TABLE OF CONTENTS Thiol group functionalization of mesoporous SiO2 SBA-15 using supercritical CO2 M.J. Tenorio, J. Morère, C. Carnerero, M.J. Torralvo, C. Pando and A. Cabañas* Highlights  Thiol functionalization of SBA-15 was performed in scCO2 at 40-150ºC and 15-29 MPa.  Grafting density increased with temperature, CO2 density, time and stirring.  SBET and pore size decreased as grafting density increased.  A compact monolayer was formed at temperatures of 80 ºC or higher.  The sample prepared in scCO2 at 80 ºC, 25.0 MPa had high grafting density and SBET. Tenorio et al. 1 THIOL GROUP FUNCTIONALIZATION OF MESOPOROUS SiO2 SBA-15 USING SUPERCRITICAL CO2 M.J. Tenorio † , J. Morère † , C. Carnerero † , M.J. Torralvo ‡ , C. Pando † and A. Cabañas †, * † Dep. Química-Física I, ‡ Dep. Química Inorgánica I, Universidad Complutense de Madrid, 28040 Madrid. SPAIN *Send correspondence to: Prof. Albertina Cabañas Departamento de Química-Física I Ciudad Universitaria s/n, 28040 Madrid SPAIN Tlf: 34 + 91 3945225 Fax: 34 + 91 3944135 e-mail: a.cabanas@quim.ucm.es mailto:Martyn.Poliakoff@nottingham.ac.uk Tenorio et al. 2 ABSTRACT Chemical modification of mesoporous SiO2 SBA-15 with thiol groups was performed using mercaptopropyltrimethoxysilane (MPTMS) dissolved in supercritical CO2 (scCO2). Thiol groups serve as adsorbents for the selective removal of contaminant metal cations and in catalysis. Functionalization was carried out in scCO2 at temperatures ranging from 40 to 150 ºC and pressures from 15.0 to 29.0 MPa. For comparison purposes, the reaction was also performed in toluene at 80 and 110 ºC. As opposed to toluene, scCO2 is considered a green solvent. Grafting of the thiol groups was confirmed by FTIR spectroscopy, thermogravimetric analysis (TGA) and elemental analysis. Grafting density and surface coverage of the materials modified using scCO2 increased with temperature, CO2 density, time and stirring and varied from 1.3 to 4.4 mmol g -1 and from 1.3 to 4.0 molecules nm -2 , respectively. On the other hand, surface area and pore size decreased as grafting density increased. At temperatures of 80 ºC or higher, the pore size remained constant, suggesting the formation of a compact monolayer. Modification at higher temperatures led to larger grafting densities but very low surface areas. Assuming total hydrolysis and condensation of the precursor, the optimum grafting density and surface coverage of 2.3 mmol g -1 and 2.4 molecules nm -2 , respectively, were obtained in scCO2 at 80 ºC and 25.0 MPa for 4 h. Grafting densities of the samples prepared in toluene were by far much lower than those obtained using scCO2 at lower temperatures and shorter times, which demonstrates the advantages of CO2 as a green functionalization medium. Keywords: Surface modification, Supercritical fluids, Sustainable Chemistry, Mesoporous, SiO2 SBA-15. Tenorio et al. 3 1. INTRODUCTION Chemical modification of metal oxide surfaces is performed to improve the physical and chemical properties of the materials such as roughness, hydrophilicity, surface energy, biocompatibility, reactivity… It is currently used to prepare stationary phases in chromatography,[1] biocompatible surfaces,[2] environmental sorbents,[3, 4] heterogeneous catalysts[5] and sensors[6] among others. Stabilization of metal nanoparticles is also performed by functionalization with organic ligands.[7] Furthermore, by proper control of the surface properties, adhesion at interfaces can be improved.[8] In particular, the surface modification of mesoporous silica materials with thiol groups has been applied to the preparation of materials for the selective removal of cations such as Hg 2+ , Cd 2+ , Pb 2+ , Cu 2+ and Cr 3+ .[9, 10] Thiol groups may also act as binding sites for catalytic complexes, allowing the incorporation of many catalytic species.[11] Through oxidation of the thiol groups to sulfonic acid groups these materials may become catalysts themselves.[12, 13] Thiols groups can be introduced by co-condensation during the synthesis[14] or by post- synthetic grafting.[15] Co-condensation of ordered mesoporous silica materials is challenging because it generally yields less ordered structures, and some of the functional groups become buried into the structure. On the other hand, the surface modification by post-synthesis covalent attachment is constrained by the number of reactive surface silanol groups, although, in principle, all the functional groups generated are accessible. Furthermore, the original support structure is preserved, although the pore size decreases. Tenorio et al. 4 Surface modification of SiO2 and other metal oxide surfaces by post-synthetic grafting has been previously carried out using supercritical CO2 (scCO2). [16, 17] scCO2 is considered a green solvent because it is nontoxic, non-flammable, has a moderate critical temperature and pressure (Tc = 31.1 ºC, Pc = 7.38 MPa) [18]. Furthermore CO2 is a gas at atmospheric pressure and does not leave any residue. Most of the CO2 used nowadays is obtained as a sub-product of many industrial processes. The large solubility of many silane precursors (alkoxy and chloro silanes) in scCO2 at moderate pressure and temperature conditions and the low viscosity, surface tension and enhanced diffusivity of the supercritical solutions, allow the efficient penetration of highly porous materials. The reaction generally involves the hydrolysis of the precursor and its condensation with the -OH groups from the SiO2 surface or other hydrolyzed molecule precursors. Tripp and Combes studied the interaction of scCO2 with fumed silica using infrared spectroscopy.[19] They pointed out the drying character of scCO2 that can remove adsorbed water from the surface increasing the amount of accessible silanol groups on the surface. Thus the surface reaction of the precursor with the metal oxide surface is favored at these conditions. Our group has recently reported the sustainable functionalization of mesoporous SiO2 SBA-15 with amine groups in scCO2. [20, 21] In comparison to the conventional functionalization process using toxic organic solvents, modification using scCO2 was performed at lower temperature and shorter times and led to larger grafting densities. The materials were used as selective CO2 sorbents in carbon capture technology. Similarly, Lopez-Aranguren et al. [22] have applied the supercritical functionalization process to the Tenorio et al. 5 modification of the internal surface of silica porous substrates of diverse pore architecture showing that functionalization in scCO2 preserved the porous character of the materials. In this communication we report the surface modification of mesoporous silica SBA-15 with thiol groups in scCO2. There are a few reports on the incorporation of thiol groups by reaction of mercaptopropyltrimethoxysilane (MPTMS) in scCO2 on different SiO2 supports. Shin et al. [23] reported the surface modification of microporous zeolites at 150 ºC and 52 MPa. The same pressure and temperature conditions were chosen by Zemanian et al. [24] to modify mesoporous SiO2 MCM-41. Rebiscoul et al. [25] modified the surface of oxidized Si wafers at much more moderate pressure and temperature conditions (40 ºC and 10.9 MPa). A small amount of cosolvent (acetone or N-methyl-2-pyrrolidone, NMP) was used to improve precursor solubility. Finally Ashu-Arrah et al. [26] recently modified different porous SiO2 commercial supports at 40-100 ºC and 41.4 MPa. The wide range of experimental conditions used in the different publications from 40 to 150 ºC and 10.9 and 52 MPa and the high pressures used in most of the reports, points out the need to perform a more systematic study of the role of the different variables on the process. We have previously studied the phase diagram of MPTMS and CO2.[27] Full miscibility of both components was achieved at 40, 50 and 60 ºC above 8.7, 10.2 and 12.3 MPa, respectively. Therefore, the high pressures employed in some of the surface modification experiments are not justified based on solubility requirements. Tenorio et al. 6 2. MATERIALS AND METHODS 2.1 MATERIALS Tetraethylorthosilicate (TEOS, 99+% pure), poly(ethyleneglycol)–block– poly(propylene glycol)–block–poly(ethylenegly-col) (Mw = 5800) (PEO–PPO–PEO), (mercaptopropyltrimethoxysilane (MPTMS, purity>95%) and toluene (>99.5% pure) were obtained from Sigma–Aldrich and used as received. CO2 (purity > 99.99%) was supplied by Air Liquide. Mesoporous silica SBA-15 was prepared following a procedure similar to that described by Zhao et al.[28, 29] Two different samples with BET surface areas of 592 and 674 m 2 g -1 were obtained. 2.2. SURFACE MODIFICATION OF SILICA SBA-15 IN SUPERCRITICAL CO2 Most experiments were conducted in a ca. 65 cm 3 custom-made stainless steel high- pressure reactor in the batch mode. A few experiments were performed in a 100 cm 3 stirred high-pressure Bolted Closure reactor (Autoclave Eng.). In the smaller reactor, the liquid MPTMS was place in a glass vial and the mesoporous silica SBA-15 was wrapped in a filter paper. In a typical experiment, ca. 100 mg of SiO2 SBA-15 was treated with 0.5 or 1 g of MPTMS. A large molar excess was stablished considering a maximum silanol content of 3.7 OH nm −2 [30]. Experiments were performed at 40, 60, 80 and 100 ºC and 15.0 and 25.0 MPa. At 40-80 ºC, the small reactor containing the support and reagent was immersed in a thermostatic bath (PolyScience) and filled with CO2 by means of a high-pressure syringe pump (Isco, Inc. Model 260D) at the desired temperature and pressure. At 100 ºC, a heating aluminum jacket connected to a PDI temperature controller (Micromega, model CN77322) was used instead. No stirring was Tenorio et al. 7 used in these experiments. Temperature was measured using a type J calibrated thermocouple. The pressure was measured using a pressure transducer (Gems). A safety valve (Swagelok) was fitted to the reactor. The reaction proceeded for 1-8 hour depending on the experiments. A few samples were prepared using the larger stirred high-pressure reactor at 80, 100 and 150 ºC. The reactor is equipped with a heating jacket connected to PDI controller to reach the desired temperature. The support was loaded into a basket fitted to the upper part of the reactor, whilst MPTMS was placed directly at the bottom of the reactor. The reactor was then filled with CO2 using the high-pressure syringe pump at the same temperature up to the given pressure. The reactor was provided with an impeller stirrer which was agitated at 200 rpm. The temperature was measured using a K-type thermocouple. The pressure was measured using a pressure gauge. A rupture disk was fitted to the reactor. Reaction time for these experiments was fixed to 4 hours. After the stablished reaction time, the reactor was extracted with scCO2 at the reaction pressure and temperature at a flow rate of 1–2 mL min −1 during 60 min and was slowly depressurized through a needle valve in about 30 min. Samples were dried at 110 ºC under vacuum for 4 h. 2.4. SURFACE MODIFICATION OF SILICA SBA-15 IN TOLUENE For comparison purposes, the SiO2 SBA-15 was functionalized in toluene at 80 ºC and 110 ºC. Approximately 200 mg of silica SBA-15 and 1 or 2 g of MPTMS were loaded into a two-neck round-bottom glass flask containing 50 cm 3 of toluene fitted to a condenser. The glass flask was then immersed in a silicon oil bath set at the working temperature Tenorio et al. 8 using a temperature-controlled stirring hot plate with an external temperature probe and was purged with N2. The reaction was allowed to proceed for 24 h with gentle stirring. After reaction, the thiol-modified SiO2-SBA-15 was filtered and washed with copious amounts of toluene and then ethanol. The recovered solid was dried at 100 ºC in a natural convection oven for 24 h. 2.4. MATERIALS CHARACTERIZATION Materials were characterized using Fourier Transform Infrared (FTIR) spectroscopy, thermogravimetric analysis (TGA), elemental analysis and N2-adsorption. FTIR spectra of the powdered materials were recorded as KBr pellets on an IR Perkin Elmer Spectrometer. At a spectral resolution of 4.0 cm −1 , 25 interferograms have been co-added. TGA of the samples was obtained on a SDT – Q600 at a heating rate of 10 ºC/min in N2 flow (100 ml/min). Elemental analysis was performed on a LECO CHNS-932 microanalyser. The uncertainties in the elemental analysis were %C ± 0.26, %H ± 0.24 %N ± 0.25 and %S 0.35. Low angle XRD patterns of the mesoporous materials were collected on a XPERT- Alpha1 diffractometer with Cu K- radiation on the transmission mode and were recorded at 2values between 0.1 and 3º. N2 adsorption/desorption isotherms at 77 K were obtained using a Micromeritics ASAP- 2020. Prior to adsorption measurements, samples were out-gassed at 110 ºC and ~10 -1 Pa for 4 h. Isotherms were analyzed using standard procedures.[31] The BET equation was used for specific surface calculations. The total pore volume was estimated from the amount adsorbed at a relative pressure of 0.995. The pore size distributions were calculated using the Barrett, Joyner and Halenda (BJH) method for a cylindrical pore model corrected Tenorio et al. 9 by the statistical thickness using the adsorption and desorption branches of the isotherms. The micropore volume was determined using the t-plot method. The actual pore size was estimated from the adsorption branch of the isotherm as was reported by Zhao et al. [29] 3. RESULTS AND DISCUSSION Silica SBA-15 was chosen as the support to be modified because of its high surface area and pore volume. The support is formed by a hexagonal array of cylindrical mesopores which are interconnected by smaller micro and mesopores. The low-angle XRD pattern of the substrate is given in the Supplementary data. Three well resolved peaks were indexed according to the two dimensional P6m hexagonal symmetry. Functionalization of mesoporous SiO2 SBA-15 with MPTMS was carried out in scCO2 at 40, 60, 80 and 100 ºC and 150 ºC and 15.0, 25.0 and 29.0 MPa. Reaction time was varied from 1 to 8 h and 5:1 and 10:1 precursor to support mass ratios (large precursor excess) were used. At 60 ºC, MPTMS and CO2 are fully miscible above 12.3 MPa. [27] For comparison purposes, samples were also modified following the conventional method in toluene at 80 and 110 ºC. The experiments performed are summarized in Table 1. Figure 1 depicts the possible reaction mechanism involving the hydrolysis of the precursor (i) and its condensation with the -OH groups from the SiO2 surface or other neighboring molecule precursors (ii). Although in most experiments, water was not deliberately introduced into the system, the presence of adsorbed water on the silica surface, which is extracted in scCO2, seems to be enough to initiate the reaction. Considering the surface reaction with the silanol groups of the surface a large number of possible structures can be envisioned, including partially hydrolyzed and condensed Tenorio et al. 10 structures (A, B and D), those fully condensed and isolated (D) and those fully condensed forming a monolayer (E), among others. The polymerization reaction in the fluid phase is also possible, leading eventually to an oligomer that can be adsorbed at the sample surface forming a thick polymer condensed layer (not shown here). Figure 1- Possible mechanism for the surface reaction between MPTMS and SiO2 SBA-15. Figure 2 shows the FTIR spectra of sample 6 functionalized in scCO2 at 80 ºC and 25.0 MPa for 4 hours and sample 13 modified in toluene at 110 ºC for 24 hours. FTIR spectra show the characteristic C-H stretching vibrations at 2962, 2937, 2892 and 2852 cm -1 indicative of both (–CH2–) and (–CH3) groups, which clearly indicate the presence of the organic groups attached to the silica surface, including unreacted precursor or partially hydrolyzed forms on the support surface. The absorption band at 1460 cm -1 characteristic Tenorio et al. 11 of the bending mode of (–CH2–) is also observed. The S-H stretching vibration at ca. 2580 cm -1 was very weak although visible in the sample functionalized in CO2. The broad band centered at ca. 3450 cm -1 was assigned to the stretching vibration of surface silanols groups and the remaining adsorbed water. The bending mode of H-O-H was also observed at 1630 cm -1 . Absorption bands at 1220, 1080 cm -1 and 808 and 457 cm -1 are due to the asymmetric and symmetric Si–O stretching vibrations and Si–O–Si bending vibrations, respectively. [32-35] Similar FTIR spectra were obtained for the different samples. Figure 2- FTIR spectra of sample 6 functionalized in scCO2 at 80 ºC and 25.0 MPa for 4 hours (—) and sample 13 functionalized in toluene at 110 ºC for 24 hours (·····). Tenorio et al. 12 Figure 3 TGA of sample 6 functionalized in scCO2 at 80 ºC and 25.0 MPa for 4 hours (—) and sample 13 functionalized in toluene at 110 ºC for 24 hours. Dash lines show the first derivative plotted as a function of temperature. TGA analysis of sample 6 modified in scCO2 at 80 ºC and 25.0 MPa and sample 13 modified in toluene at 110 ºC are given in Figure 3. Analysis revealed relatively large weight losses associated to the decomposition of the precursor. In both samples, several weight losses were observed at the following temperature intervals: 25-175, 175-310, 310- 425 and 425-700 ºC. These intervals appeared in all the samples functionalized in scCO2 and toluene. The weight loss below 175 ºC was ascribed to the loss of surface water and alcohol formed in the hydrolysis of the precursor. The weight loss in the 175-310 ºC range could be ascribed to the presence of unreacted precursor physically adsorbed on the SiO2 surface. The weight loss in this interval was close to 3 or 4% for most samples prepared in scCO2 and between 1 and 2% for the samples prepared in toluene. Weight losses above Tenorio et al. 13 310 ºC have been attributed to the thermal decomposition of the mercaptopropyl organic groups attached to the surface. The higher decomposition temperatures suggest covalent attachment of these groups to the surface. Because decomposition or desorption of adsorbed MPTMS and derived physisorbed species and decomposition of covalent attached silanes are difficult to separate and following previous reports[26], the total amount of thiol groups on the surface was established from the mass percent loss between 175 and 700 ºC. Grafting densities (𝜑𝑚) in (mmol MPTMS/g SiO2) were estimated assuming total hydrolysis of the precursor as 𝜑𝑚 = % mass loss (175 − 700 °C)/ 75 100 − % mass loss (25 − 700 °C) where 75 g mol -1 is the molar mass of the organic part -(CH2)3SH. This value is an approximation, due to the presence of physisorbed species and the large number of possible partially hydrolyzed structures (Figure 1). Surface coverage (𝜑𝑆) in (molecules nm -2 ) was calculated for 𝜑𝑚 and the BET surface area of the support, 𝑆BET, following 𝜑𝑆 = 𝜑𝑚 N𝐴 𝑆BET Elemental analysis for the different samples is also provided in Table 1. C to S mass ratios ranged from 1.3 to 2.3 in all the samples. If complete hydrolysis of the precursor took place (forms D and F in Figure 1) a 3:1 C to S molar ratio would be expected (1.13 C to S mass ratio). On the other hand, if physical adsorption with no hydrolysis was taking place, a 6:1 C to S molar ratio (2.25 C:S mass ratio) would be observed. The fact that Tenorio et al. 14 values in most cases are intermediate between the former situations suggests that hydrolysis is not complete. Results for the different experiments performed in scCO2 and toluene are given in Table 1. At 60 ºC and 15.0 MPa, reaction time was varied from 1 to 4 h. C:S mass ratios were at these conditions very high, indicating incomplete hydrolysis of the reactants. Grafting density increased as reaction time increased from 1 to 2 hours, but it decreased slightly for 4 hours. However, considering that the lowest C:S mass values were obtained at 4 h, in the following experiments the reaction time was fixed to 4 h. At 60ºC, the pressure was then increased from 15.0 to 25.0 MPa and CO2 density increased up to 0.787 g cm -3 . This pressure rise did not lead to substantial variations in the grafting density and C:S mass ratio. Therefore, the pressure effect is not significant at this temperature. At 25.0 MPa, temperature was also decreased at 40 ºC (CO2 density 0.879 g cm -3 ) leading to similar grafting densities. Although temperature in this case was much lower, grafting density remained almost constant showing how the higher pressure and higher CO2 density have a positive effect on the surface modification. Next, temperature was increased up to 80 ºC, keeping the pressure constant at 25.0 MPa for 4 h, leading to a grafting density close to 2.3 mmol g -1 , whilst C:S mass ratio increased slightly to 1.6. At a longer time, 8 hours, grafting density increased to 3.1 mmol g -1 and C:S mass ratio decreased to 1.3. A further experiment at 100 ºC and 25.0 MPa was also performed. At these conditions, grafting densities and C:S mass ratios were similar to those obtained at 80 ºC, which could be related to the lower density of the CO2 mixture. Tenorio et al. 15 A few experiments were performed in the stirred reactor (samples 9 – 11). Grafting densities in this case were always larger than those corresponding to the samples obtained in the small reactor without stirring at the same experimental conditions. For example at 80ºC and 25.0 MPa, the grafting density at 4 hours reached a value of 3.5 mmol g -1 that was even larger than the value obtained at the same pressure and temperature conditions without stirring after 8 hours. Similarly, comparison of samples 8 and 10 at 100 ºC and 25 MPa showed that a larger grafting density was obtained with stirring. A slight decrease of the grafting density was observed at 25.0 MPa when temperature increased from 80 to 100 ºC. The maximum grafting density of 4.4 mmol g -1 was obtained at 150 ºC and 29.0 MPa. C:S mass ratios for samples 9 to 11 were 1.1, 1.7 and 1.4 at 80, 100 and 150 ºC, respectively. The C:S mass value for the sample prepared at 80 ºC suggested complete hydrolysis of MPTMS. The experiments performed in scCO2 show that grafting density increases with temperature, pressure, CO2 density, time and stirring. The fact that grafting density increases with temperature clearly shows that the surface reaction is an activated process. Although solubility of MPTMS in scCO2 is very high, pressure and CO2 density also have a positive effect on the process probably helping in the dissolution process. The diffusion of the reactant to the support is also favored with increasing time and stirring. For the experiments performed using the conventional method in toluene for 24 h, grafting density also increased with temperature from 80 to 110 ºC, whilst the C:S mass ratio decreased. However, even at 110 ºC, the grafting density was by far much lower than those obtained by functionalization in scCO2 at much lower temperatures. Furthermore, Tenorio et al. 16 from a green chemistry viewpoint, reaction in scCO2 is much cleaner than that in toluene. These results demonstrate the advantages of using supercritical CO2 as a functionalization medium. Considering the maximum concentration of silanol groups[30], surface coverages between 3.7 and 1.2 molecules nm -2 were expected. These values correspond to the completely and partially hydrolyzed and condensed forms of MPTMS on SiO2 (structures E and A in Figure 1, respectively). Feng et al. [34] have reported slightly larger surface coverages of 5.0 molecules nm -2 for a mesoporous SiO2 material. The values calculated in this work are between 1.3 and 4.0 molecules nm -2 in agreement with the previous estimates. N2-adsorption-desorption isotherms were measured on the different samples. Figure 4 compares the adsorption-desorption isotherms of the bare support and the samples modified in scCO2 at 40, 80 and 100 ºC and 25.0 MPa. All the isotherms exhibited a type IV with H1 hysteresis loop characteristic of highly ordered mesoporous materials with a well-defined cylindrical pore morphology. Upon functionalization isotherms shifted downwards, due to the decrease in surface area. Values for the BET surface area (SBET) and total pore volume (VP) of the different samples are given in Table 2. Values for the surface area of the bare SiO2 support samples were 592 and 674 m 2 g -1 . The highest surface area among the functionalized samples was measured for the sample prepared in toluene at 110 ºC, due to its lower grafting density (394 m 2 g -1 ). SBET for samples modified in scCO2 at 40, 80 and 100 ºC and 25.0 MPa for 4 hours without stirring (samples 5,6 and 8) was much lower adopting values equal to 296, 245 and 221 m 2 g -1 , Tenorio et al. 17 respectively. Surface area also decreased as the reaction time increased (sample 7). The larger the grafing density, the lower the surface area. SBET for samples 9-11 modified using the stirred reactor were much smaller in agreement to their higher grafting densities. In particular sample 11, with the highest grafting density had a surface area of only 11 m 2 /g, which limits the applications of this material. Figure 4- N2 adsorption-desorption isotherms of bare SiO2 SBA-15, samples functionalized in scCO2 at different temperatures and reaction times (samples 5 to 8) and sample 13, functionalized in toluene at 110 ºC. Similar variations with grafting density were observed for the total pore volume. VP decreased from 0.73 cm 3 g -1 for the SiO2 SBA-15 support to 0.230 cm 3 g -1 for sample 7 Tenorio et al. 18 and 0.034 cm 3 g -1 for sample 11. The mesoporous support also presented a small micropore volume of 0.11-0.12 cm 3 g -1 , which disappeared completely with grafting, indicating filling of the micropores. Figure 4 shows how capillary condensation shifted gradually towards lower P/Po values due to the lower mesopore size and the disappearance of the micropores.[36, 37] Figure 5- Pore size distributions obtained from the adsorption (left) and desorption (right) branches of the isotherms of bare SiO2 SBA-15, samples functionalized in scCO2 at different temperature, precursor concentration and reaction time (samples 5 to 8) and sample 13 functionalized in toluene at 110 ºC. The pore size distributions of the adsorption and desorption branches of the isotherms were studied using the BJH method for a cylindrical pore model and are shown in Figure 5. For the bare SiO2 SBA-15 support, pore sizes of 6.5 nm and 7.0 nm were obtained from the adsorption branch of the isotherm for the two samples used. Pore sizes of the functionalized samples were always lower than those of their bare supports and also peaks heights decreased substantially, which indicated modification of the internal surface of the Tenorio et al. 19 cylindrical mesopores. The decrease in the pore size was lower for the sample modified in toluene at 110 ºC. Surprisingly, the pore size in the samples functionalized in scCO2 decreased as the grafting density increased up to 5.4 nm (adsorption) at 80 ºC and 25.0 MPa. A further increase in the grafting density did not lead to a lower pore size, but to a decrease of the pore volume. These results suggests the formation of a compact monolayer in the SiO2 surface at temperatures over 80 ºC, as it has been previously reported by other authors.[25] For a theoretical length of the rest O–Si-(CH2)3-SH equal to ca. 0.61 nm (estimated from the bond lengths and angles optimized using Hyperchem™ 7.1 software and the semiempirical method AM1) the formation of a compact monolayer would decrease the mesopore size by twice this amount, rendering a pore size of ca. 5.3 nm, which is just slightly lower than the pore size obtained. Upon increasing the temperature from 80 to 100 ºC or the time, the surface area and the pore volume decreased, suggesting partial clogging of some of the pores which are not accessible to the N2 molecules. Therefore, the best materials are those exhibiting large grafting densities and high surface areas with a pore size corresponding to the compact monolayer. Considering all the properties simultaneously, sample 6 reacted in scCO2 at 80ºC and 25 MPa for 4 hours without stirring exhibited the best properties: high surface coverage and the highest surface area. The low-angle XRD pattern of sample 6 was compared to that of the bare support (Supplementary data). Peak position did not change significantly after functionalization Tenorio et al. 20 and demonstrates that the support structure is not altered with the surface modification in scCO2. There are previous reports on the surface functionalization of silica base materials with MPTMS using supercritical CO2. Shin et al. [23] reported grating densities of 0.87 mmol g -1 zeolite for commercial zeolite beta of SBET similar to SiO2 SBA-15 although exhibiting microporous structure. Functionalization was performed in scCO2 at 150 ºC and 52 MPa for 12 hours. The much lower grafting density reported for this material could be related to the microporous structure of the support. On the other hand, Zemanian et al. [24] reported at the same temperature and pressure but much shorter time (0.5 hours) a maximum surface coverage of 6.44 molecules nm -2 on hydrated SiO2 MCM-41. Although this value is larger than the maximum value of 4.0 molecules nm -2 obtained in this work at 150 ºC and 29.0 MPa using the stirred reactor, the gas adsorption properties of the material after modification were not reported. Ashu-Arrah et al. [26] carried out the modification of silica stationary phases in scCO2. A maximum surface coverage equal to 3.04 mol m -2 (1.83 molecules nm -2 ) was obtained at 70 ºC and 41.4 MPa for 1 h with stirring. This value is much lower than the measured in our work at a similar temperature. They also reported a decrease in the surface coverage at temperatures above 70 ºC and times longer than one hour. The sometimes unclear procedure employed in the determination of the surface coverage in the different publications and the lack of the material structural characterization after modification make difficult the direct comparison. However, it seems clear that lower temperature and pressures conditions than those previously employed can be applied to the process. All these reports demonstrate that silylation in scCO2 is a clean, fast and efficient Tenorio et al. 21 route for SiO2 surface functionalization. Our study also demonstrates how microstructural parameters should be taken into account to evaluate the success of the functionalization procedure. Tenorio et al. 22 Table 1 Summary of the surface functionalization experiments in scCO2 and toluene at several temperatures, pressures and reaction times showing the weight loss at the different temperature intervals, grafting density (𝜑𝑚), surface coverage (𝜑𝑠), C:S mass ratio from microanalysis results and density of scCO2. [38] ǂ Experiments performed in the stirred reactor. Sample # T (ºC) P (MPa) time (h) Weight loss (%) 𝜑𝑚 (mmol g -1 ) 𝜑𝑠 (molecules nm -2 ) C:S mass ratio  CO2 (g·cm -3 ) 25- 175ºC 175- 310ºC 310- 435 ºC 425- 700 ºC 1 60 15.0 1 1.7 2.9 2.4 3.2 1.3 1.3 2.0 0.604 2 60 15.0 2 18 3.1 3.6 4 2.0 2.0 2.2 0.604 3 60 15.0 4 3.3 3.1 3.9 3.8 1.7 1.7 1.7 0.604 4 60 25.0 4 8 2.6 3.1 4.1 1.6 1.6 1.6 0.787 5 40 25.0 4 1.9 2.9 3.7 4.4 1.7 1.7 1.4 0.879 6 80 25.0 4 0.9 3.8 5 5.9 2.3 2.4 1.6 0.686 7 80 25.0 8 0.5 4.8 6.7 7.4 3.1 2.8 1.3 0.686 8 100 25.0 4 1 2.9 5.5 5.9 2.3 2.3 1.2 0.588 9ǂ 80 25.0 4 0.6 5.2 7.8 7.8 3.5 3.2 1.1 0.686 10ǂ 100 25.0 4 0.6 4.6 7.5 6 3.0 2.7 1.7 0.588 11ǂ 150 29.0 4 0.4 1.8 13.1 10 4.4 4.0 1.4 0.478 12 80 Toluene 24 1 1.1 1.4 1.5 0.6 0.6 2.3 - 13 110 Toluene 24 1 1.7 3.5 4.3 1.4 1.4 1.6 - Tenorio et al. 23 Table 2- N2-adsorption desorption isotherm data for the SiO2 SBA-15 support and the modified samples showing the BET surface area (SBET), the total pore volume (VP) and the pore size. Sample # SBET(m 2 /g) VP (cm 3 /g) Pore size (nm) Support Sample Support Sample Support Sample 4 592 256 0.735 0.398 6.5 5.5 5 592 296 0.735 0.468 6.5 5.9 6 592 245 0.735 0.393 6.5 5.4 7 674 137 0.733 0.230 7.0 5.4 8 592 221 0.735 0.365 6.5 5.4 9 674 85 0.733 0.159 7.0 5.7 10 674 216 0.733 0.361 7.0 5.7 11 674 11 0.733 0.034 7.0 – 13 592 394 0.735 0.365 6.5 5.9 4. CONCLUSIONS Functionalization of SiO2 SBA-15 with MPTMS has been successfully carried out in scCO2 at 40-150 ºC and 25.0-29.0 MPa for 1-8 hours. The precursor solubility data in scCO2 were considered so that moderate conditions for complete miscibility of both components were chosen. The grafting density and surface coverage increased with temperature, CO2 density, reaction time and stirring. A compromise must be reached for temperature and pressure. Higher temperatures favor an activated process but lead to a lower CO2 density at a given pressure and make less effective the supercritical medium. Tenorio et al. 24 Values for grafting density and surface coverage in scCO2 were much larger than those obtained refluxing toluene at 110 ºC after 24 hours. This seems to be due to the better transport properties of scCO2 which allow the fast penetration and diffusion of the reactant into the support pores. Furthermore, functionalization using scCO2 can be considered a sustainable process. Microstructural characterization by gas adsorption showed how surface area and pore volume decreased as the grafting density increased. Pore size also decreased up to the formation of a compact monolayer. Further increase in the grafting density led to pore blocking and a substantial reduction of the surface area. Sample 8 prepared at 150 ºC and 29.0 MPa with stirring led to the highest grafting density (4.4 mmol g -1 ) and surface coverage (4.0 molecules nm -2 ) but with almost complete pore blockage, which led to very low surface area and pore volume. Sample 6 synthesized at 80 ºC, 25.0 MPa and 4 hours without stirring exhibited the best properties: high surface coverage and high surface area and pore volume. Shorter times can be employed if stirring is incorporated. In comparison to previous reports on the surface modification of other porous SiO2 supports in scCO2, the reaction was performed at lower temperature and/or pressure conditions. It is also shown that microstructure must be always taken into account in the functionalization process to maximize the material efficiency. ACKNOWLEDGEMENTS We gratefully acknowledge the financial support of the Spanish Ministry of Economy and Competitiveness (MINECO), research project CTQ2013-41781-P. We also thank “ICTS- Centro Nacional de Microscopía Electrónica”, “Centro de Rayos X“ and “CAI Tenorio et al. 25 microanálisis elemental” at Universidad Complutense de Madrid (UCM) for technical assistance. 6. REFERENCES [1] J. Nawrocki, A. Dabrowska, Nonpolar stationary phases for HPLC - Recent achievements, Chem. Anal. (Warsaw) 46 (2001) 781-811. [2] X. Liu, P.K. Chu, C. Ding, Surface nano-functionalization of biomaterials, Mater. Sci. Eng., R 70 (2010) 275-302. [3] C. Airoldi, R.F. de Farias, The use of organofuntionalized silica gel as sequestrating agent for metals, Quím. Nova 23 (2000) 496-503. [4] P.K. Jal, S. Patel, B. Mishra, Chemical modification of silica surface by immobilization of functional groups for extractive concentration of metal ions, Talanta, 62 (2004) 1005- 1028. [5] F. Lefebvre, P. Putaj, J.-M. Basset, X. Wang, X. Fu, Modification of the adsorption and catalytic properties of micro- and mesoporous materials by reactions with organometallic complexes, Sci. China Chem., 53 (2010) 1862-1869. [6] C. Wang, L. Yin, L. Zhang, D. Xiang, R. Gao, Metal Oxide Gas Sensors: Sensitivity and Influencing Factors, Sensors, 10 (2010) 2088-2106. [7] M.-A. Neouze, U. Schubert, Surface modification and functionalization of metal and metal oxide nanoparticles by organic ligands, Monatsh. Chem., 139 (2008) 183-195. [8] J. Ducheta, J.-F. Gerardb, J.-P. Chapelc, B. Chabertd, Grafting of alkylchlorosilanes onto silica from solution for adhesion enhancement, J. Adhes. Sci. Technol. , 14 (2000) 691-718. [9] S. Wu, F. Li, R. Xu, S. Wei, G. Li, Synthesis of thiol-functionalized MCM-41 mesoporous silicas and its application in Cu(II), Pb(II), Ag(I), and Cr(III) removal, J. Nanopart. Res, 12 (2010) 2111-2124. [10] M. Barczak, M. Oszust, K. Michalak, K. Gdula, S. Pasieczna-Patkowska, E. Zieba, A. Dabrowski, Functionalized SBA-15 organosilicas as sorbents of mercury(II), cadmium(II) and copper(II), Glass Phys. Chem. , 40 (2014) 41-48. [11] L. Huang, J.C. Wu, S. Kawi, Rh-4(CO)(12)-derived functionalized MCM-41-tethered rhodium complexes: preparation, characterization and catalysis for cyclohexene hydroformylation, J. Mol. Catal. A: Chem. , 206 (2003) 371-387. Tenorio et al. 26 [12] D. Das, J.F. Lee, S.F. Cheng, Selective synthesis of Bisphenol-A over mesoporous MCM silica catalysts functionalized with sulfonic acid groups, J. Catal. , 223 (2004) 152- 160. [13] I.K. Mbaraka, D.R. Radu, V.S.Y. Lin, B.H. Shanks, Organosulfonic acid- functionalized mesoporous silicas for the esterification of fatty acid, J. Catal. , 219 (2003) 329-336. [14] M.H. Lim, C.F. Blanford, A. Stein, Synthesis of ordered microporous silicates with organosulfur surface groups and their applications as solid acid catalysts, Chem. Mater., 10 (1998) 467-+. [15] A. Walcarius, M. Etienne, B. Lebeau, Rate of access to the binding sites in organically modified silicates. 2. Ordered mesoporous silicas grafted with amine or thiol groups, Chem. Mater., 15 (2003) 2161-2173. [16] D. Sanli, C. Erkey, Silylation from supercritical carbon dioxide: a powerful technique for modification of surfaces, J. Mater. Sci., 50 (2015) 7159-7181. [17] C. Roy, A. Vega-Gonzalez, C.A. Garcia-Gonzalez, T. Tassaing, C. Domingo, P. Subra-Paternault, Assessment of scCO2 techniques for surface modification of micro- and nanoparticles: Process design methodology based on solubility, Journal of Supercritical Fluids, 54 (2010) 362-368. [18] B.E. Poling, J.M. Prausnitz, J.P. O´Connell, The Properties of Gases and Liquids, International Edition ed., McGraw-Hill Singapore, 2001. [19] C.P. Tripp, J.R. Combes, Chemical modification of metal oxide surfaces in supercritical CO2: The interaction of supercritical CO2 with the adsorbed water layer and the surface hydroxyl groups of a silica surface, Langmuir, 14 (1998) 7350-7352. [20] Y. Sanchez-Vicente, C. Pando, M. Cortijo, A. Cabañas, Chemical surface modification of mesoporous silica SBA-15 with a tertiary aminosilane using supercritical carbon dioxide, Microporous Mesoporous Mater., 193 (2014) 145-153. [21] Y. Sanchez-Vicente, L.A. Stevens, C. Pando, M. Jose Torralvo, C.E. Snape, T.C. Drage, A. Cabañas, A new sustainable route in supercritical CO2 to functionalize silica SBA-15 with 3-aminopropyltrimethoxysilane as material for carbon capture, Chem. Eng. J., 264 (2015) 886-898. [22] P. Lopez-Aranguren, J. Saurina, L.F. Vega, C. Domingo, Sorption of tryalkoxysilane in low-cost porous silicates using a supercritical CO2 method, Microporous Mesoporous Mater., 148 (2012) 15-24. Tenorio et al. 27 [23] Y. Shin, T.S. Zemanian, G.E. Fryxell, L.Q. Wang, J. Liu, Supercritical processing of functionalized size selective microporous materials, Microporous Mesoporous Mater., 37 (2000) 49-56. [24] T.S. Zemanian, G.E. Fryxell, J. Liu, S. Mattigod, J.A. Franz, Z.M. Nie, Deposition of self-assembled monolayers in mesoporous silica from supercritical fluids, Langmuir, 17 (2001) 8172-8177. [25] D. Rébiscoul, V. Perrut, O. Renault, F. Rieutord, S. Olivier, P.H. Haumesser, Alkoxysilane layers deposited by SC CO2 process on silicon oxide for microelectronics applications, J. Supercrit. Fluids, 51 (2009) 287-294. [26] B.A. Ashu-Arrah, J.D. Glennon, K. Albert, Synthesis and characterisation of bonded mercaptopropyl silica intermediate stationary phases prepared using multifunctional alkoxysilanes in supercritical carbon dioxide as a reaction solvent, J. Chromatogr. A, 1222 (2012) 38-45. [27] Y. Sanchez-Vicente, O. Alonso-Pastor, C. Pando, A. Cabanas, Phase behaviour of the two binary systems formed by CO2 and the silane precursors N- 3-(trimethoxysilyl)propyl aniline or (3-mercaptopropyl) trimethoxysilane, J.Chem. Thermodyn., 103 (2016) 152-156. [28] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Triblock copolymer syntheses of mesoporous silica with periodic 50 to 300 angstrom pores, Science, 279 (1998) 548-552. [29] D.Y. Zhao, Q.S. Huo, J.L. Feng, B.F. Chmelka, G.D. Stucky, Nonionic triblock and star diblock copolymer and oligomeric surfactant syntheses of highly ordered, hydrothermally stable, mesoporous silica structures, J. Am. Chem. Soc., 120 (1998) 6024- 6036. [30] I.G. Shenderovich, G. Buntkowsky, A. Schreiber, E. Gedat, S. Sharif, J. Albrecht, N.S. Golubev, G.H. Findenegg, H.H. Limbach, Pyridine-N-15 - A mobile NMR sensor for surface acidity and surface defects of mesoporous silica, J. Phys. Chem. B, 107 (2003) 11924-11939. [31] M. Jaroniec, M. Kruk, J.P. Olivier, Standard Nitrogen Adsorption Data for Characterization of Nanoporous Silicas, Langmuir, 15 (1999) 5410-5413. [32] L. Saikia, D. Srinivas, P. Ratnasamy, Chemo-, regio- and stereo-selective aerial oxidation of limonene to the endo-1,2-epoxide over Mn(Salen)-sulfonated SBA-15, Appl. Catal., A, 309 (2006) 144-154. [33] C.-C. Chen, S. Cheng, L.-Y. Jang, Dual-functionalized large pore mesoporous silica as an efficient catalyst for bisphenol-A synthesis, Microporous Mesoporous Mater., 109 (2008) 258-270. Tenorio et al. 28 [34] X. Feng, G.E. Fryxell, L.Q. Wang, A.Y. Kim, J. Liu, K.M. Kemner, Functionalized Monolayers on Ordered Mesoporous Supports, Science, 276 (1997) 923. [35] H. Van Duc, D. Tuyet Phuong, D. Quang Khieu, N. Huu Phu, V. Anh Tuan, The synthesis of novel hybrid thiol-functionalized nano-structured SBA-15, Adv. Nat. Sci: Nanosci. Nanotechnol. , 1 (2010) 035011. [36] V.-D. Hoang, T.-P. Dang, Q.-K. Dinh, H.-P. Nguyen, A.-T. Vu, The syntesis of novel hybrid thiol-functionalized nano-structured SBA-15, Adv. Nat. Sci.: Nanosci. Nanotechnol., 1 (2010) 1-6. [37] E.B. Celer, M. Kruk, Y. Zuzek, M. Jaroniec, Hydrothermal stability of SBA-15 and related ordered mesoporous silicas with plugged pores, J. Mater. Chem., 16 (2006) 2824- 2833. [38] NIST Webbook http://webbook.nist.gov/chemistry/fluid/. http://webbook.nist.gov/chemistry/fluid/ SUPPLEMENTARY INFORMATION THIOL GROUP FUNCTIONALIZATION OF MESOPOROUS SiO2 SBA-15 USING SUPERCRITICAL CO2 M.J. Tenorio†, J. Morère†, C. Carnerero†, M.J. Torralvo‡, C. Pando† and A. Cabañas†,* †Dep. Química-Física I, ‡Dep. Química Inorgánica I, Universidad Complutense de Madrid, 28040 Madrid. SPAIN Figure S1 compares the low angle XRD patterns of the SiO2 SBA-15 substrate and sample 6 modified with mercaptopropyltrimethoxysilane (MPTMS) dissolved in supercritical CO2 (scCO2) at 80 ºC and 25.0 MPa for 4 hours. Figure S1. Low angle XRD patterns of SiO2 SBA-15 support and sample 6 modified in scCO2. XRD patterns showed three well resolved peaks which were indexed according to the two dimensional P6m hexagonal symmetry. The first peak was assigned to the (100) planes and appeared at ca. 0.94 2 values for the SiO2 SBA-15 support. Values of 9.2 and 10.7 nm were calculated, respectively, for the spacing between (100) planes, d100, and the cell parameter, ao, corresponding to the distance between two pores centers. The slight difference in the peak position for the two samples is within the experimental error. We conclude that the modification process in scCO2 does not alter the support structure. TABLE OF CONTENTS Highlights THIOL FUNCTIONALIZATION-cabañas final SUPPLEMENTARY Material