Super-acids supported on ordered mesoporous materials (OMMs) or mesoporous molecular sieves (MMSs) exhibit exceptional chemical and physical properties that have suggested a vast number of applications, such as aromatic alkylation, isomerisation, oligomerization, acylation of arenes [1, 9]. Here we investigate such a catalytic system for the etherification of glycerol with isobutene. Because of the support properties such as regular pore structure, high surface area, narrow pore size distribution, adjustable pore size, highly dispersed active sites and uniform distribution of these on silica pore these catalysts are similar to the sulfonic-acid groups incorporated in macroporous ion exchange resins that have been successfully used in such applications [2, 5, 9]. However, the pore sizes of these mesoporous SBA-15 silica materials are usually less than 10 nm, which represents a limit for the separation macromolecules such as proteins and polymers [2- 4]. MCM-41 has thin walls of amorphous silica allowing the pore size to be varied from 2 to 10 nm, and chemical properties can be manipulated [3, 5]. This silica's are the ideal template material, which permit independent control of both composition and channel size [3].
The surfactant is the most important on preparation procedure because it defines the structure and porosity of silica, and the calcination procedure have the same effect. The increase in the chain length of the surfactant were found to be effective in the pore size control in OMMs, including silica with two-dimensional hexagonal arrays of cylindrical pores such as MCM-41 template by alkyl-ammonium surfactants(C16), and SBA-15 silica template by tri-block copolymers surfactants( Pluronic P123), or the use of novel surfactants that form very large micelles [4]. The use of surfactant is one can continuously increase the pore diametre by increasing the amount of the growing agent used.
These silica materials functionalized with organosulfonic acid groups have demonstrated an excellent catalytic behaviour on the etherification of glycerol with isobutene to yield tert-butylated derivates. The use of moderately strong acid centres, such as those located in arene-sulfonic acid modified mesostructured silica, provides improved results both in the glycerol conversion and selectivity towards the desired products [6-9].
2. Experimental
2.1. Materials used for preparation of MCM-41.
The silica synthesis sources used were Cab-O-Sil M5 from Sigma Aldrich, tetra-methyl-ammonium silicate (15-20 wt % silica, Sigma Aldrich). The CnH2n+1(CH3)3NBr (from Carl Roth) were used to form the template with n=16. The surfactant solution were prepared by ion-exchanging the 20 wt % C16 aqueous solution with equal molar exchange capacity of Amberjet 4400(OH) ion exchange resin (Sigma Aldrich) by 24 h batch mixing. The antifoaming agent was Antifoam A Concentrate (Sigma Aldrich), which is a silane polymer alkyl terminated by meth-oxy groups. Acid acetic (Roth) was used for pH adjustment of the synthesis solution.
2.2. Materials used for preparation of SBA-15.
Poly(ethyleneglycol)-block-poly(propyleneglycol)-block-poly-(ethyleneglycol) (EO20PO70EO20, Pluronic P123), 1,3,5-trimethylbenzene (TMB) and tetraethyl ortho-silicate (TEOS), all from Sigma Aldrich, hydrochloric acid(HCl), potassium chloride(KCl), both products from Carl Roth.
2.3. Common materials.
The benzenesulfonic acid (BSA), trifluoromethanesulfonic acid (Triflic acid or TFA) and tetrafluoroethanesulfonic acid (TFESA) was purchase from Sigma Aldrich. The water used in all experiments is de-ionized water.
2.4. Catalysts preparation of MCM-41.
The surfactant solution was first prepared. The powder of cetyl-tri-methyl-ammonium bromide (CTAB, 20.0 g) was dissolved in de-ionized water (80.0 g) to make a 20 wt % solution. These solution mixed by 2 h. Then Amberjet 4400 OH anion-exchange resin was added into the solution to exchange Br ions with OH ions. The ion-exchange process was performed by 24 h under vigorous stirring. The resulting solution was filtered and ready to use on the silica preparation. The pore diametre fine-tuning is acquired through the change in the amount of the surfactant, the change in the duration of the hydrothermal treatment, the calcination flow type or its calcination temperature.
The fumed silica Cab-O-Sil M5 (2.5 g) was added to the tetra-methyl-ammonium silicate aqueous solution (5 g) and de-ionized water (55 g), then the mixture was stirred vigorously for 1 h. Two drops of antifoam agent (0.2 wt % of surfactant) was added. After this, 28.79 g of the surfactant solution was added under stirring.
Then the pH was adjusted to 11.5 by adding acetic acid under agitation. If the pH of the solution is higher than 11.5, then acetic acid was added drop-wise, else maximum 5 drops of surfactant was added to adjust the pH. After 1 h of mixing, the synthesis solution was take it into the polypropylene bottle and placed into an autoclave at 373 K for 6 days.
After the solution was cooled to room temperature, the resulting solid was recovered by filtration (the residual over the bottle wall was washed with de-ionized water), washed in 50 ml de-ionized water, filtered again and dried in an oven at 348 K under ambient air for 24 h. Then the solid was pound very thin, followed by take it in to the glass cell for calcination, with layer thicknesses of the catalyst at 1 cm. The pre-dried solid was heated at a constant rate from room temperature to 813 K over 17 h under helium flow and hold for 1 h under the same conditions, followed by the calcinations at 813 for 5 h with air to remove the residual surfactant, then cooled at a constant rate (4 degrees/min) from 813 K to room temperature under air flow.
2.5. Catalysts preparation of SBA-15
The mesoporous SBA-15 silica particles were prepared using the tri-block copolymer P123 as a structure-directing agent. In a typical synthesis procedure, the P123 (2.0 g) and a given amount of the potassium chloride (KCl, 1.54 g) were dissolved in the de-ionized water (60.0 g) and hydrochloric acid (HCl conc.37%, 11.8 g) at ambient temperature until the solution became transparent. Then the 1, 3, 5 trimethylbenzene (TMB, 1.5 g) was added to the above solution under stirring. After 10 h of stirring, the tetraethyl ortho-silicate (TEOS, 4.3 g) was added drop-wise, and stirred vigorously for 30 minutes.
The molar ratio of mixture was 1TEOS:0.017P123:0.6TMB:1KCl:5.85HCl:165H2O. The obtained mixture was kept at 348 K in an oven for 24 h, then transferred to an autoclave and heated under static condition at 373 K for 24 h. The obtained solid product was filtered (the residual over the bottle wall was washed with de-ionized water), washed in the de-ionized water (50 ml), filtered again and dried at 333 K overnight in an oven.
Removal of organic template was achieved by heated under air flow condition from room temperature to 813 K, with a heating rate of 4 degrees/min, and calcined at 813 K for 10 h under the same conditions, followed by cooled the calcined SBA-15 silica spheres at a constant rate (4 degrees/min) from 813 K to room temperature.
2.6. Impregnation of silica support with super-acids.
Supported catalysts were prepared by pore volume impregnation method using aqueous solutions containing trifluoromethanesulfonic acid (triflic acid or TFA), benzene-sulfonic acid (BSA) and 1,1,2,2-tetrafluoroethanesulfonic acid (TFESA). Those silica spheres MCM-41 and SBA-15 were impregnated with 10 wt % super-acids. Before impregnation, the adsorption capacity of the silica support was checked. Alternatively, the value of the adsorption volume from the adsorption-desorption isotherm calculated with BET method could be used. The adsorption capacity was checked by the difference between the initial amount of water and the amount of water remaining after adsorption.
Adsorption Capacity = V adsorbed water = V initial water-V non-adsorbed water.
The resulting mixture (super-acid and water) was added over silica under continuous stirring for a good and uniform distribution in the silica pores. The impregnated silica was dried overnight at 373 K in an oven.
2.7. Characterization
Characterizing the pore structure of silica materials is important to check its physical and chemical properties. The liquid nitrogen isotherms provide the data to characterize the pore structure of silica. The typical adsorption/desorbtion isotherms of pure silica and impregnated silica follow the type IV isotherm. The surface area can be calculated from the nitrogen physic adsorption/desorbtion isotherms by Brunauer-Emmett-Teller (BET) methods. Pore volume and pore size distributions can be obtained by Barrett-Joyner-Halenda (BJH) methods.
The XRD measurement has an important application in this article. Although pore walls of silica are amorphous, the mesoporous material has long-range order as shown in this article, established by the XRD measurement. The two dimensional hexagonal structure can be characterized by XRD, showing a sharp (100) plane diffraction peak and also higher Miller Index diffractions (110), (200) and (210). Many researchers have used XRD measurements to characterize nanocrystalline materials. The XRD measurements were carried out using a D8 Advance, Bruker X-ray diffract meter (Cu Kα, λCu = 1.5406 Å, 40 kV, 40 mA).
The FT-IR measurements were done by using a instrument Varian 3100 FT-IR Excalibur Series.
The reaction samples were analyzed by gas chromatography (Agilent Tehnologies Network GC system) using a DB- WAX column (30 m *0,32 mm, DF = 0,32 mm) and a FID detector.
3. Results and discussion
Experience has shown that the equilibrium distribution of adsorbate molecules between the surface of the adsorbent and the gas phase is dependent upon pressure, temperature, the nature and area of the adsorbent, and the nature of the adsorbate. An adsorption isotherm shows how the amount adsorbed depends upon the equilibrium pressure of the gas at constant temperature.
The silica MCM-41 was prepared using two types of Cab-O-Sil M5, one from the UPG (ROM) and the other one from Yale University(USA), using the same method of preparation, and the SBA-15 silica was prepared. The liquid nitrogen adsorption/desorption isotherms provide the data to characterize the pore structure of the silica and check the pore size, the pore volume and the isotherm type.
The nitrogen adsorption/desorption isotherms for the SBA-15 silica and for the MCM-41 silica samples prepared with Cab-O-Sil M5 from the Yale University(USA) and with Cab-O-Sil M5 from the UPG (ROM) are presented in figure 1.
These isotherm are comparable and follow the type IV without hysteresis curves, characteristic for the mesostructured silica pores with ordinary structure.
The adsorption isotherms show four different regions. The first zone at low relative pressure p/p0 (gaps 0-0.3 for the silica MCM-41 and 0-0.18 for the silica SBA-15) is specific to highest nitrogen physical adsorption yield, associated with a thin-layer adsorbed over silica surface (at external surface and in mesoporous area). The second zone (gaps 0.3-0.38 for the silica MCM-41 and 0.18-0.2 for the silica SBA-15) describe an increase of pressure corresponding to a multi-layer adsorption given by the contribution of the external surface and mesoporous area.
The third zone (gaps 0.38-0.93 for the silica MCM-41 and 0.2-0.85 for the silica SBA-15) reflects the adsorbed nitrogen yield growing significantly, that suggest a nitrogen cappilar condensation on the mesoporous area. The lowest difference between adsorption and desorption curves for pure silica MCM-41 gives an indication of the pore diametre uniformity, without obstructions. But for pure silica SBA-15 the largest difference between these curves suggest a non-uniformity of the pore diametre.
The fourth zone, at rellative pressure close to 1, suggests an increase of the volume of adsorbed nitrogen associated with nitrogen condensation in the interparticular spaces.
The pore volume and pore size distributions can be obtained by Barrett-Joyner-Halenda (BJH) methods. The pore size distributions for the pure silica SBA-15 and for the pure silica's MCM-41 prepared with Cab-O-Sil M5 from the Yale University (USA) and Cab-O-Sil M5 from the UPG (ROM) are shown in figure 2.
The two prepared silica MCM-41 have approximately the same dimensions of the porosity (27Å for the silica MCM-41 ROM and 28Å for the silica MCM-41 USA), also for the silica SBA-15, the pore size is about 39 Å.
The X-ray diffraction is the best suited for determination of the bulk composition. From the symmetry and the intensities of the patterns we discover whether the material is highly crystalline or non-crystalline (with amorphous structure).
The X-ray diffraction exhibit several important properties. The first property, they signify whether the catalyst, or a component of it, is crystalline or non-crystalline. The second property, they yield an estimate of the size of the micro-crystallites that may be present on the samples. The third property, gives information about the crystalline structure and the unit cell dimensions, insight into the atomic constituents of the unit cell.
The XRD measurements for silica MCM-41 prepared with both type of Cab-O-Sil M5 are shown in figure 3. We observe the silica prepared with Cab-O-Sil M5 from the Yale (USA) present a higher intensity for the peaks then for the silica prepared with Cab-O-Sil M5 from the UPG (ROM), that suggest a very good defining of the structure for the first silica.
The D100, D110 si D200 peaks coresponding 2 θ angle 2.5°, 4.4° and 4.8° confirm the amorphous hexagonal structure of the support silica MCM-41 as honeycomb type.
The impregnation was done using the silica MCM-41 prepared with Cab-O-Sil M5 from the Yale (USA) and the silica SBA-15.
The nitrogen adsorption/desorption isotherms for the impregnated silica SBA-15 samples prepared are shown in figure 4. This three isotherms are symmetrical and folowing the IV type isotherm with a largest hysteresis curve characteristic for the mesostructured silica pore dimension with ordinary structure and with lowest obstruction at the pore level.
The nitrogen adsorption/desorption isotherms for the impregnated silica MCM-41 samples prepared are shown in figure 5. These isotherm are comparable and follow the type IV without hysteresis curves, the lowest difference between adsorption and desorption curves for the impregnated silica MCM-41 gives an indication of the pore diametre uniformity, without obstructions.
The pore size for the impregnated silica SBA-15 samples are shown in figure 6 and for the impregnated silica MCM-41 samples are presented in figure 7.
The pore size for those three impregnated silica SBA-15 samples are the same and equal to 38 Å, on the same time for those three impregnated silica MCM-41 samples are equal to 26 Å, but for pure silica the pore size is biggest. These suggests the impregnation process has partially affected the structure of the pure silica, probably with increasing of walls thickness because of the super-acids grafted on the surface leading to a decrease of the pore size.
The XRD measurements for the impregnated silica MCM-41 samples and for the impregnated silica SBA-15 samples are shown in figure 8.
A small shift of the first diffraction peak to lower 2θ and the lower intensity of impregnated silica compared to the pure silica is indicative of a decrease in the pore diametre values. The flattening of the peaks coresponding to 2θ angle equal to 4.4° and 4.8° confirm some obstruction of the tubular chanels of the pure silica caused by the impregnation process.
The FT-IR spectroscopy for the silica impregnated with super-acids were done to check if the mean chemical bonds existing at the surface of the pure silica support with super-acids. The vibrational spectra of different groups in a molecule give rise to absorptions at characteristic frequencies, because a normal mode of even a very large molecule is often dominated by the motion of a small group of atoms. The intensities of the vibrational bands that can be associated with the motions of small groups are also transferable between molecules.
Consequently, the molecules in a sample can often be identified by examining its infrared spectrum and referring to a table of characteristic frequencies and intensities or to values of main compounds.
The FT-IR spectra for the impregnated silica MCM-41 and for the impregnated silica SBA-15 samples are shown in figure 9.
We see that FT-IR spectra for pure silica SBA-15 doesn't present sygnificants peaks, just a simple vibration characteristic to silica at 950 cm-1. The vibration at 1500 cm-1 observed in sample BSA/SBA-15 is characteristic of the phenyl ring and the vibration at 1760 cm-1 is attribute of the phenyl ring directly bonded to silica. The intense band at 1205-1230 cm-1 is characteristic of the strong bond C-F( tri-fluoro-methyl), whereas the large band at 1023-1075 cm-1 belongs to C-F bonds, also is present the specific vibration of the Si-F bond at 1340-1380 cm-1, those vibrations are present on the FT-IR spectra for TFA/SBA-15 and TFESA/SBA-15 silicas. The bands at 1740 cm-1 found in those three spectra of the impregnated silica samples are typical of the asymmetric stretching of SO2 moieties and confirm the presence of the sulfonic acid species.
The FT-IR spectra for pure silica MCM-41 present sygnificants peaks, whereas a simple vibration characteristic to silica at 1059 cm-1. The vibration at 1400 cm-1 observed in sample BSA/MCM-41 is characteristic of the phenyl ring. The FT-IR spectra for TFA/SBA-15 and TFESA/SBA-15 silicas present intense band at 1205-1230 cm-1 is characteristic of the strong bond C-F( tri-fluoro-methyl), whereas the large band at 1023-1058 cm-1 belongs to C-F bonds, also the presence of the specific vibration of the Si-F bond at 1355-1363 cm-1. The bands at 1735-1740 cm-1 found for those three impregnated silica samples are typical of the asymmetric stretching of SO2 moieties and confirm the presence of the sulfonic acid species.
The glycerol etherification with isobutene in presence of acid catalyst to yield a mixture of mono-, di-, and tri-tert-butyl glycerol ethers (MTBG, DTBG, and TTBG, respectively) was tested with the 10 wt % TFESA/MCM-41 and with the 10 wt % TFESA/SBA-15. This reaction has been usually performed over catalyst surface also inside of the pore catalyst.
The yield on desired etherification products depends by size of the accessible surface area for the reactants, by pore size, by number of acid active site and on the same time by mass transfer. The porosity of the catalyst must be enough to allow the reactants transit to acid active site (internal diffusion), also the products off the active site to catalyst surface (external diffusion), therefore she doesn't affect mass transfer.
These highly surface materials with large interconnected mesopores and high accessibility of strong acid sites might be suitable/eligible catalysts for this etherification reaction.
The 1, 1, 2, 2-tetrafluoroethanesulfonic acid (TFESA) impregnated over mesoporous silica MCM-41 and SBA-15 are not tested on this reaction of glycerol with isobutene until now. The mesoporous silica was impregnated with 10 wt % TFESA and I test it in glycerol etherification with isobutene.
This reaction has been performed in a 600 mL stainless steel Berghoff autoclave equipped with mechanical stirring. The autoclave is electrically heated, with automatic temperature control. For all experiments performed the concentration of the catalyst respect the amount of glycerol loaded in the reactor was 4 wt %, the amount of glycerol is 130 g, running for 5 h at 353 K, at isobutene/glycerol molar ratio of 3/1 and without the pH correction of the glycerol phase. The concentration of the emulsifier in the both reaction mixture was of 0.1 wt %.
The ammonium quaternary salt (N-Benzyl-N, N di-methyl-N-[4-(1, 1, 3, 3 tetra-methyl-butyl) phenoxy-ethoxy-ethyl]-ammonium chloride or C27H42ClNO2 was added to the reaction mixture as emulsifier. The stirring rate was enhance and maintained at 1200 rot/min and the pressure in the reactor was monitored continuously during the whole duration of each experiment.
The analyses of the reaction products were performed by gas-chromatography, using an instrument from Agilent Technologies with FID detector, equipped with DB-WAX polar column of 30 m length and 0.32 mm inner diameter. The chromatographic column was operated between 40-220°C, with nitrogen as carrier gas.
Figure 10 show the glycerol and the isobutene conversions, the yields to ethers of the glycerol and the selectivity to ethers of the isobutene obtained in the presence of the emulsifier over the TFESA/MCM-41 and the TFESA/SBA-15 as catalysts.
The glycerol conversions are about the same for the two cases, but higher with 1 % when the reaction is performed in presence of the TFESA/MCM-41 as catalyst, around 95 %, and the isobutene conversion are the same for the using both catalysts, about 100 %, because both catalysts present almost the same porosity and the pore structure.
The yields in mono- and di-ethers are 3 % higher for each product, but the yield in tri-ether is roughly 4 % smaller when the reaction is performed in presence of the TFESA/MCM-41 catalyst then the case were is using the TFESA/SBA-15 catalyst (around18,6 % for the MTBG, 55,5 % for the DTBG and 20 % for the TTBG).
A posible explanation for these results is the difference between the pore size for both silica, that suggest the silica TFESA/SBA-15 with large interconnected mesopores and high accessibility who permit a rapid mass transfer of the isobutene to the glycerol phase and making the isobutene molecules ready to completing the glycerol conversion to tri-ether.
We observe the same distribution of the selectivity for the isobutene to ethers such we see in the case of the yields to ethers, a 2 % highly selectivity to the mono- and di-ethers and a smaller selectivity, roughly 2 % ,to the tri-ether when is using TFESA/MCM-41 silica then the selectivity for each products when the reaction is performed in presence of the TFESA/SBA-15 silica (about 9 % for the MTBG, 27,5 % for the DTBG and 10 % for the TTBG), because of his smaller pore size witch not permit a totaly conversion to tri-ether, a steric problem.
However, the two reagents are very different in nature. The glycerol is the continuous phase, highly polar and hydrophilic product. But the isobutene is the dispersed phase, nonpolar and hydrophobic compound. So they show low reciprocal solubility and form two liquid phases while in contact.
Hence, the performance of the etherification process has been proven to be strongly dependent on the mass transfer between the two liquid phases.
These oxygenated compounds or tert-butylated derivates, when incorporated into the standard diesel fuel, help to decreasing the amount of the emissions in particles, hydrocarbons, carbon monoxide and unregulated aldehydes, reducing its viscosity, improving the flash point.
4. Conclusions
Typical adsorption/desorbtion isotherms of the pure silica and the impregnated silica follow the IV type isotherm and characteristic for the mesostructured silica pore dimension with ordinary structure. Both silica's are shown, almost, the same uniformity of the pore size distribution.
The pore diametre fine-tuning is acquired through the change in the amount of the surfactant, the change in the duration of the hydrothermal treatment, the calcination flow type or its calcination temperature.
The pore size for the pure silica is bigger then for the impregnated silica and these suggests, the impregnation process has partially affected the structure of the pure silica, probably with increasing of walls thickness because of the super-acids grafted on the surface leading to a decrease of the pore size.
The MCM-41 silica prepared with Cab-O-Sil M5 from the Yale (USA) present a higher intensity for the XRD peaks then silica prepared with Cab-O-Sil M5 from the UPG (ROM), and suggest a very good shaping of structure for the first silica.
The FT-IR spectra show the characteristic vibrations of the phenil ring, of the strong bond C-F and of the asymmetric stretching of the SO2 moieties, who confirm the presence of the acid/sulfonic acid species.
The catalyst TFESA/SBA-15 silica with large interconnected mesopores and high accessibility permit a rapid mass transfer of the isobutene to the glycerol phase and making the isobutene molecules ready to completing the glycerol conversion to tri-ether.
The performance of the etherification process has been proven to be strongly dependent on the mass transfer between the two liquid phases.
