Selective Oxidative Cleavage of the C-C Bond in α,β-Epoxy Ketone into Carbonyl Compounds.

This method afforded aromatic carbonyl compounds under TBHP via selective oxidative cleavage of the C-C bond in α,β-epoxy ketones. Aromatic acid came from the aroyl section, and aromatic aldehyde came from the other aromatic group. TBHP acted as a free radical initiator and oxidant. The reaction within the solvent went through a ring-opening addition, cleavage of the C-C bond in the ethylene oxide section, and oxidation, affording the target compounds in moderate to good yields. The HPLC yield of aromatic aldehyde was up to 91%. The HPLC yield of aromatic acid was up to 99%. The reaction under solvent-free conditions gave two kinds of aromatic acids coming from different moieties of α,β-epoxy ketone via the further oxidation of aromatic aldehyde. The substituent effect was discussed, and the reaction mechanism was proposed. This method allowed the reaction to occur in a simple system metal-free.

Introduction

Oxidative cleavage of the C–C bond is widely presented in organic synthetic chemistry. Especially, selective oxidative cleavage has attracted many efforts in the fields of synthesis and biotechnology. Achieving efficiently selective cleavage of the C–C bond for chemical transformations remains challenging.

Epoxides, including derivatives of ethylene oxide, are an important class of building blocks. They could be converted into various functional groups. To date, much research on the transformation of α,β-epoxy ketones is generally focused on ring-opening reactions of the oxirane ring.[1] The dominant reactions of α,β-epoxy ketones involve cleavage of the C–O bond of epoxide to form 1, 2-[2] or 1, 3-[3] dicarbonyl compounds with no change in the total number of carbon. On the other hand, the breakage of the C–O bond of α,β-epoxy ketones followed by rearrangement could form carbonyl compounds with a changed skeleton.[4] However, the direct C–C bond cleavage of the epoxide motif is more difficult and has been less studied due to the harsh conditions required. The ionic liquid [bmIm]OH-catalyzed retro-Aldol cleavage reaction could catalyze the thiolysis of α,β-epoxy ketones to provide β-keto sulfides.[5] Chalcone epoxides could form β-methoxy alcohols via ring-opening reaction under iodine in methanol, and then the C–C bond cleaved to form α-ketoaldehyde, followed by acetalization to give α, α-dimethoxyacetophenones.[6] It has not been reported that cleavage of the C–C bond followed by oxidation, that is, oxidative cleavage of the C–C bond of α,β-epoxy ketone affords aromatic aldehyde and aromatic acid. Herein, we report the selective oxidative cleavage of the C–C bond of α,β-epoxy ketones (chalcone epoxides) to form the corresponding aromatic carbonyl compounds by tert-butanol hydroperoxide (TBHP) as a free radical initiator and oxidant (Scheme ).

Scheme 1

Reactions Involving the C–C Bond Cleavage in α,β-Epoxy Ketones

Results and Discussion

When chalcone epoxide (1a) was added to the TBHP system [TBHP 4.0 equiv in diethylene glycol diethyl ether (DGDE)] at 130 °C for 2 h, the reaction led to the two expected products, which were benzaldehyde (A) and benzoic acid (B) in 27% and 20% yield, respectively (Table , entry 1). At first, various free radical oxidants (4.0 equiv) were screened at 130 °C for 2 h, such as TBHP, benzoyl peroxide (BPO), lauroyl peroxide (LPO), H2O2, m-chloroperbenzoic acid (m-CPBA), dicumyl peroxide (DCP), di-tert-butyl peroxide (DTBP), and cumyl hydroperoxide (CHP). The results showed that TBHP was more suitable for the reaction as a free radical initiator and oxidant (Table , entries 1–8). Second, several protic, aprotic, polar, or nonpolar solvents with higher boiling points were screened, such as DGDE, propylene glycol (PG), o-dichlorobenzene (ODCB), and N-methyl pyrrolidone (NMP). DGDE gave a better result (Table , entries 1 and 9–12). Although the conversion rates were high in DMSO and PG, only trace or small amounts of the target compounds (A and B) were produced (Table , entries 9 and 10).

Table 1

Cleavage Reaction of 1aa

					yield (%)b
entry	T (°C)	free radical oxidant (equiv.)	solvent (10 mL)	time (min)	A	B	conv.b (%)
1	130	TBHP (4.0)	DGDE	120	27	20	55
2	130	BPO (4.0)	DGDE	120	0	0	16
3	130	LPO (4.0)	DGDE	120	0	0	0
4	130	H2O2 (4.0)	DGDE	120	4	5	16
5	130	m-CPBA (4.0)	DGDE	120	trace	trace	15
6	130	DCP (4.0)	DGDE	120	trace	trace	48
7	130	DTBP (4.0)	DGDE	120	trace	trace	28
8	130	CHP (4.0)	DGDE	120	trace	trace	5
9	130	TBHP (4.0)	DMSO	120	trace	trace	>99
10	130	TBHP (4.0)	PG	120	4	6	89
11	130	TBHP (4.0)	ODCB	120	trace	25	29
12	130	TBHP (4.0)	NMP	120	3	3	25
13	130	TBHP (2.0)	DGDE	120	3	2	13
14	130	TBHP (8.0)	DGDE	120	65	51	82
15	130	TBHP (10)	DGDE	120	76	64	94
16	130	TBHP (12)	DGDE	120	77	97	>99
17	130	TBHP (14)	DGDE	120	77	113	>99
18	130	TBHP (18)	DGDE	120	76	118	>99
19	130	TBHP (20)	DGDE	120	70	126	>99
20	90	TBHP (4.0)	DGDE	120	0	0	0
21	110	TBHP (4.0)	DGDE	120	11	3	20
22	140	TBHP (4.0)	DGDE	120	32	16	65
23	150	TBHP (4.0)	DGDE	120	22	13	47
24	160	TBHP (4.0)	DGDE	120	25	12	50
25	140	TBHP (12)	DGDE	60	81	65	88
26	140	TBHP (12)	DGDE	75	84	79	94
27	140	TBHP (12)	DGDE	90	85	94	98
28	140	TBHP (12)	DGDE	95	87	95 (83)c	>99
29	140	TBHP (12)	DGDE	105	83	97	>99

Reaction conditions: 0.5 mmol compound 1a.

Determined by HPLC with biphenyl as an internal standard.

Isolated yield.

Next, the conversion rate of 1a, as expected, enhanced with the increase of TBHP at 130 °C. The yields of A and B also increased simultaneously. When TBHP reached 12 equivalents, 1a completely conversed, giving the 77% yield of A (Table , entries 1 and 13–19). Interestingly, as shown in entries 1 and 13–15, the yield of A was usually greater than that of B. This trend continued until TBHP was over 10 equivalents. It means that the formation rate of A might be higher than that of B (Scheme ). On the other hand, the situation reversed when TBHP exceeded 12 equivalents and the yield of B exceeded A and persistently enhanced with the increase of TBHP. Increasing TBHP to 18 equivalents, the reaction only showed a 1% drop of A, but 20 equivalents of TBHP resulted in a 7% drop. That was due to further oxidation of benzaldehyde (A) into benzoic acid (B) by surplus TBHP (Scheme ), usually occurring in the final stage. It was an undesired result because too much aldehyde was oxidized unless we did not wish to get aldehyde.

Scheme 3

Proposed Mechanism of the Free Radical Selective Oxidative Cleavage of the C–C Bond in α,β-Epoxy Ketone

To further figure out how much equivalents of TBHP would result in the excessive oxidation of aromatic aldehyde, the substituted chalcone epoxide (2a) was carried on the free radical oxidative cleavage reaction under the optimized conditions mentioned above. Entry 3 in Table showed that oxidation of A had begun when TBHP closed to 10 equivalents. Now the reaction was not done yet, the conversion rate of 2a was only 76%, and the yield of 2ab was also low (69%). 2ab gradually increased with the increase of TBHP, and oxidation of A increased accordingly (Table , entries 4–6). However, TBHP over 20 equivalents was unable to further improve the yield of 2ab, instead of causing much oxidation of A (Table , entry 7), indicating that the high yield of 2ab was inevitably accompanied by the excessive oxidation of aromatic aldehyde.

Table 2

Cleavage Reaction of 2aa

			yield (%)b
entry	TBHP (equiv.)	DGDE (mL)	A	B	2ab	conv.b (%)
1	4.0	10	11	trace	8	14
2	8.0	10	59	trace	40	66
3	10	10	66	7	69	76
4	12	10	69	14	78	>99
5	14	10	61	23	86	>99
6	18	10	56	31	88	>99
7	20	10	50	40	88	>99
8	10	0	0	74	70	93
9	15	0	0	81	65	>99
10	20	0	0	84	69	>99

Reaction conditions: 0.5 mmol compound 2.

Determined by HPLC with biphenyl as an internal standard.

What follows is the expansion of substrates under the optimized conditions. Substrates, all in the cis configuration, were prepared according to the reported method[7] (see Supporting Information). Table shows that the substituents had profound effects on the C–C cleavage reaction, whether on the phenyl D-ring (black) adjacent to the epoxide or the phenyl E-ring (red) connected to the carbonyl.

Table 3

Scope of α,β-Epoxy Ketonesa

substrate	R2	product and yield (%)b			substrate	R1	product and yield (%)
3a	4-CH3	3aa (89)	3ac (7)	B (93)	2a	4-CH3	A (69)	B (14)	2ab (78)
4a	4-OCH3	4aa (74)	4ac (26)	B (88)	21a	4-OCH3	A (78)	B (18)	21ab (76)
5a	4-F	5aa (91)	5ac (8)	B (87)	22a	4-F	A (72)	B (16)	22ab (85)
6a	4-Cl	6aa (64)	6ac (14)	B (73)	23a	4-Cl	A (74)	B (14)	23ab (87)
7a	4-Br	7aa (65)	7ac (14)	B (67)	24a	4-Br	A (76)	B (16)	24ab (91)
8a	4-NO2	8aa (trace)	8ac (trace)	B (trace)	25a	4-NO2	A (trace)	B (trace)	25ab (trace)
9a	2-CH3	9aa (88)	9ac (11)	B (>99)	26a	2-CH3	A (64)	B (17)	26ab (75)
10a	2-OCH3	10aa (91)	10ac (7)	B (91)	27a	2-OCH3	A (75)	B (18)	27ab (89)
11a	2-F	11aa (trace)	11ac (trace)	B (trace)	28a	2-F	A (64)	B (15)	28ab (76)
12a	2-Cl	12aa (trace)	12ac (trace)	B (trace)	29a	2-Cl	A (61)	B (16)	29ab (75)
13a	2-Br	13aa (trace)	13ac (trace)	B (trace)	30a	2-Br	A (57)	B (14)	30ab (57)
14a	2-CF3	14aa (trace)	14ac (trace)	B (trace)	31a	3-CH3	A (78)	B (20)	31ab (90)
15a	3-CH3	15aa (73)	15ac (13)	B (84)	32a	3-OCH3	A (75)	B (18)	32ab (8)
16a	3-OCH3	16aa (66)	16ac (15)	B (97)	33a	3-F	A (74)	B (17)	33ab (90)
17a	3-F	17aa (trace)	17ac (trace)	B (trace)	34a	3-Br	A (75)	B (16)	34ab (78)
18a	3-Cl	18aa (trace)	18ac (trace)	B (trace)	35a	3-NO2	A (trace)	B (trace)	35ab (trace)
19a	3-Br	19aa (trace)	19ac (trace)	B (trace)	36a	Benzo[c]3,4	A (48)	B (11)	36ab (63)
20a	3-NO2	20aa (trace)	20ac (trace)	B (trace)

Reaction Conditions: 0.5 mmol of each substrate.

Determined by HPLC with biphenyl as an internal standard.

For D-ring, parasubstituents with electron-donating (ED) or weak electron-withdrawing (EW) effects, such as CH3-, CH3O-, F-, Cl-, and Br- (substrates 3a–7a), gave the medium to high yields of benzoic acid (B) and the corresponding aromatic aldehydes (3aa–7aa). However, the NO2 -substituted substrate did not react (substrate 8a). Interestingly, the stronger ED group CH3O- could result in heavy deep oxidation of aldehyde (4ac reached 26% for substrate 4a), indicating that the aromatic aldehyde with high electronic density could be oxidized easily. Surprisingly, these substituents, whether at ortho- or meta-, showed serious inhibitory effects toward the cleavage reaction (substrates 9a–20a), except CH3- and CH3O- (substrates 9a, 10a, 15a, and 16a). It seems like the higher electron density on D-ring could be conducive to the cleavage reaction, but not exactly (e.g., m-CH3O- with EW effect).

On the E-ring, whether the substituents (except NO2-) were at para-, ortho-, or meta-, the cleavage reaction gave good results (substrates 2a and 21a–36a). For parasubstituents, CH3O- and CH3- (with the ED effect) gave the lower yields than that of halogen- (with EW) (substrates 21a-24a), indicating that the lower electron density on the E-ring may be conducive to the cleavage reaction, but excessive low electron density (e.g., NO2- with strong EW) would not trigger the reaction. There may be a threshold value of electron density that triggers the reaction. Moreover, the ortho- and metasubstituents did not present relatively large differences and regularity (substrates 26a–34a), except the substrate 35a with NO2–. As to the E-ring replaced by naphthalene, the yield of 36ab reduced may be due to the steric hindrance.

We also tried to use the aliphatic-substituted α,β-epoxy ketone, or the aliphatic-substituted epoxy as substrates, such as 2-phenyl-1-oxaspiro[2,5]octan-4-one, or 2-(phenoxymethyl) oxirane; unfortunately, no corresponding products formed (see Supporting Information).

It was found that with no solvent, the cleavage reaction gave only two kinds of aromatic acids from different phenyl rings, and no benzaldehyde (A) was detected. Table entries 8–10 displayed the reaction results of substrate 2a under solvent-free conditions. With 20 equivalents of TBHP at 140 °C for 95 min, the reaction gave 69% 4-methylbenzoic acid (2ab) and 84% benzoic acid (B) in more than 99% conversion rate. The oxidative cleavage process with or without solvent (DGDE) should be similar: A formed first, 2ab formed slowly in multiple steps, and during this period, A was further oxidized to B by excessive TBHP (Scheme ). Under solvent-free conditions, aldehyde A was oxidized easily in a high concentration of TBHP. On the contrary, TBHP with a lower concentration in DGDE would decrease oxidizing ability, and more, TBHP may form a solvate, thereby weakening its oxidative activity. Table shows the results of most substrates’ reaction under solvent-free conditions with 20 equivalents of TBHP. The substrates of D-ring substitution gave high yields of benzoic acid (B), except the substituents NO2- and CF3- with a strong EW effect. The yield of B originating from the E-ring was slightly affected by electronic effects and positions of substituents on the D-ring, even F-, Cl-, and Br-, that were different from the reaction in DGDE (Tables vs ). However, aromatic acids yielding from the D-ring were influenced greatly by substituents. The substrates 3a, 9a, and 15a with CH3- respectively on para-, ortho-, and meta- gave only medium yields (Table , 3ac 55%, 9ac 48%, and 15ac 62%). Similarly, the substrates 4a and 10a with CH3O- respectively on para- and ortho- gave low yields of aromatic acids (Table , 4ac 30% and 10ac 10%). In theory, the yields of the two aromatic acids should be close, such as shown in substrates 5a–7a, 11a–13a, and 16a–19a (Table ). Further oxidation of aromatic acid may occur to form side products due to the high electronic density on the D-ring. Especially, substrate 16a showed a higher yield of 16ac (74%) because of the EW effect of meta-OCH3. It means that the appropriate electronic density on the D-ring was beneficial to the selective oxidative cleavage reaction. On the other hand, with no substituents on the D-ring, the substrates gave steady yields of B from 71 to 87% under solvent-free conditions. The effects of substituents on the E-ring were similar to the reaction in solvent (substrates 2a, 21a–24a, 26a, 28a–34a, and 36a).

Table 4

Scope of α,β-Epoxy Ketones with Solvent-Freea

substrate	R2	product and yield (%)b		substrate	R1	product and yield (%)
1a	H	1ac (B)	B (>99) (88)c	2a	4-CH3	B (84)	2ab (69)
3a	4-CH3	3ac (55)	B (95)	21a	4-OCH3	B (82)	21ab (88)
4a	4-OCH3	4ac (30)	B (89)	22a	4-F	B (81)	22ab (90)
5a	4-F	5ac (78)	B (89)	23a	4-Cl	B (88)	23ab (>99)
6a	4-Cl	6ac (87)	B (94)	24a	4-Br	B (80)	24ab (93)
7a	4-Br	7ac (78)	B (89)	26a	2-CH3	B (82)	26ab (88)
9a	2-CH3	9ac (48)	B (>99)	28a	2-F	B (83)	28ab (60)
10a	2-OCH3	10ac (10)	B (66)	29a	2-Cl	B (71)	29ab (77)
11a	2-F	11ac (78)	B (89)	30a	2-Br	B (87)	30ab (90)
12a	2-Cl	12ac (70)	B (90)	31a	3-CH3	B (83)	31ab (67)
13a	2-Br	13ac (72)	B (94)	32a	3-OCH3	B (80)	32ab (84)
14a	2-CF3	14ac (trace)	B (trace)	33a	3-F	B (82)	33ab (88)
15a	3-CH3	15ac (62)	B (90)	34a	3-Br	B (81)	34ab (87)
16a	3-OCH3	16ac (74)	B (94)	36a	Benzo[c]3,4	B (80)	36ab (79)
17a	3-F	17ac (79)	B (97)
18a	3-Cl	18ac (79)	B (89)
19a	3-Br	19ac (73)	B (87)
20a	3-NO2	20ac (trace)	B (trace)

Reaction Conditions: 0.5 mmol of each substrate.

Determined by HPLC with biphenyl as an internal standard.

Isolated yield.

To understand the feature of the reaction and the role of TBHP, the control experiments were carried out with different additives (Table ). With transition metal ions (Cu2+, Co2+, and Ni2+), the cleavage reaction presented significantly reduced conversion rates (Table , entries 1–3). The possible inhibitory mechanism of transition metal ions is shown in Scheme . The free radical tBuOO· from TBHP reacted with substrate 1a to form the free radical intermediate Ι via ring-opening addition, followed by oxidation by metal ions and reacted with the acetate ion (CH3CO2–) to form acetate peroxide (ΙΙ). Cleavage of the peroxy-bond resulted in re-cyclization to give the substrate 1a. The futile circle inhibited the oxidative cleavage of the C–C bond.[8] Furthermore, the reaction was completely inhibited by radical scavenger 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) (Table , entry 4). It revealed that the C–C cleavage of α,β-epoxy ketones belongs to a free radical-induced reaction, and TBHP acted as a free radical initiator.

Table 5

Control Experimenta

				yield (%)b
entry	DGDE (mL)	additive (mol %, equiv., atm)	TBHP (equiv.)	A	B	conv.b (%)
1	10	Cu(CH3COO)2•H2O 10%	4.0	trace	trace	11
2	10	Co(CH3COO)2•4H2O 10%	4.0	trace	trace	19
3	10	Ni(CH3COO)2•4H2O 10%	4.0	trace	trace	19
4	10	TEMPO 4.0 equiv	4.0	0	0	0
5	10	O2 1.0 atm.	4.0	20	27	65
6	10	N2 1.0 atm.	4.0	20	26	64
7	10	O2 1.0 atm.	0	trace	trace	trace
8	2		12	30	168	>99
9	4		12	58	141	>99
10	6		12	75	121	>99
11	8		12	85	98	>99
12	10		12	87	95	>99

Reaction Conditions: 0.5 mmol of 1a.

Determined by HPLC with biphenyl as an internal standard.

Scheme 2

Proposed Effect Mechanism of Transition Metal Ions on the Free Radical Oxidative Cleavage of the C–C Bond

Next, the reaction could occur in O2 or N2 (Table , entries 5–6). However, in O2 without TBHP, the reaction could not happen. It means that TBHP acted not only as a free radical initiator but also as an oxidant in the cleavage reaction. Moreover, entries 7–11 showed that as the amount of DGDE increased from 2 to 10 mL, the concentration of TBHP gradually decreased, the yield of A gradually increased, and B gradually decreased, indicating the concentration of TBHP and solvate greatly affected its oxidize ability.

Based on the results mentioned above, this work proposed the mechanism of the free radical selective oxidative cleavage of the C–C bond of α,β-epoxy ketone (Scheme ). At first, TBHP generated t-butyl peroxy radicals at a high temperature, followed by reacting with the substrate to form the two free radical intermediates via ring-opening addition. The free radical intermediates belong to the classical O-class radicals, and Ι was the favorable form with the ED substituent,[9] resulting in the higher electron density on the D-ring could be conducive to the cleavage reaction. The free radical α-cleavage of Ι yielded one target product--Aldehyde (A2), and an intermediate ΙΙ, which converted to α-carbonyl aldehyde ΙΙΙ by homolytic cleavage.[6] The α-carbonyl aldehyde ΙΙΙ generated a new free radical intermediate ΙV, which decomposed into V.[10] The benzaldehyde radical V was further oxidized into acid (B1) by TBHP via path a or b. Second, the product A2 was further oxidized into B2 through a path similar to ΙΙΙ. It is worth noting that α-carbonyl aldehyde ΙΙΙ had the advantage in competition with A2 for being oxidized into acids because of its higher activity than that of A2. Therefore, only at the later stages of the reaction, A2 was further oxidized into B2 with the increase of TBHP. The high yield of B1 may inevitably bring about the increased B2.

In summary, this work has developed an effective way of chemically selective oxidative C–C cleavage of α,β-epoxy ketones by TBHP, thereby converting a variety of α,β-epoxy ketones into carbonyl compounds (aromatic aldehydes and aromatic acids). This reaction system is simple, metal-free, and suitable for obtaining various aromatic carbonyls from the corresponding α,β-epoxy ketones.