ε-Caprolactone polymerization using titanium complexes immobilized onto silica based materials functionalized with ionic liquids: insights into steric, electronic and support effects

Abstract

Heterogeneous catalysts comprised of titanium immobilized on silica imidazolium-based systems have been synthesized for the ring opening polymerization of ε-caprolactone. Titanium isopropoxide as a metallic precursor was grafted onto ionic liquid-functionalized silica surfaces or alternatively first grafted onto silica followed by reaction with 1-methyl-3-[(triethoxysilyl)propyl]imidazolium chloride. These materials have been characterized by X-ray diffraction (XRD), X-ray fluorescence (XRF), N₂ adsorption–desorption, transmission electron microscopy (TEM) with microanalysis by energy-dispersive X-ray (EDX), ²⁹Si, ¹³C and ^49/47Ti solid-state NMR spectroscopy (MAS NMR), Fourier-transform infrared spectroscopy (FT-IR), UV-vis diffuse reflectance spectroscopy (DRUV-vis), thermogravimetric analysis (TG-DTG) and solid state electrochemical techniques. Different ionic liquid compounds have been tested with different silica supports (mesoporous silica nanoparticles, SBA-15 and high surface area silica). Thus, the effect of the ionic liquid in the chemical environment of the anchored titanium complex has been determined. In addition, the stability and reusability of the catalysts have been studied.

---

1. Introduction

In recent years, the use of mesoporous silica nanoparticles (MSNs) as a catalyst support has raised great interest due to their interesting properties as high surface area, nanospherical morphologies and excellent chemical and thermal stability. Among them, organic–inorganic hybrid MSNs have been used in several synthetic reactions and transformations.¹ Ionic liquids (ILs) have also received a lot of attention due to their unique chemical and physical properties. Despite ILs being expensive, supported ionic liquids have been considered as an alternative to immobilized metal species on solid supports.² In 2010, Doorslaer et al.³ described four types of catalyst-ionic liquid systems. The SILP materials defined as those materials where the metal complex and some degree of the IL are not bounded to the support. The materials named as SILC, those where, both the IL and catalyst are bounded to the support. Finally, the SCIL type system where the IL is not bounded to the support. Thus, supported ionic liquids materials present a combination of homogeneous and heterogeneous aspects of catalysis in one system and significant advantages in biphasic catalysis, such as high activity and the ease the product separation. Moreover, as the amount of published works increases the comparison between homogenous, heterogeneous and biphasic systems has become possible.

Poly(caprolactone) (PCL) a biodegradable and biocompatible polyester,⁴ is mainly produced by ring-opening polymerization (ROP) of the cyclic ester employing homogeneous catalysts.^5,6 To date, there are only a few studies reporting the use of heterogeneous systems in ring-opening polymerization processes. Some have reported the immobilization of homoleptic alkoxo or amide complexes on silica surfaces in the preparation of catalysts active in polymerization of L-lactide and ε-caprolactone.^7–10 Alternatively, catalytic supports such as magnetic nanoparticles¹¹ and aluminium and calcium-incorporated MCM-41-type silica¹² have been used. Recently, our group¹³ has synthesized titanium, zinc, aluminium and magnesium based silica mesoporous materials as heterogeneous polymerization catalysts of ε-caprolactone. Our studies concluded that the metallic precursor plays an important role in the catalytic activity and in the molecular weight and polydispersity of the polymers obtained. The material prepared by grafting titanium isopropoxide onto SBA-15 was the most effective catalyst providing 100% conversion after 24 h and producing PCL with a narrow molecular weight distribution.

Herein, immobilized titanium-containing ionic liquid materials (SILC) have been prepared by using different strategies and supports (mesoporous silica nanoparticles (MSNs), SBA-15 mesoporous silica and SiO₂ high surface area silica). The synthesized materials present immobilized moieties consisting of ionic liquid and titanium covalently grafted to a silica surface. The materials have been characterized by common techniques. In addition, the titanium atom environment has been studied in depth by DRUV-vis, ^49/47Ti MAS NMR spectroscopy and solid state electrochemical techniques. The catalytic activity and reusability of all catalysts has been investigated in the ring opening polymerization of ε-caprolactone by emphasizing the effect of the ionic liquid, the nature of the silica support as well as the titanium loading on the reaction activity.

2. Results and discussion

Heterogeneous catalysts comprised of titanium immobilized into silica imidazolium-based systems have been synthesized by using two different strategies as shown in Scheme 1. Titanium isopropoxide was firstly immobilized onto MSN (x% Ti–MSN with x = 3.7 and 9.45% titanium content) or SBA-15 (Ti–SBA-15) followed by reaction with a slight excess of 1-methyl-3-[(triethoxysilyl)propyl]imidazolium chloride (IMILCl). These materials are denoted as IMILCl–Ti–MSN and IMILCl–Ti–SBA-15 hereafter. Alternatively, titanium isopropoxide was grafted onto ionic liquid-functionalized silica surfaces to give the materials labelled as Ti–IMILCl–MSN and Ti–IMILCl–SBA-15. Furthermore, the materials denoted as Ti–IMILPF₆–MSN and Ti–IMILBF₄–MSN was synthesized by immobilization of 1-methyl-3-[(triethoxysilyl)propyl]imidazolium hexafluorophosphate (IMILPF₆) and 1-methyl-3-[(triethoxysilyl)propyl]imidazolium tetrafluoroborate (IMILBF₄) onto MSN, respectively, followed by the addition of titanium isopropoxide. The resulting white solids were thoroughly washed in order to remove any traces of unreacted titanium precursor.

Scheme 1 Preparation of immobilized titanium-containing ionic liquid materials by grafting.

For comparison purposes, the ionic liquid was incorporated into the MSN or SiO₂ framework by a one-step “in situ” synthesis procedure (Scheme 2), to obtain the hybrid materials denoted as CoIMILCl–MSN and CoIMILCl–SiO₂, respectively. In a second stage, Ti–CoIMILCl–MSN and x% Ti–CoIMILCl–SiO₂ (x = 3.5 and 5.4% titanium content) materials were obtained by reaction of titanium isopropoxide with the ionic liquid functionalized materials and subsequent washing.

Scheme 2 Preparation of immobilized titanium-containing ionic liquid materials by one step co-condensation (CTBA = hexadecyltrimethylammonium bromide, P123 = poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol)).

The mesostructure of titanium containing materials was confirmed by small-angle XRD and N₂ adsorption–desorption isotherms. Unmodified MSN shows a well-resolved pattern at low 2θ values with one strong (100) diffraction peak at 2.49 and two additional high order peaks (110) and (200) with lower intensities, corresponding to a highly ordered hexagonal mesoporous silica (Fig. 1). The titanium containing materials showed the same pattern, indicating that the ordered mesoporous structure of MSN was well retained after the grafting process (Fig. 1 and S1†). The plane (100) in the host material shows higher intensity compared with the immobilized catalyst which can be attributed to the lowering local order, such as variations in the wall thickness of the framework or reduction of scattering contrast between the channel wall and the ligands present on the inner surface of silica materials (Table 1).

Fig. 1 Small-angle XRD patters of (a) MSN (b) 9.45% Ti–MSN (c) Ti–IMILCl–MSN.

Table 1 Textural properties and metal content of titanium supported catalysts

Material	SBET (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (Å)	Ti^a (%)	Ti (mmol g^−1)	Cl^a (%)	IMILCl (mmol g^−1)
a The titanium and chloride content were calculated by X-ray fluorescence (XRF) analysis.
SBA-15	730.66	0.77	62.25	—	—	—	—
IMILCl–Ti–SBA-15	373.61	0.47	50.64	7.43	1.55	1.33	0.37
Ti–IMILCl–SBA-15	363.33	0.47	51.29	2.37	0.49	2.16	0.61
MSN	1064.71	0.96	36.05	—	—	—	—
3.7% Ti–MSN	836.22	0.59	26.17	3.70	0.77	—	—
9.45% Ti–MSN	827.69	0.59	28.66	9.45	1.97	—	—
IMILCl–Ti–MSN	627.61	0.59	28.31	7.97	0.17	1.64	0.46
Ti–IMILCl–MSN	697.12	0.53	30.61	3.14	0.66	2.64	0.74
Ti–IMILPF6–MSN	331.32	0.26	28.93	2.62	0.55	—	—
Ti–IMILBF4–MSN	466.64	0.40	34.04	1.32	0.27	—	—
Ti–CoIMILCl–MSN	601.55	0.65	43.47	5.92	1.24	0.10	0.03
CoIMILCl–SiO2	649.22	2.19	135.5	—	—	3.13	0.88
3.5% Ti–CoIMILCl–SiO2	435.19	1.59	126.4	3.50	0.73	1.29	0.36
5.4% Ti–CoIMILCl–SiO2	440.25	1.58	123.21	5.40	1.13	1.73	0.49

---

The physical parameters of the nitrogen isotherms, as the surface area (S_BET), total pore volume and BJH average pore diameter for parent supports and titanium containing materials are shown in Table 1. The parent MSN material possesses high S_BET (1064 m² g⁻¹) and a BJH pore diameter of 36 Å, these values are typical of MCM-41 type MSN. The titanium materials present a drastic decrease in the S_BET, pore volume and average BJH pore diameter in comparison with the parent support due to the presence of ionic liquid and titanium precursor anchored to the channels partially blocking the adsorption of nitrogen molecules. In the case of Ti–IMILBF₄–MSN and Ti–IMILPF₆–MSN materials, a further decrease in surface area and pore volume was observed, which can be explained by the steric hindrance imposed by the bulkier anions in the ionic liquid. In comparison, the material Ti–CoIMILCl–MSN synthesized by co-condensation method presented higher pore diameter (43.47 Å) and pore volume (0.65 cm³ g⁻¹) than the material Ti–IMILCl–MSN prepared by post-synthesis. The titanium and IMILCl loading were calculated on the basis of experimental X-ray fluorescence (XRF) analysis (Table 1). As expected, higher titanium loadings were obtained, 7.9% for IMILCl–Ti–MSN and 7.4% for IMILCl–Ti–SBA-15 materials, prepared by direct reaction of the titanium precursor with naked silica and subsequent reaction with the ionic liquid, in comparison to those prepared by immobilization of titanium onto previously ionic liquid functionalized silica; with 3.1% and 2.4% titanium loading values for Ti–IMILCl–MSN and Ti–IMILCl–SBA-15, respectively. Consequently, the amount of immobilized ionic liquid increases in the latter materials, that is 0.74 mmol g⁻¹ for Ti–IMILCl–MSN and 0.61 mmol g⁻¹ for Ti–IMILCl–SBA-15.

The ²⁹Si and ¹³C MAS NMR spectra give proof of the reaction of IMILCl and titanium precursor with the available silanol groups on the silica surface. The ²⁹Si MAS NMR spectrum of Ti–IMILCl–MSN (see Fig. S2†) shows the most abundant site Q⁴ ((SiO)₄Si) as an intensive peak at −110 ppm. The presence of Q³ ((SiO)₃SiOH) and Q² ((SiO)₂Si(OH)₂) sites is practically undetected, indicating that after immobilization the amount of unreacted silanol groups onto MSN surface has significantly decreased. In addition, two signals are observed at −63 and −59 ppm corresponding to T² (RSi(OSi)₂(OR′)) and T³ (RSi(OSi)₃) sites, which confirms the incorporation of ionic liquid within the MSN support. As an example, in the ¹³C CP MAS NMR spectrum of Ti–MSN (Fig. S3a†) the signals of the methyl and methyne groups of the isopropoxide ligand appear at 23 and 78 ppm, respectively. An additional peak is observed at 65 ppm due to the methyne carbon atom of coordinated 2-propanol.¹⁴ The ¹³C PDA MAS NMR spectrum of Ti–CoIMILCl–SBA-15 (Fig. S3b†) exhibits signals at 7, 22 and 49 ppm due to the carbon atoms of the propyl chain –Si–CH₂–, –CH₂–CH₂–CH₂– and –CH₂–CH₂–CH₂–N, respectively. The signals attributed to the imidazolium group appear at 135, 122 and 34 ppm due to the carbon atoms (N–C–N), (N–CH–CH–N) and to the methyl group. Additional signals at 16 and 58 ppm are due to the carbons of the unreacted ethoxy groups. The spectrum also shows signals at 21 and 77 ppm assigned to the methyl carbon and methyne carbon of isopropoxide groups.

The FT-IR spectrum of MSN (Fig. 2) shows characteristic bands at 3441 cm⁻¹ assigned to O–H stretching vibrations and the remaining physisorbed water molecules, at 1629 cm⁻¹ due to deformation vibrations of the adsorbed water molecules and at 1082, 959 and 806 cm⁻¹ attributed to the Si–O stretching vibrations. The titanium materials functionalized with different ionic liquids (Fig. 2) display additional bands at 2979 and 2941 cm⁻¹ due to the ν(C–H) stretching vibrations and around 1574 and 1456 cm⁻¹ corresponding to the stretching vibrations of the imidazole group (see Fig. S4†). A comparison between the FT-IR spectra of the titanium and IMILCl containing MSN materials is shown in Fig. 3. As can be seen, all spectra exhibit the characteristic peaks of IMILCl.

Fig. 2 FTIR spectra of (a) pure MSN (b) Ti–IMILCl–MSN, (c) Ti–IMILPF₆–MSN and (d) Ti–IMILBF₄–MSN.

Fig. 3 FTIR spectra of (a) Ti–IMILCl–SBA-15, (b) Ti–IMILCl–MSN (c) Ti–CoIMILCl–MSN, (d) 5.4% Ti–CoIMILCl–SiO₂ and (e) IMILCl–Ti–MSN.

DRUV-vis spectra were recorded in order to study the immobilized titanium environment. The spectrum of Ti–IMILCl–SBA-15 exhibits a predominant band at 215 nm attributed to the imidazolium group (Fig. 4b) and a broad band in the range 210–240 nm, indicating the presence of tetrahedral Ti(IV) species. Similar absorption spectra are observed for Ti–IMILPF₆–MSN and Ti–IMILBF₄–MSN (Fig. 5). According to these results the steric hindrance imposed by the ionic liquid grafted to the support prevents the oligomerization between titanium species on the silica surface. The spectra of Ti–IMILCl–MSN (Fig. 6) and 5.4% Ti–CoIMILCl–SiO₂ (Fig. S5†) are particularly interesting, they show an additional weak band at 260 nm, suggesting the presence of oligomeric titanium species in these materials. This fact is also supported by the absorption spectra recorded for IMILCl–Ti–SBA-15, IMILCl–Ti–MSN and Ti–MSN, which shows a broad band red shifted in the range 260–300 nm, typically assigned to titanium with higher coordination environments. These results suggest the increase of titanium oligomerization with the decrease of ionic liquid loading and subsequent increase of the average IMILCl intermolecular distance.

Fig. 4 DRUV-vis spectra of titanium supported SBA-15 catalysts (a) IMILCl–Ti–SBA-15 (b) Ti–IMILCl–SBA-15.

Fig. 5 DRUV-vis spectra of titanium supported MSN catalysts.

Fig. 6 ^49/47Ti MAS NMR spectra of titanium supported MSN catalysts.

The ^47/49Ti MAS NMR spectra was recorded in order to gain extra information about the titanium environment. All spectra show broadened signals and overlapping resonances from ⁴⁷Ti and ⁴⁹Ti isotopes.¹⁵ As can be seen in Fig. 6, the immobilization of ionic liquid in Ti–IMILCl–MSN and IMILCl–Ti–MSN results in an increase of the signal width compared to Ti–MSN, indicating that the presence of ionic liquid causes asymmetrical electron distribution at the titanium centre in these materials. This behaviour has been observed previously by our group,¹⁶ for materials obtained by simultaneous heterogenization of [TiCl₃(η⁵-C₅HMe₄)] and a silylating agent onto MCM-41. The lower sterical requirement of the silylating agent Me–Si–O₃≡ compared to Me₃Si–O–Si≡ produces a decrease in the line width. On comparing the spectrum of Ti–IMILCl–SBA-15 (Fig. 7) with Ti–IMILCl–MSN the signals observed for the former are slightly shielded suggesting the influence of the silica support in the electronic properties or electron density of the titanium centre.

Fig. 7 ^49/47Ti MAS NMR spectra of titanium supported SBA-15 and SiO₂ catalysts.

Voltammetry measurements have been performed to test the electro activity of the grafted titanium complexes onto silica based materials. The experiments have been carried out at room temperature, using a three-electrode-single compartment electrochemical cell. A modified carbon paste with titanium immobilized material has been used as working electrode, an Ag/AgCl/KCl 3 M as reference electrode and a platinum rod as counter electrode. In Fig. 8 the differential pulse voltammograms of the immobilized titanium silica materials synthesized in this study are shown. The measurements have been recorded in aqueous media immediately following immersion of the fresh smooth electrode surface into the electrolyte buffer phosphate solution (pH 7.4). As can be seen a well-defined reduction peak is observed in all samples. DP voltammogram of Ti–MSN material shows a reduction peak at −0.80 V attributed to the redox system Ti(IV)/Ti(III). The materials also containing ionic liquid grafted onto silica surface show similarly one peak associated to the reduction of grafted titanium(IV) complexes, but in this case there is a clear shift of the peak to more negative potential values, −1.06 V, which indicates higher electron density at the titanium atom. There is no difference between Ti–IMILCl–MSN and IMILCl–Ti–MSN suggesting that the synthetic procedure does not render different structural forms. On the contrary, the reduction potential value observed for Ti–IMILCl–SBA-15 material is higher −1.15 V (more negative) suggesting the influence of the support as previously observed in the ^47/49Ti MAS NMR studies. Under identical experimental conditions, no peaks were observed with a pure silica modified carbon paste electrode; therefore the peaks have been assigned to the redox system Ti(IV)/Ti(III) demonstrating that titanium is readily accessible to electron transfer reactions in aqueous medium.

Fig. 8 Differential pulse voltammograms (75 mV modulation amplitude) of immobilized titanium silica materials immediately after immersion in aqueous phosphate buffer pH 7.4 as electrolyte vs. Ag/AgCl/KCl (3 M) as reference electrode.

This behaviour is supported by the cyclic voltammograms recorded for the samples. This technique, less sensitive that DPV, shows in all cases, one reduction peak attributed to the irreversible reduction of Ti(IV) species. As an example, Fig. 9 shows the cyclic voltammograms of Ti–IMILCl–SiO₂ and Ti–IMILCl–SBA-15. In both samples, a well-defined reduction peak is observed around −1.20 V with an associated oxidation peak in the reverse scan in the former material. There is a linear correlation between current intensity of the peak and square root of scan rate, showing that the kinetics of the overall process is controlled by diffusion behaviour. This corresponds most probably to the pseudo-diffusion of electrons between two adjacent electroactive centres (as titanium atoms are firmly attached to the silica walls in the porous materials) and to counter ion diffusion since the reduction process of titanium requires charge balance by electrolyte cation. Therefore, the redox process is governed by the diffusion of the electrolyte cations to the pore system of the silica material, by their mobility within the channels and eventually by the location and hence the accessibility of the titanium centres.

Fig. 9 Cyclic voltammograms of Ti–IMILCl–SiO₂ and Ti–IMILCl–SBA-15 immediately after immersion in aqueous phosphate buffer pH 7.4 as electrolyte vs. Ag/AgCl/KCl (3 M) as reference electrode (scan speed from 50–400 mV s⁻¹).

It is well known that titanium atoms located at the surface of the silica wall act as Lewis acid sites and react with Lewis bases through six-fold coordinated structures. In the presence of water, 4 and 6 coordinated Ti(IV) species are converted into 6-coordinate aqua hydroxido complexes after hydrolysis of isopropoxide units bonded to titanium and water coordination. So, under the experimental conditions chosen in this study the peak observed at −1.06 V can be tentatively assigned to the reduction process Ti(IV)/Ti(III) of octahedral titanium centres. Fig. 10a shows the differential pulse voltammograms of Ti–IMILCl–MSN recorded immediately after the immersion in the aqueous phosphate buffer electrolyte solution and subsequent minutes. As can be seen, two cathodic peaks resulted as a consequence of two different types of titanium centres, probably due to distinct structural forms. The intensity and wave profile of the cathodic peaks at −1.05 and −0.88 V are time dependent, the intensity of the lowest potential peak decreases whiles the one at highest potential increases. By analogy with the behaviour established by titanium alkoxides in presence of water it is reasonable to consider firstly the formation of immobilized hydroxide titanium complexes and subsequent condensation through water release to form supported bridged μ-oxo species (Scheme 3).^17,18 This behaviour is corroborated by cyclic voltammetry (Fig. S6–S8†), the cyclic voltammograms recorded immediately after the immersion in the aqueous electrolyte solution and 10 minutes after immersion. Two cathodic peaks resulted as a consequence of two different types of titanium centres, probably owing to distinct structural forms. The intensity and wave profile of the cathodic peaks at −0.96 and −1.22 V are time dependent. Fig. 10b shows the voltammogram recorded for Ti–IMILCl–SBA-15, as can be seen the initial cathodic peak shifted from −1.22 to −1.14 V at very short times after water contact, suggesting a very fast hydrolysis and condensation reaction of titanium species grafted onto the silica surface.

Fig. 10 Differential pulse voltammograms (75 mV modulation amplitude) of Ti–IMILCl–MSN and Ti–IMILCl–SBA-15 materials at different times after immersion in aqueous phosphate buffer pH 7.4 as electrolyte vs. Ag/AgCl/KCl (3 M) as reference electrode.

Scheme 3 Hydrolysis and condensation of grafted titanium species.

The synthesized mesoporous silica nanoparticles present spherical morphology and uniform size, as can be seen by SEM analysis (Fig. S9†). Fig. 11 shows some representative TEM images for titanium containing MSNs materials. The obtained MSNs present mesoporous structure, which is maintained after the immobilization. The average size of Ti–IMILCl–MSNs was 76 nm and 84 nm in the case of Ti–CoIMILCl–MSN material. The particles of the materials containing ionic liquid are less aggregated than those of the parent support, suggesting that the ionic liquid efficiently disperses the particles and prevents them from aggregation. Energy-dispersive X-ray spectrum of 9.45% Ti–MSN (Fig. S10a†) confirms the presence of titanium and in the case of Ti–IMILCl–MSN (Fig. S10b†) the presence of titanium and chloride, the anion of the imidazolium based ionic liquid.

Fig. 11 TEM images of (a) MSN, (b) 9.45% Ti–MSN, (c) Ti–IMILCl–MSN and (d) Ti–CoIMILCl–MSN.

In thermogravimetric curves (TG) of titanium catalysts, all materials exhibit a loss of mass below 197 °C attributed to physically adsorbed water. The degradation process of isopropoxy ligands bonded to the titanium atom occurs in the range 197–350 °C for ionic liquid containing materials, Ti–IMILCl–MSN (Fig. S11b†) and IMILCl–Ti–MSN (Fig. S11c†), and between 184–316 °C for Ti–MSN material (Fig. S11a†). The differential thermogravimetric (DTG) curve for Ti–MSN shows a well defined peak at 268 °C, narrower than those of ionic liquid containing materials which suggests that the ionic liquid interacts with titanium complex immobilized on the surface as previously observed by Ti MAS NMR and electrochemical studies. In addition, the curves of Ti–IMILCl–MSN and IMILCl–Ti–MSN show at around 400 °C an additional weight loss, consequence of ionic liquid degradation process.¹⁹

The titanium materials were tested in ROP of ε-CL and the results obtained are summarized in Table 2. A control experiment was carried out by using IMILCl–SBA-15 as catalyst (entry 1), obtaining a 0.5% conversion value, which confirms that the titanium species are the active sites in the polymerization process. As can be seen, the catalytic activity is titanium loading dependent, the reaction time decreases to 2 h when the titanium content increases from 3.7% to 9.45% for Ti–MSN (entries 5 and 6) and from 3.5% to 5.4% for Ti–CoIMILCl–SiO₂ (entries 12 and 13). The materials 9.45% Ti–MSN, 3.1% Ti–IMILCl–MSN and 5.4% Ti–CoIMILCl–SiO₂ exhibit total conversion of the monomer in 2 h with molecular weights in the range 15 118–17 447 (entries 6, 8 and 13). It was observed that 3.1% Ti–IMILCl–MSN produces polymers with the lowest molecular weight distribution (PDI = 1.13). In addition, on comparing 3.1% Ti–IMILCl–MSN with 3.7% Ti–MSN (entries 6 and 5), with similar titanium content, the former material exhibits higher catalytic activity as it achieves higher conversion in a shorter time. Comparable results were obtained for the pair Ti–SBA-15 and IMILCl–Ti–SBA-15 (entries 2 and 4). Ti–SBA-15 with the highest titanium loading requires longer reaction time to reach total conversion of the monomer. These results suggest that catalytic activity can be enhanced by the presence of ionic liquid grafted on the silica surface. The effect of the ionic liquid can be discussed in more detail on the basis of the results obtained with the materials 3.14% Ti–IMILCl–MSN and 5.92% Ti–CoIMILCl–MSN (entries 8 and 11) prepared by post synthesis and one step co-condensation method, respectively, since these two procedures allow the incorporation of different amounts of titanium and ionic liquid. The former material with higher degree of functionalization achieves higher conversion values. To determine the influence of the ionic liquid size, Ti–IMILPF₆–MSN and Ti–IMILBF₄–MSN materials were also studied (entries 9 and 10). As can be seen for Ti–IMILBF₄–MSN, total conversion was achieved after 24 h and Ti–IMILPF₆–MSN only gave 26% conversion in the same 24 h. These results reveal that the sterical effects imposed by the ionic liquid substantially inhibit the catalytic activity. The catalytic activity obtained for titanium immobilized onto MSN and silica functionalized with ionic liquids is higher than the results reported by Long et al.¹¹ with magnetic nanoparticles supported aluminium isopropoxide (95% conversion in 10.5 h) and the results obtained by Khan et al.⁹ with calcium alkoxide-functionalized silica (90% conversion in 100 h).

Table 2 Polymerization of ε-CL catalysed by the different titanium supported catalysts^a

Entry	Initiator	Ti (%)	Conversion^b (%)	Time (h)	Mn^c	PDI^c
a Reaction conditions: 100 mg of catalyst, [M0/I0] = 100, temperature = 80 °C, toluene as solvent (10 mL).b Determined by ^1H NMR analysis.c Measured by GPC at 27 °C in THF relative to polystyrene standards.
1	IMILCl–SBA-15	—	0.5	24	—	—
2	Ti–SBA-15 (ref. 13)	15.5	100	24	6900	1.20
3	Ti–IMILCl–SBA-15	2.37	83	24	—	—
4	IMILCl–Ti–SBA-15	7.43	100	14	19 197	1.27
5	3.7% Ti–MSN	3.70	100	4	18 928	1.14
6	9.45% Ti–MSN	9.45	100	2	17 447	1.48
7	IMILCl–Ti–MSN	7.90	100	6	16 213	1.21
8	Ti–IMILCl–MSN	3.14	100	2	15 118	1.13
9	Ti–IMILPF6–MSN	2.62	26	24	—	—
10	Ti–IMILBF4–MSN	1.32	100	24	—	—
11	Ti–CoIMILCl–MSN	5.92	100	4	15 937	1.29
12	3.5% Ti–CoIMILCl–SiO2	3.50	100	4	13 332	1.05
13	5.4% Ti–CoIMILCl–SiO2	5.40	100	2	16 377	1.38

---

The molecular weights of PLCs produced are in the range 15 000–19 000 g mol⁻¹ and, in most cases, show a narrow molecular weight distribution (close to one). The polymer surface produced by 9.45% Ti–MSN reveals a cauliflower morphology (Fig. 12a). The surface of the particle becomes rough, comprising of subgrains with diameter of 10–40 μm, which suggests that there is replication phenomena, common in the heterogeneous polymerization of olefins.²⁰ The polymerization first takes place on the outer part of the particle, while the centre of the support remains unfragmented. On the contrary, the morphology of PCL produced by Ti–IMILCl–MSN is irregular (Fig. 12b), the particle is clearly an assembly of smaller structures. This fact supports the presence of the active sites on the silica surface; monomer diffuses through the IMILCl supported on silica nanoparticles adsorbs on the layer of the polymer surrounding the catalyst and diffuses through this layer to the active sites on the surface where polymerization takes place.²¹ Our previous work¹³ demonstrated that Ti–SBA-15 also produces irregular PCL with an assembly of smaller structures. In these studies, we established that the initiation mechanism, using the material Ti–SBA-15, proceeded via coordination–insertion reaction through an acyl–oxygen bond cleavage. Since the ¹H NMR spectra of the polymer obtained with titanium containing ionic liquid materials (Fig. S12†) exhibit signals at 3.64 and at 4.99 ppm, corresponding to the –CH₂–OH end group and the methyl groups of isopropyl ester (–COOCH(CH₃)₂), respectively; a similar coordination–insertion initiation mechanism is proposed.

Fig. 12 SEM micrographs of (a) PLC produced by 9.45% Ti–MSN and (b) PLC produced by Ti–IMILCl–MSN.

To study the reusability of the titanium catalysts, a typical polymerization reaction was quenched with isopropanol in order to regenerate the titanium catalyst. The polymer containing solution was separated by filtration and the heterogeneous catalyst washed with dichloromethane and dried under vacuum before reutilization. The structure of the catalysts is maintained, as confirmed by FT-IR (Fig. S13†) and TEM studies (Fig. S14†). The recovery efficacies were around 80% in the case of catalysts IMILCl–Ti–SBA-15–R, 5.4% Ti–CoIMILCl–SiO₂–R and 3.7% Ti–MSN–R, similar efficiency was achieved by Lee et al.¹⁰ for silica-supported tin(II) methoxide catalysts. However, for titanium immobilized onto ionic liquid-functionalized MSN catalysts the recovery efficacies were in the range of 20–40%, since the small size of the nanoparticles and the presence of the ionic liquid makes it very difficult to separate them by filtration from the polymer. The results reveal that IMILCl–Ti–SBA-15–R and 5.4% Ti–CoIMILCl–SiO₂–R can be recovered and reused under similar experimental conditions, obtaining conversion values of 94% in 4 h and 99% in 16 h, respectively. Nevertheless, for reused 3.7% Ti–MSN–R catalyst a low conversion (55% conversion) was observed, suggesting the existence of metal leaching. Accordingly, the recovery experiment carried out for the Ti–SBA-15 catalyst¹³ showed inferior conversion (72% conversion) to that of IMILCl–Ti–SBA-15–R. These results suggest that the ionic liquid may provide additional protection, stabilizing the titanium species and improving the reusability of the catalyst.

3. Conclusions

In summary, immobilized titanium-containing ionic liquid materials have been prepared by using different supports and methodologies and tested as catalysts in the ROP of ε-CL. The prepared materials have been characterized by DRUV-vis, ^49/47Ti MAS NMR spectroscopy and solid state electrochemical studies demonstrating that the ionic liquid influences the chemical environment and the electronic properties of the anchored titanium species. The studies show that the catalytic activity of the materials depends on the titanium and ionic liquid loading. Thus, the material Ti–IMILCl–MSN, with low content of titanium and containing ionic liquid shows high activity (100% conversion in 2 h) rendering poly(ε-caprolactone) with low polydispersity value (PDI = 1.13). The recovery experiments carried out for titanium supported catalysts suggest that ionic liquid helps to stabilize the titanium species improving the reusability of these materials.

4. Experimental and methods

Materials

All reactions were performed using standard Schlenk tube techniques under an atmosphere of dry nitrogen or argon. Tetraethylorthosilicate (TEOS) 98%, NaBF₄ 98% and NaOH 97% were purchased from Sigma Aldrich and used as received. Hexadecyltrimethylammonium bromide (CTAB) 99.7% and KPF₆ 99% were purchased from Acros and used as received. Titanium tetraisopropoxide 97% (Ti(OⁱPr)₄) was purchased from Sigma Aldrich distilled and stored under a nitrogen atmosphere prior to use. ε-Caprolactone was purchased from Fluka, and was refluxed over CaH₂, distilled and stored under a nitrogen atmosphere prior to use. Hydrochloric acid 35% was purchased from Panreac. Organic solvents were purchased from SDS, distilled and dried before use according to conventional literature methods. Water was obtained from a Millipore Milli-Q system (Waters, USA). The hexagonal material (SBA-15) was prepared using poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol) (P123) in an acidic medium, according to the method of Zhao et al.²² The ionic liquid 1-methyl-3-[(triethoxysilyl)propyl]imidazolium chloride (IMILCl) was synthesized according to the procedure reported in the literature.²³ The synthesis of the ionic liquids 1-methyl-3-[(triethoxysilyl)propyl]imidazolium chloride (IMILCl), 1-methyl-3-[(triethoxysilyl)propyl]imidazolium hexafluorophosphate (IMILPF₆) and 1-methyl-3-[(triethoxysilyl)propyl]imidazolium tetrafluoroborate (IMILBF₄) are described in detail in the ESI.†

Characterization

X-ray diffraction (XRD) patterns of the silicas were obtained on a Phillips Diffractometer model PW3040/00 X'Pert MPD/MRD at 45 KV and 40 mA, using Cu-Kα radiation (λ = 1.5418 Å). N₂ gas adsorption–desorption isotherms were obtained using a Micromeritics TriStar 3000 analyzer, and pore size distributions were calculated using the Barret–Joyner–Halenda (BJH) model on the adsorption branch. Infrared spectra were recorded on a Nicolet-550 FT-IR spectrophotometer (in the region 4000 to 400 cm⁻¹) as nujol mulls between polyethylene pellets and KBr disks. Proton-decoupled ²⁹Si MAS NMR spectra were recorded on a Varian-Infinity Plus 400 MHz spectrometer operating at 79.44 MHz proton frequency (4 μs 90° pulse, 1024 transients, spinning speed of 5 MHz). Cross polarization ¹³C CP/MAS NMR spectra were recorded on a Varian-Infinity Plus 400 MHz spectrometer operating at 100.52 MHz proton frequency (4 μs 90° pulse, 4000 transients, spinning speed of 6 MHz, contact time 3 ms, pulse delay 1.5 s). Room-temperature wide-NMR line solid-state ^49/47Ti MAS NMR spectra were obtained on a Varian Unity-400 spectrometer with applied field of 9.4 T using a 7 mm MAS probe in static conditions. The NMR ^49/47Ti Larmor frequency of 22.54 and 22.55 MHz, respectively was used. The spectra were obtained using a two pulses echo sequenced with phase cycling. This pulse sequence was first described by Kunwar et al.²⁴ for eliminating artefacts due to acoustic, transmitted and probe circuits and ringing. The interval between successive accumulations (1 s) was chosen to avoid saturation effects. The number of accumulations (20.000) was selected to maintain an appropriated signal-to-noise ratio (S/N = 20). Chemical shifts were referenced to titanium isopropoxide (which is shifted by −850 ppm against the conventional standard for titanium used as an internal reference (0 ppm) (TiCl₄)). The titanium content was determined by XRF. Scanning electron micrographs and morphological analysis were carried out on a XL30 ESEM Philips. The DRUV-vis spectroscopic measurements were carried out on a Varian Cary-500 spectrophotometer equipped with an integrating sphere and polytetrafluoroethylene (PTFE) as reference, with d = 1 g cm⁻³ and thickness of 6 mm. Thermogravimetric analysis was performed using a Setsys 18 A (Setaram) thermogravimetric analyzer. The cyclic voltammograms were recorded with a potentiostat/galvanostat Autolab PGSTAT302 Metrohm. A conventional three electrode system was used throughout the electrochemical experiments at the room temperature with a modified carbon paste electrode (CPE) as working electrode, a platinum wire as auxiliary electrode, and a saturated Ag/AgCl/KCl (3 M) electrode (Methrom) as reference electrode against which all potentials were measured in this paper. The phosphate buffer used as electrolyte solution in the cell was purged with highly purified nitrogen gas for at least 5 min to remove dissolved oxygen and then a nitrogen atmosphere was kept over the solution during measurements.

Preparation of catalysts

1. Synthesis of mesoporous silica nanoparticles (MSN). The mesoporous silica nanoparticles were synthesized according to the method described by Zhao et al.²⁵ To 1440 mL of nanopure water, hexadecyltrimethylammonium bromide (CTAB, 3.00 g, 8.22 mmol) and sodium hydroxide aqueous solution (2.00 M, 10.50 mL) were added and stirred. The temperature of the mixture was adjusted to 80 °C. Tetraethylorthosilicate (TEOS, 15.00 mL, 67.2 mmol) was added dropwise to the surfactant solution under vigorous stirring. The mixture was allowed to stir for 2 h to give a white precipitate. The solid crude product was isolated by filtration, washed with nanopure water and methanol and dried under vacuum. The surfactant template (CTAB) was removed by calcination at 550 °C for 12 h.

2. Functionalization of SBA-15 or MSN with ionic liquid.
2.1 Synthesis of IMILCl–SBA-15, IMILCl–MSN, IMILPF₆–MSN and IMILBF₄–MSN hybrid materials through grafting. A solution of ionic liquid 1-methyl-3-[(triethoxysilyl)propyl]imidazolium chloride (IMILCl) (0.65 g, 2 mmol) in chloroform was added to 1.0 g of SBA-15 or MSN, previously dehydrated at 150 °C under vacuum for 12 h. The mixture was refluxed at 80 °C for 24 h. The materials were filtered and washed with chloroform (3 × 30 mL). The resulting materials, designated as IMILCl–SBA-15 and IMILCl–MSN, were dried under vacuum and stored under inert atmosphere. The mesoporous silica nanoparticles were functionalized with IMILPF₆ and IMILBF₄ ionic liquid following the same procedure to obtain IMILPF₆–MSN and IMILBF₄–MSN, respectively.
2.2 Synthesis of CoIMILCl–MSN and CoIMILCl–SiO₂ hybrid materials by one step co-condensation. In a typical synthesis, to a solution of CTBA (1.0 g, 2.74 mmol) in 480 mL of ultrapure water, 3.5 mL of NaOH 2 M was added and the mixture heated to 90 °C. Then, TEOS (5 mL, 22.4 mmol) and IMILCl (3.62 g, 11.2 mmol) were added consecutively under vigorous stirring. The mixture was allowed to stir for 2 h and aged for 24 h at the same temperature. The suspension was filtered and the solid washed with water several times. The surfactant was removed by Soxhlet extraction with boiling ethanol for 48 h. The resulting material, designated as CoIMILCl–MSN, was dried under vacuum and stored under inert atmosphere.

The hybrid material CoIMILCl–SiO₂ was synthesized according to the method described by Wang et al.²⁶ To an aqueous solution (45 mL) of 4.0 g of P123, 16.3 g of HCl (35 wt%) was added and the mixture stirred until complete dissolution of the reactants. Then, TEOS (31 mL, 139 mmol) and IMILCl (4.49 g, 13.9 mmol) were added consecutively to the solution under vigorous stirring. The mixture was stirred at 45 °C for 2 h and the suspension was heated in Teflon coated stainless steel autoclave at 100 °C for 72 h. The surfactant was removed by Soxhlet extraction with boiling ethanol for 48 h.

3. Synthesis of immobilized titanium catalysts.
3.1 Immobilization of titanium isopropoxide onto MSN. 1.0 g of dehydrated MSN was suspended in 30 mL of toluene and Ti(OⁱPr)₄ (0.6 mL, 2 mmol) or (1.2 mL, 4 mmol) was added and the mixture stirred for 24 h at room temperature. Afterwards, 15 mL of hexane was added to the suspension and the mixture was stirred 10 extra minutes to achieve particle flocculation. The white solid product was isolated by filtration and washed repeatedly with hexane (3 × 30 mL). The resulting materials, designated as 3.7% Ti–MSN and 9.45% Ti–MSN containing 3.7% and 9.45% of titanium, respectively, were dried under vacuum and stored under inert atmosphere.
3.2 Immobilization of titanium isopropoxide onto ionic liquid functionalized support. Titanium isopropoxide was immobilized onto the materials IMILCl–SBA-15, IMILX–MSN (X = Cl, PF₆, BF₄), CoIMILCl–MSN and CoIMILCl–SiO₂. To a toluene suspension of dehydrated ionic liquid functionalized supports, 4 mmol g⁻¹ of Ti(OⁱPr)₄ was added. The mixture was allowed to react for 24 hours at room temperature. The mixture was separated by filtration and the solid dried under vacuum. The resulting materials were stored under inert atmosphere and labelled as Ti–IMILCl–SBA-15, Ti–IMILCl–MSN, Ti–IMILPF₆–MSN, Ti–IMILBF₄–MSN, Ti–CoIMILCl–MSN and x% Ti–CoIMILCl–SiO₂ (x = 3.5 and 5.4% titanium content), respectively.
3.3 Immobilization of ionic liquid onto titanium immobilized materials. Firstly, titanium contained materials were synthesized. To a toluene suspension of dehydrated supports (MSN or SBA-15), 4 mmol g⁻¹ of Ti(OⁱPr)₄ was added, and thoroughly washed to obtain the functionalized materials Ti–MSN and Ti–SBA-15. Secondly, the ionic liquid 1-methyl-3-[(triethoxysilyl)propyl]imidazolium chloride (IMILCl, 2 mmol g⁻¹) was immobilized onto Ti–MSN and Ti–SBA-15 materials, using the same procedure described in 2.1 Section. The resulting solids were washed and dried under vacuum to obtain the materials designated as IMILCl–Ti–MSN and IMILCl–Ti–SBA-15, respectively.

ε-Caprolactone polymerization

Polymerization of ε-caprolactone was carried out under inert atmosphere using a flask equipped with a magnetic stirrer. Predetermined amounts of purified ε-CL, dried toluene, and heterogeneous catalysts were charged and the monomer was added by syringe with vigorous magnetic stirring. The polymerization studies were carried out at 80 °C. In the kinetic and mechanism studies small aliquots were taken at different reactions times depending on the catalyst (see results and discussion section). After the reaction time had finished the ε-CL polymerization was quenched by addition of 5 mL of methanol, the solid phase was separated by filtration and the polymer precipitated with an excess of methanol. The polymer was washed with methanol and dried at 50 °C under vacuum for 12 h. The sample was analyzed by ¹H NMR spectroscopy (CDCl₃) and gel permeation chromatography (GPC) to determine the average molecular weights M_n, M_w of the produced polymer and molecular weight distribution (PDI). The monomer conversion was determined from the relative intensity of the NMR signals at δ 4.20 and 4.02 ppm corresponding to the OCH₂ resonance in ε-CL and PCL, respectively.

Electrode preparation

The carbon paste electrodes (CPE) were prepared by mixing with a pestle in an agate mortar the studied titanium silica based materials with graphite (Metrohm) (6–10% (w/w) ratio) and mineral oil as agglutinant (Aldrich) until a uniform paste was obtained. The resulting material was packed into the end of a Teflon cylindrical tube equipped with a screwing stainless steel piston providing an inner electrical contact. All of the initial electrode activity could always be restored by simply removing the outer layer of paste by treatment with polishing paper.

Acknowledgements

We gratefully acknowledge financial support from the MICINN (project CTQ2012-30762) and Universidad Rey Juan Carlos-Banco Santander (Excellence group QUINANOAP). We would like to thank Carmen Force Redondo (Universidad Rey Juan Carlos) for useful discussions in ^49/47Ti MAS NMR spectra.

References

1. R. K. Sharma, S. Sharma, S. Dutta, R. Zboril and M. B. Gawande, Green Chem., 2015, 17, 3207–3230.
2. H. Li, P. S. Bhadury, B. Song and S. Yang, RSC Adv., 2012, 2, 12525–12551.
3. C. V. Doorslaer, J. Wahlen, P. Mertens, K. Binnemans and D. De Vos, Dalton Trans., 2010, 39, 8377–8390.
4. C. G. Pitt, Poly ε-caprolactone and its copolymers, in Biodegradable polymers as drug delivery systems, ed. R. Langer and M. Chasin, Marcel Dekker, New York, 1990, pp. 71–120.
5. C.-Y. Li, C.-J. Yu and B.-T. Ko, Organometallics, 2013, 32, 172–180.
6. Y.-H. Chen, Y.-J. Chen, H.-C. Tseng, C.-J. Lian, H.-Y. Tsai, Y.-C. Lai, S. C. N. Hsu, M. Y. Chiang and H.-Y. Chen, RSC Adv., 2015, 5, 100272–100280.
7. E. Martin, P. Dubois and R. Jérôme, Macromolecules, 2003, 36, 7094–7099.
8. E. Kim, E. W. Shin, I.-K. Yoo and J. S. Chung, J. Mol. Catal. A: Chem., 2009, 298, 36–39.
9. J. H. Khan, F. Schue and G. A. George, Polym. Int., 2010, 59, 1506–1513.
10. E. J. Lee, K. M. Lee, J. Jang, E. Kim, J. S. Chung, Y. Do, S. Yoon and S. Y. Park, J. Mol. Catal. A: Chem., 2014, 385, 68–72.
11. W. Long, C. S. Gill, S. Choi and C. W. Jones, Dalton Trans., 2010, 39, 1470–1472.
12. N. Wanna, T. Kraithong, T. Khamnaen, P. Phiriyawirut, S. Charoenchaidet and J. Tantirungrotechai, Catal. Commun., 2014, 45, 118.
13. Y. Pérez, I. Hierro, L. Zazo, R. Fernández-Galán and M. Fajardo, Dalton Trans., 2015, 44, 4088–4101.
14. A. O. Bouh, G. L. Rice and S. L. Scott, J. Am. Chem. Soc., 1999, 121, 7201–7210.
15. M. E. Smith, Annu. Rep. NMR Spectrosc., 2001, 43, 121–175.
16. R. Ballesteros, M. Fajardo, I. Sierra, C. Force and I. Hierro, Langmuir, 2009, 25, 12706–12712.
17. D. Osella, M. Ravera, C. Floriani and E. Solari, J. Organomet. Chem., 1996, 510, 45–50.
18. L. R. Groom, A. F. Russell, A. D. Schwarz and P. Mountford, Organometallics, 2013, 32, 7520–7539.
19. P. Cruz, Y. Pérez, I. Hierro and M. Fajardo, Microporous Mesoporous Mater., 2016, 220, 136–147.
20. X. Zheng, M. Smit, J. C. Chadwick and J. Loos, Macromolecules, 2005, 38, 4673–4678.
21. T. F. McKenna and J. B. P. Soares, Chem. Eng. Sci., 2001, 56, 3931–3949.
22. D. Zhao, Q. Huo, J. Feng, B. F. Chmelka and G. D. Stucky, J. Am. Chem. Soc., 1998, 120, 6024–6036.
23. H. Wang, B. Wang, C.-L. Liu and W.-S. Dong, Microporous Mesoporous Mater., 2010, 134, 51–57.
24. A. C. Kunwar, G. L. Turner and E. Olfield, J. Magn. Reson., 1986, 69, 124–127.
25. Y. Zhao, B. G. Trewyn, I. Slowing and V. S.-Y. Lin, J. Am. Chem. Soc., 2009, 131, 8398–8400.
26. F. Wang, Z. Zhang, J. Yang, L. Wang, Y. Lin and Y. Wei, Fuel, 2013, 107, 394–399.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01220e

---