Structural Changes of Hierarchically Nanoporous Organosilica/Silica Hybrid Materials by Pseudomorphic Transformation

Abstract Herein, it is reported how pseudomorphic transformation of divinylbenzene (DVB)‐bridged organosilica@controlled pore glasses (CPG) offers the possibility to generate hierarchically porous organosilica/silica hybrid materials. CPG is utilized to provide granular shape/size and macroporosity and the macropores of the CPG is impregnated with organosilica phase, forming hybrid system. By subsequent pseudomorphic transformation, an ordered mesopore phase is generated while maintaining the granular shape and macroporosity of the CPG. Surface areas and mesopore sizes in the hierarchical structure are tunable by the choice of the surfactant and transformation time. Two‐dimensional magic angle spinning (MAS) NMR spectroscopy demonstrated that micellar‐templating affects both organosilica and silica phases and pseudomorphic transformation induces phase transition. A double‐layer structure of separate organosilica and silica layers is established for the impregnated material, while a single monophase consisting of randomly distributed T and Q silicon species at the molecular level is identified for the pseudomorphic transformed materials.


Introduction
Controlled pore glasses (CPGs) and ordered mesoporous silicas, which are formedb yt he liquid-crystal template mechanism and often referred to as micellar-templated silicas (MTS), are two classes of porous silica-based materials that are both established in numerous fields of applications, for example,c atalysis, [1,2] chromatography, [3,4] gas adsorption [5,6] or enzyme immobilization. [7,8] CPGs can be synthesized in versatile shapes from monoliths to granules in the micrometerrange. Thermally induced phase separation in combinationw ith selective leaching of alkali borosilicate glasses yields am onomodals pongelike pore system with narrow pore size distribution adjustable between 4-1000nm. [9,10] MTS are prepared following a common sol-gela pproach in the presence of as urfactant, where well-structured mesopores between approximately 3-10 nm are generated. [11][12][13] Yet, the synthesis of MTS is usually limitedt op roducingp owders as other morphologies are especially challenging to obtain. Hence, Galarneau and co-workers founda nalternative way for post-synthetic micellar induced pore generation in silica materials of different morphology. [14] This so-calledp seudomorphic transformation (gr. pseudos = false, morphe = form)p rocess is based on the dissolution of amorphous silica phase in basic media and its recondensation in the presence of as urfactant. Under optimal conditions these competing reactions occur in spatialp roximityg enerating mesoporous network, while the initial morphology of the silica phase can be preserved. [14][15][16] Enke et al. and Grandjean et al. adapted the process of the pseudomorphic transformation to CPGs with different initial pore diameters ranging from meso to macroscales. [17][18][19] In the case of mesoporousC PGs, the initialp ores ystem collapses during the transformation,b ut a hierarchically porous silica materialc an be generated from macroporous CPGs. [20] The mesoporesw hich are subsequently generated within the CPG pore walls are thus accessible through the macropores,a lthough swelling of the transformed materialo ccurs whichr educes the pore diameter of the initial CPG. [20,21] Furthermore, different concepts for adjustment of the transformation degree have been discussed, for example, by changing the transformation time, the pH value or the ratio of surfactant to silica, intending to preservet he initial pore system. [17,22] The impact of hierarchical porosity on variousa pplications was recently reviewed. [23] Moreover,S chwieger et al. gave ag ood overviewo nw hat makesu pahierarchical material and how hierarchical systemsc an be classified. [24] Hierarchy does not only mean pores of differentsizes but also implies interplayb etween the different pore levels that improves the flow for example, the distribution of ag as or al iquid.
The incorporationo fo rganic compounds in the silica matrix expands the range of possible applicationsa nd makes the synthesis of materials with application-adjusted features possible.
Most common examples are grafting or co-condensation of silylated organic compounds. Particularly,i nt he case of small molecules, high organosilica content and good distribution between organic moiety and silica speciesa re feasible by co-condensation. [25] In addition, organosilicas that contain up to 80 wt %o fo rganicf unction and possess variousp orosity features can be obtained from bridgeds ilsesquisiloxanes (O 1.5 Si-R-SiO 1.5 ). [26] The adaption of the micelle templating process leads to periodic mesoporous organosilicas (PMOs), which are ac lass of materials presenting both molecular scale periodicity and mesoscale ordering of the pores. These organosilica provide very diversec hemicalf unctionality but again limitedp article size and shape just as in the case of MTS. Thus, the research has been restricted to mainly nanoparticles. [27,28] Therefore, broadening the possible range of particlem orphologya nd size as well as chemical functionality in porous organosilica has been an important issue. However,o nly af ew studies have been reported up to now.R eber and Brühwiler added 3-aminopropyltriethoxysilane (APTES) during the pseudomorphic transformation of silica particles and hence,c ombined the pseudomorphic transformation with ac o-condensation process. [29] Our group has utilizedt he pseudomorphic transformationo fo rganosilica materials with different morphologies, namely of inverse opals and millimeter-sized organosilica beads. [30,31] In this work, we combine the morphological flexibility and the macroporosity of CPGs with the chemical functionality of organosilica materials, to produce hierarchically porous organosilica/silica hybrid materials. We chose the precursor 1,4-((E)-2'-bis(triethoxysilyl)vinyl)benzene (BTEVB) to incorporate chemical functionality since it has been well established for PMO synthesis for several years. [32][33][34] Moreover,v ariousf unctionalized divinylbenzene (DVB)-bridged analogues,f or example, with an amino function [35] or fluorine, [36] are also available, which makes the concept extensible for variousf unctionalities. In order to generate hierarchically porous structures, we adapt the concept of pseudomorphic transformation to the impregnated organosilica/silica hybrid material( DVB-bridged sili-ca@CPG). Using physisorption (suitable forc haracterizing meso-a nd micropores)a nd mercury intrusion porosimetry (MIP,f or characterizing macropores), pore sizes and pore volumes before and after the pseudomorphic transformation have been determined,t op rove the formation of hierarchically porouss ystems. SEM and TEM images have been obtained to follow the changes in morphology and porosity feature. However,i ti ss till not well understood whether the pseudomorphic transformationa ffects both silica and organosilica species and how the pseudomorphic transformation influences the interconnectivityo ft he silica and the organosilica phase. The chemicali dentitieso ft he newly formed mesopore surfaces are not clear as well. Solid state NMR spectroscopy has provena powerful technique to study chemical/morphological structure and interfacial interaction of variouss ilicate compounds. [37][38][39][40][41] Thus, we utilized 2-dimensional(2D) 29 Si{ 1 H} and 13 C{ 1 H} heteronuclear correlation (HETCOR)m agic angle spinning (MAS) NMR experiments to investigate the distribution of the organosilica and the silica phases in the hybrid materiala nd the interactions of both Si speciesw ith the surfactant during the micelle templating processo ft he pseudomorphic transformation. Based on the NMR results, structural model for organosilica/ silica hybrid materialr esulting from pseudomorphic transformation will be proposed.T he presence of nanopores interconnected with macropores in one particlec an provide high flow of fluids and gas, good accessibility to the specific active sites and retention of reactants at the catalytic sites. Therefore, our new approach to form hierarchically porouso rganosilica/silica hybrid materials can be applied to variousf ields of technologies from energy storage or catalysis to separation process. [24]

Results and Discussion
The synthesis process forh ierarchically porouso rganosilica/silicas is subdivided into 1) the impregnationo ft he CPGs with a solutiono ft he organosilica precursor and its subsequentc ondensation and 2) the pseudomorphic transformation of the product. We explored this two-step synthetic protocol since one step process, that is, addition of DVB-precursor in basic aqueous media to the CPGs in the presence of surfactants did not lead to one hierarchically nanoporouss tructure but to the formation of as eparateo rganosilica particle. Hence, we found out that ac ovalentb ondingo fo rganosilica to the silica support is prerequisite before the pseudomorphic transformation. Figure 1g ives an overview of the synthetic concept and possi- Figure 1. Diagram of the two-step synthesis (impregnationa nd subsequent pseudomorphic transformation)o fo rganosilica/silica hybrid materials. After pseudomorphic transformation, either multilayers tructure composedo ft wo separate phaseso fo rganosilica (blue)a nd silica (gray)o ro ne mixed phase (light blue) can be formed. ble structures. The distributiona nd interconnection of the organosilica and the silica species in one mixed or two separate phases after the pseudomorphic transformation is one of the key questions and will be investigated intensively by 2D MAS NMR experiments at the end of this work.

Impregnation of CPG silica with an organosilica phase
In order to combine CPG and DVB-organosilica, the macroporous system of CPG granules was impregnated with aB TEVBcontaining solution via incipient wetness. The formation of a homogeneous organosilica layer inside the CPG pore system was assumed, where the layer thickness shouldb ea djustable via the organosilica content in the impregnations olution. The BTEVB content in the toluene-based impregnations olution was varied between 10-75 wt %, and the resulting impregnated materials are denoted as 10BTEVB, 25BTEVB, 50BTEVB and 75BTEVB where the numbers represent the weight percentage (wt %) of BTEVB in the impregnations olution. The SEM images of the impregnated samples in Figure 2g ive af irst insight about the general trend of pore wall thickness. An SEM image of the initial CPG is given in Figure S1. Due to carbon sputtering of the samples, SEM images are not suitable to estimate the pore size in the materials, however,t hickening of the pore walls with increasing organosilica content in the impregnation solutions is clearly visible.
Quantitative information aboutm acroporous media can be obtained from mercury intrusion porosimetry( MIP). The pore diameter distributions and the textural data of the impregnated samples are shown in Figure 3a and Ta ble 1, respectively (for intrusion curvess ee the FigureS2). With increasing BTEVB content in the impregnation solution, the macropore diameter maximaa sw ella st he pore volumes decrease. For example, the pore diameter and the pore volume are reduced from 168 nm and 1.5 cm 3 g À1 (initialC PG) to 132 nm and 0.70 cm 3 g À1 for 50 BTEVB. The pore diameter of 75BTEVB is not significantly reduced in comparison to 50BTEVB which indicates that impregnationw ith highero rganosilica content causesb locking of the pores rather than forming at hicker organosilica layer.I ta ppears that the impregnationi sl ess homogeneous with increasing organosilica content in the impregnation solution. With the pore volumesa nd the weight differences before and after the impregnation, the filling degree could be calculated (see details in the Supporting Information). 75BTEVB has the highest filling degree of 33 %a nd 25BTEVB and 50BTEVB show the similar filling degree of 21-22%.I nt he case of 10BTEVB the pore volume increases slightly after the impregnationw hich might be due to washing-out of the finely dispersed silica residues or etchingo ft he silica phase during the harsh alkaline treatment. Nonetheless, shifts of the pore size towards smaller values indicatet hat divinylbenzene( DVB)organosilica is successfully incorporated into the macropores of CPG. The values of the filling degrees suggest that, although incipient wetness was used, 100 %f illing is impossible to achieve. This is, on the one hand, due to the change in the density of the organosilicap hase after condensation and might, on the other hand, indicate that the organosilica layer is also on the outer surface of the granules.
To characterizem esoporosity,t he N 2 physisorption isotherms of the materials after the impregnation have been acquired as shown in Figure 3b.T he isotherm of the initial CPG is type II as expected for am acroporous material, characterizedb yu nrestricted monolayer-multilayer adsorption with an almostl inear middle section. [43] The isotherms of 10-75BTEVBs hows an additional steep gas uptake in the lowr elative pressure range  which indicates the formation of micro-or small mesopores in the impregnatedm aterials (the isothermsa re ac ombination of type Ib and type II). Thet ype Ib characteristics showingg as adsorptioni nt he low-pressure range are more distinct with increasingo rganosilica content. Apparent specific BET surface area also increases significantly with increasing organosilica content in the impregnations olution from 137 to 596 m 2 ·g À1 (Table 1). Thep ore diameter distributions of the hybrid materials 10-75 BTEVB, calculated with aN LDFT kernel, are shown in Figure 3c.T he formation of micropores/small mesopores is notable in all samples, exhibitingi ncrease of mesopore volume upon increasing organosilica content. Here, note that the mesopore volumeo fn ewly formed pores is different from the pore volume of initial macroporous system of CPG. We define the mesopore volume( V meso )a st he pore volume originating from the pores smaller than 20 nm to avoid confusion. For condensation of the organosilica phase and its covalent bonding to the silica surface of the CPG, the materials were treated with as olution of ethanol and NaOH at 80 8C. The same treatment of pure BTEVB precursor was found to generate microporous organosilicas with very high surfacea reas up to 1300 m 2 g À1 ,w hereas the pure CPG is nearly unaffected by such treatment (see physisorption isotherms of both phases after treatment in the Supporting Information Figure S3 and S4). Likely,t he additional micropores/smallm esopores of the impregnated materialr esults from an anoporous organosilica layer.
The organosilica contento ft he impregnated samples can be estimated from the mass loss in TG analysis. As can be seen in the TGA plots in Figure S5, the main mass loss occurs in the temperature range of 300-800 8C, where breakdown of organic moieties and formation of pure silica phase are assumed. The mass losses and the calculated organosilica contentso ft he samples are given in Table 1. Therei salinear relation between the specific BET surface area of the impregnated materials and the DVB-organosilica content, confirming that the porosity is dominantly influenced by the organosilica content.
The results show that the impregnationo fm acroporous CPG with BTEVB-containing solutions and subsequentc ondensation is an effective methodf or the synthesis of organosilica/ silica hybrid materials.I tc ombines the flexible morphologyo f the CPG, in this case granulate, with the chemical flexibility of organosilicas. Due to the formation of an anoporous organosilica layer inside the initial CPG pore system, ah ierarchical pore structure is obtained, where the organosilica layer is easily accessible. Hence, this material is of interest in versatile applications for example, for gas adsorption. However,c ertain limitation still exists due to the very small pore sizes of the nanopore systems.

Pseudomorphic transformation of impregnated CPG
It has been shown that the pseudomorphic transformation can be used to generate second homogeneous mesopores in CPG (pure silica phase). [18,[20][21][22] On the contrary,the effect of pseudomorphict ransformation on the organosilica phase and the difference in the reactionk inetics between silica and organosilica phases are not well known. Therefore, we aimt oi nvestigate the influenceo fp seudomorphic transformation on the organosilica phase and to explore the feasibility of improving the porosity features of organosilica/CPG hybrid materials by pseudomorphict ransformation. 50BTEVB was chosen exemplarily and two types of surfactants with different aliphatic chain lengths (C 16 TAOH andC 10 TAB) were utilized to demonstrate the tunability of the pore diameter of the second mesopore system. 50BTEVB was treated with C 16 TAOH, which combines the surfactanta nd the hydroxide ion in one molecule. This enables the synthesis of sodium free materials which are hydrothermally more stable. [17,45] The transformation time was varied between 4days (50BTEVB_4 d_C 16 TAOH) and 6hours (50BTEVB_ 6h_C 16 TAOH) to examine the adjustable transformation degree over time. [18] The pseudomorphic transformation was also carried out with C 10 TABi nN aOH solutionf or 24 hours (50BTEVB_ 24 h_C 10 TAB). As shown in SEM images (Figure 4), overall, the morphology of organosilica/CPG hybrid materials is preserved after the pseudomorphic transformation.H owever, swelling of the materiala nd shrinking of the initial pores are also noted. Significant swelling is observable after four days of reaction time (50BTEVB_4 d_C 16 TAOH), but it is less distinctive with shorter reaction time (50BTEVB_6 h_C 16 TAOH) and negligible in the case of smaller surfactant (50BTEVB_24h_C 10 TAB). This can be confirmed by the pore diameterd istributions from MIP, showni nF igure 5a (the intrusion curves are given in Figure S2). The pore diameter maximaa sw ell as the pore volumes of macropores decrease significantly after pseudomorphic transformation with C 16 TAOH solution:5 0BTEVB_4 d_ C 16 TAOH exhibits highest reduction to 100 nm and 0.43 cm 3 g À1 ,r espectively,b ut for 50BTEVB_6 h_C 16 TAOH with [a] ApparentB ET specific surface area from N 2 physisorption (determined considering an adjusted relativep ressure range according to the IUPAC recommendation). [44] [b] Mesopore volumesf rom N 2 physisorption (calculated with NLDFT kernelf or silicaw ith cylindrical pores from the adsorption branch).
lower transformation degree the macropore diameter decreases less (110 nm). In the case of pseudomorphic transformation with C 10 TAB, the macropore diameterd istribution is nearly unchanged compared to the sample before the pseudomorphic transformation (50BTEVB). Hence, the transformation with C 10 TABc auses lesss welling of the materialw hich might be due to either the formation of smaller mesopores or incomplete transformation of the hybrid material. The textural data are summarized in Table 2.
To investigate the change in the mesoporosity caused by pseudomorphic transformation, N 2 and Ar physisorption measurements are carriedo ut. The N 2 physisorption data of the samples treated with C 16 -surfactants show type IVb isotherms with as teep step due to pore condensationa tarelative pressure of 0.3-0.5 and no hysteresis (Figure 5b). The specific BET surfacea rea increases from 476 m 2 g À1 beforet ransformation to 1007 m 2 g À1 and 835 m 2 g À1 after 4days and 6hours transformation, respectively (Table 2). In both cases, mesopores of 4.2 nm in diameter are generated, as shown in the pore diameter distributionsi nF igure 5c.I na ddition, the powder X-ray diffraction patterns (p-XRD, Figure 6) of both materials show one broad reflection at 2q = 1.98 (d = 4.7 nm) which indicatess uc-cessfult emplating with the surfactant, although the low resolution of the p-XRD patterns do not allow assignment to as pecific phase for example, an MCM-41-types tructure. Thisi sc onsistent with the results of pseudomorphic transformed silica phases in the literature. [46,47] The product treated with the surfactantw ith shorter aliphatic chain length (50BTEVB_24h_ C 10 TAB) shows meso-a nd microporosity and was thus characterizedwith Ar physisorption (Figure 5b). The isotherm exhibits high gas uptake at low relative pressure( type Ib) and transits to as teep step due to capillary condensation in small mesopores.H ence, the isotherm is ac ombination of types Ib and IVb. In comparison to the materialb efore pseudomorphic transformation,i ncreased specific BET surface area of S BET = 668 m 2 g À1 is obtained( following the IUPAC recommendation in the Supporting Information). [44] Hysteresis occurs which can be due to either interparticle cavitation or desorptionf rom some of the large initial pores through the newly formed mesopores.S imilarc avitation effects have also been observed in pseudomorphic transformedC PGs before. [16,17] The shape of the isotherms indicates the presence of small mesopores and micropores, as can be seen in the pore diameter distribution with two maximaat0 .6 nm and 2.9 nm (Figure 5c). In addition, p-XRD patterns indicate formation of smaller mesoporesw hen C 10 TABs urfactant is used in comparison to C 16 TAOH ( Figure 6).  and Ar physisorption isotherms (77 Kf or N 2 showni ncircles,87Kfor Ar shown in triangles) and (c) pore diameter distributions of 50BTEVB before and after pseudomorphic transformation with C 16 TAOH and C 10 TABs urfactants deducedf rom physisorption (NLDFT kernel from adsorption branchf or silica with cylindrical pores in case of N 2 and for silica/zeolites in case of Ar physisorption).The transformeds amplesa re denoteda s5 0BTEVB_6 h_C 16 TAOH, 50BTEVB_4 d_C 16 TAOH,and 50BTEVB_24 h_ C 10 TABa ccording to the surfactant and transformation time. These results are consistentw ith our expectation, demonstrating that the surfactant with shorter aliphatic chain length generates smaller mesopores and thus, the size of mesoporesc an be tuned by the choice of the surfactants for pseudomorphic transformation.
Additionally,t he presence of macro-and mesopores in the hybrid materiala fter pseudomorphic transformationi sc onfirmed by TEM, as shown exemplaryf or 50BTEVB_24 h_C 10 TAB in Figure 7. To obtain TEM images, the materialw as ground and thus, only fragments of the initial granulates are observed. Figure 7a shows af ragment with one macropore, exhibiting rough surface and irregular pore shape due to the transformation process. TheT EM images indicates full transformationo f the complete material and homogeneous distribution of mesoporousn etwork (Figure 7b). This image was convertedi nto a diffraction pattern by fast Fourier transformation (FFT,F igure 7c), wherein aperiodic distance of around 3nmcan be detected.T hisi si ng ood agreement with the results for the pore diameter (2.9 nm) and the d-value from pXRD (3.9 nm) for this material.F or TEM images of the initial CPG and 50BTEVB see Figure S6.

Chemical structure changes by pseudomorphic transformation
Chemical structure and the structural stabilityo fo rganic bridge and siloxane linkageb efore and after pseudomorphic transformation are examined by 13 Ca nd 29 Si CP MAS NMR spectroscopy. Figure 8a comparest he 13 CCPMAS NMR spectra of 50BTEVB, 50BTEVB_6 h_C 16 TAOH and 50BTEVB_4 d_C 16 TAOH with the liquid-state NMRo ft he precursorm olecule BTEVB in CDCl 3 .T he spectra are in good agreement with each other,i ndicatingt he integrity of the organic bridge after the treatments despite the harsh alkaline conditions. Signals between 150-120ppm result from the divinyl-benzene bridge. The signals at 17 and 59 ppm are assigned to the ethoxy group bound to partially condensed siloxane species (ex, Q 3 ,Q 2 ,Q 1 , T 2 ,T 1 ). Additional weak signals between 10 and 80 ppm are observed for transformed samples, which can be assigned to the aliphatic chain of the surfactant. The intensity of these additional signals is very low,i mplying that the amount of the remaining surfactant after extraction is not significant.   The 29 Si CP MAS NMR spectra (Figure 8b)give semi-quantitative information about the condensation degree and structural integrity of siloxane network. Contraryt oC PG presenting only Qs ignals of SiO 2 ,t he impregnated samples exhibit both T (RSi(O) 3 ) [48] and Q( Si(O) 4 ) [49] signals as expected.I np ure organosilica materials, the integrity of the SiÀCb ond can be provenb yt he absence of Qs ignals. However,t his cannot be appliedt ot he organosilica/silica hybrid materials since Qs pecies is already present due to the silica specieso fC PG. In the case of 50BTEVB, T 1 (RSi(OSi)(OR') 2, R' = Ho rC H 3 CH 2 ) , T 2 (RSi(O-Si) 2 (OR')) and T 3 (RSi(OSi) 3 )s peciesa re observed at À59, À67 and À77 ppm. [32,33] After the pseudomorphic transformation, the T 1 signal disappears and relative signal intensity of T 3 over T 2 speciesi ncreases, which indicates increase in the condensation degree and formationo fh ighly networked organosilica structure.I ta ppears that the mechanical stability of the organosilica phase can benefit from the transformation process. Additionally,s ignificant change is seen in Qs pecies. Before the pseudomorphic transformation, the ratio of Q 3 (Si(OSi) 3 (OR')) to Q 4 (Si(OSi) 4 )s ignals is nearly 1:1, whereas after transformation, the intensity of Q 3 signal increases significantly and becomes dominant. This behavior can also be observedf or pure CPG after pseudomorphic transformation,i mplying that the silica phase is affected by the transformation process as well (Figure S8). The spectra of 50BTEVB_6 h_C 16 TAOH and5 0BTEVB_ 4d_C 16 TAOH are very similar.
In the impregnateds amples, the mass loss in TG was used to evaluate the organosilica content in the hybrid materials, but after the transformation, this method turned out to be insufficiently reliable (TG/MS data in Figure S9): the mass loss in the considered temperature range is higher (20-21 %) than before the transformation( 16 %), which might be due to the presenceo fr emaining surfactant and increase in ethoxyg roup as suggested in 13 Ca nd 29 Si MAS NMR. Instead, 29 Si MAS NMR with direct excitation can be used for reliable quantification of the different silica species. Note that CP MAS NMR results do not provide quantitative information because signal intensity dependsonthe functional group and mobility of the molecule. Since direct excitation method is time consuming, only 50BTEVB and 50BTEVB_6 h_C 16 TAOH were measured. The spectral deconvolution of signals and the quantitative results can be found in Figure S10 and Ta ble S3. DVB-organosilica content in the hybrid material can be deduced from the intensity ratio of T/Q signals, showing decrease of organosilica content from 16 to 8.1 mol %a fter the transformation.T his reduction can be due to leaching of the organosilica phase during the transformation or cleavage of the SiÀCb ond. Partial cleavage of the SiÀCb ond can occur,l eading to higher ratios of Qs ignals and lower ratios of Ts ignals at the same time. Other examples in the literature also suggestt hat the small fraction of SiÀCb ond undergoes cleavage and formation into Qs pecies after pseudomorphic transformation of organosilica materials. [30,31] Overall, 1D NMR results confirm that impregnationo fC PG leads to the formation of DVB-bridgedo rganosilica phase inside the CPG macropores and after pseudomorphic transformation, DVB-bridgedo rganosilica phase remains incorporated in the organosilica@CPG hybrid materials. Changes in the rela-tive ratio among Tg roups suggest that mesoporesa re formed in organosilica phase by pseudomorphic transformation. However,w ec annotr ule out the possibility that the mesoporous structures are also generated in silica phase. The exact structural changes are investigatedb ym ulti-dimensional NMR spectroscopya sb elow.

Phase transition by pseudomorphic transformation
Now,aq uestion arises:h ow are each Tand Qs peciesand pore systemsd istributed in the hybridm ateriala fter the pseudomorphic transformation?T he hybrid materialc an either form multi-layer structure consistingo fs eparate organosilica and silica phases or as ingle layer structure composedo fr andom distribution of both Ta nd Qs ilicon species. To answert hese questions,2 Df requency-switched Lee-Goldburgh eteronuclear correlation (FSLG-HETCOR)M AS NMR spectroscopy is utilized since HETCOR is ap owerful technique to detect correlation of nuclears pin pairs which reside in closep roximity in space, thus providing information about the molecular assembly structure and geometrical arrangements. [50][51][52] First, the impregnatedC PG material (50BTEVB) was characterizedb y2 D 29 Si{ 1 H} FSLG-HETCOR NMR measurements as showni nF igure 9a.A se xpected, intramolecular correlation signals belonging to DVB-bridged silica species are predominantly observed:s ignals of T 2 ( 29 Si chemical shift at À68 ppm) and T 3 ( 29 Si at d = À77 ppm) speciesa re correlated with 1 Hs ignals of divinylbenzene( DVB) at d = 6a nd 7ppm, as indicated by orange and red lines in the Figure, respectively.T he signal at 6ppm corresponds to the vinyl protons which are closer to silicon and the correlation at 7ppm can be assigned to the second vinyl proton and the aromatic protons. Thes econd vinyl protons and the aromatic protons cannot be distinguished clearly since the difference in NMR frequency of these moieties is smaller than the linewidth of solid-state NMR (e.g. 7.19 and 7.45 ppm for 2 nd vinyl and aromatic protons, respectively, of the precursor BTEVB in CDCl 3 ). The correlation signal between QS is peciesa nd the protons of the organic bridgea re of very low intensity (marked by arrow). This indicates that only small amountso ft he organosilica bridge are in contact with pure silica, for example, at the phase boundary,b ut the two species are not mixed.T his confirms our expectations for the formation of two-layers tructure consisting of two independents ilica and organosilica phases (structure I in Figure 1). Correlation between isolated silanol groups and T 2 Si occur between 29 Si at d = À68 ppm and 1 Ha td = 1.8 ppm (marked by a grey line), whichi sacommon observation for zeolitic materials. [38,40,53,54] Figure9bs hows 29 Si{ 1 H} FSLG-HETCOR NMR of 50BTEVB_ 6h_C 16 TAOH. After pseudomorphict ransformation, DVB-bridge group ( 1 H, at d = 6-7 ppm) yields additional strong correlation peaks with Q 3 ( 29 Si at À110ppm) and Q 4 (À110ppm) Si species (marked by arrow). This strongc orrelation between DVBbridge and Qs peciesi ndicates that Tand Qs pecies are now in close contact with each other,s uggestingt hat the material does not form two separate layers,b ut these two speciesa re randomly mixed at the molecularl evel. During the pseudomor-phic transformation,QSi species from the inner part of the particles can be dissolved andp rogressively recondensate with TS is pecies at the pore surfaces, whichi ss uggested by the swelling of the pore wall structure as seen in the SEM images in Figure 4. In addition, cleavage of the SiÀCb ond and recondensation of the resulting Qs pecies with uncleaved Tspecies possibly occur,y ieldingt he mixture of Tand Qs pecies. Additional cross peaks at d( 1 H) = 3.8 ppm and 0.9 ppm are caused by correlation of the Qs peciesw ith ethoxyg roup,e thanol and water.

Micellet emplating of organosilica and silica species
The formation of as econd mesopore system shown by N 2 physisorption, TEM and p-XRD measurements demonstrates successful pseudomorphic transformation of the organosilica/ silica hybrid material. However,a mong the organosilica and the silica species, it is not clear which speciesa re affected by the micelle templating process and of which species the pore walls are composed. To answer this question,apseudomorphic transformed sample before the surfactant extraction has been prepared and the interfacial structure between the silica species and the surfactant has been investigated. The signals of the surfactants are dominantly seen in 1 Ha nd 13 CC PM AS NMR spectra,c onfirming the presence of the surfactant (NMR spectra with the peak assignments are shown in FigureS11). In particular, 13 CNMR signal of inner methylene chain (-CH 2 -) and chain-endm ethyl group appear at 30.3 and 14.2 ppm, respectively,i mplying that the aliphatic chains of the surfactants assume gauche-rich conformation in contrast to the trans conformation of bulk-phase. This is consistent with the gel-likeb ehavior of the surfactants confined in nanopores. [38,55] No additional signals that can be assigned to the surfactants in the bulk phase are seen, indicating that the surfactants are located predominantly within the pores.
Spatial proximity between the surfactants and each Si groups are investigated by the 2D 29 Si{ 1 H} FSLG HETCOR MAS NMR of 50BTEVB_6 hb efore extraction (incl.C 16 TAOH) as shown in Figure 10 a. Again, correlations with DVB group ( 1 HNMR, d = 6-7 ppm) are seen for Qs pecies( marked by arrow), confirming mixed phase at the molecular level insteado fs eparate domain structures. In addition, strong correlation of Q 4 and Q 3 Si species are observed with 1 Ha t3 .2 and 1.3 ppm, which can be assigned to the methyl/ethyl group (CH 3 N/NCH 2 -) of the surfactant's head group and inner methylene chain, respectively (denoted by blue and green lines in Figure). The correlation between Qg roup and surfactant indicates that silica species build pore walls and interact with the surfactant inside the pore. In particular,t he correlation with the head group is very intense in comparison to that with the inner methylene chain, which is in agreement with the suggested micelle structure: polar head groupso ft he surfactant are directly interacting with the silicap ore wall due to the electrostatic interaction and the long hydrocarbon chains face towardt he centero f the pores. [11,38] By contrast, weaker correlation signals of T 3 and T 2 groups appear at slightly different chemical shift positions ( 1 H, d = 2.8 and 1.0 ppm) from those of the correlation with Q species. These correlation peaks cannot be attributed to SiOH, SiOCH 2 CH 3 and residual water,s ince silanola nd water exhibit correlation mainly with T 2 and T 1 species as is seen in Figure 9a.T hus, we assign this correlation to the interaction between Tspeciesa nd the surfactant. Since NMR is sensitivet o the local environments, the surfactants nearby Tf unctional group and Qg roup can yield signals at different frequencies due to the differentc hemical environments of the hydrophobic DVB group in comparison to the hydrophilic Qs ilica species.
Ac lose spatialp roximity between the surfactant and T groups are further verified by the 13 C{ 1 H} FSLG HETCOR MAS NMR as show in in Figure 10 b. As expected, intramolecular correlationsb etween aromatic carbons( 13 C, d = 120-150 ppm) and aromatic protons ( 1 H, d = 6-7 ppm) belonging to the DVBbridgeg roup andt hose between aliphatic carbons ( 13 C, d = 10-70 ppm) and protons ( 1 H, d = 1-4 ppm) belonging to the surfactants are observed. More interestingly,i ntermolecular correlation signals between DVB and the head group as well as the inner methylene chain of surfactants are visible as represented by the colored lines in the Figure 10 b. 1 Hc hemical shift values of these intermolecular correlations ignals are in agreement with those between TS is pecies and surfactants observed in 29 Si{ 1 H} FSLG HETCOR ( Figure 10 a), whichc onfirms our assignment of surfactants interacting with Ta nd organic bridge groups. The correlation of the surfactant with Tspecies as well as with the DVB group indicates that organosilica species and the surfactants are in close contact with each other and thus, pore walls also consist of organosilica species.
Overall, 2D NMR results clearly confirmt hat both silica and the organosilica speciesa re affected by the micellet emplating, resultingi nt he formationo fm esopores which are composed of both Tand QS is pecies. Further, silica and organosilica species co-condense during the pseudomorphic transformation, yieldingasingle mesophase of which pore walls consist of a random mixture of Ta nd Qs pecies at the molecular level (structure III in Figure 1). Thus, anew type of single-phase mesoporous morphology is obtained by pseudomorphic transformation from two-layer system of separate silica and organosilica phases. Two-layer structure could suffer from peeling-off and fragmentation of organosilical ayer from the silica framework which would deteriorate the performance in actual applications.L ikely,s ingle-phase structurei sb eneficial since the mixing and co-condensation of organosilica and silicas pecies could enhance the mechanical strengtho ft he functional organosilica phase and stabilize hierarchically porous structures.

Conclusions
Hierarchically porous organosilica/silica hybrid materials have been successfully synthesized in at wo-step approachb yi mpregnation and pseudomorphic transformation. First, incipient wetnessi mpregnationw as found as as uitable methodt of orm al ayer of an anoporous divinylbenzene-bridged organosilica phase in the macropores of controlledp ore glasses,w hile maintaining the morphologyo fC PG. The product exhibits high specific BET surfacea rea of up to 600 m 2 g À1 resulting from micro-or small mesopores of organosilica layer.I tw as shown that the macropore size and the apparent specific BET surfacea rea are adjustable by the organosilica content. Second, the concept of pseudomorphic transformation, which is well known for pure CPGs,w as adapted to form hybrid materials.A fter pseudomorphic transformationo fi mpregnated materials, the hybrid materials show significant increase in the specific BET surfacea reas up to 1000 m 2 g À1 and an ordered second mesopore system of which pore size and surface area are tunable with the choice of the cationic surfactant. Mesopores of d = 4.2 nm and 2.9 nm are generated with C 16 TAOH and C 10 TABs urfactants, respectively.T he 3-dimensional shape of CPG remains the same, however,t he macropore size shrinks slightly which can be controlled by the transformation parameters. Homogeneous distribution of mesopores throughout the hybrid materiali sd emonstratedb yT EM images. Two-dimensional 29 Si{ 1 H} FSLG HETCORN MR spectroscopy provides direct evidence for different phase behavioro ft heseh ybrid materials by probing correlation between organosilica and silica group. Double layer structure composedo fs eparate organosilica and silica layers( structure I) is formed for impregnatedm aterial, while as ingle mesophasew here both Ta nd QS is pecies are randomly distributed in the framework is identified for pseudomorphic transformed material( structureI II). 29 Si{ 1 H} and 13 C{ 1 H} FSLG HETCORN MR of the pseudomorphic transformed hybrid materialb efore the surfactant extraction shows correlation signals of surfactant with both Ta nd Qs pecies, indicating that the micellar templating affectsb oth organosilica and silica phases and leads to structuring of pore walls via co-condensation of Ta nd Qs peciesa tt he molecular level. This demonstratest hat the pseudomorphically transformed materials provide the accessibility to the organic functionality,w hichi so f crucial importance for its potential application, for example, in gas separation processes and catalytic reaction. Additionally, the co-condensation structure can have ar einforcing effect on the organosilica phase. Our synthetic protocol could be expanded to various bis-silylated organosilica precursors as well as different porousg lasses and hence,i to pens an ew field of variable hybrid materials with hierarchical porosity and chemical functionality.
CPG synthesis:As odium borosilicate glass with ac onstitution of 62 wt %S iO 2 ,3 0wt% B 2 O 3 ,6 wt %N a 2 Oa nd 1wt% Al 2 O 3 was milled and fractionated according to the grain sizes by sieve. The fraction with ag rain class of 40-100 mmw as annealed at 620 8Cf or 24 hours in am uffle furnace for phase separation. For acidic leaching the glass was treated in HCl (3 mol L À1 )a t9 0 8Cf or 6h ours, and then washed until the pH of the washing solution became neutral and dried. For alkaline extraction the glass was stirred in NaOH solution (0.5 mol L À1 )at3 08Cf or 5hours (volume ratio of solution/glass = 1/8). Subsequently,i tw as treated in NaOH solution (3 mol L À1 )u nder static condition at reduced pressure (0.2 bar) and at 0 8Cf or two hours (volume ratio solution/glass = 1:2). Again, the glass was washed until neutral and dried at 120 8C.
Impregnation of CPG:B efore impregnation the respective CPGs were heated up to 200 8Cf or at least 4h ours and cooled down to room temperature. 1,4-((E)-2'-Bis(triethoxysilyl)vinyl)benzene (BTEVB) was synthesized according to the literature [33] (detail in ESI). In at ypical mixture of the impregnation solution, 100 mg of the BTEVB was mixed with 900 mg of toluene in order to obtain a 10/90 (w/w) mixture under stirring for 10 minutes. For 25/75, 50/50 or 75/25 (w/w) mixtures the masses were adjusted, respectively, and the resulting samples are denoted as 10BTEVB, 25BTEVB, 50BTEVB and 75BTEVB. From this solution 1.5 mL were added dropwise to 1.00 gC PG in ar ound bottom flask. Impregnation was carried out in ar otary evaporator (Büchi)u sing ad efined pressure protocol under rotation:f or 5minutes at 300 mbar,2 5minutes at 100 mbar,1 5minutes at full evaporation power (approx. 9mbar) at room temperature and another 15 minutes at full power in aw ater bath at 50 8C. After impregnation, the composite of CPG and the precursor was placed in as crew cap bottle. As olution of 7.5 mL ethanol (99.8 %) and 2.5 mL NaOH solution (0.31 mol L À1 ,p H13.5) was prepared. 1.5 mL of this solution were added to the composite before it was treated in as ealed screw cap bottle at 80 8Cf or 20 hours. Afterwards the obtained material was filtered, washed with water and ethanol and finally dried at 80 8C. Pseudomorphic transformation of organosilica/silica hybrid materials:C etyltrimethylammonium hydroxide (C 16 TAOH) solution was prepared in NaOH medium according to the literature. [17] 500 mg of the organosilica/silica hybrid materials from impregnation were treated in 20 mL C 16 TAOH solution (0.08 mol L À1 )a t 100 8Cf or 6h ours (_6 h_C 16 TAOH) or 4d ays (_4 d_C 16 TAOH) under static conditions. Alternatively,9 10 mg C 10 TAB( 3.2 mmol) in 24 mL NaOH solution (0.1 mol L À1 )w as stirred at room temperature for 30 minutes before 600 mg of the organosilica/silica hybrid materials were added and the mixture was treated at 100 8Cf or 24 hours under static conditions (_24 h_C 10 TAB). All materials were filtered, washed with water and ethanol and dried at 80 8C. For extraction of the surfactant the samples were treated in ethanol/hydrochloric acid (32 %) mixture (v/v = 97/3) under reflux conditions for three days, and subsequently the materials were filtered, washed with water and ethanol and dried at 80 8Cagain.

Characterization
Physisorption:A ll samples were outgassed on a Quantachrome Degasser Masterprep under vacuum at 80 8Cf or 16 hours. N 2 physi-sorption data were recorded at 77 Kw ith a Quantachrome Quadrasorb-SI-MP/ Quadrasorb evo or Quantachrome Autosorb 6B. Ar physisorptions were recorded at 78 Kw ith Quantachrome Autosorb-iQ-MP using Quantachrome Cryocooler. The specific surface areas were determined using the method by Brunauer,E mmett and Te ller (BET). [42] The relative pressure range which is considered for determination of the specific BET surface area was adjusted according to the IUPAC recommendation so that it is referred to as the apparent specific BET surface area. Pore diameter distribution was calculated using the non-local density functional theory (NLDFT) kernel for silica with cylindrical pores from the adsorption branch of N 2 physisorption, and using NLDFT kernel for silica/zeolites with cylindrical pores from the adsorption branch of Ar physisorption. The mesopore volume was calculated using the same NLDFT kernel, considering pores smaller than 20 nm. Solid-state nuclear magnetic resonance (NMR) spectroscopy: Solid-state NMR experiments were performed on Bruker AvanceI I 400 spectrometer equipped with a4mm double resonance magic angle spinning (MAS) probe. 1 HM AS NMR spectra were acquired with an operating frequency of 400.28 MHz, MAS frequency of 13 kHz, 908 pulse length of 4.2 msa nd repetition delay of 3s. 13 C cross-polarization (CP) MAS spectra were collected with 13 Co perating frequency of 100.66 MHz, MAS frequency of 13 kHz, contact time of 1ms, repetition delay of 4s,r amped polarization transfer and two-pulse phase-modulated (TPPM) proton decoupling. 29 Si CP MAS NMR spectra were acquired with 29 Si frequency of 79.52 MHz, contact time of 2ms, MAS frequency of 5kHz, recycle delay of 5s and continuous wave proton decoupling. 2D 13 C{ 1 H} and 29 Si{ 1 H} frequency-switched Lee-Goldburg (FSLG) heteronuclear correlation (HETCOR) MAS NMR spectra were obtained using ar amped polarization transfer with 13 Cr ff ield strength of 60 kHz, FSLG homonuclear decoupling during the proton spin evolution time and TPPM heteronuclear 1 Hd ecoupling during the acquisition. 1 Hr ff ield strength of 66 kHz were used for both homonuclear and heteronuclear decoupling. Contact times were varied between 2a nd 5ms. 200 increments were recorded in t1 with States-TPPI method for phase-sensitive detection. MAS frequency of 13 kHz and repetition delay of 3swere used.