acetone (21–23), NaOH/1,4-dioxane (24), KOH/pyridine (25), and LiHMDS/THF followed by treatment with hydrochloric acid (26).

Flavones have been found in several plants, while 3-aroylflavones are only obtained from chemical processes. Hence, in this paper, sequential 'one-pot' syntheses of 3-aroylflavones from the reaction of 2′-hydroxyacetophenones and benzoyl chlorides promoted by choline hydroxide in triethylamine were developed with the respect to preventing the presence of acidic catalysts in the cyclodehydration step (Scheme 1). (2-Hydroxyethyl)trimethylammonium hydroxide (called choline hydroxide, ChOH), an eco-friendly and benign ionic liquid, and strong base catalyst with highly recyclable capability (27–29), has been used for a variety of synthetic processes such as the synthesis of 2-amino-3-nitro-4H-chromene derivatives (27), 2-amino-4H-chromenes (28), tetrahydrobenzo[b]pyrans (29), etc … .

Scheme 1. Sequential ‘one-pot’ synthesis of 3-aroylflavones via Baker–Venkataraman rearrangement.

Materials and methods

Materials and instrumentation

Chemical reagents were purchased from commercial sources (Sigma Aldrich, Acros, TCI, Thermo Scientific). Column chromatography was performed on silica gel (Himedia, India, 0.037–0.063 mm). TLC was performed on Merck precoated aluminum plates (Silicagel 60 F254). The composition of crude product was analyzed by HPLC–UV (Phenomenex ultracarb™ 0.005 mm ODS 30 (60) Å column, 150 × 4.6 nm LC column with UV detector) and HPLC–ELSD (Phenomenex ultracarb™ 0.005 mm ODS (30) 150 × 4.6 nm with ELSD detector). Melting point data were determined by using Büchi B-545 equipment. NMR spectra were recorded with Brüker Avance 500 II, Brüker Avance 600 NEO spectrometers, and Brüker Avance III 500 MHz (SFO1 NUC ¹⁹F 470 MHz) spectrometers. HRMS spectra were recorded with Brüker Daltonics–micrOTOF-QII–ESI-Qq-TOF mass spectrometer, Exion LC-X500R QTOF (AB Sciex, USA), and Shimadzu LCMS-IT-TOF.

General procedure for synthesis of 6a-6aa

2′-Hydroxyacetophenones (1.0 mmol), triethylamine (2.5 mmol, 0.2530 g), and aroyl chlorides (2.3 mmol) were respectively added to a 5 mL two-necked round-bottom flask assembled with condenser and stirred at room temperature for 30 minutes, then, choline hydroxide (0.5 mmol, 0.0606 g) and triethylamine (11.0 mmol, 1.1131 g) were added. After that, the reaction mixture continued being performed under reflux for the appropriate time. After cooling down to room temperature, the product mixture was extracted with methylene chloride (70 mL) and water (3 mL) with a separatory funnel. Subsequently, the water layer was extracted with methylene chloride (10 mL x 3) and water was removed under reduced pressure to recover choline hydroxide. Besides, the methylene chloride layer was neutralized with aqueous 0.1% hydrochloric acid (2 mL), washed with water (10 mL x 3), and dried by anhydrous Na₂SO₄ before removing the solvent under reduced pressure. The product mixture composition was analyzed by HPLC-ELSD or HPLC-UV. The products 6a-aa were purified by column chromatography using n-hexane:ethyl acetate as an eluent system.

Spectroscopic data

Full spectral data of all compounds can be found in Supporting Information via the ‘Supplementary Content’ section of this article’s webpage.

3-Benzoyl-6-chloroflavone (6a)

Light yellow solid; yield: 212.9 mg, 0.59 mmol (59%); mp: 170–172°C (31); R_f = 0.09 (n-hexane/EtOAc, 9.6:0.4).

¹H NMR (500 MHz, CDCl₃): δ (ppm) = 8.20 (d, J = 2.5 Hz, H-5), 7.90 (dd, J = 7.5 , 1.5 Hz, H-2″, H-6″), 7.70 (dd, J = 9.0, 3.0 Hz, H-7), 7.64 (dd, J = 8.25, 1.5 Hz, H-3″, H-5″), 7.56-7.52 (m, H-8, H-4″), 7.45-7.34 (m, H-2′, H-3′, H-4′, H-5′, H-6′) (31).

¹³C NMR (125 MHz, CDCl₃), HSQC and HMBC: δ (ppm) = 193.2 (C-9), 175.5 (C-4), 162.9 (C-2), 154.6 (C-8a), 137.0 (C-1″), 134.7 (C-7), 134.0 (C-4″), 131.9 (C-6), 131.8 (C-4′), 131.6 (C-1′), 129.6 (C-2″, C-6″), 129.0 (C-3″, C-5″), 128.9 (C-2′, C-6′), 128.7 (C-3′, C-5′), 125.7 (C-5), 124.4 (C-4a), 122.8 (C-3), 120.0 (C-8).

HRMS (ESI): m/z [M + Na]⁺ cacld for C₂₂H₁₃O₃ClNa: 383.0445; found: 383.0436.

Result and discussion

At the beginning of this work, several base catalysts consisting of triethylamine, pyridine, CaO prepared from egg-shell (called CaO-ES), KOH/pyridine, and choline hydroxide with the amount in the range of 2.5–3.0 mmol were respectively used for the ‘one-pot’ reaction between 5′-chloro-2′-hydroxyacetophenone (1.0 mmol, 1a) and benzoyl chloride (2a, 2.1 mmol) under reflux within 5 hours to produce 3-benzoyl-6-chloroflavone (6a). The results showed that most surveyed base catalysts which did not have the ability to promote the formation of the expected product (6a), only produced 2-acetyl-4-chlorophenyl benzoate (3a) from the esterification reaction and 1-[2-(benzoyloxy)-5-chlorophenyl]vinyl benzoate (4a) from the enolization and esterification reactions (Table 1); except that KOH/pyridine was able to catalyze the reaction in order to afford 3-benzoyl-6-chloroflavone (6a) in the yield of 38%.

Table 1. The nature of base catalysts influenced the product formation in ‘one-pot’ Baker–Venkataraman rearrangement.^a

Entry	Base catalyst (mmol)	Yield^b (%)
		3a	4a	6a
1	Et3N (2.5)	61	39	0
2	Pyridine (2.5)	46	20	0
3	CaO-ES (3.0)	63	21	0
4	KOH (3.0)/pyridine (11.0)	0	0	38
5	ChOH (2.5)	12	0	0
6	ChOH (3.0)	9	0	0

^a 5′-Chloro-2′-hydroxyacetophenone (1a, 1.0 mmol), benzoyl chloride (2a, 2.1 mmol), and amount of surveyed catalyst were performed under magnetic stirring and heating at 110°C for 5 h.

^b Yields were calculated based on HPLC-ELSD analyses.

With the points of view on green chemistry, KOH/pyridine should be replaced with a green and recyclable catalyst; therefore, choline hydroxide/triethylamine and CaO-ES/triethylamine continued being selected to find the better catalyst. With the unsuccessful results from the ‘one-pot’ reaction, it forced us to change our mind to carry out a sequential ‘one-pot’ reaction via two steps including (i) esterification reaction to give 2-acetyl-4-chlorophenyl benzoate catalyzed by triethylamine at room temperature within 30 minutes and (ii) cyclodehydration reaction catalyzed by choline hydroxide/triethylamine or CaO-ES/triethylamine to produce 3-benzoyl-6-chloroflavone under reflux within 5 hours. Consequently, choline hydroxide/triethylamine was selected as the better catalyst than CaO-ES/triethylamine for this reaction owing to its higher product selectivity.

To optimize the reaction, a series of experiments were performed under reflux, wherein the molar ratio between 5′-chloro-2′-hydroxyacetophenone and benzoyl chloride was screened by fixing the molar amount of 5′-chloro-2′-hydroxyacetophenone (1.0 mmol) and varying the molar amount of benzoyl chloride from 2.1 mmol up to 2.5 mmol with 0.2 mmol increasing rate. Subsequently, the amount of benzoyl chloride (2.3 mmol) was chosen for the next experiments on varying the volume of triethylamine. The results also illustrated that the amount of triethylamine influenced the yield of the expected product considerably. The absence of triethylamine in step 1 or step 2 led to the failure of 3-benzoyl-6-chloroflavone synthesis; therefore, the amount of triethylamine was found to be 2.5 mmol for step 1 and 11.0 mmol for step 2 (Scheme 2).

Scheme 2. Sequential ‘one-pot’ synthesis of 3-benzoyl-6-chloroflavone via Baker–Venkataraman rearrangement.

The next experiments on the amount of choline hydroxide surveyed in the range of 0.25–3.50 mmol also demonstrated that an increasing amount of choline hydroxide obviously affected the reaction conversion; however, an excessive amount of choline hydroxide catalyst led to the reversed formation of 5′-chloro-2′-hydroxyacetophenone from the hydrolysis of 2-acetyl-4-chlorophenyl benzoate under base medium. Consequently, the appropriate amount of choline hydroxide promoted a sequential 'one-pot' reaction of 5′-chloro-2′-hydroxyacetophenone and benzoyl chloride to take place most efficiently was 0.5 mmol to get a better yield of 3-benzoyl-6-chloroflavone at 59% (Figure 1).

Figure 1. Influence of choline hydroxide amount in the synthesis of 3-benzoyl-6-chloroflavone from the reaction of 5′-chloro-2′-hydroxyacetophenone (1.0 mmol) and benzoyl chloride (2.3 mmol). Reaction conditions: step 1: Et₃N (2.5 mmol) under r.t. stirring for 30 min., and step 2: ChOH (varied amount) in Et₃N (11.0 mmol) under reflux for 5 hours.

To enlarge the scope of our synthetic pathways, altogether eleven 2′-hydroxyacetophenones and seven benzoyl chlorides were subjected to the synthesis of 3-aroylflavones. Besides, the effects of substituents of six 2′-hydroxyacetophenones were screened for their reactivity with benzoyl chloride (Entries 1–6, Table 2), and of six aroyl chlorides were surveyed for their reactivity with 5′-bromo-2′-hydroxyacetophenone under optimized conditions (Entries 7–12, Table 2). The results illustrated that the substituents on 2′-hydroxyacetophenones and benzoyl chlorides influenced the yield of 3-aroylflavones, especially substituents on benzoyl chlorides. Aroyl chlorides possessing electron-withdrawing substituents become more reactive than their electron-donating substituents, therefore, the more reactive the aroyl chloride, the shorter the optimized time reaction.

Table 2. Yields of 3-aroylflavones from the reactions between 2′-hydroxyacetophenones and benzoyl chlorides.^a

Entry	6	R^1	R^2	R^3	R^4	R′	Yield^b % (LC^c, time^d) of 6
1	6a	H	Cl	H	H	H	59 (80, 8)
2	6b	H	F	H	H	H	52 (85,8)
3	6c	H	Br	H	H	H	56 (91, 8)
4	6d	H	OH	H	H	H	24 (44, 15)
5	6e	H	OMe	H	H	H	55 (90, 5)
6	6f	H	NO2	H	H	H	40 (66, 9)
7	6g	H	Br	H	H	4-F	30 (68, 15)
8	6h	H	Br	H	H	3-Cl	31 (62, 4)
9	6i	H	Br	H	H	4-Br	30 (57, 36)
10	6j	H	Br	H	H	4-Me	32 (59, 48)
11	6k	H	Br	H	H	4-CF3	33 (62, 2)
12	6l	H	Br	H	H	4-NO2	23 (30, 5)
13	6m	H	F	H	H	4-F	35 (62, 8)
14	6n	H	F	H	H	3-Cl	33 (57, 7)
15	6o	H	F	H	H	4-Br	29 (51, 15)
16	6p	H	F	H	H	4-NO2	46 (65, 3)
17	6q	H	H	F	H	4-Br	41 (53, 16)
18	6r	H	H	F	H	4-F	32 (66, 12)
19	6s	H	Cl	H	H	3-Cl	40 (61, 4)
20	6t	OH	H	H	H	H	39 (59, 9)
21	6u	OH	H	H	H	3-Cl	72 (100, 3)
22	6v	OH	H	H	H	4-CF3	48 (92, 4)
23	6w	H	H	OH	H	H	11 (24, 15)
24	6x	H	Cl	H	Cl	3-Cl	32 (64, 1.5)
25	6y	H	Cl	H	Cl	4-Me	39 (70, 14)
26	6z	H	OMe	H	H	4-NO2	35 (41, 10)
27	6aa	H	H	[h]-benzo		H	46 (53, 24)

^a 2′-Hydroxyacetophenone (1.0 mmol), benzoyl chloride (2.3 mmol), and choline hydroxide (0.5 mmol) in triethylamine (11.0 mmol) were performed under reflux for an appropriate time.

^b Isolated yields.

^c LC yields were calculated based on HPLC analyses.

^d Time: reaction time in hour.

According to previous literature on the synthesis of 3-aroylflavones from the reaction of 2′-hydroxyacetophenones and benzoyl chlorides (21,23,25), the reaction mechanism was proposed in two paths: formation of flavones and 3-aroylflavones. Due to using an excess amount of aroyl chlorides, 3-aroylflavones became major products (path a, Scheme 3) besides flavones were minor products with yields in the range of 0–15% (path b, Scheme 3).

Scheme 3. A plausible mechanism for the synthesis of 3-aroylflavones via sequential ‘one-pot’ Baker – Venkataraman rearrangement catalyzed by choline hydroxide in triethylamine.

With the tendency of green catalysts and the advantages of choline hydroxide, the recovery and recycling of choline hydroxide must be investigated. The aqueous solution containing choline hydroxide collected after separation from the previous reaction was washed with methylene chloride and evaporated under reduced pressure at 80°C for 30 minutes. Subsequently, choline hydroxide after sucking reached 98% of recovery yield. The recovered choline hydroxide was tested for recyclability in the reaction of 5′-chloro-2′-hydroxyacetophenone and benzoyl chloride under the optimal condition (Entry 1, Table 2). The yields of reactions promoted by recovered choline hydroxide were not changed considerably from 80% to 73% after seven catalyst runs (Figure 2). The structure of the recovered catalyst at the first and sixth recycle times was analyzed by FT-IR and compared with that of the fresh choline hydroxide. The results displayed that the sixth recycled catalyst was contaminated by little organic impurities (Figure 3), however, it has not influenced remarkably the yield of the main product.

Figure 2. Reusability of choline hydroxide in sequential ‘one-pot’ synthesis of 3-benzoyl-6-chloroflavone. Yields were calculated based on HPLC-UV analyses.

Figure 3. FT–IR spectra of fresh ChOH (a), ChOH after 1st recycle (b), and ChOH after 6th recycle (c).

Furthermore, the FT-IR spectroscopy of 3-aroylflavone has not been announced in previous literature. Consequently, in order to master the FT-IR of 3-aroylflavone, the FT-IR comparison between the C = O stretching vibration of flavone and of 3-aroylflavone was studied, for instance, the FT-IR couple of 5-hydroxyflavone and 3-benzoyl-5-hydroxyflavone (6t) as well as of 6,8-dichloroflavone and 6,8,3′-trichloro-3-(3″-chlorobenzoyl)flavone (6x) (Figure 4(a,b)). The results showed that 3-aroylflavones possessed several stretching bands in the carbonyl region (1680–1580 cm⁻¹). These bands result from the asymmetrical and symmetrical C = O stretching modes. The C = O stretching at C-4 in the cyclic ring was located at lower frequencies (lower wavenumbers) than the C = O stretching of the 3-aroyl group owing to more planar leading to more conjugated resonance effects. After that, FT-IR spectra of six 3-benzoylflavones and five 3-aroylflavones were arbitrarily selected to illustrate the range of C = O stretching vibrations (Figure 4(c,d)).

Figure 4. FT–IR spectra of pairs of 6,8-dichloroflavone (Fl-3,5Cl) and 6,8,3′-trichloro-3-(3″-chlorobenzoyl)flavone (BzFl-3,5Cl) (a), pairs of 5-hydroxyflavone (Fl-6OH) and 3-benzoyl-5-hydroxyflavone (BzFl-6OH) (b), six 3-benzoylflavones (c), and five 3-aroylflavones (d).

The sequential ‘one-pot’ synthetic protocol of 3-aroylflavones via Baker–Venkataraman rearrangement of 2′-hydroxyacetophenones and benzoyl chlorides catalyzed by choline hydroxide in triethylamine provides several advantages in terms of low-cost, easily prepared catalyst, catalytic recyclability, and yields in comparison with the results already described in the previous literature (Table 3).

Table 3. Comparison of previous synthesis of 3-aroylflavones from 2′-hydroxyacetophenones and benzoyl chlorides.

Condition	Catalyst	Time	Catalyst runs	Overall yield (%)
Heat (80–90°C)	DBU/pyridine	6–7 h	None	47–67^19
r.t. MW (400 W, 120°C)	DCC, 4-pyrrolidinopyridine, CH2Cl2 K2CO3/pyridine	12–20 h 10 min	None	54–58^20
r.t. then reflux	K2CO3/acetone	24 h	None	35–46^21–23
Heat (120–140°C) r.t.	DBU/pyridine 10% NaOH/dioxane, MeOH, H2O	24 h 3–4 h	None	5–23^24
Heat (60–110°C)	DBU/pyridine; KOH/pyridine; K2CO3/pyridine; K2CO3/acetone or wet K2CO3/acetone	24 h	None	11–55^25
0°C to r.t. r.t.	LiHMDS in THF, toluene 20% or 37% HCl	20 min to 24 h 1–12 h	None	51–72^26
r.t. Heat (reflux)	Et3N ChOH/Et3N	30 min 2–48 h	4	11–72 (our work)

Conclusion

Summarily, a sequential ‘one-pot’ synthetic procedure for the access to 3-aroylflavones via Baker–Venkataraman rearrangement catalyzed by choline hydroxide in triethylamine has been developed. Several unknown 3-aroylflavones were obtained and their structures were characterized by 1D-NMR, 2D-NMR spectroscopies, and HRMS spectrometry. Moreover, choline hydroxide is a green catalyst that is easy to prepare, store, and can be recycled up to six times without any remarkable loss of activity.

Acknowledgements

All authors are grateful for all the support from Gam Thi Hong Ho, Cong-Danh Pham, Minh-Thu Thi Nguyen, Cong-Thang Duong, Vy Quynh Mong Le, Khang Nguyen, Ha Nguyen Phuong Tran, Khang Gia Tran, Tuyen Thi Ngoc Pham (University of Science, VNU-HCMC, Vietnam) and Prof. Dr. Poul Erik Hansen (University of Roskilde, Denmark).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research is funded by Vietnam National University, Ho Chi Minh City (VNU-HCM ) [grant number 562-2024-18-06].