Study of the adsorption of an organic pollutant onto a microporous metal organic framework

As a result of the global industrial development, water pollution has become a serious threat and critical for the environment (Michael et al. 2013). Two percent of the dyes produced are usually discharged into aquatic systems as effluents without any control (Khan et al. 2013). The dyes' removal from contaminated water is important because even a small amount is undesirable and dramatically affects the ecosystem by weakening the light intensity, thus blocking the photosynthetic process. Moreover, many dyes employed nowadays are considered to be highly toxic and even carcinogenic (Michael et al. 2013). Methyl orange (MO) is a well-known acidic/anionic dye belonging to the azo groups, widely used in the textile, printing, paper, food and pharmaceutical industries (Mittal et al. 2007). The presence of azo group (N = N) on MO and its low biodegradability make it an issue for the environmental science (Mittal et al. 2007). Therefore, its elimination from water is an important necessity owing to its high toxicity (Mittal et al. 2007; Chen et al. 2010). For the dyes removal, many physical, chemical and biological methods have been used (Chen et al. 2010) and thus, adsorption is considered competitive because it's efficient, economically feasible, and requires a simple design (Chen et al. 2010). This is possible by exploring novel and efficient materials introduced as adsorbents, which has important practical significance and wide application prospects for environmental protection (Singh & Arora 2011).

Metal organic frameworks (MOFs) are the result of the reaction between organic and inorganic species; they can be described as an organization of secondary units and exhibit a crystalline porous structure (Lee Kim & Ahn 2013). Compared to the traditional porous materials, MOFs possess numerous interesting and appealing features, including a large internal surface area and pore volume, perfectly ordered and well defined pore structures, functionalized pore surfaces, and adjustable pore dimensions (Bao et al. 2016). In addition, the structural diversity present in MOFs as new materials has greatly expanded their applications (Khan et al. 2013; Fang et al. 2018). They are the most studied molecular materials, since they are porous coordination polymers with a great potential for different applications (Khan et al. 2013; Ramanayaka et al. 2019). Among the applications that have received considerable attention, is the adsorption behavior characteristics of metal–organic frameworks from both fundamental aspects and practical water treatment applications (Khan et al. 2013; Fang et al. 2018).

The stability of MOFs upon exposure to humidity or aqueous environments is one of the most important barriers to applying several types of MOFs in practical water treatment applications (Tan et al. 2015). Among many widely investigated water-sensitive MOFs, particularly built up from carboxylate acid ligands (Tan et al. 2015), MOF-5, is a well-known MOF made up by Zn₄O clusters linked by the conjugated base of terephthalic acid (Eddaoudi et al. 2002). In fact, the stability of MOF-5 under such conditions is a topic of renewed interest. Rodriguez et al. showed that the exposure of MOF-5 to water produces a relatively stable new phase, MOF-5 W, which slowly transforms to more stable phase MOF-5H, a mixture of MOF-5 W and zinc terephthalate dihydrate (ZTDH) (Rodríguez et al. 2015).

The main objective of this study was to evaluate the efficiency of MOF-5, which is considered water sensitive in practical applications of water treatment. As applications, using MOF-5 to remove MO by adsorption in an aqueous environment, the effect of operating parameters like the contact time, the pH, the temperature, the initial dye concentration and catalyst loading on the adsorption were investigated, it was after the synthesis of metal organic framework-5 and defined using different characterization techniques. Equilibrium, kinetic, mass transfer and thermodynamic studies were also performed with this purpose.

All chemicals used in this work were of analytical grade quality purchased from Merck and applied without any purification. MOF-5 was synthesized according the method reported elsewhere (Mueller et al. 2006; Lee et al. 2013). Briefly, a mixture of 9.65 g of zinc nitrate tetrahydrate Zn(NO₃)₂ 4H₂O and 2.05 g of terephthalic acid (BDC) C₈H₆O₄ were dissolved in 297 mL of dimethylformamide (DMF) C₃H₇NO as organic solvent and homogenized under moderate stirring at ambient conditions. Then, the solution was agitated in a glass reactor equipped with a reflux condenser and heated at 130 °C for 4 h. After 1 h and slow cooling, white crystals of MOF-5 emerged; the precipitate was filtered and immersed in DMF (100 mL) overnight. It was filtered again and immersed in chloroform CHCl₃ (150 mL). The solvent was changed three times over 7 days. Finally, the obtained powder was activated for 3 h at 60 °C under a reduced pressure (∼0.2 mbar).

The MOF-5 sample was characterized by X-ray diffraction (XRD) using a Bruker D8-Advance instruments operating at 40 kV and equipped with CuK_α (λ = 1.5405 A) radiation, at a scan rate of 1°/min and 2θ step size of 0.02°. The morphology of the crystalline powder was analyzed with a scanning electron microscope (SEM). Fourier transformed infrared spectroscopy (FT-IR), in the range (400–4,000 cm⁻¹) using a Bruker ALPHA one FT-IR-ATR spectrometer to identify the functional groups. The N2 adsorption/desorption isotherms were obtained at liquid N2 temperature of 77 K using a Micromeritics ASAP 2010 system.

The XRD pattern of MOF-5 (Figure 1(a)) agrees with that reported elsewhere (Hafizovic et al. 2007; Rodríguez et al. 2015; Jiang et al. 2016). The synthesized samples still exhibit several characteristic peaks of MOF-5, and the crystal faces could still be found, such as (200), (220), (400), (420) (Jiang et al. 2016). The important diffraction peaks of MOF-5 are observed at (2θ = 8.8°) with the presence of peaks (marked with hashtags) indicating the presence of zinc terephthalate dihydrate ZTDH (Rodríguez et al. 2015).

The morphology of MOF-5 was studied by SEM analysis and the micrographs were taken with different magnifications (Figure 1(b)) indicate a cubic morphology and with an average size of 20 μm.

The FT-IR spectra for the synthesized MOF-5 (Figure 1(c)) confirmed the presence of several functional groups whose characteristic bands at 1,386 and 1,578 cm⁻¹ correspond to the asymmetric and symmetric stretching of the C-O bond of the carboxylate group directly bonded to Zn (Zhang et al. 2019). The small bands in the region (656–1,145 cm⁻¹) are assigned to the C-H group of the benzene ring of terephthalic acid. The characteristic peak at 500 cm⁻¹ is caused by Zn-O vibration (Iswarya et al. 2012). Furthermore, the wide band (3,390 and 3,606 cm⁻¹) and the sharp narrow peak are attributed to the O-H group of adsorbed water. This indicates the presence of humidity that is frequently located in the metal coordination, due to its high specific surface area (Phan et al. 2010).

The BET surface area was measured by N₂ adsorption/desorption at 77 K (Figure 1(d)). As can be observed, the MOF-5 exhibits a microporous nature with an isotherm of type I (Langmuir type) adsorption without hysteresis, according to the IUPAC classification and this indicates that our adsorbent has a microporous nature. As derived from the adsorption data, the properties of the porous structure of MOF-5 are summarized in Table 1.

Effect of contact time. The contact time is an important parameter for practical application. The results (Figure 2(a)) show that MO uptake by MOF-5 is rapid in the first 15 min, reaching ∼53% of the total MO removal, with the adsorption capacity reaching ∼260 mg g⁻¹ due to the large surface area of the adsorbent (BET surface area = 532.86 m² g⁻¹), and then reaching equilibrium after 30 min. This observation is due to the number of available sites for adsorption of MO being large in the initial step and tends to saturation (Bakhtiari & Azizian 2015). The rapid uptake of dye molecules from aqueous solution indicates the efficiency of the adsorption process and the advantage of the use of this adsorption system in effluent decolorization.

Effect of pH. The influence of pH on the MO adsorption was investigated over a broad pH range (3–10); Figure 2(b) shows that the maximum adsorption percentage and capacity of MO are obtained at pH ∼3 (74.53% and 279.5 mg g⁻¹ resp.). However, with increasing pH both the percentage and adsorption capacity decrease slightly to reach 70.3% and 263.5 mg g⁻¹ at pH ∼10. This is due to the positive charge of the adsorbent MO, which decreases with increasing pH. The maximum adsorption at lower pH may be due to the electrostatic attractions between negatively charged functional groups of MO and the positive charge of the adsorbent surface (Mittal et al. 2007; Gu et al. 2014). Hydrogen ions also act as a bridging ligand between the adsorbent and MO molecules (Gu et al. 2014). At high pH, the adsorbent acquires a negative charge and no hydrogen bonds can be formed between MO and the MOF-5 surface. In this case, only the Van der Waals force contributes to MO adsorption.

Effect of the adsorbent dose. The adsorption is mainly a surface phenomenon and the adsorbent dose is a significant parameter. The surface available for the adsorption process, and consequently the MOF-5 dose, can considerably affect adsorption efficiency. Its influence on the MO adsorption was examined by the addition of different doses (0.1–0.6 g L⁻¹) to 100 mL of MO (Figure 2(c)). The MO removal increases to 78.6% when raising the MOF-5 dose to 0.6 g L⁻¹ and remains almost unchanged thereafter. However, the amount of MO adsorbed, q_t (mg g⁻¹), decreased from 619.53 to 196.6 mg g⁻¹. This is due to the high number of unsaturated sorption sites during the adsorption process.

Effect of initial concentration of MO. The influence of the MO concentration on adsorption is determined by fixing the dose of MOF-5 (0.1 g L⁻¹) and taking a different MO dye concentrations ranging between 10 and 250 mg L⁻¹. Figure 2(d) shows clearly three distinct steps of adsorption: The first step, the adsorption capacity and the MO removal efficiency are weak; in the second step the adsorption capacity qe (mg g⁻¹) increases from 11.37 to 1,270.8, when the C₀ increases from 30 to 200 mg L⁻¹ and the MO removal efficiency increases from 3.4 to 60.6%. Finally, in the third step, the adsorption capacity value reaches a plateau for the higher concentrations, while the MO removal efficiency decreases. Therefore, it can be said that the weak of adsorption capacity and the MO removal efficiency in low concentrations can be explained by the lack of adsorption affinity for MO on the surface of MOF-5, by the fact that water adsorption is more important than MO on MOF-5, because of the nature of MOF-5 and smaller size of the water molecule, which expands the hydrogen bonding network and fills the free space more easily than MO. After reaching a certain concentration, the adsorption of MO is increased due to adsorbent-adsorbent cooperative interactions. This leads to the formation of multilayer adsorption through van der Waals reactions and competitive water substitution. However, in the third step the MO removal efficiency decreases, which may be due to the fact that at higher initial concentrations of the MO dye, the number of MO dye molecules is greater than the number of MOF-5 surface sites available for adsorption.

Effect of temperature. The thermal effect on the removal of MO by MOF-5 has been studied in series of experiments that were carried out in the range (303–333 K) (Figure 2(e)). The amount of MO adsorbed and the percentage removal increases only slightly from 128.87 mg g⁻¹ and 64.4% to 133.6 mg g⁻¹ and 66.8% with raising the temperature, indicating an endothermic process. Such results can be explained by the increased solution viscosity, leading to improved exposure to active adsorption sites, on the one hand, and on the other hand the increase of the intraparticle diffusion rate (Mekatel et al. 2015).

For the solid-liquid system, the equilibrium is an important physico-chemical aspect in the adsorption mechanism (Barkat et al. 2015). In this work, the four models of Langmuir, Freundlich, Toth, Temkin, Dubinin-Radushkevich (D-R) and Dubinin-Astakhov isotherms (Table 2) were tested to investigate the relationship between the amount of MO adsorbed by MOF-5 and their equilibrium concentration in aqueous solution (Figure 3). The results gathered in Table 3 display the non-linear fitting along with the corresponding constant and correlation coefficient (R²). The non-linear regression analysis of the isotherm models was performed with the Origin Pro 9 software running on Windows 10 platform.

According to the R² values and resemblance of the theoretical and experimental values, these indicate that the Dubinin-Astakhov model gave a good fit with the MO adsorption isotherms onto the MOF-5. The obtained value of n = 9.9 was consistent with the reported n range (1–35) of the Dubinin-Astakhov model (Wood 2001), an indication of the homogeneity of the MOF-5 pore structure in comparison to other materials. For instance, the exponent values proposed are n > 4 for zeolites, n = 2–3 for homogeneous carbons, and n < 2 for heterogeneous active carbons. As has been pointed out, n values depend on both the adsorbate and the adsorbent and it is very difficult to have a universal exponent (Inglezakis 2007). The adsorption energy E gives information about the chemical or physical properties of the sorption. The calculated adsorption energy value of adsorption of the MO by MOF-5 is very small (E = 0.055 KJ mol⁻¹ < 8 kJ mol⁻¹) and this implies that the type of adsorption is physical. In this model, q_m represents the maximum adsorption capacity, C_e is the equilibrium concentration of solute in the solution, E is the free energy for a given adsorbate-adsorbent system, which is related to pore dimensions, and the exponent n is the degree of heterogeneity of the micropore system. It is characteristic of the adsorbent, and related to its structure; the more homogeneous the pores the greater its value (Inglezakis 2007). R (KJmol⁻¹ K⁻¹) is the universal gas constant and T (k) is temperature.

The adsorption study was undertaken to evaluate both the efficiency of the process and the rate limiting step that controls the adsorption uptake of MO onto MOF-5, and it also aids in determining the overall reaction time. To determine the MO adsorption mechanism; that is, the transfer and chemical reaction, the pseudo first order, pseudo second order, elovich, intra-particle diffusion and film diffusion models were tested for MO adsorption as a function of time (Table 4). The plots of model form are shown in Figures 4 and 5, while the determined constants are presented in Tables 5 and 6.

As seen in Table 5, the pseudo first and the pseudo second order models exhibited relatively higher values of R², which are superior to 0.97. However, it should be mentioned that the correlation coefficients R² corresponding to the pseudo first order kinetic model were higher in comparison with the pseudo-second order kinetic model for high concentrations. In addition, the calculated from the pseudo-first order kinetic mode were in excellent compliance with the experimental ⁠. It can imply that the adsorption phenomenon at high concentration is reversible (Ahmaruzzaman & Reza 2015), which means that the adsorption mechanism is subject to physical absorption between layers.

As for the low concentrations, the pseudo second order kinetic model gave a better fit, which may indicate that the adsorption mechanism involves electrostatic interactions between the adsorbent and the first layer of adsorbate, and may be considered a chemisorption process (Vargas et al. 2011; Laib et al. 2019). As can be observed from the Elovich results in Table 5, which exhibited the highest values of R² for the low concentrations compared to the high concentrations. The latter is used to describe chemical adsorption on heterogeneous surfaces (Wu et al. 2009).

The modeling of the data obtained by the intra-particle diffusion model reveals a correlation of multi-linearity, as shown in Figure 5(a). This proves the presence of more than one sequential step in the adsorption phenomenon (Haddad et al. 2018). The k_int and C parameters corresponding to each step are reported in Table 6. According to Weber's model (intra-particle diffusion model), the non-null values of the intercept C indicate that during adsorption of MO onto MOF-5 not only is the intra-particle diffusion controlled by them, but also by other diffusion processes. Moreover, if the C values are equal to zero (the plots that pass by the origin), this indicates the adsorption is controlled by the intra-particle diffusion (Suteu & Malutan 2013; Wakkel et al. 2019). One can see from the data of Table 6 that with increasing concentration, the boundary effects represented by the intra-particle diffusion constant C increase (Suteu & Malutan 2013), whereas the intra-particle rate constant k_int also increases with an increase of concentration. This in turn illustrates an enhanced rate of adsorption (Haddad et al. 2018). Moreover, the opposite variation of these two parameters shows that the mass transfer was controlled at the beginning of MO adsorption by a fast intra-particle diffusion process followed by a slow combination diffusion process. In addition, the horizontal portions of the plot reflect that intra-particle diffusion started to slow down and the adsorption of MO reached equilibrium (Haddad et al. 2018).

Boyd's model (film diffusion model) is based on the assumption that particle diffusion is not the sole rate-controlling process. So kinetic data is analyzed using this model (Figure 5(b) and Table 6), in order to identify the rate-limiting step involved in the MO adsorption film diffusion and intra-particle diffusion. The Boyd model is applied by plotting the fractional approach to equilibrium F and time t (Tsibranska & Hristova 2011). As shown in Figure 5(b), the plotted values of B against the experimental values of t present multi-linearity correlation that reveals the existence of more than one sequential step in the adsorption phenomenon (El-Khaiary & Malash 2011). We can also see plots with two linear segments (with one breakpoints) for the concentrations from 10 to 100 mg L⁻¹, and plots with three linear segments (with two breakpoint) for the high concentrations. Whereas the first plots of any concentration and the last plots of high concentration are very near to passing through the origin (see the intercept value from Table 6). According to Boyd's model (El-Khaiary & Malash 2011), film diffusion is not the sole controlling step, whereas intra-particle diffusion is considered more rate-controlling. This finding strengthens the above results of the intra-particle diffusion model.

To consider that the intra-particle diffusion is rate limiting, the intra-particle diffusion coefficient D should be ranged between 10⁻¹¹ and 10⁻¹³ cm² s⁻¹(Karthikeyan et al. 2010). As for whether the film diffusion coefficient D_f is in the interval of 10⁻⁶–10⁻⁸ cm² s⁻¹ then film diffusion can be considered as the rate controlling phenomenon (Karthikeyan et al. 2010). As shown in Table 6, the calculated value of the intra-particle diffusion coefficient D is high in the range of D ≥ 10⁻¹¹, as the film diffusion coefficient D_f values are low in the range of D_f ≤10⁻⁸. Consequently, and based also on the intercept values (Table 6), it can be declared here that the adsorption processes of MO on MOF-5 are controlled by both film and intra-particle diffusion steps, this latter being more significant.

The stability of MOF-5 was investigated by means of recycling experiments, different solvents were tested (Figure 7(a)) such as H₂O, 5% solutions of NaOH, H₂SO₄, NaCl and 5–50% CH₃OH. The desorption of MOF-5 was done in the most efficient eluent (CH₃OH 50%) for three adsorption/desorption cycles (Figure 7(b)). As can be observed in Figure 7(b), there was loss of MO removal efficiency during the third cycle process. The reduction of MO removal efficiency can be explained by the following reasons: (1) there was inevitable loss of adsorbent during the whole adsorption/desorption process; (2) the adsorbed fraction of contaminant on the surface of the adsorbent inhibited the interaction of MOF-5 and MO; (3) there is a possibility of deteriorating structure of the MOF-5.

In this comparison, the maximum adsorption capacities of MO were pooled by different absorbents at the optimum condition in which it occurred, like the pH, adsorbent dose required, initial concentration and the results of the equilibrium isotherm, kinetic, and thermodynamic parameters. Comparing the q_max values of MO adsorption onto MOF-5, our adsorbent exhibits a high adsorption capacity (q_max = 1,248.35 mg g⁻¹) at pH of the solution within a high initial concentration of dye of 200 mg/L and a very low adsorbent dose, this comparison revealed that MOF-5 is a promising adsorbent for the removal of organic dyes. Moreover, the adsorption equilibrium is well fitted by the Dubinin-Astakhov model, suggesting that MO adsorption is a physical adsorption process (free energy E = 0.055 kJ mol⁻¹) onto MOF-5 heterogeneous surfaces and in the presence of a more homogeneous pore structure (n = 9.9). This is also shown in Table 8. The values of k_L, k₁ and k₂ determined from the adsorption of MO onto MOF-5 have high value in almost all cases compared to other studies. Therefore, not only did MOF-5 take up large amounts of the dye, but it also adsorbed MO quickly. However, it should be mentioned that the pseudo-first-order kinetic model fitted well the adsorption of MO onto MOF-5. This can imply that the adsorption phenomenon is reversible, and the value of ΔG° and ΔH° also confirmed that adsorption is a physisorption process. An additional advantage of MOF-5 is that preparation is economic as MOFs are synthesized by ‘one-pot’ solvothermal methods under mild conditions. Herein, we report a facile route for low cost and low temperature synthesis of MOFs. However, the higher adsorption capacity of the adsorbents may compensate for the cost for such a process.

The adsorption of MO onto microporous MOF-type materials has been studied. The adsorption capacity (1,248.35 mg g⁻¹) and adsorption kinetic constant of MOF-5 have high values in almost all cases compared to other studies, showing the importance of porosity and pore size for adsorption. The results of equilibrium studies show a sigmoidal course (S-shaped), identical to type V, which is well fitted by the Dubinin-Astakhov model applicable for physical adsorption processes (free energy E = 0.055 kJ mol⁻¹ < 8 kJ mol⁻¹) onto heterogeneous surfaces and a more homogeneous pore structure (n= 9.9). Based on the rate constant (pseudo-second or pseudo-first-order kinetics for adsorption) and adsorption capacity, it can be suggested that there is a specific interaction like electrostatic interaction between MO and the adsorbent for rapid and high uptake of the dye. As can be observed from the evaluation of the kinetic data suggested by PFO, PSO and Elovich, the surface of the adsorbent is heterogeneous with different active sites for the MO adsorption. Whereas the adsorption phenomenon at high concentrations is reversible, as for the low concentrations, the data indicate that the adsorption mechanism involves electrostatic interactions between the adsorbent and the first layer of adsorbate. The adsorption mechanisms were well described by both internal and external diffusion models, in which it was found that intra-particle diffusion was the main controlling step. In addition, according to the calculated intra-particle diffusion coefficient and the film diffusion coefficient, one can consider that intra-particle diffusion was associated with the control of film diffusion phenomenon. The adsorption of MO over MOF-5 at various temperatures shows that the adsorption is spontaneous and endothermic and the randomness increases with the adsorption of MO. The driving force of MO adsorption over MOF-5 is mainly due to an entropy effect rather than an enthalpy change. Finally, this study demonstrates the efficiency of microporous MOF-type materials that are sensitive to water in practical water treatment applications, and provides a feasible route to prepare efficient MOFs adsorbent for environmental treatment.

All relevant data are included in the paper or its Supplementary Information.