In the sustainable development of selective oxidative transformations, it is necessary to consider both the environmental impacts and economic benefits. The selective catalytic oxidation of available hydrocarbons under mild conditions with molecular oxygen to give valuable fine chemicals and building blocks as feedstocks for the chemical, agricultural, and pharmaceutical industries under mild conditions is therefore attracting growing interest [1-3]. Molecular oxygen, especially in air, is the most suitable and attractive oxidant because it is cheap, clean, environmentally benign, abundantly available, and non-toxic. However, it has to be activated for direct use in the oxo-functionalization of hydrocarbons because of kinetic hindrance by its triplet ground state [4].

A literature survey showed that many excellent catalysts have been developed for promoting the activation of molecular oxygen and its incorporation into hydrocarbons. These catalysts can be roughly divided into two classes. Catalysts such as transition-metal salts [3, 5], biomimetic catalysts [6, 7], enzymatic catalysts [8, 9], organometallic catalysts [10, 11], and carbon materials [12-14] preferentially activate molecular oxygen followed by hydrocarbon oxidation. Other catalysts such as halogen-based catalysts [15, 16] and hydroxyimide organocatalysts [17-27] activate the hydrocarbon prior to incorporation of molecular oxygen into the hydrocarbon. An efficient selective oxidation strategy needs synergistic effects among many fundamental parameters, such as the catalyst, hydrocarbon, solvent, pressure, temperature, and reaction mechanism, to achieve the optimal performance [26]. Current catalytic oxidation methods use relatively high temperatures and large amounts of toxic solvents or metal cocatalysts; this inevitably causes over-oxidation or side reactions because of the relatively weak allylic/benzylic C-H bonds in hydrocarbons and does not achieve the activities and selectivities that enzymes provide. Further improvements in engineering catalytic oxidation systems to improve their practical applications are therefore needed.

Since Ishii's group [28] first reported the use of N-hydroxyphthalimide (NHPI; Fig. 1) in this field in the 1990s, innovative oxidation strategies involving hydroxyimide organocatalysts have gained much attention because of their high catalytic activities in various selective oxidation processes. Several reviews have summarized their applications [17-27]. The main mechanistic pathways of these catalytic reactions, based on extensive experimental [17-27] and theoretical [29-40] studies, are shown in Fig. 2; hydrogen abstraction from C-H bonds by in situ-generated nitroxy radicals, with the assistance of metallic or metal-free initiators [21, 26, 36], is a vital step during catalysis. Facile decomposition to phthalimides, phthalic anhydrides, and other species at high temperatures (>80 ℃) are their main limitations [21, 41, 42]. There are also many technical obstacles that limit their sustainable industrial scale-up, including the frequent need of polar cosolvents (e.g., acetonitrile) and sacrificial initiators, poor solubility in hydrocarbons, difficult recovery from product mixtures, and the large amounts (typically 10 mol%) required to achieve good catalytic performances. These features have stimulated the development of a wide array of strategies for targeted improvements in the involved chemical processes through the design of tailored catalysts and catalytic systems. These strategies include decoration of NHPIs with lipophilic chains [43-46] to improve solubility, immobilization of NHPIs on solid metal-organic framework solids or silica [47-49] for solvent-free use, NHPI modification with ion-pair compounds [38, 50, 51] to give facile recovery and increased reactivity, use of NHPIs with electron-withdrawing substituents [34, 46] and multi-nitroxy organocatalysts [52-69] for enhanced activity, combined use of NHPI and photoinitiators such as naphthalene imides [70] or carbon materials [71] to avoid the use of sacrificial initiators, use of N-alkoxyphthalimides[40, 72] for thermal catalysis above 100 ℃ or light-induced catalysis at room temperature under initiator/metal-free conditions, and physical adsorption [73] and polymer-supported [74, 75] methods to improve recovery and recycling. Despite the continuing and extensive improvements in catalysis, efficient catalytic methods that have negligible environmental impacts, are economical, and address issues connected with long-term sustainability of these industrial processes have not yet been developed.

This review focuses on selective oxidation protocols with multi-nitroxy organocatalysts developed in the last 15 years. The catalytic oxidation efficiencies of multi-nitroxy organocatalysts such as N, N-dihydroxypyromellitimide (NDHPI) [53-59] and N, N', N''-trihydroxyisocyanuric acid (THICA) [52, 60-69] (Fig. 1) are higher than that of NHPI in most cases, even if smaller amounts of catalysts are used. These catalysts can directly use pressurized-air rather than pure oxygen, and can be used at relatively high temperatures without significant loss of catalytic reactivities, compared with NHPI, because of their multiple hydroxyimide groups and good thermal stabilities [52]. Their specific performances and tunable structures have led to many experimental and theoretical studies to probe the reaction mechanisms and to enable rational design of highly efficient analogous catalysts [37, 39]. To the best of our knowledge, a complete review focusing on multi-nitroxy organocatalysts is still unavailable. The important features described above are the main reasons for reviewing fundamental contributions in this area.

Our review is organized as follows. The first and second sections highlight advances in aerobic oxidation mediated by NDHPI and THICA, respectively. The third section summarizes major advances based on combinations of experimental and theoretical results, and provides rational guidance for designing new multi-nitroxy organocatalysts with diverse functions. The final section discusses potential multi-nitroxy organocatalysts based on the previous examples and their potential practical applications. This arrangement enables a reasonable understanding of the advantages and limitations of these catalysts in oxidation, and provides information on controlling their catalytic efficiencies. We hope that our review will serve as a convenient reference source for both academic and industrial chemists. The methods described offer good alternatives to classical oxidation processes in terms of reactivity. We also explore the problems that still need to be solved such as the need for harsh conditions and the use of polar solvents.

A similar procedure to that for preparing NHPI [21] can be used for the easy preparation of NDHPI (Fig. 1) by treatment of pyromellitic anhydride with hydroxylamine hydrochloride (NH₂OH·HCl) (Eq. (1) ) [56]. NDHPI can therefore be widely used in various oxidation processes.

(1)

Early studies of NDHPI-promoted oxidation were reported by Ishii's group [53-55] between 2002 and 2003. Their use of air instead of pure oxygen is attractive for industrial applications. 3-Methylquinoline was oxidized to 3-quinolinecarboxylic acid [53] in 80% yield using 5 mol% NDHPI; this yield is 5% higher than that obtained using 20 mol% NHPI. These catalysts were used in the presence of the same small amounts of Co(OAc)₂ and Mn(OAc)₂ as cocatalysts and NO₂ as a radical initiator under dioxygen at normal pressure (Eq. (2) ). This NDHPI-catalyzed oxidation was completed with 68% yield in 5 h using air at 2 MPa at 120 ℃. This yield is 8% lower than that achieved using NHPI, but the amount of catalyst was much smaller, which is important in practical synthetic applications. The NHPI-catalyzed aerobic oxidation of methylquinolines suggests that NDHPI would perform well in various oxidations.

(2)

This conjecture was verified by the aerobic oxidation of ethane to acetic acid [54]. In the presence of Co(OAc)₂ at 150 ℃, NDHPI solely afforded 830 μmol of acetic acid; this is 200 μmol more than that obtained using twice the amount of NHPI. This is mainly because of the facile thermal decomposition of NHPI [21, 41, 42]. The yield of acetic acid increased to 888 μmol when the temperature was increased from 120 to 170 ℃. CoCl₂ was found to be the best cocatalyst among the tested Co(II) salts (Eq. (3) ). The possible hydrolysis of acetonitrile to yield acetic acid was eliminated based on experiments using deuterated acetonitrile in the oxidation. The yield and turnover number of this oxidation in propionic acid were about twice those in acetonitrile, and only a small amount of propionic acid was aerobically oxidized to acetic acid. This indicates that aliphatic carboxylic acids are good solvents for the conversion of light alkanes to the corresponding carboxylic acids.

(3)

Ishii's group [55] tested the catalytic performances of NDHPI in the aerobic oxidation of nitrotoluenes in acetic acid (Eq. (4) ). For all o-/m-/p-nitrotoluenes, nitrobenzoic acids were the main products along with small amounts of nitrobenzaldehydes; the yields were similar for both NHPI and NDHPI. Oxidation of the methyl group of o-nitrotoluene is difficult because of steric hindrance by, and the electron-withdrawing effect of, the adjacent NO₂ group. A catalytic amount of NO₂ was therefore added to accelerate the oxidation of o-nitrotoluene by NDHPI/Co(OAc)₂/Mn(OAc)₂; the yield of o-nitrobenzoic acid was 29% lower than those of p-/m-nitrobenzoic acids. This method provides an alternative practical green route to nitrobenzoic acids.

(4)

The use of NDHPI has been extended to the oxidation of various benzylic hydrocarbons under dioxygen at normal pressure. In the oxidation of p-xylene (Eq. (5) ), Espenson's group [56] found that under comparable conditions, the initial reaction rate with NDHPI was 33% higher than that with NHPI, and the same yield (85%) of terephthalic acid obtained using 6.5 mmol NDHPI was achieved using only 10 mmol NHPI. However, the catalytic performance of NDHPI was lower than that of NHPI in the oxidation of cumene to acetophenone and cumyl alcohol with Cu(I) or Co(II) salts as cocatalysts in the presence or absence of azodiisobutyronitrile (AIBN) as a radical initiator (Eq. (6) ) [58, 76]. This may be partly caused by the use of acetonitrile rather than acetic acid for the NDHPI catalysis or the particularity of the isopropyl group in cumene.

(5)

(6)

Shibamoto and Iwahama [57] have patented many gradient heating processes for producing aromatic polycarboxylic acids via NDHPI-catalyzed oxidation of durene (Eq. (7) ), which affords pyromellitic acid in yields up to 91%. This method is also applicable to the oxidation of 3, 3', 4, 4'- tetramethylbenzophenone to benzophenone-3, 3', 4, 4'-tetracarboxylic acid, giving 80% yield. Yuan's group [59] immobilized NHPI and NDHPI on mesoporous SBA-15 molecular sieves, namely NHPI/SBA-15(ip) and NDHPI/propyl-2-propanol-SBA-15, and tested their catalytic performances in the oxidation of toluene with or without Co(salen) (Eq. (8) ). The yield of benzoic acid with NDHPI/Co(salen) was higher than that with NHPI/Co(salen), and the selectivity for benzoic acid was higher. Without acetonitrile, the selectivities for benzoic acid and benzaldehyde were almost equal, and the conversion was low. After catalyst immobilization, the performance of NDHPI/propyl-2-propanol-SBA-15 was similar to that of NHPI/SBA-15(ip) and was unchanged after recycling, unlike the latter case. However, the higher product selectivity changed from benzoic acid to benzaldehyde, along with a significant decrease in conversion, when Co(salen) was not used. When NDHPI/propyl-2-propanol-SBA-15 was used with Co(salen), the conversion and the selectivity for benzoic acid increased greatly, and increased to some extent after the catalyst was recycled. The effects of the cocatalysts, molecular sieves, and solvent on these catalytic systems are therefore significant, and further mechanistic insights are needed to fully understand their exact roles.

(7)

(8)

In summary, a combination of NDHPI with small amounts of metal cocatalysts in an atmosphere of pure oxygen or pressurized air gives facile selective oxidation of hydrocarbons to the corresponding carboxylic acids with good efficiencies. This method is more practical for large-scale applications than use of NHPI. The radical-chain properties of NDHPI catalysis also improve its other applications such as the stereospecific polymerization of methyl methacrylate [77] and the selective metal-free benzylic monofluorination of aryl hydrocarbons in the presence of superstoichiometric amounts of Selectfluor [78].

A literature survey showed that there are many routes for preparing THICA from its precursors, namely alkoxy or aryloxy derivatives of THICA, followed by hydrogenation with H₂ on Pd/C; these have been briefly reviewed [60, 61]. Selected representative synthetic routes are given in Eq. (9). The availability of synthetic routes for THICA enables its wide application in practical oxidation. THICA is structurally similar to a reported safe and efficient reagent for chlorination and oxidation, namely trichloroisocyanuric acid [79], and X-ray diffraction has been used to determine the structures of its dihydrate 4-aminouracil and 4-aminouracil monohydrate [80].

(9)

The promotional effects of THICA in catalytic oxidation of hydrocarbons have mainly been reported since 2003 by Ishii's group [52]. In a typical THICA-catalyzed oxidation of toluene at 80 ℃ in acetic acid, the yield of benzoic acid was two or three times higher than that obtaining using NHPI in the same catalytic amount. The yield of benzoic acid increased to 99% with increasing catalytic amount from 1 to 5 mol% or increasing temperature from 80 to 100 ℃ (Eq. (10) ). Unlike the case for NHPI, no oxidation occurred at 25 ℃ for THICA; this is because of its strong O-H bond. THICA was also found to be efficient in the oxidation of various m-/p-substituted toluenes to the corresponding benzoic acids in excellent yields at 100 ℃; the yield of terephthalic acid from oxidation of p-xylene was over 95%, i.e., 10% higher than that obtained using NDHPI [56]. Time-dependence curves indicate that the nitroxy radicals derived from THICA were more reactive and stable than those derived from NHPI.

(10)

Based on these pioneering studies, the use of THICA has been extended to the selective oxidation of other benzylic hydrocarbons. An almost quantitative yield (>99%) of p-anisic acid was obtained in the oxidation of p-methylanisole catalyzed by 3 mol% THICA [60]; this is ca. 16% higher than that achieved using 10 mol% NHPI. THICA promoted the oxidation of 2-methylnaphthalene to its carboxylic acid, with air as the sole oxidant (Eq. (11) ) [60], with a yield 66% higher than that achieved using NHPI. THICA was also efficient in the oxidation of 1, 2, 3-, 1, 2, 4-, and 1, 3, 5-trimethylbenzenes under air at 2 MPa, giving the corresponding benzenetricarboxylic acids in good yields (81%-97%) (Eq. (12) ) [62]. ZrO(OAc)₂, in particular, greatly accelerated the oxidation of 1, 2, 4-trimethylbenzene to give a high yield (97%) of 1, 2, 4-benzenetricarboxylic acid. This effect was not clearly observed in the oxidation of 1, 3, 5-trimethylbenzene. In addition, the THICA/Co(OAc)₂ catalytic system was superior to conventional methods for producing acetylbenzoic acids. Acetophenones substituted with alkyl, alkoxy, acetoxy, and halogen groups were selectively converted to the corresponding benzoic acids in yields ranging from 82% to 97% [63]. The NHPI/Co(OAc)₂ and THICA/Co(OAc)₂ catalytic systems both improve the yields of acetylbenzoic acids. A combination of THICA and a small amount of Mn(OAc)₂ can selectively promote conversion of the acetyl group of a substrate to the corresponding carboxylic acid; for example, a 93% yield of terephthalic acid was obtained using THICA/Co(OAc)₂/Mn(OAc)₂ (Eq. (13) ).

(11)

(12)

(13)

In subsequent studies, various cocatalysts were tested in oxidation with THICA to identify alternatives to the commonly used Co(OAc)₂ and/or Mn(OAc)₂. In 2010, Zhou's group reported a metal-free catalytic system with THICA for the oxidation of various substituted toluenes in the presence of HNO₃ as a radical initiator. They also developed a Pd-free approach to preparing THICA, with a total yield of 46% [64]. The use of a combination of THICA and concentrated HNO₃ promotes the oxidation of nitrotoluenes, e.g., 2, 4, 6-trinitrotoluene (Eq. (14) ), to the corresponding carboxylic acids under O₂ at 0.2 MPa and 100 ℃, with yields up to 99%. Acetic acid and perfluoronaphthalene are excellent solvents in the oxidation of 2, 4-dinitrotoluene and replacing pure dioxygen by low-pressure air did not affect the acid selectivity. The radical signal of THICA (g = 2.00075, A_N = 4.86G) was observed in electron spin resonance (ESR) measurements, which confirmed the production of THICA radicals in HNO₃ during oxidation.

(14)

In 2009, oxime initiators, especially dimethylglyoxime (DMG), were used for the first time in NHPI-catalyzed oxidation [81]. Although DMG is structurally similar to NHPI, the feasibility of its use as a radical initiator for NHPI is mainly based on its favorable production of nitroxy radicals because of its weaker >NO-H bond. This result also suggests that a synergistic combination of two or more hydroxyimide organocatalysts with different >NO-H bond strengths could enable oxidation without the addition of any radical initiators; this was verified recently by Li's group [82]. Based on these results, Lu's group [65] evaluated the performance of a THICA/DMG catalytic system in the oxidation of toluene derivatives, and found that the performance of the PEG₁₀₀₀-based dicationic acidic ionic liquid (PEG₁₀₀₀-DAIL, PEG is poly(ethylene glycol)) was better than that of other common solvents and PEG_n-DAILs (n = 400, 600, and 2000) (Eq. (15) ). The replacement of DMG by anthraquinone, AIBN, or HNO₃ greatly reduced the catalytic performance of THICA. We conclude that the complex interactions between PEG₁₀₀₀-DAIL and these radical initiators decrease the activity of THICA. THICA and PEG₁₀₀₀-DAIL can be easily recovered and reused without any loss, and the conversion was unchanged after at least eight cycles, except in terms of the acid selectivity. Similarly, they found that MnO₂ promotes the THICA-catalyzed oxidation of toluene derivatives to the corresponding acids in PEG₁₀₀₀-DAIL under mild conditions (Eq. (16) ) [67]. Both THICA and PEG₁₀₀₀-DAIL can be reused after simple separation, and no noticeable changes in the conversions, except in terms of acid selectivity, were observed during eight consecutive cycles. It should be noted that the performances of PEG_n-DAILs (n = 400, 600, 1000, and 2000) were better than those of regular imidazolium-cation-based ionic liquids.

(15)

(16)

The above review shows that the number of radical initiators suitable for THICA is greater than that for NDHPI, and the catalytic efficiency of THICA is higher than that of NDHPI in most cases. The use of diethylglyoxime as a non-metallic radical initiator and polymeric ionic liquids as recoverable solvents for THICA catalysis not only facilitates the selective oxygenation of hydrocarbons to carboxylic acids in excellent yields, but also reduces the environmental impact and costs, and avoids the corrosion caused by metals and acetic acid; these factors contribute to the sustainability of large-scale practical oxidation processes. Preparations of phenols and ketones by THICA- catalyzed oxidation of alkylbenzenes have been patented [68, 69]. In addition to its use in aerobic oxidation of hydrocarbons, THICA can also be used in the dehydrogenation of dihydropyrimidinone to pyrimidinone [83] and the metal-free aerobic oxidation of aromatic primary amines to the corresponding oximes in presence of acetaldoxime (Eq. (17) ) [66].

(17)

Although THICA and NDHPI are more efficient than NHPI in aerobic oxidation, few examples of such reactions have been reported. It is therefore important to establish a strategy for designing multi-nitroxy organocatalysts with improved catalytic properties. Advances in computational chemistry have provided a promising method for achieving this goal and enabling a better understanding of the experimental observations and a better overview of their intrinsic advantages and limitations; these have been the subjects of several recent theoretical studies [36, 37, 39].

The reactivities of nitroxy radicals of catalysts in hydrogen abstraction from model substrates, i.e., toluene or propane, follow the order THICA > NDHPI > NHPI [31, 36]. This can be explained by the observed higher catalytic efficiencies of THICA and NDHPI compared with that of NHPI [52-69]. It has been proposed that the multi-radicals upon removal of multiple O-H hydrogen atoms from THICA and NDHPI may significantly enhance their catalytic efficiencies because of their electron-withdrawing effects, and also increase their opportunities for substrate activation. However, further experimental evidence of the effects of these multi-radicals on the catalytic kinetics is still needed. Toluene is not oxidized by THICA at room temperature [52]; this is probably because radical generation is thermodynamically and kinetically unfavorable, and coexistence of the planar conformer in aprotic solvents is favored [36]. Additionally, the enthalpies of formation (Δ_fH°_{298.15 K}) and atomization (Δ_aH°_{298.15 K}) of THICA were accurately calculated for the first time; the values are -120.9 ± 1.4 and -1487.3 ± 1.4 kcal/mol, respectively.

Because the structure of THICA is too specific for further optimization, and the structural differences among THICA, NDHPI, and NHPI are responsible for their different performances, a series of multi-nitroxy organocatalysts (M-n, n = 3-10; Fig. 3) were designed, and their catalytic activities and structures were systematically investigated theoretically by Li's group [37, 39]. These catalysts can be easily prepared by treatment of the corresponding anhydrides with NH₂OH·HCl; for example, the reactions between mellitic trianhydride [84] or naphthalene[1, 8:4, 5]tetracarboxylic acid dianhydride [85] and NH₂OH·HCl afford M-6 (Eq. (18) ) and N, N- dihydroxynaphthalenediimide (M-3), respectively. THICA and M-3, unlike NHPI, NDHPI and M-n (n = 4-10), but like N-hydroxynaphthalimide (NHNI) [86, 87], are unsuitable for use in solvent- or initiator-free catalysis because of the strong intramolecular hydrogen-bonding interactions between the H atoms of their OH groups and O atoms of the adjacent carbonyl groups. As in the case of NHNI [58], oxidation by M-3 may hardly occur at 40 ℃. The >NO-H bond strengths in most of the catalysts (Fig. 3) are higher than those in NHPI and NDHPI, or comparable to that of THICA. Intensive studies of the catalytic efficiencies of hydroxyimide organocatalysts without radical initiators have shown that a suitable catalyst needs an appropriate >NO-H bond strength for both high H-abstraction reactivity and sufficiently reactive nitroxy radicals [22, 30, 31]; this has been reported to be 85-95 kcal/mol based on the C-H bond strengths of most organic compounds [23]. A combination of the experimental and theoretical differences (ca. 7 kcal/mol) among the >NO-H bond strengths shows that the estimated experimental >NO-H bond strengths of the compounds in Figure 3 are appropriate, except in the cases of THICA and M-3. The catalytic activities of THICA and M-3 can be improved by using radical initiators or high temperatures, based on experimental results [52, 58, 60-69, 86, 87].

There is a negative linear relationship between the H-abstraction reactivity and endothermicity of H abstraction. The H-abstraction reactivities of most catalysts (Fig. 3) are higher than that of NDHPI, or even comparable to that of THICA. These reactivity profiles indicate that these catalysts could significantly lower the temperature at which similar performances to those of known catalysts can be achieved, and can more efficiently promote oxygenation of less reactive substrates. The ion-pair-modified case (M-5) could serve as a potential new model catalyst for further improvements because of its recycling properties and high reactivity toward toluene, or even inert propane. Design strategies based on the following points will be helpful in improving H-abstraction activities: (1) Increasing the number of intramolecular conjugated hydroxyimide or nitroxy radical moieties on the same aromatic ring, or doped N atoms and ion-pair groups on the aromatic ring between hydroxyimide groups can increase the catalytic activity; (2) Appropriate enlargement of the conjugated aromatic systems between hydroxyimide groups does not change the activity, which is still higher than that of NHPI. These strategies also indicate the feasibility of including hydroxyimide groups on heterogeneous carbon materials to further increase their oxidation power.

Based on the above proposed design strategies, several novel model multi-nitroxy organocatalysts with potentially high performances have been designed. Their potential practical applications are discussed in this section.

The widespread occurrence of enzymes such as cytochrome P450 and peroxidase, which can catalyze important redox reactions by activating molecular oxygen in biological systems, has resulted in the development of bioinspired strategies for selective catalytic oxidation [88]. Biomimetic catalysts such as metalloporphyrins [88, 89], metal acetylacetones [90], and metal Salens [91, 92] have simple active sites based on those in enzymes, and are efficient in various oxidations under mild conditions. Their catalytic performances can be tuned based on feasible combinations of different macrocyclic ligands, axial ligands, and ion-pair or regular substituents on the ligands and central metals. Modification of these biomimetic catalysts with chemically conjugated hydroxyimide organocatalysts affords biomimetic multi-nitroxy organocatalysts, which combine the merits of each catalyst, and have good potential for future developments in the oxidation field. Two model examples (M-11 and M-12) of this type of catalyst are shown in Fig. 4. There are a few reports of studies of relevant catalytic systems such as Hemin/NHPI [93, 94] and silica-supported cobalt(II)salen/ NHPI [95, 96], which are highly selective in the oxidation of benzylic hydrocarbons to the corresponding ketones. However, chemical combinations of hydroxyimide organocatalysts and biomimetic catalysts have not yet been reported. Further theoretical and experimental studies of the performances of biomimetic multi-nitroxy organocatalysts are therefore needed.

Initial attempts have been made to immobilize NHPI on heterogeneous polymers to improve its recovery and recyclability [74, 75]. The earliest example was polystyrene-bound NHPI resin for preparing complex alkoxyamines [97]. Since then, NHPI has been immobilized on cross-linked copolymer microspheres of glycidyl methacrylate and methyl methacrylate [74], aminomethyl- or chloromethyl-substituted polystyrene [75], and cross-linked polystyrene microspheres [98, 99]. These polymer-supported NHPIs have been successfully used in the oxidation of compounds such toluene, ethylbenzene, p-methoxytoluene, and cyclohexane with molecular oxygen. However, their catalytic performances are lower than that of NHPI, possibly because of complex interactions between the polymers and NHPI or monosubstitution of NHPI. We therefore suggest that in the future hydroxyimide organocatalysts could be linked to polymer units to give polymulti-nitroxy organocatalysts. A relevant model example (M-13) is shown in Fig. 4; modification by ion-pair substituents on NHPI greatly improves both the reactivity and ketone selectivity [38].

Metal-free carbon materials such as mesoporous g-C₃N₄ [71, 100], graphene [13, 101, 102], carbon nanotubes [103], and mesoporous carbon nitrides [104] have high catalytic efficiencies in the high-temperature (120-160 ℃) oxidation of compounds such as cyclic olefins, toluene derivatives, ethylbenzene, and fluorene to the corresponding aldehydes or ketones, mainly with selectivities greater than 99%; these have been reviewed recently [26]. Based on the theoretical studies discussed in section 4, we suggest that hydroxyimide organocatalysts could be incorporated into heterogeneous carbon materials to give heterogeneous carbo-multi-nitroxy organocatalysts. A model example (M-14) is shown in Fig. 4. The performances of the NHPI/g-C₃N₄ photocatalytic system in the selective allylic oxidation of cholesteryl acetate [71] and the facile activation of molecular oxygen on carbon nanotubes [103] have already been studied. We therefore believe that the catalytic power of such new catalysts would benefit from the tunable porous structures, good chemical and thermal stabilities, and environmental compatibility of carbon materials and the high reactivity of multi-nitroxy organocatalysts.

Strategy for designing N-alkoxy multi-nitroxy organocatalysts for metal-free oxidation, based on modification of the alkyl substituents on the active >NOH groups of multi-nitroxy organocatalysts, is also attractive. Two model examples (M-15 and M-16) are shown in Fig. 4. Recently, Lucarini's group reported the performances of N-benzyloxy-, N- diphenylmethyloxy-, and N-triphenylmethyloxy-phthalimides in the oxidation of cumene and benzyl alcohol under photoirradiation at room temperature [72]. None of these catalysts undergo thermal homolysis of their >NO-C bonds to afford benzylic radicals and nitroxy radicals at 100 ℃, and only the last two undergo light-induced >NO-C bond homolysis. It should be noted that N-acetyloxyphthalimide undergoes thermal homolysis at 190 ℃ [105]. Structure-performance relationships, based on our recent theoretical studies of the structural and thermochemical properties of a series of N-alkoxyphthalimides and N- acetyloxyphthalimides [40], show that most such catalysts can promote oxidation via homolysis to coform nitroxy radicals and alkyl radicals above 100 ℃ or by photoirradiation at room temperature. This strategy is therefore important in various practical applications, e.g., by lowering the reaction temperature, using visible light, improving the solubility in hydrocarbons, avoiding the use of sacrificial initiators, metals, and polar solvents, accelerating the initial oxidation, and improving the selectivity as a result of coformation of alkyl radicals of the targeted substrate, or in advancing their industrial applications. Many analogous N-alkoxyphthalimides have already been synthesized [106-109], but their catalytic applications in aerobic oxidation have not yet been reported.

Innovative oxidation strategies using hydroxyimide organocatalysts are useful in various selective oxo functionalizations of hydrocarbons, and provide alternative routes for establishing efficient catalytic oxidation systems that are both green and economically viable. Efficient hydroxyimide organocatalysts for targeted substrates will be developed in the future, and the gap between academic studies and industrial applications will gradually decrease as fundamental knowledge based on complete mechanistic understanding is acquired.

This review illustrates the significant experimental and theoretical advances that have been made in aerobic oxidation of hydrocarbons under air or molecular oxygen catalyzed by multi-nitroxy organocatalysts, one type of highly active hydroxyimide organocatalyst. Their potential applications were predicted, with the aim of stimulating further work in this area. On the basis of the examples given, theoretical studies, and predicted applications, appropriate catalytic systems for the selective oxidation of specific substrates can be identified, e.g., THICA/Co(OAc)₂ [52], THICA/HNO₃ [64], and THICA/DMG [65] catalytic systems for the oxidation of toluene, 2, 4, 6- trinitrotoluene, and p-methylanisole to produce the corresponding carboxylic acids in yields above 99%. The catalytic efficiencies can also be controlled by using different combinations of catalysts, solvents, temperatures, and other parameters, and rationally designing catalysts with specific experimental properties, e.g., enhanced reactivity or facile recovery. The use of polymeric ionic liquids as solvents [65, 67], small organic compounds [65] as radical initiators, and N-alkoxyphthalimides in light-induced oxidation [72] has opened up promising routes for the further development of numerous highly efficient metal-free or initiator-free catalytic systems with multi-nitroxy organocatalysts; this would decrease the high dependence on the specific instruments or reactors that are required for traditional oxidation methods. We are confident that our review will be helpful for all scientists working in the field of selective oxidation and will advance nitroxy radical chemistry.

Although the reported multi-nitroxy organocatalysts (Fig. 2) show excellent activities in a diverse array of selective oxidations, and theoretical calculations show that the H-abstraction reactivities of the designed catalysts are higher than those of most hydroxyimide organocatalysts, relevant examples are limited. Systematic investigations are therefore still needed. Further attention should be paid to their applications at relatively high temperatures or in acidic media because thermal self-decomposition is a common limitation of hydroxyimide catalysts [21, 41, 42]. Further experimental evidence of the presence of reactive multi-radicals derived from multi-nitroxy organocatalysts during oxidation and their effects on catalytic kinetics is needed. This could be obtained using spectroscopic methods such as ESR under certain conditions. The acidities induced by certain solvents and their impacts on catalysis need to be resolved. It has been reported that the acidities of hydroxyimide organocatalysts in different solvents vary, with pK_a values ranging from 3.28 to 41 [110-112], and the deprotonation of NHPI may lead to its deactivation during oxidation [30, 42]. In addition, although the model multi-nitroxy organocatalysts shown in Figs. 3 and 4 were designed to have convenient recyclability, enhanced reactivity, good stability, or good photocatalytic activity at room temperature, their performance needs to be verified experimentally and their structures should be improved accordingly. Global intensive studies of their physical chemistry and catalytic performances are important, and use of these catalysts in selective oxidation may become mainstream. We believe that the use of multi-nitroxy organocatalysts will be significant in the scaling-up of other processes, e.g., polymerization [77], fluorination [78], and dehydrogenation [83].