Modified Silica Adsorbents for Toluene Adsorption under Dry and

Jun 14, 2019 - On the other hand, the dynamic adsorption behavior of VOCs and the ..... triphenyl groups exhibit the best improvement for toluene adso...
0 downloads 0 Views 2MB Size
Download PDF
Article Cite This: Langmuir 2019, 35, 8927−8934

pubs.acs.org/Langmuir

Modified Silica Adsorbents for Toluene Adsorption under Dry and Humid Conditions: Impacts of Pore Size and Surface Chemistry Shuai Liu,† Yue Peng,*,† Tao Yan,† Yani Zhang,† John Crittenden,‡ and Junhua Li† †

Downloaded via BUFFALO STATE on July 27, 2019 at 15:55:30 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

State Key Joint Laboratory of Environment Simulation and Pollution Control, National Engineering Laboratory for Multi Flue Gas Pollution Control Technology and Equipment, School of Environment, Tsinghua University, Beijing 100084, China ‡ Brook Byers Institute for Sustainable Systems and School of Civil and Environmental Engineering, Georgia Institute of Technology, 828 West Peachtree Street, Atlanta, Georgia 30332, United States S Supporting Information *

ABSTRACT: The pore surface structure and chemistry of the ordered mesoporous silica adsorbent (KIT-6) were modified for toluene removal, including the substituent groups on the silica surface and the pore size uniformity. Uniform pore size was obtained by low-temperature heat treatment of the template at 200 °C. In the dynamic adsorption process, better channel uniformity led to higher adsorption capacity under no humidity condition because of faster pore diffusion. However, better channel uniformity resulted in poor hydrophobicity under high humidity condition because it favors the adsorption of water vapor to a greater extent. The result of Biot number indicated that triphenyl-grafted KIT-6 had a faster intraparticle mass transfer rate than pure KIT-6. Triphenyl-grafted KIT-6 had a higher adsorption capacity for toluene as compared to phenylgrafted KIT-6 under no humidity condition because of its higher surface area (SA). The higher SA was owing to the low modification of phenyl, which was caused by the isolated grafting of silicon triphenyl rather than a more even coverage by silicon phenyl. As a result, triphenyl-grafted KIT-6 exposed more hydrophilic Si−O−Si groups and therefore was less hydrophobic than phenyl-grafted KIT-6.

■

INTRODUCTION Volatile organic compounds (VOCs), serving as a significant precursor of aerosol particles and photochemical smog in the atmosphere, are discharged from various industrial fields and therefore need to be controlled without delay to avoid more damage to human health and ecological environment.1−4 Aromatic VOCs are important because they are present in gasoline and at fossil fuel extraction sites which can be close to populated areas.5 Therefore, various control methods have been developed for the removal of VOCs, such as thermal catalysis, plasma oxidation, condensation recovery, membrane separation, liquid absorption and adsorption, and so forth. Among these, adsorption is a widespread control method because of its convenient operation, low-energy consumption, and efficient purification. Numerous adsorbents such as activated carbon,6,7 zeolites,8,9 and metal−organic frameworks10 have been used for VOC adsorption, but all have their own disadvantages such as bad thermal stability and hydrophilicity.11 Recently, ordered mesoporous organosilica has been found to be a promising adsorbent to solve these problems, as reported in the previous work.11−14 © 2019 American Chemical Society

However, there are still some issues to improve in the VOC adsorption process using mesoporous organosilica. Preparing mesoporous organosilica by grafting organic groups is a useful way to improve the hydrophobicity of the adsorbent.11,13,15 In general, the molecular formula of the organosilica precursor is R−Si(OC2H5)3 (R: organic group), which has one organic group and three Si−OH bonds after hydrolysis. The precursor such as R3−SiOC2H5 has also been used for the preparation of hydrophobic organosilica adsorbents.16 However, the difference is that the latter has three organic functional groups and one Si−OH bond after hydrolysis. This difference has a significant influence on the surface chemistry and textural property of the organosilica. However, the influence of the difference has not been elucidated. On the other hand, the dynamic adsorption behavior of VOCs and the impact factor have been studied recently.17−22 Qin et al.21 and Wang et al.22 both reported that the Received: April 8, 2019 Revised: June 11, 2019 Published: June 14, 2019 8927

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Article

Langmuir

distribution and pore diameter (DP) were established by the nonlocal density functional theory method. The total pore volume (VP) was worked out in the p/p0 of about 0.99. The micropore surface area (SMA) and micropore volume (VMP) were ensured by the t-plot method. The thermolysis behaviors (TG/DTG) were recorded on a TGA/DSC 1 STARe under an air flow of 50 mL min−1. The temperature range was: (1) from 30 to 800 °C for phenyl-modified silica with an increasing heating rate of 10 °C min−1; (2) from 30 to 1100 °C for pure silica with an increasing heating rate of 10 °C min−1; and, (3) from 30 to 550 °C for pure silica with the P123 template at an increasing heating rate of 2 °C min−1. The desorption behaviors of saturated adsorbents were under a N2 flow of 50 mL min−1 from 30 to 400 °C with an increasing heating rate of 5 °C min−1. The functional groups of silanol and phenyl on the adsorbent surface were observed with Fourier transform infrared (FT-IR) spectra of a Nexus870, and the testing sample was prepared by mixing 1 mg of adsorbent and 100 mg of KBr powder, and the samples went through a heating treatment of 120 °C for 5 h before the measurement. Dynamic Adsorption Measurements. The breakthrough curves of toluene adsorption were investigated by FID in a 7890A gas chromatograph (GC) system. The particle size of testing adsorbents was 40−60 meshes, and all samples would be pretreated at a vacuum environment of 200 °C for 4 h before measurement. The outgassed sample (100 mg) was loaded into the fixed bed (the internal diameter of 6 mm and the height of 15 mm). The testing conditions: 25 °C, 1000 ppm of toluene with 100 mL min−1 of N2 flow, and the relative humidity (RH) was 90% for the measurement under high humidity conditions. The total dynamic adsorption capacity of toluene [Qdry or wet (mmol g−1)] was established by the following equation

morphology and the pore size of the porous silica dramatically affected the breakthrough behavior during the dynamic adsorption process of VOCs. The differences of pore structures, such as crystalline phase,17 spatial group structure,13 and hierarchical pore structure,18 also had a great impact on the dynamic adsorption behavior. Another difference of the pore structure is in the pore size uniformity; however, this difference of the nanostructure has not been examined. Besides, the study on the mass transfer mechanism is significant for the adsorption removal of VOCs, such as the analysis of the homogeneous surface diffusion model (HSDM). We can determine whether the VOC adsorption is limited by the external or intraparticle mass transfer by using this model.36−38 However, the analysis of the rate-determining step has rarely been explored in recent research. Based on this literature review, we created several new nano silica adsorbents. On the one hand, pure silica substrates with different pore size uniformity were created. First, it has been reported that washing with ethanol to remove the template could decrease pore homogeneity.23 Consequently, to obtain an adsorbent with uniform pores, it was essential to eliminate the template using calcination and avoid the decrease of pore homogeneity which is caused by washing. Moreover, a lowtemperature heat treatment below 200 °C should be used to prevent the condensation of surface silanol. Yang et al.24 reported that the template in the silica adsorbent could be removed at a very low temperature (200 °C). Therefore, we used thermogravimetric (TG) analysis (Figure S1) to help guide the process using low-temperature heat treatment at 200 °C for obtaining pore size uniformity. On the other hand, we also used two types of precursors, phenyltriethoxysilane (PTES) and triphenylethoxysilane (TPES) for the synthesis of mesoporous organosilica. In this work, all these impacts on adsorbent characteristics, adsorption behavior of toluene, and hydrophobicity were investigated, and the analysis of the ratedetermining step has been explored too.

■

Q dry or wet =

FAtq (1)

W

The tq (min) was calculated by the following equation tq =

i

C y

∫ jjjjj1 − CA zzzzzdt − tD k

0{

(2)

−1

where FA (mmol min ) represents the toluene molar flow, W (g) represents the mass of the sample, CA and C0 (ppm) represent the outlet and inlet toluene concentration, respectively. The correction factor tD serves as the dead time (around 10 s). The effective adsorption capacity of toluene [Qdry(0.1) or wet(0.1) (mmol g−1)] was established by the following equation

MATERIALS AND METHODS

Materials. Pure silica KIT-6 with calcination at 550 °C was prepared by Ryoo’s method,25 named 550 °C-cal silica. Moreover, the other two types of pure silica KIT-6 would be prepared for the grafting experiment in our work. The differences were in the removal methods of template P123. The first pure KIT-6 was obtained by removal of the template by washing with hydrochloric acid solution of ethanol,14 named after-wash silica. The other pure KIT-6 was prepared by low-temperature heat treatment at 200 °C for 6 h without washing, named after-200 °C-cal silica. For the typical grafting process of phenyl-modified silica, 1 g of pure silica and 6 mmol organosilica precursors were added into 50 mL of toluene with 0.2 mL of H2O, and the mixture was stirred for 24 h at 90 °C under a hermetic environment. Then, the product was filtered with toluene three times and deionized H2O three times. Finally, the obtained powder suffered the calcination at 450 °C for 4 h. When PTES (98 wt %) served as the organosilica precursor, the acquired samples with the wash silica and the 200 °C-cal silica as grafting substrates were named after phnl-wash and phnl-cal, respectively. Compared with the synthesis process of phnl-cal, the sample was named after 2phnl-cal when the quantity of added PTES was doubled; furthermore, the sample was named after triphnl-cal and 2triphnl-cal when the PTES was replaced with TPES (98%). Characterization. The textural information was obtained by N2 adsorption/desorption isotherms (77 K) performed on BELSORP MaxII; all samples went through a vacuum heating treatment of 200 °C for 4 h before measurement. The surface area (SA) was calculated by the Brunauer−Emmett−Teller (BET) method, and the pore size

Q dry(0.1) or wet(0.1) =

FAtq(0.1) (3)

W

The tq(0.1) (min) was calculated by the following equation

tq(0.1) =

∫0

t0.1 i

C yz jj jj1 − A zzzdt − t D j C0 z{ k

(4)

where tq(0.1) (min) represents the effective breakthrough time, when the outlet concentration equaled 10% of the inlet concentration. All toluene breakthrough curves would be simulated by the universal and semiempirical Yoon−Nelson model26 (Y−N model) as the following equation t=τ+

CA 1 ln k C0 − CA

(5)

where τ (min) is the time when CA = 50% C0 and k (min−1) represents the rate constant. The details of the calculation of Biot number and HSDM fitting for breakthrough curves are shown in Section S2. Static Adsorption Measurements. Static adsorption isotherms of toluene and water over adsorbents were measured on a BELSORP MaxII. The pretreatment method was similar to that of the dynamic adsorption measurement. The saturated vapor pressures calculated by the Antoine equation27 are 2.276, 2.997, 3.901, and 6.410 kPa for 8928

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Article

Langmuir toluene and 1.704, 2.337, 3.167, and 5.624 kPa for water at 15, 20, 25, and 35 °C, respectively. The Henry constants (KH) (mol g−1 Pa−1) calculated from adsorption isotherms were obtained to estimate the adsorption heat Qad (kJ mol−1), and the following Virial equations28 were requisite

ln(n/p) = A 0 + A1n + A 2 n2 + . . .

(6)

KH = exp(A 0)

(7)

where n (mol g−1) is the adsorption amount at the pressure of p (Pa) and Ai (i = 0, 1, 2...) represents the Virial constants. In eq 6, the higher order terms (A2, etc.) would not be considered in the lower p/p0 than ∼0.1. The virial graph in the p/p0 of 0.01−0.1 could be utilized to avoid higher errors in the lower p/p0 than 0.01.29 The adsorption heat Qad (kJ mol−1) could be worked out by Arrhenius equation30 as follows ln KH = ln A −

ΔHm RT

Figure 1. FT-IR results of pure silica (550 °C-cal silica and 200 °C-cal silica) and typical phenyl-modified silica (phnl-cal, triphnl-cal, 2phnlcal, and 2triphnl-cal).

(8)

Q ad = −ΔHm

(9) −3

−1

−1

where R (8.314 × 10 kJ mol K ) is the ideal gas constant. Distribution Density of Functional Groups. The quantity of silanol on pure silica was the weight loss between 200 and 1100 °C. The quantity of phenyl on phenyl-modified silica was the loss difference from pure silica 550 °C-cal silica to phenyl-modified silica between 500 and 750 °C. The distribution density of silanol or phenyl was equal to the quantity of groups divided by the SA of the adsorbent. The calculation methods were established as follows Density of silanol (nm−2) =

2 × (mass loss 200−1100° C) ×

1 100

the characteristic vibration of the silicon triphenyl instead of the silicon phenyl. This shows the successful surface modification of phenyl on 200 °C-cal silica. In addition, the 2phnl-cal silica exhibited the weakest peaks at 950 and 3420 cm−1 among the phenyl-grafted silicas, owing to its least silanols. The thermogravimetric analyses were used to quantify the amount of silanol or phenyl on the silica adsorbents as shown in Figure 2. The mass drop between 200 and 1100 °C for pure silica was attributed to the condensation of silanol,35 and the weight loss between 500 and 750 °C for phenyl-modified silica was ascribed to the decomposition of phenyl. Using eqs 10 and 11, the density of silanol and phenyl were determined and are listed in Tables S1 and 1 respectively. The coverage density of silanol on 200 °C-cal silica was similar to the wash silica but it was more than that of the 550 °C-cal silica, coinciding with the result of FT-IR. The phnl-cal sample had a similar coverage of phenyl to the phnl-wash sample, and this relates to the coverage density of silanol on their own pure silica substrates. However, the triphnl-cal sample had the lowest coverage of phenyl, perhaps because not enough precursor TPES was added. With doubling the addition of PTES, the coverage of phenyl on the 2phnl-cal sample was nearly double that of the phnl-cal sample. Moreover, the 2triphnl-cal sample had less phenyl coverage as compared to the 2phnl-cal sample in spite of the fact that the phenyl groups added for TPES were equal to the added PTES. These results revealed that silicon phenyl could be modified to pure silica more easily as compared to silicon triphenyl. This is because the precursor of silicon phenyl, which has three Si−OH bonds after the hydrolysis of PTES, could interconnect with each other. However, the precursor of silicon triphenyl possessed only one Si−OH bond after the hydrolysis of TPES. Therefore, with bonding with each other, the former could form stable bonds with the silica substrate more easily. The nano textural properties of adsorbents were investigated using N2 adsorption/desorption isotherms. They displayed type IV isotherms with typical H1-type hysteresis loops as shown in Figure S2 and Table S2. The 550 °C-cal and 200 °Ccal silica both possessed a narrower pore size distribution compared with the wash silica, as can be seen from Figure 3a. Because the pore channel of the wash silica was destroyed by the washing of ethanol solution containing HCl, consequently, pore size uniformity could not be preserved as well as the 550

× NA

M H2O × SA

(10) −5

Density of phenyl (10 =

−2

gm )

[(mass loss500−750° C)phenyl−silica − (mass loss500−750° C)pure silica ] ×

1 100

SA

(11) −1

where NA (6.02 × 10 mol ) is the Avogadro number, MH2O (18 g mol−1) is the molar mass of water, and SA (m2 g−1) is the surface area of the adsorbent. 23

■

RESULTS AND DISCUSSION Characterization of the Adsorbents. The experimental process consisted of two parts, that is, the preparation of pure silica adsorbents and the adsorbents that are modified with phenyl groups. We used FT-IR spectra to confirm the alteration of nanocomponents (surface functional group), as can be seen from Figure 1. The 200 °C-cal pure silica possessed two obvious peaks at 95031,32 and 374033,34 cm−1, which are attributed to Si−OH groups. The 550 °C-cal silica does not have the two peaks. The 200 °C-cal silica has the peak at 3740 cm−1 and the strongest peak at 950 cm−1. This proves that the most abundant silanol groups are found on the 200 °C-cal pure silica after a low-temperature heat treatment at 200 °C, and the surface is well prepared for the postgrafting process. For the second part, there are many Si−OH groups on the silica support surface, and there are also Si−OH bonds on grafted molecules [(C6H5)3−Si−OH or C6H5−Si−(OH)3]. The condensation of the two types of Si−OH groups resulted in the successful grafting. Therefore, the chemical bond between the support and the grafted molecule is the Si−O− Si bond. The peaks at 698 and 740 cm−1 could be assigned to the phenyl groups on the 2phnl-cal and 2triphnl-cal samples as the phnl-wash sample;11 the accessional peak at 715 cm−1 is 8929

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Article

Langmuir

Figure 2. TG/DTG curves of pure silica (a) and phenyl-modified silica (b).

Table 1. Mass Loss and Density of Phenyl on Typical Silica Adsorbents sample

550 °C-cal silica

phnl-wash

phnl-cal

triphnl-cal

2phnl-cal

2triphnl-cal

mass loss at 500−750 °C (wt %) mass loss of phenyl (wt %) density of phenyl (10−5 g m−2)

0.6 0 0

10.5 9.9 15.2

10.6 10.0 13.8

6.7 6.3 8.4

17.0 16.4 24.7

12.2 11.6 15.4

Scheme 1. Abridged Diagrammatic Drawings of Typical Phenyl-Modified Silica Adsorbents

coverage of phenyl and the highest value of SA among the phenyl-modified silica, and it also possesses a very uniform pore size. Notably, the use of expensive organosilanes such as PTES or triphenyltriethoxysilane for the grafting process implies the production of an expensive adsorbent. At first, the grafting solution could be reutilized, so that the PTES or TPES precursors would not be wasted. Then, the organosilica adsorbent can be used not only to remove aromatic VOCs but also to collect and recover aromatic VOCs. Moreover, the adsorbents can be also used over and over again. Static Adsorption Behavior. Static adsorption behavior is affected by several factors, including VOC or water vapor pressure, surface chemistry, and textural property (surface area, pore volume, pore size distribution). For the static adsorption capacity of toluene (Figure 4a), the 550 °C-cal silica had the highest uptake of toluene among the adsorbents at the medium pressure of 100−1500 Pa, whereas the uptakes of toluene at the low pressure of 0−100 Pa followed this order: 2triphnl-cal > triphnl-cal > phnl-cal > 550 °C-cal silica >2phnl-cal ≈ phnlwash. This is because multilayer adsorption appears at medium pressure, where the adsorption capacity was mainly affected by the surface area and pore volume instead of surface phenyl. However, at low pressure, the adsorption behavior was dominated by toluene phenyl interactions, and the adsorbent with more surface area containing more phenyl groups would adsorb more toluene. Hence, in this case, the adsorption capacity was affected by both the surface area and surface phenyl groups. Thus, some phenyl-modified silica prepond-

Figure 3. Pore size distributions of pure silica (a) and phenylmodified silica (b).

°C-cal and 200 °C-cal silica. Compared to 200 °C-cal silica, the narrower pore size of 550 °C-cal silica would be primarily owing to shrinkage of the silica wall by the hotter treatment. Additionally, as shown in Table S2, the 550 °C-cal silica had a lower value of SA than the wash silica or 200 °C-cal silica. The decrease in the BET surface area would be attributed to loss of micropores because condensation of the silica network was promoted by a thermal treatment at higher temperatures, resulting in micropore plugging. After the grafting of phenyl, all phenyl-modified silica showed lower values of SA compared with pure silica, and all 200 °C-cal-based organosilica also displayed better pore size uniformity than the phnl-wash sample (Figure 3a,b). Notably, the SA value of the 2phnl-cal sample was lower than that of the phnl-cal and 2triphnl-cal samples. In accordance with TG analyses of the 2phnl-cal sample, many silicon phenyl groups were modified onto the silica substrate, resulting in greater decrease of surface area. In contrast, a handful of phenyl groups were modified onto the triphnl-cal silica, thusresulting in the highest surface area among the phenyl-grafted silica. Based on the characterizations, the diagrammatic drawings of the two types of phenyl-modified silica are shown in Scheme 1. Benefiting from the isolated grafting of silicon triphenyl on the silica surface, the 2triphnl-cal sample had both a high 8930

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Article

Langmuir

Figure 4. Static adsorption isotherms of toluene (a) and water (b) over typical silica adsorbents at 25 °C.

550 °C-cal silica (41.5 kJ mol−1).11 The results confirmed that both silicon phenyl and silicon triphenyl increase the heat of adsorption for toluene. In addition, the 550 °C-cal silica had the highest Qad values (50.0 kJ mol−1) for water, demonstrating that the hydrophobicity of phenyl-modified silica was increased. Interestingly, the Qad value of water over the 2triphnl-cal adsorbent was higher than that of the 2phnl-cal adsorbent. Maybe it was because of the isolated grafting of triphenyl on the adsorbent surface, which led to the exposure of more Si−O−Si groups. Therefore, the hydrophilic nature of the exposed Si−O−Si groups reduced the hydrophobicity of the 2triphnl-cal adsorbent. Dynamic Adsorption Behavior. Toluene dynamic adsorption behaviors under no humidity condition were investigated, and the breakthrough curves are shown in Figure S5a−c. The adsorption capacity, Qdry (Figure 6b) was calculated using eqs 1 and 2. The effective adsorption capacity Qdry(0.1) (Figure 6b) was calculated using eqs 3 and 4 and is equal to the adsorption capacity when the outlet concentration equals 10% of the inlet concentration. The Qdry(0.1) was more in accordance with practical working conditions and therefore more significant than Qdry. As shown in Figure S5a, the wash silica and the 200 °C-cal silica had the same value of Qdry (1.93 mmol g−1); however, the 200 °C-cal silica had an obviously higher value of Qdry(0.1) than the wash silica (1.49 vs 1.32 mmol g−1). Because the 200 °C-cal silica has a narrower distribution of pore size than the wash silica, the nonuniform pore channel of the wash silica resulted in slower pore diffusion. Therefore, when the toluene molecules broke through the adsorption bed (CA/C0 = 0.1) of the wash silica, there were more pores which had not adsorbed toluene than those in 200 °C-cal silica. Hence, the 200 °C-cal silica exhibited higher Qdry(0.1) than the wash silica. Therefore, the 200 °C-cal silica could be used for a longer time under practical conditions. The constant rate k was acquired by fitting the breakthrough curves with the Y−N model. The k value of the wash silica was lower than that of the 550 °C-cal or 200 °C-cal silica, and the k value of the phnlwash adsorbent was also lower than that of the phnl-cal adsorbent. The results indicate that the uniform pore size distribution has a higher k value due to the increase of diffusion rate. Furthermore, all k values of pure silica were lower compared to those of phenyl-modified silica. This can be due to its bad surface diffusion ability, that is, adding phenyl increases the kinetics by surface diffusion (more phenyl results in stronger affinity with toluene, resulting in better surface diffusion).

erated in the uptake of toluene compared with the 550 °C-cal silica, and the 2triphnl-cal adsorbent showed the highest uptake of toluene. In addition, the toluene isotherm (P/P0 < 0.1) of the 2triphnl-cal adsorbent was fitted by the Dubinin−Radushkevich (D−R) model as shown in Section S1 and Figure S3. The values of maximum amount of adsorbed adsorbates (W0) and microporosity constant (B) are 2.02 mmol g−1 and 8.73 × 10−6 cm6 J−2, respectively. With the two numbers, we could predict its adsorption capacity for any other VOCs (the polarizability is determined by the adsorbate), whose dipole moments are less than 2 debye. The static adsorption isotherms of water are shown in Figure 4b. In the whole pressure range, all phenyl-modified silica showed a lower uptake of water as compared to the 550 °C-cal pure silica, suggesting the increase in hydrophobicity. Besides, the adsorption capacity for water followed the order: triphnlcal > 2triphnl-cal ≈ phnl-cal > 2phnl-cal ≈ phnl-wash. Moreover, more phenyl groups led to the lower uptake of water; the poor pore size uniformity for the phnl-wash adsorbent resulted in the decrease of water adsorption capacity for it. The toluene and water isotherms were measured at four temperatures (Figure S4) to calculate the Henry constant (KH) and subsequently established the adsorption heat (Qad) as shown in Figure 5. Combined with our previous work, the Qad values of toluene were in the order: 2triphnl-cal ≈ 2phnl-cal >

Figure 5. Arrhenius plots for the adsorption of water over 2phnl-cal (black closed symbols) and 2triphnl-cal (olive closed symbols), and toluene over 2phnl-cal (black open symbols) and 2triphnl-cal (olive open symbols). The values close to the lines corresponded to adsorption heat. 8931

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Article

Langmuir

Figure 6. Relationship between the rate constant k and the values of Qdry − Qdry(0.1) (a). Values of Qdry and Qdry(0.1) (b), Qdry(0.1) and SA (c), and Qwet and Qwet(0.1) (d) over typical silica adsorbents.

Figure 7. HSDM fitting lines for toluene dynamic adsorption on 550 °C-cal silica (a) and 2triphnl-cal (b) at 25 °C.

The relationship between the rate constant k and the difference between no humidity capacity and the effective adsorption capacity Qdry − Qdry(0.1) as shown in Figure 6a. It is worth mentioning that the k values almost have a negative linear correlation with the values of Qdry − Qdry(0.1). The results indicate that uniform pore size and the modification of phenyl groups could both decrease the no-effective adsorption capacity, thereby causing improvement of the effective adsorption capacity for toluene. Moreover, the adsorption capacities (Figure 6b) have also been normalized by surface area as shown in Figure 6c. All phenyl-modified silica have higher values than pure silica due to the stronger π−π adsorption interaction for toluene. 2triphnl-cal silica has the highest values of Qdry(0.1) and Qdry(0.1)/SA, indicating that the triphenyl groups exhibit the best improvement for toluene

adsorption capacity. Hence, the 2triphnl-cal adsorbent would serve as a suitable adsorbent for toluene under no humidity practical conditions. The HSDM fitting lines for toluene dynamic adsorption on the 550 °C-cal silica and the 2triphnl-cal adsorbent are shown in Figure 7. The HSDM fitting lines show a good match with the measured points. As can be seen from Table S3, the Biot numbers of 550 °C-cal silica and 2triphnl-cal are both greater than 1, which indicated that the adsorption rate was limited by the intraparticle mass transfer rather than external mass transfer.36−38 In addition, the Biot numbers are in the order: 550 °C-cal (18.7) > phnl-wash (18.5) > 2triphnl-cal (18.2) > 2phnl-cal (17.7). The result showed that the phenyl-grafted adsorbent had a faster intraparticle mass transfer rate than pure silica, coinciding with the result of the Y−N model. 8932

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Langmuir

■

The samples saturated with toluene were studied by TG as shown in Figure S6. The mass loss between 30 and 400 °C is attributed to the volatilization of the adsorbed toluene. Actually, most of the adsorbed toluene was lost before 160 °C. The result indicates that the interaction between all adsorbents and toluene is mainly related to physical interaction, including ordinary van der Waals force for pure silica and π−π interaction for phenyl-grafted silica. Furthermore, dynamic breakthrough curves were measured for a high RH value and are shown in Figure S7, and the values of Qwet and Qwet(0.1) are plotted in Figure 6d. All phenylmodified silica possessed higher values of Qwet and Qwet(0.1) (i.e., better hydrophobicity) than the 550 °C-cal silica because of the hydrophobic phenyl groups. The 2phnl-cal adsorbent showed better hydrophobicity than the 2triphnl-cal adsorbent, and the phnl-cal adsorbent also showed a better hydrophobicity than the triphnl-cal adsorbent; these agree with the results of the larger equilibrium uptake of water, which was caused by greater exposure of hydrophilic Si−O−Si groups on the 2triphnl-cal adsorbent. Notably, the phnl-wash adsorbent had higher values of Qwet and Qwet(0.1) as compared with the 2phnl-cal adsorbent, even though their static adsorption capacity for water and toluene were both similar. Because the pore size uniformity of the 2phnl-cal adsorbent appears to play an important role in dynamic adsorption of binary components (i.e., toluene and water). As discussed previously, a uniform pore size results in an increase in the internal diffusion rate. In other words, the uniform pore size of the 2phnl-cal adsorbent was more favorable for the dynamic adsorption of water as compared with the phnl-wash adsorbent. Because there are huge differences between the concentration of toluene and water (the RH is 90% at 25 °C, i.e., around 28 000 ppm water, the toluene concentration is just 1000 ppm), water was influenced preferentially rather than toluene. Furthermore, the BET surface areas (Swater) calculated from water vapor adsorption isotherms at 25 °C (Section S3) are shown in Figure S8. The lower Swater value represents better hydrophobicity. The order of the Swater values is 550 °C-cal > triphnl-cal > phnl-cal > 2triphnl-cal > 2phnl-cal > phnl-wash. Therefore, the phnl-wash adsorbent showed the best hydrophobicity among all phenyl-modified silica, consistent with the adsorption behaviors in the dynamic binary adsorption tests.

■

Article

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.9b01031. Method of D−R model, calculation of Biot number and the HSDM fitting, BET surface area calculated from the water isotherms, TG/DSC curves of the mesoporous silica with template P123, N2 adsorption/desorption isotherms of pure silica and phenyl modified silica, D−R isotherm plots obtained for the adsorption of toluene on the 2triphnl-cal adsorbent, static adsorption isotherms of toluene and water over 2phnl-cal and 2triphnl-cal adsorbents, toluene dynamic adsorption curves and the simulation curves of Y−N model, TG/DTG curves of pure silica and phenyl-modified silica saturated with toluene, toluene dynamic adsorption curves of typical silica adsorbents under high humidity conditions, BET surface area calculated from water vapor adsorption isotherms, mass loss and the density of silanol on pure silica, and textural properties of the pure and phenylmodified silica adsorbents (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +86 10 62771093. Fax: +86 10 62771093. ORCID

Yue Peng: 0000-0001-5772-3443 John Crittenden: 0000-0002-9048-7208 Junhua Li: 0000-0001-7249-0529 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the national key research and development program of China (2016YFE0126600 and 2016YFC0209203). The authors appreciate the support from the Brook Byers Institute for Sustainable Systems, Hightower Chair and the Georgia Research Alliance at Georgia Institute of Technology.

■

REFERENCES

(1) Huang, R.-J.; Zhang, Y.; Bozzetti, C.; Ho, K.-F.; Cao, J.-J.; Han, Y.; Daellenbach, K. R.; Slowik, J. G.; Platt, S. M.; Canonaco, F.; Zotter, P.; Wolf, R.; Pieber, S. M.; Bruns, E. A.; Crippa, M.; Ciarelli, G.; Piazzalunga, A.; Schwikowski, M.; Abbaszade, G.; Schnelle-Kreis, J.; Zimmermann, R.; An, Z.; Szidat, S.; Baltensperger, U.; Haddad, I. E.; Prévôt, A. S. H. High secondary aerosol contribution to particulate pollution during haze events in China. Nature 2014, 514, 218−222. (2) Niu, H.; Mo, Z. W.; Shao, M.; Lu, S. H.; Xie, S. D. Screening the emission sources of volatile organic compounds (VOCs) in China by multi-effects evaluation. Front. Environ. Sci. Eng. 2016, 10, 1. (3) McDonald, B. C.; de Gouw, J. A.; Gilman, J. B.; Jathar, S. H.; Akherati, A.; Cappa, C. D.; Jimenez, J. L.; Lee-Taylor, J.; Hayes, P. L.; McKeen, S. A.; Cui, Y. Y.; Kim, S.-W.; Gentner, D. R.; IsaacmanVanWertz, G.; Goldstein, A. H.; Harley, R. A.; Frost, G. J.; Roberts, J. M.; Ryerson, T. B.; Trainer, M. Volatile chemical products emerging as largest petrochemical source of urban organic emissions. Science 2018, 359, 760−764. (4) Ma, P. K.; Zhao, Y.; Robinson, A. L.; Worton, D. R.; Goldstein, A. H.; Ortega, A. M.; Jimenez, J. L.; Zotter, P.; Prévôt, A. S. H.; Szidat, S.; Hayes, P. L. Evaluating the impact of new observational constraints

CONCLUSIONS

A new nano silica adsorbent with uniform pore size was synthesized by the low-temperature heat treatment and the surface grafting of phenyl groups. This adsorbent had a faster intraparticle mass transfer rate than pure silica and therefore possessed a higher effective adsorption capacity. Nonetheless, the better pore size uniformity resulted in poor hydrophobicity because this benefited the dynamic adsorption of water to a greater extent. Compared with phenyl-grafted silica, triphenylgrafted silica had a higher adsorption capacity for toluene because of its higher surface area, but poor hydrophobicity due to more exposed hydrophilic Si−O−Si groups. These results showed a significant reference value for the practical application of nano silica adsorbents in the field of environmental governance. 8933

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934

Article

Langmuir on P-S/IVOC emissions, multi-generation oxidation, and chamber wall losses on SOA modeling for Los Angeles, CA. Atmos. Chem. Phys. 2017, 17, 9237−9259. (5) Bolden, A. L.; Kwiatkowski, C. F.; Colborn, T. New Look at BTEX: Are Ambient Levels a Problem? Environ. Sci. Technol. 2015, 49, 5261−5276. (6) Lin, P.; Zhang, Y.; Zhang, X.; Chen, C.; Xie, Y.; Suffet, I. H. The influence of chlorinated aromatics’ structure on their adsorption characteristics on activated carbon to tackle chemical spills in drinking water source. Front. Environ. Sci. Eng. 2015, 9, 138−146. (7) Kim, J. M.; Kim, J. H.; Lee, C. Y.; Jerng, D. W.; Ahn, H. S. Toluene and acetaldehyde removal from air on to graphene-based adsorbents with microsized pores. J. Hazard. Mater. 2018, 344, 458− 465. (8) Cheng, H.; Reinhard, M. Sorption of trichloroethylene in hydrophobic micropores of dealuminated Y zeolites and natural minerals. Environ. Sci. Technol. 2006, 40, 7694−7701. (9) Zhou, Y.; Zhou, L.; Zhang, X.; Chen, Y. Preparation of zeolitic imidazolate framework-8/graphene oxide composites with enhanced VOCs adsorption capacity. Microporous Mesoporous Mater. 2016, 225, 488−493. (10) Liao, P.-Q.; Huang, N.-Y.; Zhang, W.-X.; Zhang, J.-P.; Chen, X.-M. Controlling guest conformation for efficient purification of butadiene. Science 2017, 356, 1193. (11) Liu, S.; Peng, Y.; Chen, J.; Shi, W.; Yan, T.; Li, B.; Zhang, Y.; Li, J. Engineering surface functional groups on mesoporous silica: towards a humidity-resistant hydrophobic adsorbent. J. Mater. Chem. A 2018, 6, 13769. (12) Hu, Q.; Li, J. J.; Hao, Z. P.; Li, L. D.; Qiao, S. Z. Dynamic adsorption of volatile organic compounds on organofunctionalized SBA-15 materials. Chem. Eng. J. 2009, 149, 281−288. (13) Dou, B.; Hu, Q.; Li, J.; Qiao, S.; Hao, Z. Adsorption performance of VOCs in ordered mesoporous silicas with different pore structures and surface chemistry. J. Hazard. Mater. 2011, 186, 1615−1624. (14) Liu, S.; Chen, J.; Peng, Y.; Hu, F.; Li, K.; Song, H.; Li, X.; Zhang, Y.; Li, J. Studies on toluene adsorption performance and hydrophobic property in phenyl functionalized KIT-6. Chem. Eng. J. 2018, 334, 191−197. (15) Hoffmann, F.; Cornelius, M.; Morell, J.; Fröba, M. Silica-based mesoporous organic-inorganic hybrid materials. Angew. Chem., Int. Ed. 2006, 45, 3216−3251. (16) Wang, H.; Wang, T.; Han, L.; Tang, M.; Zhong, J.; Huang, W.; Chen, R. VOC adsorption and desorption behavior of hydrophobic, functionalized SBA-15. J. Mater. Res. 2016, 31, 516−525. (17) Rungsirisakun, R.; Nanok, T.; Probst, M.; Limtrakul, J. Adsorption and diffusion of benzene in the nanoporous catalysts FAU, ZSM-5 and MCM-22: A molecular dynamics study. J. Mol. Graphics Modell. 2006, 24, 373−382. (18) Kosuge, K.; Kubo, S.; Kikukawa, N.; Takemori, M. Effect of pore structure in mesoporous silicas on VOC dynamic adsorption/ desorption performance. Langmuir 2007, 23, 3095−3102. (19) Liu, P.; Long, C.; Li, Q.; Qian, H.; Li, A.; Zhang, Q. Adsorption of trichloroethylene and benzene vapors onto hypercrosslinked polymeric resin. J. Hazard. Mater. 2009, 166, 46−51. (20) Serra, R. M.; Miró, E. E.; Sapag, M. K.; Boix, A. V. Adsorption and diffusion of toluene on Na and Cs mordenites for hydrocarbon traps. Microporous Mesoporous Mater. 2011, 138, 102−109. (21) Qin, Y.; Wang, Y.; Wang, H.; Gao, J.; Qu, Z. Effect of morphology and pore structure of SBA-15 on toluene dynamic adsorption/desorption performance. Procedia Environ. Sci. 2013, 18, 366−371. (22) Wang, H.; Wang, T.; Yu, M.; Huang, X.; Zhong, J.; Huang, W.; Chen, R. Elaborate control over the morphology and pore structure of porous silicas for VOCs removal with high efficiency and stability. Adsorption 2017, 23, 37−50. (23) Thielemann, J. P.; Girgsdies, F.; Schlögl, R.; Hess, C. Pore structure and surface area of silica SBA-15: influence of washing and scale-up. Beilstein J. Nanotechnol. 2011, 2, 110−118.

(24) Yang, C.-M.; Zibrowius, B.; Schmidt, W.; Schüth, F. Stepwise removal of the copolymer template from mesopores and micropores in SBA-15. Chem. Mater. 2004, 16, 2918−2925. (25) Kleitz, F.; Hei Choi, S.; Ryoo, R. Cubic Ia3d large mesoporous silica: synthesis and replication to platinum nanowires, carbon nanorods and carbon nanotubesElectronic supplementary information (ESI) available: TEM images of mesoporous cubic silica and Pt networks, XRD patterns during formation of the cubic phase. See http://www.rsc.org/suppdata/cc/b3/b306504a/. Chem. Commun. 2003, 2136−2137. (26) Ruthven, D. M. Principles of Adsorption and Adsorption processes; John Wiley & Sons: New Jersey, 1984. (27) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties of Gases and Liquids; McGraw Hill: New York, 1986. (28) Cole, J. H.; Everett, D. H.; Marshall, C. T.; Paniego, A. R.; Powl, J. C.; Rodriguez-Reinoso, F. Thermodynamics of High Temperature Adsorption of Some Permanent Gases by Porous Carbons. J. Chem. Soc., Faraday Trans. 1 1974, 70, 2154−2169. (29) Srikanth, C. S.; Chuang, S. S. C. Infrared Study of Strongly and Weakly Adsorbed CO2 on Fresh and Oxidatively Degraded Amine Sorbents. J. Phys. Chem. C 2013, 117, 9196−9205. (30) Fan, W.; Chakraborty, A. Isosteric heat of adsorption at zero coverage for AQSOA-Z01/Z02/Z05 zeolites and water systems. Microporous Mesoporous Mater. 2018, 260, 201−207. (31) Yuan, L.-Y.; Liu, Y.-L.; Shi, W.-Q.; Li, Z.-j.; Lan, J.-H.; Feng, Y.X.; Zhao, Y.-L.; Yuan, Y.-L.; Chai, Z.-F. A novel mesoporous material for uranium extraction, dihydroimidazole functionalized SBA-15. J. Mater. Chem. 2012, 22, 17019−17026. (32) Lin, O. H.; Md Akil, H.; Ishak, Z. A. M. Characterization and Properties of Activated Nanosilica/Polypropylene Composites With Coupling Agents. Polym. Compos. 2009, 30, 1693−1700. (33) Yuan, L.-Y.; Bai, Z.-Q.; Zhao, R.; Liu, Y.-L.; Li, Z.-J.; Chu, S.Q.; Zheng, L.-R.; Zhang, J.; Zhao, Y.-L.; Chai, Z.-F.; Shi, W.-Q. Introduction of Bifunctional Groups into Mesoporous Silica for Enhancing Uptake of Thorium(IV) from Aqueous Solution. ACS Appl. Mater. Interfaces 2014, 6, 4786−4796. (34) Zhang, J.; Han, F.; Wei, X.; Shui, L.; Gong, H.; Zhang, P. Spectral Studies of Hydrogen Bonding and Interaction in the Absorption Processes of Sulfur Dioxide in Poly(ethylene glycol) 400 + Water Binary System. Ind. Eng. Chem. Res. 2010, 49, 2025− 2030. (35) Ek, S.; Root, A.; Peussa, M.; Niinisto, L. Determination of the hydroxyl group content in silica by thermogravimetry and a comparison with 1H MAS NMR results. Thermochim. Acta 2001, 379, 201−212. (36) Crittenden, J. C.; Trussell, R. R.; Hand, D. W.; Howe, K. J. MWH’s Water Treatment-Principles and Design, 3rd ed.; John Wiley & Sons: New Jersey, 2012. (37) Hand, D. W.; Crittenden, J. C.; Thacker, W. E. Simplified Models for Design of Fixed-Bed Adsorption Systems. J. Environ. Eng. 1984, 110, 440−456. (38) Crittenden, J. C.; Cortright, R. D.; Rick, B.; Tang, S.-R.; Perram, D. Using Gac to Remove Vocs from Air Stripper Off-Gas. J.Am. Water Works Assoc. 1988, 80, 73−84.

8934

DOI: 10.1021/acs.langmuir.9b01031 Langmuir 2019, 35, 8927−8934