High temperature synthesis of interfacial functionalized carboxylate mesoporous TiO2 for effective adsorption of cationic dyes

Chemical Engineering Journal 281 (2015) 20–33

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

High temperature synthesis of interfacial functionalized carboxylate mesoporous TiO2 for effective adsorption of cationic dyes Minh-Tri Nguyen-Le, Byeong-Kyu Lee ⇑ Department of Civil and Environmental Engineering, University of Ulsan, Nam-gu, Daehak-ro 93, Ulsan 680-749, Republic of Korea

h i g h l i g h t s  The mesoporous TiO2 was grafted with carboxylate groups at high temperature.  Bidentate carbonate species were also detected.  The surface chemical properties impact the dye adsorption behaviors of TiO2.  The MB adsorption was highly dependent on the electrostatic interaction.  TiO2 can be easily regenerated up to 9 times by high temperature treatment.

a r t i c l e

i n f o

Article history: Received 7 April 2015 Received in revised form 15 June 2015 Accepted 16 June 2015 Available online 23 June 2015 Keywords: Adsorption Carboxylate Electrostatic interaction Intraparticle diffusion

a b s t r a c t This study reported a facile method to enhance the adsorption capacity of TiO2-based mesoporous material for the removal of methylene blue (MB) as a cationic dye from aqueous solution. The mesoporous surface of the modified TiO2 was successfully grafted with interfacial carboxylate groups by using the sol–gel method with ethylenediaminetetraacetic acid (EDTA) as a template at high temperature. Properties of the as-synthesized adsorbents were characterized by FTIR, XPS, FE-SEM, TGA and BET analysis. The MB uptake by adsorption was highly dependent on the electrostatic interaction between the adsorbent and the adsorbate. Basic media was favorable for greater MB deposition on the modified TiO2 surface, which related to uncoordinated carbonyl groups. The maximum capacity was reported as 32.15 mg/L. The removal efficiency was maintained at 83% even after 9 adsorption–desorption cycles, which supported the reusability of the adsorbent for cationic dyes removal from aqueous solutions. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction The rapid growth of many industries such as textile, paper, printing, leather, food and cosmetics has resulted in the extensive use of dyes for coloring substrates [1]. Treatment of discharged wastewater containing dye compounds has become challenging due to their stable and complex chemical structures, and their low biodegradability. Qada et al. reported that less than 1 ppm of dyes can make colored wastewater [2]. Moreover, the presence of dyes in industrial effluents has seriously deteriorated the environment. Once dyes enter aquatic liquefiers, they can prevent the sunlight from penetrating into the water and hence impede the photosynthesis of the aqueous ecosystem [3]. Some dyes are also carcinogenic, mutagenic, teratogenic and harmful to the human respiratory system [4,5]. Therefore, the removal of dyes has

⇑ Corresponding author. Tel.: +82 52 259 2864; fax: +82 52 259 2629. E-mail address: [email protected] (B.-K. Lee). http://dx.doi.org/10.1016/j.cej.2015.06.075 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.

attracted much attention. Methylene blue (MB) is a cationic dye with a heterocyclic aromatic structure. Like other dyes, its adverse impacts on the aqueous media are inevitable. Hence, studies on the decolorization of MB in aqueous solutions are of significant practical importance. The numerous methods used to treat dye – loaded effluents can be divided into three categories: chemical (e.g., oxidation, photochemical), physical (e.g., adsorption, membrane filtration, ion exchange) and biological (e.g., aerobic and anaerobic biodegradation) [6–9]. Among them, adsorption has been found to be superior to others in terms of its simplicity, low cost, acceptable removal efficiency and ease of operation. Various adsorbents have been used for dye removal from wastewater. From a purely practical point of view, adsorbents themselves should be effective and nontoxic. TiO2, an inexpensive and nontoxic material, has been propagated for use as a photocatalyst in environmental and energy-related fields, particularly in air and water purification-related fields. However, even when TiO2 is used as a

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

photocatalyst, the adsorption of pollutants on TiO2 plays an important role as a prerequisite for the effective photodegradation [10]. The adsorption capacity of TiO2 is limited due to their intrinsic disadvantages such as narrow pore diameter, low surface area, thermal instability and less active sites [11]. Recently, several studies have focused on methods to enhance the adsorption ability of TiO2 for effective removal of pollutants by functionalizing the TiO2 surface with carboxylic acids [12,13]. However, most of synthesis have been performed at low temperature. Synthesis, in which interfacial functional groups can be survival at high temperature, is essential for an extension of TiO2 applications as an adsorbent and a photocatalyst under visible light. Mesoporous TiO2 has been investigated with the expectation of improved performance in terms of photocatalytic activity [14–16]. However, extensive studies on the adsorption performance of functionalized mesoporous are favored in a viewpoint of application to water treatment. Therefore, the current study aimed to synthesize interfacial functionalized mesoporous TiO2 (hereafter, modified TiO2) with carboxylate groups at high temperature by using EDTA as a template. Effects of functional groups on the enhanced MB adsorption of the modified TiO2 were discussed in detail. The feasibility of thermal treatment for regenerating the adsorbent was also investigated. 2. Methods and materials

21

of solution containing 0.06, 0.1, 0.2, 0.3, 0.4, 1.2, 1.6, and 2.0 g of MB at 30 ± 1 °C at a shaking speed of 150 rpm for 210 min. The solution pH was adjusted in the range of 1–12 by adding a small amount of 0.1 M HCl or 0.1 M NaOH while other parameters were kept constant. The adsorption isotherm studies were also carried out by mixing 0.02 g of modified TiO2 with 20 mL of MB solution in which the amount of MB was varied (0.06–2.0 g), for 210 min at pH = 10 and 30 ± 1 °C. The same experimental conditions were applied when investigating the adsorption capacity as a function of time. The remaining MB – concentration was determined using a UV–vis spectrophotometer. Before the measurement, aliquots of MB were taken out at specific intervals of time, and centrifuged at 3000 rpm for 5 min to remove solids. The amount of MB removed per gram of adsorbent (qt) and the removal ratio (H%) were obtained according to Eqs. (1) and (2), respectively [17]:

qt ¼

V s  ðC 0  C t Þ 1000  m

H% ¼

100  ðC 0  C t Þ C0

ð1Þ

ð2Þ

where C0 and Ct (mg/L) are the concentrations of MB before and after the adsorption, respectively, Vs (mL) is the volume of MB solution, and m (g) stands for the mass of the adsorbent.

2.1. Material synthesis and analysis methods First, 3.60 g of EDTA (99%, Sigma–Aldrich) was dissolved in 100 mL of deionized water. A 50.0 mL mixture (1:2) of titanium (IV) isopropoxide precursor (97%, Sigma–Aldrich) and isopropanol was then added dropwise to the prepared EDTA solution to make the sol form under vigorous stirring at 4 °C for 3 h. The as-prepared solution was aged in the dark for 24 h to induce gelation. Afterwards, the solvent was evaporated at 70 °C for 1 h to obtain the modified TiO2 sample. The sample was further calcinated at 400 °C for 5 h and ground into powders. For comparison, the pristine TiO2 was synthesized using the same process but in the absence of EDTA. All reagent-grade chemicals were used without pretreatment. To study the nature of the chemical bonding, Fourier Transform Infrared (FTIR) transmittance spectra in the region 4000–400 cm1 were recorded on a Perkin–Elmer Nicolet Nexus 470 FTIR spectrometer at room temperature. The thermogravimetric (TG) analyses were performed on Setaram Labsys Evo in N2 flow and at a heating rate of 10 °C/min. The chemical states of elements in the samples was analyzed based on X-ray photoelectron spectroscopy (XPS) data recorded on a Thermo Scientific Sigma Probe spectrometer with a monochromatic AlKa source (photon energy 1486.6 eV), spot size 400 lm, pass energy of 200 eV and energy step size of 1.0 eV. The specific Brunauer–Emmett–Teller (BET) surface area and pore size measurements of the samples were conducted on a Micromeritics ASAP 2020 apparatus under nitrogen air. Field Emission – Scanning Electron microscopy (FE-SEM) was used to check the morphology of the as-synthesized materials. The solution pH was monitored through a pH meter. A Genesis 10S UV–Vis spectrophotometer was used to obtain the absorbance of the MB solution at 665 nm. 2.2. MB adsorption study Batch mode was applied throughout triplicate MB adsorption experiments under dark condition. To optimize the adsorption conditions, 0.01–0.10 g of the adsorbent was shaken with 20 mL

Fig. 1. (a) FTIR spectra of modified and pristine TiO2. (b) FTIR spectra of modified TiO2 before and after adsorption.

22

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

O

Ti O

O

O

Ti O O O

Ti

O Ti O O O

Ti OH

3. Results and discussion 3.1. Material characteristics

O

Fig. 2. Proposed structures of functional groups on the modified TiO2 surface.

2.3. Regeneration of the material The regeneration was performed through calcination at high temperature. After adsorption, the MB-loaded TiO2 was separated from the solution by centrifugation at 3000 rpm for 5 min. The supernatant solution was then removed and the adsorbent was transferred into a furnace for calcination at 400 °C for 3 h with a heating rate of 6 °C/min. The pollutant-free material was introduced to the next cycle and acted as an adsorbent. The process was repeated up to 9 cycles. A blank sample was also subjected to the same procedure for comparison.

The FTIR spectra of both modified and pristine TiO2 were analyzed to quantitatively determine the characteristics of existing functional groups (Fig. 1a). For the untreated TiO2, the bands at 3414 and 1629 cm1 are ascribed to the stretching and deforming vibrations of adsorbed water molecules, respectively. A broad band centered at 473 cm1 belongs to Ti–O–Ti stretching peaks. For the modified TiO2, other surface species are also characterized: vasym(CH2) (2917 cm1), vsym(CH2) (2842 cm1) and v(C–C) (1061 cm1), v(CO3) (1347 cm1) and v(COO) (1449 cm1 and 1563 cm1) [18,19]. These peaks suggest the co-existence of succinic acid and carbonate species on the modified TiO2 surface. Du et al. reported the value of v(bidentate CO3) is 1347 cm1, which also observed in this study [20]. The coordination of carboxylate groups of succinic acid on the modified TiO2 surface can be clarify by examining the value of Dv(COO). The value of Dv(COO) is

Fig. 3. XPS spectra of the modified and pristine TiO2 for (a) Ti 2p level and (b) O 1s level.

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

23

Fig. 4. XPS spectra of the modified TiO2 for (a) C 1s level and (b) N 1s level.

114 cm1, which indicates a bridging structure. In combination with the results of pHpzc analyses (discussed in Section 3.2.1), both carboxylate groups are involved in bridging coordination, which makes a highly basic TiO2 surface due to uncoordinated carbonyl groups. The proposed characteristics of carboxylate groups and carbonate species are displayed in Fig. 2. Fig. 1b shows the FTIR spectra of the modified TiO2 before and after adsorption. After adsorption of MB, the adsorption peaks at 1593 cm1, 1386 cm1 and 1322 cm1 indicate the vibration of the aromatic ring, C–N bond, and CH3 group for MB, respectively [21]. The intrinsic peaks of functional groups of modified TiO2 seem not to be observed due to the interaction between MB and the surface of modified TiO2. Therefore, the high adsorption capacity can be attributed to such a interaction. Fig. S1 shows the TG results of the pristine and modified TiO2. For the pristine TiO2, the first thermal degradation step below 200 °C shows 2.0 wt.%, which is attributed to the loss of adsorbed water molecules. The second thermal degradation step between 300 °C and 400 °C is about 3.8 wt.%, possibly ascribed to the decomposition of the organics from the hydrolysis of titanium (IV) isopropoxide precursor. For the modified TiO2, 3.6 wt.% loss of adsorbed water molecules is observed below 200 °C. A dramatic weight loss (30.3 wt.%) is observed over a wide range of temperature (300–600 °C) due to both the thermal decomposition of the organics from the hydrolysis of titanium (IV) isopropoxide

precursor and carboxylate groups on the modified TiO2 surface, which suggests some of functional groups still remain on the modified TiO2 surface at high temperature calcination. The XPS spectra of the samples were examined to reveal the chemical states of elements contained in the pristine and modified samples. In the high resolution XPS spectra of Ti (Fig. 3a) shown in the pristine sample, two peaks at 459.4 eV and 464.8 eV are the typical Ti2p3/2 and Ti2p1/2 peaks of pure TiO2, respectively [22]. The spectra of the modified TiO2 exhibit a large shift of about 1.5 eV towards lower binding energies of two typical peaks for Ti, indicating the more electronegative states (Ti3+ or Ti2+) of Ti. Lee et al. [23] reported binding energies for Ti3+ and Ti2+ of 457.8 eV and 456.2 eV, respectively. Accordingly, the ionic state of Ti atoms in this study should be Ti3+ (457.9 eV). The lower oxidation state of Ti is probably due to the formation of oxygen vacancies. As far as O1s spectra is concerned, Fig. 3b presents a wider and more asymmetric O1s peak of the modified TiO2 than that of the pristine TiO2, indicating the presence of more than one valence state of O atoms according to the binding energy. The peaks at 529.8 eV, 531.2 eV, 532.6 eV and 533.6 eV are assigned to Ti–O–Ti (lattice O), C@O (oxygen bound species), Ti–N–O (interstitial N) and NOx species, respectively [23,24]. To verify the state of C and N, the core levels of C1s and N1s are reported in Fig. 4a and b, respectively. In the C1s spectra, the peaks at 284.6 eV and 286.4 eV are attributed to Csp2 and C–O bond

24

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Fig. 5. FE-SEM images of the pristine TiO2 (a–b) and modified TiO2 (c–d) at magnifications of 300 (left) and 1500 (right), respectively.

(oxygen bound species) in the modified sample, respectively [25]. The peak observed at 288.6 eV is assigned to C@O bonding, which mostly corresponds to carbonate species [25]. Moreover, the carboxylate groups were also detected in the FTIR analysis. Therefore, these groups are also related to the C@O bonding. It can be inferred that carbonate species and carboxylate group may be bound to the TiO2 surface via C–O–Ti bonds. From the N1s high resolution spectra, it is obvious that several of N atoms are incorporated into TiO2 lattice via both interstitial N (Ti–N–O) and substitutional NO (N–Ti–O), whose XPS peaks were observed at 399.8 eV and 398.2 eV, respectively [23]. Moreover, the low-intensity peak at 404.8 eV represents surface-adsorbed nitrogen species (NOx) [23]. The morphological structure of the modified and pristine TiO2 are confirmed by FE-SEM, as shown in Fig. 5a–d. The size distribution of the rock-like crystals of modified TiO2 is clearly heterogenous: around 10–50 lm in length and 10–30 lm in width. Moreover, the 1500-times-magnification image shows the surface of large aggregates of particles is made up of much smaller crystallites. The same characteristics of morphology are also observed from FE-SEM images of the pristine TiO2. However, the surface of large aggregates of particles of the pristine TiO2 is likely to be made up of much more smaller crystallites than that of the modified TiO2. Fig. 6a and b shows the N2 adsorption–desorption isotherms and pore size distribution of both modified and pristine TiO2, respectively. The isotherms of both samples are type IV adsorption isotherms with hysteresis loops in the relative range of 0.4–1.0 (Fig. 6a), which is indicative of mesoporous structures [26]. The isotherm of pristine TiO2 shows a H4-type hysteresis loop, which may arise from the presence of large mesopores connected to surrounding pores of much smaller size, whereas the isotherm of the modified TiO2 displays a H3-type hysteresis loop, which indicates that mesopores have cylindrical pore geometries. The isotherm of the pristine TiO2 is also shifted downward compared to that of

the modified TiO2, indicating the higher BET surface area of the modified TiO2. This finding is consistent with the calculated BET areas and total pore volumes of pristine (29.809 m2/g and 0.016 cm3/g, respectively) and modified TiO2 (91.927 m2/g and 0.276 cm3/g, respectively). The average pore size of modified TiO2 is 20.48 nm, which is larger than that of pristine TiO2 (3.60 nm) (Fig. 6b). This may be due to the large grain size of the modified TiO2 as a result of the less thermally stable TiO2 precursor in the presence of EDTA. The high surface area and high mesopore volume of the modified TiO2 could benefit the contact between adsorption sites and adsorbate molecules, thus contribute to enhancing the MB adsorption. 3.2. Adsorption condition optimization The pH of the media, initial amounts of adsorbate and adsorbent dose are the main discernible factors contributing to the effectiveness of the aqueous adsorption of MB at a given temperature [27]. 3.2.1. pH Fig. 7a shows the fluctuation of MB uptake various pH values. For the pristine TiO2, no MB uptake was observed at pH < 4. Only small amount of MB was adsorbed in the pH range of 4–6 (0.40 mg/g). When pH changes from 6 to 7, the adsorption rapidly increases (0.40–0.87 mg/g), and remains constant afterwards. For the modified TiO2, as the pH is increase from 1 to 12, the MB uptake initially slightly increases (2.51–2.67 mg/g) in the pH range of 1–7, then rapidly rises (2.67–3.24 mg/g) over pH 7–10, and then remains constant over pH 10–12. Those trends in MB uptake behavior can be explained by closer inspection of the point of zero charge (pHpzc) at which the electrical net charge is zero. The pHpzc of modified TiO2 determined by the solid addition method is about 9.5 (Fig. 7b) which is higher than that of pristine TiO2 (pHpzc = 6.8) and commercial Degussa P25 (pHpzc = 6.3) [28,29]. A solution pH

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

25

Fig. 6. Nitrogen adsorption–desorption isotherms (a) and pore size distribution (b) of both modified and pristine TiO2.

lower than 7 ( 7 due to the deprotonation of superficial OH groups, which can strongly attract cationic MB molecules. The surface of modified TiO2 might, however, be partially negatively charged in the pH range of 7–9.5 ( 9.5, the surface becomes mostly negatively charged, which would attract more cationic MB molecules and thereby slightly increases the MB adsorption (3.02–3.24 mg/g). The coordination of carbonate species would become monodentate mode in which free oxygen atoms are negatively charged (Fig. 8c). Du et al. reported the FTIR value of v(monodentate CO3) is 1385 cm1, which also observed in the FTIR spectra of the modified TiO2 after adsorption (Fig. 1b)

[20]. Obviously, pH significantly affects the adsorption of MB. Basic media seems to benefit the deposition of MB on the modified TiO2 surface. Therefore, pH 10 was chosen for further adsorption study. 3.2.2. Initial amounts of adsorbate The effect of initial MB amount on the adsorption is illustrated in Fig. 9a. An increase in MB concentration significantly increases the adsorption capacity at low MB concentrations (3–20 mg/L). The plateau at higher concentrations (80–100 mg/L) reveals the saturation of active sites occupied by adsorbate molecules. A high initial MB concentration also provides a driving force that benefits the mass transfer between the solid–liquid interface. The adsorption capacity of MB was maximized at 31.67 mg/g at the given adsorption conditions. The adsorption ratio, however, revealed an opposite trend. The removal efficiency of MB per gram of the modified TiO2 for a given dose of adsorbent (0.02 g) first reaches 90% at an MB concentration of 3 mg/L but then drastically decreases at higher MB concentrations. MB molecules have more chance to

26

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Fig. 7. (a) Effect of pH on MB removal (MB initial concentration = 3 mg/L; mass of adsorbent = 0.02 g; temperature = 30 °C; contact time = 60 min). (b) Point of zero charge determination of modified TiO2 and pristine TiO2.

come into contact with active sites at low adsorbate concentrations than at higher concentrations, when the competition for adsorption between adsorbate molecules is remarkable. Fig. 9b show the images of MB solution before and after the adsorption. 3.2.3. Adsorbent dose optimization Adsorbent dose is an important parameter for determining the capacity of the adsorbent under given experimental conditions [27]. Fig. 10 demonstrates the effect of adsorbent dose on MB adsorption. Increasing the amount of modified TiO2 utilized increases the number of adsorption sites on the adsorbent surface. The percentage MB removal for a given amount of MB solution usually increases with increasing adsorbent dose. However, the adsorption ratio slightly declines at adsorbent dose above 0.02 g, which may be ascribed to the heterogeneous distribution of active sites between the particle surfaces. Conversely, the MB loading capacity per gram of adsorbent dose decreases continuously from 6.26 mg/g to 3.24 mg/g, which is attributed to a decrease in the accessibility of MB molecules to adsorption sites. Excessive amount of adsorbent for a given MB concentration may decrease the availability of adsorption sites by blocking active sites. For economic removal efficiency, an adsorbent dose of 0.02 g was selected for further study. 3.3. Adsorption isotherm studies of equilibrium Langmuir, Freundlich and Temkin models were exploited to elucidate the adsorption characteristics. The Langmuir isotherm

is the simplest homogeneous monolayer adsorption model, while the other two isotherms are heterogeneous multilayer coverage models. Moreover, unlike the first two models, the Temkin adsorption isotherm model explicitly accounts for adsorbent–adsorbate interaction. The linear forms of these models are expressed as Eqs. (3)–(5), respectively [33]:

  Ce 1 1 ¼ Ce þ KL qe qm

ð3Þ

log qe ¼ log K F þ n1 log C e

ð4Þ

RT ðln aT þ ln C e Þ bT

ð5Þ

qe ¼

where Ce (mg/L) and qe (mg/g) are the residual MB concentration and adsorbent capacity at equilibrium, respectively, qm (mg/g) is the maximum coverage of MB monolayer, KL (L/mg) is a Langmuir constant related to adsorption energy, KF (mg/g) (L/mg)1/n is a Freundlich constant related to adsorption capacity, n is the adsorption intensity, and bT (J/mol) and aT (L/g) are Temkin constants related to the heat of adsorption and equilibrium binding constant, respectively. Three adsorption models were plotted, as seen in Fig. 11(a)–(c). All relevant parameters are also summarized in Table 1. Among the three models, the Langmuir model showed the best fit (R2 = 0.9923) with a saturated coverage of 32.15 mg/g, which is close to the experimental data (31.67 mg/g). This result implies a monolayer adsorption of MB. The moderate fit (R2 = 0.9904) is

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

27

Fig. 8. Schematic illustration of the proposed mechanism for the adsorption of MB at different pH values.

obtained with the Temkin model whose heat of adsorption is around 464 J/mol. This reaction heat suggests a main adsorption mechanism via electrostatic interaction [34]. Moreover, the value of n (2.36) obtained from the Freundlich model lies in a range of 2–10, which indicates a favorable MB adsorption. Table 2 compares the maximum capacity among the different adsorbents, and verifies the potential application of the modified TiO2 as an effective adsorbent for MB removal. To assess the feasibility of MB adsorption, a dimensionless separation factor, RL, was calculated from Eq. (6) [40]:

RL ¼ 1=ð1 þ C 0 K L Þ

ð6Þ

Different values of RL indicate various characteristics of the adsorption: RL = 0 (irreversible), 0 < RL < 1 (favorable), RL = 1 (linear), RL > 1 (unfavorable). The obtained RL values are in range of 0.611–0.052 over the MB concentration range of 3–100 mg/L. Hence, the on-going adsorption is favorable, particularly, at higher MB concentration where the mass transfer is fast. 3.4. Adsorption kinetic and mechanism The MB uptake was analyzed in respect to time to determine an appropriate contact time at which the adsorption reaches

equilibrium. As seen in Fig. 12, the adsorption easily reaches equilibrium after only 30 min for relatively low adsorbate concentrations (3–20 mg/L). However, as the concentration is increased, equilibrium is obtained only slowly. At relatively high concentrations (60–100 mg/L), it took as long as 150 min for most of the vacant active sites to be occupied by MB dye molecules. There is no significant difference in adsorption capacity between 80 and 100 mg/L of MB after 120 min because of the strong repulsion between the dye molecules adsorbed onto the adsorbent surfaces for a given amount of the adsorbent. To determine the rate that controlled the adsorption, the pseudo-first and -second order models were investigated. These two models are described by Eqs. (7) and (8), respectively [41,42]:

lnðqe  qt Þ ¼ ka t þ lnqe

ð7Þ

t 1 t ¼ þ qt q2e kb qe

ð8Þ

where qe and qt (mg/g) are the mass of MB adsorbed per mass of adsorbent at equilibrium and at time t (min), respectively, and ka (min1) and kb (g mg1 min1) are rate constants of the pseudo-first and -second order models, respectively.

28

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

(a)

(b)

Fig. 9. (a) Effect of MB concentration on the adsorption (mass of adsorbent = 0.02 g; temperature = 30 °C; pH = 10.0; contact time = 210 min). (b) Image of MB solution before and after the adsorption; the inset: image of MB-loaded modified TiO2 (after adsorption, MB solution was centrifuged to remove the supernatant, followed by drying at 80 °C in 1 h).

Fig. 10. Effect of adsorbent amount on the removal of MB (MB initial concentration = 3 mg/L; temperature = 30 °C; pH = 10.0; contact time = 60 min).

29

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Fig. 11. Plots of Langmuir (a), Freundlich (b) and Temkin (c) isotherms for the adsorption.

Table 2 Comparison of saturated adsorption capacities for MB on different adsorbents at the condition of 30 °C and a pH range of 8–10.

Table 1 Parameters of three isotherm models. Models

Parameters

Langmuir

qm = 32.15 mg/g KL = 0.18 L/mg R2 = 0.9923

Freundlich

KF = 5.72 mg/g n = 2.36 R2 = 0.9423

Temkin

aT = 2.85 bT = 464 J/mol R2 = 0.9904

Adsorbents

Adsorption capacity (mg/g)

References

Carbon nanotubes Zeolite Titania Rice husk Magnetic graphene-carbon nanotube composite Magnetic graphene Modified TiO2

35.2 10.86 5.98 9.83 32.47

[35] [36] [37] [38] [39]

24.88 32.15

[39] This study

30

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Fig. 12. Dependence of adsorption on the contact time (mass of adsorbent = 0.02 g; temperature = 30 °C; pH = 10.0; contact time = 210 min).

Fig. 13. Pseudo-first order (a), pseudo-second order (b) and Weber’s intraparticle (c) models (mass of adsorbent = 0.02 g; temperature = 30 °C; pH = 10.0; contact time = 210 min).

31

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Table 3 The values of qe,cal, qe,exp, k and R2 for the pseudo-1st and -2nd order kinetic models at each given MB concentration (mass of adsorbent = 0.02 g; temperature = 30 °C; pH = 10.0; contact time = 210 min). Pseudo-1st order

Pseudo-2nd order

Concentration (mg/L)

qe,cal (mg/g)

qe,exp (mg/g)

ka (min1)

R2

qe,cal (mg/g)

qe,exp (mg/g)

kb  102 (min1)

R2

3 5 10 15 20 60 80 100

0.32 0.39 7.02 4.94 13.11 18.55 25.39 24.44

3.18 4.54 8.42 10.24 14.04 23.51 29.88 30.51

0.007 0.024 0.036 0.028 0.025 0.019 0.024 0.025

0.7413 0.9351 0.7499 0.9816 0.9607 0.6190 0.8959 0.8195

3.18 4.56 8.73 11.44 16.02 24.29 32.55 32.49

3.18 4.54 8.42 10.24 14.04 24.52 29.88 30.51

21.82 17.92 1.35 1.17 0.26 0.25 0.16 0.19

0.9996 0.9999 0.9985 0.9998 0.9969 0.9515 0.9887 0.9893

Table 4 The values of k and R2 for the Weber’s kinetic model at each given MB concentration (mass of adsorbent = 0.02 g; temperature = 30 °C; pH = 10.0; contact time = 210 min). Concentration (mg/L)

kint,1 (kint,2) (mg/g min1/2)

R2, kint,1 (R2, kint,2)

3 5 10 15 20 60 80 100

0.051 0.071 0.568 0.763 1.225 1.386 1.846 1.589

0.9775 0.9921 0.9821 0.9840 0.9784 0.9495 0.9782 0.9317

(0.037) (0.020) (0.200) (0.101) (0.604) (1.813) (1.685) (0.636)

(0.9495) (0.9511) (0.9745) (0.9747) (0.9896) (0.9077) (0.9230) (0.9171)

The italics in the 2nd and third column represent the values of k (called kint,2) and R2 (called R2, kint,2) of the second intraparticle diffusion step, respectively.

Fig. 13a and b show the distribution of plotted points related to these two models at various MB concentrations. All values of qe,cal (calculated from the models), qe,exp (experimental data), ka, kb and determination coefficients (R2) of the two rate-controlled adsorption models are listed in Table 3. In the analysis of R2 values, the linearity of points plotted from the pseudo-second model showed a better fit (R2 = 0.9515–0.9999) than that from the pseudo-first model. Moreover, the deviation between the calculated qe,cal and experimental qe,exp values of the pseudo-first model was large, whereas that of the pseudo-second model was not significant. We therefore concluded that MB is deposited on the modified TiO2 in accordance with the pseudo-second order kinetic model.

Furthermore, low MB concentrations favor the adsorption in term of kinetics due to the high kb value. Through Weber’s intraparticle diffusion, we studied the steps of the adsorption. The amount of MB uptake at any time (qt) is believed to be proportional to t0.5 rather than to the contact time t, as shown in Eq. (9) [43]:

qt ¼ kint t0:5 þ b

ð9Þ 1

0.5

where kint (mg g min ) is the intraparticle diffusion rate constant and b is the intercept of the intraparticle plot. Fig. 10c presents two broken straight lines with different slopes at each MB concentration in the plot. This multilinearity clearly reveals a two-step adsorption mechanism [44]. At first, MB molecules may diffuse from the bulk of the solution to the adsorbent–adsorbate interface by the driving force derived from the initial dye concentrations, and accumulate on the external surface of the adsorbent by the driving force resulting from electrostatic interaction. Later, the intraparticle diffusion through the pores leads to equilibrium. Values of intraparticle rate constants are listed in Table 4. Generally, increasing adsorbate concentration increases kint due to the increased driving force, which is consistent with the earlier discussion [45]. Furthermore, kint,1 is clearly greater than kint,2, revealing the first step to be faster than the second step. However, the slow-rate step is not the sole rate-determining step because the plot does not pass through the origin in the coordinate [45]. Other unknown processes may also influence the adsorption rate.

Fig. 14. Plot of free Gibbs energy DG0 versus reaction temperature T(K).

32

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Fig. 15. Removal efficiency at each regeneration cycle of the adsorbent.

Furthermore, the determination coefficients (R2) for the intraparticle diffusion model are in range of 0.90 and 0.99, which demonstrates the plausible feasibility of Weber’s model.

3.5. Thermodynamic analysis Thermodynamic parameters related to intrinsic energy change of the adsorption, such as standard Gibbs free energy of adsorption DG0 (kJ/mol), standard enthalpy DH0 (kJ/mol), and standard entropy DS0 (J/mol K), were obtained according to Eqs. (10) and (11) [46]:

DG0 ¼ RT ln K L ;

ð10Þ

DG0 ¼ DH0  T DS0

ð11Þ

where KL (L/mg) is the Langmuir constant, R the gas constant (8.314  103 kJ/mol K), and T the Kelvin temperature (K). The values of DH0 and DS0 were determined from the slope and intercept of the plots of DG0 versus T (Fig. 14), respectively. The DG0 value is negative over the entire temperature range (298–323 K), which demonstrated the underlying spontaneity and thermodynamic favorability of the adsorption. Also, the DG0 value is less negative at high temperature, suggesting that the adsorption is more spontaneous at low temperatures. Ozcan et al. claimed that if DG0 lies in range of 20 to 0 kJ/mol, the adsorption is physisorption [47]. Otherwise, if the range is between 80 and 400 kJ/mol, the adsorption is chemisorption. In this study, the DG0 values range from 27.80 to 25.98 kJ/mol, indicating that physisorption is dominant. Also, the DH0 value is negative (50.74 kJ/mol) and smaller than 40 kJ/mol, suggesting an exothermic physisorption [46]. Further, the negative DS0 (76.75 J/mol K) reveals the decrease in the degree of freedom at the adsorbent–adsorbate interface and the stability of the internal structure through MB deposition onto the adsorbent [48]. These thermodynamic analysis results demonstrate the affinity of the cationic dye towards the functional groups of modified TiO2 [49].

3.6. Sustainable adsorption ability of the modified TiO2 Restoration of adsorption ability is an important issue for economic utilization. Most previous studies have used the wet treatment method for the desorption of dyes with the aid of washing reagents such as ethanol, acids, bases or their mixtures [50–52]. However, this requires an abundance of washing reagents, and the dyes still need to be disposed of appropriately after desorption. Therefore, in this study, we used thermal treatment for removing the organic dyes deposited on modified TiO2. Wang et al. reported an MB adsorption efficiency of more than 88% when the synthetic zeolite was heated at 540 °C for 1 h. However, chemical regeneration recovered only 60% of the original adsorption efficiency [53]. In the present study, a removal efficiency of 83% was achieved even after 9 adsorption–desorption cycles of adsorbent regeneration (Fig. 15). This demonstrates the reusability feasibility of the modified TiO2 for MB removal. 4. Conclusions The interfacial functionalized mesoporous TiO2 with carboxylate groups synthesized at high temperature by using EDTA demonstrated an enhanced adsorption capacity of TiO2 for removal of MB from aqueous phase through electrostatic interaction. The properties of functional groups on modified TiO2 surface mainly depend on pH solution. The removal efficiency was maximized at 90% after 60 min when 0.02 g of the adsorbent was applied to 20 mL of 3 mg/L MB solution at pH 10 and 30 °C. The material was feasibly regenerated with an acceptable loss of only 17% of the original removal efficiency even after 9 adsorption–desorption cycles. These results suggest that the adsorption capacity of modified TiO2 is as good as other selective adsorbents in the removal of cationic dyes. Acknowledgment This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science, ICT and Future Planning (2013R1A2A2A03013138).

M.-T. Nguyen-Le, B.-K. Lee / Chemical Engineering Journal 281 (2015) 20–33

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2015.06.075. References [1] J.Z. Guo, B. Li, L. Liu, Kangel Lv, Removal of methylene blue from aqueous solutions by chemically modified bamboo, Chemosphere 111 (2014) 225–231. [2] E.N. El Qada, S.J. Allen, G.M. Walker, Adsorption of basic dyes from aqueous solution onto activated carbons, Chem. Eng. J. 135 (2008) 174–184. [3] K.G. Bhattacharyya, A. Sharma, Kinetics and thermodynamics of methylene blue adsorption on Neem (Azadirachta indica) leaf powder, Dyes Pigm. 65 (2005) 51–59. [4] R.O.A. de Lima, A.P. Bazo, D.M.F. Salvadori, C.M. Rech, D.P. Oliveira, G.A. Umbuzeiro, Mutagenic and carcinogenic potential of a textile azo dye processing plant effluent that impacts a drinking water source, Mutat. Res., Genet. Toxicol. Environ. Mutagen. 626 (2007) 53–60. [5] P.A. Carneiro, G.A. Umbuzeiro, D.P. Oliveira, M.V.B. Zanoni, Assessment of water contamination caused by a mutagenic textile effluent/dyehouse effluent bearing disperse dyes, J. Hazard. Mater. 174 (2010) 694–699. [6] B.V.D. Bruggen, C. Vandecasteele, Removal of pollutants from surface water and groundwater by nanofiltration: overview of possible applications in the drinking water industry, Environ. Pollut. 122 (2003) 435–445. [7] T.H. Kim, C. Park, J. Yang, S. Kim, Comparison of disperse and reactive dye removals by chemical coagulation and Fenton oxidation, J. Hazard. Mater. 112 (2004) 95–103. [8] N. Daneshvar, M. Ayazloo, A.R. Khataee, M. Pourhassan, Biological decolorization of dye solution containing malachite green by microalgae Cosmarium sp., Bioresour. Technol. 98 (2007) 1176–1182. [9] J. Labanda, J. Sabate, J. Llorens, Experimental and modeling study of the adsorption of single and binary dye solutions with an ion-exchange membrane adsorber, Chem. Eng. J. 166 (2011) 536–543. [10] L. Xiong, W.L. Sun, Y. Yang, C. Chen, J.R. Ni, Heterogeneous photocatalysis of methylene blue over titanate nanotubes: effect of adsorption, J. Colloid Interface Sci. 356 (2011) 211–216. [11] T. Nguyen-Phan, M.B. Song, E.W. Shin, Removal efficiency of gaseous benzene using lanthanide-doped mesoporous titania, J. Hazard. Mater. 167 (2009) 75– 81. [12] Y. Liu, C. Liu, Z. Zhang, Effects of carboxylic acids on the microstructure and performance of titania nanocrystals, Chem. Eng. J. 138 (2008) 596–601. [13] J. Feng, Jinwei Zhu, Wei Lv, Jingjing Li, Wei Yan, Effect of hydroxyl group of carboxylic acids on the adsorption of Acid Red G and Methylene Blue on TiO2, Chem. Eng. J. 269 (2015) 316–322. [14] F. He, J. Li, T. Li, G. Li, Solvothermal synthesis of mesoporous TiO2: the effect of morphology, size and calcination progress on photocatalytic activity in the degradation of gaseous benzene, Chem. Eng. J. 237 (2014) 312–321. [15] S. Zhan, Y. Yang, X. Gao, H. Yu, S. Yang, D. Zhu, Y. Li, Rapid degradation of toxic toluene using novel mesoporous SiO2 doped TiO2 nanofibers, Catal. Today 225 (2014) 10–17. [16] B. Li, Z. Zhao, F. Gao, X. Wang, J. Qiu, Mesoporous microspheres composed of carbon-coated TiO2 nanocrystals with exposed (0 0 1) facets for improved visible light photocatalytic activity, Appl. Catal. B 147 (2014) 958–964. [17] I. El Saliby, L. Erdei, J.H. Kim, H.K. Shon, Adsorption and photocatalytic degradation of Methylene Blue over hydrogen-titanate nanofibres produced by a peroxide method, Water Res. 47 (2013) 4115–4125. [18] D.T. Sawyer, M. McKinnie, Properties and infrared spectra of ethylenediaminetetraacetic acid complexes. III. Chelates of higher valent ions, Inorg. Chem. 82 (1960) 4191–4196. [19] A.W. Jackson, O. Shebanova, A.L. Hector, P.F. McMillan, Amorphous and nanocrystalline titanium nitride and carbonitride materials obtained by solution phase ammonolysis of Ti(NMe2)4, Solid State Chem. 179 (2006) 1383–1393. [20] H. Du, T.W. Christopher, D.E. Armin, A.R. James, In situ FTIR spectroscopic analysis of carbonate transformations during adsorption and desorption of CO2 in K-promoted HTlc, Chem. Mater. 22 (2010) 3519–3526. [21] L.H. Ai, C.Z. Zhang, F. Liao, Y. Wang, M. Li, L.Y. Meng, J. Jiang, Removal of methylene blue from aqueous solution with magnetite loaded multi-wall carbon nanotube: kinetic, isotherm and mechanism analysis, J. Hazard. Mater. 198 (2011) 282–290. [22] S.P. Chenakin, G. Melaet, R. Szukiewicz, N. Kruse, XPS study of the surface chemical state of a Pd/(SiO2; TiO2) catalyst after methane oxidation and SO2 treatment, J. Catal. 312 (2014) 1–11. [23] S.H. Lee, E. Yamasue, K.N. Ishihara, H. Okumura, Photocatalysis and surface doping states of N-doped TiOx films prepared by reactive sputtering with dry air, Appl. Catal. B 93 (2010) 217–226. [24] R. Jaiswal, N. Patel, D.C. Kothari, A. Miotello, Improved visible light photocatalytic activity of TiO2 co-doped with Vanadium and Nitrogen, Appl. Catal. B 126 (2012) 47–54. [25] W. Ren, Z. Ai, F. Jia, L. Zhang, X. Fan, Z. Zou, Low temperature preparation and visible light photocatalytic activity of mesoporous carbon-doped crystalline TiO2, Appl. Catal. B 69 (2007) 138–144.

33

[26] R.S. Nasar, M. Cerqueira, E. Longo, J.A. Varela, Analysis of Ca-PZT powder obtained by the Pechini and partial oxalate methods, Cerâmica 54 (2008) 38– 42. [27] M.T. Yagub, T.K. Sen, S. Afroze, H.M. Ang, Dye and its removal from aqueous solution by adsorption: a review, Adv. Colloid Interface Sci. 209 (2014) 172– 184. [28] L.S. Balistrieri, J.W. Murray, The surface chemistry of goethite (alpha FeOOH) in major ion seawater, Am. J. Sci. 281 (1981) 788–806. [29] J. Ryu, W. Choi, Substrate-specific photocatalytic activities of TiO2 and multiactivity test of water treatment application, Environ. Sci. Technol. 42 (2008) 294–300. [30] D.A. Dougherty, The cation–p interaction, Acc. Chem. Res. 46 (2013) 885–893. [31] Z. Chen, J. Zhang, J. Fu, M. Wang, X. Wang, R. Han, Q. Xu, Adsorption of methylene blue onto poly(cyclotriphosphazene-co-4,4’-sulfonyldiphenol) nanotubes: kinetics, isotherm and thermodynamics analysis, J. Hazard. Mater. 273 (2014) 263–271. [32] K.S. Tong, M.J. Kassim, A. Azraa, Adsorption of copper ion from its aqueous solution by a novel biosorbent Uncaria gambir: equilibrium, kinetics, and thermodynamic studies, Chem. Eng. J. 170 (2011) 145–153. [33] Y. Gao, Y. Li, L. Zhang, H. Huang, J.J. Hu, S.M. Shah, X.G. Su, Adsorption and removal of tetracycline antibiotics from aqueous solution by graphene oxide, J. Colloid Interface Sci. 368 (2012) 540–546. [34] S.C. Smith, F. Ahmed, K.M. Gutierrez, D.F. Rodrigues, A comparative study of lysozyme adsorption with graphene, graphene oxide, and single-walled carbon nanotubes: potential environmental applications, Chem. Eng. J. 240 (2014) 147–154. [35] C.H. Wu, Adsorption of reactive dye onto carbon nanotubes: equilibrium, kinetics and thermodynamics, J. Hazard. Mater. 144 (2007) 93–100. [36] C. Woolard, J. Strong, C. Erasmus, Evaluation of the use of modified coal ash as a potential sorbent for organic waste streams, Appl. Geochem. 17 (2002) 1159–1164. [37] M.L. Fetterolf, H.V. Patel, J.M. Jennings, Adsorption of methylene blue and acid blue 40 on titania from aqueous solution, J. Chem. Eng. Data 48 (2003) 831– 835. [38] P. Sharma, R. Kaur, C. Baskar, W.J. Chung, Removal of methylene blue from aqueous waste using rice husk and rice husk ash, Desalination 259 (2010) 249–257. [39] P. Wang, M. Cao, C. Wang, Y. Ao, J. Hou, J. Qian, Kinetics and thermodynamics of adsorption of methylene blue by a magnetic graphene–carbon nanotube composite, Appl. Surf. Sci. 290 (2014) 116–124. [40] T.W. Weber, R.K. Chakravorti, Pore and solid diffusion models for fixed-bed adsorbers, AIChE J. 20 (1974) 228–238. [41] E. Demirbas, N. Dizge, M.T. Sulak, M. Kobya, Adsorption kinetics and equilibrium of copper from aqueous solutions using hazelnut shell activated carbon, Chem. Eng. J. 148 (2009) 481. [42] Z.A. AL-Othman, R. Ali, M. Naushad, Hexavalent chromium removal from aqueous medium by activated carbon prepared from peanut shell: adsorption kinetics, equilibrium and thermodynamic studies, Chem. Eng. J. 184 (2012) 242. [43] G.F. Malash, M.I. El-Khaiary, Piecewise linear regression: a statistical method for the analysis of experimental adsorption data by the intraparticle-diffusion models, Chem. Eng. J. 163 (2010) 256–263. [44] A.E. Ofomaja, Intraparticle diffusion process for lead (II) biosorption onto mansonia wood sawdust, Bioresour. Technol. 101 (2010) 5868–5876. [45] M. Alkan, Ö. Demirbas, M. Dog˘an, Adsorption kinetics and thermodynamics of an anionic dye onto sepiolite, Microporous Mesoporous Mater. 101 (2007) 388–396. [46] C.Y. Kuo, C.H. Wu, J.Y. Wu, Adsorption of direct dyes from aqueous solutions by carbon nanotubes: determination of equilibrium, kinetics and thermodynamics parameters, J. Colloid Interface Sci. 327 (2008) 308–315. [47] A. Ozcan, E.M. Oncu, S. Ozcan, Kinetics, isotherm and thermodynamic studies of adsorption of acid blue 193 from aqueous solutions onto natural sepiolite, Colloids Surf. A: Physicochem. Eng. Aspects 277 (2006) 90–97. [48] T. Aman, A.A. Kazi, M.U. Sabri, Q. Bano, Potato peels as solid waste for the removal of heavy metal copper(II) from waste water/industrial effluent, Colloids Surf. B 63 (2008) 116–121. [49] M.T. Sikder, Y. Mihara, M.S. Islam, T. Saito, S. Tanaka, M. Kurasaki, Preparation and characterization of chitosan–carboxymethyl-b-cyclodextrin entrapped nanozero-valent iron composite for Cu(II) and Cr(IV) removal from wastewater, Chem. Eng. J. 236 (2014) 378–387. [50] L. Zhou, J. Jin, Z. Liu, X. Liang, C. Shang, Adsorption of acid dyes from aqueous solutions by the ethylenediamine-modified magnetic chitosan nanoparticles, J. Hazard. Mater. 185 (2011) 1045–1052. [51] Y.R. Zhang, S.Q. Wang, S.L. Shen, B.X. Zhao, A novel water treatment magnetic nanomaterial for removal of anionic and cationic dyes under severe condition, Chem. Eng. J. 233 (2013) 258–264. [52] J. Liu, G. Liu, W. Liu, Preparation of water-soluble b-cyclodextrin/poly(acrylic acid)/graphene oxide nanocomposites as new adsorbents to remove cationic dyes from aqueous solutions, Chem. Eng. J. 257 (2014) 299–308. [53] S. Wang, H. Li, S. Xie, S. Liu, L. Xu, Physical and chemical regeneration of zeolite adsorbents for dye removal in wastewater treatment, Chemosphere 65 (2006) 82–87.