Aqueous synthesis of highly adsorptive copper–gallic acid metal–organic framework

A greener route to synthesize mesoporous copper–gallic acid metal–organic framework (CuGA MOF) than the conventional method using harmful DMF solvent was proposed in this study. Various synthesis attempts were conducted by modifying the synthesis conditions to produce CuGA MOF with comparable physical properties to a reference material (DMF-synthesized CuGA NMOF). The independent variables investigated include the molar ratio of NaOH to GA (1.1 to 4.4) and the synthesis temperature (30, 60, 90 °C). It was found that proper NaOH addition was crucial for suppressing the generation of copper oxide while maximizing the formation of CuGA MOF. On the other hand, the reaction temperature mainly affected the stability and adsorption potential of CuGA MOF. Reacting Cu, GA, and NaOH at a molar ratio of 1:1:2.2 and a temperature of 90 °C, produced mesoporous MOF (CuGA 90–2.2) with a surface area of 198.22 m2/g, a pore diameter of 8.6 nm, and a thermal stability of 219 °C. This MOF exhibited an excellent adsorption capacity for the removal of methylene blue (124.64 mg/g) and congo red (344.54 mg/g). The potential usage of CuGA 90–2.2 as a reusable adsorbent was demonstrated by its high adsorption efficiency (> 90%) after 5 adsorption–desorption cycles.

The ever-growing population and developing industries cause major impact on the environment. Water emerges as one of the highly affected resources due to the increasing usage of synthetic dyes in various sectors, including food processing, pharmaceutical, cosmetic, plastic and textile industries 1 . Annually, over 7•10 5 tons of synthetic dyes have been generated and approximately 10-15% of it was leached out during coloring process and may end up in the water bodies. Anionic dyes (i.e., congo red, CR) and cationic dyes (i.e., methylene blue, MB) are the most frequent synthetic dyes found in dye-contaminated water 2,3 . Earlier studies have reported the adverse effect of these dyes for environment and human health, where extended contact may trigger kidney disease, cancer and allergies 4,5 . Due to the severity of dye contamination, several removal methods, such as electrolysis, membrane filtration, photo-assisted degradation, and adsorption, have been developed [6][7][8] . Among these methods, adsorption is a favorable approach due to its low energy consumption and processing cost, safer by-product, and simple regeneration techniques 9,10 . Therefore, it triggers researches on the development of various adsorbent material with efficient dye removal ability.
In the past decades, metal-organic frameworks (MOFs) have emerged as one of the extensively studied adsorbents. MOFs offer better adsorptive performance than the commonly used adsorbent such as bentonite, activated carbon, and zeolite [11][12][13][14] . This superiority is attributed to their high surface area, large pore volume, and numerous adsorption sites 15 . Copper-gallic acid MOF (CuGA MOF) is a bio-MOF since it composes biologically active phytochemical (i.e., GA) and essential metal element (i.e., Cu) 16,17 . To date, the applicability of CuGA MOF is inferior to other extensively studied MOFs such as MIL-100, MIL-101, UiO-66, ZIF-8 18 . While the application of CuGA MOFs as drug carrier with superior antitumor activity has been reported previously 19 , their adsorptive behavior remains elusive. The preparation of CuGA MOF was first introduced by Sharma et al. who reacted Cu(II) and GA with a mixture of solvents containing Aerosol OT, n-butanol, and N,N-dimethylformamide (DMF) 19 . However, the use of DMF is not desirable due to its harmful potential to the environment 20 . Hence, it is necessary to develop a greener route for synthesizing CuGA.
In this work, the aqueous synthetic method involving the use of NaOH is explored. NaOH was used to create alkaline condition to trigger the dissociation of organic ligands making it accessible for interaction with the metal core 16,21,22 . The optimum molar ratio of NaOH to GA which is a prerequisite for the formation of CuGA MOF was determined. This optimum ratio restricts the formation of undesirable product, such as metal hydroxide and oxide species, which were commonly found in NaOH-rich condition 22,23 . Concurrently, the developed synthesis method eliminates the use of organic solvent and reduces the reaction time to 2 h. In addition, the effect of synthesis temperature on the adsorption ability of the CuGA MOFs was highlighted and the reusability of CuGA MOF for the removal of CR and MB was assessed in this study.

Results and discussion
In the following sections, the synthesized CuGA MOF samples were designated as CuGA T-X; with T refers to the synthesis temperature, and X refers to the molar ratio of NaOH to GA.
Effect of NaOH to GA molar ratio on the characteristics of CuGA. To obtain the optimum conditions for CuGA MOFs synthesis, reactions were carried out at different reaction temperatures and molar ratios of the reactants. The first set of reactions were done by varying X with temperature fixed at 60 °C. The purpose was to determine the proper amount of NaOH required to produce CuGA MOF with comparable characteristics to that of the reference material, CuGA NMOF 19 . Product yield and XRD pattern were employed as the basis for this screening. As shown in Fig. 1a, the highest yield (34.66%) was obtained at X = 2.2.
A comparison of the PXRD patterns between CuGA 60-X and reference CuGA NMOF is given in Fig. 1b  www.nature.com/scientificreports/ The formation of Cu 2 O can be attributed to the oxidation activity that occurs due to the presence of excessive NaOH 24 . These results suggest that X = 2.2 is the most suitable NaOH to GA molar ratio for the synthesis of CuGA MOF, and therefore, this NaOH concentration was chosen for further studies. The amount of NaOH involved in the synthesis also significantly affected the crystallinity of the CuGA. Based on the PXRD pattern analysis, it was observed that CuGA 60-1.1 exhibited lower crystallinity (23.9%) than that of CuGA 60-2.2 (40.1%) and CuGA 60-3.3 (35.4%). These results illustrate the importance of NaOH in the formation of CuGA. NaOH has a modulatory role in promoting the coordination between Cu and GA. As a base, NaOH triggers the deprotonation of carboxylic acid and hydroxyl group of GA, thus produces electron-rich GA. This negatively-charged GA then provides better attraction towards the surrounding metal ions and facilitates interaction of Cu and GA 19 . At a lower NaOH concentration (X = 1.1), there was only a limited amount of OH − ions available for GA deprotonation. Therefore, the affinity of GA molecules toward the metal ions was low. On the other hand, at a higher NaOH concentration (X = 3.3 and 4.4), excessive OH − ions may inhibit the Cu to GA coordination bonds and triggers the formation of metal hydroxide compound instead 22 .
Effect of temperature on the synthesis of CuGA. To 25 . This result suggests that higher synthesis temperature is preferred to form more stable coordination between Cu and GA.  Fig. 3c. A hysteresis loop was observed at P/P 0 = 0.5, and the curve is a type IV isotherm curve which indicates that CuGA 90-2.2 has a mesoporous structure with an H3 hysteresis loop caused by a delay of the desorption process [27][28][29] . The calculated BET surface area, total pore volume, and average pore diameter of CuGA 90-2.2 are 198.22 m 2 /g, 0.4262 cc/g, and 8.6 nm, respectively which are comparable to the CuGA NMOF reported by Sharma et al. (172 m 2 /g, 0.73 cc/g, and 2.2 nm, respectively) 19 . These results suggest that the newly developed synthetic method at relatively low temperature (90 °C) and in the absence of solvents can be used to produce CuGA MOF with similar characteristics as the DMF-synthesized CuGA NMOF. The SEM image of CuGA (Fig. 3d) shows the irregular granular shaped particles with rough surface morphology. Meanwhile, no specific geometry shape can be observed.
Effect of pH on the adsorption of dye by CuGA 90-2.2. One of the critical variables that affect the adsorption process is the initial pH of solution since it affects the surface charge of adsorbent and adsorbate 30 . Figure 4a shows the effect of initial pH value on the adsorption efficiency of CuGA 90-2.2 for MB and CR removal. The percent removal of MB and CR peaked at pH 6 and 7, respectively. This pH-dependent adsorption behavior of MB and CR onto the adsorbent can be explained from the pH PZC and zeta potential of CuGA 90-2.2, as shown in Fig. 4b. The pH PZC of the CuGA 90-2.2 is about 4.41; thus, at pH 6 and 7, the surface charge of this MOF is negative. www.nature.com/scientificreports/ The negative surface charge is also confirmed from the zeta potential value (Fig. 4c). At pH 6, CuGA 90-2.2 has a surface charge that is opposite from the cationic MB, suggesting that their interactions was driven by electrostatic forces 31 . Meanwhile, binding mechanisms such as π-π stacking 32 , might be the contributing force for the interaction of CuGA 90-2.2 and CR since both are negatively charged at pH 7.
Adsorption isotherm. The adsorption isotherm gives the relationship between adsorbate and adsorbent in equilibrium condition. The adsorption isotherm curves of MB and CR at 303, 313, and 323 K on the CuGA 90-2.2 are presented in Fig. 5.
Based on the shape of the initial slope, the curves belong to class H-curve with subclass 2; which indicate that the adsorbate molecules have high affinity toward the adsorbent 33 . Furthermore, a long plateau was observed in the curve, which shows that the adsorption of adsorbate reached saturation 34 . 2-Parameter models (i.e., Langmuir, Freundlich) and 3-parameter Sips model were used to fit the adsorption data. The calculated parameters, along with R 2 values, are presented in Table 1. The R 2 values are close to 1 for Langmuir and Sips models, indicating good fittings. Furthermore, the calculated q max of Langmuir and Sips (q max,L and q max,S ) models also show good agreement with the highest adsorption capacity obtained from experimental data (q exp ). The dimensionless n parameter of the Sips model can show the heterogeneity tendency of the adsorbent. The value of n parameter in Sips, which is close to 1 implies a homogeneous adsorption process. Although a lower R 2 was obtained for the Freundlich model, the value of n in the Freundlich model can show the favorability of the adsorption. The n parameter of the Freundlich model falls between 2-10 for all systems, indicating that the adsorption is favorable 34 .
CuGA 90-2.2 can adsorb a higher amount of CR than MB as indicated by the q max and q exp values. While the adsorbent-adsorbate electrostatic attractions cannot explain this phenomenon, the interaction can be related to the presence of Cu-nodes in CuGA 90-2.2. In hard-soft acid-base theory, Cu is classified as borderline-soft acid that has better interaction with borderline base 35,36 . Based on its molecular feature, CR possesses sulfonate groups, a borderline base, thus provide better interaction to the Cu-nodes of CuGA 90-2.2. In addition, the presence of benzene rings in CR and MB structures may contribute to dye adsorption onto CuGA due to the occurrence of π-π interaction (Fig. 6). The greater number of benzene rings in the CR structure facilitates π-π interactions with the CuGA 90-2.2 32 .
The FTIR, XRD, and SEM analysis were conducted on the post adsorption CuGA 90-2.2. The FTIR chromatogram of the dye-loaded CuGA 90-2.2 (referred to as MB@CuGA and CR@CuGA) shows the alteration on the spectral shape compared to the spectra of CuGA 90-2.2 before adsorption. Several new absorption bands appear in the spectra of MB@CuGA and CR@CuGA (Fig. 7a), which confirms the adsorbate and adsorbent  www.nature.com/scientificreports/ and -N=N-stretching vibration of CR, respectively 38 . SEM images of the MB@CuGA and CR@CuGA was also observed. However, no apparent alteration was observed from the SEM images of the dye-loaded CuGA (Fig. 7c,d).
It worths noting that the synthesized CuGA 90-2.2 has higher adsorption capacity toward MB and CR than the other MOF-adsorbents reported in literatures, as shown in Table 2. The maximum adsorption capacity of CuGA 90-2.2 is 1.4 to 5.7 times higher for MB, and 1.4 to 5.3 times higher for CR than the other reported MOFs. This shows that the synthesized CuGA 90-2.2 can be a promising adsorbent for either anionic or cationic dye adsorption. Furthermore, the aqueous synthesis of CuGA 90-2.2 can be considered as an eco-friendly process.
Adsorption thermodynamics. The thermodynamic parameters, including the changes in enthalpy (ΔH), free energy (ΔG), and entropy (ΔS), in the adsorption of MB and CR onto CuGA 90-2.2 were determined. According to the linear relationships, the parameter ΔH and ΔS can be obtained as the slope and intercept of 1/T versus ln (K C ) plot which obtained based on the following equations: where R is the universal gas constant (8.314 J/mol K), T is the temperature (K), and K C is the dimensionless distribution coefficient for the adsorption. The value of K C can be calculated as the ratio between the equilibrium concentration on the solid phase (C a , mg/L) and in the solution (C e , mg/L) 47 . ΔG was calculated using the equation below: The calculated values of ΔH, ΔS, and ΔG are given in Table 3. The negative ΔH of the adsorption systems denote that the heat was released during the process and that the adsorption occurred exothermically 48 . The negative ΔS implies the decrease in the degree of freedom of the molecules due to the fact that the bonded adsorbate molecules have less freedom of movement than when they are free in solution. Compared to the MB system, the less negative of ΔS value for the CR system can be attributed to the softer binding potential in this system 49 . The negative ΔG values of MB and CR adsorption indicate the thermodynamic favorability and spontaneous nature of the adsorption process 50 .   www.nature.com/scientificreports/ Reusability. Adsorbents with high reusability can benefit from the aspects of process-cost savings, environmental friendliness, and practicality 51 . Ethanol was employed as the desorption agent for the regeneration of CuGA. The CuGA showed excellent recyclability, retaining high adsorption efficiency (> 90%) after five adsorption-desorption cycles (Fig. 8). After fifth adsorption-desorption cycles, the efficiency of CuGA is 94.96% and 93.82% for MB and CR, respectively-with the corresponding adsorption capacity for MB and CR removal of 115.51 mg/g and 310.67 mg/g, respectively. To confirm the stability of CuGA 90-2.2 after 5 adsorption-desorption cycles, the PXRD patterns of the CuGA were recorded. As shown in Fig. 7b, there is no significant change in the PXRD pattern of the adsorbent before and after 5 cycles of dye removal.

Conclusion
An aqueous synthesis method of CuGA MOF has been developed to replace the conventional synthesis method. Two important conditions in the synthesis that affect the successful formation of CuGA MOF are the synthesis temperature and the molar ratio of NaOH to GA. A proper amount of NaOH addition is important in promoting the formation of CuGA MOF while preventing metal oxide formation. A synthesis temperature of 90 °C was found to facilitate the formation of CuGA with stable coordination. Synthesis temperature also profoundly affected the adsorption ability of the resulting CuGA MOF for dye removal. Compared to other extensively studied MOFs, the aqueous-synthesized CuGA MOF possesses higher adsorption capacity for MB and CR. The CuGA MOF showed high reusability with adsorption efficiency decreased less than 10% after 5 adsorption-desorption cycles. The dark brown solids obtained from the reaction were collected then rinsed with DI water 5 times. Finally, the solids were freeze-dried using a Labconco freeze dryer for 24 h. The as-synthesized solids were referred to as CuGA T-X (T refers to the reaction temperature; X refers to the molar ratio of NaOH to GA, as listed in Table 4).

Materials and methods
Characterization of CuGA MOF. Elemental analysis of the complex was carried out using a Thermo Flash 2000 CHNS/O Analyzers. The crystallinity pattern was determined by using a Powder X-Ray Diffraction (PXRD, Bruker D2 PHASER XE-T XRD) operated at 30 kV; 10 mA using Cu-Kα radiation (λ = 0.154060 nm) at 3° PSD opening, 0.5 times per step and 2Theta from 5° to 60°. The surface functional group analysis was performed by using a Fourier-Transform Infrared Spectroscopy (FTIR, SHIMADZU Tracer-100); the analysis was carried out in the range of 400-4000 cm −1 wavenumbers and KBr pellets were applied as the background. Thermogravimetric analysis (TGA, TA instruments/TGA 550) was carried out at temperature range 30-650 °C with a heating rate of 10 °C/min. The isotherms of N 2 adsorption-desorption were measured at 77 K using a BELSORP-max analyzer. The samples were degassed for 4 h at 383 K before analyzing. The specific surface area was calculated from the adsorption branch using the Brunauer-Emmett-Teller (BET) model. The pH PZC and zeta potential measurement were done according to the previous report 34,52 .  where Q e (mg/g) is the equilibrium amount of adsorbate adsorbed, C 0 and C e (mg/L) represents the initial and final (post-adsorption) concentrations of dye, respectively, V (L) is the volume of dye solution, and W (g) is the weight of CuGA MOF.
Adsorption modeling. SigmaPlot 12.5, a Systat software, was used for adsorption data modeling and fitting. The models used include 2-parameters model (i.e., Langmuir, Freundlich) and 3-parameters Sips models. The Langmuir model can be expressed as: where K L (L/g) is the Langmuir affinity constant related to the energy of adsorption, and Q max (mg/g) is the maximum adsorption capacity of the CuGA MOF with assumption of monolayer surface coverage 53 . Freundlich isotherm model was used to describe the multilayer adsorption with the interaction of adsorbent and adsorbed molecules. Freundlich model is represented by the following equation: where K F ((mg/g)(L/mg) −1/n ) is the Freundlich constant related to the adsorption capacity of the adsorbent, and n (dimensionless) is the heterogeneity and adsorption intensity of adsorbent 54 .
Sips isotherm model is a combined form of Langmuir and Freundlich expressions designed for predicting the heterogeneous adsorption systems while evading the limitation of rising adsorbate concentration. At low adsorbate concentrations, it reduces to Freundlich isotherm; and it predicts monolayer adsorption, a characteristic of the Langmuir isotherm, at high adsorbate concentrations. Sips isotherm model is represented by following equation: where Q ms (mg/g) is the Sips maximum adsorption capacity of the adsorbent, Ks (L mg −1 ) is the Sips equilibrium constant, and β is the Sips model exponent, which can also be employed to describe the system's heterogeneity when is between 0 and 1. When β = 1, the Sips equation reduces to the Langmuir equation which implies a homogeneous adsorption process 34 .