Fabrication of Carbon-Like, π-Conjugated Organic Layer on a Nano-Porous Silica Surface

This paper presents a new type of black organic material-porous silica composite providing an extremely highly selective adsorption surface. This black composite was prepared by lamination on nano-sized pores with a carbon-like, π-extended structure, which can be converted via the on-site polymerization of 1,5-dihydroxynaphthalene with a triazinane derivative and a thermally induced condensation reaction with denitrification. This bottom-up fabrication method on porous materials had the great advantage of maintaining the pore characteristics of a raw porous material, but also the resultant black surface exhibited an extremely high molecular-shape selectivity; for example, that for trans- and cis-stilbenes reached 14.0 with the black layer-laminated porous silica, whereas it was below 1.2 with simple hydrophobized silica.


Introduction
For a sustainable environment, there has been a universal requirement for selectivity enhancement in adsorbents in a wide range of research and industrial fields, and therefore the surface modification of porous carrier materials has been widely investigated. The dense immobilization of functional groups such as ionic or polar groups containing heteroatoms in porous carrier materials is one of the simplest ways to achieve this purpose. Further selectivity enhancement can be successfully observed in macrocyclic compound-immobilized adsorbents [1][2][3][4][5][6][7][8][9][10], whose high selectivity is achieved via the cyclic ordering of heteroatoms as interaction sources with a molecular pocket. However, owing to its complex synthetic process and difficult purification process, this approach has limited applicability in molecular design and synthesis. Nevertheless, selectivity enhancement has been achieved by non-macrocyclic linear compounds when a weak interaction site is integrated with highly oriented structures. For instance, significant selectivity enhancement is observed for structural isomers of polycyclic aromatic hydrocarbons when heteroatom-based non-ionic polar groups such as carbonyl units are integrated in polymer side chains that are in highly ordered states. Typical examples have been reported in long chain alkylated polyacrylate [11][12][13], alternating copolymer systems phases. These successful results are because of the promotion of multiple carbonyl-π interactions [22][23] with guest molecules.
In this study, we present a facile and versatile fabrication method for the formation of π-electronrich, carbon-like black organic phases on the surface of porous carrier particles. Our method can be emphasized by its distinct advantages compared with conventional carbon sources such as carbon black and graphite, which are the most readily available carbon sources, but their large size makes them unsuitable for the present purposes. Graphene is significantly small in thickness, and hence it can be used to wrap large surfaces. However, it cannot be used for coating at the nanosized inner interface of the pore [40]. Therefore, developing a bottom-up carbon structure creation method through which small molecules are adsorbed into the pores and that results in carbonization via posttreatment is necessary. Summarizing our method: (1) Long-ranged π-extended structures are created through the polymerization of small monomers and heat treatment, and then the resultant carbonlike layer can be fabricated on the surface of the nanosized pores in carrier particles ( Figure 1). (2) The monomers used in this method are anionic and cationic. Therefore, the resultant carbon-like layer can be easily attached to various carrier surfaces via electrostatic interaction. (3) The degree of extension of the π-conjugated structure can be adjusted by the temperature and time of the heat treatment. Additionally, the adsorption ability can be tuned via this process. In this paper, we also demonstrate the strong absorptivity and high selectivity of the black layer-laminated porous silica as a new class of adsorbent. Figure 1. Schematic illustration of black layer-laminated porous silica through bottom-up carbon structure creation from 1,5-dihydroxynaphthalene (DHN) and 1,3,5-trimethyl-1,3,5-triazinane (TMTA), and its estimated reaction mechanism.

Preparation of π-Conjugated Polymer-Silica Composites
Process I with microwave heating: The procedure is as follows: 1 g of porous silica (YMC SIL-120-S5) and 8 mL of 2-propanol were mixed in the reaction vessel of the micro-wave synthesizer (Monowave 300, Anton Paar Gmbh, Graz, Austria) [41][42][43], and then the mixture was dispersed for 1 min using an ultrasonic bath (SONO CLEANER 100D, Kaijo Corporation, Osaka, Japan). Thereafter, 89.7 mg (0.56 mmol) of DHN and 72.4 mg (0.56 mmol) of TMTA were added to the dispersed mixture, and then the mixture underwent ultrasonication. The reaction vessel was placed in a micro-wave synthesizer and heated at 100 • C for 3 min while stirring. The obtained slurry was filtered, and the residue was thoroughly washed with 2-propanol and EtOH, and then dried in vacuo at room temperature. Process I without microwave heating: The procedure was as follows: 5 g of porous silica (YMC SIL-120-S5) was ultrasonically dispersed in 230 mL of EtOH in a 500 mL flask, and then 0.55 g (3.5 mmol) of DHN and 0.45 g (3.5 mmol) of TMTA were added to the dispersed silica. The mixture was gently stirred at room temperature for 1 min, heated to 75 • C, and then stored at 75 • C for 4 h. The obtained greenish product was filtered, and then the residue was thoroughly washed with EtOH and dried in vacuo at room temperature.

Blackening of Polymer-Silica Composites
The blackening of the polymer-silica composites, which corresponded to carbonization, was carried out in the presence and absence of a microwave synthesizer. Process II with microwave heating: a mixture of 2 g of the polymer-silica composites and 20 mL of ethylene glycol was redispersed in a reaction vessel of the microwave synthesizer, and then the mixture was heat-treated at 200 or 300 • C for 10 min in a microwave synthesizer. A dark-colored product was collected via filtration, and then it was thoroughly washed with EtOH and dried in vacuo at room temperature. Process II without microwave heating: 5 g of the polymer-silica composite was placed in a crucible (10 mL) and heat-treated in a nitrogen atmosphere at a predetermined temperature ranging from 200 to 900 • C for 2 h.

Instrumentations
The UV-visible and fluorescence spectra were measured using V560 (JASCO Co., Ltd., Tokyo, Japan) and FP-6500 (JASCO Co., Ltd., Tokyo, Japan), respectively. The reflective UV-vis spectroscopy was measured using USB4000 with a UV-vis-NIR light source DH-2000 and optical fiber (Ocean Optics, Largo, FL, USA). The obtained particles were observed via SEM Hitachi SU-8000 (Hitachi-Hightech Corp., Tokyo, Japan). Osmium tetroxide samples were treated using a Filgen osmium plasma coater OPC60A before observation. Thermogravimetric analysis (TGA) was conducted using the TG/DTA6200 (Seiko Instruments Inc., Chiba, Japan) in atmospheric air. The heating rate was set at 10 • C min −1 . Elemental analyses were carried out on Yanaco CHN Corder MT-6 apparatus (YANACO Co., Ltd., Kyoto, Japan). Raman spectroscopy was measured using a laser Raman spectrophotometer NRS-5100 (JASCO Co., Ltd., Tokyo, Japan). The Raman spectra were recorded at a 532 nm laser excitation wavelength with a power of 5.5 mW. Each spectrum was recorded for 40 s with 10 accumulations and with a spectral resolution of 2 cm -1 . The specific surface area and pore properties of the prepared particles were measured using NOVA2200 (Quantachrome, Boynton Beach, FL, USA) based on the Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) methods. Liquid nitrogen was used as the coolant.

Evaluation of Absorptivity
The HPLC columns were obtained by packing the greenish and black-colored silicas in a stainless-steel column (150 × 4.6 mm i.d.). A 20 µL loop JASCO 980 pump with a Reodyne Model 7725 injector (JASCO Co., Ltd., Tokyo, Japan) and a JASCO MD-2010 plus UV-vis photodiode array detector (JASCO Co., Ltd., Tokyo, Japan) were used for the HPLC measurements. The column temperature was maintained using a column jacket with a heating and cooling system. A personal computer connected to the detector, and a pump with ChromNAV (Ver 1.17, JASCO Co., Ltd., Tokyo, Japan) software was used to control the system and analyze the data. Chromatography-grade solvents were used to prepare the mobile phase. The mobile phases used in HPLC were methanol-water mixtures, ethanol, and ethanol-THF mixtures. The separation factor (α) was obtained from the retention factor (k ) ratio. The retention time of water was used as the void volume (t 0 ) marker (the absorption for methanol was measured at 220 nm). The water/1-octanol partition coefficient (log P) was determined using the retention factor with an ODS column: logP o/w = 3.759 + 4.207logk (r 2 = 0.99997) [44].

Characterization of the Polymerization-Induced Assembly Method
To achieve a bottom-up carbon structure creation using small molecules in the inner pores, we applied the polymerization-induced self-assembling approach using hydroxyaromatic as a key monomer with a triazinane derivative as a cross-linker, which was developed to produce monodisperse nano-sized to micro-sized spherical particles [42,45]. Because of its adjustability on the assembling size and easy carbonization via simple heat treatment at a relatively low temperature, this approach is relevant to the present purpose. In this study, 1,5-dihydroxynaphthalene (DHN) was selected as a key monomer, and the polymerization was initiated by heating a DHN-containing solution in the presence of one equimolar of 1,3,5-trimethyl-1,3,5-triazinane (TMTA). As shown in the scanning electron microscope (SEM) of Figure 2A, the well-size-controlled spherical solid particles were produced spontaneously without any surfactant when a proper solvent such as ethanol and similar polar media (Table 1) were selected. The polymerization conditions are summarized in Table 1. Figure 2A-1,2 show the SEM images from the particles prepared with microwave heating at 150 • C, and the SEM images in Figure 2A-3,4,5,6,7 were from the particles prepared without microwave heating at 75 • C. The polymerization rate was significantly accelerated by microwave heating. However, a conventional batch method without microwave heating was useful for large-scale production. The SEM images showed that the obtained solids exhibited a spherical shape with a good monodispersity regardless of the heating methods. In addition, all the particles were insoluble and less swelled in any solvent presumed to be a high-density crosslinked resin-type polymer. Therefore, it is estimated that the polymerization-induced aggregation is accompanied by a cross-linking reaction for solidification. rate was significantly accelerated by microwave heating. However, a conventional batch method without microwave heating was useful for large-scale production. The SEM images showed that the obtained solids exhibited a spherical shape with a good monodispersity regardless of the heating methods. In addition, all the particles were insoluble and less swelled in any solvent presumed to be a high-density crosslinked resin-type polymer. Therefore, it is estimated that the polymerizationinduced aggregation is accompanied by a cross-linking reaction for solidification. Through the dynamic light scattering (DLS) method, the polymerization progress was monitored by the particle diameter. Figure S1 shows the typical time courses of the increase in the particle diameter during polymerization. Considering this result and the fact that the particle diameter can be affected by the reaction conditions (Table 1 and Figure 2A), various factors can be expected to play a key role in the laminating process on a pore surface with a polymer layer.
To estimate the reaction mechanism of this polymerization, the time course of the carbon and nitrogen content ratio (CN ratio) in elemental analysis was examined without microwave heating. The CN ratio increased from 8.5 to 10.0 at a time course ranging between 6 and 8 h; the detailed data are given in Figure S2. These values are within 6.8 and 11.1, which were calculated as one-to-two and one-to-one addition products (whose structures are estimated in Figure S2), respectively. In contrast, we detected a significant increase in fluorescence at the initial stage. Figure 3 contains the time courses of the fluorescent intensity at the emission maximum at 455 nm upon excitation at 400 nm and the absorption increase at 500 nm when the reaction was monitored at a low concentration condition of 1/30 to reduce the reaction rate; the detailed spectra are given in Figure S3.
The P m and P h particles were prepared with and without microwave heating, respectively. The concentration indicates that of each monomer. The diameter and CV were determined by a SEM analysis of the corrected solids after the proper washing and drying processes.
No similar emission was observed in the raw material DHN. Finally, the reaction mixture changed to a non-fluorescent and dark-colored suspension as the reaction progressed. These results indicate that the fluorescent species are formed as an intermediate during polymerization, resulting in a π-extended structure. According to the literature [46], the cross-linking polymerization of DHN has been obtained by mixing ammonia with formaldehyde and not TMTA. Additionally, it explains that polymerization can be performed via the formation of the naphthoxide precursor (given in Figure S2), although it is not mentioned that fluorescent species are generated. Therefore, we estimate that our polymerization system includes a similar precursor formation with a fluorescent structure. are given in Figure S2. These values are within 6.8 and 11.1, which were calculated as one-to-two and one-to-one addition products (whose structures are estimated in Figure S2), respectively. In contrast, we detected a significant increase in fluorescence at the initial stage. Figure 3 contains the time courses of the fluorescent intensity at the emission maximum at 455 nm upon excitation at 400 nm and the absorption increase at 500 nm when the reaction was monitored at a low concentration condition of 1/30 to reduce the reaction rate; the detailed spectra are given in Figure S3.

Formation of a Carbon-Like Black Layer on Porous Carrier
Based on the fundamental observations in the polymerization progress, we carried out this polymerization in the presence of porous silica (5 µm and 12 nm in average diameter and pore size; specific surface area of 330 m 2 g −1 in the catalog specifications), which had been mostly used as carrier particles for high-performance liquid chromatography (HPLC). Table 2 summarizes the preparation conditions with the abbreviation of the products obtained. The change in color during polymerization can be used to estimate the reaction progress. As shown in Figure 4, the white suspension owing to raw silica turned greenish and then deeply dark. Figure 2B-2,3 show SEM images of the colored silica (abbreviated as Sil-P 17 /H 75 -0), obtained using 20 wt % of DHN of silica. Compared to the raw silica ( Figure 2B-1), the colored silica maintained a good sphericity, with a smooth surface after polymerization ( Figure 2B-3). Additionally, the average diameter slightly increased by approximately 9%. In contrary, non-uniform surfaces ( Figure 2B-4) were observed when approximately 50 wt % monomers were used in the polymerization process. This indicates that the excessive use of monomers promotes the undesirable self-growth of polymer particles on the silica surface. Nanomaterials 2020, 10, x FOR PEER REVIEW 7 of 14 as a solvent in process I. The initial wt. ratio corresponds to the weight ratio of the monomers for silica. c: In the microwave method, ethylene glycol (EG) was used as the suspension solvent. d: Sil-0 indicates the raw porous silica. As shown in Figure S4, thermogravimetric analysis (TGA) of the colored silica (Sil-P17/H75-0) indicated that the amount of organic component attached on silica was ca. 15.2 wt %, considering the loss from silica because of the heat treatment. This value corresponds to 99% of the initially added DHN being attached as a polymer when it was assumed that the DHN was converted to a one-to-one addition polymer. The progress of such highly efficient polymerization also indicates that the cationic TMTA is effectively concentrated on the pore surface of the silica interface due to the electrostatic interaction. However, the slight elution of brown substances was observed by washing, while the polymer-coated silica did not change the color. This may be due to the by-production of a small polymer during the microwave heating, but this is not suitable for further application with HPLC because HPLC is very sensitive to such a slight elution. Therefore, we used only the polymer-coated silica obtained without microwaves for the HPLC study.

Blackening and Carbonization of the Polymer Component
The significant color changed from dark green to complete black when Sil-P17/H75-0 was incubated at higher temperatures, such as 200, 400, and 560 °C, under a nitrogen atmosphere. As shown in Figure 4d, Sil-P17/H75-H560 did not show any reflection in the visible light range. In addition, a further weight reduction was observed from 14.7 to 8.6 wt % (see Figure S4). Consistent with this result, Figure 2B indicates that the particle size gradually decreased with the increase in the incubation temperature.
The derivative thermogravimetric analyses (Figure 5a) of the TGA profiles provide important information on the weight loss leading to the blackening process. Figure 5a shows that the decomposition profiles of the Sil-P17/H75-H200 and Sil-P17/H75-H400 prepared via incubation at 200 °C and 400 °C were similar to that without additional heat treatment. The thermal decomposition of these composites occurs at wide temperatures ranging from 230 to 650 °C . This profile can be understood by combining the small changes around 230-310 °C (Stage II) and 310-420 °C (Stage III) with the large changes observed at about 500 °C (Stage IV). Therefore, it is estimated that the effective blackening due to the formation of π-expanded structures is promoted by heat treatment at temperatures above 400 °C . Elemental analysis supports the structural change in the attached polymer by the heat treatment. The CN ratio of the Sil-P17/H75-0 without heat treatment was calculated to be 9.74 from 10.5 and 1.08 wt % in carbon and nitrogen, respectively. This CN value is close to the 9.90 of the Ph-3 particles prepared without silica. In contrary, the incubation of Sil-P17/H75-0 at 560 °C evidently increased the CN ratio to 37.3, with a significant decrease in nitrogen. This In addition, owing to successive washing with various solvents such as ethanol, DMF, and chloroform, the colored silica had neither a significant color change nor reduction in the organic components. Therefore, it can be established that the polymer from DHN and TMTA was formed and tightly attached to the silica surface. This is probably due to the electrostatic interaction between the TMTA-derived imino groups of the polymer and the silanol groups of silica.
As shown in Figure S4, thermogravimetric analysis (TGA) of the colored silica (Sil-P 17 /H 75 -0) indicated that the amount of organic component attached on silica was ca. 15.2 wt %, considering the loss from silica because of the heat treatment. This value corresponds to 99% of the initially added DHN being attached as a polymer when it was assumed that the DHN was converted to a one-to-one addition polymer. The progress of such highly efficient polymerization also indicates that the cationic TMTA is effectively concentrated on the pore surface of the silica interface due to the electrostatic interaction. However, the slight elution of brown substances was observed by washing, while the polymer-coated silica did not change the color. This may be due to the by-production of a small polymer during the microwave heating, but this is not suitable for further application with HPLC because HPLC is very sensitive to such a slight elution. Therefore, we used only the polymer-coated silica obtained without microwaves for the HPLC study.

Blackening and Carbonization of the Polymer Component
The significant color changed from dark green to complete black when Sil-P 17 /H 75 -0 was incubated at higher temperatures, such as 200, 400, and 560 • C, under a nitrogen atmosphere. As shown in Figure 4d, Sil-P 17 /H 75 -H 560 did not show any reflection in the visible light range. In addition, a further weight reduction was observed from 14.7 to 8.6 wt % (see Figure S4). Consistent with this result, Figure 2B indicates that the particle size gradually decreased with the increase in the incubation temperature.
The derivative thermogravimetric analyses (Figure 5a) of the TGA profiles provide important information on the weight loss leading to the blackening process. Figure 5a shows that the decomposition profiles of the Sil-P 17 /H 75 -H 200 and Sil-P 17 /H 75 -H 400 prepared via incubation at 200 • C and 400 • C were similar to that without additional heat treatment. The thermal decomposition of these composites occurs at wide temperatures ranging from 230 to 650 • C. This profile can be understood by combining the small changes around 230-310 • C (Stage II) and 310-420 • C (Stage III) with the large changes observed at about 500 • C (Stage IV). Therefore, it is estimated that the effective blackening due to the formation of π-expanded structures is promoted by heat treatment at temperatures above 400 • C. Elemental analysis supports the structural change in the attached polymer by the heat treatment. The CN ratio of the Sil-P 17 /H 75 -0 without heat treatment was calculated to be Nanomaterials 2020, 10, 1882 8 of 14 9.74 from 10.5 and 1.08 wt % in carbon and nitrogen, respectively. This CN value is close to the 9.90 of the P h -3 particles prepared without silica. In contrary, the incubation of Sil-P 17 /H 75 -0 at 560 • C evidently increased the CN ratio to 37.3, with a significant decrease in nitrogen. This phenomenon was common to the non-silica polymer particles from DHN and TMTA. Therefore, it is presumed that heat treatment enhanced the thermally induced denitrification reaction, and it developed into a long-range π-conjugated structure to become black via multi-step reactions.
Nanomaterials 2020, 10, x FOR PEER REVIEW 8 of 14 phenomenon was common to the non-silica polymer particles from DHN and TMTA. Therefore, it is presumed that heat treatment enhanced the thermally induced denitrification reaction, and it developed into a long-range π-conjugated structure to become black via multi-step reactions. The Raman spectroscopy results support a thermally induced increase in the sp 2 carbon content. From Figure 5b, it is clear that the incubation of Sil-P17/H75-0 at 560 °C increased the typical absorption intensities around 1590 and 1350 cm −1 , which were attributed to the G and D bands based on the sp 2 and sp 3 carbons, respectively [47][48][49] The spectral pattern and the ID/IG ratio (ca. 0.5) indicate that the blackened component on the silica is likely to be amorphous carbon [50]. However, further detailed investigation is needed to determine the exact structure.

Pore Characterization of Black Polymer-Laminated Silica
The maintainability of porosity after polymer lamination is an important requirement for a variety of applications. In this study, the particle characterization of the raw silica, the polymerlaminated silica, and the blackened silica was examined using SEM and Brunauer-Emmett-Teller (BET) analyses. The comparison of Sil-P17/H75-0 and Sil-P50/H75-0 before heat treatment summarized in Table 3 shows that polymerization increased the particle size by decreasing both the pore size and the specific surface area. In particular, Sil-P50/H75-0, which obtained a polymer content of 48.3 wt %, showed an undesired loss of surface area, corresponding to a decrease in the pore volume to 21%.
In contrary, Sil-P17/H75-0, in which 15.2 wt % of the polymer was introduced, maintained a good porosity of 229 m 2 g −1 in their specific surface area, corresponding to 69% of the raw silica. Further improvement was confirmed via heat treatment. In the case of Sil-P17/H75-0, it restored the pore characteristics by slightly decreasing the particle size. For instance, incubation at 200 and 560 °C increased the porosity to 235 and 279 m 2 g −1 in their specific surface areas, which corresponded to 71% and 85% of the raw silica, respectively. Notably, the average pore diameter was larger than that of the raw silica after heat treatment at 560 °C. It is difficult to determine if the original porosity is maintained by thin layer formation on the pore surface or because a newly created polymer layer increases the surface roughness. However, it is noted that the non-silica-supported polymer particles exhibited no significant specific surface area, and when the polymerization with silica was carried out using low concentrations, the resulting surface was smoother than at high concentrations. Therefore, these facts indicate that our method suggests the possibility of maintaining a good initial porosity. The Raman spectroscopy results support a thermally induced increase in the sp 2 carbon content. From Figure 5b, it is clear that the incubation of Sil-P 17 /H 75 -0 at 560 • C increased the typical absorption intensities around 1590 and 1350 cm −1 , which were attributed to the G and D bands based on the sp 2 and sp 3 carbons, respectively [47][48][49] The spectral pattern and the I D /I G ratio (ca. 0.5) indicate that the blackened component on the silica is likely to be amorphous carbon [50]. However, further detailed investigation is needed to determine the exact structure.

Pore Characterization of Black Polymer-Laminated Silica
The maintainability of porosity after polymer lamination is an important requirement for a variety of applications. In this study, the particle characterization of the raw silica, the polymer-laminated silica, and the blackened silica was examined using SEM and Brunauer-Emmett-Teller (BET) analyses. The comparison of Sil-P 17 /H 75 -0 and Sil-P 50 /H 75 -0 before heat treatment summarized in Table 3 shows that polymerization increased the particle size by decreasing both the pore size and the specific surface area. In particular, Sil-P 50 /H 75 -0, which obtained a polymer content of 48.3 wt %, showed an undesired loss of surface area, corresponding to a decrease in the pore volume to 21%.
In contrary, Sil-P 17 /H 75 -0, in which 15.2 wt % of the polymer was introduced, maintained a good porosity of 229 m 2 g −1 in their specific surface area, corresponding to 69% of the raw silica. Further improvement was confirmed via heat treatment. In the case of Sil-P 17 /H 75 -0, it restored the pore characteristics by slightly decreasing the particle size. For instance, incubation at 200 and 560 • C increased the porosity to 235 and 279 m 2 g −1 in their specific surface areas, which corresponded to 71% and 85% of the raw silica, respectively. Notably, the average pore diameter was larger than that of the raw silica after heat treatment at 560 • C. It is difficult to determine if the original porosity is maintained by thin layer formation on the pore surface or because a newly created polymer layer increases the surface roughness. However, it is noted that the non-silica-supported polymer particles exhibited no significant specific surface area, and when the polymerization with silica was carried out using low concentrations, the resulting surface was smoother than at high concentrations. Therefore, these facts indicate that our method suggests the possibility of maintaining a good initial porosity. The surface characterization was evaluated via ζ-potential measurement. Figure 6 shows the pH dependencies of the raw silica (Sil-0) and the polymer-laminated silica (Sil-P 17 /H 75 -0 and Sil-P 17 /H 75 -H 200 ) on the ζ-potential profiles. From Figure 6, the raw silica exhibited negative values at pHs above 5. Because the bending point is obtained at a pH ranging between 5 and 8, the pH dependency of Sil-0 can be attributed to the pKa = 5.6 of Si-OH [51]. In contrary, the polymer lamination of silica (Sil-P 17 /H 75 -0) resulted in a positive charge under acidic conditions, while the negative charge remained in the alkaline pH. From the estimated structure of the polymer in Figure 1, such an ampholytic property is presumed to be derived from an imino group of TMTA as a cross-linker and a phenolic group of DHN.  ; and PV: pore volume (mg L −1 ) were determined by BET.
The surface characterization was evaluated via ζ-potential measurement. Figure 6 shows the pH dependencies of the raw silica (Sil-0) and the polymer-laminated silica (Sil-P17/H75-0 and Sil-P17/H75-H200) on the ζ-potential profiles. From Figure 6, the raw silica exhibited negative values at pHs above 5. Because the bending point is obtained at a pH ranging between 5 and 8, the pH dependency of Sil-0 can be attributed to the pKa = 5.6 of Si-OH [51]. In contrary, the polymer lamination of silica (Sil-P17/H75-0) resulted in a positive charge under acidic conditions, while the negative charge remained in the alkaline pH. From the estimated structure of the polymer in Figure 1, such an ampholytic property is presumed to be derived from an imino group of TMTA as a cross-linker and a phenolic group of DHN. On the other hand, heat treatment at 560 °C (Sil-P17/H75-H560) reduced the positive charge in an acidic pH, which can be explained by a thermally induced denitrification reaction to become a carbonlike structure. Looking at the pH profile in detail, the negative charge is seen in a wide range of pHs, but also it seems to be a two-step transition. In addition, we confirmed both the reduction of positive charge in an acidic pH and the remaining of the negative charge in an alkaline pH for the heat treatment of non-silica-supported DHN-TMTA particles. Therefore, the negative charge of Sil-P17/H75-H560 in a wide range of pHs indicates the existence of a phenolic OH, although it is indistinguishable from that of a Si-OH moiety from the matrix silica.

Detection of Specific Interface Property by Selective Adsorption
First, through column methodology, a fundamental property of the adsorptivity and selectivity of the polymer-laminated silica was examined using hydrophobic substances such as alkylbenzene homologues and simple aromatic hydrocarbons such as benzene and naphthalene as elutes. When the Sil-P17/H75-H200 and Sil-P17/H75-H560 columns were used for the separation of alkylbenzene homologues with alkyl chains of different lengths using polar solvents such as methanol, ethanol, and acetonitrile as the mobile phase, their elution orders were identical to those of octadecylated silica (ODS), which has been mostly used in HPLC. This indicates that the Sil-P17/H75-H200 and Sil-P17/H75- On the other hand, heat treatment at 560 • C (Sil-P 17 /H 75 -H 560 ) reduced the positive charge in an acidic pH, which can be explained by a thermally induced denitrification reaction to become a carbon-like structure. Looking at the pH profile in detail, the negative charge is seen in a wide range of pHs, but also it seems to be a two-step transition. In addition, we confirmed both the reduction of positive charge in an acidic pH and the remaining of the negative charge in an alkaline pH for the heat treatment of non-silica-supported DHN-TMTA particles. Therefore, the negative charge of Sil-P 17 /H 75 -H 560 in a wide range of pHs indicates the existence of a phenolic OH, although it is indistinguishable from that of a Si-OH moiety from the matrix silica.

Detection of Specific Interface Property by Selective Adsorption
First, through column methodology, a fundamental property of the adsorptivity and selectivity of the polymer-laminated silica was examined using hydrophobic substances such as alkylbenzene homologues and simple aromatic hydrocarbons such as benzene and naphthalene as elutes. When the Sil-P 17 /H 75 -H 200 and Sil-P 17 /H 75 -H 560 columns were used for the separation of alkylbenzene homologues with alkyl chains of different lengths using polar solvents such as methanol, ethanol, and acetonitrile as the mobile phase, their elution orders were identical to those of octadecylated silica (ODS), which has been mostly used in HPLC. This indicates that the Sil-P 17 /H 75 -H 200 and Sil-P 17 /H 75 -H 560 components in silica can act as a hydrophobic phase. However, a detailed analysis of the retention behavior evokes a new perspective for understanding specific behavior, especially in the P 17 /H 75 -H 560 moiety. For instance, Sil-P 17 /H 75 -H 560 exhibited the fact that the selectivity (α) calculated from each retention factor was 10.7, whereas the ODS was 1.25 when the retention factor of 1-butylbenzene (C 10 H 14 ) was compared with that of tert-butylbenzene (C 10 H 14 ). Additionally, a similar difference was observed in the comparison of naphthalene (C 10 H 8 ) with 1-butylbenzene (C 10 H 14 ): α = 26.5 and 1.80 for Sil-P 17 /H 75 -H 560 and ODS, respectively (see Table S1). ODS exhibited small separations because these elutes have the same number of carbon atoms and there is no significant difference in their log P as an indicator of hydrophobicity. Therefore, to explain the unusual selectivity enhancement owing to heat treatment, we should consider a different mechanism that recognizes the other physical properties of elutes. To investigate the specific retentivity and selectivity of Sil-P 17 /H 75 -H 560 , geometrical isomers of stilbenes were selected as elute samples. As summarized in Table 4, monomeric ODS shows a small selectivity (α= 1.04) for the transand cis-isomers of stilbenes, whereas a small increase to 1.19 can be observed in polymeric ODS. These small values are because the isomers have no significant hydrophobicity difference, which can be characterized by their log P = 4.71 and 4.81 in the cisand trans-forms, respectively. In contrast, a higher selectivity (α = 1.68-3.65) can be observed for the special separating agents that are expected to have additional interaction mechanisms, such as the carbonyl-π [15,19,52], π-π [53][54][55][56], and polar-π [57,58] interactions.
Notably, this value is the highest. According to our previous studies, multiple interactions between the adsorbent and solute increase the retentivity and selectivity, even if it is based on weak interactions such as the π-π interaction. In the case of stilbenes, it is understandable that Sil-P 17 /H 75 -H 560 with carbon-like π-electron planes promote multiple interactions with trans-stilbene with a planar structure that is much stronger than that of a non-planar cis-stilbene.

Conclusions
In this study, we have presented a versatile method for introducing a carbon-like π-electron-rich organic phase into porous silica. Our method can be emphasized by its applicability for on-site polymerization without the pre-modification of the carrier particle surface, the consequent facile carbonization process, and the good maintainability of the initial porosity of the carrier particles. These advantages are derived from a bottom-up creation method for a carbon structure where the small key molecules are adsorbed into the pores. In this study, we focused on the application of high-selective adsorption using the surface modification of porous silica. However, because of the change in monomers and carriers, our method experienced limited applicability. For instance, through a preliminary test we have confirmed that fluorescent particles can be created by changing the monomer from 1,5-to 2,6-DHN as a starting monomer. Moreover, other applications include the modification of the catalyst surface, masking, and stability enhancement. We hope to report these findings in the near future.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/10/9/1882/s1: Figure S1: Time courses of the diameter change of the particles generated in the polymerization of DHN with TMTA in ethanol at 75 • C. Figure S2: Time course of the carbon-nitrogen (CN) ratio of the products determined by elemental analysis, and the estimated structures of precursors and polymers as possible intermediates in the reaction mixture. Figure S3: UV-visible and fluorescence spectra monitored during the polymerization. Figure S4: Thermal decomposition profiles of Sil-P 17 in air. Table S1: Separation factors (α) of hydrocarbons with C 10 -hydrocarbons with ODS and Sil-P 17 /H 75 -H 560 columns.