Fabrication of silica aerogel and carbon–silica composite for dynamic adsorption of benzene from dry and wet gas streams

ABSTRACT In the present study, the white kaolin-derived silica (SA) and silica-carbon aerogel composite (CSA) were successfully synthesised as useful adsorbents for the effective removal of benzene from dry and wet gas streams. The physicochemical features of the adsorbents were characterised using SEM, BET, XPS, and TGA, and the adsorption properties of benzene were investigated through breakthrough time curves and equilibrium study. The specific surface areas of SA and CSA were 628 and 783 m2 g−1, respectively, demonstrating the high porosity of the synthesised adsorbents. The dynamic adsorption and isotherm performances of SA and CSA confirmed that both adsorbents have superior adsorption properties and Khan isotherm for both developed adsorbents were the fittest model among the two- and three-parameter isotherm models studied and provided a maximum adsorption capacity of 366.2 and 224.4 for SA and CSA mg g−1, respectively. Furthermore, the reusability tests revealed that the SA and CSA could be efficiently recovered at 180°C. Our experimental outcomes indicated that the prepared SA and CSA can be applied as efficient and inexpensive adsorbents for the removal of benzene from gaseous streams.


Introduction
BTEX, the notorious subgroup of volatile organic compounds (VOCs), including benzene and derivatives from benzene, is released from natural and man-made sources and has severe consequences on the environment [1].The emissions of benzene from motor vehicles, industrial plants, solvent spills, and solid waste disposal systems could negatively affect human health [2,3].Benzene has been classified by the Agency for Toxic Substances and Disease Registry (ATSDR) as a carcinogen for humans (Group 1) [4].Additionally, benzene and its derivatives have significant impacts on atmospheric chemical reactions, and play a remarkable role as precursors in the formation of peroxyacetyl nitrate (PAN) and secondary organic aerosols (SOA) [5].For instance, tropospheric O 3 is a secondary pollutant produced in the presence of sunlight through chemical reactions between oxides of nitrogen (NOx) and BTEX.The European Union (EU) has established 5 µg m −3 as a reference value for the annual average of benzene in outdoor air and suggested the decrement in benzene percentage in petrol from 5% to 1% [6].
There are various technologies extensively applied for benzene removal from air flow, such as catalytic and photocatalytic degradation, oxidation, biological remediation, and adsorption [7,8].Among these removal technologies, the most well-known and widely used method is adsorption, exclusively owing to its low cost, simple application and design, and excellent efficacy in the removal of benzene [9].Obtaining the ideal adsorbents with low-cost precursors, the proper proportion between microporous and mesoporous networks, and appropriate reusability are crucial for the adsorption process [10].
Several adsorbents have been studied for the adsorption of benzene air samplesfor example, activated carbon (AC), nano adsorbents, carbon nanotubes (CNTs), silica aerogel (SA), ordered mesoporous carbon (OMC), carbon aerogel (CA), and carbon-silica composites (CSCs) [11][12][13][14][15]. Carbonaceous adsorbents are recognised as qualified adsorbents because of their unique properties, ability to adsorb a wide range of pollutants, stability in numerous cycles, and environment-friendly nature.Nevertheless, some serious obstacles decrease their commercial applications, such as pore obstruction and complexity in activated carbon, weak transmission for VOCs with larger molecular sizes in Biochar, bulk density problem in CNT and low water-resistant capability related to functional groups in OMC [16,17].
SA is a unique adsorbent from silica-based mesoporous family with the specific surface area in the range of approximately 500-1200 m 2 g −1 , flexibility in hydrophobicity, a low density of about 0.003 g cm −3 and high porosity [18].These characteristics result in easy accessibility of the inner surfaces and being effective for benzene removal from the air stream.Alkoxides (RO − ), for instance, tetramethoxysilane and tetraethoxysilane, are the primary precursors for the synthesis of silica aerogels.The main limitations of their synthesis with alkoxides are the toxicity and high price of these organic precursors [19].
To address these problems, sodium silica solution as a substitute for the alkoxides has been suggested in previous studies.White kaolin clay (WKC) is a clay mineral mainly composed of SiO 2 (almost 80%), which could be utilised as a low-priced precursor for the production of sodium silicate [20].For increasing the structural strength and overcome diffusional paths, CSCs with large specific surface area and elevated ignition temperature could enhance the adsorption process performance [21].Finding the superior adsorbent or modification techniques in existing adsorbents by considering factors such as lower cost, higher efficiency, and effective physico-chemical properties is essentially required to improve the performance of the benzene adsorption process [22].This paper aims to investigate the low-cost synthesis and characterisation of silica and silica-carbon aerogels for measuring breakthrough curves of benzene adsorption at dry and wet air streams in a fixed-bed column.SA and CSA were prepared by employing WKC and WKC with resorcinol-formaldehyde (RF), respectively.Furthermore, several tests have been performed to achieve equilibrium adsorption isotherms, by determining the optimal conditions from the breakthrough curves.

Silica aerogel
To achieve hydrophobic SA, WKC was used as a precursor in the sol-gel procedure [23].Figure 1 depicts a scheme of SA preparation.WKC was primarily washed with distilled water and screened with the mesh of 140 to remove impurities.Subsequently, it was dried at 100°C for 2 h.In the second step, 200 ml of 1 M sodium hydroxide solution and 4 g of the dried WKC were mixed under 110°C for 8 h.Afterwards, the temperature of the mixture decreased to 25°C and the obtained mixture underwent vacuum filtration with CHM® filter papers (F2040 grade) to separate fine particles.In the next step, the filtered solution (sodium silicate) was applied for the synthesis of SA.Consequently, for replacing sodium ions with OH ions and obtaining the main precursor (orthosilicic acid: SiOH 4 ) for aerogel synthesis, strongly acidic resin with a 1:1 volume/volume proportion was added to the sodium silicate solution.Following the ion-exchange phase, the solution was stirred for 8 min and simultaneously, pH gradually declined from 13.5 to 2.9.The used amberlite resin was then separated and washed with distilled water until the pH of the effluent increased to neutral pH.In addition, the pH of silica sol was adjusted to 4 by adding 1 M sodium hydroxide for gelation.Thereafter, silica sol was transferred to cube polyethylene container with a tight lid, and then, 5.5 M ammonium solution, as the base catalyst, was added to the container and underwent ageing in 60 min at 65°C to form the hydrogel.In the final step, due to silylating groups, which are unsolvable in water, the hydrogel for the solvent exchange step was submerged in acetone at 25°C for 24 h in which the fresh solvent was substituted every 8 h.Furthermore, the ETMS:ethanol as silylating agents (replacing Si-OH to Si-R 3 ) were used in a volume ratio of 2:1 at 45 °C for 36 h to improve surface properties.Subsequently, the wet gel was prepared for the drying step by immersing two times in hexane solution to make sure of elimination of the modification factors.Ultimately, the gels were dried via supercritical drying with CO 2 gas in 60 bar.

Carbon silica aerogel
In a unique approach to the synthesis of carbon-silica adsorbent, two solutions that are precursors of carbon aerogel and SA were vigorously mixed to obtain a homogeneous mixture, as shown in Figure 2. At first, solution one was prepared with R:F:C in 2:1:300 molar proportion that was added to AEtOH:ETMS with 2:1 molar ratio under magnetic steering at 25°C.Prior to beginning the RF gel formation, solution one was combined with solution two, which was prepared using silicic acid derived from WKC and sodium hydroxide.Thereafter, acetic acid (C 2 H 4 O 2 , 0.1 g of 50% by weight) was provided as a secondary catalyst to decrease the pH of the final solution to nearly 4. Before gelation, the mixture was stirred until a light reddish-brown transparent solution appeared.Subsequently, the mixed sol was transferred into a PFA container with a screw cap (with a diameter of around 30 mm) and then aged at 80°C for 12 h in an oven for gradual gelation [24].Following the ageing step, the hydrogel was concurrently washed and immersed in acetone for 24 h.The fresh solvent was replaced three times a day.For the drying step, the gel was dried at −0.05 MPa and 120 °C in a vacuum drying oven (OV-12, Jiao tech, South Korea) for 12 h.The dried gel was placed in a tube furnace for carbonisation and activation at 950°C for 2 h under a flowing CO 2 high-purity gas (80 ml min −1 ) [25].

Characterisation of fabricated adsorbents
The morphology of the adsorbents was investigated utilising Hitachi scanning electron microscopy (SU3500, Japan).The pore volume, specific surface area, and pore size distributions of aerogels were measured via a nitrogen adsorption−desorption setup (Tristar II plus, Micro metrics, USA) at 77°K using the Bruner-Emmett-Teller (BET) and Barrett-Joyner -Halenda (BJH) analysis.Fourier transform infrared spectroscopy (FT-IR) was applied using an IRTracer-100 (Shimadzu Corporation, Japan) utilised to review the chemical bonding state of SA and CSA.Energy-dispersive X-ray analyser (EDX)-mapping images were obtained employing scanning electron microscopy (SEM) equipped with an energy-dispersive spectrometer (EDS) (XFlash 6130, Bruker Co., USA).The XPS (X-ray photoelectron spectroscopy) spectrophotometer (K-alpha+, Thermo Fisher Scientific, USA) was used for providing the details related to the chemical environment and elemental composition of SA and SCA employed with an Al Kα micro-focused (radiation of 1486.6 eV).The thermal stability properties were investigated by thermal gravimetric analysis (TGA).Moreover, for measuring the heat flow rate, the differential scanning calorimetry analysis (DSC) was performed with a thermogravimetric analyser (SDT-Q600, TA Instruments, USA) under the heating rate of 10°C min −1 to 1000°C.Furthermore, the pilot-scale design of dynamic adsorption, benzene measurement and specific details of the utilised devices are located in the supplementary file.

Dynamic adsorption model fitting and equilibrium isotherms
In the present study, three parameters were investigated under dynamic conditions, namely temperature, concentration, and water vapour.Furthermore, breakthrough curves for benzene removal under these various conditions, as well as the effect of these parameters on the adsorption capacity, were discussed in detail.For the calculation of the area above the breakthrough curves, the dynamic adsorption capacity (q) equation was used.The equation and parameters of the adsorption capacity are defined in Eq.1 [26].
Where C 0 is the inlet concentration (mg m −3 ), C is the outlet concentration (mg m −3 ), Q is the gas flow rate (l min −1 ), t s is the saturation time (min), t b is the initial time (min), m is the mass of adsorbent (g), and q is the dynamic adsorption capacity (mg g −1 ads).
The equilibrium adsorption tests were performed by filling the determined adsorbent quantity into the column at the optimal temperature and concentration under variable pressures of the carrier gas.Eventually, the equilibrium adsorption data were used by employing the two-and three-parameter isotherm model to investigate the fit with models for the adsorption of benzene [27,28].Six non-linear isotherm models (Langmuir, Freundlich, Dubinin-Radushkevich, Redlich-Peterson, Sips, and Khan) were studied to predict the experimental data.The equations associated with the isotherms are listed in the supplementary data as Table S6.

Morphologies of fabricated adsorbents
Figure 3(a,b) shows SEM micrographs of developed adsorbents with supercritical and vacuum drying in this study.Figure 3(a) depicts the modified SA as a porous and nonagglomerated structure, which is composed of spherical nanoparticles with a heterogeneous appearance with spongy structure.The micrograph of the SA also indicated the modification phenomenon (last step in SA preparation) as an impressive surface adjustment that decreased the number of cracks and prevented condensation.The mesoporous structure would be formed because of the spring back effect (reversible shrinkage) that happened in the wet gel of SA during supercritical drying due to ETMS modification of the surface [29].structure of disordered particles with a highly porous look and numerous deep cavities [30].However, the nanoparticles of CSA were non-spherical and had a partial aggregation, in fact, possibly due to the presence of Si-C bonds, which expedites carbothermal reduction reactions [31].The particle size of CSA was larger than that of SA, which represented a compressed mesoporous network, whereas silica and carbon frameworks were segregated and uniformly distributed.

Pore Characteristics of fabricated adsorbents
In order to understand the N 2 adsorption-desorption isotherms, pore structures and BJH methods were used for the pore size distribution of SA and CSA, as presented in Figure 4  (a,b).The isotherms of SA and CSA in Figure 4(a) could be classified as type IV isotherm, according to the BDDT (Brunauer, Deming, Deming, and Teller) classification [32].Both isotherms presented a parallel and vertical adsorption-desorption model, yet their different pore characterisation due to the hysteresis curve classification in CSA and SA, which was related to H1 and H3 loops, is apparent in Figure 4(a).In both CSA and SA adsorbents, the presence of H1 and H3 hysteresis curves corresponding to pore type is attributed to tubular and wedge-shaped structure pores, respectively [33].The SA sample had a hysteresis loop at a relatively high pressure (P/P 0 > 0.7); therefore, the existence of a B point at this pressure indicated high monolayer-multilayer capacity.The steep increment and capillary condensation step at high relative pressure (P/P 0 < 0.5) in CSA indicated the formation of worm-like mesopore channels in this synthesised adsorbent.Additionally, Figure 4(b) shows a significant alteration related to the major pores of the synthesised SA known to be mesopore channels; these pores were under 20 nm and the dominant pores were around 11 nm.Meanwhile, the highest pores of CSA, known to be mesopores, were above 30 nm and the prevailing pore diameter size was about 38 nm.As previously mentioned, high capillary mesoporous condensation and unrestricted multilayer adsorption after completing the monolayer adsorption were expected in SA and the nitrogen adsorption-desorption outcomes demonstrated pieces of evidence on the presence of both macrosporous and mesoporous structure in CSA.Table 1 summarises the physical properties of the SA and CSA samples.Based on Table 1, despite the higher specific surface area in the CSA adsorbent, the micropore volume in CSA is much lower than that in the SA adsorbent.Accordingly, due to the greater contribution of micropore volume in the uptake of benzene and the main role of a proportion of micropores to mesopores in benzene adsorption, more adsorption capacity is expected in the SA adsorbent.

Fourier transform infrared and X-ray photoelectron spectroscopy
The FTIR Spectroscopy graphs for both adsorbents are given in Figure 5. Figure 5 exhibits the characteristic peaks in SA to be about 449 cm −1 and 1091 cm −1 , which were due to Si-O-Si for the general silica network, denoting the asymmetric and bending forms of Si-O-Si vibrations [34].Considering SA spectra, three absorption signatures could be recognised at 756 cm −1 , 848 cm −1 , and 1261 cm −1 , and these three peaks appeared due to surface modification, which was affected by the Si-C bond.The peaks around 2982, 2891, and 862 cm −1 were assigned to the symmetric and antisymmetric of the -CH 3 groups, which indicated the presence of the silane (hydrophobic agent) in the silica gel structure, thereby providing the aerogel hydrophobic behaviour.The narrow peaks that appeared at 1628, 953, and 3600 cm −1 are referred to as the O-H groups and were related to the physically adsorbed water molecules on the surface of the aerogel.The spectrogram of CSA contains a characteristic signature at the wavenumbers of 1076 cm −1 and 450 cm −1 that were attributed to siloxane bands [35].However, the absorption peak arose from Si-O-Si to around 1076 cm −1 in CSA, probably covered by the broadband of C-O at 1100 cm −1 .In addition, the CSA spectrum is illustrated with two peaks near 3600 cm −1 and a peak at 1650 cm −1 , which could be associated with O-H groups.
To investigate the quantitative data of the compositions of aerogels, Figure 6(a,b) illustrates high-resolution XPS spectra of SA and CSA samples, respectively.The data obtained employing the Lorentzian-Gaussian function for SA and CSA aerogels are presented in Table 2.The information from the survey spectra primarily included the main elements, namely Si, O, and C, with a limited amount of Na in the SA sample and C, O, and Si in the CSA sample.As shown in Figure 6(a), the SA sample had a C peak with a binding energy of 284.4 eV and with the O peak at 534 eV along with the 103.5 and 155 eV peaks, which were related to Si [36].It was observed that in the SA sample with    a characteristic C1s peak, the C-O(-Si) bond at 284.2 eV occupies 81.2%, whereas the ratio of the C-C(-H) bond was approximately 0. This proved that -CH 3 groups existed in silica, was also confirmed by the FTIR analysis as explained previously [34].We revealed that in the Si2P chart, the peaks at 103.6 eV and 104.6 eV corresponded to Si-O-Si and Si-O bonds in the SA sample, respectively.
Figure 6(b) illustrates the survey scan of CSA.It proved the incorporation of silica and carbon aerogel precursors in the CSA sample.Given the Si2P graph in Figure 6(b), the characteristic peaks at 102 eV and 103.2 eV were associated with C-O-Si and O-Si-O bonds, respectively.The C1s peak observed at 284.6 eV corresponded to the three components that were in 284.2, 285.2, and 286.2 eV binding energies attributed to sp2, sp3, and C-O bonds, respectively [37].It shows that the presence of an sp3 peak at 285.2 eV in the CSA sample is associated with the disordered structure of the sample [38].

Identifying the elemental compositions in the adsorbents
The EDX analysis graphs for both SA and CSA samples are presented in Figure 7.   solvent exchange process, the impurities were successfully eliminated.Furthermore, according to SEM images of the SA sample in supporting information, after the elimination step, the residuals had a pearl-like morphology, which is expected for water glassbased aerogels [39].A limited portion of Na appears in Figure 7(a), denoting that the quantity of Na inside the SA sample was quite low. Figure 7(b) shows the Si, C, and O elements distribution in the CSA sample; they are interlinked on the microscopic scale.Element mappings of CSA exhibited the main elements that were relatively uniformly distributed in the aerogel matrix with some agglomeration in Si situation.Quantitative raw data on chemical elements of both samples are presented as Supplementary Information files.

Thermal properties and stability
Thermal stability analysis was conducted at a heat rate of 0 °C to 1000 °C.The reason for using this temperature range is the frequent use of these adsorbents in reusability.This is because in practical scenarios, the amount of heat applied during desorption should be much higher than the adsorption operation.The TGA and DSC plots of SA and CSA samples are shown in Figure 8(a,b), respectively.In this figure, the loss of weight in the first level (25 °C to 105 °C) was 1.8% to 2% due to the exhaust of water adsorbed and the remaining solvents from the surface of the SA sample.A steep slope was then observed in the curve because of the oxidation (at temperatures above 200°C) of the residual organic materials in the SA [40].Afterwards, with the rise in the temperature, a considerable weight loss occurred during calcination, which was related to the oxidation of CH 3 groups on SA [34].In addition, the exothermic peak at 479°C indicated that the SA sample could change to be hydrophilic by calcining over this high point.Figure 8(b) represents the TGA and DSC curves of the CSA sample.The first stage of 4% weight loss, which was due to the release of residual organic substance and adsorbed water from the surface of the sample, occurred from 56°C to 190°C.Moreover, a weight loss of around 28% was observed, whereas the temperature was incremented to approximately 700°C.This weight loss essentially resulted from the pyrolysis process of substances that have NH2, C = C, and C-H groups [22].Weight loss at temperatures of over 750°C of CSA sample could be attributed to the carbothermal reaction, which causes phase alteration [41].

Dynamic adsorption behaviours
Breakthrough curves illustrated the breakthrough time and saturation time in the plot, which was the mass transfer zone reaching the end of the adsorbent bed depending on the time [42].To investigate the impact of the concentration of benzene on the adsorption capacity and breakthrough time, the concentrations were gradually changed from 16, 25.5, 32, and 48 mg m −3 at the flow rate of 0.4 l min −1 , and the temperature was 25°C.
The related data are given in Figure 9(a,d), and Table 3. Normally, the curves with a good 'S' shape and a longer breakthrough time possess a greater dynamic adsorption capacity.
In Figure 9(a), we could see the expected shifts in the shape of the curves from 16 to 48 mg m −3 possibly related to the interlayer interactions of molecules and intermolecules that correspond to passing specified concentrations through the column [43].As it can be observed from the concentration curves in Figure 9(a,d), at low concentrations in both SA and CSA adsorbents, the bed was saturated more slowly than at high concentrations.As observed in Figure 9(a), once the concentration increased from 16 to 48 mg m −3 , there was just a shift towards the lowest saturation time of the filter, with no alteration in the shape of the curve.Based on Figure 9(d), it is obvious that the breakthrough curves of CSA had different shapes and trends from SA curves.The CSA breakthroughs were saturated very quickly and had a sudden and drastic behaviour to reach the saturation time point.Consequently, due to the pressure gradient in macropores of the CSA, gas flow passed more expeditiously into the microfractures based on Fick's diffusion law [44].It could be concluded that the CSA with a high total pore volume does not necessarily have better performance compared to SA and the presence of micropores and mesopores in the adsorbent structure is essential, which is directly leads to the increase in efficiency [24].As mentioned in Table 3, with the increase in concentration from 16 to 48 mg m −3 , some phenomena, like molecules branching or adhering to other molecules, might occur; the adsorption capacity then incremented.

Effect of temperature
Figure 9(b,e) display the correlation between the sorption of benzene and temperature conducted at different temperatures of gas flow from 25, 40, 50, and 65°C at 0.4 l min −1 and 48 mg m −3 .It could be observed that with the increment in the temperature from 25°C to 65°C, the adsorption ability reduced, followed by a decrease in the adsorption capacity in both adsorbents.Normally, lower feed concentration and increased operating temperature result in a decrease in the absorption capacity, which can be seen in Table 3.
The increased efficiency of the adsorption process in reliance on the diffusion of adsorbate molecules on the adsorbent layers was considerably influenced by high temperatures [45].Reduction of the uptake of benzene by the SA and CSA adsorbents at higher temperatures could be assigned to two main reasons: increasing volatility and possibly the loss of active sites [46].Therefore, based on the outcomes of the present tests, 25°C was found to be an efficient temperature.Accordingly, the experiments and isotherms were performed at this temperature.

Effect of water vapour
To figure out the impact of different levels of humidity on the adsorption capacity and breakthrough of benzene adsorption, the relative humidity of 0, 45, 55, 65% and 48 mg m −3 of benzene passed at a constant gas flow rate of 0.4 l min −1 at 25°C.The data are displayed in Figure 9(c,f).The breakthrough curves explain that the SA sample in a high moisture content environment had a sharp shape and immediately moved upwards, which shows that the bed was saturated very quickly.The difference in the capacity of benzene adsorption concerning the humidity of the environment in Table 3 between the SA and CSA samples is clearly found and feasible to compare.As presented in Table 3, the presence of carbon-silica in CSA provided a large improvement in the adsorption amounts in humidity conditions, while the adsorption capability of the SA sample decreased significantly compared to CSA.In addition, the better adsorption of benzene in different humidities by the CSA was affected by the hydrophilic characteristics of carbon aerogel, which indicated benzene has enough time to diffuse into the pores [24].This result, coupled with a high proportion of mesoporous structure and other outcomes, revealed the highest performance of CSA under higher humidity than SA.In SA, with the increase in the level of water vapour, the adsorption capacity reduced, which presumably corresponded to the decrease in specific surface areas.Adsorption on micropores was restricted through the ultrafine microporous presence, while they did not absorb humidity from the gas flow [46,47].In contrast to SA adsorbent, the CSA sample with a higher BET surface area had a higher adsorption capability of benzene.All these declared that SA and CSA can be employed as high-protentional adsorbents to eliminate benzene and other aromatic compounds from gaseous streams.

Reusability
The ability to reutilise the SA and CSA was investigated for several adsorption-desorption cycles as shown in Figure 10.Furthermore, Table 4 lists the achieved parameters.All adsorption-desorption cycles were conducted under the optimal temperature and initial concentration conditions of 25°C and 48 mg m −3 at the dry flow gas.In Figure 10(a,b), the benzene adsorption capacities of the SA and CSA adsorbents declined considerably in the second and third cycles.The inert gas (N 2 gas, 25°C) for purging physically adsorbed benzene was used in cycles 2 and 3 for 10 min.The adsorbate benzene molecules, which are held by physical forces such as van der Waals forces, separated from the surface of the studied adsorbents.Meanwhile, heat treatment dramatically improved the active sites and discharged the chemisorbed molecules, which were applied to SA and CSA desorbing after the third cycle.Moreover, based on FTIR and XPS analysis, the hydrogen bond and π- π bond (which are mainly done by a phenyl ring) have been confirmed in the interaction between benzene and the studied adsorbates [52].Besides, pore-filling can be another mechanism involving the sorption process, in which mesopores and micropores have an important contribution; and due to the porous characteristics of our sorbents, the mentioned mechanism is definitely one of the main mechanisms of the adsorption of benzene molecules into the SA and CSA.The benzene adsorption capacities of SA and CSA adsorbents incremented continuously with rising thermal desorption temperature, confirming chemisorption and physisorption practice that plays the main role in the adsorption process in the present work.
In addition, the benzene desorption capability of both SA and CSA rose when the thermal desorption temperature increased.An increase in the temperature caused the promotion of the dispersion and migration of benzene and the increment in stimulation of molecular movement [53].Additionally, there is a kind of interaction between benzene and adsorbents that is restrained by high temperatures [40].Nevertheless, in Figure 10(a), in the sixth cycle, capillary condensation expanded in micropore structures, which increased the amount of remaining benzene and affected the recovery of SA.Nevertheless, the rate of reduction in the benzene adsorption capacity on the CSA sample was obviously more than that of SA at equal temperatures and required a higher temperature for desorption of benzene.The reason for the difficult separation of benzene in CSA could be the synergistic effect of partial replacement of C-C bonds with Si-C bonds and SiO2 compounds, which improve mechanical strength, physical properties and thermal insulation characteristics [46].The considerable reusability of SA and CSA due to the high capacity and high thermal stability in adsorption-desorption cycles provides a feasible strategy for utilising these adsorbents in the removal of benzene from gas flows.

Conclusions
We prepared silica and silica-carbon aerogel from natural white kaolin clay and resorcinolformaldehyde with supercritical and vacuum drying methods, respectively.The fabrication of a SA adsorbent with a pearl-like shape and graphite nature of CSA adsorbent with an average pore diameter around 8.4 and 24.3 nm were approved by SEM and BET analysis.Furthermore, surface analytical techniques and surface property analysis revealed that the breakthrough time curves were almost associated with both graded pore size, being longer for micropore presence, and surface chemistries.The adsorption isotherms achieved from the SA and CSA adsorbents were mainly fitted to the experimental data and represented that the dynamic adsorption process was favourable.The highest adsorption capacities of the Khan isotherm pattern for SA and CSA were 366.2 and 224.4 mg g −1 , respectively, which are higher than those investigated in the previous literature.The degree of decrease in adsorption capacity in adsorption-desorption performance evaluation of the reused SA and CSA after sixth and seventh cycles were 1.7% and 3.5%, respectively.It can be concluded that the prepared adsorbents were appropriate yet effective for air treatment based on their higher adsorption capacities and low desorption temperatures.Furthermore, low-cost synthesis leads to a lower impact on the environment and significant economic savings.However, a detailed study on the application of the SA and CSA for the removal of other volatile organic compounds (VOC) must be performed.

Figure 1 .
Figure 1.Flowchart of the synthesising process of the SA sample.

Figure 2 .
Figure 2. Flowchart of the synthesising process of the CSA sample.

Figure 3 .
Figure3(a,b) shows SEM micrographs of developed adsorbents with supercritical and vacuum drying in this study.Figure3(a) depicts the modified SA as a porous and nonagglomerated structure, which is composed of spherical nanoparticles with a heterogeneous appearance with spongy structure.The micrograph of the SA also indicated the modification phenomenon (last step in SA preparation) as an impressive surface adjustment that decreased the number of cracks and prevented condensation.The mesoporous structure would be formed because of the spring back effect (reversible shrinkage) that happened in the wet gel of SA during supercritical drying due to ETMS modification of the surface[29].Figure3(b) demonstrates the CSA sample that presented a typical porous network structure similar to SA; meanwhile, CSA had the framework

Figure 4 .
Figure 4. (a) The N 2 adsorption-desorption isotherms and (b) BJH pore size distribution of SA and CSA adsorbents.
Sample BET surface area (m 2 g −1 ) a Average pore diameter (nm) Characterised by using Brunauer-Emmett-Teller (BET) method.b Calculated by using of the t-plot -De Boer method.

Figure 5 .
Figure 5. FTIR spectrum of SA and CSA adsorbents.
Figure 7 (a,b) exhibit the elemental mapping for SA and CSA samples.In the SA sample, only three elements of Si, O, and Na could be detected.Therefore, in surface modification and

Figure 7 .
Figure 7. SEM and EDX spectrum of SA (a) and CSA (b).

Figure 8 .
Figure 8. TG-DSC curves of the SA and CSA adsorbents.

Figure 9 .
Figure 9. Breakthrough curves of benzene adsorption for SA and CSA adsorbents; (a) effect of concentration on benzene adsorption at 25°C and dry stream for SA adsorbent; (b) effect of temperature on benzene adsorption at 48 mg m −3 and dry stream for SA adsorbent; (c) effect of water vapour on benzene adsorption at 25°C and 48 mg m −3 for SA adsorbent, (d) effect of concentration on benzene adsorption at 25°C and dry stream for CSA adsorbent, (e) effect of temperature on benzene adsorption at 48 mg m −3 and dry stream for CSA adsorbent, (f) effect of water vapour on benzene adsorption at 25°C and 48 mg m −3 for CSA adsorbent.

Figure 10 .
Figure 10.Benzene breakthrough time curves on the SA (a) and CSA (b) adsorbents through continuous adsorption-desorption cycles.

Table 1 .
Pore characteristic comparison of the SA and CSA samples.

Table 2 .
Surface chemical composition of the SA and CSA samples from X-ray Photoelectron Spectroscopy analysis.

Table 4 .
Adsorption-desorption performance evaluation of the reused SA and CSA adsorbents.Degree of decrease in the adsorption capacity (mg g −1 ) CSA* Degree of decrease in the adsorption capacity (mg g −1 )