Journal of Colloid and Interface Science 344 (2010) 603–610

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier .com/locate / jc is

Physicochemical characterization of silylated functionalized materials

Tiago Borrego a, Marta Andrade b, Moisés L. Pinto b,c, Ana Rosa Silva a,1, Ana P. Carvalho b, João Rocha c,Cristina Freire a,*, João Pires b,*

a REQUIMTE/Departamento de Química e Bioquímica, Faculdade de Ciências, Universidade do Porto, Rua do Campo Alegre, 4169-007 Porto, Portugalb Departamento de Química e Bioquímica e CQB, Faculdade de Ciências, Universidade de Lisboa, Ed. C8, Campo Grande, 1749-016 Lisboa, Portugalc Departamento de Química, CICECO, Universidade de Aveiro, 3810-193 Aveiro, Portugal

a r t i c l e i n f o a b s t r a c t

Article history:Received 1 November 2009Accepted 8 January 2010Available online 18 January 2010

Keywords:SilylationMesoporous silicasClaysGraftingOrganosilanes

0021-9797/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.jcis.2010.01.026

* Corresponding authors.E-mail addresses: acfreire@fc.up.pt (C. Freire), jpsil

1 Present address: Laboratório Associado CICECO, UnUniversitário de Santiago, 3810-193 Aveiro, Portugal.

Silylation of several materials where the surface area arises from the internal pores (MCM-41 and FSM-16) or is essentially external (silica gel, and clays) was performed using three organosilanes: (3-amino-propyl)triethoxysilane (APTES), 4-(triethoxysilyl)aniline (TESA) and (3-mercaptopropyl)trimethoxysilane(MPTS). The materials were characterized by nitrogen adsorption–desorption at �196 �C, powder XRD,XPS, bulk chemical analysis, FTIR and 29Si and 13C MAS NMR.

For MCM-41 and FSM-16 the highest amounts of organosilane are obtained for APTES, while for theremaining materials the highest amounts are for MPTS; TESA always anchored with the lowest percent-age. In terms of surface chemical analysis, TESA anchored with the highest contents irrespectively of thematerial, and the opposite is registered for MPTS. Comparison of bulk vs surface contents indicate thatTESA is mainly anchored at the material external surface. Moreover, with N or S (surface and bulk) con-tents expressed per unit of surface area, MCM-41 and FSM-16 (internal porosity) show the lowestamounts of silane; the highest amounts of silane per unit of surface area are obtained for the clays.

Grafting of the organosilanes to the surface hydroxyl groups was corroborated by FTIR and 29Si and 13CMAS NMR. Furthermore, NMR data suggested that TESA and APTES grafted mostly through a bidentateapproach, whereas MPTS grafted by a monodentate mechanism.

� 2010 Elsevier Inc. All rights reserved.

1. Introduction

Porous materials modified with an immobilized organic layer arewidely used as adsorbents, chromatography phases and catalysts[1]. In this type of materials, which can be considered hybrid mate-rials or polymer nanocomposites [2,3], the useful properties arehighly conditioned by the nature of the organic species, type of bondbetween these species and the solid, and the chemical groups thatare then left to interact or, in some cases, covalently react with a thirdchemical entity. This latter aspect can be used in the heterogeniza-tion of metal complexes that are catalytically active for various reac-tions in the homogeneous phase, aiming a more efficient use of thecatalyst, namely by allowing its reuse [4,5]. In fact, the leaching ofthe active phase during the catalytic reaction is still a major draw-back in the immobilization of both, metal complexes or metalphases, in porous supports such as porous silicas [4,5]. A numberof species has been immobilized at silica surfaces, to serve as linkersor scavengers of the actual catalytic phase [1,6–14].

ll rights reserved.

va@fc.ul.pt (J. Pires).iversidade de Aveiro, Campus

The present work studies the funtionalization of varioustypes of silica based mesoporous materials via covalent attach-ment of organosilanes, which is an important step in the prep-aration of materials for the above-mentioned catalyticapplications. We endeavor to highlight the factors that contrib-ute to the more adequate conjugation organosilane-porous sup-port to be used in the heterogenization of metal complexes. Forthis, a systematic study was made using three silanes with ter-minal ANH2 or ASH that are very important reactive groups toanchor catalytic phases: (3-aminopropyl)triethoxysilane (APTES),4-(triethoxysilyl)aniline (TESA) and (3-mercaptopropyl)trimeth-oxysilane (MPTS). As the porous supports, five silica-basedmaterials were used: MCM-41, FSM-16, silica gel (SiO2) andclays. These supports have different porosity: (i) internal, forthe materials with well defined regular mesoporosity, MCM-41and FSM-16; and (ii) external for silica gel and clays. For theclays used in this work – Laponite (LAP) and K10 – there arealso differences in their chemical composition and morphology,particularly in the crystallite sizes. The physicochemical charac-terization of all materials was made by powder X-ray diffraction(XRD), low-temperature nitrogen adsorption, FTIR, X-ray photo-electron spectroscopy (XPS), bulk chemical analysis, 29Si and13C CP MAS NMR.

Si

O

O

O CH3

CH3

H3C

Si

O

O

O CH3

CH3

H3C

Si

O

O

OCH3

CH3

CH3

H2N

H2N

HS

(a)

(b)

(c)

Scheme 1. Species used in the silylation reactions: (a) (3-aminopropyl)triethox-ysilane – APTES; (b) 4-(triethoxysilyl)aniline – TESA; (c) (3-mercaptopropyl)tri-methoxysilane – MPTS.

604 T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610

2. Experimental section

2.1. Materials preparation

MCM-41 was obtained as described in detail elsewhere [15].The synthesis involved the dissolution of CTAB (Aldrich) in deion-ized water and aqueous ammonia (Merck, 25%). The silica sourcewas tetraethyl orthosilicate (Aldrich, 98%). The solid was dried at90 �C and heated at 550 �C for 5 h, after a ramp of 1 �C min�1. Fromthe powder XRD peaks the value 3.78 nm was obtained for the d100

of MCM-41, with a0 parameter considering a hexagonal symmetryof 4.37 nm which is within the range of published values [15]. Thesecond regular mesoporous material, the FSM-16 sample, waskindly given by Dr. Shinji Inagaki, from Toyota Central R&D Labs,and used as received. From the XRD data, and considering the hex-agonal symmetry, the a0 parameter is 4.41 nm slightly higher thanthe value obtained for MCM-41. Silica gel-60 (0.040–0.063 mm)was purchased from Merck.

K10 is a synthetic montmorillonite (from Sigma–Aldrich) withthe average one half unit cell Na0.6K0.12Ca0.02[(Al1.78Fe0.12Mg0.10)VI

(Si3.89Al0.11)IVO10(OH)2]. This clay belongs to the family of so-called2:1 clays [16]. In this solid, and as denoted by its composition, thetetrahedral layers (depicted by the superscript IV) are essentiallymade of silica, only partially substituted by alumina, that make a‘‘sandwich” of the inner layer where Al, Fe and Mg are in octahedralcoordination (denoted by the superscript VI) [16]. LAP is also asynthetic clay but from Laporte Industries Ltd. The formula of itsone half unit cell, as indicated by the supplier, is Na0.7[(Mg2.75Li0.15)VI(Si4)VIO10(OH)2]. This synthetic material is similar to the montmo-rillonite type clays [16], but the octahedral aluminum was substi-tuted by magnesium. Additionally, LAP has regular crystallites ofsmall size, in the range of 25 nm. These crystals are smaller thanthe crystals of K10, which are in the range of 80 nm [17].

The organosilanes used in the silylation reactions, APTES, TESAand MPTS, were from Aldrich with purities of 99, 97 and 95%,respectively. The structures of these molecules are given inScheme 1. Typically, 2 g of material were refluxed with 100 ml ofdry toluene (Aldrich – 99.8%) containing 6 mmol of the organosi-lane for 24 h in argon atmosphere: it is worthwhile mentioningthat the dryness conditions of the reaction media are of key impor-tance to prevent large extension of lateral silane polymerizationwithin the materials and to allow for an efficient silane functional-ization. The resulting materials were purified by Soxhlet extractionwith dry toluene and finally dried in an oven, at 120 �C for 24 h.The functionalized materials will be denoted as ‘material@orga-nosilane’: for example, K10@TESA identifies the material K10 uponfunctionalization with TESA.

2.2. Characterization methods

Low-temperature nitrogen (Air Liquid, 99.999%) adsorption iso-therms were determined in a volumetric automatic apparatus(Quantachrome, Nova 2200e or Micromeritics, ASAP), at �196 �Cusing a liquid nitrogen cryogenic bath. The samples, between 50and 100 mg, were degassed for 2.5 h at a pressure lower than0.133 Pa. The degassing temperature was 150 �C to avoid decom-position of the organic groups.

XRD patterns were collected on a Philips PX 1730 instrumentusing Cu Ka radiation. For the clays, oriented mounts were used.Typical experimental conditions used a time per step of 2.5 s anda step size of 0.005 2h degrees.

Thermogravimetry was performed under helium flux (1 L h�1)with a ramp of 5 �C min�1 in a TG-DSC apparatus, model 111 fromSetaram.

FTIR spectra of the materials were obtained with a Jasco FT/IR Plusspectrophotometer in the 400–4000 cm�1 region, with a resolution

of 4 cm�1 and 32 scans. The samples were diluted with KBr (Aldrich,spectroscopic grade) and studied as pellets.

Nitrogen and sulfur elemental analyses (EA) were performed at‘Laboratório de Análises’, IST, Lisbon, Portugal. XPS was performedat ‘‘Centro de Materiais da Universidade do Porto”, Portugal, in aVG Scientific ESCALAB 200A spectrometer using non-monochro-matic Al Ka radiation (1486.6 eV). The materials were compressedinto pellets prior to the XPS studies. To correct possible deviationscaused by electric charge of the samples, the C 1 s band at 285.0 eVwas taken as internal standard.

Solid state NMR spectra were recorded on a 9.4 T Bruker Avance400P spectrometer. 29Si CP MAS NMR spectra were recorded with4 ls 1H 90� pulses, 8 ms contact time, a spinning rate of 5 kHz,and 4 s recycle delays [18]. 13C CP MAS NMR spectra were recordedwith a 4.5 ls 1H 90� pulse, 2 ms contact time, a spinning rate of 7–9 kHz, and 4 s recycle delays. 29Si and 13C chemical shifts arequoted in parts per million from TMS.

3. Results and discussion

3.1. Nitrogen adsorption and XRD

Fig. 1 shows the nitrogen adsorption–desorption isotherms at�196 �C for the original and the surface functionalized silica and

0

5

10

15

20

25

0.0 0.2 0.4 0.6 0.8 1.0

p/p 0

na (

mm

ol/g

)

SiO2

SiO2@TESA

SiO2@MPTS

SiO2@APTES

(a)

0 5 10 15 20 25

pore diameter (nm)

SiO2

SiO2@TESA

SiO2@MPTS

SiO2@APTES

(b)

(a) (b)

0

5

10

15

20

25

30

35

0.0 0.2 0.4 0.6 0.8 1.0

p/p 0

na (

mm

ol/g

)

FSM-16

FSM-16@TESA

FSM-16@MPTS

FSM-16@APTES

1 2 3 4 5 6

pore diameter (nm)

V/

r (a

.u.)

FSM-16

FSM-16@TESA

FSM-16@MPTS

FSM-16@APTES

δδ

V/

r (a

.u.)

δδ

Fig. 1. (a) Nitrogen adsorption–desorption isotherms at �196 �C, and (b) respective mesopore-size distributions, for the SiO2 and the FSM-16 samples. Open points foradsorption and closed points for desorption.

Table 1Specific surface areas (ABET) and the mesoporous volumes (Vmeso) determined fromthe nitrogen adsorption, at �196 �C. Nitrogen elemental analysis (E.A.) and surfacecontents obtained from XPS for nitrogen, except for the samples with MPTS wherevalues are for sulfur.

Material ABET

(m2 g�1)Vmeso

(cm3 g�1)E.A.(mmol g�1)

XPS(mmol g�1)

SiO2 536 0.83 – –SiO2@TESA 533 0.75 0.5 0.74SiO2@MPTS 381 0.53 1.59 0.54SiO2@APTES 288 0.41 1.23 1.63LAP 378 0.28 – –LAP@TESA 318 0.25 0.25 0.60LAP@MPTS 40 0.03 2.94 1.03LAP@APTES 19 0.02 2.28 1.93K10 223 0.32 – –K10@TESA 168 0.23 0.29 0.70K10@MPTS 80 0.12 1.88 0.96K10@APTES 81 0.16 1.12 2.00MCM-41 1087 0.86 – –MCM-41@TESA 869 0.62 0.36 0.44MCM-41@MPTS 840 0.67 1.50 0.26MCM-41@APTES 330 0.44 1.60 1.49FSM-16 1323 1.18 – –FSM-16@TESA 887 0.69 0.36 0.72FSM-16@MPTS 808 0.54 1.59 0.47FSM-16@APTES 669 0.44 2.14 1.77

T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610 605

FSM-16, as selected examples. The respective mesopore-size distri-butions are also given. The data for the remaining samples are gi-ven in Fig. S1 of the Supporting Information. Mesopore-sizedistributions were obtained from the low-temperature nitrogenadsorption data using the Broekhoff–de Boer method with theFrenkel–Halsey–Hill equation (BdB–FHH) [19]. This method givespore-size distributions more reliable than the more currently usedBarret–Joyner–Halenda (BJH) method [20] and the results obtainedgive mesopore-size distributions that match well compared withthose obtained by more elaborated methods based on the densityfunctional theory [21]. The values for the specific surface areas(ABET) were obtained in the relative pressure range between 0.05and 0.15 and are registered in Table 1 for all materials. The porousvolumes given in this table are the mesoporous volumes (Vmeso) ofthe respective samples since the t-method analysis [22] did notshow the presence of relevant microporosity in the samples. There-fore, Vmeso were estimated from the amounts adsorbed at the rela-tive pressure of 0.95 [22].

The mesoporous nature of the studied materials is clearly evidentfrom the shape of the nitrogen isotherms in Fig. 1 and S1. In the caseof the SiO2, the isotherms are of type IV according to the IUPAC clas-sification [22,23] and the maximum in the mesopore-size distribu-tion of the initial sample is centered at 7 nm. This value remainsalmost unchanged after functionalization, although in the case ofthe MPTS based material, the distribution of pore sizes is broader.The fact that SiO2 has the widest pores of the series of materials stud-ied it is most probably the cause for showing the lowest relative de-crease in the specific surface after functionalization.

In the case of the parent clays – LAP and K10 – (Fig. S1) the ob-served mesoporosity comes from the aggregation of the finely di-vided clay particles, and is not due to the expansion of the claylayers. In this sense, the type IV [23] isotherm obtained for the par-ent LAP is due to the mesoporosity from the particle–particleaggregation and is centered at 3 nm. Upon functionalization, the

mesopore-size distribution widens and changes to higher values,although with MPTS the maximum that was observed for the par-ent LAP is partially retained. The XRD patterns for the LAP samples(Fig. 2) do not allow a clear determination of the interlayer spacing,that is the d001 value [24], due to the broad peaks that are observed.This was previously reported in the literature and is due to the lowdimension of the LAP crystallites, which results in this materialforming partially disordered aggregates through edge-to-edge

2 4 6 8 10 12

2θ /º

inte

nsit

y (a

.u.)

LAP

LAP@TESA

LAP@APTES

(a)

2 4 6 8 10

2θ /º

inte

nsit

y (a

.u.)

K10

K10@TESA

K10@APTES

(b)

Fig. 2. X-ray powder diffraction patterns of original and functionalized materials:(a) LAP and (b) K10.

0

2

4

6

TE

SA

MPT

S

APT

ES

-OH

TE

SA

MPT

S

APT

ES

-OH

TE

SA

MPT

S

APT

ES

-OH

TE

SA

MPT

S

APT

ES

-OH

TE

SA

MPT

S

APT

ES

-OH

SiO2 LAP K10 MCM-41 FSM-16

sila

ne m

olec

ules

or

OH

gro

ups

/ nm

2 Bulk XPS -OH

Fig. 3. Bulk and surface nitrogen (for the TESA and APTES based materials) andsulfur (for the MPTS based materials) contents, or OH groups, expressed per unit ofsurface area of the respective non-functionalized material.

606 T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610

and edge-to-face interactions [25–27]. Nevertheless, Fig. 2a con-firms qualitatively the above discussion on the mesopore-size dis-tribution of LAP, since the peaks in the diffractograms of thefunctionalized materials are shifted to lower angles (higher inter-layer distances). Upon the funtionalization with APTES the LAPparent clay particle organization is disrupted almost completely,as shown by the low specific surface area of LAP@APTES (Table1), which is more characteristic of a non-porous material.

The nitrogen isotherms of K10 (Fig. S1) can not be classified pre-cisely as type IV [23] since they do not show a definite plateau inthe adsorbed amounts as the relative pressure approaches 1. Thecurves may tentatively be classified as a mixed between types IIand IV, since they do show hysteresis that denotes the presenceof mesoporosity. The mesopore-size distribution of the parentK10 presents a small maximum near 3 nm and a more significantone around 5 nm, that is, for higher values than that found forthe parent LAP. As in the latter, the porosity of K10 comes mainlyfrom the aggregation of the fine clay particles. The difference be-tween the maxima in the mesopore sizes for LAP and K10 maybe related with the differences in the sizes of their crystallites,since the K10 clay is less finely divided than LAP. As in LAP, thefunctionalization with the silanes decreases the specific surfacearea of K10 (Table 1) and the mesopore-size distributions areshifted to high values, particularly in the samples with TESA andAPTES. The XRD of the parent K10 (Fig. 2b) shows a peak at 2h de-grees of 8.9 which is commonly observed for the montmorillonitetype clays to which K10 belongs, being ascribed to the d001 reflec-tion, that is, to the basal spacing of the clay. Upon functionalizationa shoulder is noticed in the diffractograms for lower 2h values,indicating a possible partial intercalation of the silane moleculesin the interlayer region of the K10 clay.

In the case of the MCM-41 based materials, with the exception ofthe sample modified with APTES, the nitrogen adsorption–desorp-tion isotherms (Fig. S1) do not present a hysteresis loop and aretherefore classified of type IVc. Nitrogen adsorption isotherms oftype IVc were previously reported for MCM-41 samples with lowmesopore sizes [28], and reflect the existence of uniform near-cylin-drical pores [22]. The same classification can be applied to the nitro-

gen isotherms in the FSM-16 based materials (Fig. 1). Uponfuntionalization with the silanes, and similarly to the other materi-als, the specific surface areas and the porous volumes of MCM-41and FSM-16 decrease. Contrarily to what occurred for the clay mate-rials, the mesopore-size distributions are shifted to the lowest valuesfor the regular mesoporous silicas (Fig. 1 and S1). In fact, the values ofthe maxima initially centered near 3 nm for the parent FSM-16 andMCM-41 show the highest reduction for FSM-16@APTES (2.2 nm). Inthe case of the MCM-41@APTES, the shape of the isotherm is quitedifferent from the remaining and can be classified as type IV [23].According to the mesopore-size distribution (Fig. S1), this materialdoes not present pores within the expected values for the MCM-41materials: the mesopore-size distribution is flat for pore sizes near3 nm and below, and only for values higher than 4 nm, pores seemto exist. This can be related to a high degree of pore blocking and par-tial disruption of the structure of MCM-41 after the reaction withAPTES. The XRD data of the various MCM-41 samples (Fig. S2, Sup-porting Information,) show that the MCM-41@APTES is the onlysample of the MCM-41 series that presents clear differences to theinitial material. Additionally, and as discussed below, the 29Si CP/MAS NMR spectrum of MCM-41@APTES also presents the more sig-nificant changes in relation to the parent material. The reason whythe APTES funtionalization of FSM-16 does not have a similar effecton its porous structure can be related with the higher stability of theFSM-16 structure relatively to that of MCM-41 [29] as referred in theliterature [14].

3.2. Chemical analysis of N and S

The results for the bulk and the surface chemical analysis, nitro-gen for TESA and APTES samples, and sulfur, for the MPTS samples,are given in Table 1. From N bulk analysis (EA), it is possible to seethat the amounts of grafted TESA are the lowest. For the materialswhere the available surface comes from the internal porosity(MCM-41 and FSM-16) the highest amounts of silane are obtainedfor APTES, while for the remaining materials the highest measuredamounts are for MPTS. In terms of surface analysis (XPS data) it canbe concluded from Table 1 that all the materials with TESA presentthe highest amounts of N, the opposite situation being registeredfor MPTS. Most probably, the highest amounts at the surface ob-served for TESA can be due to pore diffusion constrains imposedby its high rigidity induced by the aromatic ring. Furthermore,APTES shows an intermediate behavior between the cases of TESAand MPTS.

In Fig. 3 the amounts of N or S (bulk and surface chemical anal-ysis) are plotted for the various materials after being expressed bythe specific surface area of the respective initial material (note that

250026002700280029003000

wavenumber (cm-1)

Tra

nsm

itanc

e (a

.u.)

FSM-16

FSM-16@TESA

FSM-16@MPTS

FSM-16@APTES

(a)

250026002700280029003000

wavenumber (cm-1)T

rans

mita

nce

(a.u

.)

TESA

MPTS

APTES

(b)

Fig. 4. FTIR spectra of: (a) the parent and functionalized FSM-16 and (b) theorganosilane species.

T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610 607

each N or S atom corresponds to one silane molecule and that1 lmol m�2 corresponds to 0.6 molecules�nm�2). This figure allowsthe interpretation of the data in terms of the differences of the sur-face chemistry and texture of the parent supports. It is clearly no-ticed that in the materials where the porosity is essentially internalthe amounts of silane are the lowest. Furthermore, the highestamounts of silane per area are obtained for the clays. This resultis corroborated by the already mentioned shoulder in the X-ray dif-fractogram (Fig. 2), that appears for lowest angles in the function-alized samples, suggesting at least partial intercalation of thesilane molecules within the clay.

The degree of functionalization depends not only on the avail-able surface, but also on the quantity of reactive surface OH groups.A first approach to estimate the density of OH groups (OH contentper area) at the surface in silica-based materials can be made fromthermogravimetric analysis by the weight loss for temperaturesabove 200 �C, since below this temperature the mass decrease ismainly from the physisorbed water [30]. Examples of the thermo-gravimetric data obtained are presented in the Supporting Infor-mation, Fig. S3, and the values of the density of OH groupsexpressed by number of OH groups per nm2, are presented inFig. 3. For SiO2, MCM-41 and FSM-16 materials, the values comparewell with the maximum number of grafted APTES and MPTS, indi-cating that in these cases almost all available OH groups were usedin the grafting process, whereas for TESA the number of graftedmolecules is lower than the estimated OH groups. As discussedabove, the amounts obtained for TESA by chemical analysis werealso the lowest of the series, suggesting that the anchoring of TESAis more difficult than for MPTS and APTES, due to steric hindranceinduced by its high rigidity, as referred above. In the case of theclays the amounts of grafted APTES and MPTS are higher thanthe OH density (Fig. 3), particularly for LAP. Therefore, in the caseof the clays some silane molecules are stabilized within the claylayers or polymerized with other silane molecules that are cova-lently bonded to the surface.

3.3. FTIR and NMR

The FTIR spectra of the various materials display the characteris-tic strong and broad band in the 1250–1000 cm�1 region assigned tothe asymmetric stretching modes of Si–O–Si and the band near800 cm�1 assigned to the symmetric stretching mode [31–33]. Theyalso show a broad band in the range 3750–3000 cm�1, due to OHstretching vibrations from silanol groups and adsorbed water anda band in the range 1650–1630 cm�1 due to the OH bending from ad-sorbed water [34]. Upon silylation the FTIR spectra of all materialsshow bands characteristic of the grafted organosilanes, typically inthe regions 3000–2500 and 1650–1300 cm�1. The FTIR spectra ofFSM-16 based materials in the high energy range are given in Fig. 4as an example. The alkane chains present in the organosilanes resultin the appearance of typical bands at 2940–2930 and 2870–2860 cm�1, attributed to asymmetric and symmetric CH2 stretch-ings, respectively; for TESA the stretching vibrations characteristicof the aromatic CAH (3000–3100 cm�1) are not seen probably be-cause they are overlapped with the strong OH stretching vibrations.In the region of 1650–1630 cm�1 (not shown) the band due to OHbending from adsorbed water is still present in all materials but withlower intensity. In the case of TESA based materials, the latter bandhas some shoulders in the low energy side due to the partial overlapwith vibrations modes of para-substituted benzene ring (1600 and1607 cm�1) and NH2 bending vibrations (1625 and 1510 cm�1).For APTES and MPTS based materials, the bands corresponding tothe CH2 bending vibrations are also detected with low intensity inthe range 1470–1400 cm�1 (not shown). Furthermore, for APTESbased materials, there is a band in the range 1560–1520 cm�1 thatcan be tentatively assigned to NH2 bending vibrations, but in the

MPTS based materials it was not possible to detect, as usual, thebands due to SH stretching in the range 2590–2570 cm�1.

Although the FTIR data showed the presence of grafted silanesin the various materials, 29Si and 13C CP MAS NMR can give addi-tional information on surface silicon environments and attachedsilanes, specifically in their grafting mechanisms.

The 29Si CP MAS NMR spectra are shown in Fig. 5a, for SiO2 andclays and Fig. 5b for the regular mesoporous silicas MCM-41 andFSM-16; peaks are identified according to the commonly used Qn

and Tn notation [35] by the vertical broken lines. The precise peakpositions corresponding to these vertical lines and the Q2/Q4 andQ3/Q4 peak area ratios obtained after deconvolutionwith three Gauss-ian peaks are presented in the Supporting Information, Table S1.

In general, the Q4, Q3 and Q2 resonances of the siliceous matriceswere detected (except for LAP as discussed in more detail below). Itshould be emphasized that in the context of silane functionaliza-tion, the Q3 and Q2 sites have a major importance since they pro-vide the OH groups (one or two, respectively) for silane grafting.

For SiO2, MCM-41 and FSM-16 samples the Q2 signals areclearly weaker than the Q4 and Q3, in accordance with the fact thatfew geminal OH groups (Q2) are expected in this type of solids [34].Because the Q4 sites do not have OH groups and, thus, they do notreact with the silane molecules, the area of this resonance wasused as an internal reference for normalizing the areas of the othersites of each material. In this way, the decrease of the Q2/Q4 and Q3/Q4 ratios relatively to the ratios obtained for the parent materialsindicates which type of silanol, geminal (Q2) or single (Q3) site, re-acts preferably with a given silane. Although the individual popu-lation ratios obtained using 29Si CP MAS are not the ‘real’ valuesthat would be obtained with MAS, we are only interested in theirevolution upon silylation, for a given class of materials.

-150-100-500

δ(ppm)

-150-100-500

δ(ppm)

-150-100-500

δ(ppm)

SiO2

SiO2+APTES

SiO2+TESA

SiO2+MPTS

Q4

Q3

Q2

T3

T2

T1

LAP

LAP+APTES

LAP+MPTS

LAP+TESAT

3T

2T

1

Q3

Q2

K10

K10+APTES

K10+MPTS

K10+TESAT

3T

2T

1

Q4

Q3

Q2

-150-100-500

δ(ppm)

-150-100-500

δ(ppm)

MCM-41

MCM-41+APTES

MCM-41+MPTS

MCM-41+TESAT

3T

2T

1

Q4

Q3

Q2

FSM-16

FSM-16+APTES

FSM-16+MPTS

FSM-16+TESAT

3T

2T

1

Q4

Q3

Q2

(a) (b)

Fig. 5. 29Si CP MAS NMR spectra of (a) SiO2 and the clays (LAP and K10) samples and (b) hexagonal mesoporous silicas MCM-41 and FSM-16.

608 T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610

The decrease of the Q2/Q4 and Q3/Q4 ratios is compared to theinitial ratio (parent sample) which is considered to have the start-ing value of 100%; this can not be done for the clay materials sincethese materials have no Q4 sites (see below) [3,35]. As can be seenfrom Fig. 6, the decrease in the ratios is lowest for TESA and highestfor APTES, in accordance with the results discussed in Sub-section3.2. The decrease of the ratios Q2/Q4 and Q3/Q4 is different for thethree porous materials, but is highest for the former one, that is,the geminal OH will react preferably, presumably because the Q2

sites are more at the surface [35] and, thus, more exposed tosilylation.

29Si CP MAS NMR spectrum of the parent LAP (Fig. 5) showsessentially a single 29Si peak, in accordance with its natural claycounterpart hectorite [3], because all silicon sites in the clay havethe same Q3 environment [3,35].

0

25

50

75

100

TE

SA

MPT

S

APT

ES

TE

SA

MPT

S

APT

ES

TE

SA

MPT

S

APT

ES

SiO2 MCM-41 FSM-16

Are

a ra

tio (

%)

Q2/Q4 Q3/Q4

Fig. 6. Area ratios for the Q3 and Q4 signals of the 29Si CP/MAS NMR spectra.

The very low intensity signal attributed to Q2 in LAP, is not usu-ally detected in the natural hectorite [3]. The LAP gives this peakprobably due to the very small dimensions of the crystallites whichconsiderably increases the number of surface terminal OH groups,but the presence of a low amount of impurities from the synthesiscan not be entirely ruled out. In the case of the K10, Q2 and Q3 sig-nals are detected in the range of chemical shifts reported in the lit-erature for other 2:1 type clays [3,35–37]. The low quality of theK10 spectrum is attributed to the presence of Fe(III) in the frame-work. The Q4 signal was detected but not expected since in clayminerals the tetrahedral Si is not connected to four other tetrahe-dral Si [36] but a similar result was already reported in the litera-ture for other synthetic clays and interpreted in terms of impuritiesof amorphous silica from the synthesis procedure [37].

Upon silylation, T1, T2 and T3 sites, from the reacted silane mole-cules, were detected in the 29Si CP MAS NMR spectra [35,38] (Fig. 5).For TESA and APTES, the signals are essentially of T2 and T3 type, sug-gesting that these molecules have reacted in a bi- or tri-dentate man-ner, respectively. Nevertheless, the surface grafting mechanism mayhave occurred mostly through a bidentate chelation because higherintense peaks are observed for T2 (these results are corroborated bythe fact that Q2 show the highest decreases, see above) and since thepresence of T3 sites may also correspond to bidentate grafting andsome lateral condensation of grafted silanes. The same may also ap-ply to some T2 sites that may correspond to monodentate silanegrafting and lateral condensation [10].

Only in the case of MPTS, T1 sites were detected along with T2

and T3. The different behavior between MPTS and the other silanessuggests a higher tendency for monodentate grafting. This may bedue to the steric hindrance for methoxyl derivative to act as abidentate ligand, due to its smaller alkyl chain. Similarly to theother organosilanes, some T2 and T3 sites may also be related to lat-eral condensation.

050100150200

δ(ppm)

FSM-16

MCM-41

K10

LAP

SiO2TESA

0102030405060

δ(ppm)

FSM-16

MCM-41

MPTS

020406080

δ(ppm)

FSM-16

MCM-41

K10

LAP

SiO2

APTES

Fig. 7. 13C CP MAS NMR for the functionalized samples.

T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610 609

The 13C CP MAS NMR spectra for the functionalized samplesare shown in Fig. 7, for each type of functional group. The mate-rials with TESA display three aromatic ring resonances at 145,136, 114 ppm, confirming the presence of the aniline group. Inthe case of the clays (K10 and LAP) these peaks are faint, dueto the lowest absolute amounts of TESA on these samples (Table1) and the partial intercalation of this molecule between the claylayers as shown by XRD (Fig. 2). The strong peaks at ca. 15 and57 ppm are due to ACH3 and AOCH2A carbons, respectively, andshow that the ethoxyl groups did not react completely in allmaterials.

For MPTS, the spectra show three peaks for MCM-41 at 48, 26,9 ppm attributed to ACH2ASH, ACH2A and ACH2ASiA carbons,confirming the presence of the thiopropyl chain. For FSM-16 onlypeaks at 26 and 9 ppm are detected. This observation can not beentirely understood at present, but it may be related with the unu-sual reactivity of MPTS in FSM-16 also detected in 29Si CP MASNMR, which shows a Q3/Q4 ratio quite different from that of othersolids. Materials with APTES exhibit three peaks at 43, 26, 9 ppmattributed to, respectively, the ACH2ANH2, ACH2A and ACH2ASiAcarbons of the aminopropyl group [39] confirming the presence ofthe organic part of the molecule. Small peaks at 15 and 57 ppm dueto, respectively, ACH3 and AOCH2A are also observed in MCM-41and FSM-16, showing that the ethoxyl groups did not react com-pletely. No 13C NMR spectra were obtained for SiO2, LAP andK10-based materials since enough sample was not available.

The dashed lines in Fig. 7 indicate the estimated chemical shiftsfor the organic groups of the anchors, using the GAUSSIAN 03 soft-ware [40]. The models used were the aminopropylsilicate,NH2C3H6SiO3, aminophenylsilicate, NH2C6H5SiO3, and thioprop-ylsilicate, SHC3H6SiO3, connected to a Si3O3(OH)3 ring by SiAObonds. Geometry optimizations have been computed using thenonlocal hybrid three parameters B3LYP density functional ap-proach [41,42] and the 6-31G basis set. NMR shielding tensors of

the optimized geometries have been computed with the Gauge-Independent Atomic Orbital (GIAO) method [43] using the B3LYPdensity functional and the 6-311 + G(2d,p) basis set. The calcula-tion was centered on the amine (or thiol) free side of the anchorand not on the alkoxide part, since the latter does not raise majordoubts on the assignment of the peaks. The observed deviationsbetween the calculated and experimental values clearly indicatethat the models used in the calculations are oversimplificationsof real solids. In fact, the models only considered tridentate graft-ing to the silica structure, whereas the experimental results dis-cussed above indicated a considerable fraction of bidentate andmonodentate bonding. Nevertheless, the possibility of the forma-tion of charged groups (ANHþ3 or ASHþ2 ) in the organic part usingthe same models and methods was investigated. The calculatedchemical shifts for the MPTS (50, 29 and 16 ppm) and APTES (58,28 and 12 ppm) cases agree less with the experimental values thanthe shifts of the non-charged species and, therefore, do not supportthe presence of charges on the amine or thiol groups. However, forthe TESA case, this agreement is better (calculated: 147, 134 and124 ppm; experimental: 145, 136, 114 ppm), indicating that a par-tial positive charge may be present in the amine group due tohydrogen bonding effects.

4. Conclusions

Materials with different pore structures (internal versus exter-nal porosity) were successfully functionalized with amino (APTESand TESA) and mercapto (MPTS) silanes. The results showed a rela-tion between the silane grafting extension and the type of materialand silane structure; a correlation between the grafting mecha-nism and the organosilane was also detected.

MCM-41 and FSM-16 (mainly internal porosity) anchored thehighest APTES amounts (N bulk content), whereas silica and theclays (mainly external porosity), were the best at anchoring MPTSwith the highest amounts; TESA was anchored in the lowestamounts for all materials. Surface analysis of N and S revealed thatTESA anchored with the highest percentages for all materials, andthe reverse situation was observed for MPTS. Combination of bulkand surface results indicates that TESA anchoring occurred prefer-ably at the external surface. When N or S contents are expressedper unit of specific surface area, MCM-41 and FSM-16 showedthe lowest amounts of organosilane, while the highest amountsof silane per area were obtained for the clays.

Grafting of organosilanes to silanol groups was supported byFTIR and 29Si and 13C CP MAS NMR which suggested a surface graft-ing mechanism involving mostly bidentate chelation for APTES andTESA, whereas for MPTS, a higher tendency to grafting by a mono-dentate mechanism was detected.

Finally, the prepared silylated materials have free ANH2 and SHgroups that can be used to immobilize metal complexes throughdirect coordination to the metal centers or through reaction withreactive groups within the ligands. In this context, this work sug-gests that the materials whose surface area is due to externalporosity may be the best supports for the immobilization of bulkymetal complexes with catalytic properties because they allow ahigher content of grafted organosilanes and a high degree of struc-tural flexibility for the catalyst, a key aspect for chiral catalysts inenantioselective catalytic reactions.

Acknowledgements

This work was funded by FCT Fundação para a Ciência e aTecnologia (FCT) and FEDER, through the project PPCDT/CTM/56192/2004. Moisés L. Pinto acknowledges FCT for a post-doctoralgrant (BPD/26559/2006).

610 T. Borrego et al. / Journal of Colloid and Interface Science 344 (2010) 603–610

Appendix A. Supplementary material

Nitrogen adsorption-desorption isotherms at �196 �C, andrespective mesopore size distributions, for the clays and MCM-41; XRD for the MCM-41 samples; thermogravimetric data; chem-ical shifts of the 29Si CP MAS NMR spectra. Supplementary dataassociated with this article can be found, in the online version, atdoi:10.1016/j.jcis.2010.01.026.

References

[1] S.A. Alekseev, V.N. Zaitsev, J. Fraissard, Chem. Mater. 18 (2006) 1987.[2] A.M. Shanmugharaj, K.Y. Rhee, H.R. Sung, J. Colloid Interface Sci. 298 (2006)

854.[3] S.S. Hou, F.L. Beyer, K. Schmidt-Rohr, Solid State Nucl. Mag. 22 (2002) 110.[4] C.M. Crudden, M. Sateesh, R. Lewis, J. Am. Chem. Soc. 127 (2005) 10045.[5] C.E. Song, S. Lee, Chem. Rev. 102 (2002) 3495.[6] R. Ishii, Y. Shinohara, J. Mater. Chem. 15 (2005) 551.[7] P.A. Wheeler, J.Z. Wang, J. Baker, L.J. Mathias, Chem. Mater. 17 (2005) 312.[8] T. Yokoi, H. Yoshitake, T. Yamada, Y. Kubota, T. Tatsumi, J. Mater. Chem. 16

(2006) 1125.[9] J. Madejova, Vib. Spectrosc. 31 (2003) 1.

[10] H.P. He, J. Duchet, J. Galy, J.F. Gerard, J. Colloid Interface Sci. 288 (2005) 171.[11] G. Schottner, Chem. Mater. 13 (2001) 3422.[12] K.J. Shea, D.A. Loy, Chem. Mater. 13 (2001) 3306.[13] N.N. Herrera, J.M. Letoffe, J.P. Reymond, E. Bourgeat-Lami, J. Mater. Chem. 15

(2005) 863.[14] P. Das, A.R. Silva, A.P. Carvalho, J. Pires, C. Freire, Catal. Lett. 129 (2009) 367.[15] M. Grün, K.K. Unger, A. Matsumoto, K. Tsutsumi, in: B. McEnaney, T.J. Mays, J.

Rouquérol, F. Rodriguez-Reinoso, K.S.W. Sing, K.K. Unger (Eds.), Characteri-sation of Porous Solids IV, Royal Society of Chemistry, London, 1997, p. 81.

[16] R.E. Grim, Clay Mineralogy, McGraw-Hill, New York, 1968.[17] C.-C. Chou, M. -Li Chiang, J.-J. Lin, Macromol. Rapid Commun. 26 (2005) 1841.[18] J. Rocha, J. Klinowski, J. Magn. Reson. 90 (1990) 567.[19] W.W. Lukens Jr., P. Schmidt-Winkel, D. Zhao, J. Feng, G.D. Stucky, Langmuir 15

(1999) 5403.[20] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373.[21] J. Pires, A. Carvalho, M. Pinto, J. Rocha, J. Porous Mater. 13 (2006) 107.[22] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids,

Academic Press, San Diego, CA, 1999.[23] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.

Siemieniewska, Pure Appl. Chem. 57 (1985) 603.

[24] B. Velde, Introduction to Clay Minerals Chemistry, Origin Uses andEnvironmental Significance, first ed., Chapman & Hall, London, 1992.

[25] I.K. Biernacka, A.R. Silva, A.P. Carvalho, J. Pires, C. Freire, Langmuir 21 (2005)10825.

[26] M.L. Occelli, S.D. Landau, T.J. Pinnavaia, J. Catal. 90 (1984) 256.[27] P. Cool, E.F. Vansant, Microporous Mater. 6 (1996) 27.[28] R. Schmidt, M. Stöcker, E. Hansen, D. Akporiaye, O.H. Ellestad, Microporous

Mater. 3 (1995) 443.[29] N. Igarashi, K.K. Koyano, Y. Tanaka, S. Nakata, K. Hashimoto, T. Tasumi,

Microporous Mesoporous Mater. 59 (2003) 43.[30] L. Peng, W. Qisui, L. Xi, Z. Chaocan, Colloid Surf. A: Physicochem. Eng. Aspects

334 (2009) 112.[31] F.L. Galeener, Phys. Rev. B 19 (1979) 4292.[32] A. Fidalgo, L. Ilharco, Chem. Eur. J. 10 (2004) 392.[33] J. Pires, M. Pinto, J. Estella, J.C. Echeverría, J. Colloid Interface Sci. 317 (2008)

206.[34] X.S. Zhao, G.Q. Lu, A.K. Whittaker, G.J. Millar, H.Y. Zhu, J. Phys. Chem. B 101

(1997) 6525.[35] G. Engelhardt, D. Michel, High-Resolution Solid State NMR of Silicates and

Zeolites, John Wiley & Sons Ltd., New York, 1987.[36] S.K. Lausen, H. Lindgreen, H.J. Jakobsen, N.C. Nielsen, Am. Mineral. 84 (1999)

1433.[37] K.S.M. Corrêa, R.B. Bernini, M.C.S. Mattos, M.R.M.P. Aguiar, A.W.S. Guarino, J.

Braz. Chem. Soc. 18 (2007) 1509.[38] M. Geppi, S. Borsacchi, G. Mollica, C.A. Veracini, Appl. Spectrosc. Rev. 44 (2009)

1.[39] S. Ek, E.I. Iiskola, L. Niinisto, J. Vaittinen, T.T. Pakkanen, A. Root, J. Phys. Chem. B

108 (2004) 11454.[40] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,

Jr., J.A. Montgomery, T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar,J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A.Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox,H.P. Hratchian, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E.Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y.Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S.Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K.Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J.Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L.Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M.Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A.Pople, Gaussian 03, Revision B.03 ed., Gaussian Inc., Wallingford, CT, 2003.

[41] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.[42] C.T. Lee, W.T. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.[43] K. Wolinski, J.F. Hinton, P. Pulay, J. Am. Chem. Soc. 112 (1990) 8251.