Detection of Homogeneous Distribution of Functional Groups in

ARTICLE pubs.acs.org/Langmuir

Detection of Homogeneous Distribution of Functional Groups in Mesoporous Silica by Small Angle Neutron Scattering and in Situ Adsorption of Nitrogen or Water Monir Sharifi,† Roland Marschall,†,‡ Michaela Wilhelm,§ Dirk Wallacher,|| and Michael Wark*,†,‡ †

Institute of Physical Chemistry and Electrochemistry, Leibniz University Hannover, Callinstr. 3A, D-30167 Hannover, Germany Laboratory of Industrial Chemistry, Ruhr University Bochum, Universitaetsstr. 150, D-44801 Bochum, Germany § Ceramic Materials and Components, University of Bremen, Am Biologischen Garten 2/IW3, D-28359 Bremen, Germany Berlin Neutron Scattering Center (BENSC), Helmholtz-Zentrum Berlin f€ur Materialien und Energie GmbH, Hahn-Meitner-Platz 1, D-14109 Berlin, Germany

)

‡

bS Supporting Information ABSTRACT: The distribution of SO3H-functional groups attached to the ordered inner pore walls of mesoporous Si-MCM-41 materials based on SiO2 was investigated by gas adsorption combined with in situ small angle neutron scattering (SANS). The functionalization was performed by two different methods, (i) grafting and (ii) co-condensation. The adsorbates N2 at 77 K or a H2O/D2O mixture of 42:58 at 298 K possess neutron scattering length densities (SLD) similar to that of SiO2 and therefore quench the diffraction signals of the nonmodified silica. SANS measurements show that N2 matches completely not only with the pristine mesoporous Si-MCM-41 but also with Si-MCM-41-SO3H functionalized by grafting. Thus, full access of adsorbate into the entire length of the pores is proven. For the analysis of the distribution of functional groups within the pores in dependence on the used functionalization method, grafting or co-condensation, however, the more specific adsorbate H2O/D2O (42:58) is necessary, because it reacts more sensitively toward small changes in the SLD of the host material. For grafted Si-MCM-41-SO3H materials, an incomplete quenching was observed, indicating that only some regions, probably the pore mouths, have been modified. For a sample functionalized by co-condensation, almost no quenching of the neutron diffraction was found, indicating a very homogeneous distribution of the functional groups along the entire pores.

1. INTRODUCTION Since their discovery 1992, highly ordered mesoporous materials based on SiO2 remain a prominent topic of current research.1,2 Si-MCM-41 in particular, with a two-dimensional hexagonal pore structure, high surface area (∼1000 m2/g), high pore volume, and pore diameters of ca. 3 nm, has found many applications in optical gas sensing, catalysis, separation, and nanotechnology.3
5 However, for many applications, the modification of mesoporous silica materials with organic functional groups is mandatory.6,7 Three general routes are available to incorporate organic moieties into the framework of the mesoporous silica materials: (i) subsequent modification of the pore surface (grafting), (ii) simultaneous condensation of the silica source and the organosilica (cocondensation), and (iii) incorporation of organic components by using bissilylated-organosilica precursors. The latter is not comparable with the first ones since the texture of mesoporous framework obviously differs. Hence, only the grafting and the co-condensation methods are topics in this work and are described in our previous papers in detail.7
9 Both techniques exhibit advantages and disadvantages with respect to functionalization degree, structural changes compared to the pristine material, and the applicability of r 2011 American Chemical Society

the desired organic functional group.10,11 In order to improve the properties of the mesoporous materials sufficiently for the final application, such as the conduction of charge carriers, e.g., protons,7,8,12 an extensive characterization of the designed samples with respect to the homogeneity of the modification is of high importance. Typical analysis methods like nitrogen sorption,13 X-ray diffraction (XRD), or IR spectroscopy14 only give information regarding structure, surface area, and degree of functionalization of the inorganic
organic materials. However, those analysis methods fail regarding the determination of the distribution of functional groups within the pores of the host material. To learn about the distribution of functional groups and the mechanism of adsorption and desorption in porous materials, it is necessary to specify any subsiding process such as micropore filling, capillary condensation, pore blocking effects, surface properties, and interparticle effects. For this reason, it is helpful to implement more advanced techniques. Recently, Gartmann and Br€uhwiler have succeeded in imaging aminopropylsilane with different alkoxy leaving groups Received: October 4, 2010 Revised: March 9, 2011 Published: April 11, 2011 5516

dx.doi.org/10.1021/la2000188 | Langmuir 2011, 27, 5516–5522

Langmuir

ARTICLE

using confocal laser scanning microscopy. They showed that coupling on the outer surface is preferred with increased length of the leaving groups.15 Yokoi et al. made comparable experiments and showed that the adsorption behavior of Co2þ and Fe3þ on mono-, di-, and triamino-alkoxysilane functionalized materials differs if a grafting or co-condensation method was used.16 Another advanced technique is the combination of in situ small-angle neutron (or X-ray) scattering (SANS or SAXS) measurements with isothermal gas adsorption.17 The intensities of the X-ray or neutron scattering reflections of porous solids are altered if the pores are filled with suitable gases. This effect, called contrast matching, was already introduced by Bragg et al. in 1952.18 The degree of matching, however, depends strongly on the chemical nature of the pore walls and, thus, already small changes at the interface between the host material (the adsorbent) and the adsorbate can be directly monitored.19 Smarsly et al. presented a concept to describe the subsequent stages of nitrogen sorption (i.e., micropore filling, formation of nitrogen layers, and capillary condensation) for nonmodified porous silica materials.20 The application of contrast matching in SANS experiments is superior to SAXS because it provides insight into both the sorption mechanism and the structure properties.17
22 Furthermore, there is a multiplicity of opportunities in the choice of the adsorbate (e.g., hydrocarbon, water, benzene, etc.).12,23
25 These adsorbates are favored due to the fact that hydrogen atoms can be isotopically substituted by deuterium atoms and, hence, the coherent scattering length of condensed gas or vapor can be adjusted easily to match that of the solid host material.26 Additionally, the large number of possible adsorbates allows use of a wide temperature range for the measurement. By adsorbing suitable mixtures of H2O/D2O on SiMCM-41 in situ in SANS equipment, Mansour et al. showed that the resulting population of water in the cylindrical pores after desorption at room temperature rules on a radial function with a minimum at the center and a maximum near the pore wall.27

Figure 1. Proton conductivity of SO3H-functionalized Si-MCM-41 using different methods.

However, in the previous studies, the combination of SANS with gas sorption was only used for pristine mesoporous materials;23 there is no study on functionalized mesoporous materials available yet. When the proton conductivity of SO3H-functionalized mesoporous materials is determined, large differences occur in dependence on the functionalization method used, i.e. grafting and cocondensation,9 although the loadings with sulfonic acid groups (∼1.5 mmolHþ/g) are almost identical (Figure 1). The samples synthesized by co-condensation exhibit higher proton conductivities along the whole temperature range (Table 1). The obtained values make them attractive for use as additives in fuel cells membranes, because commercially used Nafion membranes mostly also exhibit proton conductivities around 10
1 S/cm, whereas the conductivity drops down at temperatures above 373 K.28 On the basis of the synthesis procedure and theoretical models, this could be explained by assuming that in the case of co-condensation, the SO3H groups are more homogenously distributed along the pores,8 but an experimental proof is needed. Thus, we investigated functionalized mesoporous materials with in situ SANS experiments combined with N2 sorption at 77 K and H2O/D2O (42:58) sorption at 298 K in order get detailed information regarding pore blocking effects, possible structural changes as well as degree and distribution of the SO3H-groups within the Si-MCM-41 pores.

2. SYNTHESES AND CHARACTERIZATION 2.1. Syntheses. 2.1.1. Synthesis of Si-MCM-41. Following the recipe by Cai et al. a solution of 0.7 mL NaOH (2 M) and 96 mL water were mixed and heated upon 353 K.29 After a clear solution was obtained 0.2 g of CTAB was added and 1 mL tetraethoxyorthosilica (TEOS) was dropped to the reaction mixture. Stirring was continued for 2 h followed by filtration and subsequent washing with water and ethanol. The template was removed via ethanol/ HCl extraction. The final molar composition of the solution was 1200 H2O: 0.31 NaOH: 0.125 CTAB: 1 TEOS. 2.1.2. Functionalization of Si-MCM-41 by Grafting. The template-free host material was placed in a flask and dried under vacuum for 3 h. Under argon atmosphere, 0.5 g Si-MCM-41 powder was suspended in 20 mL dry dichloromethane. The suspension was cooled down to a temperature of 265 K followed by the addition of 3-mercaptopropyltrimethoxysilane (MPMS). Different amounts of functionalization agent can be added to realize different degrees of functionalization;9 in this work 20 mmol (3.76 mL MPMS) per gram host material was used. The reaction mixture was stirred for 24 h followed by filtration and washing with dichloromethane and ethanol. 2.1.3. Functionalization of Si-MCM-41 by Co-condensation. Functionalized Si-MCM-41 samples were prepared by co-condensation as previously described.9 Following the procedure already mentioned for the synthesis of pure Si-MCM-41, a solution of 0.7 mL NaOH (2 M) and 96 mL water was mixed and heated to 353 K. However, 15% of the silica source TEOS was replaced by

Table 1. Texture Properties of Investigated Samplesa sample

functionalization method

Hþ

SBET

pore volume

pore diameter

proton conductivity

(mmol/g)

(m2/g)

(cm3/g)

(nm)

at 140 °C (S/cm)

1108

1.102

2.8

1.95  10
6

828 737

0.615 0.580

2.3 2.1

1.45  10
3 1.12  10
2

Si-MCM-41 Si-MCM-41-SO3H Si-MCM-41-SO3H a

grafting co-condensation

1.4 1.5

The error in the proton conductivity measurements is 8% at maximum. 5517

dx.doi.org/10.1021/la2000188 |Langmuir 2011, 27, 5516–5522

Langmuir MPMS. Thus, 0.86 mL TEOS was dropped into the solution and stirring was performed for 15 min. Then, 0.131 mL MPMS was added to the mixture, and stirring was continued for 105 min. The final molar composition of the solution was 1200 H2O: 0.31 NaOH: 0.125 CTAB: 0.8 TEOS: 0.2 MPMS. 2.1.4. Thiol-Oxidation to SO3H. In order to obtain SO3H-groups the attached SH-groups were oxidized with hydrogen peroxide. 0.3 g of the MPMS functionalized sample prepared by grafting or by co-condensation was suspended in 10 mL of H2O2 solution (30 wt.-%) and stirred for 48 h at room temperature. The powder was filtered and washed with ethanol and water. Then the oxidized solid was suspended in 30 mL of a 2 M H2SO4 solution, stirred for 2 h at room temperature, and finally filtered and washed with ethanol and water. The loading with exchangeable Hþ ions was determined by titration with NaOH (1 M), resulting 1.4 mmolHþ/g for the grafted and 1.5 mmolHþ/g for the co-condensation sample, respectively. 2.2. Characterization/In Situ SANS. Neutron diffraction patterns with in situ gas adsorption were collected at the V1 diffractometer of the Helmholtz Center Berlin, located at a cold neutron source (= λ5.23 Å) with a sample-to-detector distance of 102.52 cm. The scattering intensity was collected in the range q = 0.01
0.7 Å by varying the detector position. The scattering vector q is defined as q = 4π/λ sin(θ), with λ being the wavelength and θ the Bragg angle. Every sample was degassed in vacuum and the measuring cell was assembled in the glovebox. The temperature during the neutron experiments was 77 and 298 K, depending on the used adsorbate, respectively. Before taking a neutron scattering pattern, the sample was filled with a certain amount (n/n0) of adsorptive, by applying a gas adsorption setup that was connected to the measuring cell and enables a direct in situ SANS measurement for each point of the p-V adsorption isotherm. SANS measurements were only performed after a constant cell pressure was obtained which ensures an adsorption equilibration. Nitrogen adsorption isotherms at 77 K were determined on a Quantachrome Autosorb 3 apparatus. Prior to each adsorption measurement, the samples were outgassed at 433 K for 24 h. The Brunauer
Emmett
Teller (BET) method was used to determine the specific surface area. The pore volumes and pore diameters were calculated according Barrett
Joyner
Halenda (BJH) method.30 Water sorption isotherms at 295 K were obtained by volumetric vapor adsorption on a BELSORP 18
3 (Bel Japan Inc.) measurement device using an equilibration time of 500 s. The samples were dried at 395 K in flowing argon for 24 h before measurement.28 To determine the proton conductivities, impedance spectroscopy measurements were performed on a Zahner electrochemical workstation IM6e in a frequency range from 1 to 106 Hz with an oscillating voltage of 100 mV. The powder samples were prepared as described elsewhere.9 For the measurements, the pellets were placed between two thin graphite slices and then put into a holder made of Teflon where the pellet is clamped between two sintered metal electrodes. This holder was finally put in the temperature cell of the impedance spectrometer. The relative humidity (RH) can be adjusted using a connected water tank which can be heated individually.

3. RESULTS AND DISCUSSION The surface areas, pore volumes, and pore diameter of the studied samples were determined from N2 adsorption/desorption isotherms (Figure 2); the obtained values being listed in

ARTICLE

Figure 2. Nitrogen adsorption isotherms and pore diameters (inset) for Si-MCM-41(2), Si-MCM-41-SO3H-1.4 mmolHþ/g (9, grafting), and Si-MCM-41-SO3H-1.5 mmolHþ/g, (b, co-condensation).

Figure 3. Neutron diffraction patterns of Si-MCM-41 at different stages of adsorbed nitrogen.

Table 1. Pristine Si-MCM-41 material exhibits a strong increase in the adsorbed amount at relative pressures of p/p0 = 0.2
0.4 due to the reversible capillary condensation of nitrogen in the ca. 3 nm mesoporous channels.31 The shapes of the isotherms for SiMCM-41 and the grafted Si-MCM-41-SO3H are characteristic for mesoporous materials and are attributed to the type-IV isotherm according to IUPAC classification.32 The values obtained for the pristine Si-MCM-41 with hexagonal pore arrangement are typical for this mesoporous silica material. Due to partial pore filling by organic chains the surface area, pore volume and pore diameter decrease in the Si-MCM-41 modified by grafting compared to pristine host material from 1108 m2/g to 808 m2/g, from 1.102 cm3/g to 0.615 cm3/g, and from 2.8 to 2.3 nm, respectively. However, it should be noted that the decreased pore size after the grafting is not only caused by the organic chains directing into the pores but also by the synthesis conditions (s. 2.1) during grafting, i.e., the treatment with H2O2 and H2SO4. As shown by N2 adsorption isotherms (Supporting Information, Figure S1) even for pure Si-MCM-41 the average pore diameter decreased by 0.34 nm under these conditions. In case of Si-MCM-41-SO3H prepared by co-condensation, the changes in the texture properties are even more pronounced (Table 1). As a consequence of the partial replacement of a part of the TEOS by MPMS, the isotherms show slight 5518

dx.doi.org/10.1021/la2000188 |Langmuir 2011, 27, 5516–5522

Langmuir

ARTICLE

Figure 4. (1) Pore model for Si-MCM-41 at different stages of adsorption, (2) pore model for functionalized Si-MCM-41-SO3H synthesized by grafting, and (3) pore model for functionalized SiMCM-41-SO3H synthesized by co-condensation.

differences in the shape of the hysteresis loops. For the sample with 15% MPMS added to the synthesis mixture, the slope is weaker than for pure Si-MCM-41 due to a less uniform pore distribution. Nevertheless, the mesoporous structure with hexagonal pore ordering is still well developed.9 Figure 3 displays characteristic changes of the neutron scattering intensity for pristine Si-MCM-41 at different stages of nitrogen adsorption. This waterfall-diagram was chosen since no characteristic peak is shifted in position, indicating that the periodic sequence of pore-center-distance remains unchanged during the adsorption. The observed neutron scattering for empty mesopores at p/p0 = 0 documents the hexagonal order of the mesopores, because all characteristic scattering signals [100], [110], [200], and [210] at q = 0.17, 0.27, 0.31, and 0.40 Å, respectively, are present. Complete filling of the pores with nitrogen at p/p0 = 0.869 leads to the disappearance of all the signals. This proves that the scattering length density (SLD) of the Si-MCM-41 framework based on SiO2 and that of nitrogen in the condensed state are quite similar. The signal intensity in SANS I(q) is proportional to the square of the difference of the scattering length densities of the host material F1 and the empty pores F2, respectively. IðqÞ  kðF1
F2 Þ2 ¼ kðΔFÞ2 IðqÞ  kðFsolid Þ2 ,

if

ð1Þ

Fsolid > > Fevacuated, for evacuated pores ð2Þ

where k is a constant that depends on P(q), a form factor of the pores which is determined by the shapes of the individual pores, and S(q) which results from the correlation function of the pores.33 Porous SiO2 and nitrogen as adsorbate at T = 77 K are particularly suitable for contrast matching in SANS measurements, since the scattering length densities of amorphous silica (FSiO2 = 3.43  1010 cm
2) and liquid nitrogen (FN2 = 3.22  1010 cm
2) are almost equal.17 So, for further interpretation, the assumption of a two-phase-system was made, where each phase differs only in its SLD although it might consists of several compounds. Phase one consists of a SiO2 framework and condensed nitrogen (relates to F1 in eq 1) and phase two consists of empty pores (relates to F2 in eq 1), since both phases exhibit a different SLD for all stages of adsorption. As a result of completely filled pores, the reflections of Si-MCM-41 are erased in SANS measurements if no pores remain empty.20 However, also for completely filled pores a diffraction signal can remain, if the SLD of the framework was altered from that of the adsorbate. As will be demonstrated below, this is the case for organically functionalized pore walls.

Figure 5. Neutron diffraction patterns of Si-MCM-41-SO3H synthesized by grafting at different stages of adsorbed nitrogen.

The SANS curves in Figure 3 show some significant specifics in scattering properties while nitrogen is adsorbed. The intensity of the main reflection at q = 0.17 nm
1 increases by 35% at a relative pressure of p/p0 = 0.054. This can be explained by the fact that the filling of voids with condensed nitrogen smoothes the rough inner pore surfaces and some irregularities in the pore wall thickness which might maintain after calcination (Figure 4-1b).24,34 Thus, the periodicity is improved and consequently the intensity of the main reflection increases. Simultaneously, a decrease of the three signals at higher q values is obtained. The signal appearing at q = 0.40 Å disappears already at a very low relative pressure of p/p0 = 0.054, those at q = 0.27 and 0.31 Å decrease with increased amount of adsorbed nitrogen and disappear completely at p/p0 = 0.292. With nitrogen adsorption at p/p0 > 0.26, the intensity I(q) of the [100] reflection, representing the long-range order of the mesopores and thus in general the periodic contrast of wall and pore, drops abrupt by a factor more than 2. This is caused by condensation of nitrogen within the mesopores as shown by the isotherm in Figure 2 and in the pore model in Figure 4-1c/d. As expected, almost no scattering signals are obtained after complete capillary condensation since all pores are filled with nitrogen at a relative pressure of p/p0 = 0.399, which relates to the relative pressure point above the capillary condensation step in the isotherm (Figure 3). Figure 5 shows neutron scattering curves of functionalized mesoporous Si-MCM-41-SO3H synthesized by grafting. The obtained neutron diffraction pattern in vacuum confirms that the mesoporous structure of Si-MCM-41 is not damaged due to the grafting process, since almost all characteristic peaks are visible. Only the fourth signal at higher scattering vector q, which was already found to be most sensitive to changes in the pore walls in pristine Si-MCM-41, is lost indicating an alteration of the SLD of the pores by the introduced propyl-SO3H chains. With progressive N2 adsorption, the sample exhibits analogue results to pristine Si-MCM-41, for example the intensity of the main signal at q ≈ 0.16 Å increases at low adsorption levels,34 whereas simultaneously the signals at higher q values indicating the presence of voids disappear fast. At relative pressures of p/p0 g 0.21 capillary condensation takes place (inset in Figure 5) causing at p/p0 = 0.26 the intensity-drop in the neutron scattering reflection of the main signal. For a better understanding of the distribution of functional groups on the inner surface of the mesoporous material, a mixture of H2O/D2O (42:58) with FH2O/D2O = 3.48  1010 cm
2 was used as a second adsorbate. Hence water adsorption measurements 5519

dx.doi.org/10.1021/la2000188 |Langmuir 2011, 27, 5516–5522

Langmuir

ARTICLE

Figure 7. Neutron diffraction patterns of Si-MCM-41-SO3H synthesized by grafting at different stages of water adsorption.

Figure 6. Water adsorption (filled symbols) and desorption (open symbols) isotherms for Si-MCM-41(2), Si-MCM-41-SO3H-1.4 mmolHþ/g (9, grafting), and Si-MCM-41-SO3H-1.5 mmolHþ/g, (b, co-condensation).

were performed on the samples to ensure pore filling. Figure 6 shows the water adsorption isotherms for pure Si-MCM-41 (2) and SO3H-functionalized Si-MCM-41 synthesized by grafting (9) and by co-condensation (b), respectively. The shape of water isotherms of pure Si-MCM-41 is widely described as type-V isotherm16 according to IUPAC classification,32 which indicates a weak interaction of the water molecules with the siliceous surface particularly at low relative pressure. In spite of the hydrophobic surface characteristic, the adsorbed water amount increases at medium pressure via water
water interaction (H2O cluster formation) followed by capillary condensation in mesopores. The capillary condensation for the investigated materials occurs in a range of 0.35 and 0.5 and depends on the different pore diameters, respectively (Figure 6). The shift of the onset of capillary condensation from a p/p0 value of 0.54 for Si-MCM-41 toward lower relative pressures (0.34 for Si-MCM-41-SO3H modified by grafting and 0.29 modified by co-condensation) can be attributed to the narrowing of the pore width caused by functionalization (Table 1). However, the isotherms of the functionalized grafting (9) and co-condensation samples (b) show both a higher adsorption at low relative pressure of 0.0
0.4 compared to pure Si-MCM-41 (2), which is caused by stronger interactions of water with the functionalized surface and indicates a more hydrophilic nature compared to pristine Si-MCM-41. The shape of the isotherms of the modified samples indicates a transition from a type-V to a type-IV isotherm due to the functionalization altering the surface toward a higher hydrophilicity. Despite a much higher pore volume analyzed by N2 adsorption for pure Si-MCM-41 the increased hydrophilicity of the functionalized samples results in similar maximum water uptake of 54.2% (pure), 50.3% (grafting) and 52.2% (co-condensation) at p/p0 = 1. For the combined adsorption/SANS measurements different amounts of a H2O/D2O (42:58 mixture with FH2O/D2O = 3.48  1010 cm
2) at 298 K were adsorbed and SANS measurements were performed for each adsorption point. Figure 7 shows the diffraction patterns obtained for SO3H-Si-MCM-41-functionalized by grafting at different relative pressures of in situ H2O/ D2O adsorption. Analogous to the nitrogen adsorption experiment, the intensity of the main peak increases at low relative

Figure 8. Neutron diffraction patterns of Si-MCM-41-SO3H synthesized by co-condensation at different stages of water adsorption.

pressures due to the formation of a liquid H2O/D2O film on the slightly rough inner pore surface.34 But this increase is less pronounced, indicating that either smoothing the surface by water adsorption is not as effective as with N2 or the water film that is formed on the surface exhibits some unevenness compared to the condensed nitrogen film. However, the signals at higher scattering of surface functionalized Si-MCM-41 at q = 0.27 and 0.31 Å, respectively, are still present. They only disappear completely when the main channels are filled due to capillary condensation at a relative pressures p/p0 > 0.4 as shown by water adsorption measurements for pristine and SO3H-functionalized MCM-41 materials in previous work (Figure 6). Furthermore, the intensities for q < 0.1 Å increase clearly with increasing water adsorption, due to the presence of liquid water between the interparticle space and the strong incoherent scattering of hydrogen. At relative pressures of p/p0 > 0.42 capillary condensation as shown by water adsorption measurements starts (Figure 6) and causes the drop in intensity for q = 0.16 Å. Further filling with H2O/D2O mixture leads to a decreasing intensity of all signals. It is known from SANS studies that the adsorbing water (42:58 H2O/D2O) seeps into the entire length of the unfunctionalized Si-MCM-41 pores and no voids or central hollow spaces. As a result, all scattering signals of Si-MCM-41 disappear since the pores are filled with matching water (42:58 H2O/D2O).22,23,35 However, in contrast to pure Si-MCM-41, the first scattering signal of a grafted Si-MCM-41-SO3H in Figure 7 does not disappear completely even when the pores are completely filled with water. Hence, the H2O/D2O mixture for which a composition was used that tunes the SLD to be equal to that of SiO235 is not able to match completely, indicating that regions with SLDs different from that of SiO2 are present in the sample. That altered SLD results from the organic propyl chains 5520

dx.doi.org/10.1021/la2000188 |Langmuir 2011, 27, 5516–5522

Langmuir with the SO3H end groups attached to the SiO2 walls. Most of them will be located at the pore opening because during the grafting process, the OH groups sitting there can most easily react with the offered MPMS.35,36 The anchoring of the propyl silane at the pore mouths decreases the width of the pores and, thus, entering the pores is much hindered or even prevented for further silane molecules. Thus, deeper in the pores the degree of functionalization of the walls is quite low (Figure 4-2) and the smaller H2O or D2O, whose kinetic diameter is small enough to pass the partly blocked pore mouths, find many adsorption places on pure SiO2 wall, leading to the obtained almost 90% of neutron scattering matching (Figure 7). Hence, in SANS measurements with in situ gas adsorption, eq 1 does not hold for functionalized SiO2 materials prepared by grafting, since the intensity (I) depends not only from F1 and F2 but also from a third component F3 which reflects the SLD of the functionalized region in the framework. SANS measurements combined with H2O/D2O (42:58) adsorption on a Si-MCM-41 sample functionalized with MPMS by co-condensation, however, document a different situation (Figure 8). Although obviously the quality of the mesoporous structure is slightly decreased compared to the samples functionalized by grafting, because the small diffraction signals have already disappeared in the pattern taken in vacuum, the presence of the strong main signal proves the existence of the mesoporous structure as also confirmed by nitrogen adsorption and X-ray diffraction measurements.28 Again, adsorption of the H2O/D2O molecules first smooth the inner pore walls and cause a slight increase in the intensity of the SANS diffraction signal. Further adsorption and finally complete filling of the mesopores does not lead to any changes in the SANS patterns; the signal intensity remains almost constant. Hence no matching of neutron diffraction was found. This indicates that by co-condensation, the entire inner pore surface is altered by functional groups that are distributed much more homogeneously than in samples functionalized by grafting37 (Figure 4-3), proving the conclusions made from detailed nitrogen adsorption studies and XRD.28 As a result in the whole mesoporous framework, the SLD of the adsorbed H2O/D2O is different from that of the (modified) pore walls impeding contrast matching. Hence, an assumption of a twophase-system, which was made for pure Si-MCM-41 previous, is not correct for Si-MCM-41-SO3H functionalized by co-condensation. If the MPMS is directly added to the synthesis sol and if it is ensured that the hydrolysis rates of the MPMS and the TEOS are almost equal, then the probability for a TEOS to react with the MPMS is the same as to react with another TEOS unit. Thus, no phase separation during the synthesis procedure takes place and a homogeneously altered network results. The synthesis mechanism of the co-condensation reactions has been discussed before by many authors6,8,37,38 but here a new direct experimental evidence for the uniform spatial distribution of the functional groups is given. Only because the SO3H-groups are well-dispersed along the inner pore channels the factor of 10 increase in proton conductivity of the co-condensation sample, compared to the grafted sample (Figure 1), is possible. In the grafted sample, the protons mainly move along the external surface of the particles because they cannot overcome the large distances between adjacent SO3H-groups in the pore channels. In contrast, in the co-condensed samples the SO3H-groups are very close together additionally guiding protons through the pore channels thus leading to much higher proton conductivity especially at low humidities.28

ARTICLE

4. CONCLUSIONS In the present work, small angle neutron scattering experiments combined with in situ adsorption of an N2 at 77 K or H2O/D2O mixture (42:58, at 298 K) were used to describe pristine mesoporous Si-MCM-41 and especially proton conducting SO3H-functionalized Si-MCM-41 prepared by grafting or by co-condensation. The obtained neutron diffraction patterns showed that (i) nonmodified as well as modified Si-MCM-41 possess some slight roughness of the inner surfaces of the pore walls, (ii) N2 accesses into the entire length of the pores, and (iii) N2 matches completely with pristine mesoporous Si-MCM-41 as well as with Si-MCM-41-SO3H functionalized by grafting. In order to get a more detailed insight into the distribution of SO3H-groups within the pore channels and the proton conduction mechanism SANS with in situ H2O/D2O (42:58) adsorption at 298 K was investigated. An observed incomplete quenching for grafted Si-MCM-41-SO3H materials indicates the presence of modified regions, which exist, however, mainly at the pore mouths. This inhomogeneity hinders an effective proton transport. Samples functionalized by co-condensation, however, exhibit almost no quenching, proving a very homogeneous distribution of the SO3H-groups along the channels explaining the higher proton conductivities compared to grafted samples. ’ ASSOCIATED CONTENT

bS

Supporting Information. Nitrogen adsorption isotherms and pore diameters of pure Si-MCM-41 without functional groups before and after H2O2 oxidation and H2SO4 treatment. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail:[email protected].

’ ACKNOWLEDGMENT The work was supported by the Deutsche Forschungsgemeinschaft (DFG) (WA 1116/15, SPP1181). The authors thank Dr. Thomas Hauss (Helmholtz-Zentrum Berlin f€ur Materialien und Energie GmbH, Berlin, Germany) for technical support during the SANS measurements and Prof. J€urgen Caro (Institute of Physical Chemistry, Leibniz University Hannover, Germany) for general support. ’ REFERENCES (1) Cho, M. S.; Choi, H. J. Langmuir 2004, 20, 202–207. (2) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B. J. Am. Chem. Soc. 1992, 114, 10834–10843. (3) Taguchi, A.; Sch€uth, F. Micropor. Mesopor. Mater. 2005, 77, 1–45. (4) Vallet-Regi, M.; Ramila, A.; Del Real, R. P.; Perez-Pariente, J. J. Chem. Mater. 2001, 13, 308–311. (5) Lai, C.; Trewyn, B. G.; Jeftinija, D. M.; Jeftinija, K.; Xu, S.; Jeftinija, S. J. Am. Chem. Soc. 2003, 125, 4451–4459. (6) Sadasivan, S.; Khushalani, D.; Mann, S. J. Mater. Chem. 2003, 13, 10231029. (7) Marschall, R.; Sharifi, M.; Wark, M. Micropor. Mesopor. Mater. 2009, 123, 21–29. (8) Hoffmann, F.; Cornelius, M.; Morell, J.; Fr€oba, M. Angew. Chem., Int. Ed. 2006, 45, 3216–3251. 5521

dx.doi.org/10.1021/la2000188 |Langmuir 2011, 27, 5516–5522

Langmuir

ARTICLE

(9) (a) Marschall, R.; Rathousky , J.; Wark, M. Chem. Mater. 2007, 19, 6401–6405. (b) Marschall, R.; Bannat, I.; Caro, J.; Wark, M. Micropor. Mesopor. Mater. 2007, 99, 190–196. (c) Cavalcanti, W. L.; Marschall, R.; T€olle, P.; K€ohler, C.; Wark, M.; Frauenheim, T. Fuel Cells 2008, 8, 244–253. (10) Ying, J. Y.; Mehnert, C. P.; Wong, M. S. Angew. Chem., Int. Ed. 1999, 38, 57–77. (11) Ziolek, M. Catal. Today 2004, 90, 145–150. (12) Gomes, D.; Marschall, R.; Nunes, S. P.; Wark, M. J. Membr. Sci. 2008, 322, 406–415. (13) Kruk, M.; Jaroniec, M.; Sayari, A. Langmuir 1997, 13, 6267–6273. (14) Kumar, D.; Schumacher, K.; Du Fresne von Hohenesche, C.; Gr€un, M.; Unger, K. K. Colloids Surf., A 2001, 187, 109–116. (15) Gartmann, N.; Br€uhwiler, D. Angew. Chem., Int. Ed. 2009, 48, 6354–6356. (16) Yokoi, T.; Yoshitake, H.; Tatsumi, T. J. Mater. Chem. 2004, 14, 951–958. (17) Ramsay, J. D. F.; Kallus, S. J. Non-Cryst. Solids 2001, 285, 142–147. (18) Bragg, W. L.; Perutz, M. F. Acta Crystallogr. 1952, 5, 277–283. (19) Sel, O.; Brandt, A.; Wallacher, D.; Thommes, M.; Smarsly, B. Langmuir 2007, 23, 4724–4727. (20) Smarsly, B.; G€oltner, C.; Antonietti, M.; Ruland, W.; Hoinkis, E. J. Phys. Chem. B 2001, 105, 831–840. (21) Gethner, J. S. J. Appl. Phys. 1986, 59, 1068–1085. (22) Smarsly, B.; Thommes, M.; Ravikovitch, P. I.; Neimark, A. V. Adsorption 2005, 11, 653–655. (23) Hoinkis, E. Part. Part. Syst. Charact. 2004, 21, 80–100. (24) Tun, Z.; Mason, P. C.; Mansour, F. K.; Peemoeller, H. Langmuir 2002, 18, 975–977. (25) Li, J. C.; Ross, D. K.; Hall, P. L.; Heenan, R. K. Physica B 1989, 156, 185–188. (26) Ramsay, J. D. F.; Wing, G. J. Colloid Interface Sci. 1991, 141, 475–485. (27) Mansour, F. K.; Peemoeller, H. Langmuir 2002, 18, 975–977. (28) (a) Wilhelm, M.; Jeske, M.; Marschall, R.; Cavalcanti, W.; T€olle, P.; K€ohler, C.; Koch, D.; Frauenheim, T.; Grathwohl, G.; Caro, J.; Wark, M. J. Membr. Sci. 2008, 316, 164–175. (b) Marschall, R.; Wark, M.; Jeske, M.; Wilhelm, M.; Grathwohl, G.; Caro, J. Stud. Surf. Sci. Catal. 2007, 170, 1540–1545. (29) Cai, Q.; Luo, Z.-S.; Pang, W.-Q.; Fan, Y.-W.; Chen, X.-H.; Cui, F.-Z. Chem. Mater. 2001, 13, 258–264. (30) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373–376. (31) Mokaya, R. Catal. Lett. 1996, 37, 1891–1894. (32) Kruk, M.; Jaroniec, M. Chem. Mater. 2003, 15, 2942–2947. (33) Li, J. C.; Ross, D. K.; Howe, L. D.; Heeman, R.; Ibel, R. Phys. Rev. B 1994, 49, 5911–5916. (34) Tun, Z.; Mason, P. C. Acta Crystallogr. 2000, 56, 536–539. (35) Glinka, C. J.; Nicol, J. M.; Stucky, G. D.; Ramli, E.; Margolese, D.; Huo, Q.; Higgins, J. B.; Leonowicz, M. E. J. Porous Mater. 1996, 3, 93–98. (36) Br€uhwiler, D. Nanoscale 2010, 2, 887–892. (37) Lim, M. H.; Stein, A. Chem. Mater. 1999, 11, 3285–3294. (38) Huh, S; Wiench, J. W.; Yoo, J.-C-; Pruski, M.; Lin, V.S.-Y. Chem. Mater. 2003, 4, 4247–4256.

5522

dx.doi.org/10.1021/la2000188 |Langmuir 2011, 27, 5516–5522