Preparation of cyclic imides from alkene-tethered amides: application of homogeneous Cu(ii) catalytic systems

A Cu-based homogeneous catalytic system was proposed for the preparation of imides from alkene-tethered amides. Here, O2 acted as a terminal oxidant and a cheap and easily available oxygen source. The cleavage of C 
<svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="13.200000pt" height="16.000000pt" viewBox="0 0 13.200000 16.000000" preserveAspectRatio="xMidYMid meet"><metadata>
Created by potrace 1.16, written by Peter Selinger 2001-2019
</metadata><g transform="translate(1.000000,15.000000) scale(0.017500,-0.017500)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z"/></g></svg>
 C bonds and the formation of C–N bonds were catalyzed by Cu(ii) salts with proper nitrogen-containing ligands under 100 °C. The synthesis approach has potential applications in pharmaceutical syntheses. Moreover, scaled-up experiments confirmed the practical applicability.


Introduction
Cyclic imides are becoming increasingly popular in the pharmaceutical eld. [1][2][3][4][5] Some drug molecules bearing cyclic imide structures are shown in Scheme 1. Among them, lenalidomide, carmofur, uorouracil, and aminoglutethimide are antineoplastic drugs; utamide is an antiu drug; and phensuximide, phenytoin, and glutethimide are antiepileptic, antiarrhythmic, and sedative-hypnotic drugs, respectively. The preparations of imides and cyclic imides have been paid much attention over the years. [6][7][8][9] As shown in Scheme 2, certain methods were gradually proposed. In 2005, Higuchi et al. proposed an oxidation method of amides with 2,6-dichloropyridine N-oxide catalyzed by ruthenium porphyrin in benzene at a low temperature of 40 C overnight (Scheme 2a). 10 Aer that, more nonnoble metals were employed in this type of transformation. In 2009, Beller et al. translated hex-3-yne into cyclic imides with CO and NH 3 catalyzed by [Fe 3 (CO) 12 ] in THF at 120 C for 16 h, which was the rst report on the iron-catalyzed synthesis of succinimides via the carbonylation of diverse internal and terminal alkynes with amines or ammonia to achieve good selectivity and high activity (Scheme 2b). 11 Later, metallic oxides were also found to be able to achieve such preparations. In 2016, Shimizu et al. completed the direct synthesis of cyclic imides by using carboxylic anhydrides and amines using Nb 2 O 5 as a watertolerant Lewis acid catalyst without solvents (Scheme 2c). 12 In 2017, Gaunt et al. set up a Co-catalyzed carbonylative cyclization procedure of unactivated, aliphatic C-H bonds with a combination of Co(acac) 2 , PhCOONa, Ag 2 CO 3 , PhCl, and CO (Scheme 2d). 13 In addition, other methods such as electrocatalysis and photocatalysis could also be used in this transformation (Scheme 2e and f). 14,15 The valid and selective oxidation of organic compounds offers the opportunity for the streamlined conversion of simple precursors into value-added products. 16 Oxidation reactions, together with polymerizations and carbonylations, constitute the largest industrial applications in the eld of homogeneous catalysis, and substantial value-added bulk and ne chemicals can be fabricated through this technology. 17 With the superiority of environmental compatibility, low cost, and high efficiency, O 2 is becoming a frequently used oxidant in both experimental and industrial scenarios. 18,19 With regard to green Scheme 1 Pharmaceutical molecules with cyclic imide structures. and sustainable chemistry, oxidants such as inorganic salts, IBX, oxone, BQ TBHP, DDQ, or PhI(OAc) 2 suffer from many problems such as waste disposal, high cost, and poor atom economy.
Encouraged by the recently published articles and our interest in nonnoble-metal-catalyzed oxidization and cyclization reactions, herein, we developed an efficient, copper-based catalytic system comprising commercially available Cu salts and a bidentate nitrogen ligand for the preparation of cyclic imides from alkene-tethered amides in acetonitrile (MeCN) under atmospheric O 2 at 100 C. [20][21][22] 2. Results and discussion

Screening of metal salts and ligands
Metal salts coordinated with specic ligands exhibit excellent catalytic abilities. [23][24][25][26][27][28][29] In the initial set of experiments, a series of frequently used metal salts coordinated with ligands containing N, P, or other elements were tested, and the results are shown in Table 1. Here, 10 types of transition metal salts (Fe, Co, Ni, Cu, Zn, Mn, Pd, Ru, Rh, and Ir salts) with 22 types of ligands (for detailed information, see ESI, Table S1 †) were employed in the oxidation of alkene-tethered amide 1a into cyclic imide 2a in MeCN at 100 C in an O 2 environment under atmospheric pressure. Cu(acac) 2 with ligand A (neocuproine) exhibited outstanding catalytic performance with the highest 2a yield of 85%, while the others behaved moderately or even badly under otherwise identical conditions. Phosphine ligands were not expected to be efficient in an O 2 atmosphere due to inactivation caused by oxidation. 18 Further, they indeed provided no yields (Table 1; for detailed yields, see ESI, Tables S2-S11 †). In addition, all the tested Cu salts satisfactorily performed with yields over 75% (e.g., Cu(BF 4 ) 2 $2H 2 O: 82%; CuF 2 : 81%; CuCl 2 -$2H 2 O: 80%), among which Cu(acac) 2 gave the highest yield of 2a (Table S12 †).

Optimized dosage of Cu(acac) 2 and neocuproine
Aer conrming the suitable combination of Cu(acac) 2 and neocuproine, the optimized ratio of Cu(acac) 2 and neocuproine ligand was explored, as shown in Fig. 1A. Originally, the relative molar ratio of [Cu] vs. ligand was set at 1 vs. 1. As expected, the yields were higher when the loading values were higher (from 20% at 0.4 equiv. : 0.4 equiv. to 46% at 1 equiv. : 1 equiv.). However, aer each loading was above 1 equiv., the yields stopped increasing (maintaining 46% at 1.5 equiv.). Then, the loading of neocuproine was adjusted further as the dosage of Cu(acac) 2 was xed at 1 equiv. As shown in Fig. 1A, when the loadings of neocuproine were 0.8, 1.0, 1.2, and 1.5 equiv., the yields were 31%, 46%, 68%, and 85%, respectively. This implied the strong dependency of neocuproine dosage in the system. However, a further increase in the neocuproine dosage to 1.8 equiv. did not help in increasing the yield. Finally, the loading was doubled as 2 equiv. : 3 equiv., and the yield was still unimproved. Therefore, the optimized dosage of Cu(acac) 2 and neocuproine was 1 equiv. vs. 1.5 equiv.

Effects of solvents
Solvents acted as the reaction media and intensively inuenced the catalytic reactions. [30][31][32][33][34] The effects of nine representative solvents on the catalytic reaction were tested, and the results were summarized, as shown in Fig. 1B. Five nonpolar solvents were tested. Tetrahydrofuran (THF) and acetone showed nearly no activity in the system. Nitromethane (MeNO 2 ) and 1,4dioxane started to show some reactive abilities. Surprisingly, when the solvent was switched to MeCN, the highest yield was 85%. In addition, three high-boiling-point polar aprotic solvents, namely, dimethyl sulfoxide (DMSO), dimethyl formamide (DMF), and 1,3-dimethyl-2-imidazolidinone (DMI), were tested, but they yielded nearly no products. Finally, we tried ethanol (EtOH) as a protic solvent, and it showed a medium yield. Without doubt, MeCN was considered to be the most suitable solvent for this system.

Effects of temperature
To gain more specic information regarding the catalytic system, the reaction temperatures were set from 30 to 110 C (shown in Fig. 1C). Here, 30 and 40 C were too low to provide sufficient energy for the reaction. Temperatures ranging from 50 to 100 C started to give satisfactory yields: a higher temperature was associated with a higher yield. However, when the temperature was further improved to 110 C, the yield was slightly reduced perhaps due to the side reactions induced by the higher temperature. Therefore, 100 C was considered to be the optimal temperature for this system.

Kinetic study
We conducted kinetic observations; the amounts of 1a and 2a were recorded and an aldehyde intermediate (1-ethoxy-4-ethyl-5-oxo-4-phenylpyrrolidine-2-carbaldehyde) was detected and recorded (shown in Fig. 1D). A clear initial increase in the aldehyde concentration was observed, and consumption subsequently occurred. In addition, the amount of 1a sharply decreased in the initial phase (for detailed information of the quantitative data, see ESI, Table S15 †).

Effects of additives
In order to further increase the yield and efficiency of the reaction, 10 types of additives were tested (shown in Table 2). Unfortunately, none of them yielded better results. However, the incorporation of H 2 O did not reduce the yield, which implied that the catalytic system was robust against water and therefore could be applied to more industrial scenes.

Authentication of component necessity in the catalytic system
A series of control experiments were conducted with the lack of Cu salt, ligand, or O 2 to conrm the indispensability of each component (Table S14 †). Further, the results showed that without the Cu salt, ligand, or O 2 , the reaction could not proceed at all.

Scaled-up experiment
Importantly, we conducted a large-scale experiment (larger by 10 times) and obtained a yield of 66% over 24 h and 78% over 36 h, indicating that our system can nd valuable applications in industrial production, as shown in Scheme 3.

Substrate scope
On the basis of the optimized reaction conditions, the scope of this transformation was explored, as shown in Scheme 4. To our delight, ethoxy could be well preserved during the entire process. Moderate to good yields were obtained, with the highest isolated yield of 84% (2a). A phenyl group in the molecular chain seemed to be in favor of the transformation (2b, 2c, and 2d vs. 2e and 2g) perhaps due to the stabilization effect of the aromatic ring in the reaction course. 35 Substitution beta to the amide led to a substantial decrease in the reaction yield even aer longer reaction times (2f), presumably due to steric hindrance. Substrates with 4-, 5-, and 6-membered spirocycles in the alpha position could be tolerated, but with relatively lower yields (2h, 2i, 2j, and 2k). The 1l-containing conjugated alkene structure could be translated into 2l, but with a low yield (48 h) maybe because of the unanticipated oxidation process in the O 2 atmosphere at 100 C.
To explore the versatility of this catalytic system, it was used to prepare glutarimide derivatives. Before that, we retested the copper salts and found that the best one was CuF 2 (Table S13, ESI †). Further, a new round of substrate scope was demonstrated in Scheme 3. Products with a 6-membered ring exhibited lower yields (4a-4g, 49-63%) owing to the lower stability than the 5-membered ones. 36 Similarly, 3h bearing a conjugated alkene structure still performed poorly (Scheme 5).

Reaction mechanism
In order to investigate the possible reaction mechanism, some control experiments and isotope-labeling experiments were conducted (Scheme 6). Two types of radical scavengers, namely, (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) and 1,1-diphenylethylene, were employed to explore whether the transformation involved free radicals or not. The extremely low yields of 2a indicated that the free radicals participated in the process, which was doubtlessly consistent with earlier reports (Scheme 6a and b). 35,[37][38][39] Compound E was detected in the system and was considered to be an intermediate. Therefore, E was used as the substrate and O 2 was replaced by 18 O 2 to track the O source. Further, product 2a 0 with isotope-labeling was produced, as detected by HR-ESI (Scheme 6c). In addition, C 18 O was conrmed using GC-MS. Finally, a proof experiment was conducted (Scheme 6d). In conclusion, O 2 took part in the reaction and became a carbonyl group in the nal product; here, C 18 O was detected again. Considering the experimental results and earlier reports involving this type of metal/ligand system, 35,40-45 a possible reaction pathway was proposed, as shown in Scheme 7. For precision and emphasis, the acac À anion was omitted and therefore the electric charges were not shown. Copper salts coordinated with the N-ligand neocuproine to form the catalytically active species [Cu(II)]/L. [Cu(II)]/L replaced a proton in substrate 1a generating A, and the cis-amidocupration of the alkene occurred, affording organo-copper(II) intermediate B. In the next step, oxygen molecules participated in the process and radical (C) was formed via the homolysis of the C-Cu bond. 37,39 The generation of C from B was achieved via a radical species (for detailed process, see ESI, Scheme S1 †). Then, the 1,3hydrogen migration and homolysis of the O-O bond generated intermediate E, which could be detected in the system. A combination of [Cu(II)]/L and transfer of electrons afforded G. Finally, another oxygen molecule was inserted, and product 2a was formed aer intramolecular electron transfer, releasing CO at the same time. 38

Conclusions
Summarily, a homogeneous Cu(II) catalytical system was set up for the preparation of cyclic imides from alkene-tethered amides. In this environmentally friendly reaction pathway, no expensive catalyst was employed. O 2 was used as an efficient and easily available terminal oxidant. Products containing succinimide and glutarimide structures can nd wide applications in pharmaceutical syntheses along with successfully scaled-up experiments. Finally, a possible reaction pathway was proposed and veried by control experiments and isotope-labeling experiments. We believe that our catalytic system has academic and practical worth, and further investigations on the cascade amidoarylation of unactivated olens catalyzed by copper complexes are ongoing.

Materials
Metal salts, ligands, additives, raw materials, and solvents were purchased from Sinopharm Chemical Reagent Co, J&K Chemicals, or Sigma-Aldrich. Materials obtained from commercial resources were used without further purication unless otherwise noted.

Instrumentation
Liquid 1 H NMR spectra were obtained in CDCl 3 using the residual CHCl 3 as the internal reference (7.26 ppm) using a Bruker 400 spectrometer. 1 H NMR peaks were labeled as singlet (s), doublet (d), triplet (t), quartet (q), and multiplet (m). The coupling constant values were reported in Hertz (Hz). Liquid 13 C NMR spectra were conducted at 100 MHz in CDCl 3 using residual CHCl 3 as the internal reference (77.0 ppm). GC-MS analysis was performed using gas chromatography-mass spectrometry (GC-MS, 7890A and 5975C, Agilent). Highresolution electrospray ionization mass spectrometry (HR-ESI-MS) was performed on a Bruker FT-ICR-MS instrument (Solarix 9.4T).

Synthesis of substrates
Taking the synthesis of 2-ethyl-N-ethyoxy-2-phenylpent-4enamide (1a) as the example, the procedure was as follows. According to the reported processes, 41,46 freshly made lithium diisopropylamide (LDA, 12.5 mmol) was mixed with a preprepared solution of 2-phenylbutanoic acid (10 mmol) in 10 mL extra dry THF at 0 and 40 C and maintained for 2 h. Aer that, allyl bromide (18 mmol) was added dropwise and the solution was stirred for 2.5 h. Liquid separation aer dilution (diethyl ether/water), extraction aer acidication (water phase), and column chromatography on the silica gel afforded 2-ethyl-2-phenylpent-4-enoic acid for use in the subsequent step. Oxalyl chloride (12.5 mmol) was added dropwise into a solution of 2-ethyl-2-phenylpent-4-enoic acid (10 mmol) in CH 2 Cl 2 (10 mL) followed by adding a catalytic amount of DMF. Aer stirring for 2 h, the solvent was removed via a rotary evaporator. The remnant solid was slowly added to a mixture of EtONH 2 $HCl (15 mmol) and K 2 CO 3 (20 mmol) in EtOAc and H 2 O (2 : 1). Aer another 2 h, the organic phase was collected and the target compound in the aqueous phase was extracted by EtOAc. Washing, desiccation, ltration, and column chromatography afforded the target compound of 2-ethyl-N-ethyoxy-2phenylpent-4-enamide (1a).

General procedure for the reaction of alkene-tethered amides to cyclic imides
Substrate (0.1 mmol), copper salts (0.01 mmol), ligand (0.015 mmol), and solvent (1.5 mL) were successively loaded into a reactor; then, the reactor was connected to an O 2 balloon. Next, the reactor was moved into an oil bath maintained at the desired temperature (e.g., 100 C) and stirred for 24 h. Aer this reaction, the reactor was cooled down to room temperature. The products were isolated by column chromatography on silica gel using n-hexane/ethyl acetate as the eluent and their NMR and MS spectra were obtained.