Fabrication of Functionalized Porous Silica Nanocapsules with a Hollow Structure for High Performance of Toluene Adsorption-Desorption.

Functionalized mesoporous silicas are an emerging kind of adsorbents for the removal of volatile organic compounds (VOCs). Breaking the limitations of traditional mesoporous silica, in this study, porous silica nanocapsules (PSNs) functionalized with phenyl and n-octyl groups (named as p-PSN and n-PSN, respectively) were developed for the first time. Under dry conditions, the PSNs exhibited highest dynamic adsorption capacity and desorption efficiency among the ever-reported typical adsorbents (i.e., SBA-15, KIT-6, silicalite-1, and activated carbon). Under wet conditions, the functionalized PSNs made up the defects of pure PSNs, displaying excellent hydrophobicity. The Q WET for n-PSN and p-PSN increased by 44 and 76%, respectively, as compared with that of pure PSNs in 50% relative humidity. The Henry constant of static adsorption demonstrated that p-PSN had a better capture ability for toluene, which was owing to the π-interaction between the phenyl groups and the toluene molecules. In addition, p-PSN showed considerable stability after six consecutive dynamic adsorption-desorption cycles in 50% relative humidity.

Introduction

With rapid industrial development, the heavy emission of waste volatile organic compounds (VOCs) results in serious environmental hazards, such as chemical smog, secondary organic aerosol, and formation of photochemical ozone.[1,2] Many technologies have been applied for controlling VOC emission, such as membrane separation,[3,4] photocatalytic oxidation,[5] adsorption,[6,7] and combustion (including thermal incineration[2] and catalytic combustion[8,9]). Among these techniques, combustion is a widely used technology to completely convert VOCs into carbon dioxide and water without secondary pollution. However, for treating the emission of low-concentrated VOCs in real conditions, combustion alone causes enormous loss of energy.[10] Therefore, hybrid technologies in combination with adsorptive concentration and combustion are found effective to solve this problem. An adsorptive concentration system can concentrate the low-concentrated VOCs using adsorption, followed by thermal desorption.[11] For quasi-continuous adsorption–desorption operation, the crucial part of such system is the adsorbent, which requires a high VOC adsorption capacity, rapid desorption rate, and relatively low desorption temperature. Additionally, humid gas is ubiquitous in the VOC stream; thus, in order to improve the applicability, an adsorbent with high hydrophobic property needs to be urgently developed.[12]

Activated carbon (AC) has been conventionally selected as a gas adsorbent owing to its high adsorption capacity with low cost.[2,13,14] Unfortunately, several serious drawbacks limited its wide applications, mainly including poor thermal stability and high desorption temperature in regeneration.[15,16] Hence, other adsorbents such as zeolites and mesoporous silica are desired to replace it. Among zeolites, aluminum-free MFI-type silicalite-1 is widely used owning to its high hydrophobic property and thermal stability.[17] However, microporous structures hinder diffusion of molecules in both adsorption and desorption.[17,18] Mesoporous silicas outperform zeolites because of their mesoporous channels, improving the mass transferring performance. In previous studies, among ordered mesoporous silicas, three-dimensional (3-D) mesostructured materials were found to be better than one-dimensional mesostructured materials, that is, KIT-6 showed a better adsorption performance than either SBA-15 or MCM-41.[19] Thus, further research has always paid more attention on KIT-6.[20] In order to improve the adsorption ability in humid environments, tailored surface modification was required to enhance the hydrophobicity. Dou et al.[19] prepared four-ordered mesoporous silicas grafted with phenyl groups. The phenyl-grafted KIT-6 offered the best adsorption of toluene under less wet conditions. Liu et al.[21] synthesized phenyl-KIT-6 by a cocondensation one-step method, and they exhibited high adsorption capacity of toluene in a high humidity environment. Liu et al.[22] further synthesized triphenyl-grafted KIT-6 and phenyl-grafted KIT-6, exhibiting different adsorption performances because of the surface area and exposed organic groups. From all of the above research studies, it can be concluded that surface modification with organic groups for silica-based materials could enhance the adsorption performance under wet conditions, improving the applicability. Nevertheless, these improvements were only limited into the field of KIT-6. For industrial applications, there are still some issues that inherent in KIT-6, such as high cost of manufacturing and poor adsorption capacity compared with commercial AC. Consequently, for making up the above defects of KIT-6-based materials, there is a need of identifying a desirable new functionalized silica-based adsorbent that is economically viable and has excellent adsorption performance. However, as far as we know, a new kind of functionalized silica-based adsorbent for VOC adsorption has rarely been investigated in recent research studies.

Based on the literature review, a bimodal pore system was essential for adsorption of VOCs.[23] We speculated that a hollow structure combined with sufficient small mesopores was an advanced form of a bimodal pore system. Moreover, for further practical applications, it was expected that this new silica-based mesoporous adsorbent should be equipped with a hydrophobic surface. Herein, in the present study, we developed for the first time hollow-structured porous silica nanocapsules (PSNs) functionalized with phenyl and n-octyl groups, which possessed many mesopores and a functional hydrophobic surface. Toluene was selected as a probe VOC molecule, and a continuous-flow adsorption method was established to evaluate the adsorption performance. The dynamic adsorption–desorption behaviors for PSNs were systematically evaluated and compared with those of other typical adsorbents (silicalite-1, KIT-6, SBA-15, and AC). The hydrophobicity of PSN-based materials was investigated under different relative humidities (11–50% RH) in the test conditions. The aim of this study was to design an adsorbent possessing superior VOC adsorption capacity and high desorption efficiency under both dry and wet conditions.

Results and Discussion

Adsorbent Characterization

The PSN-based materials were synthesized via a dynamic self-assembly method by using diethyl ether as the cosolvent. Transmission electron microcopy (TEM) analysis was conducted to show the morphology and structure of the adsorbents. All the particles showed the same spherical shapes with inner cavities as shown in Figure a,b, which suggested a successful synthesis of hollow nanospheres. Figure c,d illustrated that the hollow structure was maintained after the organic groups were grafted onto the silica surface. The mean diameter of these spherical nanocapsules was estimated between 150 and 200 nm. According to the high-magnification TEM images, each nanocapsule presented a shell wall of around 15 nm. Interestingly, many small mesopores of 2–4 nm were observed on the wall of the nanocapsules. Pleats consisting of channel-like mesopores were found on the nanocapsules, which were caused by simultaneous self-assembly of condensed silica and heterogeneous gasification of diethyl ether.[24]

Figure 1

TEM images of PSNs (a,b), p-PSN (c), and n-PSN (d).

The morphologies of other silica adsorbents for comparison were also analyzed. As shown in Figure a, silicalite-1 consisted of nanoparticles with an average diameter of 70 nm. The hexagonally ordered pores of SBA-15 can be distinguished in Figure b, and uniformly straight two-dimensional channels were also observed. Likewise, as shown in Figure c, KIT-6 exhibited a 3-D cubic structure with ordered bimodal mesopores.

Figure 2

TEM images of other silica adsorbents: silicalite-1 (a), SBA-15 (b), and KIT-6 (c).

The nitrogen adsorption/desorption isotherms of PSN-based materials are shown in Figure . According to the IUPAC nomenclature,[25] both pure PSNs and functionalized PSN materials exhibited typical type IV isotherms with H3 hysteresis loops, indicating the presence of slit-shaped pores. The initial rapid increase of adsorbed N2 at relatively low pressure (P/P0 < 0.1) was caused by the micropores on the wall. Interestingly, no adsorption plateau was observed when the relative pressure was approaching 1.0, suggesting that the adsorbents had large mesopores,[26] which was attributed to the hollow structure, consistent with the direct observation of TEM. The results of nitrogen adsorption were summarized in Table . With the incorporation of phenyl and n-octyl groups, the Brunauer–Emmett–Teller (BET) surface area of p-PSN and n-PSN reduced by 13.1 and 19.8%, respectively, while the obvious decrease of total pore volume was 68.1 and 65.4%, respectively. This phenomenon was caused by grafting of organic groups on the surface. As expected, functionalized organic groups would fill part of the hollow cavity. The pore distribution of all PSN-based materials showed a similar distribution, which agreed well with the TEM result, indicating the adsorption sites on the wall preserved after functionalized with organic groups.

Figure 3

N2 adsorption–desorption isotherms and pore size distribution of PSNs, p-PSN, and n-PSN.

Table 1

Textural Properties of Adsorbents Obtained from Nitrogen Adsorption/Desorption Isotherms

	SBETa (m2 g–1)	Dporeb (nm)	Vtotalc (cm3 g–1)	Smicrod (m2 g–1)	Vmicrod (m3 g–1)
PSNs	1018	5.4	1.5	645	0.31
p-PSN	885	3.7	0.5	602	0.29
n-PSN	816	4.1	0.5	540	0.26
silicalite-1	486	4.4	0.4	340	0.14
SBA-15	745	5.7	1.2	195	0.09
KIT-6	850	5.8	1.3	243	0.11
AC	505	2.9	0.2	408	0.17

BET surface areas.

Average pore diameter.

BJH desorption pore volume.

t-Plot micropore surface area (Smicro) and micropore volume (Vmicro).

The incorporation of phenyl and n-octyl groups into the PSN frameworks was investigated by Fourier transform infrared (FT-IR). As shown in Figure , after grafting with organic groups, several new peaks appeared in the spectra compared with pure PSNs. Notably, p-PSN showed two peaks at 698 and 740 cm–1, which were assigned to the benzene rings.[27,28] The peaks at 2974, 2933—, and 2886 cm–1 were attributed to the −CH3 and −CH2– overlap stretching vibration of the n-octyl groups, while as for p-PSNs, the intensities of these peaks decreased because of less C–H bonds on the phenyl groups.[29,30] The broad peaks at 3420 cm–1 could be assigned to the −OH stretching vibration, which exhibited lower intensity, indicating that most of hydroxyl groups were grafted by organic groups.[27]

Figure 4

FT-IR spectra of PSNs, p-PSN, and n-PSN.

The thermogravimetry (TG)/derivative thermogravimetry (DTG) curves obtained from the thermogravimetric analysis (TGA) experiments are shown in Figure . p-PSN and n-PSN displayed an increasing weight loss compared with PSNs. A significant peak at temperature lower than 200 °C appeared in the DTG curves, which is attributed to the loss of surface-adsorbed water.[27] As compared with the amount of adsorbed water, p-PSN (8.39%) and n-PSN (7.54%) have lesser amount than PSNs (10.87%), indicating that functionalized PSNs exhibited stronger surface hydrophobicity. Compared with PSNs, p-PSN and n-PSN showed additional peaks on the DTG curves in the range of 500–700 and 230–600 °C, respectively, which were attributed to the decomposition of phenyl groups and n-octyl groups, respectively. Moreover, calculating from the weight loss data, the amount of grafting organic groups for p-PSN and n-PSN was 6.67% (phenyl groups) and 7.32% (n-octyl groups), respectively.

Figure 5

TG (solid lines) and DTG (dashed lines) of the PSNs, p-PSN, and n-PSN.

Above all, the PSN materials consisting of hollow spheres were fabricated, and organic groups were incorporated on the surface of the PSNs successfully. Our first priority was to investigate the adsorption performance of pure PSNs and compare with other adsorbents to confirm the superiority of these kinds of hollow-structured materials.

Dry Condition Adsorption–Desorption Behavior of Pure Silicas

To evaluate the dynamic adsorption performance of pure PSNs, we adopted the breakthrough measurements[31] of low-concentrated toluene vapor. The obtained experimental breakthrough curves and the curves simulated with the Yoon–Nelson model are shown in Figure . Important adsorption parameters such as the first adsorption capacity (QDRY) and the first breakthrough time (tB, C/C0 = 0.05) are summarized in Table . Under dry conditions, compared with traditional silica adsorbents and AC, PSNs exhibited excellent adsorption ability. The first adsorption capacity decreased in the order of PSNs (2.10 mmol g–1) > AC (1.78 mmol g–1) > silicalite-1 (1.48 mmol g–1) > KIT-6 (1.34 mmol g–1) > SBA-15 (1.10 mmol g–1). It was important to point out that the first adsorption capacity of PSNs increased by 91, 57, and 42% as compared with that of SBA-15, KIT-6, and silicalite-1 (typical silica materials for adsorption), respectively. Furthermore, the first adsorption capacities of PSNs surpassed that of AC, at around 18%, overcoming the low capacity disadvantage of mesoporous silicas. As a more detailed comparison, Tables S1 and S2 listed the adsorption capacities of other reported adsorbents. As we can see, PSNs exhibited the maximal adsorption capacity among silica-based adsorbents, and the capacity was comparable to several carbon-based materials. The half-life τ obtained from the Y–N model also represented the adsorption ability, and a greater τ demonstrated a superior adsorption capacity. The τ of PSNs (42.79 min) was higher than those of silicalite-1 (30.52 min), KIT-6 (27.75 min), SBA-15 (22.25 min), and AC (36.48 min), in accordance with the order of the actual adsorption capacity.

Figure 6

Toluene breakthrough curves on all the adsorbents under dry condition.

Table 2

Adsorption Capacity and Simulation Adsorption Parameters of All Adsorbents under dry Conditions

	QDRY (mmol g–1)	tBa (min)	τ (min)	k (min–1)	R2
PSNs	2.10	38	42.79	0.5605	0.99
silicalite-1	1.48	24	30.52	0.4642	0.99
KIT-6	1.34	25	27.75	1.3387	0.99
SBA-15	1.10	20	22.25	1.2668	0.99
AC	1.78	32	36.48	0.8223	0.99

tB was the time when C/C0 = 0.05.

In general, the rate constant value k represented the diffusion and mass-transfer characteristic of toluene in the fixed bed. Among pure silica materials, a higher k represented a quicker breakthrough process, indicating a lower diffusion resistance.[32] KIT-6 showed the highest k value of 1.3387 min–1, indicating the least resistance in the diffusion process. SBA-15 also showed a high k value of 1.2668 min–1, indicating that the ordered mesoporous silica adsorbents enhanced the mass transfer in the adsorption process. In particular, the breakthrough time of silicalite-1 was the longest, corresponding to a k value of 0.4642 min–1, demonstrating the limited diffusion in microporous silica adsorbents. Compared with these three kinds of silica materials, the diffusion resistance of PSNs was between those of silicalite-1 and SBA-15, which was most likely caused by the hollow structure.

As previously reported by Kosuge,[23] the VOC molecules were more likely to be adsorbed into small mesopores (2.1 nm in size). For mesoporous adsorbents, the adsorption process was controlled by the diffusion of VOC molecules into the pore channels of the adsorbents. KIT-6 offered better adsorption performance than SBA-15 because of its bicontinuous cubic pore structure and large mesopores.[19] The cubic mesopore structure was favorable to diffusion of toluene molecules, leading to higher accessibility of mesopores. As for PSNs, the crackles on the shell offered channels for VOC molecules to enter the space underneath. Additionally, the spherical hollow structure created 3-D interconnected channels, providing better accessibility for the VOC molecules through the micropores and small mesopores on the shell.[33] The intrinsic pore connection between the mesopore channels and the micropores was the key factor in VOC adsorption.[23] In other words, the hollow structure dramatically enhanced the utilization efficiency of the adsorption sites in the shell. This unique structure could be responsible for the larger adsorption capacity of toluene.[34,35]

In order to further investigate the desorption efficiency, all of the adsorbents were regenerated by thermal treatment (80 °C). The desorption curves (shown in Figure ) show a sharp increase at the beginning of the desorption process. High-concentration toluene streams came out, and over time, the outlet toluene concentration gradually decreased and the value of Qdesorption increased relatively gently. In Table , the desorption performance on various adsorbents were compared in terms of E5 (the desorption efficiency in the first 5 min) and the desorption ratio. A higher E5 typically indicated a better VOCs concentrated process. The desorption efficiency of PSNs (6.27 mmol s–1 g–1) was far more than those of the other adsorbents. This difference was caused by the pore structure and adsorption capacity simultaneously. For ordered mesoporous materials such as KIT-6 and SBA-15, after thermal treatment, the desorbed toluene quickly passed through the ordered channels and finally leaving the adsorption bed. However, the amount of desorbed toluene was limited because of the low adsorption capacity, and that did not work for an effective concentrated process. The PSN-based materials did not contain any ordered mesopores, and thus, the excellent desorption performance was due to the unique hollow structure. All adsorption sites in the spherical shell were easily exposed to the high-temperature carrier gas, and then numerous toluene molecules desorbed into the hollow or interstitial hole, promoting the transport diffusion and boosting the process of concentration. As to silicalite-1 and AC, the values of E5 and the desorption rate were very low. The desorption curves showed that the toluene adsorbed by silicalite-1 and AC desorbed very slowly in the initial stage, and it took a much longer time for them to reach a steady state, suggesting that the narrow micropores hindered the regeneration process.

Figure 7

The comparison of the desorption amount of all the adsorbents under dry condition.

Table 3

Comparison between Dynamic Desorption Performance for Various Adsorbents under Dry Conditions

	PSNs	silicalite-1	KIT-6	SBA-15	AC
E5 (mmol s–1 g–1 × 103)a	6.27	2.87	3.96	3.38	2.65
desorption ratio (%)b	95.7	83.7	96.2	96.9	59.7

E5 was the desorption efficiency in the first 5 min during the desorption process.

Desorption ratio (%) was equal to the value of Qdesorption/Qadsorption for the first desorption cycle under dry conditions.

The adsorption capacities of all adsorbents are depicted in Figure . PSNs, SBA-15, and KIT-6 remained almost unchanged after six cycles. The adsorption capacity of AC decreased dramatically in the second adsorption cycle, like the result reported by Wang,[36] indicating that the toluene molecules adsorbed in the micropores of AC were difficult to be desorbed under mild conditions (80 °C). The adsorption capacity of all adsorbents remained the same after the second-regeneration process.

Figure 8

Equilibrium adsorption capacities for six cycles of all the adsorbents under dry condition.

Wet Condition Adsorption Behavior of PSN-Based Materials

As expected, pure PSNs exhibited the best adsorption performance. However, hydrophilic silicon hydroxyl groups on the surface restricted the application under wet conditions. In this study, we fabricated p-PSN and n-PSN with stable hydrophobic surface by replacing the silicon hydroxyl groups with phenyl and n-octyl groups, respectively. As discussed before, after functionalization, the uniform hollow structure was preserved, which was beneficial to adsorption. Thus, combined with the results of physicochemical techniques, functionalized PSNs may possess excellent adsorption performance under wet conditions.

To investigate the effects of the grafted hydrophobic surface, the PSN-based materials were tested under different RHs. As can be seen from Figure , it was noted that the obtained experimental breakthrough curves of PSNs under 50% RH was quite different from that under dry conditions. The phenomenon of roll-up (C/C0 > 1) appeared in the breakthrough curves of PSNs as is mentioned in previous reports.[13,21] Water and toluene molecules competed for adsorption at the active sites. Part of adsorbed toluene molecules were replaced by water vapor and eventually the outlet concentration of toluene increased. The good news was that this phenomenon disappeared in the breakthrough curves of p-PSN and n-PSN.

Figure 9

Toluene breakthrough curves on PSN-based materials under 50% RH.

The breakthrough times (tB) acquired from the breakthrough curves under different RHs are listed in Table . The corresponding breakthrough curves are provided in the Supporting Information. As expected, the tB values of p-PSN and n-PSN surpassed those of pure PSNs under 11 and 27% RH, respectively, in spite of comparatively lower pore volume and surface area of functionalization. This phenomenon was attributed to the abundance of grafted organic groups, indicating that phenyl and n-octyl groups might resist to water vapor. Therefore, the organic groups on the surface were a crucial factor to improve the hydrophobicity. A comparison of the adsorption capacity under different wet conditions (QWET) is shown in Figure . The values of QWET decreased with increasing RH, a sharper decline always illustrated its poor hydrophobicity. The QWET of pure mesoporous silica adsorbents (PSNs and KIT-6) decreased more rapid with RH compared with functionalized PSNs, revealing significantly poor humidity tolerance. The adsorption capacities under different conditions of PSNs were still better than those of KIT-6. In contrast, the QWET of p-PSN and n-PSN decreased gradually, and the values of QWET for p-PSN and n-PSN increased up to 1.44–1.76 times than that of the pure PSNs in 50% RH, demonstrating that p-PSN and n-PSN possessed excellent hydrophobicity under high wet condition.

Table 4

Breakthrough Time tB of all Adsorbents When C/C0 = 0.05 under Different Humidity Conditions

			PSNs	p-PSN	n-PSN	KIT-6
tB (min)a	RH	11%	30	32	28	23
		19%	24	30	23	21
		27%	20	29	23	18
		37%	18	26	21	15
		50%	15	25	20	10

tB was the time when C/C0 = 0.05.

Figure 10

Values of QWET of all the adsorbents under identical humidity conditions of 0–50% RH.

To describe the interaction between the toluene molecules and the adsorbent surface at room temperature, we calculated the Henry constant from the water and toluene static adsorption isotherm (Figure ) under relatively low pressure (P/P0 < 0.1), where the interaction between the adsorbed molecules could be ignored.[15,18] As listed in Table , the H values of water vapor followed the order PSNs (66.76) > n-PSN (18.19) ≈ p-PSN (16.69), whereas the H values of PSNs were 3 times higher than that of functionalized PSNs. The big difference demonstrated that the interaction between water molecules and the PSNs surface was strongest. The functionalized surface of p-PSN and n-PSN exhibited a smaller interaction with water molecules, further proving that the hydrophilic silicon hydroxyl groups were grafted by organic groups.[27] For static adsorption of toluene, the H values increased in the order of: p-PSN (264.20) > PSNs (224.18) > n-PSN (150.09). p-PSN presented the strongest interaction with toluene molecules. This phenomenon was caused by the π-interactions between the surface functional phenyl groups and toluene,[19,21] which was stronger than that between toluene and n-octyl or silicon hydroxyl groups. The H ratio (toluene/water) combined the influence of water and toluene; as expected, p-PSN showed the highest value. This result signified the best selectivity of toluene for phenyl-grafted surface under wet conditions, in agreement with the dynamic adsorption performance.

Figure 11

Toluene (a) and water vapor (b) static adsorption isotherms (P/P0 < 0.1) of PSN-based materials.

Table 5

Henry Constants (H) of Toluene and Water on PSN-Based Materials

	H
	toluene	water	ratioa
PSNs	224.18	66.76	3.36
p-PSN	264.20	16.69	15.83
n-PSN	150.09	18.19	8.25

The ratio was equal to values of H (toluene)/H (water).

As discussed above, for industrial applications, rapid desorption was also a critical factor. To test the desorption efficiency under 50% RH, PSN-based materials were regenerated by heating to 80 °C, keeping the air flow constant for 20 min. As shown in Figure , the QWET of all materials remained relatively steady after six adsorption/desorption cycles. p-PSN exhibited superior adsorption performance in every cycle, which was considerably higher than that of pure PSNs, demonstrating excellent stability. Grafting of organic groups on PSNs not only preserved the hollow structure, which enabled the high efficiency of toluene removal, but also improved the hydrophobicity. In summary, p-PSN showed a wide scope in toluene removal under both dry and wet conditions. The effect of structure and morphology, ways of surface modification, the amounts of organic groups, and even the regeneration conditions are worth studying in future works.

Figure 12

Comparison for the adsorption capacities of PSN-based materials and KIT-6 of six cycles under 50% RH.

Conclusions

In this study, PSN-based materials with a uniform hollow structure were successfully synthesized through a dynamic self-assembly method, and then, their hydrophobicity was further modified by postgrafting phenyl and n-octyl groups to yield p-PSN and n-PSN, respectively. The resultant PSN-based materials showed uniform morphologies with hollow spherical shapes after grafting. As supported by the dynamic adsorption–desorption performance of toluene, the as-synthesized PSN-based materials exhibited high adsorption capacities, desorption efficiency, and good stability. Under dry conditions, the dynamic adsorption capacity and the desorption efficiency (E5) of PSNs were far better than those of the traditional mesoporous silicas. Under wet conditions, functionalized PSN materials made up the defects of pure PSNs, holding the excellent adsorption capacity and reusability. The Henry constants from water vapor or toluene static adsorption of p-PSN demonstrated the best capture ability for toluene at room temperature. These advantages of p-PSN can be due to the hollow spherical structure and the phenyl groups grafted on the surface, which improved the utilization of the adsorption sites and presented stronger interaction with toluene molecules, respectively. With excellent adsorption–desorption performance and stability, the future looks bright for the application of p-PSN as an adsorbent for controlling VOCs.

Experimental Section

Synthesis of Porous Silica Materials

PSNs were synthesized via a dynamic self-assembly method.[37] CTAB (cetyltrimethylammonium bromide, 99%) was used as the surfactant, and water and diethyl ether (98%) were used as cosolvents. In a typical synthesis, 1.0 g of CTAB was dissolved in 140 mL of H2O, then 1.6 mL of aqueous ammonia (25–28%) and 40 mL diethyl ether were added into the solution, and stirred vigorously for 0.5 h at room temperature. Then, 5 mL of tetraethoxysilane (98%) was added dropwise to the solution under vigorous stirring and the mixture was continuously stirred for 4 h at room temperature. A white precipitate was obtained by filtration, dried in 60 °C, and eventually calcined at 550 °C in air for 6 h. Silicalited-1, KIT-6, and SBA-15 were prepared by the method reported in the literature.[19,32] Commercial AC was purchased from Yantai Kelier Co., Ltd.

Functionalized PSNs were performed by adding 1 g of pure PSNs to a mixture of 50 mL of toluene, 0.2 mL of H2O, 5 mL of triethoxyphenylsilane (98%), or triethoxy(octyl)silane (97%). The mixture was stirred at 90 °C for 24 h and then filtered. The resulting white solid was washed with toluene, ethanol, and deionized water and then air-dried at 100 °C. The samples grafted with phenyl and n-octyl were named as p-PSN and n-PSN, respectively.

Adsorbent Characterizations

FT-IR spectra were collected on a Thermo Nicolet 6700 with pellets consisting of accurate adsorbent samples (1 mg) and KBr (100 mg). TEM images were recorded on a JEOL 2100F microscope operating at an accelerating voltage of 200 kV. The thermal decomposition behavior (DTG) was performed on TGA8000. Each adsorbent was heated in the range of 30–800 °C with a liner heating rate of 10 °C min–1 under the flow of air. Nitrogen sorption isotherms were acquired through an ASAP 2020 gas sorption analyzer. All of the adsorbents were degassed at 120 °C for 6 h before measurements. The specific surface area was calculated by using the BET equation. The pore diameter was established based on the desorption branch of N2 isotherms by using the Barrett–Joyner–Halenda (BJH) method. The pore size distribution was estimated by the density functional theory method. The micropore surface area and the micropore volume were obtained from the t-plot method.

Dynamic Toluene Adsorption and Desorption Measurements

For the dynamic adsorption process, pelletized samples (40–60 meshes) were degassed under atmospheric pressure with heating (120 °C) for overnight. Next, about 0.1 g of the treated samples was packed in a U-type quartz tube, and quartz wool was plugged at the top and bottom of the tube as a support for the adsorption bed. During the adsorption process, air was used as the carrier gas with a total flow rate of 100 mL min–1 at 25 °C. The inlet concentrations of toluene were controlled at 1000 ppmv under both dry/wet conditions (11, 19, 27, 37, and 50% RH at 25 °C).

After the adsorption, the regeneration experiment was carried out by heating the samples with simultaneous purge air of 100 mL min–1 for 20 min at 80 °C and then rapidly cooled to 25 °C for the next continuous adsorption test. The entire process was repeated 4–6 times to investigate the desorption efficiency and reusability of the samples. The concentration at the outlet of the reaction bed was analyzed by an online gas chromatograph equipped with a flame ionization detector.

The dynamic adsorption capacity (Q, mg/g) of the samples was calculated by the following equations[21]where FA and W are the molar flow of toluene gas and the weight of the adsorbents, respectively, CA/C0 is the relative concentration of toluene at time t, respectively, and tD is the dead time of the experimental setup.

The total desorption capacity for each cycle was calculated as followswhere fB and W are the volume flow of air and the weight of the adsorbents, respectively, CD is the concentration of toluene outside the adsorption bed at time t, respectively, and tD is also the dead time. The desorption efficiency was equal to the value of Qd/Qa.

In this study, the dynamic breakthrough curves of toluene are analyzed with the Yoon–Nelson model (the Y–N model) to evaluate the adsorption performance of the mesoporous silicas. The Y–N model is a widely used kinetic theory model for VOC adsorption with the equation as follows[32]where C0 and C (ppmv) represent the inlet and outlet concentrations at time t (min), respectively, and τ (min) and k (min–1) refer to the half-life for C = 50% C0 and the rate constant, respectively.

Static Adsorption Measurements

Static adsorption performance of toluene and water for PSN-based materials was evaluated by using an intelligent gravimetric analyzer (IGA-100B, Hiden Isochema Ltd) at 25 °C. This equipment has an ultrahigh vacuum system, determining isotherm and the corresponding adsorption kinetic by setting exact pressure steps. Before testing, the samples were degassed under vacuum condition with heating (120 °C) for 8 h.

Under the relatively low pressure (P/P0 < 0.1), Henry’s law was used as follows[18]where q (%) and H represent the adsorbed amount of molecular and the Henry constant, respectively.