Hybrid Materials Based on Imidazo[4,5-b]porphyrins for Catalytic Oxidation of Sulfides

Abstract

Heterogenized metalloporphyrin catalysts for oxidation reactions are extensively explored to improve chemical production. In this work, manganese meso-tetraarylporphyrins were immobilized on hydrated mesoporous titanium dioxide (S_BET = 705 m² g⁻¹) through carboxylate or phosphonate anchoring groups separated from the macrocycle by the 2-arylimidazole linker fused across one of the pyrrolic rings of the macrocycle. The element composition of two mesoporous hybrid materials thus obtained were investigated and the integrity of the immobilized complexes was shown by different physicochemical methods. Finally, the catalytic efficiency of the more stable material Mn(TMPIP)/TiO₂ with the phosphonate anchor was evaluated in the selective oxidation of sulfides to sulfoxides by molecular oxygen in the presence of isobutyraldehyde (IBA). The heterogenized complex has shown excellent catalytic activity exhibiting a turnover (TON) of ~1100 in a single catalytic run of the sulfoxidation of thioanisole. The catalyst was successfully reused in seven consecutive catalytic cycles.

1. Introduction

Oxidation reactions are key technological processes for converting petroleum into pharmaceuticals, agrochemicals, and fine chemicals [1,2]. Among all transformations of organic compounds, oxidation reactions are especially challenging because they often have low selectivity and require stoichiometric amounts of toxic and hazardous oxidizing reagents such as oxometal oxidants or peracids [3]. Thus, these processes are among the most dangerous and pollutant transformations in chemical production. The development of reactions in which molecular oxygen is used as a terminal oxidant, together with a reusable catalyst and a nontoxic solvent, is extremely important for both economic and environmental reasons [4]. It is not surprising that catalytic aerobic oxidation reactions performed with both non-metal and metal catalysts have received great attention [5,6].

In nature, such processes are carried out in a highly selective manner by mono- and dioxygenases under mild conditions [7,8]. This stimulated investigations of biomimetic porphyrin catalysts pioneered by Groves and co-workers, who studied the oxidation of organic compounds by strong oxidants in the presence of iron(III) porphyrinates in the late seventies of the last century [9,10,11]. Later extensive studies on the metalloporphyrin-catalyzed oxidation have demonstrated that the catalytic efficiency of manganese(III), ruthenium(III) and cobalt(III) porphyrins can even exceed that of iron(III) complexes in part due to the high stability of these complexes [12,13,14,15,16]. Another remarkable feature of metalloporphyrins is that these catalysts are efficient in different types of oxidation processes, such as the epoxidation and hydration of alkenes, the hydroxylation of alkanes, the dehydrogenation of amines, and the oxidation of arenes, amines, sulfides, alcohols, and aldehydes and so forth [8,17]. The reaction rate and selectivity can be controlled by varying the nature of the central metal ion and substituents at the periphery of the porphyrin macrocycle. The peripheral substituents also influence the catalyst stability, and meso-tetraarylporphyrins bearing ortho-disubstituted aryl residues and electron-deficient aryl groups exhibit high efficiency in part due to their increased chemical stability [18,19]. Lately, major interest in these reactions appeared when it was shown that metalloporphyrins could catalyse reactions with environment-friendly oxidants such as hydrogen peroxide or molecular oxygen which can be used in the presence of isobutyraldehyde (IBA) as a sacrificial reagent or in photooxidation processes [8].

Unfortunately, the use of metalloporphyrins in practical organic synthesis is still limited because these catalysts are expensive and often have a disappointingly low TON and TOF. It was recognized that the heterogenization of homogeneous catalysts significantly increased their efficiency. In fact, solid catalysts based on metalloporphyrins commonly display higher stability than their homogeneous counterparts because the deactivation of metalloporphyrin catalysts mainly proceeds through the formation of catalytically inactive dimeric complexes. The catalyst dimerization can be decreased when porphyrin molecules are attached to a solid support, in particular through covalent bonding [20]. Being heterogenized, the expensive metalloporphyrins can be recovered after the reaction completion and reused in the next catalytic cycle. This simplifies product isolation and could decrease the cost of industrial production. Thus, recent research was mainly focused on preparing efficient heterogenized porphyrin catalysts [21].

Several different strategies were investigated for the immobilization of porphyrin complexes onto inorganic and organic supports [22,23,24,25,26,27,28,29], such as zeolites [30,31], clays [32,33,34,35], silicas [36,37,38,39], polymers [20,40], metal-organic frameworks (MOF) [21,41,42,43,44] and covalent organic frameworks (COF) [45]. Metalloporphyrin catalysts were also grafted onto photoactive anatase and a nanocrystalline anatase/hydrozincite composite material using triethoxysilane anchoring group [46] or electrostatic interactions [47], respectively. Surprisingly, the photocatalytic decomposition of porphyrins immobilized on these photoactive supports was slow, and the materials thus obtained were stable enough to be used as reusable catalysts. Recently, the method for the preparation of hybrid materials bearing two porphyrin complexes incorporated into charged layer supports was developed [47,48]. Most of the previously reported heterogenized complexes were efficient as catalysts, and a synergetic action of support and transition metal catalysts was observed in some cases [43,44,48]. It was also shown that the immobilization of metalloporphyrin catalysts could switch the chemoselectivity of oxidation reactions [47,49].

Porous materials with high surface areas are widely explored in catalysis because they allow us to increase catalyst loading and diminish the influence of the physical phenomenon of mass transfer on reaction rates [50,51]. Different types of porous materials were investigated, but all of them have serious drawbacks. Mesoporous silicas commonly display low chemical stability [52,53]. MOF- and COF-bearing catalytic centers in porphyrinic linkers are expensive because a high amount of metalloporphyrins is required for their preparation, and drying procedures leading to porous crystalline networks are very troublesome and time-consuming [54]. Microporous zeolites are also inconvenient as solid supports for porphyrins because the size of porphyrin molecules is close to that of most available zeolite Y supercages (~1.3 nm in diameter) and the diffusion of organic compounds through zeolite lattice apertures is sterically hindered in these hybrid materials [55]. There is, thus, a need for more efficient immobilization strategies for preparing porous catalysts based on metalloporphyrins.

In this work, we report grafting porphyrin complexes on the surface of hydrated mesoporous titanium(IV) oxide (TiO₂, S_BET = 650–705 m² g⁻¹). This cheap and photocatalytically inactive support was recently prepared by us using a non-templating sol-gel process. Our grafting methodology is based on the use of imidazo[4,5-b]porphyrins as molecular precursors (Figure 1). These fused porphyrin-heterocycle compounds, which were pioneered by Crossley et al. [56], were widely studied but were not used for the preparation of heterogenized catalysts [57]. The 2-aryl-imidazole residue is of interest as a linker in the synthesis of functional materials because the introduction of different substituents in the aryl residue of this linker is straightforward [56,58]. This long and rigid spacer provides a perpendicular orientation of the porphyrin ring with respect to the solid surface. The synthetic strategy based on the introduction of the 2-arylimidazole residue fused across one pyrrolic ring of the macrocycle opens a way for covalent grafting meso-tetraarylporphyrins, in particular their derivatives with sterically hindered o-disubstituted aryl groups which are known to be most efficient catalysts for the oxidation reactions [59]. In this work, this strategy was used for grafting porphyrin catalysts onto titanium dioxide through a carboxylate or phosphonate anchoring group, as shown in Figure 1. The element composition of two mesoporous hybrid materials thus obtained were investigated, and the integrity of the immobilized molecules was shown using different physicochemical techniques. Finally, the catalytic efficiency of the more stable material Mn(TMPIP)/TiO₂ was evaluated in the selective oxidation of sulfides to sulfoxides by molecular oxygen in the presence of IBA. The Mn(TMPIP)/TiO₂ material has shown excellent catalytic activity and was successfully reused in seven catalytic cycles.

2. Results and Discussion

2.1. Synthesis of (2-arylimidazo[4,5-b]porphyrinato)manganese(III) Chlorides and Their Characterization

meso-Tetramesityl-substituted imidazo[4,5-b]porphyrins were chosen for investigation in this work as their synthesis is reproducible, and they can be obtained in a 100 milligram scale using available laboratory equipment. The free base 2-arylimidazo[4,5-b]porphyrins 2H(TMPIC) and 2H(TMPIP) bearing carboxylate and phosphonate anchoring groups on the aryl substituent of imidazole were prepared through the oxidation of corresponding β-aminoporphyrin to 2,3-dioxochlorin followed by the Debus–Radziszewski condensation of this dione with aromatic aldehydes (Scheme 1) [58].

Insertion of manganese ions was carried out using synthetic procedures developed for 5,10,15,20-tetraphenylporphyrin [60]. Free base porphyrins 2H(TMPIC) and 2H(TMPIP) were heated with an excess of manganese(II) chloride (five equiv equivalents) in DMF under air until the complete consumption of the starting compound. Complexes Mn(TMPIC) and Mn(TMPIP) were isolated by column chromatography on silica gel and characterized by HRMS-ESI spectrometry, FTIR and UV-vis spectroscopies, with the last one being the most useful spectroscopic technique. The complexes display UV-vis spectra characteristic of manganese(III) porphyrins showing specific shapes compared to other porphyrin complexes due to strong metal-porphyrin π interactions [61,62]. The mixing of metal and porphyrin orbitals leads to the appearance of a number of charge-transfer bands which alter classical electron absorption spectra of the porphyrin complexes in which only π–π* transitions of the macrocycle are observed. For example, in the spectrum of Mn(TMPIP) in CHCl₃, several bands appear in the region of 300–450 nm with the maxima at 312, 382, and 402 nm (the region of VI and V bands) and a sharp charge-transfer band is observed at 484 nm (log ε = 4.92, V band; see Supporting Information (SI), Figure S1). Moreover, less intensive bands appear at 576 nm (IV band) and 622 (III band) and 749 (II band). The successful insertion of manganese ion was also confirmed by observation of [M+H]⁺, [M–Cl]⁺ and [M–Cl+H]²⁺ ion peaks with matching isotopic patterns as the most prominent peaks in the high-resolution mass spectra of complexes Mn(TMPIP) and Mn(TMPIC) (see Supplementary Materials, Figures S2 and S3).

DFT calculations were employed to gain insight into electronic structures for the frontier orbitals of complexes Mn(TMPIC) and Mn(TMPIP) and (5,10,15,20-tetraphenylporphyrinato)manganese(III) chloride (Mn(TPP)). The computations were performed using B3LYP functional using 6-31G* basis set for all atoms, including all-electron set for Mn. As shown in Figure 2, efficient mixing of metal and porphyrin orbitals is observed in HOMO and LUMO orbitals. In complexes Mn(TMPIC) and Mn(TMPIP), the electronic density of HOMO is also located on the fused imidazole fragment. In contrast, the 2-aryl residue is not involved in the HOMO and LUMO electronic states of both imidazoporphyrins. Thus, the introduction of any anchoring group does not significantly change the electronic structure of 2-arylimidazo[4,5-b]porphyrins and energies of frontier molecular orbitals. This appears to indicate that electronic communication between the catalytic site and any solid support may be weak when a 2-arylimidazole linker is used for preparing heterogenized catalysts.

Complexes Mn(TMPIC) and Mn(TMPIP) were found to be soluble in methanol and chlorinated solvents such as chloroform and dichloromethane, which is important for their use in material chemistry.

2.2. Immobilization of Manganese(III) Complexes Mn(TMPIC) and Mn(TMPIP) on TiO₂

Recent studies of hybrid organic-inorganic materials based on organophosphonates have revealed the great potential of metal oxide and phosphonate networks for preparing functional materials, in particular, due to their thermal and chemical stability [63,64,65,66]. Nevertheless, studies on TiO₂-supported transition metal catalysts are limited [67,68,69,70,71,72,73,74], probably because the synthesis of porous TiO₂ is challenging and time-consuming. On the other hand, the immobilization of organic chromophores onto porous titania films was widely explored in view of the production of clean energy from the sun in dye-sensitized solar cells [75]. Among different anchors, the carboxylate group was most commonly used and sometimes gave better results as compared to the phosphonate anchoring group. In this work, we compared the efficiency of the carboxylate and phosphonate anchors for the fabrication of heterogenized catalysts.

Mn(TMPIC) and Mn(TMPIP) were grafted onto the surface of hydrated amorphous titanium dioxide, which does not exhibit any photocatalytic activity on its own. Such cost-effective mesoporous TiO₂ (S_BET = 650–705 m² g⁻¹) is readily available by a template-free sol-gel method developed by us recently [76].

To prepare target materials from phosphonate-substituted complex Mn(TMPIP), we employed a sol-gel process in organic solvents [77,78] since the immobilization of phosphonic acids in aqueous media probably occurs through monodentate P–O–Ti interactions and yields less-stable materials [71]. Dialkyl phosphoester Mn(TMPIP) was first transformed into more reactive bis(trimethylsilyl)phosphonate by reacting the complex with a large excess (30 equivalents; a large excess was used to accelerate the reaction performed in a diluted solution) of bromotrimethylsilane (TMSBr) in CH₂Cl₂ for 2 d at room temperature (Scheme 2). The evaporation of the reaction mixture to dryness gave the target compound in a quantitative yield.

Then this moisture-sensitive complex was engaged in the next step, dissolving the solid in dry CH₂Cl₂ and adding this solution to a suspension of mesoporous titania in the same solvent at room temperature under Ar. The material was isolated by centrifugation after stirring for 2 d, thoroughly washed with THF, water, MeOH, and ether and dried under reduced pressure. In these experiments, the Mn/Ti molar ratio was varied between 1:100 and 1:15 to prepare a solid with spatially separated catalytic sites (1:100, Mn(TMPIP)/TiO₂) and to get insight into maximal loading of catalytic sites on this support (1:15, Mn(TMPIP)/TiO₂-1).

The studies on the element composition of hybrid materials thus obtained were started by EDX chemical analysis in STEM mode of solid Mn(TMPIP)/TiO₂ because this technique is well suited for the determination of the axial ligand in the heterogenized Mn(TMPIP). It was previously demonstrated that the treatment of porphyrinatogallium(III) chlorides with TMSBr in toluene led to the exchange of axial chloride ligands by bromide ions [79]. We hypothesized that this exchange reaction could also be observed under our experimental conditions. The EDX spectrum confirmed successful grafting of the Mn complex and the presence of all expected elements, with the only exception being chlorine (Figure 3 and Figure S5 (see Supplementary Materials)). We also observed an EDX peak corresponding to bromine. This indicated that the ligand exchange reaction was observed during the activation of the diethoxyphosphoryl group by TMSBr. The elemental mapping images shown in Figure 3 revealed the homogeneous distribution of all elements in the studied samples.

The empirical formulas of materials thus obtained were derived from the content of six elements (C, H, N, Ti, Mn and P; see Supplementary Materials, Table S1). The Mn/Ti ratio in the materials Mn(TMPIP)/TiO₂ and Mn(TMPIP)/TiO₂-1 was found to be 1:125 and 1:28, respectively. Thus, the content of heterogenized Mn(TMPIP) was lower than the expected values in both cases despite the fact that colored filtrates were obtained only during the synthesis of Mn(TMPIP)/TiO₂-1. To prove quantitative immobilization of Mn(TMPIP) onto the solid Mn(TMPIP)/TiO₂, the combined organic and aqueous filtrates obtained during washing the solid were evaporated and dried under reduced pressure. The residue thus obtained was dissolved in methanol and analyzed by UV-vis spectroscopy to check the presence of starting complex Mn(TMPIP) and corresponding phosphonic acid Mn(TMPIP-OH). These two compounds are the only porphyrin products that can be present in the solution if grafting is incomplete. To perform the qualitative analysis, complex Mn(TMPIP-OH) was prepared, as shown in Scheme 2. Based on the results of this UV-vis study, we safely concluded that more than 98% of complex Mn(TMPIP) was immobilized on a titania support. In contrast, the filtrates were intensively colored when Mn(TMPIP) and titania were reacted in a 1:15 molar ratio, and the experimental content of Mn in the target solid was about twice lower than was expected. This likely indicates that maximal catalyst loading (~1:30) was achieved under these experimental conditions.

To enable the immobilization of Mn(TMPIC), the complex was dissolved in dry CH₂Cl_2, and this solution was added to mesoporous titania suspended in the same solvent and stirred for 2 d at room temperature (Scheme 3).

The Mn/Ti molar ratio also varied between 1:100 and 1:15 for comparative reasons. The empirical formulas of Mn(TMPIC)/TiO₂ and Mn(TMPIC)/TiO₂-1 thus obtained were derived from the content of five elements (C, H, N, Ti and Mn) and were very similar to those of materials based on Mn(TMPIP) (Table S1). Quantitative immobilization was observed when complex and inorganic support were reacted in the ratio of 1:100, and the maximal catalyst loading was about 1:32.

The intact structure of the heterogenized complexes Mn(TMPIP) and Mn(TMPIC) was confirmed by UV-vis diffuse reflectance and FTIR spectroscopies. In brief, the spectral signatures of both manganese complexes before and after immobilization on TiO₂ were remarkably similar except for the characteristic vibration bands of carboxylic and phosphonic anchors. Kubelka–Munk transformed UV-vis diffuse reflectance spectra of the studied solids show a strong and broad absorption feature in the high energy region (<300 nm). This band was attributed to the TiO₂ phase. The spectral shape of materials in the 350–750 nm region resembles closely to those observed for Mn(TMPIC) and Mn(TMPIP) complexes, confirming that the metalloporphyrin residue keeps its integrity after immobilization (Figure 4).

In the FTIR spectra (see Supplementary Materials, Figures S6 and S7), vibration bands assigned to the porphyrin macrocycle (720–850 and 1000–1200 cm⁻¹) and aryl residues (1350–1610 cm⁻¹) are all much weaker as compared to the Ti–OH and O–H stretching modes (3500–3700 cm⁻¹) even in the spectra of Mn(TMPIP)/TiO₂-1 and Mn(TMPIC)/TiO₂-1 materials with high Mn/Ti ratios. Notably, the shape of Mn(TMPIP)/TiO₂-1 spectrum in the region of 850–1260 cm⁻¹, where characteristic bands associated with vibrations of the phosphonate substituent are observed, is significantly different from those of diethoxyphosphoryl-substituted complex Mn(TMPIP) and phosphonic acid Mn(TMPIP-OH). Valence vibration bands of P=O residue of the diethoxyphosphoryl group (1250 and 1277 cm⁻¹) and of P–O–H of phosphonic acid (916 cm⁻¹) are absent. This likely reveals the formation of metal–O–P bonds whose stretching vibrations are overlapped with those of the porphyrin macrocycle, giving a rather strong band at 1003 cm⁻¹. In the spectrum of Mn(TMPIC)/TiO₂-1, stretching vibrations of the carboxylic acid group (1719 cm⁻¹ for Mn(TMPIC)) were not observed, pointing out the involvement of this anchoring group in the interaction with the titania support.

The morphology of Mn(TMPIP)/TiO₂ was studied by SEM and compared to that of the bare TiO₂ (Figure 5 and Figure S8). The latter is composed of strongly aggregated nanoparticles of similar shape and narrow grain-size distribution. The incorporation of Mn(TMPIP) does not have any influence on the cauliflower-like morphology of the solid. The mesoporous nanospheroids, with diameters ranging from 5 to 20 nm, are irregularly distributed in space and separated by large holes of hundreds of nanometers, thus providing good accessibility to the bound complexes of catalytic sites. This morphology is, therefore, perfectly suited for application of the material in catalysis.

N₂ adsorption-desorption isotherms of all four hybrid materials were recorded at 77 K, and the results of these analyses were summarized in Table S2 (see Supplementary Materials). Upon the derivatization of TiO₂, no change in the shape of the isotherms was observed, whereas a marked decrease in the BET surface area and volume of pores for all samples was noted, which was consistent with the presence of grafted complexes in the mesopores. It’s worth noting that materials Mn(TMPIP)/TiO₂-1 and Mn(TMPIC)/TiO₂-1 with high porphyrin loading display smaller BET surface areas and pore volumes as compared to their counterparts Mn(TMPIP)/TiO₂ and Mn(TMPIP)/TiO₂ charged less in porphyrin complexes. These data agree well with the hypothesis that both complexes are located not only at the outside titania surface but also in its mesopores.

Powder X-ray diffraction measurements confirmed that the pristine mesoporous titania and all samples of functionalized materials prepared in this work are amorphous solids.

The materials with high catalysts loading Mn(TMPIP)/TiO₂-1 and Mn(TMPIC)/TiO₂-1 were chosen for comparative studies of their stability in different organic solvents and in aqueous media with variable pH. A suspension of the solid in toluene, chlorinated (CH₂Cl₂, CHCl₃), polar (CH₃CN, DMF), and protic (MeOH, EtOH) solvents were centrifugated after 1 week, and the solutions thus obtained were analyzed visually or using a laboratory spectrophotometer. Only Mn(TMPIP)/TiO₂-1 with the phosphonate anchor was stable in all these solvents, which is commonly used in organic synthesis. Mn(TMPIC)/TiO₂-1 was rapidly washed off in protic solvents (>50%) and also in acetonitrile (~10%).

The stability of Mn(TMPIP)/TiO₂-1 was also explored in aqueous media in all pH ranges. The leaching of the complex was observed only in basic solutions (pH > 9). We also found that Mn(TMPIP)/TiO₂-1 was stable in the presence of triethylamine in CHCl₃ and DMF. This is important for organic synthesis because tertiary amines are common reagents in catalytic and photocatalytic transformations.

The more stable heterogenized complex with spatially separated catalytic sites (Mn(TMPIP)/TiO₂) was next studied as a catalyst in oxidation reactions.

2.3. Catalytic Reactions

To get insight into the catalytic properties of Mn(TMPIP)/TiO₂, we studied its catalytic efficiency and reusability in the oxidation reactions using molecular oxygen as a terminal oxidant and isobutyraldehyde (IBA) as a sacrificial reagent (Mukaiyama conditions [80]) because these mild oxidation conditions are suitable for transformations of different compounds such as alkanes, alkenes, aldehydes, cyclic ketones, sulfides and alcohols [81,82,83,84,85]. Metalloporphyrins are known as active catalysts for these reactions under homogeneous conditions [12,16,81,83,86,87]. In our work, the oxidation of sulfide to sulfoxide was chosen as a model reaction. Selective sulfoxidation attracts considerable interest due to its relevance to biochemistry, warfare agent disposal, the environmental consequences of fuel desulfurization, and organic synthesis [88,89,90,91,92,93,94,95]. The overoxidation of sulfides to sulfones and the cleavage of S–C and (S)C–H bonds are commonly observed as side reactions [96,97,98], and thus selective methods for transformation of different sulfides are in high demand, in particular in the asymmetric synthesis of ligands and in the development of pharmaceuticals involving selective late-stage oxygenation [93,95,99,100,101].

In the preliminary experiments, the oxidation of thioanisole with molecular oxygen (1 atm) was performed in the presence of 0.09 mol% of TiO₂-supported complex Mn(TMPIP) and 6 equivalents of IBA in toluene at 80 °C and monitored by GC-MS (

Table 1. Oxidation of thioanisole by molecular oxygen in the presence of Mn(TMPIP)/TiO₂ and IBA ¹.

Entry	IBA (Equiv.)	Temperature (°C)	Time (h)	Conversion 2 (%)	Selectivity 2 (%)
					Sulfoxide	Sulfone
1	6	80	0.5	100	58	42
2	6	40	2	100	63	37
3	6	30	2.5	92	99	1
			3	100	53	47
4	6	20	2	85	100	0
			2.5	100	82	12
5	3	80	1.5	76	96	4
			3.5	86	86	14
6	3	50	1.5	52	100	0
			8	94	95	5
7	5	20	3	100	98	2
8 3	5	20	3	0
9	0	20	3	0
10 4	5	20	3	0
11 5	5	20	3	0
12 6	5	20	3	0

¹ Reaction conditions: 2.5 mmol of thioanisole, molecular oxygen (1 atm), IBA, 0.09 mol% of heterogenized Mn(III) porphyrin (based on the formula of Mn(TMPIP)/TiO_2, which was calculated from the data of elemental analysis and ICP-OES (Table S1)), toluene (10 mL) were stirred at a given temperature. ² Conversion and selectivity were determined by GC-MS analysis of reaction mixtures. Naphthalene was used as an internal standard. ³ The reaction was performed without a transition metal catalyst. ⁴ The experiment was performed in Ar (1 atm). ⁵ The catalyst was replaced by bare TiO₂. ⁶ The reaction was carried out in the presence of 0.009 mol% of the heterogenized complex.

). Starting sulfide was consumed in less than 30 min, but the overoxidation of sulfoxide to methyl phenyl sulfone (42%) was observed as a side reaction (entry 1). To diminish the overoxidation, the temperature was decreased to 40 °C (entry 2). Under these conditions, the reaction proceeded more slowly, but sulfone was obtained in a similar yield (37%). Comparing the amount of sulfone in the reaction mixture before and after complete consumption of sulfide in the reaction performed at 30 °C (entry 3), we concluded that the overoxidation was slow until the starting sulfide was present in the reaction mixture but was significantly accelerated after the sulfide consumption. The results obtained at 20 °C confirmed this hypothesis, but the control of the overoxidation was difficult, even under these mild conditions, because the reaction was too rapid (entry 4). Thus, the amount of IBA was decreased up to three equivalents. The rate of oxidation was decreased, and methyl phenyl sulfoxide was the main product (96%) when 76% of starting sulfide was consumed (entry 5). However, when this reaction was prolonged for additional 2 h, only 10% of starting sulfide was reacted, probably due to the complete consumption of IBA. Oxidation at 50 °C also gave only sulfoxide when 52% of sulfide was reacted (entry 6), but 6% of starting compound was still observed after 8 h of reacting. Finally, adjusting the amount of IBA to reach a complete conversion and an appropriate reaction rate (5 equivalents), the sulfoxide was obtained in about quantitative yield (entry 7).

It is noteworthy that the oxidation did not proceed at all in the absence of Mn(TMPIP)/TiO₂, IBA, or oxygen under these conditions (entries 8–10). Our attempts to use non-functionalized TiO₂ as a catalyst (entry 11) or to decrease catalyst loading 10 times (entry 12) failed.

Thus, as low as 0.09 mol% of TiO₂-supported Mn(TMPIP) were required for the selective sulfoxidation of thioanisole with molecular oxygen in the presence of five equivalents of IBA (TON = ~1100, TOF = ~370 h⁻¹) in a single catalytic run of the sulfoxidation of thioanisole. The overoxidation of sulfide to sulfone in the presence of this catalyst proceeds even at room temperature, but sulfoxides can be obtained in a quantitative yield controlling the amount of IBA, temperature, and the time of reacting.

To determine whether the catalyst leaching was observed under reaction conditions and show that the reaction was not catalyzed by the non-supported manganese complex in solution, a “hot filtration test” was performed (see Supplementary Materials, Figure S9). Half of the solution was filtered when 50% of the starting sulfide was consumed, and both reactions were continued under identical conditions. In the filtered solution, oxidation was also observed but was much slower than in the suspension. To prove the absence of catalyst leaching, we determined the content of manganese-containing compounds in the residue obtained after evaporation of the homogeneous reaction mixture using the ICP-OES technique. The manganese content was found to be less than 5 ppm. Based on these data, we assumed that the oxidation reaction observed in the filtrated solution proceeds without the participation of transition metal catalysts. This can be the sluggish autoxidation of sulfides in the presence of O₂ and IBA, as reported previously [102,103].

Encouraged by these results, we investigated the scope of substrates for this reaction. To our surprise, various sulfides can be selectively oxidized to sulfoxides under ambient conditions controlling the amount of IBA introduced into the reaction (

Table 2. Oxidation of sulfides by molecular oxygen in the presence of Mn(TMPIP)/TiO₂ and IBA ¹.

Entry	Sulfide	IBA (Equiv.)	Time (h)	Conversion 2 (%)	Selectivity 2 (%)
					Sulfoxide	Sulfone
1		5	2.5	100	98	2
2		5	2	100	98	2
3		5	7	100	97	3
4		3.3	6	100	92	8
5		6	4	94	98	2
6		9	2.5	92	92	8
7		3	3	99	94	6
8		10	4	98	97	3
9		7	4	98	89	11

¹ Reaction conditions: 2.5 mmol of tioanisole, IBA, 0.09 mol% of heterogenized complex (calculations were based on the data of ICP analysis of Mn(TMPIP)/TiO₂), toluene (10 mL) were stirred at a given temperature under molecular oxygen (1 atm). ² Conversion and selectivity were determined by GC-MS analysis of reaction mixtures. Naphthalene was used as an internal standard.

). The rate of oxidation of aryl methyl sulfides is dependent on the nature of substituents on the phenyl ring. Electron-rich derivatives are more reactive than compounds with electron-deficient aryl groups. For instance, 4-methoxythionisole was converted to sulfoxide in 2 h, but the reaction time should be prolonged to 7 h if 4-nitrothioanisole was oxidized. Nevertheless, all target products were obtained in almost quantitative yield. The oxidation of more sterically hindered diphenyl sulfide, which is known to be resistant to oxidation by O₂ under photocatalytic conditions [104], can be performed with excellent selectivity using only 3.3 equivalents of IBA and 6 h of reacting under otherwise identical conditions to give the target product in a 92% yield.

Linear and cyclic dialkyl sulfides were found to be less reactive than aryl sulfides and were oxidized in the presence of 6–9 equivalents of IBA. The oxidation of cyclic derivative seems to be less selective because sulfoxide and sulfone were obtained in a 92:8 ratio when only 92% of the sulfide was consumed.

Finally, we explored the sulfoxidation of compounds bearing additional oxidizable sites such as vinyl, hydroxyl and cyano substituents (entries 7–9). To our great satisfaction, all target sulfoxides were obtained in a high yield, and the overoxidation of sulfoxide was only observed as a side reaction. Despite the reactions being optimized only briefly, even in the case of the most problematic phenyl vinyl sulfide, the sulfoxide:sulfone ratio was 89:11 when 90% of starting compound was consumed.

The mechanistic pathways of oxidation by the O₂-IBA reagent system were debated in the literature [12,105,106,107,108,109]. Several active species with and without transition metal atoms were identified. The relative efficacy by which a sulfide molecule is oxidized by one versus the other pathways appears to be dependent on the nature of the transition metal catalyst, solvent and temperature. Despite mechanistic studies of porphyrin catalysts being very limited, high-valent metalloporphyrin–oxo intermediates were considered as main active oxygen-containing intermediates [108,110,111,112]. Our data do not contradict this hypothesis because the oxidation is more rapid in the presence of the Mn(III) complex, as was shown in the “hot filtration test.” However, other less-active oxygen-containing intermediate species can also contribute to the formation of the target product. Detailed mechanistic studies of this reaction are beyond the scope of this work.

Catalyst recovery and refining were straightforward. After completion of the reaction, the catalyst was isolated by centrifugation, washed with toluene and introduced in the next cycle. Seven consecutive oxidation reactions of thioanisole were carried out, giving the target sulfoxides in comparable yields (94–98%; Figure 6 and Table S3 (see Supplementary Materials)). The reaction time was different in these experiments, but the changes were non-systematic, probably because this reaction parameter is highly dependent on the efficiency of stirring heterogeneous reaction mixtures. The content of resting manganese derivatives in the crude products was found to be less than 5 ppm by using the ICP-OES analysis (Table S3 (see Supplementary Materials)). This proves that catalyst leaching in the reaction mixture was negligible. The integrity of the Mn(TMPIP)/TiO₂ catalyst was also confirmed by studying the solid recovered after the last catalytic cycle by diffuse reflectance spectroscopy. As shown in Figure S10 (see Supplementary Materials), the solid exhibits absorption features characteristic of Mn(TMPIP) in CHCl₃ solution and Mn(TMPIP)/TiO₂ material. A small bathochromic shift of spectral band positions was observed for several bands in the spectrum of the recovered solid. This likely indicates a difference in the adsorbed solvent molecules in the starting and reused catalysts or the chemical modification of the TiO₂ surface (the esterification by isobutyric acid, for instance) during the oxidation reaction. Nevertheless, based on these spectral data, we can safely conclude that the heterogenized porphyrin macrocycle is not decomposed after seven consecutive oxidation cycles.

When oxidation of thioanisole was performed under homogeneous conditions replacing Mn(TMPIP)/TiO₂ by its soluble counterpart Mn(TMPIP) or commercially available (meso-tetramesitylporphyrinato)manganese(III) chloride (Mn(TMP)) only starting thioanisole was observed in the reaction mixture after 24 h of reacting. When Mn(TMPIP) loading was increased up to 0.5 mol%, only 15% of thioalisole was converted to sulfoxide after 24 h. Thus, immobilization of Mn(TMPIP) on hydrated titania significantly increases its catalytic efficiency allowing us to perform sulfoxidation under ambient conditions. The reason for the synergetic effect is unclear at this point of investigation. We cannot exclude that the adsorption of sulfides on a large surface of inorganic support accelerates the oxidation reaction, but most probably, peroxide intermediates formed on the surface of hydrated TiO₂ [113] also participated in the oxidation.

Comparing Mn(TMPIP)/TiO₂ with porphyrin catalysts reported previously, we can conclude that this material is situated among the best catalysts known for the sulfoxidation reaction. Indeed, the oxidation of sulfides by sodium periodate in the presence of Mn(III) porphyrins grafted on polystyrene [114] or silica [115] led to a mixture of sulfoxides and sulfones. When Mn(III) porphyrins were grafted on graphene oxide and used as catalysts for the sulfoxidation by hydrogen peroxide-urea adduct, about 2 mol% of catalysts were required, and the reactions were less selective, in particular when aromatic sulfides were used as substrates [116]. Fe(III) porphyrins grafted on activated multi-walled carbon nanotubes were efficient in ethanol when tetra-n-butylammonium peroxomonosulfate was used as an oxidant, but some of the studied reactions were difficult to control [117]. Several heterogenized porphyrin catalysts were also reported for the oxidation of sulfides to sulfoxides under Mukaiyama conditions. Manganese(III) complex with 5,10,15,20-tetrakis(N-methylpyridinium-4-yl)porphyrin immobilized in montmorillonite gave the selective transformation only when the reaction was conducted at 50–80 °C [34]. Only COP based on Fe(III) porphyrin has shown catalytic activity similar to that of Mn(TMPIP)/TiO₂ [45]. However, this immobilization procedure requires a large amount of the complex and yields hybrid materials in which the number of catalytic sites cannot be optimized.

It’s also worth noting that porphyrin catalysts bearing phosphonate anchors were successfully immobilized onto the zinc phosphonate network with low porosity [118] and incorporated onto zirconium phosphonate film materials [119]. Our immobilization strategy is complementary to these approaches and gave porous porphyrin catalysts.

3. Materials and Methods

3.1. General

Unless otherwise noted, all chemicals and starting materials were obtained commercially from Acros (via Thermo Fisher Scientific, Illkirch, France) or Aldrich-Sigma Co. (via Merck Co., Darmstadt, Germany) and used without further purification. (5,10,15,20-Tetramesitylporphyrinato)manganese(III) chloride (Mn(TMP)) was provided by PorphyChem (Dijon, France). 2-Nitro-5,10,15,20-tetramesitylporphyrin was prepared according to the literature protocol [120]. 5,10,15,20-Tetramesityl-2-(4-carboxyphenyl)-1H-imidazo[4,5-b]porphyrin (2H(TMPIC)) and 5,10,15,20-tetramesityl-2-[4-(diethoxyphosphoryl)phenyl]-1H-imidazo[4,5-b]porphyrin (2H(TMPIP)) were synthesized according to a procedure developed by us previously [58,121]. Mesoporous hydrated titanium(IV) oxide was prepared using a surfactant-free sol-gel procedure reported by us previously [76]. The content of Ti in the solid thus obtained was determined by ICP-OES analysis. Preparative column chromatography was carried out using silica gel 60 (40–63 μm) from Merck Co. (Darmstadt, Germany). CH₂Cl₂ was distilled over CaH₂.

UV-vis absorption spectra were collected with an Agilent Cary 60 (Massy, France) using Suprasil 300 cuvettes (Hellma, l = 1 cm). Diffuse reflectance spectra of materials were acquired on a Cary 5000 (Massy, France) UV-vis-NIR spectrophotometer outfitted with a praying Mantis^TM accessory (Harrick, via Agilent, Massy, France). The baseline was recorded with a Spectralon^® pellet. Corrected reflectance data (R) were converted to f(R) values using the Kubelka–Munk function expressed as f(R) = (1 − R²)/2R.

FTIR spectra were registered either on a Nicolet iS 5 (Thermo Ficher Scientific, Illkirch, France) or a Bruker Vector 22 (Wissembourg, France) spectrophotometer. A micro-ATR accessory (Pike) was used in order to obtain FTIR spectra of solid polycrystalline complexes.

MALDI-TOF mass-spectra were obtained on a Bruker Ultraflex II LRF 2000 (Wissembourg, France) mass-spectrometer in positive ion mode with the dithranol matrix. Accurate mass measurements (HRMS-ESI) were performed with a Thermo Scientific Orbitrap Elite high-field Orbitrap hybrid mass spectrometer (Thermo Ficher Scientific, Illkirch, France). The samples were dissolved in a methanol/chloroform (1:1, v/v) solvent mixture and analyzed in positive mode. Microanalyses (CHN) were performed using a Thermo Electron Flash EA 1112 analyzer (Thermo Ficher Scientific, Illkirch, France). Mn, P and Ti contents were measured by inductively coupled plasma optical emission spectrometry (ICP-OES DUO) on an ICAP 7400 instrument from Thermo Ficher Scientific (Illkirch, France). The samples were mineralized using standard procedures.

Field-emission scanning electron microscopy (FESEM) was performed using a JEOL JSM 7600F (JEOL (Europe) SAS, Croissy-sur-Seine, France) instrument located in the ARCEN analysis center of the University of Bourgogne (Dijon, France). Images were acquired using GentleBeam-High SEM mode. Transmission electron microscopy (TEM) analyses were conducted using a JEOL JEM-2100F microscope ((JEOL (Europe) SAS, Croissy-sur-Seine, France) operating at 200 kV and located in the ARCEN analysis center of the University of Bourgogne (Dijon, France). EDX spectroscopy in STEM mode was used for chemical mapping and qualitative and quantitative elemental analysis using a Bruker XFlash Detector 5030 spectrometer fitted on the JEM-2100F microscope (Bruker, Wissembourg, France).

Nitrogen adsorption-desorption isotherms were measured with a BELSORP max analyzer (BEL Japan, INC, via Mercer Instruments, Passy, France) at 77 K with samples degassed for 6 h at 80 °C under reduced pressure (10⁻⁵ torr). Specific surface areas (S_BET) were calculated according to the Brunauer–Emmett–Teller (BET) method.

Powder X-ray diffraction experiments were performed with an Empyrean diffractometer from the PANalytical company (Palaiseau, France) in the range 3° < 2θ < 50°, using a copper anticathode X-ray tube (Cu Kα₁ = 1.54060 Å and Cu Kα₂ = 1.54443 Å) and an X’Celerator detector outfitted with an anti-scattering slit of 5 mm. The uncrushed samples were placed between 2 Mylar sheets, and the analysis was carried out in transmission mode using a focusing X-ray mirror equipped with fixed divergent and anti-scattering slits (aperture 0.5°) and 0.02 rad Soller slits.

All elemental analyses except EDX analyses were performed at the “Pôle Chimie Moléculaire,” the technological platform for chemical analysis and molecular synthesis (http://www.wpcm.fr, accessed on 30 December 2022) of the Institut de Chimie Moléculaire de l’Université de Bourgogne and Welience^TM, a private subsidiary of the Université de Bourgogne.

3.2. Synthesis of Manganese(III) Complexes Mn(TMPIC), Mn(TMPIP) and Mn(TMPIP-OH)

General procedure. To a solution of free base porphyrin (0.1 mmol) in dimethylformamide (13 mL), manganese dichloride (63 mg, 0.5 mmol, 5 equivalents) was added at 150 °C. The reaction mixture was stirred at this temperature and monitored by TLC. After complete consumption of starting porphyrin (4 h), the reaction mixture was cooled to room temperature and evaporated under reduced pressure. The residue thus obtained was dissolved in dichloromethane (30 mL), and the solution was washed with water (3 × 20 mL), dried over MgSO₄ and evaporated under reduced pressure. The residue thus obtained was chromatographed on silica gel using CH₂Cl₂/MeOH as eluent.

[5,10,15,20-tetramesityl-2-(4-carboxyphenyl)-1H-imidazo[4,5-b]porphyrinato]mangane-se(III) chloride (Mn(TMPIC)). The complex was prepared according to the general procedure from 2H(TMPIC) (120 mg, 0.127 mmol) and isolated as a green solid using CH₂Cl₂/MeOH (95:5 v/v) as an eluent. Yield 92% (116 mg); mp > 320 °C, decomp. UV-vis (CHCl₃, λ_max, nm (log(ε (cm⁻¹M⁻¹))): 333 (4.60), 382 (4.80), 403 (4.75), 481 (5.00), 540 (sh), 573 (4.09), 622 (4.04), and 748 (3.04). FTIR (neat, ν, cm⁻¹): 3410 (w, NH), 2913 (m), 2849 (m), 1719 (m, C=O), 1610 (m), 1437 (w), 1378 (w), 1331 (w), 1250 (w), 1225 (w), 1195 (m), 1003 (s), 862 (w), 851 (w), 833 (m), 802 (m), and 721 (m). HRMS-ESI: m/z [M-Cl]⁺ calcd for C₆₄H₅₆MnN₆O₂: 995.38398; found: 995.38814.

[5,10,15,20-tetramesityl-2-[4-(diethoxyphosphoryl)phenyl]-1H-imidazo[4,5-b]porphyrin-ato]-manganese(III) chloride (Mn(TMPIP)). The complex was prepared according to the general procedure from 2H(TMPIP) (554 mg, 0.535 mmol) and isolated as a green solid using CH₂Cl₂/MeOH (95:5 v/v) as eluent. Yield 60% (330 mg); mp > 320 °C, decomp. UV-vis (CHCl₃, λ_max, nm (log(ε (cm⁻¹M⁻¹))): 327 (4.46), 383 (4.72), 407 (4.65), 484 (4.92), 576 (3.93), 626 (3.89), and 749 (2.92). FTIR (neat, ν, cm⁻¹): 3385 (w, NH), 2916 (w), 2854 (w), 1609 (w), 1475 (w), 1437 (w), 1410 (w), 1377 (w), 1333 (w), 1248 (m, P=O), 1227 (m, P=O), 1196 (m), 1130 (m), 1051 (m), 1016 (s, P–O–C), 1003 (s), 953 (m), 851 (w), 833 (m), 800 (s), 762 (w), 719 (m), and 665 (w). HRMS-ESI: m/z [M-Cl+H]²⁺ calcd for C₆₇H₆₆MnN₆O₃P: 544.21518; found: 544.21526; [M-Cl]⁺ calcd for C₆₇H₆₅MnN₆O₃P: 1087.42308; found: 1087.42690; [M+H]⁺ calcd for C₆₇H₆₆ClMnN₆O₃P: 1123.39976; found: 1123.39579.

[5,10,15,20-tetramesityl-2-[4-(dihydroxyphosphoryl)phenyl]-1H-imidazo[4,5-b]porphyrin-ato]manganese(III) bromide (Mn(TMPIP-OH)).

A dry Schlenk tube was charged with diethyl phosphonate ester Mn(TMPIP) (15 mg, 0.0134 mmol) and dry CH₂Cl₂ (5 mL) under argon. Then, TMSBr (53 mL, 0,4 mmol, 30 equivalents) was added via a syringe, and the resulting mixture was stirred for 3 d at r. t. and monitored by MALDI-TOF mass spectrometry. After the complete consumption of ethyl phosphoesters, methanol (2 mL) was added to the reaction mixture, and stirring was continued for 30 min. Evaporation of volatiles under reduced pressure afforded the target phosphonic acid as a dark green solid. Yield: 90% (14 mg); mp > 350 °C, decomp. UV-vis (CHCl₃, λ_max, nm (log(ε (cm⁻¹M⁻¹))): 303 (4.40), 385 (4.78), 419 (4.73), 452 (4.53), 491 (4.81), 587 (3.90), 633 (3.92), and 760 (2.86). FTIR (neat, ν, cm⁻¹): 2918 (m), 2853 (m), 1609 (w), 1437 (m), 1408 (w), 1379 (w), 1335 (w), 1196 (m), 1134 (w), 1003 (s), 916 (m, P–OH), 851 (w), 831 (s), 802 (s), 719 (s), and 663 (w). HRMS-ESI: m/z [M-Br]⁺ calcd for C63H57MnN6O3P: 1031.36048; found: 1031.36328; [M-HBr+Na]⁺ calcd for C63H56MnN6NaO3P: 1053.34242; found: 1053.34345. HRMS-ESI spectrum is shown in Figure S4.

3.3. Synthesis of Heterogenized Catalysts

Grafting complex Mn(TMPIC). A dry Schlenk tube was charged with carboxylic acid Mn(TMPIC) (18 mg, 0.018 mmol) and dry CH₂Cl₂ (50 mL) under argon. Then mesoporous hydrated TiO₂ (18 mmol, 100 equivalents) was added under an argon stream. The suspension was stirred for 2 d at r.t. The solid was collected by centrifugation and thoroughly washed with THF (2 × 20 mL), water (15 mL), MeOH (3 × 15 mL), and ether (2 × 15 mL). The material was finally dried for 24 h at 80 °C under reduced pressure (2 mmHg). Yield: 165 mg.

The solid Mn(TMPIC)/TiO₂-1 with a high content of Mn(TMPIC) (the Mn(TMPIC):TiO₂ ratio was 1:15) was also prepared according to this procedure.

Grafting complex Mn(TMPIP). A dry Schlenk tube was charged with diethyl phosphonate ester Mn(TMPIP) (360 mg, 0.32 mmol) and dry CH₂Cl₂ (50 mL) under argon. Then, TMSBr (30 equiv equivalents) was added via a syringe, and the resulting mixture was stirred for 48 h at r. t. The reaction was monitored by MALDI-TOF mass spectrometry to ensure a complete conversion had occurred before the evaporation of volatiles under reduced pressure. Then, 50 mL of dry CH₂Cl₂ was introduced into the Schlenk tube with a syringe, followed by the addition of mesoporous hydrated TiO₂ (32 mmol, 100 equivalents) under an argon stream. The suspension was stirred for 2 d at r. t. The solid was collected by centrifugation and thoroughly washed with THF (2 × 20 mL), water (15 mL), MeOH (3 × 15 mL), and ether (2 × 15 mL). The material was dried for 24 h at 80 °C under reduced pressure (2 mmHg). Yield: 3050 mg.

The solid Mn(TMPIP)/TiO₂-1 with a high content of Mn(TMPIP) (ratio of Mn(TMPIP):TiO₂ was 1:15) was prepared according to this procedure.

3.4. Catalytic Reactions

General procedure. A 25 mL two-necked flask equipped with a magnetic stir bar and a back-flow condenser was charged with 0.09 mol% of heterogenized Mn(TMPIP) (calculations were based on the data of ICP-OES analysis of Mn(TMPIP)/TiO₂, Table S1). The reaction vessel was evacuated and purged by oxygen 3 times. Subsequently, 2.5 mmol of sulfide, the calculated amount of IBA (Table 1 and Table 2), and 10 mL of toluene were added by syringe, and the reaction mixture was stirred under oxygen (filled balloon), controlling the temperature (see Table 1 and Table 2). The reaction was periodically monitored by GC-MS, and quantitative analyses were performed using naphthalene as an internal standard. When the reaction was complete, the mixture was centrifugated, and the filtrate was washed with water (3 × 5 mL), dried over MgSO₄ and evaporated under reduced pressure at room temperature. The structure of the target product was confirmed by ¹H NMR, comparing its spectral data with those reported previously in the literature.

Turnover number (TON = Number of product molecules/Number of active catalytic sites) and turnover frequency (TON in an hour) were estimated using a calculated value of catalyst loading (0.09 mol%).

In the experiment with bare TiO₂ (Table 1, entry 11), the amount of hydrated titania was identical to that which was used in the experiments with Mn(TMPIP)/TiO₂. The calculations were performed using the empirical formula obtained from the data of elemental analysis and ICP-OES.

Methyl phenyl sulfoxide (Table 2, entry 1) [122]. ¹H NMR (CDCl₃, 400 MHz): 2.72 (s, 3H, CH₃), 7.50–7.58 (m, 3H, Ar), 7.63–7.72 (m, 2H, Ar).

4-Methoxyphenyl methyl sulfoxide (Table 2, entry 2) [122]. ¹H NMR (CDCl₃, 300 MHz): 2.70 (s, 3H, CH₃), 3.82 (s, 3H, OCH₃), 7.04−7.10 (m, 2H, Ar), 7.59−7.68 (m, 2H, Ar).

Methyl 4-nitrophenyl sulfoxide (Table 2, entry 3) [123]. ¹H NMR (CDCl₃, 300 MHz): 2.81 (s, 3H, CH₃), 7.79–7.88 (m, 2H, Ar), 8.37–8.44 (m, 2H, Ar).

Diphenyl sulfoxide (Table 2, entry 4) [124]. ¹H NMR (CDCl₃, 300 MHz): 7.29–7.43 (m, 6H, Ar), 7.56 (m, 4H, Ar).

Dibutyl sulfoxide (Table 2, entry 5) [122]. 0.96 (t, J = 7.3 Hz, 3H, CH₃), 1.36–1.61 (m, 2H, CH₂), 1.65–1.83 (m, 2H, CH₂), 2.60–2.78 (m, 2H, CH₂).

Thian-4-one S-oxide (Table 2, entry 6) [125]. ¹H NMR (CDCl₃, 300 MHz): 2.43–2.58 (m, 2H, CH₂), 2.78–2.95 (m, 2H, CH₂), 3.19–3.39 (m, 4H, 2CH₂).

(Phenylsulfinyl)acetonitrile (Table 2, entry 7) [126]. ¹H NMR (CDCl₃, 300 MHz): 3.72 (s, 2H, CH₂), 7.49–7.75 (m, 5H, CH).

(2-Phenylsulfinyl)ethanol (Table 2, entry 8) [127]. ¹H NMR (CDCl₃, 300 MHz): 3.07–3.13 (t, 2H, CH₂S), 3.72–3.77 (t, 2H, CH₂O), 7.19–7.28 (m, 3H, p-, m-CH), 7.35 (d, J = 7 Hz, 2H, o-CH).

Phenyl vinyl sulfide (Table 2, entry 9) [128]. ¹H NMR (CDCl₃, 300 MHz): 5.88 (d, J = 9.7 Hz, 1H, CH₂), 6.18 (d, J = 16.4 Hz, 1H, CH₂), 6.65 (dd, J = 16.4, 9.7 Hz, 1H, CH), 7.44–7.77 (m, 5H, Ar).

Hot filtration test. A 15 mL 2-necked flask equipped with a magnetic stir bar was charged with 0.09 mol% of heterogenized Mn(TMPIP) (calculations were based on the data of ICP-OES analysis of Mn(TMPIP)/TiO₂). The reaction vessel was evacuated and purged with oxygen 3 times. Subsequently, 117 μL (1 mmol) of thioanisole, 456 μL (5 mmol) of IBA, and 4 mL of toluene were added by syringe, and the reaction mixture was stirred at room temperature under oxygen (filled balloon) and periodically monitored by GC-MS. When half of the sulfide was consumed, half of the solution (2.2 mL) was withdrawn and filtered using a membrane filter (0.22 μm). The second 15 mL 2-necked flask equipped with a magnetic stir bar was evacuated and purged by oxygen 3 times. This reaction vessel was charged with the filtered solution obtained as described above and an oxygen-filled balloon. Both reaction mixtures were stirred at room temperature and monitored by GS-MS. The results of GS-MS analyses are shown in Figure S9.

Catalyst recycling. A 25 mL 2-necked flask equipped with a magnetic stir bar was charged with 0.09 mol% of heterogenized Mn(TMPIP) (calculations were based on the data of ICP-OES analysis of Mn(TMPIP)/TiO₂). The reaction vessel was evacuated and purged by oxygen 3 times. Subsequently, 293 μL (1 mmol) of thioanisole, 1.139 mL (5 mmol) of IBA, and 10 mL of toluene were added via a syringe, and the reaction mixture was stirred at room temperature under oxygen (filled balloon) and monitored by GC-MS. After complete consumption of thioanisole, the reaction mixture was centrifugated, and the catalyst was washed with toluene (3 × 10 mL) and dried for 3 h at r.t. The solid thus obtained was used as a catalyst in the next consecutive oxidation cycle. The filtrate was analyzed using GC-MS with naphthalene as an internal standard, evaporated under reduced pressure and analyzed by ICP-OES to determine the content of Mn. The results of these experiments are summarized in Table S3 and shown in Figure 6.

3.5. Computational Details

The quantum calculations were carried out with Spartan’10 software (build 1.1.0, Wavefunction Inc., Tokio, Japan) running on a desktop computer with the Windows operating system. The structures were found by convergence to equilibrium geometry at an energy minimum with default values of gradient tolerance (4.5 × 10⁻⁴ Hartree.Bohr⁻¹) and distance tolerance (1.8 × 10⁻³Å). The DFT calculations were performed at the B3LYP level of theory with a 6-31G* basis set. The starting structures for calculation were prepared by minimization with the semi-empirical PM6 method. The vibrational frequencies were calculated to prove the found structures to be minimal and showed no imaginary frequencies.

4. Conclusions

meso-Tetraarylporphyrins were immobilized on hydrated mesoporous titanium dioxide (S_BET = 705 m² g⁻¹) using carboxylate or phosphonate anchoring groups attached to 2-arylimidazole linker fused across one of the pyrrolic rings of the macrocycle. Two types of hybrid materials with different loading of metalloporphyrins were prepared and characterized as both a bulk solid and at the molecular level using different physicochemical methods, including elemental analysis, UV-vis diffusion and infrared spectroscopies, SEM microphotography and nitrogen sorption isotherms. The morphology of these hybrid materials is similar to that of the pristine TiO₂ and is perfectly suited for application of these materials in catalysis. The mesoporous nanospheroids, with a diameter ranging from 5 to 20 nm, are irregularly distributed in space and separated by large holes of hundreds of nanometers, thus providing good accessibility to catalytic sites. The homogeneous distribution of catalytic sites inside mesopores was proven by EDX analysis in STEM mode and studies of N₂ adsorption isotherms. Quantitative immobilization of porphyrin catalysts can be achieved using both carboxylate and phosphonate anchors, but only the latter one gives hybrid materials which are stable in polar and protic solvents commonly used in organic synthesis.

Finally, catalytic properties of the more stable material Mn(TMPIP)/TiO₂ with the phosphonate anchor were evaluated in the selective oxidation of sulfides to sulfoxides by the molecular oxygen/IBA system. The heterogenized complex has shown excellent catalytic activity exhibiting a turnover (TON) of ~1100 in a single catalytic run of the sulfoxidation of thioanisole. TON can be increased by recycling the catalysts at least seven times, owing to the exceptional stability of the solid. This catalyst allows for mild reaction conditions, a wide range of substrates, and excellent product yields in the reactions performed in toluene which is tolerated in industry.

Thus, we have developed a novel synthetic approach to the heterogenization of metalloporphyrins. This strategy has many potential benefits for sustainable chemistry technologies. Structural parameters of grafted metalloporphyrins can be fine-tuned by changing the nature of the central metal ion and four aryl substituents attached to the meso-positions of the porphyrin core, which is important for the optimization of their catalytic activity. The heterogenized catalysts thus prepared also benefit from the rational design of the spacer, which is a long and rigid residue and provides a perpendicular orientation of the mean porphyrin plane with respect to a solid surface which is highly desirable for the easy accessibility of catalytic sites. The mesoporous titania is a cost-efficient, thermally and chemically stable support that could decrease production costs. Last but not least, the quantity of the catalyst grafted onto the TiO₂ surface can be easily controlled and adapted to the target catalytic reaction allowing for the economic use of expensive porphyrin complexes. We believe that this immobilization strategy will be useful in catalysis because the scope of reactions catalyzed by metalloporphyrins is constantly increased.