Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane

Microporous and Mesoporous Materials xxx (2016) 1e11

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane Frank Bauer a, *, Saskia Czihal a, Marko Bertmer b, Ulrich Decker c, Sergej Naumov c, Susan Wassersleben a, Dirk Enke a, ** €t Leipzig, D-04103 Leipzig, Germany Institute of Chemical Technology, Universita €t Leipzig, D-04103 Leipzig, Germany Institute of Experimental Physics II, Universita c Leibniz Institute of Surface Modification, D-04318 Leipzig, Germany a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 November 2015 Received in revised form 16 January 2016 Accepted 25 January 2016 Available online xxx

Surface modification of mesoporous biogenic silica (rice husk ash) by aqueous and ethanol-based 3aminopropyltriethoxysilane (APTES) grafting solutions has been investigated using 29Si NMR, ESI-MS, TGA, surface-sensitive NH2 titration, and DFT method. Prior to grafting, a rapid formation of ladderlike aminosilane oligomers has been observed in both solvents limiting the grafting time for the mostly desired formation of a monomolecular organosilane layer to few minutes. An excess of water yielded a long-term stable equilibrium between the oligomerized APTES species within the grafting solution promising a reproducible surface modification. After curing at 120
C, washing studies showed that these oligomeric aminosilane clusters have been successfully grafted but yielding a heterogeneous surface morphology with even pore blocking rather than a dense, uniform aminosilane monolayer. In addition, the optimized structures of three different grafting modes between APTES and a silica cluster model were obtained by DFT calculations. The results indicated that bond lengths and angles of the different grafting modes reveal no substantial structural distortions. Nevertheless, the ladder-like grafting mode is given preferential consideration because the DFT results are in accordance with the ESI-MS and 29Si NMR findings. © 2016 Elsevier Inc. All rights reserved.

Keywords: Biogenic silica Modification 3-Aminopropyltriethoxysilane 29 Si NMR spectroscopy DFT calculations

1. Introduction For the preparation of organic-inorganic composite materials, coupling agents such as the very versatile organo-substituted trialkoxysilanes are commonly used to accomplish a durable molecular bridge between the polymeric matrix and the inorganic reinforcement. In addition, these bifunctional organosilanes can be used as “primer” molecules providing an anchorage point on the surface of oxide particles and a linkage for the immobilization of different kinds of biomolecules such as enzymes [1], dyes [2e4], and metal nanoparticles [5]. Thus, surface functionalization of silica fibers [6,7], silica nanoparticles [8,9], ordered mesoporous materials [10], porous glasses [11], and biosilica materials [12] by various trialkoxysilanes is a crucial issue for their numerous industrial,

* Corresponding author. Tel.: þ49 341 9736300; fax: þ49 341 9736349. ** Corresponding author. Tel.: þ49 341 9736302; fax: þ49 341 9736349. E-mail addresses: [email protected] (F. Bauer), dirk.enke@uni-leipzig. de (D. Enke).

biological, and environmental applications. Furthermore, the stability of the grafted trialkoxysilane layer on the inorganic surface is an important question for the success of multi-step functionalization reactions and the long-term stability of composite systems. At this point it has to be mentioned that in case of modern photochemistry relatively large dye molecules derived from azobenzene/ stilbene [13] and perylene [14] have been firmly incorporated only by adsorptive forces into molecular sieves such as AlPO4-5, ZSM-5 and zeolite L. For the large-pore zeolite L having one-dimensional channels of 12-membered rings with channel openings of 0.71 nm in diameter, however, arising dye leakage could only be strictly avoided by sealing the pore aperture with a polymer coating [15]. But, the effectiveness of such sealing techniques is highly questionable for many other macro/meso-porous systems, but most important for all non-porous host materials. Therefore, chemical anchoring of active guest components onto the surface of inorganic host systems is the base of many composite applications and a permanent interest on the morphology of each individual molecular bridge within new (but also properly investigated) hybrid materials seems to be justified.

http://dx.doi.org/10.1016/j.micromeso.2016.01.046 1387-1811/© 2016 Elsevier Inc. All rights reserved.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

2

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

Among the organosilanes, aminosilanes (in particular 3aminopropyltriethoxysilane (APTES)) are of special interest owing to the reactive amino group and their outstanding water solubility which significantly facilitates the application of any surface modification reaction on an industrial scale. Accordingly, considerable effort has been spent on optimizing silanization conditions in aqueous and non-aqueous solutions [16e18] early on and is still invested in this fundamental field [19e21]. The main objectives of such studies are the influence of silane concentration, water content, pH value, type of auxiliary solvent, temperature, reaction time, and drying/curing conditions on silane surface density and stability as Zhu et al. [22] recapitulated in their work on aminosilanederived layers on silica. The two classical mechanisms for surface modification by trialkoxysilanes involve: (i) a one-step solvolysis reaction under exceptionally water-free conditions and (ii) a two-step reaction via (partial or total) hydrolysis of the silane agent to the various intermediate organosilanols followed by condensation reactions with surface OH groups and/or neighboring silanols. While trace quantities of water can most likely be avoided in gas-phase reactions, it is difficult to exclude hydrolysis of alkoxysilanes in solutions even when working in dry solvents because water molecules are efficiently adsorbed on hydrophilic oxide surfaces [22]. On the other hand, excess water is reported to result in both uncontrolled lateral/ vertical crosslinking of silane molecules on the surface [23e25] and formation of silane oligomers within the grafting solution [26,27], which can also react with and attach to the surface. This polymerization problem of hydrolyzed trialkoxysilanes is associated with the ability of organosilanols to interact also with silanols of neighboring organosilane molecules, which give rise to a number of linear, branched, and cyclic siloxane structures. In addition to such disordered layers of oligomeric organosilanes, the selfcondensation of trialkoxysilanes grafted in form of ladder-like structures has been reported [28,29]. Using a modified LC-MS technique, Beari et al. [26] observed siloxanes with up to nine SiO units formed within a 10% aqueous solution of APTES. Obviously, the covalent attachment of such polycondensed organosilanes (similar to silsesquioxanes) results in a heterogeneous surface morphology and it may significantly influence the application of the modified surfaces. For investigating the molecular structure of grafted trialkoxysilanes with 29Si NMR spectroscopy, unfortunately, there is uncertainty with regard to the assignments of signals (Fig. 1). But, the relative abundance of the various silane species formed within the grafting solution and/or in surface-near interphase layer obviously controls the structure and morphology of the subsequent silane deposits on the surface. To gain more insight into the composition of eco-friendly aqueous APTES grafting solutions, this work on surface modification of rice husk ash (RHA) combines 29Si NMR spectroscopic studies with ESI-MS measurements. After curing and washing, the APTES-modified RHA samples have been characterized by sorption measurements, CHN analysis, thermogravimetry, 29 Si CP MAS NMR, and surface-sensitive NH2 titration. In addition, quantum chemical calculations were performed to obtain a more indicative image of the different APTES grafting modes rather than given by the schematic representations. 2. Experimental methods 2.1. Materials Rice husk ash (RHA) used in this study has been prepared due to a protocol of Alyosef et al. [30] including hot acid leaching of rice husk prior to burning at 600
C. The samples were composed of 99 wt.% SiO2 and showed less than 0.1 wt.% carbon. SEM pictures of

Fig. 1. Schematic representation of covalently bonded APTES on silica surfaces: (a) tridental grafting of aminosilane, (b) linear cross-linked oligomeric aminosilanes, and (c) latter-like arranged oligomeric aminosilanes; all of these structures may be responsible for dominating T3 signals observed in 29Si MAS NMR spectra.

the siliceous material derived from RHA, as was demonstrated earlier by Alyosef [30], reveal its biological origin as well as intraparticle meso/macropores with diameters between 40 and 60 nm (responsible for the remarkably high specific surface area of the treated RHA material) and additional macropores in the range of 300 and 500 nm. To increase the density of OH surface groups, the burnt RHA samples has been hydrothermally treated at 120
C for 2 h using a Teflon autoclave and finally dried at 120
C overnight. To obtain a silane surface concentration of one molecule 3aminopropyltriethoxysilane (APTES, obtained from SigmaeAldrich) per nm2, a certain amount of APTES (depending on the specific RHA surface after rehydroxylation) was dissolved in water, ethanol, and an ethanol/water mixture. The solvent ethanol (analytical grade, AnalaR NORMAPUR) contains max. 0.2% water. After solvent evaporation, the modified RHA samples were dried at 120
C overnight. Finally, the functionalized RHA has been washed three times with the same solvent used in the grafting process. 2.2. Characterization techniques The hydrolyzed and partially condensed APTES was characterized by NMR and MS spectroscopy. High-resolution 1H NMR and 29 Si NMR spectra were recorded using a Bruker Avance-600 IIþ spectrometer (Germany). Deuterated water and methanol were used as solvents. Especially, the 29Si NMR spectroscopy gives useful information on the coordination at the Si atom of trialkoxysilanes (T i signals) as well as of Si atoms of the silica material (Qi signals) [8]. The percentages of the 29Si species were determined by signal integration and thus, the extent of silane hydrolysis (T0 signals) and the different steps of silane condensation reactions (T1, T2, and T3 signals) have been estimated. The active silanol reactivity (SR) of the grafting solution were calculated according to SR ¼ (3 T0 þ 2 T1 þ 1 T2)/3 assuming that T3 structures do not contribute to SR because they do not bear any silanol groups [26]. Using standard ionization conditions and recording the positive ions, ESI-MS spectra of the grafting and washing solutions were measured on a Bruker Daltonics Esquire 3000þ spectrometer (Germany). The content of aminosilanes within the washing solutions has been

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

3

determined by salicylaldehyde modification of the amino groups and succeeding UV/vis measurements at a wave length of 404 nm using Spectronic Genesys 5 spectrophotometer (Milton Roy, USA). The pristine and modified RHA particles were characterized by nitrogen sorption, elemental and thermogravimetric analysis, surface-sensitive NH2 titration, and 29Si CP MAS NMR spectroscopy. Nitrogen sorption measurements were performed on a ASAP 2010 apparatus (Micromeritics) to determine the specific surface area (SBET), pore volume (VP) and pore diameter (dP) from the BJHmethod using the desorption branch. Elemental analyses were performed with a vario Max CHN (Elementar Analysensysteme GmbH) instrument. In addition, the amino groups present on the RHA surface were determined via titration [31]. The amount of organics on the modified RHA was measured by thermogravimetric analysis (TGA 7, PerkinElmer). The 29Si CP-MAS solid state NMR experiments were recorded on a Bruker Avance 750 spectrometer. 2.3. Quantum chemical calculations Quantum chemical calculations were performed using the Density Functional Theory (DFT) B3LYP method [32,33] by means of software Jaguar version 8.5 [34]. The silica cluster model and the different APTES grafting structures were optimized at B3LYP/631(d) level of theory, which seem to be a reliable method for studying the structures and stabilities of silica materials [35]. The frequency analysis was made at the same level of theory to obtain thermodynamic parameters such as total enthalpy (H) and Gibbs free energy (G) at 298 K. A 32T (tetrahedral) silica cluster model (cell size of about 1 nm  1 nm) in its hydrophilic form (about 4 OH groups/nm2) has been used as model system for calculations which resulted in a Si32O52H12 composition (96 atoms as a whole). The calculated OeH distance of the silica cluster model is about 0.09 nm, which is in good agreement with the literature [36]. The same also applies to the SieO bond lengths of about 0.166 nm and SieOeSi angles in the range of 129e160
. 3. Results and discussion 3.1. Pre-grafting investigation of APTES hydrolysis and condensation reactions Hydrolysis and self-condensation of a 4:1 (v/v) APTES/water mixture solved in methanol is visualized in Fig. 2. This water content of 20% (v/v) meets the stoichiometric amount of water needed for complete APTES hydrolysis. Peaks assignments of APTES and the various structures resulting from the interaction of APTES with water are as follows for 29Si NMR: 41 to 44 ppm (T0), 44.6 ppm (APTES), 50 to 54 ppm (T1), 56 to 62 ppm (T2), and 62 to 72 ppm (T3). Hydrolysis begins immediately on initiation of the reaction. All three silanol species (T0) are visible in the NMR spectrum at 30 min reaction time. The product of the first hydrolysis step (signal at 43.4 ppm) exhibits a higher concentration than the other silanols with NMR signals displaced to lower field. After 60 min reaction time, pristine APTES remains detectable but its content has been reduced from 93% to 8%. The silanol species are still present in small quantities while the progress of the condensation is quite clearly evident by the NMR spectrum. The totally condensed T3 species are the most abundant class in the reaction mixture, accounting for 74% of all silane species. Irrespective the grafting rate of those oligomerized organosilanes, their anchoring onto the silica surface will result in a non-uniform distribution of silane clusters and possible pore blocking rather than APTES monolayer formation [37,38]. As the concentrations of the intermediate T0 and T1 species remain low in the grafting solution it can

Fig. 2. 29Si NMR spectra of a 4:1 (v/v) APTES/water mixture after a reaction time of 30 min (d) and 60 min ( ) at room temperature.

be concluded that the self-condensation of APTES is faster than its hydrolysis. Thus, the fast condensation reaction limits the reaction time of such grafting processes to values < 60 min under these conditions, otherwise condensed aminosilane T3 species are predominantly available for surface functionalization. To get a close relationship with the industrial use of organofunctional trialkoxysilanes [26,39], the further studies focus in particular on dilute aqueous systems. The results of 29Si NMR investigations of APTES solved in pure water (APTES/water mixture of 1: 9 (v/v)) are shown in Fig. 3. This water content of 90% (v/v) exceeds the stoichiometric amount of water needed for complete APTES hydrolysis by a factor of 30. Because aqueous solutions of amino-functional silanes exhibit a pH of 11, the autocatalyzed hydrolysis is typically complete in a few minutes [26,27] and unhydrolyzed APTES is no more detectable after a reaction time of 1 h. In contrast, hydrolyzed APTES species (T0) can furthermore be observed in small quantities, while the quantity of the T1, T2, and T3 species even reaches 16%, 36%, and 44%, respectively. After a reaction time of about 57 h, no significant changes are observed in the reaction solutions. As demonstrated by Beari et al. [26], aqueous APTES solutions remain stable even over several weeks to months.

Fig. 3. 29Si NMR spectra of a 1:9 (v/v) APTES/water mixture after a reaction time of 1 h (d) and 57 h ( ) at room temperature.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

4

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

These findings prove that a purely aqueous medium suppresses the self-condensation reaction of hydrolyzed silane structures in contrast to alcoholic solvents [27]. To estimate the reactivity of grafting solutions as a function of reaction time and/or reaction conditions, the silanol content has been proposed as a measurement parameter [26,27]. The values of the active silanol reactivity (SR) of the 4:1 (v/v) APTES/water solution (Fig. 2) are 5% and 8% after a reaction time of 30 min and 60 min, respectively. For the fairly dilute 1:9 (v/v) APTES/water system (Fig. 3), the SR value of about 24% is considerably larger and almost independent of the reaction time. For comparison, Beari et al. [26] reported a maximum value of silanol groups of 55% after 7 h in the case of a 4 wt.% APTES/water solution and still 47% of remaining reactive SieOH functions after a standing time of 172 days. Thus, in highly diluted aqueous APTES solutions, reactive silanols are maintained at higher concentrations over longer periods in contrast with grafting solutions having only the stoichiometric amount of water. The convincing conclusion that condensation reactions are suppressed in highly diluted aqueous APTES solutions is also supported by the lower amount of T3 species formed. This equilibrium between hydrolyzed and condensed APTES species in a purely aqueous medium permits longer reaction times. The relatively fast self-condensation reactions within grafting solutions having the stoichiometric amount of water (or less) can even inevitably restrict an effective grafting process to only few minutes. It should be pointed to the fact that low SR values need not mean a low grafting yield of the trialkoxysilane applied because typical organosilane T3 moieties formed within the grafting solution (however, pure polysilsesquioxanes are excluded from this statement) can be covalently anchored via few, but existing, reactive OH and/or alkoxy end groups of such oligomerized silane clusters as will be illustrated in the following section. The NMR findings on self-condensation reaction of APTES in aqueous solutions are supported and rendered more precisely by ESI-MS studies. Such ESI-MS experiments became necessary because the low sensitivity of 29Si NMR spectroscopy is not suitable for the investigation of typical, highly diluted grafting solutions of about 1e2% APTES. The aim of the comparative measurements was also to detect the detailed build-up of oligomerized aminosilanes (including the number of their individual SieOC2H5 or SieOH end groups which should depend on the aqueous and alcoholic APTES solutions applied). After a reaction time of 1 h, the mass spectrum of a 1:74 (v/v) APTES/water mixture shows both a MS signal for unconverted APTES molecules at 221 Da and several higher molecular weight signals with a characteristic MS difference pattern of 119 Da pointing to a NH2C3H6eSiO(OH) repeat unit of APTES oligomers (Fig. 4). Please note, the molecular peak of APTES in purely aqueous solutions has been observed non-typically at 221 Da (assuming radical cation formation due to a very low ionization energy) while in alcoholic solutions the positive charges are mostly contributed by protons, i.e. the typical (M þ Hþ) signal at 222 Da appeared (not shown). We assume that this different, solvent-dependent ionization behavior holds also for all APTES oligomers. While no fragmentation was found using a standard ESI source, H2O elimination under positive ion ESI conditions has particularly been detected for siloxane species with OH end groups. Using the aminopropylhydroxysilyl repeat unit of 119 Da, the observed MS signals prominently indicate the formation of siloxanes with two to seven SiO units. Because MALDI-TOF MS studies of silica nanoparticles modified with several trialkoxysilanes propose a ladder-like structure of siloxane moieties [40], it is obvious to assign the molecular structures of a tetramer and a pentamer (as shown in Fig. 5) to the dominating MS signals at 560 Da and 661 Da,

Fig. 4. ESI mass spectrum of a 1:74 (v/v) APTES/water mixture recorded after a reaction time of 1 h.

respectively. By adjusting the number of SieOC2H5 and SieOH end groups and probably taking any H2O elimination (one or two-fold) into account, all of the observed MS signals can be attributed to such ladder-like siloxane structures. For example, the signal at 762 Da is the result of H2O elimination from a totally ethoxylated aminosilane hexamer ((NH2C3H6eSiO3/2)6(C2H5)2(OC2H5)2 having a molar mass of 808 Da) after hydrolysis of one ethoxy group (MS signal at 780 Da). The mass spectrum of a 1:74 (v/v) APTES/ethanol mixture (Fig. 6) reveals signals of APTES oligomers with a repeat unit of 119 Da as well. Hence, the trace water content of commercial ethanol is sufficient for the aminosilane hydrolysis and the build-up of ladder-like siloxane structures. The difference between aqueous and alcoholic APTES solutions consists essentially in the degree of substitution of non-hydrolyzed ethoxy groups by OH groups both attached to the aminosilane oligomer. For example, the formation of aminosilane pentamers (the totally ethoxylated species (NH2C3H6eSiO3/2)5(C2H5)2(OC2H5)2 has a molar mass of 661 Da) is able to prove in aqueous solutions with a dominating MS signal at 661 Da and a weak signal of the hydrolyzed oligomer at 633 Da (Fig. 4). In alcoholic solutions, the intensity of the MS signal nonhydrolyzed aminosilane oligomer at 662 Da (M þ Hþ) is weak while the signals of the hydrolyzed species at 616 Da and 570 Da are

Fig. 5. Proposed structures of APTES oligomers with MS signals at 560 Da and 661 Da.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

Fig. 6. Detail of ESI mass spectrum of a 1:74 (v/v) APTES/ethanol mixture recorded after a reaction time of 1 h.

obviously stronger (Fig. 6). The characteristic MS difference pattern of 46 Da is due to the substitution of one ethoxy end group by an OH group and the subsequent elimination of one water molecule from the hydrolyzed aminosilane induced by ESI-MS conditions. In addition, the MS spectra of alcoholic APTES solutions need some artifact correction. For example, the signal at 573 Da can be assigned to a original double positive charged adduct formed from APTES ion (222 Da, M þ Hþ) and its tetraethoxylated dimer ion (368 Da, M þ Hþ) followed by NHþ 4 elimination (18 Da). Similar concerns the signal at 527 Da after hydrolysis of one ethoxy group of the dimeric APTES including the subsequent elimination of one water molecule. Although the mass spectra can only be used for the qualitative detection of aminosilane oligomers and do not significantly reflect the percentage distribution of the individual components, it can be assumed that the ladder-like APTES oligomers are mainly formed as ethoxylated species in different solvents (water being prerequisite) prior to surface grafting. The following hydrolysis into aminosilane oligomers with OH end groups (probably more appropriate for their anchoring on silica surfaces) seems to be more likely in alcoholic solutions rather than in pure water. On the basis of these MS results on fast oligomer build-up in both aqueous and alcoholic solutions, grafting of trialkoxysilanes onto surfaces will take place by oligomerized organosilanes in the form of clusters and islands which do not permit the formation of homogeneous, monomolecular organosilane coverages [21,41]. Finally, the formation of totally condensed polysilsesquioxanes (NH2C3H6eSiO3/2)n (n ¼ even) with a molar repeat unit of 110 Da has been expected in both aqueous and alcoholic APTES solutions. Their presence on the surface of silica nanoparticles modified by trialkoxysilanes has been reported in previous MALDI-TOF MS studies [29,42]. Unfortunately, appropriate molecular weight signals (M or M þ Hþ) such as 440/441 Da, 660/661 Da, and 880/ 881 Da are not conclusive. Although the reasons for that discrepancy between MALDI-TOF MS and ESI-MS studies are still unknown, both techniques have proved to reveal detailed phenomena of surface modification by organosilanes. 3.2. Hydrolytic and solvolytic studies of APTES surface modifications The long-term stability of any surface modification is crucial as the performance and durability of industrially manufactured

5

composite materials depend to a large degree on it [43]. For APTES modified silica, it is a well-known fact that the chemical stability in aqueous medium is rather limited due to the hydrolysis of siloxane bonds owing to the basic properties of the amine function [44,45]. Unlike the stability studies mentioned above in which the modified surfaces are exposed to a wide range of conditions on prolonged application times, our investigations concern the amount and nature of non- or weakly bounded organosilane species after annealing/curing the APTES loaded samples at 120
C for 12 h. A simple washing procedure using different solvents has been applied for the release of the chemically non-bonded organosilane species. To measure the organosilane content washed off, both the aminosilane concentration within the washing solutions and the amount of organics on the silica surface before and after washing have been determined. As shown in Table 1, washing with water and ethanol resulted in a mass loss of 2.5 and 2.1 wt.%, respectively. Similar contents of organosilanes removed by both solvents have been detected within the washing solutions by salicylaldehyde modification of the released amino groups and succeeding UV/vis measurements. To elucidate the chemical nature of organosilanes detached from RHA surfaces, ESI-MS spectroscopy has been exclusively applied because their concentrations in the washing solutions have been below the NMR detection limit. For water and ethanol as washing solvents, Fig. 7 reveals that preferably APTES (molar mass ¼ 221/222 Da) and its hydrolyzed species have been released rather than APTES oligomers (please note that up to aminosilane heptamers were observed in the grafting solutions). While such signals of high-molecular weight oligomers are missing in the washing solutions, APTES monosilanol (M þ Hþ  H2O; 176 Da), aminosilanediol (M þ Hþ  H2O; 148 Da), and aminosilanetriol (M þ Hþ  H2O; 120 Da) can clearly be detected. However, it cannot be generally ruled out that a formation of hydrolyzed APTES species as well as dimeric aminosilanols, e.g., signal of the totally hydrolyzed species at 239 Da (M þ Hþ  H2O), takes place during the washing procedure, i.e. the detected aminosilanols were not originally adsorbed on the RHA surface. Because the hydrolyzed aminosilanes washed off by both water and ethanol are primarily from monomeric APTES species, it can be concluded that single APTES molecules seem to be predominantly weakly adsorbed on RHA surfaces (and not covalently bonded even after a thermal treatment at 120
C) while most of the aminosilane oligomers formed within the grafting solutions are firmly anchored. For the same reason, the assumption that APTES is primarily grafted in tri-dental mode on silica surfaces which can form a homogeneous monolayer seems to be hardly plausible. Furthermore, the preorganization of aminosilane molecules in form of ladder-like oligomers even in highly diluted aqueous and alcoholic grafting solutions and their sufficient hydrolytic stability after a typical curing step corroborate the assumption of an inhomogeneous aminosilane coverage on RHA surfaces by cluster-like organosilane moieties [41]. Table 1 Data of CHN elementary analysis of aminofunctionalized RHA samples before and after washing in different solvents. Solvent

C (wt.%)

H (wt.%)

N (wt.%)

Dmwashing (%)

Water After washing Ethanol/water After washing Ethanol After washing

1.69 1.63 2.13 1.95 2.15 2.07

1.21 1.20 1.34 1.26 1.31 1.29

0.67 0.65 0.71 0.72 0.68 0.69

e 2.5 e 5.9 e 2.1

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

6

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

Fig. 7. ESI mass spectra of the aqueous and alcoholic washing solution after annealing/ curing the APTES loaded sample at 120
C for 12 h.

3.3. Characterization of APTES-modified rice husk ash The hydrolysis/condensation studies of APTES grafting solutions discussed above reveal that water in traces, stoichiometric amounts or in excess promotes the interaction of aminosilanes for crosslinking, inevitably leading to the formation and deposition of organosilane clusters at typical reaction times of 1 h or longer. Due to the stable equilibrium between the oligomerized aminosilane species using an excess of water, the 1:74 (v/v) APTES/water mixture has been preferably applied for obtaining a reproducible surface modification of RHA. Nevertheless, the 4:1 (v/v) and the 1:9 (v/v) APTES/water solutions showed similar features regarding nitrogen sorption characteristics, the thermogravimetric behavior, and the distribution of Ti structures in 29Si CP MAS NMR spectra. 3.3.1. Sorption measurements The presence of aminosilane clusters formed within the grafting solutions results not only in a non-uniform distribution of the grafted amino groups, but also can render the pores partially inaccessible for additional APTES moieties or even block the pore entries. Fig. 8 shows the nitrogen sorption isotherms of RHA samples before and after APTES grafting whereas data in Table 2 present details of the textural parameters. For the pristine RHA, the values are in agreement with the structural data as given by Alyosef et al. [30] who described the benefit of acid leaching prior to burning in the preparation of porous biogenic silica extracted from rice husks. The structural defects resulting from the recommendable hydrothermal rehydroxylation process after RHA burning are within admissible range of <10% (Table 2). Both the rehydroxylated and the aminofunctionalized sample exhibits the common characteristics of a type II isotherm with closed hysteresis loops revealing a broad pore distribution of mesoand macropores. After modification with APTES, rice husk ash shows a substantial reduction of BET surface area and total pore volume by 54% and 45%, respectively, in comparison with the rehydroxylated sample. The blockage of RHA pores, especially of micro- and mesopores by APTES clusters, is demonstrated by the lower amount of nitrogen adsorbed. The assumed blockage of smaller pores by APTES clusters can be supported by the following results of quantum chemical calculations. Thus, grafting of a single APTES molecule at the pore opening will reduce the pore diameter only by ~0.5 nm. An APTES tetramer properly anchored at the pore aperture should cover an area of about 1.4  1.0 nm2, most likely

Fig. 8. Nitrogen-sorption isotherms of rehydroxylated (d) and aminofunctionalized ( ) RHA samples.

sufficient for preventing nitrogen probe molecules to enter into larger micropores (1e2 nm) and smaller mesopores (2e2.5 nm) but nearly ineffective for typical mesopores. That is why, the behavior of the isotherms in Fig. 8 is significant different between the pristine and modified RHA sample in the low pressure region up to 0.05 p/p0. Beyond the sterical hindrance, a specific feature of the smaller pores, e.g. a higher density of OH groups, that results in their preferential blockage is not assumed. In accordance with these findings, the mean pore diameter (dP) determined from the desorption branch of the isotherms using the BJH method increases (Table 2). Similar conclusions of APTES grafting on mesoporous silica, i.e. clustering of APTES in the presence of water and the nonuniform distributions of amino groups with higher grafting densities at the pore entrances or eventually pore blocking, have been reported for SBA-15 [46] and MCM-41 [47]. 3.3.2. CHN/TG analysis and surface-sensitive NH2 titration Elementary and thermogravimetric analysis were applied to determine the organosilane coverage of modified rice husk ash before and after the washing procedure. In the case of elementary analysis (Table 1), the hydrogen content determined can be enhanced due to dehydroxylation reactions of the siliceous material while the TGA mass loss can additionally be influenced by incomplete combustion of the grafted aminosilanes at temperatures < 450
C (see below). For aqueous APTES grafting solutions, the averaged values of CHN/TGA measurements resulted in a content of organics present on the RHA surface of about 2.5 wt.% after the drying/curing step at 120
C. Using water and ethanol as washing solvents yielded a mass loss of 2e6% indicating a minor need for washing procedures with these more or less polar solvents. Transformed on an atomic base, the carbon and nitrogen contents observed reflect a carbon-to-nitrogen ratio of about 3:1 as expected for the aminopropyl moieties of APTES. Based on BET

Table 2 Specific surface area (SBET), total pore volume (VP), and pore diameter (dP) for parent and treated RHA samples obtained by BET method. Sample

SBET [m2/g]a

VP [cm3/g]a

dP [nm]b

Pristine RHA Rehydroxylated RHA APTES modified RHA

314 305 140

0.439 0.402 0.219

3.5 3.8 7.3

a b

Nitrogen adsorption. Nitrogen adsorption, mean pore diameter, desorption branch, BJH method.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

surface area and carbon content, the surface coverage of APTES species has been calculated and displayed in Fig. 9. Due to the silane loading applied, the calculations unsurprisingly confirm surface densities of about 1 APTES molecule/nm2. In addition, titration has been used to quantitatively determine the free accessible amino groups grafted on the surface of rice husk ash (Fig. 9). It has been commonly known that large discrepancy can exists between the results obtained by thermogravimetric and elemental analysis vs. surface-sensitive, chemical assay methods involving direct reaction with amino groups [48e51]. As measured by elementary analysis (Table 1), the nitrogen content of the aminofunctionalized sample is about 0.7 wt.%, corresponding to about 500 mmol/g of total amino groups. The amino group density as obtained by the surfacesensitive method via titration [31] is in the range between 325 and 410 mmol/g of NH2 groups corresponding to surface densities of 0.65e0.79 NH2 groups/nm2 (Fig. 9). Consistent with the reports in literature, our result detected by the surface-sensitive titration method shows that a large fraction of about 30% of total amino groups seems to be not really accessible as anchoring sites in further functionalization reactions. Non-free amino groups, i.e., totally occluded or only partially obstructed, can be the result of pore blockage and/or strong interactions with neighboring atoms. For example, the assumption of hydrogen bonds between amino groups and residual silanols as well as hydrogen bonds between neighboring amino groups have been supported by quantum chemical argumentations [52]. Potential interactions of the aminopropyl tail with surface atoms are indicated by our quantum chemical calculations for the monodental grafting of linear/cyclic and ladder-like APTES polymers (see Section 3.4). In both cases, aminopropyl groups can be aligned nearly parallel to the silica surface at a 0.17 nm distance and thus obviously hindered by hydrogen bonding while the aminopropyl group in the tridental grafting mode is orientated more or less perpendicular to the silica surface and should be accessible without restrictions. Thermogravimetric analysis of pristine and rehydroxylated RHA shown in Fig. 10 indicates that the initial weight loss (up to 150
C) results from desorption of physisorbed water molecules. At temperatures higher than 200
C, a more or less distinctive water evolution follows originating from dehydroxylation reactions of silanol groups. Compared to the pristine RHA, the higher amount of adsorbed water and the more extensive mass loss by dehydroxylation between 200 and 750
C can reasonably be expected by an increased hydrophilicity of the rehydroxylated sample. The continuous dehydroxylation of these RHA samples complicates,

Fig. 9. Comparison of surface densities of aminopropyl groups obtained by elementary analysis ( ) and surface-sensitive NH2 titration ( ) for different solvents.

7

Fig. 10. TGA profiles of pristine rice husk ash (—), after rehydroxylation (/), after grafting of APTES including a thermal treatment at 120
C (d), and after grafting and washing procedures ( ).

however, a precise determination of the amount of organosilane grafting by temperature-programmed oxidation. A thermal decomposition of grafted APTES in air takes place between 300 and 380
C and all organosilane species appear to be entirely removed around 410
C. Similar organosilane decomposition temperatures of 280e400
C have been observed for APTES functionalized SBA15 samples [53] and silica nanoparticles modified by different trialkoxysilanes [40]. The thermogravimetric curves of the sample before and after washing show no significant difference and the TGA mass loss is estimated to be about 0.8 wt.%. Unfortunately, this is not in agreement with the carbon contents of about 1.6 wt.% as determined by elementary analysis. However, the functionalized RHA samples show a steeper TGA gradient at temperatures > 200
C compared with the rehydroxylated sample. An obvious explication is a decomposition of the organosilane layer over a longer period of 300e750
C as observed for aminofunctionalized mesoporous silica [54,55]. The weight loss of about 2.5 wt.% thus determined is far better in agreement with the original load of aminopropyl groups as already determined by elementary analysis. 3.3.3. 29Si CP MAS NMR spectroscopy 29 Si CP MAS NMR technique was applied for better understanding the grafting of APTES. The spectrum of the untreated RHA shows three signals at 91.8 (Q2), 101.9 (Q3), and 113.1 (Q4) ppm (Fig. 11a), which are usually assigned to geminal silanols, isolated silanols, and siloxane groups, respectively. The fact that the crosspolarization technique applied prefers silicon nuclei near to protons, i.e., Si atoms at or near the surface, makes this procedure a sensitive method for detecting SiOH groups but only allows a semiquantitative determination of silicon atoms on solid surfaces. Using Q4 silicon for normalization, the peak-area ratio of the Q2, Q3, and Q4 units in the pristine RHA is 15:53:32, as summarized in Table 3. This ratio indicates that the RHA surface is mostly composed of isolated silanol groups (Q3) and siloxane units (Q4). As expected, the grafting process reduces the intensities of the NMR signals of geminal and isolated silanol groups in comparison with those of the siloxane groups (Fig. 11c). The peak-area ratio of the functionalized RHA is Q2:Q3:Q4 ¼ 6:50:44, still revealing a major content of isolated silanol groups. It is interesting to note that the peak assigned to Q2 units, which are present as a minor SiOH component on the surface of RHA, disappears to a greater extent than that of isolated

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

8

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

Fig. 11. 29Si CP MAS NMR spectra of pristine rice husk ash (a) and after grafting of APTES (c). For comparison, 29Si NMR spectrum of an aqueous APTES solution (b) after 5 h reaction time at room temperature.

silanols, thus indicating a higher grafting yield of geminal silanols (Table 3). At a given surface concentrations of about 1 APTES molecule/ nm2, the moderate reduction in the surface concentration of isolated SiOH groups of about 31% (Table 3) indicates that a tridentate anchoring of trifunctional silanes (Fig. 1a) is very unlikely. Such morphology of silane layers would yield a severe decrease in isolated silanols and a more intense redistribution of silicon sites in favor of Q4 units assuming a surface concentration of 2.25 OH groups/nm2 for biogenic silica [56] and that all the silanol groups are surface functionalization sites. Similar 29Si NMR findings of a silanol conversion of merely 34% due to trialkoxysilane grafting even at a high surface coverage of about 5.5 silane molecule/nm2 [8] support the assumed anchoring of APTES oligomers arranged in form of linear (Fig. 1b) and/or ladder-like (Fig. 1c) polymeric chains. For APTES grafting solutions, a high-resolution 29Si NMR signal at about 45 ppm can be detected. After bonding on RHA surfaces, the very broad 29Si CP MAS NMR signal between 55 ppm and 75 ppm corresponds to overlapping T2/T3 structures with a dominating T3 content at about 67 ppm (Fig. 11c). Even though often presented as crucial argument [36,57,58], the mere observation of T3 signals in the solid-state 29Si NMR spectrum does not ensure a tridentate bonding of trialkoxysilane molecules on silica surfaces. Early on in the application of the 29Si CP MAS NMR technique on surface functionalization, the assignment of T3 signals to the formation of oligomerized organosilane chains before grafting rather than to a tridentate anchoring mode of monomeric trifunctional silanes has been discussed in literature [59e61]. As Fig. 11b demonstrates in an exemplary manner, the preorganization of APTES species in aqueous solutions is

Table 3 29 Si CP MAS NMR signal areas IQx (in a.u.). 29

Si NMR signals

Geminal silanol (Q2) Free silanol (Q3) Siloxan (Q4) a

yQ x ¼ I0Q x  IQ x =I0Q x $100.

Pristine RHA

RHA after APTES treatment

Grafting yield yQxa (%)

45.9 166.9 100

15.0 115.2 100

67.3 30.9 e

characterized by signals of T2 units between 58 and 62 ppm and signals of T3 units between 65 and 72 ppm in dominance while the broad signal in the Qi region has to be assigned to the instrumental background. The formation of such APTES oligopolymers can take quickly place in the presence of water (even traces of water adsorbed on the surface of miscellaneous oxide and non-oxide materials are sufficient [22]) via silane hydrolysis and chain forming reactions between neighboring silane species. By adjusting the water content in the grafting solution, the populations of the T2 and T3 bonding mode of the APTES modified sample has been suggested to be governed [62]. At this point it has to be mentioned that the low sensitivity of 29Si CP MAS NMR spectroscopy makes this excellent technique practically unsuitable for investigations at APTES loadings of <5 wt.%. For example, Weichold et al. [63] doesn't reported on any Ti signals but only changes in the Q3:Q4 ratio at a loading of 1.7 wt.% APTES on silica nanoparticles although the presence of amino groups on the surface has been proven by a ninhydrin test solution. Thus, it remains an open question whether oligomeric silane structures are also present on silica surfaces at very low surface concentrations of grafted aminosilane. For surface modification by trialkoxysilanes, Fig. 11 convincingly shows that there are no significant differences between the 29Si chemical shift values of T3 structures characteristic of silane selfcondensation (T-O-T links) and silane grafting reaction with the surface (T-O-Q links) [64]. Furthermore, the detected T3 signals could include silsesquioxane species formed by self-condensation of trialkoxysilanes [42]. These fully condensed APTES species will cover the silica surface (with covalent links with it or not) and may not be soluble enough to be efficiently extracted by washing procedures applied. Finally, a veritable NMR spectroscopic evidence for the presence of covalent bonds between trifunctional organosilanes and the silica surface is difficult to offer, e.g. by twodimensional 1He29Si CP MAS NMR correlation experiments [60], and is commonly based on the decrease in the surface concentration of SiOH groups after grafting. 3.4. DFT study on the interaction between APTES and silica The most frequently asked questions about surface grafting of trialkoxysilanes concerns their grafting mode and surface morphology. Obviously, schematic views as shown in Fig. 1 have to withstand impartial scrutiny. In this section, theoretical investigations on the possible grafting modes of APTES on the surface of a silica cluster model are performed to reveal a more realistic impression of the surface arrangement of aminosilane species rather than to obtain thermodynamic data necessary to determine the grafting mode. To this end, density functional theory (DFT) is used to calculate the optimized structures of aminosilane species grafted to the surface of silica. The application of reaction energies for differentiating between the grafting modes has been regarded too doubtful due to the complexity of the reaction system. To adopt DFT-based methods to silica surface modeling [35], a simple 32T silica slab model with a thickness of about 0.3 nm (see Fig. 12) seems to be adequate when considering the rigidity of the silica cluster model necessary for studying the interaction between silica and the different grafting structures of APTES. In fact, the interaction of all aminosilane species anchored influences the 32T silica cluster structure only marginally, because it is important to maintain the structure of the silica cluster model for discussing the APTES grafting modes thoroughly [36]. Fig. 12 shows the most illustrated tri-dental APTES grafting mode in its optimized molecular structure. As the DFT calculations have shown, the structural distortions of both the tri-dentally grafted APTES and the silica cluster model are minor even in circumstances when there is a hydrogen bonding of the aminopropyl

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

Fig. 12. Optimized molecular structure for APTES tri-dentally grafted on a hydrophilic 32 Si-member silica cluster model and hydrogen-bonded to a neighboring OH group at the B3LYP/6-31(d,p) level of theory. For reasons of demonstrability, protons at the silica cluster's boundaries are omitted.

group to a surface hydroxyl group [45,65]. Largely independent of the potential interaction between amino and surface OH groups, the non-hydrolyzable SieC bond of the organosilane moieties is perpendicular to the silica surface. The related bond length and angle results are summarized in Table 4. The intermolecular angles :(OieSiSilaneeOi) for the tri-dental grafting mode of 102.5…115.7


9

are in sufficient agreement with the corresponding angles of APTES. In addition, the surface angles :(SiSilaneeOieSiSurf) of 122.7… 129.2
are equivalent to the APTES angles :(SiSilaneeOieCi). The calculated bond length of SiSilaneeOieSiSurface is 0.290…0.300 nm which is slightly longer than the bond length of the silica cluster model of about 0.285 nm. For 3-mercaptopropyltrimethoxysilane grafted tri-dentally onto an 8T silica cluster model, Wu et al. [36] obtained bond lengths SiSilaneeOieSiSurf in the range of 0.274… 0.300 nm and surface angles :(SiSilaneeOieSiSurf) between 110.0… 123.3
. Thus, the tri-dental grafting mode of trialkoxysilanes cannot be ruled out simple due to geometric reasons. But, the probability of an appropriate arrangement of 3 or even 4 OH groups per nm2 surface area is small. The optimized structure of a cross-linked, mono-dentally grafted aminosilane oligomer is visualized in Fig. 13a. For the sake of simplifying the polymeric arrangement, the interaction of a cyclic APTES tetramer on the silica surface has been calculated (Table 4). Similar to the tri-dental grafting mode, the characteristic angles :(OieSiSilaneeOi) of 105.1…109.3
reveal no serious structural distortions. In the case of mono-dental polymer bonding, the bond length SiSilaneeOieSiSurface of 0.309…0.319 nm is slightly longer than the bond length of the tri-dental grafting mode. Unsurprisingly, the elongated surface angles :(SiSilaneeOieSiSurf) of 140.0…153.9
are in sufficient agreement with the corresponding bond lengths. The most remarkable result of the quantum chemical calculations is, however, that the monodentally grafted aminosilane moieties lie flat to surface (but not

Table 4 The angle data (
) and bond length (Å) of APTES and different grafting modes. APTES & silica model

Tri-dental grafting mode

Linear/cyclic polymer grafting mode

Ladder-like polymer grafting mode

:(OieSiSilaneeOi) 105.5…111.1 :(SiSilaneeOieCi) 124.4…129.1 SiSilaneeOieCi 2.75…2.79 SiSurfeOieSiSurf 2.84…3.01

:(OieSiSilaneeOi) 102.5…115.7 :(SiSilaneeOieSiSurf) 122.7…129.2

:(OieSiSilaneeOSurf) 105.1…109.3 :(SiSilaneeOieSiSurf) 140.0…153.9

:(OieSiSilaneeOSurf) 108.4…109.6 :(SiSilaneeOieSiSurf) 142.7…160.9

SiSilaneeOieSiSurface 2.90…3.00

SiSilaneeOieSiSurface 3.09…3.19

SiSilaneeOieSiSurface 3.12…3.22

Fig. 13. Optimized geometries for APTES oligomers arranged in (a) a 4 Si-member linear/cyclic form and (b) a 6 Si-member latter-like form and grafted on a hydrophilic 32 Simember silica cluster model at the B3LYP/6-31(d,p) level of theory. For reasons of demonstrability, protons at the silica cluster's boundaries are omitted.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

10

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11

perpendicular as typically assumed in schematic illustrations such as Fig. 1b), inhibiting a free access to the amino group due to possible hydrogen bonding as indicated by surface-sensitive NH2 titration. For the ladder-like grafting mode as shown in Fig. 13b, detail information of bonding lengths and angles (see Table 4) does not indicate any substantial structural distortions as well. Compared to the tri-dental grafting mode, the surface angles :(SiSilaneeOieSiSurf) of 142.7…160.9
are conspicuously elongated but in the range of the intramolecular angles :(SiSilaneeOieSiSilane) of 146.7… 159.7
for a ladder-like silane arrangement. Similar to the monodentally grafted aminosilane, steric hindrance of the two surfacenear amino groups can be expected. In general terms, the bond lengths and angles of the different grafting modes show small deviations from the primary values in APTES and the 32T silica cluster. But, these differences cannot justify stating an order of stability for the different grafting modes. The application of reaction energies for differentiating between the grafting modes has been regarded too doubtful due to the complexity of the reaction system. Nevertheless, the calculated reaction enthalpy (DH) and Gibbs free energy (DG) per one new SiSilaneeOeSiSurface grafting bond are exothermic with 5 ± 4 kcal mol1 and 11 ± 4 kcal mol1, respectively, indeed irrespective for all grafting modes. For comparison, Wu et al. [36] calculated the tri-dental grafting reaction on a 8T silica cluster to be strong endothermic with DG ¼ þ242 kcal mol1 which obviously reduces the effective ability of this grafting mode in APTES anchoring. It can be assumed that the 8T silica cluster model is too rigid for modeling such multi-dental bonding structures of trialkoxysilanes. The ladder-like grafting mode is given preferential consideration because the DFT results are in accordance with the findings of 29Si MAS NMR spectroscopy, thermogravimetry, and atomic force microscopy that only few OH groups are required for anchoring large organosilane clusters [21,40].

References [1] [2] [3] [4] [5] [6] [7] [8]

[9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30]

4. Conclusions A high degree of silane self-condensation in eco-friendly aqueous APTES grafting solutions has been determined by 29Si NMR and ESI-MS measurements. While the stoichiometric amount of water resulted in a quick irreversible build-up of condensed aminosilane species limiting the grafting time for the mostly desired formation of a monomolecular organosilane layer to few minutes, an excess of water yielded a long-term stable equilibrium between the oligomerized APTES species promising a reproducible surface modification procedure. Detailed analysis of the MS data points to a ladder-like structure of the polycondensed aminosilanes formed even in highly diluted solutions. These findings are particularly interesting in view of a heterogeneous surface morphology of the biogenic silica to be expected after APTES modification. Quantum chemical calculations provided a deeper insight into the surface functionalization by APTES as well as a plausible explanation of the smaller amounts of amino groups determined by surface-sensitive NH2 titration. With respect to the different APTES grafting modes, the ladder-like grafting mode is given preferential consideration because the DFT results are in accordance with the MS/NMR spectroscopic and thermogravimetric findings. Acknowledgments The technical assistance of Mses. R. Oehme, I. Reinhardt, S. Schulze, and Mr. Dr. W.-D. Einicke is gratefully acknowledged.

[31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42]

[43]

[44] [45] [46] [47] [48] [49] [50] [51] [52] [53]

R.A. Sheldon, Adv. Synth. Catal. 349 (2007) 1289e1307. J. Lin, J.A. Siddiqui, R.M. Ottenbrite, Polym. Adv. Technol. 12 (2001) 285e292. M. Guli, Y. Chen, X. Li, G. Zhu, S. Qiu, J. Lumin. 126 (2007) 723e727. N. Al Dahoudi, J. Xi, G. Cao, Electrochim. Acta 59 (2012) 32e38. A.R. Morrill, D.T. Duong, S.J. Lee, M. Moskovits, Chem. Phys. Lett. 473 (2009) 116e119. A.T. DiBenedetto, Mater. Sci. Eng. A 302 (2001) 74e82. D.M. Laura, H. Keskkula, J.W. Barlow, D.R. Paul, Polymer 43 (2002) 4673e4687. €sel, H. Langguth, F. Bauer, H. Ernst, U. Decker, M. Findeisen, H.J. Gla E. Hartmann, R. Mehnert, C. Peuker, Macromol. Chem. Phys. 201 (2000) 2654e2659. R.P. Bagwe, L.R. Hilliard, W.H. Tan, Langmuir 22 (2006) 4357e4362. S. Wang, Microporous Mesoporous Mater. 117 (2009) 1e9. D. Enke, F. Janowski, W. Schwieger, Microporous Mesoporous Mater. 60 (2003) 19e30. P. Yuan, D. Liu, D.Y. Tan, K.K. Liu, H.G. Yu, Y.H. Zhong, A.H. Yuan, W.B. Yu, H.P. He, Microporous Mesoporous Mater. 170 (2013) 9e19. K. Hoffmann, F. Marlow, J. Caro, Adv. Mater. 9 (1997) 567e570. M. Busby, A. Devaux, C. Blum, V. Subramaniam, G. Calzaferri, L. De Cola, J. Phys. Chem. C 115 (2011) 5974e5988. A. Devaux, G. Calzaferri, P. Belser, P. Cao, D. Bruehwiler, A. Kunzmann, Chem. Mater. 26 (2014) 6878e6885. E.P. Plueddemann, in: E.P. Plueddemann (Ed.), Silane Coupling Agents, Plenum Press, New York, 1982, pp. 49e73. G.S. Caravajal, D.E. Leyden, G.R. Quinting, G.E. Maciel, Anal. Chem. 60 (1988) 1776e1786. F.D. Osterholtz, E.R. Pohl, J. Adhes. Sci. Technol. 6 (1992) 127e149. J. Huser, S. Bistac, M. Brogly, C. Delaite, T. Lasuye, B. Stasik, Appl. Spectrosc. 67 (2013) 1308e1314. P. Rostamzadeh, S.M. Mirabedini, M. Esfandeh, J. Coat, Technol. Res. 11 (2014) 651e660. G. Jaksa, B. Stefane, J. Kovac, Appl. Surf. Sci. 315 (2014) 516e522. M.J. Zhu, M.Z. Lerum, W. Chen, Langmuir 28 (2012) 416e423. K.M.R. Kallury, P.M. Macdonald, M. Thompson, Langmuir 10 (1994) 492e499. A.V. Krasnoslobodtsev, S.N. Smirnov, Langmuir 18 (2002) 3181e3184. V. Palanivel, D.Q. Zhu, W.J. van Ooji, Prog. Org. Coat. 47 (2003) 384e392. F. Beari, M. Brand, P. Jenkner, R. Lehnert, H.J. Metternich, J. Monkiewicz, H.W. Siesler, J. Organomet. Chem. 625 (2001) 208e216. O. Paquet, M.C.B. Salon, E. Zeno, M.N. Belgacem, Mater. Sci. Eng. C 32 (2012) 487e493. €sel, U. Decker, H. Ernst, A. Freyer, E. Hartmann, V. Sauerland, F. Bauer, H.J. Gla R. Mehnert, Prog. Org. Coat. 47 (2003) 147e153. L. Nothdurft, T. Glueck, W. Dempwolf, G. Schmidt-Naake, Macromol. Mater. Eng. 293 (2008) 132e139. H.A. Alyosef, A. Eilert, J. Welscher, S.S. Ibrahim, R. Denecke, W. Schwieger, D. Enke, Part. Sci. Technol. 31 (2013) 524e532. O.I. Voroshilova, A.V. Kiselev, Y.S. Nikitin, Colloid J. USSR 42 (1980) 223e229. A.D. Becke, J. Chem. Phys. 104 (1996) 1040e1046. C.T. Lee, W.T. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785e789. version 8.5, Jaguar, Schrodinger, Inc., New York, NY, 2014. A. Rimola, D. Costa, M. Sodupe, J.-F. Lambert, P. Ugliengo, Chem. Rev. 113 (2013) 4216e4313. J. Wu, L. Ling, J. Xie, G. Ma, B. Wang, Chem. Phys. Lett. 591 (2014) 227e232. T. Yokoi, H. Yoshitake, T. Tatsumi, J. Mater. Chem. 14 (2004) 951e957. H. Ritter, M. Nieminen, M. Karppinen, D. Bruehwiler, Microporous Mesoporous Mater. 121 (2009) 79e83. M. Brand, A. Frings, P. Jenkner, R. Lehnert, H.J. Metternich, J. Monkiewicz, J. Schram, Z. Naturforsch. 54 (1999) 155e164. F. Bauer, H. Ernst, D. Hirsch, S. Naumov, M. Pelzing, V. Sauerland, R. Mehnert, Macromol. Chem. Phys. 205 (2004) 1587e1593. €sel, S. Naumov, R. Mehnert, Macromol. F. Bauer, V. Sauerland, H. Ernst, H.A. Gla Chem. Phys. 204 (2003) 375e383. F. Bauer, V. Sauerland, H.J. Gl€ asel, H. Ernst, M. Findeisen, E. Hartmann, H. Langguth, B. Marquardt, R. Mehnert, Macromol. Mater. Eng. 287 (2002) 546e552. N.S. Bhairamadgi, S.P. Pujari, F.G. Trovela, A. Debrassi, A.A. Khamis, J.M. Alonso, A.A. Al Zahrani, T. Wennekes, H.A. Al-Turaif, C. van Rijn, Y.A. Alhamed, H. Zuilhof, Langmuir 30 (2014) 5829e5839. B.V. Zhmud, J. Sonnefeld, J. Non-Cryst. Solids 195 (1996) 16e27. M. Etienne, A. Walcarius, Talanta 59 (2003) 1173e1188. M. Moritz, M. Laniecki, Appl. Surf. Sci. 258 (2012) 7523e7529. N. Gartmann, C. Schutze, H. Ritter, D. Bruhwiler, J. Phys. Chem. Lett. 1 (2010) 379e382. A. del Campo, T. Sen, J.P. Lellouche, I.J. Bruce, J. Magn. Magn. Mater. 293 (2005) 33e40. J.M. Rosenholm, M. Linden, Chem. Mater. 19 (2007) 5023e5034. Y. Zhang, Y. Chen, IET Nanobiotechnol. 6 (2012) 76e80. Z. Qu, H. Xu, P. Ning, H. Gu, Surf. Interface Anal. 44 (2012) 1309e1313. A.A. Golub, A.I. Zubenko, B.V. Zhmud, J. Colloid Interface Sci. 179 (1996) 482e487. A.S.M. Chong, X.S. Zhao, A.T. Kustedjo, S.Z. Qiao, Microporous Mesoporous Mater. 72 (2004) 33e42.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046

F. Bauer et al. / Microporous and Mesoporous Materials xxx (2016) 1e11 [54] Z.H. Luan, J.A. Fournier, J.B. Wooten, D.E. Miser, Microporous Mesoporous Mater. 83 (2005) 150e158. [55] X.E. Zhang, W.J. Wu, J.F. Wang, X.Z. Tian, Appl. Surf. Sci. 254 (2008) 2893e2899. [56] S. Dixit, P. Van Cappellen, Geochim. Cosmochim. Acta 66 (2002) 2559e2568. [57] T. Tang, Y. Zhao, Y. Xu, D. Wu, J. Xu, F. Deng, Appl. Surf. Sci. 257 (2011) 6004e6009. [58] G.F. Andrade, D.C.F. Soares, R.K.D. Almeida, E.M.B. Sousa, J. Nanomater. (2012). Article ID 816496. [59] J.W. Dehaan, H.M. Vandenbogaert, J.J. Ponjee, L.J.M. Vandeven, J. Colloid Interface Sci. 110 (1986) 591e600. [60] C.A. Fyfe, Y. Zhang, P. Aroca, J. Am. Chem. Soc. 114 (1992) 3252e3255.

11

[61] K.C. Vrancken, L. Decoster, P. Vandervoort, P.J. Grobet, E.F. Vansant, J. Colloid Interface Sci. 170 (1995) 71e77. [62] Q. Tao, H.P. He, T. Li, R.L. Frost, D. Zhang, Z.S. He, J. Solid State Chem. 213 (2014) 176e181. [63] O. Weichold, B. Tigges, M. Bertmer, M. Moeller, J. Colloid Interface Sci. 324 (2008) 105e109. [64] S. de Monredon-Senani, C. Bonhomme, F. Ribot, F. Babonneau, J. SoleGel Sci. Technol. 50 (2009) 152e157. [65] E. Vansant, P. VanDerVoort, K. Vrancken, Chemical Modification of Silica: Applications and procedures, Characterization and Chemical Modification of the Silica Surface, 1995, pp. 149e192.

Please cite this article in press as: F. Bauer, et al., Water-based functionalization of mesoporous siliceous materials, Part 1: Morphology and stability of grafted 3-aminopropyltriethoxysilane, Microporous and Mesoporous Materials (2016), http://dx.doi.org/10.1016/ j.micromeso.2016.01.046