David
Skoda
ab,
Ales
Styskalik
ab,
Zdenek
Moravec
a,
Petr
Bezdicka
c,
Jiri
Bursik
d,
P. Hubert
Mutin
e and
Jiri
Pinkas
*ab
aMasaryk University, Department of Chemistry, Kotlarska 2, CZ-61137 Brno, Czech Republic. E-mail: jpinkas@chemi.muni.cz
bMasaryk University, CEITEC MU, Kamenice 5, CZ-62500 Brno, Czech Republic
cInstitute of Inorganic Chemistry ASCR, CZ-25068 Husinec-Rez, Czech Republic
dInstitute of Physics of Materials ASCR, Zizkova 22, CZ-616 62 Brno, Czech Republic
eInstitut Charles Gerhardt UMR 5253 CNRS-UM-ENSCM, Université de Montpellier, Place Eugène Bataillon, 34095 Montpellier Cedex 05, France
First published on 13th July 2016
A novel non-hydrolytic sol–gel (NHSG) synthesis of mesoporous tin silicate xerogels is presented. The polycondensation between silicon tetraacetate, Si(OAc)4, and tetrakis(diethylamido)tin, Sn(NEt2)4, resulting in acetamide elimination leads to tin silicate xerogels containing Si–O–Sn linkages. The addition of Pluronic P123 or F127 templates provides homogeneous stiff gels that are, after template removal by calcination at 500 °C in air, converted to stable mesoporous silica xerogels with large surface areas (476 m2 g−1) and dispersed SnO2 nanoparticles (6–7 nm). Heat treatment of the as-prepared tin silicate gels in an inert N2 atmosphere leads to reduction and transformation to Sn nanoparticles (70–150 nm) embedded in a silica–carbon matrix. The composition and morphology of the xerogels, volatile reaction byproducts, and thermal transformations were followed by elemental analyses, IR spectroscopy, thermal analysis TG-DSC, nitrogen adsorption measurements, solid-state NMR spectroscopy, DRUV-vis spectroscopy, electron microscopy, and HT powder XRD. The SnO2–SiO2 xerogels were tested as potential catalysts for aminolysis of styrene oxide with aniline and for the Meerwein–Ponndorf–Verley reduction of 4-tert-butylcyclohexanone. The resulting reaction systems displayed good activity and selectivity.
Many reports are devoted to porous materials such as Sn-ZSM12, Sn-beta zeolites, Sn-MFI, Sn-MCM, Sn-SBA, and Sn-KIT-5.2,4,8,12,15–17 Preparation procedures for mesoporous tin silicate materials involve addition of a tin precursor (SnCl2, SnCl4·5H2O) into an appropriate mesoporous silica (MCM, SBA, KIT-5) reaction mixture. Aqueous solutions of TEOS with HCl and Pluronic P123 or F127 are used in the case of SBA or KIT-5 supports.8,18 The resulting mesoporous tin silicates were used in a variety of catalytic reactions and they displayed high activities. For example, Endud used Sn-MCM-48 molecular sieve (1291–1585 m2 g−1) as a catalyst for oxidation of benzyl alcohol with TBHP.17 An interesting method for the preparation of nanocrystalline Sn-beta zeolite catalysts for ring-opening hydration of epoxides was published by Tang.16 This approach is based on dealumination of commercial beta aluminosilicate zeolite with nitric acid and a subsequent reaction with (CH3)2SnCl2. Recently, Reddy successfully applied mesoporous (582–732 m2 g−1) tin silicates as catalysts for acyl Sonogashira coupling reactions with yields up to 97%.12
The major problem of hydrolytic sol–gel processes is the disparity in the hydrolysis reaction rates of different precursors. The non-hydrolytic method19–21 involves completely different mechanism of polycondensation reactions in non-aqueous media and offers decisive advantages, such as control of composition, homogeneity, structure, and texture.22 However, the non-aqueous synthesis of SnO2–SiO2 materials is not well developed and only two reports were published on this topic. The first describes a procedure based on the reaction of TEOS with tin(IV) acetate in anhydrous acetic acid.23 The second report introduces the twin polymerization method for the production of SnO2–SiO2 materials.24 When compared to conventional hydrolytic methods, non-hydrolytic sol–gel chemistry provides effective routes for the synthesis of homogeneous monolithic gels, nanoparticles25 and heterogeneous catalysts.19,20
We have recently reported that non-hydrolytic acetamide elimination26 can be successfully used for the synthesis of mesoporous titanosilicate, zirconium silicate, and aluminosilicate catalysts.27–29 In the work presented here, this reaction was used for the preparation of tin silicate xerogels. Templated one-pot sol–gel method provided homogeneous stiff gels with a high content of Sn. We were also able to prepare SnO2 nanoparticles embedded in mesoporous silica matrix which prevented them from aggregation. Resulting SnO2–SiO2 xerogels were tested as catalysts for aminolysis of styrene oxide. Heating of dried gels in nitrogen at 500 °C caused reduction of Sn(IV) species and Sn–silica–carbon nanocomposites were obtained.
Chemicals, such as Pluronic P123 (EO20PO70EO20, Mav = 5845 g mol−1), Pluronic F127 (EO100PO65EO100, Mav = 12600 g mol−1), tetrakis(diethylamido)tin Sn(NEt2)4 (99%), styrene oxide (97%), aniline (97%), 4-tert-butylcyclohexanone (99%), and nonane (99%), were purchased from Sigma-Aldrich. Si(OAc)4 was synthesized according to the published procedure.30 Toluene, dichloromethane and isopropanol were dried by standard methods and distilled before use. Pluronic P123 and F127 were dried under vacuum at 60 °C and dissolved in dry toluene and stored as 20 wt% solutions.
TG/DSC: (air, 5 °C min−1) weight loss at 1000 °C 69.40%; (N2, 5 °C min−1) weight loss at 1000 °C 69.21%.
SiSnF dried xerogel was calcined in a tube furnace at 400 (SiSnF-400) and 500 °C (SiSnF-500) for 3 h in air to remove the template. Carbonization of template and reduction of Sn species was performed by heat treatment under N2 at 400 (SiSnF-400N) or 500 °C (SiSnF-500N) for 3 h.
MAS NMR SiSnF-500 (500 °C, air): 29Si CPMAS (ppm), δ −110, −102, −92.
BET SiSnF-500 (500 °C, air) 191 m2 g−1, C = 94, Vtot = 0.39 cm3 g−1, d = 9.8 nm.
TG/DSC: (air, 5 °C min−1) weight loss at 1000 °C 61.71%; (N2, 5 °C min−1) weight loss at 1000 °C 61.45%.
MAS NMR: 29Si CPMAS (ppm), δ −109, −102.
BET: 476 m2 g−1, C = 80, Vtot = 0.58 cm3 g−1, d = 6.6 nm.
The MPV reduction31 of 4-tert-butylcyclohexanone was carried out in a 50 cm3 round-bottom Schlenk flask equipped with a reflux condenser connected to a dry N2 source. The reaction mixture containing the calcined tin silicate catalyst (25 mg), 4-tert-butylcyclohexanone (500 mg, 3.54 mmol), and dry 2-propanol (15 cm3, 0.20 mol) was stirred under reflux for 1 h.
Aminolysis of styrene oxide32 was performed in a 25 cm3 round-bottom Schlenk flask connected to a dry N2 source. The reaction mixture consisted of a calcined tin silicate xerogel (25 mg), dry toluene (5 cm3), aniline (0.456 cm3, 5.00 mmol), and styrene oxide (0.587 cm3, 5.00 mmol). This reaction mixture was heated at 50 °C for 1 h.
The products were analyzed by 1H NMR spectroscopy and the GC-MS method using 0.100 cm3 of nonane as an internal standard.
In the template-assisted reactions between Si(OAc)4 and Sn(NEt2)4, the yield of the product as well as the mass of starting precursors were precisely weighed to allow gravimetric estimation of the degree of condensation, DC = 100(ntotal − nresidual)/ntotal, where ntotal is the molar amount of organic groups in the starting materials and nresidual is molar amount of residual organic groups in the xerogel is based on the difference of theoretical and experimental yield. As the condensation reactions were never quantitative, the degree of condensation represents the relative difference between the maximum theoretical loss of Et2NC(O)CH3 (eqn (1)) in comparison to what is experimentally observed. This difference also defines the number of acetoxy groups on silicon and diethylamide groups on tin that are left in the matrix.
Si(OAc)4 + Sn(NEt2)4 → SiO2–SnO2 + 4Et2NAc | (1) |
The degrees of condensation (DC, %) for SiOSn xerogels (eqn (1)) determined by gravimetry reached 69% (Table 1).
All reactions proceed with the formation of transparent yellowish gels which are dried under vacuum to dark orange rubbery products. GC-MS analysis of volatile byproducts confirmed N,N-diethylacetamide as the only reaction byproduct of heterocondensation (Fig. 1S†). Templates were removed by calcinations in air at 400 and 500 °C. Heat treatment in the atmosphere of N2 at 400 and 500 °C was employed as an effective method for carbonization and reduction of Sn species to metallic tin nanoparticles.
IR spectra of dried gels presented in Fig. 1 show vibrational bands corresponding to Si–O–Sn linkages18,36,37 at 940–956 cm−1, Si–O–Si at 793–799 and 1014–1030 cm−1, and Sn–O–Sn38 at 667 cm−1. The residual acetoxy groups are represented by asymmetric and symmetric COO stretches at 1564 and 1450 cm−1, respectively. The difference between symmetric and asymmetric carboxylate vibrational bands is 110–120 cm−1 and according to Deacon–Phillips rules, this is indicative of their bidentate bridging mode on metal centers.39 Diethylamido groups display an absorption band at 1370 cm−1. An intense band at 1100 cm−1 is attributed to C–O–C vibrations of templates.
Fig. 1 IR spectra of dried tin silicate xerogels. Vibrational band labeled with T is characteristic for Pluronic templates. |
TG/DSC analyses of dried products performed under air atmosphere show the highest mass loss between 200 and 400 °C (54–60%). This decrease is caused by oxidation and decomposition of the residual organic groups and templates. An interesting behavior is observed by the DSC method where samples show exothermic peaks between 350 and 400 °C. These effects are attributed to oxidation of copolymer template. Exothermic peak between 400 and 530 °C can be connected with the crystallization of SnO2 in silicate matrix (Fig. 2 and 2S†). This crystallization effect displayed by the sample SiSnPA (Fig. 3S†) prepared in an autoclave at 110 °C is weaker than in the case of samples SiSnP and SiSnF (Fig. 2S† and 2) synthesized at 80 °C. This may imply better stabilization of Sn in the Sn–O–Si matrix in xerogels prepared by polycondensation at higher temperature. A higher degree of condensation and crosslinking agrees with a lower organic content and thus a higher residual mass left after heating to 1000 °C during the TG/DSC experiments (Table 1). DSC measurements performed in nitrogen atmosphere revealed endothermic peaks between 390 and 410 °C. These effects could correspond to reduction of Sn–O–Si species and subsequent formation of Sn metallic nanoparticles.40 In the case of SiSnP sample synthesized with P123, the maximum of endothermic peak is located at 399 °C (Fig. 4S†). The DSC curve of SiSnF sample prepared with F127 template shows an exothermic effect followed by an endothermic peak with the maxima at 406 °C (Fig. 2). The SiSnPA DSC curve (Fig. 5S†) of TG/DSC analysis performed under N2 illustrates weak exothermic and endothermic effects.
Fig. 2 TG/DSC curves of SiSnF xerogel. Analyses performed in air (left); analyses performed under N2 (left). |
The presence of SnO2 and metallic Sn particles mentioned above was established by the powder XRD diffraction analysis. The tin silicate xerogels calcined in air at 400 and 500 °C display diffractions of tetragonal cassiterite SnO2 (PDF 41-1445) (Fig. 3 and 6S†). Considering the width of SnO2 diffraction lines, the size of crystallites is very small and remains almost constant on heating up to 1200 °C (Fig. 7S, Table 1S†). The Rietveld refinement of the high temperature powder XRD patterns shows an average size of crystallites from 7 to 12 nm (Fig. 7S and 8S†). The size of about 7 nm is retained to 950 °C. A significant increase of the SnO2 nanoparticle size is observed between 1000 and 1200 °C when the size is increased from 8 to 12 nm (Fig. 7S†). In the case of the autoclave-synthesized sample, SiSnPA-500, SnO2 nanoparticles with the size of 5.9 nm were observed (Fig. 9S and 10S†). After calcination at 1200 °C the SnO2 nanocrystal size increased to 8.3 nm (Fig. 11S†). From this behavior we can conclude that the SnO2 nanocrystals are stabilized to high temperatures as they are enclosed in the pores of silicate matrix and their growth is largely prevented. The powder XRD patterns of the samples heated in N2 at 500 °C show diffractions assigned to metallic Sn (PDF 04-0673), SnO2 (PDF 41-1445), and also SnO (PDF 06-0395). The diffractions of metallic tin prove that partial reduction by carbon takes place during heating under N2 (Fig. 3 and 6S†). The size of Sn nanoparticles in the nanocomposites SiSnF-500N and SiSnPA-500N computed from Rietveld refinement is about 146 and 107 nm, respectively (Fig. 12S and 13S†). Powder XRD diffractogram of the SiSnF-400N sample (heated in N2 at 400 °C) (Fig. 14S†) does not display the lines of metallic Sn and only weak diffractions of SnO2 were observed. This result is in an agreement with the DSC measurement where the endothermic effect of Sn reduction is observed at 406 °C. The SnO2 phase in samples heated under N2 can be formed by a phase separation from tin silicate during heat treatment.
The solid state NMR spectroscopy was used to study structural sites of calcined xerogels. The 29Si CPMAS NMR spectrum (Fig. 4) of calcined sample SiSnF-500 displays resonances with chemical shifts of −110, −102, and −92 ppm. These signals can be attributed to silicon environments represented by Si(OSi)4, Si(OSi)3(OSn), and Si(OSi)2(OSn)2 species (and possibly also Q3 and Q2 sites).41,42 These resonances represent additional hint pointing to the presence of Si–O–Sn bonds in calcined xerogels.
The coordination environment of tin centers in the air calcined xerogel SiSnP-500 was examined by DRUV-vis spectroscopy (Fig. 5). Absorption maxima observed at 222 nm are attributed to tetrahedrally coordinated tin atoms.17,36,37,41 The shoulder at about 266 nm is indicative of hexacoordinated polymeric Sn species. This spectrum indicates that a part of the tin atoms in the calcined xerogels are tetrahedrally coordinated. These species are not separated as SnO2 after calcination, but form a solid solution of SnO2 in SiO2.
Air calcined tin silicate xerogels exhibit surface areas from 164 to 476 m2 g−1 and total pore volume from 0.28 to 0.58 cm3 g−1 (Table 2). Adsorption/desorption isotherms belong to type IV with H2 hysteresis at higher p/p0 pressures, characteristic of mesoporous solids. The average pore diameters reach up to 9.8 nm. The mesopores can be in this case occupied by SnO2 nanoparticles and this probably accounts for the relatively low pore volumes. On the other hand, their embedding in the pores prevents growth of SnO2 crystals during heat treatment. The pore diameters are in close agreement with the sizes of SnO2 nanoparticles computed by Rietveld refinement33 (Table 1S†). The highest surface area of 476 m2 g−1 with pore volume of 0.58 cm3 g−1 and pore diameter of 6.6 nm (Fig. 15S,†Table 2) was reached with the calcined product of the reaction performed in an autoclave (SiSnPA-500). This could be a result of a lower Sn content, more extensive crosslinking due to higher reaction temperature, and resulting higher homogeneity. Moreover, the different polarity of CH2Cl2, as compared to toluene, can affect the arrangement of the template in the reaction mixture. After heating under N2 (500 °C) the sample SiSnPA-500N exhibits specific surface area of 123 m2 g−1. In this case mesoporosity is retained the average pore diameter is surprisingly higher (9.8 nm) than for the sample calcined in air (SiSnPA-500) (Fig. 16S and 17S†), and also higher than the pore diameters of other N2 heated samples. This difference could be caused by a lower content of Sn, higher condensation degree and resulting stability of this sample during the heating in N2, and/or by the formation of a lower amount of SnO2 as compared to the sample calcined in air. The tin silicate xerogel SiSnF-500N heated under nitrogen at 500 °C exhibits a lower apparent surface area (121 m2 g−1, 0.14 cm3 g−1) than the sample calcined in air (Fig. 6). This decrease is caused by reduction of Sn species and separation of Sn nanoparticles. The pore diameter according to QSDFT is 3.4 nm and it shows that mesoporous character is retained (Fig. 7). In this case, the hysteresis type is H4 which is characteristic for slit-shaped pores and/or plate-like particles.43 The same xerogel SiSnF heated in N2 atmosphere at 400 °C reached a surface area of 223 m2 g−1 (Fig. 6, Table 2). At 400 °C, the xerogel framework does not collapse and Sn species are not reduced to metallic tin nanoparticles. This observation is in a good agreement with the powder XRD pattern where diffractions of metallic tin is not observed at this temperature and the DSC endothermic peak for the metallic Sn formation appears above 406 °C (Fig. 2).
Sample | Template | SSA (BET) [m2 g−1] | V tot [cm3 g−1] | d [nm] |
---|---|---|---|---|
SiSnP-400 | P123 | 189 | 0.35 | 9.7 |
SiSnP-500 | P123 | 167 | 0.28 | 9.1 |
SiSnP-500N | P123 | 58 | 0.06 | 3.8 |
SiSnF-400 | F127 | 164 | 0.32 | 9.8 |
SiSnF-500 | F127 | 191 | 0.39 | 9.8 |
SiSnF-400N | F127 | 223 | 0.39 | 5.8 |
SiSnF-500N | F127 | 129 | 0.14 | 3.4 |
SiSnPA-500 | P123 | 476 | 0.58 | 6.6 |
SiSnPA-500N | P123 | 123 | 0.40 | 9.8 |
Fig. 6 Adsorption/desorption N2 isotherms of SiSnF xerogel calcined in air and heated under N2 at 400 and 500 °C. |
Fig. 7 Pore diameter histograms of SiSnPA and SiSnF xerogels calcined in air (NLDFT) and heated under N2 (QSDFT). |
TEM images illustrate the morphology and structure of air calcined and N2 heated xerogels (at 500 °C). Air calcined xerogels display SnO2 nanoparticles embedded in silica (Fig. 8 left, Fig. 18S†). Composites formed during heating in N2 display relatively uniform Sn metallic particles in silica/carbon matrix (Fig. 8 right). The size of these particles is approximately 70–150 nm.
The presence of Lewis and Brønsted acid sites on the surface of air-calcined xerogel (SiSnP-500) was examined by pyridine adsorption followed by IR spectroscopy. Vibrational bands in the IR spectra (Fig. 19S†) were assigned to Lewis (1454 cm−1), Lewis + Brønsted (1489 cm−1) and Brønsted (1540 cm−1) acidic sites.44 These data indicate that calcined tin-silicate materials possess mixed Lewis and Brønsted surface acidity.
Sample | Sn [wt%] (ICP) | n Sn in cat [mmol] | SSA (BET) [m2 g−1] | d [nm] | t [min] | Conversion [%] | Selectivity [%] | TOFa [mmol mmol−1 h−1] | |
---|---|---|---|---|---|---|---|---|---|
I | II | ||||||||
a Apparent turnover frequency TOF [mmol mmol−1 h−1]. | |||||||||
SiSnP-500 | 52.4 | 0.11 | 167 | 9.1 | 5 | 20 | 97 | 3 | 109 |
60 | 50 | 96 | 4 | 23 | |||||
120 | 63 | 95 | 5 | 14 | |||||
SiSnF-400 | 43.2 | 0.09 | 164 | 9.8 | 60 | 34 | 95 | 5 | 19 |
SiSnF-500 | 43.2 | 0.09 | 191 | 9.8 | 60 | 30 | 96 | 4 | 17 |
Catalyst | n Si/nM | Conversion [%] | Selectivity [%] | TONb [mmolEpo mmolM−1] | |
---|---|---|---|---|---|
I | II | ||||
a Reaction conditions: 5 mmol epoxide, 5 mmol amine, 25 mg catalyst, temperature = 35 °C, reaction time = 0.5 h. b Calculated as the number of moles of styrene oxide converted per mole of the metal center. | |||||
Sn-beta45 | 97.0 | 25.4 | 91.0 | 9.0 | 305.0 |
Meso-Sn-beta45 | 96.8 | 41.0 | 91.1 | 8.9 | 476.0 |
On the other hand, heating in N2 atmosphere resulted in the Sn–silica–carbon nanocomposites. These materials display mesoporous character with a small amount of micropores. The size of Sn particles is in the range of 70–150 nm. Potential application of these nanocomposites could be found in electrode materials.47
From these data we can conclude that non-hydrolytic templated acetamide elimination can be used as an effective one pot synthetic method for the preparation of SnO2–SiO2 and Sn–silica–carbon nanocomposites.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra16556g |
This journal is © The Royal Society of Chemistry 2016 |