Enhancing Acidic Dye Adsorption by Updated Version of UiO-66

ABST-CT: In this study, two improved versions of UiO-66 were successfully synthesised. Modified UiO-66 and UiO-66-Ce were characterised to confirm the integrity of the structure, the stability of functional groups on the surface and the thermal stability. Activated samples were used for removal harmful anionic dye (methyl orange) (MO) from wastewater. Batch adsorption process was relied to investigate the competition between those MOFs for removing MO from aqueous solution. Based on the results, at a higher initial concentration, the maximum MO uptake was achieved by UiO-66-Ce which was be9er than modified-UiO-66. :ey adsorbed 71.5 and 62.5 mg g-1 respectively. Langmuir and Freundlich isotherms were employed to simulate the experimental data. In addition, Pseudo first order and Pseudo second order equations were used to describe the dynamic behaviour of MO through the adsorption process. :e high adsorption capacities on these adsorbents can make them promised adsorbents in industrial areas.


INTRODUCTION
Nowadays, significant amount of wastewater with colour has been created by various industries including textile, leather, paper, printing, dye stuff, plastic and so on [1]. Based on the high impact of the quality of water by colour, the removal of dyestuff from polluted water is becoming very important issue [1]. Besides, the high visibility of dye even with a small amount is undesirable. Additionally, many dyes are toxic and even carcinogenic [1][2][3].
According to the stability to light and oxidation reactions, the degradability of synthetic dyes become one of the most challenges to the environmentalists [3]. It is reported that physical, chemical and biological methods have been applied for the remediation of dyes in aqueous solution [1,[3][4]. One of the most feasible techniques among the proposed methods, is adsorption process due to its high efficiency, economic, and simplicity of design as well as operation [3][4].
Methyl orange (MO) is classified as a common acidic (anionic) dyes, which has been broadly used in many industries including textile, printing, paper, food and pharmaceutical industries, wood, silk, and research laboratories [2,4]. e toxicity of MO has drawn the a ention of researcher to the importance of the removal of this synthetic dye [2][3][4]. According to the above-mentioned explanation, MO was chosen to be examined in present study as a typical acidic dye. For the purpose of removal this kind of cationic dye, many solid materials were used as adsorbents to decolour the contaminated wastewater like activated carbon [5][6], graphene oxide [7], Layered double Hydroxide (LDH) [8] and zeolites [9] .
Metal-organic frameworks (MOFs) (hybrid materials) are class of extended structure crystalline solid porous materials which have been used widely in many applications [10]. e characteristics of MOFs including structural and functional tunability have led to explosive growth in their applications via extensive research during the last two decades [11]. erefore, they have been used in several applications like gas and vapour storage, molecular separations, chemical catalysis, chemical sensing, ion exchange, drug delivery and hazardous materials [12][13]. ey are good alternative adsorbents for removal of MO from aqueous solutions [14].
Because of lack of studies in MO removal using MOFs, the present study used the modified UiO-66 and UiO-66-Ce as structural sable porous materials to eliminate MO from wastewater. ese samples were well characterised. en, they are tested for adsorption of anionic dye (MO) at room temperature. e results proved high adsorption affinity of MO toward these adsorbents. e experimental data were fi ed by kinetic and equilibrium adsorption models to investigate the type of the adsorption process and the energetic characteristics of the adsorbent's surface respectively. at can suggest an opportunity of using them as alternative sorbents instead of the conventional adsorbents for removal ionic dyes from wastewater.

Synthesis and Activation
All chemicals were supplied by Sigma Aldrich-Australia without further purifications. UiO-66-Ce was solvothermally synthesised under autogenous pressure. An exact amount of terephthalic acid (0.3771 g, 2.27 mmol; Sigma-Aldrich, 98%) was mixed with 31.387 mL of DMF (Sigma-Aldrich, 99%). en the mixture was stirred for 10 min till a clear solution was formed. A er that ZrCl 4 (0.529 g, 2.27 mmol; Sigma-Aldrich, 99%) was added into the solution and stirred for 5 min. en CeN 3 O 9 .6H 2 O (0.1971 g, 0.000454 mmol; Sigma-Aldrich, 99%) was added to the reactants and stirred for 15 min. e mixture of reactants was loaded into a Teflon-lined autoclave, sealed and placed in a preheated oven at 393 K for 24 hr. White powder of UiO -66-Ce was collected by centrifugation and washed with a DMF at least three times. en the product was dried and immersed in absolute methanol (100%, Sigma Aldrich) for 5 d. eventually, it was filtered, dried and heated under vacuum at 473 K overnight prior to use for adsorption.
Modified UiO66 was synthesised solvothermally according to a reported procedure by a single-solvent method [15]. A solution of 32 mL of DMF was divided equally into two batches. In the first batch, 2.27 mmol of terephthalic acid was added and mixed for 15 min and then NH 4 OH (0.4 mL, 2 M) was added dropwise to this mixture. In the second batch, (0.53 g, 2.27 mmol) of ZrCl 4 was mixed with the solvent for approximately 30 min. A er that, both batch solutions were mixed in a single pot for approximately 20 min. Finally, the resulting solution was placed inside a 45 mL Parr PTFE-lined stainless-steel vessel (Parr Instrument Company, USA) and placed in an oven at 393 K for 24 hr. A er cooling to room temperature, the centrifugation was used to separate white gel-like material, which was dried in an oven at 353 K for 24 hr. For activation of modified-UiO-66, a solvent exchange method by chloroform was used by mixing 0.1 g of a synthesised material in 50 mL of chloroform for 30 min, and then keeping soaked for 5 d.
en, the product was filtered by vacuum filtration and dried in the oven at 373 K for 2 hr. e final product was heated under vacuum at 473 K for 2 d.

Characterization
ermogravimetric analysis of the adsorbents were checked by a thermogravimetric analysis (TGA) instrument (TGA/DSC1 STARe system-ME LER TOLEDO). e samples were loaded into an alumina pan and heated to 1173 K at a rate of 5 K min -1 . e air flow rate was maintained at 10 mL min -1 .
FTIR spectra (Spectrum 100-FT-IR Spectrometer, Perkin-Elmer) were obtained to check the stability of the functional groups on the organic ligands. e spectrums were scanned from 650 to 4000 cm -1 with a resolution of 4 cm -1 by using an a enuated total reflectance (ATR) technique. X-ray powder diffraction pa erns were obtained with an X-ray diffractometer (D8 Advance-Bruker aXS) using Cu Kα radiation (λ = 1.5406 Å). Quantachrome instrument (Autosorb-1) was used to determine the N 2 adsorption/ desorption isotherm, BET surface area, pore size and pore volume of activated samples. e samples were firstly evacu-ated at 473 K for 24 hr prior to secondly use in the analysis at 77 K. Elemental analysis was done by e NexION® 300/350 -ICP-MS (PerkinElmer) and Optima 8300-ICP-OES (PerkinElmer).

Adsorption Process
An aqueous stock solution of MO (1000 ppm) was prepared by dissolving MO (C 14 H 14 N 3 NaO 3 S, MW: 327.34, Sigma-Aldrich) in deionized water. Aqueous solutions with different concentration of MO were prepared by successive dilution of the stock solution with water. Spectrophotometer (UV spectrophotometer-Jasco V-670) was used to determine the actual absorbance at the wavelength of 464 nm for each standard concentration. en the calibration curve was obtained for the concentrations ranged from 5-100 ppm.
Prior to use in adsorption study, the adsorbents (modified UiO-66 and UiO-66-Ce) were dried overnight under vacuum at 373 K. e several glass containers were cleaned, dried and filled up to 20 mL with MO of different concentrations ranged from 10 to 75 ppm. A er that, the exact amount of an adsorbent (20 mg) was mixed with MO in the containers by magnetic stirrer and maintained at a time from 5 min to 24 hr at 298 K. e samples for analysis were collected by syringe filter at different sampling interval. A supernatant was used in UV spectrometer for analysis. e uptake of an adsorbent at any time and equilibrium time is expressed as shown in Equation (1) and (2), accordingly. However, removal efficiency of MO is calculated based on Equation (3) [16].
Where: q t : Amount of MB adsorbed in a specific weight of the adsorbent at any time t (mg. g -1 ), q e : Amount of MB adsorbed in a specific weight of the adsorbent at equilibrium (mg. g -1 ), C 0 : Initial concentrations of MB solution at time zero (mg L -1 ), C t : Concentration of MB solution at time t (mg L -1 ), C e : Concentration of MB solution at equilibrium (mg L -1 ), V : volume of MB solution in the adsorption batch process (L), R% : Removal efficiency MB, m : e specific weight of the modified UiO-66 or UiO-66 -Ce in (g) used in the adsorption batch process.
Experimental adsorption capacities were fi ed by two kinetic models including pseudo-first order and pseudosecond models to determine the adsorption rate and identify the type of MO adsorption [17][18][19][20]. Moreover, the experimental data were also fi ed by the Langmuir and Freundlich adsorption isotherm to describe the adsorption equilibrium isotherm of MO on the modified UiO-66 and UiO-66-Ce samples [17][18][19]21].

RESULT AND DISCUSSION
Activation process by chloroform exchange process was very active to remove all unwanted chemicals from the pores. Also, It mostly removed the whole content of the second metal which was expected to be weakly coordinated on the structure of UiO-66 due to their high incompatibility with Zr +4 . erefore, trace of Ce +2 0.1% has been investigated by the elemental analysis.  samples, it was demonstrated that the structures were the same, similar to that in previous reported articles [15,22]. In addition, FTIR spectra of the samples in this study was shown in Figure1 (b), all functional groups which were seen on the surface of UiO-66 in previous studies [23] were observed and maintained on the surface of UiO-66-Ce and modified UiO-66. Formation of the crystalline UiO-66 was confirmed by presenting the peaks at the band of 1550-1630 cm − 1 which referred to coordinated carboxylates with metal centers. On the other hand, the stretching variation band of CO bond in free carboxylic acid throughout the region of 1640-1670 cm − 1 were mostly disappeared on all samples that can be a good indication for a good activation procedure used for the removing non-reacted carboxylic acid and free solvent from the pores. Furthermore, the thermal stability of those samples was investigated as demonstrated in Fig-ure1 (c). All samples have the same thermal stability as that was mentioned in reported articles [15,24]. ey are stable up to about 750K. Figure 2 shows the N 2 adsorption/ desorption isotherms on the modified UiO-66 samples confirming their porosity. It seems these samples have pores in the long size range extending from micropore to macropore. Primary fillings in narrow micropores were achieved at low relative pressure up to 0.01 as shown in Figure 2. While the secondary fillings in wider micropore were filled on the long range of relative pressure up to 0.9 therefore this behaviour was represented by relatively horizontal line. However, hysteresis of desorption isotherm over adsorption isotherm referred to the presence of neck-bo le mesopores. In addition, rapid increasing of adsorption rate at higher pressure close to 0.999 can be a good evidence for presence of macropores.   Table 1 displays the textural properties of the updated UiO-66 samples. Specific surface area (BET) was reduced in the structure of UiO-66-Ce, it was 911.85 m 2 . g -1 . Whereas, the specific surface area of modified UiO-66 was approximately the same as that was published for UiO-66. It was 1226 m 2 . g -1 . e pore volume and pore size were enhanced in all the samples, the largest pore volume and pore size were exposed by modified UiO-66 of 1.65 cc.g -1 and 5.4 nm correspondingly. While they were 0.91 cc. g -1 and 4.00 nm in UiO-66-Ce. Consequently, the external surface area was 331 and 224.45 m 2 . g -1 on the modified UiO-66 and UiO-66-Ce respectively.

Kinetics Adsorption Models
Kinetics adsorption studies is crucial part due to the giving a good indication about one of the criteria for efficiency of adsorbent which is the rate of adsorption as well as give clear picture about the mechanism of adsorption. e linear relationship of pseudo first order model can be expressed in eq. 4 [25]: While e linear form of pseudo second order model is given in eq. 5 [19]: Where: q e : amount adsorbed at equilibrium (mg. g -1 ). q t : amount adsorbed at time t (mg. g -1 ). t : contact time (min). k 1 : First order kinetic constant (min -1 ) k 2 : Second order kinetic constant (g. mg -1 min -1 ) ln (q e − q t ) versus t plot can lead to a straight line from which the value of k 1 and q e can be found from the slope and intercept value, accordingly [26][27][28]. In the same way, t/q t versus t can be plo ed to give linear relationship, from which q e and k 2 can be computed [29][30][31].
To compare the adsorption kinetics precisely, the experimental adsorbed amount changed with time are fi ed with pseudo-first-order and pseudo-second-order kinetic models [17][18][19][32][33][34] as shown in Figure 3 ( a, b, c, d). Tables 2 and 3 are tabulated the obtained parameter of the rate constant studies for various initial concentrations of MO by fi ing the experimental data to the pseudo-firstorder and second-order equations. e results of the correlation coefficient (R 2 ) of the pseudo-first-order and pseudo-second-order adsorption models were revealed that R 2 values of pseudo-second-order equation has higher values (greater than 0.99) compared to R 2 values of pseudo-firstorder equation, and its calculated equilibrium adsorption capacity (q e,cal ) is consistent with experimental one (q e ). Consequently, the pseudo-second-order adsorption mechanism is predominant for the adsorption of dye. erefore, the overall rate of MO uptake on updated UiO-66 samples seems to be controlled by the chemisorption processes [35] including valence forces through sharing or exchange of electrons [20,36].   Table 2 Pseudo-First-order kinetics constants k 1 (min -1 ) and correlation coefficients (R 2 ) for adsorption of MO on updated UiO-66 samples at mechanism had been detected in the removal of MB and MO by adsorption onto an activated carbon [37]. It is also notable that in general the rate constant of the pseudo-second-order model (k 2 ) is decreases with increases of the initial MO concentration for the samples in this study. It appears from Table 3 that the rate constant (k 2 ) increased on the adsorbents when initial concentration was set on 30 mg.L -1 while it decreased at lower and higher concentration.
at can be interpreted according to abundance of negative charges of MO which increases with increasing an initial concentration. Hence, the mobility of MO molecules may be affected by the initial concentration as it is first can be increased with increasing the concentration up to the overcharge limit then was decreased with raising the initial concentration further up. occurred in the first 30 min of the contacting time indicating that in the beginning of the adsorption process, the diffusion rate of MO toward active sites on the adsorbents was increased and consequently the adsorption uptake of MOD was rapidly enhanced. Furthermore, with increasing the contact time, the uploading of MO was slower due to lack availability of adsorptive sites as well as the likelihood of repulsive force between the molecules of MO themselves in the bulk solution. e adsorption mechanism of MO on the updated UiO-66 samples is controlled by the external surface area and pore size because MO has a large hydrophobic molecule [38]. Herein, all updated UiO-66 had larger average pore size which can be facile access for the molecules of MO. Consequently, the interaction between MO and the organic linkers is more likely to occur by π-π stacking interactions. In addition, present of trace of Ce in the metal center of Zr enhanced its positive charge density as a result of that, the affinity of MO adsorption toward the metal center of the UiO-66-Ce sample was increased.

Equilibrium adsorption models
Equilibrium status of MO on the adsorbents at constant temperature is required because it is very important to determine the isotherm data and then to design the appropriate adsorption system. Many isotherm adsorption models were established to describe sorption equilibrium isotherms. e most important isotherm adsorption models that were widely employed to fit the equilibrium data were Langmuir [39] and Freundlich [40].

Freundlich isotherm
It is commonly used to describe a heterogeneous adsorption systems by using the following empirical formula; Equation (6) is linearised and simplified in the following form of Equation (7); Where: q e : amount of dye sorbed at equilibrium per unit mass of sorbent (mg. g -1 ). C e : the equilibrium concentration of dye in solution (mg.L -1 ) k F : Freundlich equilibrium constants (mg. g -1 . (L. mg -1 ) 1/n ) is an indicator of the adsorption capacity. : an empirical parameter refers to the adsorption intensity while n measures the deviation from linearity of the adsorption. e favourability of the adsorption systems may be tested by determining the reciprocal of the magnitude of the exponent n value. When the value of n ranged from 1 to 10, it can be concluded that the normal adsorption process is favourable. is behaviour is most common physical adsorption because of a decrease in adsorbate-adsorbent interactions with increasing the surface density [41].
From the straight-line plot of ln (q e ) vs ln (C e ), the values of n and k F can be simply found from the slop and intercept, accordingly. e values of these parameters were significantly affected by the type of adsorbate and adsorbent. erefore, they are obviously changed with using modified UiO-66, and UiO-66-Ce, as displayed in Table 4.

Langmuir isotherm
Langmuir model [41] is built on many assumptions; the adsorbent has a uniform surface and its adsorptive sites are equivalent; and the molecules of adsorbates do not interact with themselves therefore a monolayer is made. [42]. Based on this theory, Langmuir empirically derived an equation which correlates between the numbers of active adsorption sites on the surface of adsorbent and the concentration of adsorbate.
is correlation is called Langmuir Equation which is expressed as follow: Whereas the linear form of Langmuir model may be wri en as below: Where: qm : Langmuir maximum loading capacity (mg. g -1 ), k L : Langmuir constant related to the energy of adsorption and affinity of binding sites (L. mg -1 ), C e : the equilibrium concentration of dye in solution (mg. L -1 ) qe : amount of dye sorbed at equilibrium per unit mass of sorbent (mg. g -1 ). For obtaining the q m and k L from the experimental data, C e /q e can be plo ed against C e yielding linear relationship with slop (1/q m ) and intercept (1/k L q m ). e values of these parameters as well as the parameters of the Freundlich model were found for modified UiO-66 and UiO-66-Ce and listed in Table 4.
According to the Langmuir model Table 4 demonstrates that, the maximum sorption capacity of MO (qm) on the modified UiO-66 and UiO-66-Ce was 67.56 and 208 mg. g -1 respectively. e maximum adsorption capacity was a hugely deviated away from the experimental data in UiO-66-Ce while that was in a good agreement in the modified UiO-66, In addition, K L values in modified UiO-66 was much higher than that in UiO-66-Ce, it was 1.56 and 0.256 respectively. at indicates to strong interaction is likely to happen between MO and the active site on the surface of modified UiO-66 whereas this interaction is likely to be very weak in UiO-66-Ce. Furthermore, according to the Freundlich model, Table 4 displays that K F and n were 40.8 and 1.18; and 31.13 and 3.75 in UiO-66-Ce and modified UiO-66 respectively.
at indicates that the adsorption rate of MO toward the surface of UiO-66-Ce was higher than that toward modified UiO-66. Also, UiO-66-Ce has a lower deviation from the linearity.
In Figure 4 displays that, adsorption of MO on UiO-66-Ce and modified UiO-66 was well fi ed by Langmuir model. Our results are distinguished among other results from different researches as listed in Table 5 showing a good performance of updated UiO-66 samples. Figure 5 demonstrates that the removal efficiencies of MO on the modified UiO-66 and UiO-66-Ce was the highest (99%) at the lowest initial concentration used then it is slightly decreased on UiO-66-Ce with increasing the initial concentration. However, the removal efficiency was sharply dropped on modified UiO-66. e expression of essential characteristics of the Langmuir isotherm is very important indication for the favourability of adsorption process which may be wri en based on a dimensionless constant separation factor (R L ) as follow: Where: C 0 : the initial concentration of adsorbate (mg. L -1 ), k L : (L. mg -1 ) is Langmuir constant.
Type of the isotherm can be predicted from the R L value indicating the favourability of the process of adsorption. It may be equal to zero, one, greater than one, or greater than zero and less than one proofing that the adsorption process is either irreversible, linear, unfavourable, or favourable, respectively. Figure 6 shows that all values of R L are between 0 < R L < 1 confirming that the normal adsorption of MO on UiO-66 samples are favourable. RL approached to zero at high concentration indicated that a further increasing in initial concentrations might lead to irreversible adsorption due to enhancing the a raction force with increasing the initial concentrations [22].    ey were synthesised in single pot solvothermal methods. e textural properties were improved and enhanced. e batch adsorption experiment was achieved at constant temperature to remove MO from the wastewater using the activated updated UiO-66 samples. e experimental data were fi ed for kinetic adsorption and adsorption equilibrium isotherms. In regards to the kinetic adsorption, it was confirmed that the pseudo-second-order kinetics was dominated, indicating that the MO uptake on modified UiO -66 samples is controlled by the chemisorption processes. However, the adsorption isotherm of MO on all updated UiO-66 samples was in accordance with the Langmuir isotherm model verifying the adsorption of MO on adsorbents is monolayer adsorption and the adsorbent s are structurally homogeneous. Moreover, modified UiO-66 was the best adsorbent for MO removal at highest maximum adsorption capacity of 68.49 mg. g -1 . ese updated UiO-66 samples can be selected as candidates for industrial application for removal other contaminants from wastewater.