Effects of Al3+ Ions on Formation of Silica Framework and Surface

Jun 28, 2016 - Al3+ ions were introduced into silica framework at 318 K in order to make active Al sites for SO42– by the addition of aqueous sodium...
2 downloads 10 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Effects of Al3+ Ions on Formation of Silica Framework and Surface Active Sites for SO42− Ions Shigeo Sasahara†,‡,§ and Sumio Ozeki*,‡ †

Team21, Fuji Chemical Co. Ltd., 1683-1880 Nasubigawa Nakagaito, Nakatsugawa, Gifu, 509-9132, Japan Department of Chemistry, Faculty of Science, Shinshu University, 3-1-1 Asahi, Matsumoto, Nagano 390-8621, Japan

‡

S Supporting Information *

ABSTRACT: Al3+ ions were introduced into silica framework at 318 K in order to make active Al sites for SO42− by the addition of aqueous sodium silicate solution to aqueous sulfuric acid solution of Al2(SO4)3. The 27Al and 29Si NMR spectra of aluminosilicates were measured at 278 K with reaction time. 29Si NMR spectra were analyzed by the multivariate curve resolution. The addition of Al3+ ions to aqueous silicate solution promoted gel formation. Small amounts of Al3+ ions were incorporated as a four-coordinated complex at early stage of polymerization reaction of silicates and during subsequent reaction six-coordinated Al complex increased, suggesting reversible conversion between 4- and 6coordinated complexes. SO42− ions interact with positive surfaces of aluminosilicates and are specifically adsorbed on the surface sites of 6-coordinated Al3+ species, which may be stabilized on silicate surfaces as [Al(H2O)5SO4]+.

■

INTRODUCTION Silicate sol is formed by hydrolysis and polymerization of tetraethyl orthosilicate (TEOS) and alkaline silicate ions.1−3 Monomeric silicate ions react into chains and ring oligomers through condensation reaction of eq 1 to grow up to primary particles 2Si − OH → Si − O − Si +H 2O

SIMPLISMA was applied to Raman spectra in a sol−gel hydrolysis and condensation reaction of tetramethyl orthosilicate.15 When Al3+ ions coexist in the gelation process of aqueous silicate solutions, siloxane bond formation, formation of primary particle and their aggregation, and surface charge of silica gel particles should be modified. It is well-known that Al3+ ions adsorbed on silica particles give positive charges to their surfaces, which increases the point of zero charge (PZC). Such aluminosilicate dispersions were used commercially as Ludox AM.3 When sodium silicate is used for silica gel synthesis, silicas prepared will include trace amount of Al3+ ion to change their properties. Such Al effects on polymerization of silicates also could be observed in the formation process of zeolites.1 Harris et al. reported the incorporation of Al into silicate network in synthesizing ZSM-18 with octahyrohexamethyl benzotripyrrolium: Al reacted with silicate ions through rapid addition reaction and subsequent slow chelation.17 Miyazaki et al. reported the effects of counterions on the reaction between silicate and Al 3+ ions at pH 2. 18 Sulfate ions gave montmorillonite-like aluminosilicate, and chloride and nitrate ions gave a mixture of aluminum hydroxide, silica gel, and montmorillonite-like aluminosilicate. Limited papers, however, report effects of negative ions on Al surface site formation in forming silica network in acidic conditions.

(1)

Recently, reaction mechanisms of polymerization of silicates were discussed by NMR and Raman spectra and ab initio calculation4−6 and NMR spectra and small-angle X-ray scattering.7−10 Many peaks of 29Si NMR spectra are assigned to Qn structure, Si(OH)4−n(OSi)n (n = 0, 1, 2, 3, and 4). However, in a mixed silicate species system it is difficult to estimate the structure of each silicate species, such as linear chains and cyclic origomers, from Qn structure contents. Chemometrics is very useful for analysis of concentration profile of a multicomponent system. In particular, a curve resolution technique gives the component spectra and concentration profiles of reactants, intermediates, and products in a reaction system. By analyzing the 29Si NMR data set of a reaction system using the curve resolution, we can discuss how the silicate structure changes with reaction time.11 Lawton and Sylvestre described a curve resolution method in 1971,12 referred to self-modeling curve resolution (SMCR). Two decades later, SMCR was applied to various fields.13,14 Simple-to-use interactive self-modeling mixture analysis (SIMPLISMA)15 and orthogonal projection approach (OPA)16 were proposed by Windig et al. and Sanchez et al., respectively. © 2016 American Chemical Society

Received: May 21, 2016 Revised: June 25, 2016 Published: June 28, 2016 7079

DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085

Article

Langmuir

Figure 1. Variation of 29Si NMR spectra of aqueous silicate (panel A, n = 0.00) and aluminosilicate solutions (panel B, n = 0.04) with reaction time, tR/min. The reaction was performed at 318 K and NMR spectra were measured at 278 K. The silica gels prepared were crushed into the size of 500 μm and washed with abundant distilled water, which was enough to decrease down to 0.1 mS/cm in electric conductivity. The water content of gel was determined from weight loss after heating at 433.2 K for 2 h. Al(OH)3 (n = 1) also was synthesized from a 19.2 wt % aqueous solution of KAl(SO4)2 (Wako Pure Chemical Industries Ltd.) adjusted with 10 wt % NH3 water to pH 10.0 at 333.2 K. The precipitates were filtered by a filter paper (#5) and washed with distilled water as in washing gels. Water content of gels and Al(OH)3 dried at room temperature was estimated from the weight loss due to heating at 433.2 K for 2 h. NMR Measurements. 27Al and 29Si NMR spectra of silicate sols containing Al3+ ions were measured at 278 K to which the sol was quenched from 318 K, using a JEOL ECA500 (500 MHz) and JEOL JNM-LA400 (400 MHz) NMR spectrometer, respectively. The measurement conditions for 27Al and 29Si NMRs were the 90° pulse of 11.7 and 15.0 μs, repetition delay of 0.4 and 20 s, and scan of 1024 and 512 times, respectively. Aluminum nitrate and tetramethylsilane (TMS) were used as the external references for 27Al and 29Si, respectively. Chemometrics Procedures. 27 The time dependence of peak intensity in 29Si NMR spectra was analyzed by principal component analysis (PCA) in order to determine the number of pure component. MCR-ALS with OPA was applied to the data of 29Si NMR spectra during gelation under the number of pure components. Data matrix X contains spectra of mixed materials at several sampling points. MCRALS decomposes X as eq 3

The role of Al 3+ ions in cement and geopolymer formation19−22 and the diffusion of various negative ions included in soil23−25 were also reported. Because Al and Fe ions in silicate sources for silica synthesis affect the size of silica particle,26 silicate sources should be purified in order to control the size of silica particles. Thus, for the sake of functionalization of silica gels, it is important to elucidate how cations such as Al3+ are incorporated into silica networks and what kind of surface active sites are produced. It is well-known that ettringite is generated on concrete surfaces to break the concrete, when sulfate ions attack concrete structure. Therefore, the control of diffusion of sulfate ions is important for some use of materials with concrete, such as chemical grouting and concrete structure near sea. In this paper, we investigated how Al3+ ions were incorporated into silica networks and fixed at silica surfaces during silica gel formation. 27Al and 29Si NMR spectra of silicate solutions containing aluminum sulfate were measured in the gel formation process. The 29Si NMR spectra were analyzed by chemometrics procedure. Pure component spectra and concentration profiles of mixed aluminosilicate species were estimated with multivariate curve resolution−alternative leastsquares (MCR-ALS) with OPA. Also, we determined surface charge density of silica particles from pH titration and sulfate ion adsorption in order to discuss Al active sites for SO42− ions on silica surfaces.

■

X = CST + E

where S and C contain the pure component spectra and pure concentration profiles, respectively, and E is an error matrix. We found initially S with OPA and resolved NMR data sets, matrix X, with MCR-ALS to obtain matrix C. Potentiometric Titration. The pH titration of gels prepared was carried out in order to determine their surface charge density.28 A 250 mg sample of silica gels in a dry state was added to 0.1 mol dm−3 HNO3 solution of 15 mL. The total volume of solvent was adjusted to 30 mL by the addition of distilled water. The gel suspensions stirred were titrated with 0.15 mol dm−3 NaOH solution at 17 μL/s using a Kyoto Denshi AT-610 titrator. As a reference, 0.1 mol dm−3 HNO3 solution of the same volume also was titrated in the same procedure. Surface charge density of gels was calculated by eq 4

EXPERIMENTAL SECTION

Sample Preparation. A diluted sulfuric acid solution was mixed with an aqueous aluminum sulfate solution (13 wt % in Al2O3; Hokuriku Kasei Industry Co. Ltd.) to make mixtures of Al3+ and SO42− in the range 0−0.9 and 6.0−13.9 mol dm−3, respectively. A 1.1 mol dm−3 silicate solution, prepared by dilution of a standard sodium silicate solution (SiO2 28.9 wt %, Na2O 9.5 wt %; SiO2/Na2O = 3.1 mol/mol; Fuji Chemical Co, Ltd.), was added slowly (0.3 mL/sec) to the mixture with stirring. The pH of the solutions was adjusted to 2.8 using sulfuric acid. The Al content in alminosilicates prepared ranged from n = 0 up to 0.14, where n is the molar fraction of alumina defined by eq 2

n=

[Al 2O3] [SiO2 ] + [Al 2O3]

(3)

Γ=

(2) 7080

C bFv msS

(4) DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085

Article

Langmuir where Cb mol dm−3 is the concentration of the titrant, v dm3 is the difference between titrant volumes giving a same pH in titrated solutions with and without silica gels (ms/g), F is the Faraday constant, and S is the specific surface area of gel. The specific surface area S/m2 g−1 was determined from N2 adsorption isotherm at 77 K of the gel pretreated at 393 K and 1 mPa for 2 h, using a Quantachrome Quadrasorb SI. Sulfate Adsorption. Sulfuric acid of various concentrations of up to 1.2 mol dm−3 was added to the dried silica gel of 250 mg in a polypropylene container. After the container was kept at 293.2 K over 2 h, silica gels were separated by centrifugation and the concentration of its supernatant was examined by the inductively coupled plasma spectrometry with a Shimadzu ICPS-8100. The amount W of SO42− adsorbed on gels was calculated by eq 5

W≈

(VC0 − (V + m w )C) ms

(5)

where C0 and C are the initial and final concentrations of SO42−, and mw and ms are the weights of water and silica contained in the wet gel, respectively.

Figure 2. Pure component spectra obtained using MCR-ALS with OPA for 29Si NMR data set (Figure 1) of silicate (A) and aluminosilicate solutions (B). Components: 1, red; 2, blue; 3, black.

RESULTS AND DISCUSSION Figure 1 shows variations of 29Si NMR spectra of aqueous silicate and aluminosilicate solutions with reaction time. Since it is possible for an Al atom to replace a Si atom, the number (m) of AlO4 tetrahedra bound to a Qn unit is given in parentheses, Qn(mAl); that is, each of m atoms of Al connects to the central Si atom of Qn through an oxygen atom. Each peak is assigned to silicate species Qn and Qn(mAl), as shown in Figure 1.22,29,30 The Q0 peak at the chemical shift of δ = −72 ppm decreased rapidly and the Q1 and Q2 peaks at −81 and −91 ppm also decreased, but gradually. The broad peak of Q4 at −110 ppm increased monotonically. On the other hand, the Q3 peak at −100 ppm increased at the initial stage and over around 50 min decreased through a broad peak that suggests existence of various structural species. In the aluminosilicate system of n = 0.04, Q4(3Al) peak at −89 ppm, and Q4(2Al) peak at −98 ppm appeared obviously even within 5 min, suggesting that Al insertion to siloxane skelton should be rapid. The broad peaks may arise from overlap of peaks from many kinds of silicate species and amorphous silica. For example, the peaks of linear trimer, prismic hexamer, tricyclic hexamer, pentacyclic heptamer, and so forth31 may appear in the region of −88∼−91 ppm, which may be assigned to Q2 peak. When PCA was performed for the data set of 29Si NMR spectra, the eigenvalues of the first three pure components accounted for more than 98% of total variability. In the constraint conditions of three pure components obtained from PCA, the pure component spectra and their profiles were estimated by MCR-ALS with OPA, as shown in Figures 2 and 3. In the silicate system (n = 0.00), pure component 1 having large amount of Q1 (14.1%), Q2 (31%), and Q3 structure (32.5%) may be oligomeric species1 containing many silanol groups and decreased monotonically to change into pure component 2 within about 60 min. The pure component 2 includes less Q1 (6.5%), Q2 (24.7%) than those in the pure component 1, and more Q3 (45.8%) and Q4 structures (23.0%) than those in the component 1. The component 2 should be an intermediate because it increased steeply within 60 min and then decreased slowly. Pure component 3 containing mainly Q4 structures (73.4%), along with Q3 structure (19.2%), may be assigned to microgel,1 which includes many kinds of siloxane bonds such as cyclic species, as suggested by broad Q3 and Q4

Figure 3. Time course of concentrations of pure components obtained using MCR-ALS with OPA for 29Si NMR data set of silicate solutions. Circle and red line, n = 0.00; Square and blue line, n = 0.04.

■

bands. The component 3 increased monotonically with reaction time. The intensity of Q1, Q2, and Q3 peaks in all pure component spectra of the aluminosilicates of n = 0.04 (except for the Q3 structure of the component 1) increased, demonstrating that the coexistence of Al3+ ions promotes silanol group formation. The increment of active silanol groups led to more rapid condensation reaction between them than that of the silicate system, as shown in Figure 3, which indicates more rapid transformations of component 1 to 2 and 2 to 3. Pure component spectrum 3 in the aluminosilicate was similar to that in the silicate of n = 0.00, suggesting that gels as final products should have similar framework even under coexistence of Al3+ ions. The Q4(3Al) peak at −89 ppm, Q4(2Al) peak at −98 ppm, and Q4(1Al) peak at −104 ppm appeared only in pure component spectra 1 of the aluminosilicate. This means that silicate framework incorporated Al3+ ions easily in the early reaction stage to be modified by depression of condensation 7081

DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085

Article

Langmuir

Figure 4. (A) 27Al NMR spectra of silicate and aluminosilicate solutions at various reaction time, tR/min. (B) The part in the region of 0 ppm of panel A is magnified.

reactions between silanol groups. Moreover, no Qn(mAl) peaks in the pure component spectra 2 and 3 suggest that 4coordinated Al3+ species was effused gradually from frameworks in the acid solution, which was confirmed directly by disappearance of the peak at 55 ppm in 27Al NMR spectra with reaction progress. 27 Al NMR spectra of aluminosilicate (n = 0.04) as a function of reaction time are shown in Figure 4. Almost all Al species (>95%) in the reaction solution will be incorporated during the gelation process, which was checked by feeble 27Al NMR peaks of water extracted from aluminosilicate gels. The weak, broad peak at around 55 ppm and two sharp peaks at 0.2 and −3.3 ppm appeared. Each peak is assigned to 4-coordinated Al3+ in silicate frameworks, 6-coordinated Al3+, and 6-coodinated Al3+ with SO42− or [Al(H2O)5SO4]+,32,33 respectively. The area of each peak is normalized by the total area of all peaks and plotted as a function of reaction time, as shown in Figure 5. The amount of 4-coordinated Al3+ decreased gradually with reaction time. On the other hand, the amount of 6-coordinated Al3+ increased, corresponding to the decrease of 4-coordinated Al3+. Since all Al3+ ions in the acidic Al2(SO4)3 solution were 6-

coordinated species, the Al2(SO4)3 solution added to a silicate solution (pH 12) within 5 min should react around basic silicate droplets to form 4-coordinated tetrahedral species. After that, the 4-coordinated species decreased exponentially with reaction time, because the pH of the entire solution changed toward 2.8. The amount of [Al(H2O)5SO4]+ at −3.3 ppm was almost unchanged with gelation. It seems reasonable that the formation ratio of [Al(H2O)5SO4]+ to total Al3+ ions should be constant, because the ratio is proportional to SO 4 2− concentration that is practically constant. Figure 6 shows 27Al NMR spectra of an aluminosilicate gel prepared for 420 min in the same conditions as Figure 4. Almost all Al atoms in an as-synthesized gel from aqueous sulfonic acid solution were located in 6-coordinated sites, as shown by the peaks of 0 and −3.3 ppm (a). Washing the gel with distilled water, almost all Al in the washed gel (b) became 4-coordinated species, as shown by appearance of the peak at around 53 ppm and disappearance of the peaks at 0.2 and −3.3 ppm. Adding 0.52 mol/dm−3 aqueous sulfuric acid solution to the washed gel, the spectrum c was very similar to the spectrum a of as-synthesized gel, except for no peak at 55 ppm. Because there were no Al3+ ions in the sulfuric acid solution, 6coordinated Al species should be formed by the transformation of 4-coordinated species in the washed gel. The formation of 6coordinated species in sulfonic acid may result in low pH. These results indicate that Al atoms in aluminosilicates exist as 4- and 6-coordinated species at neutral and acidic conditions, respectively, and the coordination number can be transformed reversibly. The 6-coordinated Al will be formed at low pH on the surfaces of silicate gels and comprised of 4-coordinated Si species. When primary particles associate with each other, the surface 6-coordinated Al may be incorporated into walls of narrow pores and their boundary, where relatively large SO42− cannot access. On the other hand, some portion remained at outer surfaces of associated particles and adsorb SO42−. The 4coordinated Al3+ formed by displacement of Si4+ will give negative points that become the adsorption sites for cations. Surface charge density of aluminosilicate gels is plotted as a function of pH (Figure 7). The PZC shifted to higher pH with an increase in n. In the pH region of 1.5 to 3, where SO42− adsorption was carried out, the positive surface charge of the gels increased with an increase in n.

Figure 5. Reaction time dependence of the relative area intensity of 27 Al NMR peaks of an aluminosilicate solution (n = 0.04). Black triangle, 55 ppm (4-coordinated Al); blue square, 0.2 ppm (6coordinated Al); red circle, −3.3 ppm (6-coordinated Al with SO42−). 7082

DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085

Article

Langmuir

Figure 6. (A) 27Al NMR spectra of an as-synthesized aluminosilicate gel (tR/min = 420) before (a) and after washing with distilled water (b). The 0.52 mol dm−3 aqueous H2SO4 solution was added to the washed gel (c). (B) The relative intensity is magnified in order to show the peaks at around 55 ppm.

Figure 7. Surface charge density of aluminosilicate gels as a function of pH.

Figure 8. Amounts of SO42− adsorbed on aluminosilicate gels at 298.2 K. The solid lines are fitted by the Langmuir equation. n = [Al2O3]/ ([SiO2] + [Al2O3]): circle, 0.00; square, 0.02; diamond, 0.04, triangle, 0.14, inverted triangle, 1.00.

Figure 8 shows adsorption isotherms for SO42− on aluminosilicate gels at pH 2.8 and 293.2 K. The isotherms were fitted by the solid lines calculated with the Langmuir equation, except for silica gel. The saturation adsorption amounts increased with an increase in n, as expected by the surface charge density, which suggests that 6-coordinated Al sites should work as adsorption sites for SO42−, as shown by the −3.3 ppm peak of 27Al NMR. The adsorption equilibrium constant, a, was almost independent of n (Figure 9), indicating that the active sites for SO42− should be energetically uniform. The empty 6-coordinated Al (at 0.2 ppm) may exist in confined space to which SO42− cannot access. On the other hand, surface 6-coordinated Al sites with SO42− (at −3.3 ppm) are active adsorption sites. Because SO42− adsorption occurred in the Langmuir type, the active site should be filled at high concentration, that is, there are no empty sites. This means that all 6-coordinated Al sites are not adsorption sites for SO42−, because some part of the sites cannot be accessed by SO42−. Although silica surfaces also adsorb SO42−, small amounts of Al3+ ions incorporated into silica surfaces were much more effective for adsorbing SO42− ions, as shown by the Ws/n profile in Figure 9.

Figure 9. Langmuir parameters obtained from the data of SO42− isotherms (Figure 8) on various aluminosilicates. 7083

DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085

Article

Langmuir

■

catalyzed hydrolysis and polycondensation of tetramethylorthosilicate. J. Phys. Chem. C 2011, 115 (22), 11077−11088. (6) Halasz, I.; Kierys, A.; Goworek, J.; Liu, H.; Patterson, R. E. 29Si NMR and Raman Glimpses into the Molecular Structures of Acid and Base Set Silica Gels Obtained from TEOS and Na-Silicate. J. Phys. Chem. C 2011, 115 (50), 24788−24799. (7) Gaboriaud, F.; Nonat, A.; Chaumont, D.; Craievich, A.; Hanquet, B. 29Si NMR and small-angle X-ray scattering studies of the effect of alkaline ions (Li+, Na+, and K+) in silico-alkaline sols. J. Phys. Chem. B 1999, 103 (12), 2091−2099. (8) Gaboriaud, F.; Nonat, A.; Chaumont, D.; Craievich, A. Aggregation and Gel Formation in Basic Silico-Calco-Alkaline Solutions Studied: A SAXS, SANS, and ELS Study. J. Phys. Chem. B 1999, 103 (12), 5775−5781. (9) Gaboriaud, F.; Nonat, A.; Chaumont, D.; Craievich, A. Structural model of gelation processes of a sodium silicate sol destabilized by calcium ions: combination of SAXS and rheological measurements. J. Non-Cryst. Solids 2005, 351 (4), 351−354. (10) Tognonvi, M. T.; Massiot, D.; Lecomte, A.; Rossignol, S.; Bonnet, J. P. Identification of solvated species present in concentrated and dilute sodium silicate solutions by combined 29Si NMR and SAXS studies. J. Colloid Interface Sci. 2010, 352 (2), 309−315. (11) Sasahara, S.; Kaida, K.; Ozeki, S. Chem. Lett. 2016, 45, submission for publication. DOI: 10.1246/cl.160499. (12) Lawton, W. H.; Sylvestre, E. A. Self Modeling Curve Resolution. Technometrics 1971, 13 (3), 617−633. (13) Tauler, R. Multivariate curve resolution applied to second order data. Chemom. Intell. Lab. Syst. 1995, 30 (1), 133−146. (14) de Juan, A.; Maeder, M.; Martinez, M.; Tauler, R. Combining hard- and soft-modeling to solve kinetic problems. Chemom. Intell. Lab. Syst. 2000, 54 (2), 123−141. (15) Windig, W.; Guilment, J. Interactive self-modeling mixture analysis. Anal. Chem. 1991, 63 (14), 1425−1432. (16) Sanchez, F. C.; Khots, M. S.; Massart, D. L.; De Beer, J. O. Algorithm for the assessment of peak purity in liquid chromatography with photodiode-array detection. Anal. Chim. Acta 1994, 285 (1−2), 181−192. (17) Harris, R. K.; Samadi-Maybodi, A.; Smith, W. The incorporation of aluminum into silicate ions in alkaline aqueous solutions, studied by aluminum-27 NMR. Zeolites 1997, 19 (2-3), 147−155. (18) Miyazaki, A.; Yokoyama, T. Effects of Anions on Local Structure of Al and Si in Aluminosilicates. J. Colloid Interface Sci. 1999, 214 (2), 395−399. (19) Bortolotti, V.; Brizi, L.; Brown, R. J. S.; Fantazzini, P.; Mariani, M. Nano and sub-nano multiscale porosity formation and other features revealed by 1H NMR relaxometry during cement hydration. Langmuir 2014, 30 (36), 10871−10877. (20) Rejmak, P.; Dolado, J. S.; Stott, M. J.; Ayuela, A. 29Si NMR in Cement: A Theoretical Study on Calcium Silicate Hydrates. J. Phys. Chem. C 2012, 116 (17), 9755−9761. (21) Urbanova, M.; Kobera, L.; Brus, J. Factor analysis of 27Al MAS NMR spectra for identifying nanocrystalline phases in amorphous geopolymers. Magn. Reson. Chem. 2013, 51 (11), 734−742. (22) Davidovits, J. Geopolymer Chemistry and Applications, 3rd ed.; Geopolymer Institute: Saint-Quentin, France, 2011. (23) Meng, X.; Letterman, R. D. Effect of component oxide interaction on the adsorption properties of mixed oxides. Environ. Sci. Technol. 1993, 27 (5), 970−975. (24) Meng, X.; Letterman, R. D. Modeling ion adsorption on aluminum hydroxide-modified silica. Environ. Sci. Technol. 1993, 27 (9), 1924−1929. (25) Lin, Y. F.; Chen, H. W.; Chang, C. C.; Hung, W. C.; Chiou, C. S. Application of magnetite modified with aluminum/silica to adsorb phosphate in aqueous solution. J. Chem. Technol. Biotechnol. 2011, 86 (11), 1449−1456. (26) Pagliaro, M. Silica-Based Materials for Advanced Chemical Applications; RSC Publishing: London, 2009. (27) Wehrens, R. Chemometrics with R; Springer: New York, 2011.

CONCLUSION Al ions promoted silica gel formation and produced adsorption sites for SO42−. 29Si NMR spectra of aluminosilicates in the formation process were analyzed by chemometrics, MCR-ALS with OPA. Three principal spectra or three kind of aluminosilicate species, such as oligomeric, primary particles, and microgels, were obtained and their formation ratios changed with reaction time (Figure S1). From the results of 27 Al spectra, along with 29Si NMR spectra, Al3+ ions were incorporated into silica skeleton as 4-coordinated structure at the early stage of reaction. As the reaction proceeds, silicate species became the principal species 3 (microgels) having Q3 and Q4 structures. The gel formation reaction was promoted by coexistence of Al3+ ions. The 4-coordinated Al species may be reversibly changed into 6-coordinated species, which may be formed on silica surfaces and pore walls. The introduction of Al to silica gels shifted the PZC to higher pH and thus the positive charge density of aluminosilicate gels increased at the adsorption condition of pH = 2.8. The 6-coordinated Al species adsorbed on gel surfaces may bring about positive sites for SO42− ions to form [Al(H2O)5SO4]+, as indicated in 27Al NMR spectrum. 3+

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.6b01940. Figure S1 showing a schematic model of formation process of aluminosilicate gel.(PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address

§ Shigeo Sasahara, Team21, Fuji Chemical Co. Ltd., 1683-1880 Nasubiga-wa Nakagaito, Nakatsugawa, Gifu, 509-9132, Japan.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Mr. Naoki Ohno (Fuji Chemical Co. Ltd.) for assistance of sulfate adsorption measurements. This research was supported by JSPS Grant-in-Aid for Scientific Research Number 25288003.

■

REFERENCES

(1) Iler, R. K. The Chemistry of Silica; Wiley-Interscience: New York, 1979; pp 220 and 231. (2) Bergna, H. E. The Colloid Chemistry of Silica; American Chemical Society: Washington, DC, 1994. (3) Bergna, H. E.; Roberts, W. O. Colloidal Silica. Fundamentals and Applications; CRC Press: Boca Raton, FL, 2005. (4) Depla, A.; Lesthaeghe, D.; van Erp, T. S.; Aerts, A.; Houthoofd, K.; Fan, F.; Li, C.; Van Speybroeck, V.; Waroquier, M.; Kirschhock, C. E. A.; Martens, J. A. 29Si NMR and UV-Raman Investigation of Initial Oligomerization Reaction Pathways in Acid-Catalyzed Silica Sol-Gel Chemistry. J. Phys. Chem. C 2011, 115 (9), 3562−3571. (5) Depla, A.; Verheyen, E.; Veyfeyken, A.; Van Houteghem, M.; Houthoofd, K.; Van Speybroeck, V.; Waroquier, M.; Kirschhock, C. E. A.; Martens, J. A. UV-Raman and 29Si NMR spectroscopy investigation of the nature of silicate oligomers formed by acid 7084

DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085

Article

Langmuir (28) Hunter, R. J. Introduction to Modern Colloid Science; Oxford Science Publications: New York, 1993. (29) Klinowski, J. Nuclear magnetic resonance studies of zeolites. Prog. Nucl. Magn. Reson. Spectrosc. 1984, 16 (3−4), 237−309. (30) Lippmaa, E.; Magi, M.; Samoson, A.; Engelhardt, G.; Grimmer, A. R. Structural studies of silicates by solid-state high-resolution silicon-29 NMR. J. Am. Chem. Soc. 1980, 102 (15), 4889−4893. (31) McCormick, A. V.; Bell, A. T.; Radke, C. J. Quantitative determination of siliceous species in sodium silicate solutions by 29Si n.m.r, spectroscopy. Zeolites 1987, 7, 183−190. (32) Mason, J. Multinuclear NMR; Plenum Press: New York, 1987. (33) Akitt, J. W.; Farthing, A. Aluminum-27 nuclear magnetic resonance of sulfato complexes of the hexaaquoaluminum ion. J. Chem. Soc., Dalton Trans. 1972, 12, 1226−1229.

7085

DOI: 10.1021/acs.langmuir.6b01940 Langmuir 2016, 32, 7079−7085