Addition of graphene oxide to ZIF-8/HKUST-1 composite for enhanced adsorptive and photocatalytic removal of congo red in wastewater

Ternary composites comprised of graphene oxide (GO), ZIF-8


Introduction
The development of the textile industry has resulted in large amounts of dye waste, causing environmental pollution due to the poisonous and hard to decompose characteristics of the waste, and also the limitation of water resources (He et al., 2014).One of the most commonly used dye is Congo Red (CR), a sodium salt from benzidinediazo-bis-1-naphthylamine-4-sulfonic acid which has a molecular formula of C 32 H 22 N 6 Na 2 O 6 S 2 .The complex chemical and molecular structure of CR may cause dangerous effects towards humans, such as respiratory, skin, and eye diseases, in addition to digestive discomforts.Thus, countermeasures against CR dye waste are needed (Liu et al., 2022).
Researchers have put great effort into removing organic pollutants from the environment in order to maintain a clean environment.Various methods have been proposed for the removal of dye waste such as adsorption (Wang et al., 2022), photocatalytic treatment (Liu et al., 2022), electrochemical treatment (Di et al., 2020), and microbial treatment (Rizqi and Purnomo, 2017).However, there are several disadvantages in the use of these methods.In microbial treatments, the degradation process is very slow and can take up to several days (Adewuyi, 2020;Rizqi and Purnomo, 2017).In adsorption methods, organic pollutants are moved from one phase (adsorbate) to another (adsorbent) without degrading hazardous compounds and further problems may arise due to the production of secondary waste (Ediati et al., 2023).Meanwhile, the application of electrochemical treatments is expensive (Di et al., 2020).Photocatalytic treatment is considered to be the best method to degrade organic pollutants due to the effective conversion of harmful dye to H 2 O and CO 2 .Furthermore, photocatalytic treatment is environmentally friendly, does not produce secondary pollutants, and is relatively low cost.In the process, photocatalytic uses light as a catalyst to generate highly reactive radicals that can then be used to break down organic pollutants into less harmful products in the environment through redox reactions (Liu et al., 2022).Owing to their optical characteristics, semiconductors are commonly used as a photocatalyst.Furthermore, by combining different semiconductors, charge recombination can be suppressed and the absorption of photons can be increased (Zulfa et al., 2023).
Metal Organic Frameworks (MOFs) are porous crystals consisting of metal clusters coordinated by organic linkers through coordinated bonds.MOFs have properties that are needed in adsorption, namely dispersed metal atoms, a uniform pore size, and a large specific surface area.Moreover, MOFs have been shown to exhibit excellent photocatalytic performance based on its visible light response due to the presence of lowest unoccupied molecular orbital (LUMO) and highest occupied molecular orbital (HOMO) in their orbital.Several studies have shown the adequate performance of many different MOFs in degrading contaminants in aquatic environments (Bahmani et al., 2020;Moradi et al., 2017;Mosleh et al., 2017Mosleh et al., , 2016aMosleh et al., , 2016b)).
HKUST-1 is a type of MOF composed of copper ions coordinated by 1,3,5-benzenetricarboxylate (BTC) ligands in a cubic lattice (Lin et al., 2012).HKUST-1 is generally used in the removal of contaminants due to its relatively easy synthesis, large surface area, crystallinity, porosity, high thermal stability, high availability, and low cost (Ediati et al., 2021;Li et al., 2022).Combining two types of MOF to produce a heterogenous composite can increase the performance of MOFs in removing organic pollutants (Ling et al., 2021).The resulting heterogenous composite possess a higher specific surface area, thermal stability, chemical stability, and can also further reduce electron-hole recombination compared to pristine MOFs.Azhar et al. (2017) had previously reported that UiO-66/HKUST-1 exhibited better performance than pristine UiO-66 and HKUST-1 in removing MB dye.Furthermore, ZIF-8 is a type of MOF comprised of Zn metal clusters that has been widely used as a photocatalyst.Previous studies have reported that ZIF-8 synthesized with TEA exhibited good photocatalytic performance due to TEA acting as an agent and improving the optical properties of ZIF-8 (Rodríguez et al., 2020;Zulfa et al., 2023).
Co-catalysts also play a crucial role in photocatalysis.Several cocatalysts have been reported such as graphene oxide (GO) (Narindri Rara Winayu et al., 2022), reduced GO (Packialakshmi et al., 2023) and metal (Li et al., 2023).Combining MOF composites with co-catalysts that serve as a bridge for charge separation is a superior strategy to enhance the photocatalytic performance of MOF composites.In recent years, limited research has been carried out on the performance of GO in removing pollutants within wastewater (Chen et al., 2019;Ju et al., 2021;Pan et al., 2021).GO is a semiconductor which has several properties, namely a two-dimensional structure, large surface area, good conductivity, and high electron mobility, that are beneficial in the field of photocatalytic applications (Chen et al., 2019;Ju et al., 2021).Furthermore, GO contains various functional groups such as -OH, -COOH, epoxide and having the sp 2 hybridization from their functional group, which functions to increase adsorption performance through π-π interactions (Chen et al., 2019).The application of HKUST-1/GO composite as an adsorbent to capture SO 2 was reported by Pan et al. (2021).In the study, the addition of GO was shown to result in a better adsorption performance.Liu et al. (2012) reported the application of ZIF-8/GO composite to dissolve Cu(II) in an aqueous solution, in which GO was shown to have a positive impact towards Cu(II) adsorption.In this study, a novel GO/ZIF-8/HKUST-1 composite was synthesized and used to remove CR dye from wastewater.The physical, chemical and optical structure of the composite was characterized using several characterizations.Furthermore, the mechanism of degradation and adsorption behaviour of the proposed composite was analyzed.

Synthesis of HKUST-1
Synthesis of HKUST-1 was initiated by dissolving 1.0333 g of Cu (NO 3 ) 2 .3H 2 O in 7.5 mL of demineralized water prior to being stirred for 15 min.Next, 0.5 g of H 3 BTC was dissolved in 7.5 mL mixture of ethanol: DMF with a ratio of 1:1 (v:v) and stirred for 15 min to obtain a homogeneous mixture.The mixture was then stirred for 30 min before carrying out the solvothermal process for 10 h at a temperature of 100 ⁰C.The mixture was then cooled at room temperature and left idle for 48 h.The obtained solid was washed with DMF and methanol in an oil bath at a temperature of 50-55 ⁰C for 3 days.Subsequently, the solids were dried at a temperature of 80 ⁰C for 4 h.

Synthesis of ZIF-8
Synthesis of ZIF-8 was initiated by dissolving 1.835 g of Zn (NO 3 ) 2 .6H 2 O in 21 mL of 2% acetic acid and stirred for 15 min, which was then labelled as mixture A. Subsequently, 4.5149 g of MeIM was dissolved in a mixture of 7.8 mL TEA and 15 mL demineralized water in a separate container before being stirred for 20 min, which was then labelled as mixture B. Mixture B was dripped into mixture A while being stirred using a magnetic stirrer for 1 h before being left idle at room temperature for 24 h.Next, the mixture was decanted prior to being washed with demineralized water and methanol several times.The mixture was then filtered for its filtrate and solids using a filter paper.The obtained solids were dried in an oven at a temperature of 50 ⁰C for 18 h.

Synthesis of GO/ZIF-8/HKUST-1
The ZIF-8/HKUST-1 composite was synthesized using the same method as the HKUST-1 synthesis.As much as 0.058 g (5%) of ZIF-8 was added into the ligand (H 3 BTC).The ZIF-8/HKUST-1 synthesis method was resumed in accordance to the previously explained HKUST-1 synthesis method.Synthesis of GO/ZIF-8/HKUST-1 was performed by following the synthesis method of ZIF-8/HKUST-1 and adding varying amounts of GO, namely 5, 10 and 20%.The addition of varying amounts of GO was carried out after the mixture was left idle for 48 h, mixed for 30 min, and left idle again for 24 h.

Characterization
The crystallinity of the synthesized materials were analyzed by means of X-ray Diffraction (XRD) in the range between 5 and 50 • using the Philips X'Pert PN-1830 X-ray diffractometer (Bridge Tronic Global Inc, USA), with Cu Kα radiation (λ = 1,5406 Å), at a voltage of 40 kV and a current of 30 mA.The bonding structure of the synthesized materials were analyzed by means of Fourier-transform Infrared Spectroscopy (FTIR) using 8400S SHIMADZU spectrometer (Shimadzu, Japan) at wavenumbers between 400 and 4000 cm − 1 .Morphology and element distribution of the synthesized materials were analyzed by means of Scanning Electron Microscopy and Energy Dispersive X-ray Spectroscopy (SEM-EDX) where the SEM images were taken using Zeiss EVO MA10 scanning electron microscope (Zeiss, Germany).The surface area and pore distribution of the MOF composites were examined based on N 2 adsorption-desorption isotherms measured using a Quantachrome NovaWin gas sorption analyzer (Quantachrome Instrument, USA).The samples were degassed at a temperature of 300 • C for 2 h.Photoluminescence (PL) was measured by means of fluorescence spectroscopy using the Shimadzu RF-5310 spectrofluorometer (Shimadzu, Japan).UV-Vis diffuse reflectance spectroscopy (UV-Vis DRS) was measured using the Shimadzu UV2600-I spectrophotometer (Shimadzu, Japan) to study the optical absorption properties.

Adsorption and photocatalytic performance
The adsorption process at various contact times was performed with an adsorbent dose of 10 mg in 20 mL of CR solution with an initial concentration of 50 mg/L.The adsorption process was performed in a dark environment at a constant stirring speed of 300 rpm.The contact times in this study was varied at 10, 20, 30, 40, 50, and 60 min.After the adsorption process, the mixture was centrifugated, then the remaining concentration of CR in the solution was measured using UV-Vis spectroscopy at a maximum wavelength of 498 nm.The data obtained was plotted on a graph of adsorption capacity (mg/g) against time (minutes).The obtained optimum time was used as the contact time in the CR dye photocatalytic process under UV-LED irradiation.
The photocatalytic activity test towards the elimination of CR dye under UV-LED irradiation was initiated by weighing 0.01 g of HKUST-1, ZIF-8, ZIF-8/HKUST-1, GO(5)/ZIF-8/HKUST-1, GO(10)/ZIF-8/HKUST-1, and GO(20)/ZIF-8/HKUST-1.Each material was then added into a glass beaker containing 20 mL of CR dye.The photocatalytic test was conducted in a photocatalytic reactor under UV-LED irradiation with a constant stirring speed of 300 rpm.The contact times in this study was varied at 10, 20, 30, 40, 50, and 60 min.After carrying out the photocatalytic test, the mixture was centrifugated at 1500 rpm for 20 min.The filtrate obtained from the photocatalytic process was analyzed using UV-Vis spectroscopy at a maximum wavelength of 498 nm.The data obtained was plotted on a graph of Ce/C 0 against time (minutes).

Adsorption kinetics
The adsorption results of the synthesized materials were analyzed using the pseudo first and second order equations to determine the kinetics of adsorption.The pseudo first order equation is shown in Eq. (1).
The results of the calculations using Eq. ( 1) were plotted on a graph of time t (minutes) as the x-axis against ln (Qe -Qt) as the y-axis.As a result, the K 1 values (the pseudo first order adsorption rate constant) were obtained from the slope of the linear equations and from the calculated Qe (Qe cal ) based on the 2.303 log intercept.The pseudo second order equation is shown in Eq. ( 2).
By plotting the results of the calculations using Eq. ( 2) on a graph of time t (minutes) as the x-axis against t/Qt as the y-axis, the value of calculated Qe (Qe cal ) can be obtained from 1/gradient of the slope of the linear equation.Simultaneously, the K 2 values (pseudo second order adsorption rate constant) can also be obtained from slope 2 /intercept.The acquired correlation coefficient value (R 2 ) can be used to determine the suitability of the adsorption kinetics.

Characterization of materials
The diffractogram of the synthesized materials are displayed in Fig. 1a.The characteristic peaks of HKUST-1 were observed at 2θ = 6.57• and 9.33 • with a moderate intensity, 2θ = 11.45 • with the highest intensity, and 2θ = 13.28 • with a low intensity.These results are consistent to that of previous research (Ediati et al., 2023(Ediati et al., , 2019)).The characteristic peaks of ZIF-8 were observed at 2θ = 7. 23, 10.30, 12.63, 16.34, and 17.97 • , which are in accordance to the results of previous research (Zulfa et al., 2023).The characteristic peaks of the ZIF-8/HKUST-1 composite were observed at 2θ = 6.60, 9.40, 11.53, 13.34, and 18.96 • .These characteristic peaks are similar to those of ZIF-8 and HKUST-1, demonstrating that ZIF-8 had been successfully synthesized into the ZIF-8/HKUST-1 composite material.The GO/ZIF-8/HKUST-1 composite exhibited similar peaks to those of the ZIF-8/HKUST-1 composite.However, characteristic peaks of GO were not observed in all the ternary composites.In addition, the intensity of all the ternary composites decreased drastically as the amount of GO increased, which is correlated with the amorphous properties of GO as shown in Fig. 1a (Gabal et al., 2022).The decrease in intensity is caused by less light reflection due to the amorphous surface of a material.
The functional groups of the synthesized materials were analyzed using FTIR.In the FTIR spectra shown in Fig. 1b, the 5 characteristic absorption bands of HKUST-1 can be observed in the spectrum of all the materials containing HKUST-1.The characteristic peak at 727 cm − 1 corresponds to the stretching vibration of Cu-O.The bands at 1367, 1447, and 1640 cm − 1 are assigned to the stretching vibration of C-O, bending vibration of the OH group, and stretching vibration of the C=O group, respectively.Furthermore, the characteristic peak of ZIF-8 at 420 cm − 1 represents the Zn-N stretching vibration.The bands at 995 cm − 1 R. Ediati et al. and 1145 cm − 1 are attributed to the bending vibration of C-N bending and aromatic vibration of C-N.In addition, the band at 1581 cm − 1 corresponds to the bending vibration of C=N.The bands at 2930 and 3132 cm − 1 are ascribed to the stretching vibration of C-H sp 3 and the presence of C-H sp 2 , respectively.The FTIR spectrum of the ZIF-8/ HKUST-1 and GO(10)/ZIF-8/HKUST-1 exhibited characteristic peaks of pristine HKUST-1 and ZIF-8.Fig. 2a shows the SEM image of pristine HKUST-1.It can be seen in Fig. 2a that HKUST-1 possess an octahedral morphology with sharp edges and smooth surface, which correlates with the high crystallinity exhibited in the XRD spectrum of HKUST-1.The addition of ZIF-8 to HKUST-1 changed the morphology of HKUST-1.It can be observed in Fig. 2b that the surface of HKUST-1 becomes slightly covered with the spherical morphology of ZIF-8, demonstrating the independent growth of ZIF-8 particles in the HKUST-1 solution.Based on a previous study, ZIF-8 possess a spherical morphology with nanosized particles when synthesized in an acetic acid solution (Santoso et al., 2021).The GO/ZIF-8/HKUST-1 composite possess an irregular morphology (Fig. 2c-e).Compared to GO(5)/ZIF-8/HKUST-1 and GO (10)/ZIF-8/HKUST-1, GO(20)/ZIF-8/HKUST-1 possess a smoother morphology due to the higher loading of GO.This is related to the presence of GO, which makes the ZIF-8/HKUST-1 structure thinner through a coating process.The EDX spectra shown in Fig. 3 confirms the element distribution of GO/ZIF-8/HKUST-1, in which it can be seen that Cu, Zn, O and C exist in all the GO/ZIF-8/HKUST-1 composites with varying amounts of GO.Table 1 shows that increasing the amount of GO leads to an increase in the amount of the C and O elements in the composite.The amount of C and O in GO(20)/ZIF-8/HKUST-1 are higher compared to GO(5)/ZIF-8/HKUST-1 and GO (10)/ZIF-8/HKUST-1, owing to an increase in the amount of C and O due to the increasing GO loading.
To further examine the crystal structure and microstructure of ZIF-8/ HKUST-1 and GO(10)/ZIF-8/HKUST-1, the materials were characterized by means of transmission electron microscopy (TEM).The TEM image of ZIF-8/HKUST-1 and GO(10)/ZIF-8/HKUST-1 are shown in Fig. 4a and Fig. 4b, respectively.It can be seen in Fig. 4a that ZIF-8 is uniformly dispersed in the pores of HKUST-1.Furthermore, the black dots show the distribution of Zn atoms in the pores of HKUST-1.Fig. 4b exhibits the distribution of ZIF-8 and HKUST-1 particles on the surface of GO, which further proves that GO(10)/ZIF-8/HKUST-1 was successfully synthesized (Cao et al., 2021).These results correlate with the FESEM images where the addition of GO causes the surface of the composite to look smoother.
The presence of GO in the GO/ZIF-8/HKUST-1 composites were further confirmed using Raman analysis, as shown in Fig. 2f.The characteristic raman bands of GO revealed two main bands at 1347 and 1559 cm -1 , corresponding to D and G bands, respectively.Significant changes were observed in the D band as the amount of GO was increased (Rout et al., 2022).The occurrence of the stronger D band caused defects in the basal graphene plane, which became stronger due to the higher interaction with the functional groups in the ZIF-8/HKUST-1 composite.On the other hand, the low D band intensity in GO(5)/ZIF-8/HKUST-1 is due to the low interaction between graphene and the ZIF-8/HKUST-1 composite.Furthermore, there was a slight shift to a higher wavenumber for the D and G bands in the GO/ZIF-8/HKUST-1 composites caused by the addition of GO.

Optical properties of materials
The electronic structure on the recombination rate of the photodriven electron-hole pairs was analyzed using photoluminescence (PL) spectroscopy.Fig. 5a shows the PL spectra of HKUST-1, ZIF-8, and GO (10)/ZIF-8/HKUST-1.The HKUST-1, ZIF-8, and GO(10)/ZIF-8/HKUST-1 composites exhibited emission peaks at 362, 355 and 358 nm, respectively.Although HKUST-1 showed a lower PL intensity, it cannot be concluded that HKUST-1 has a low charge recombination.It can be seen in Fig. 5b that HKUST-1 has the lowest absorbance.Therefore, it can be concluded that the low PL intensity produced by HKUST-1 correlates with the low absorbance of HKUST-1.Furthermore, a lower PL intensity indicates a higher electron-hole separation efficiency (Sheikhsamany et al., 2021).Combining HKUST-1, ZIF-8 and GO to generate a heterojunction was proven to decrease the PL emission  UV-Vis DRS was carried out to study the optical properties of HKUST-1, ZIF-8, and the GO/ZIF-8/HKUST-1 composites.The UV-Vis spectra are shown in Fig. 5b.It can be seen that HKUST-1 has a lower intensity than the other materials.This correlates with the low PL intensity of HKUST-1.By extrapolating the curves toward the X-axis, the absorption edge wavelengths of each photocatalyst were calculated (Wu et al., 2022).The reduction in particle size may be the cause of the move of the absorption maxima (blue shift) towards shorter wavelengths (Ndikau et al., 2017).According to the Debye Scherrer equation, the particle size of HKUST-1 (99 nm) is larger than that of GO(5)/ZIF-8/HKUST-1 (48 nm) and GO(10)/ZIF-8/HKUST-1 (93 nm), causing a blue shift.GO (10)/ZIF-8/HKUST-1 has a larger particle size compared to GO (5)/ZIF-8/HKUST-1, causing a red shift.Based on Fig. 5b, HKUST-1, ZIF-8, GO(5)/ZIF-8/HKUST-1, and GO(10)/ZIF-8/HKUST-1 showed an adsorption band at 365, 391, 326, and 332 nm, respectively.The band gap energy (E g ) can be estimated using the Kubelka-Munk equation.The calculated E g value for HKUST-1, ZIF-8, GO(5)/ZIF-8/HKUST-1, and GO (10)/ZIF-8/HKUST-1 are 3.4, 3.17, 3.8 and 3.73 eV, respectively.The valence band edge (E VB ) and conduction band edge (E CB ) potential can be calculated using the Butler and Ginley equation as follows: where, X is the geometric average of the semiconductor's component atoms' electronegativity and E e is the free electrons energy on hydrogen scale (around 4.5 eV vs. SHE).The calculated X value for HKUST-1 and ZIF-8 are 4.95 and 5.895 eV, respectively (Li et al., 2019;Sofi et al., 2018).Based on Eq. ( 3), the E CB value for pristine HKUST-1 and ZIF-8 were calculated to be − 0.305 and − 1.137 eV, respectively.Meanwhile, based on Eq. ( 4), the E VB value for pristine HKUST-1 and ZIF-8 were calculated to be 3.095 and 2.035 eV, respectively.

Adsorption, kinetic and photocatalytic behavior on CR adsorption 3.3.1. Adsorption-photocatalytic process of CR dye removal
Based on the N 2 adsorption-desorption result, it is known that the photocatalyst is a porous material.Therefore, it is important to carry out the adsorption process before the photocatalytic process to reach adsorption-desorption equilibrium.After 10 min up to 20 min, a drastic increase in adsorption capacity occurred, as shown in Fig. 6a.Subsequently, after 20 min up to 50 min, a non-significant increase in  adsorption capacity occurred.It can be observed in Fig. 6a that the adsorption capacity decreased after 50 min until 60 min.Therefore, it can be concluded that that 50 min is the optimum time in dark conditions, where the optimum adsorption capacity of 83.92 mg/g was on the GO(10)/ZIF-8/HKUST-1 composite.The drastic increase of adsorption capacity within the early minutes of the adsorption process was caused by the high availability of active sites.As the adsorption process further progressed, many of the active sites had already been filled with CR dye leading to the decrease in adsorption capacity (Aulia et al., 2020;Ediati et al., 2023).
Adsorption kinetics plays an important role in studying the adsorption mechanism.Contact time is a vital factor in evaluating the adsorption kinetics for each photocatalyst.In this study, the adsorption kinetics were modeled using the pseudo-first order (Fig. 6b) and pseudosecond order (Fig. 6c) models, previously expressed in Eq. ( 1) and Eq. ( 2), respectively.The kinetic parameters for CR removal in each photocatalyst are shown in Table 2.The pseudo-second order model fitted the experimental data better than the pseudo-first order model, with a correlation coefficient (R 2 ) value for each photocatalyst > 0.96 and a Qe calculation close to the experimental Qe.The pseudo-second order model indicated that the adsorption process occurred chemically and was determined by two variables, namely the concentration of CR and the concentration of active sites on the surface of the photocatalyst.Additionally, the pseudo-second order model demonstrated that the simplicity of resolution and susceptibility to experimental error is minimum (de Oliveira et al., 2023).The correlation coefficient value for each catalyst in the pseudo-first order model was < 0.85, which completely deviates from the experimental data.
After obtaining the optimum time in dark conditions (50 min), the next step is to carry out the photocatalytic process under UV-LED irradiation to degrade the CR dye.Even though the percentage of removal was 83.92% in adsorption conditions, the photocatalytic process was still needed due to the photocatalytic active nature of all the materials which had been previously proven by the optical characterization.All the photocatalysts possess a band gap in the band gap range of semiconductors that are commonly used in photocatalysis.It can be seen in Fig. 6d that the photodegradation of CR within 60 min was 57.23%, 80.21%, and 81.83% for HKUST-1, ZIF-8, and ZIF-8/HKUST-1, respectively.The higher degradation of CR for ZIF-8 than HKUST-1 is due to the lower band gap value of ZIF-8 compared to HKUST-1.In general, a lower band gap value correlates with a higher degradation performance (Li et al., 2021;Liu et al., 2019).On the other hand, the ZIF-8/HKUST-1 composite showed a higher degradation performance than pristine HKUST-1 and ZIF-8 due to the formation of a binary heterojunction which can suppress the recombination rate of electron-hole pairs and increase the charge separation.The addition of GO to the ZIF-8/HKUST-1 binary heterojunction can improve the degradation performance through the formation of a ternary heterojunction, despite the fact that the band gap value is higher than that of pristine HKUST-1 and ZIF-8.GO can act as an electron mediator which can increase the degradation performance by improving the charge separation of the binary heterojunction (Huang and Li, 2023;Wu et al., 2018).The degradation performance of GO(5)/ZIF-8/HKUST-1, GO(10)/ZIF-8/H-KUST-1 and GO(20)/ZIF-8/HKUST-1 were 85.77, 91.81 and 85.54%, respectively.The higher degradation performance of GO (10)/ZIF-8/HKUST-1 was also indicated by the low PL intensity due to the effective charge separation and low charge recombination.Based on these results, the optimal amount of GO added to ZIF-8/HKUST-1 is 10%.The addition of higher amounts of GO, namely over 10%, results in the active sites of the photocatalyst being covered and in turn decreases the degradation performance.Furthermore, the excess of GO weakens the synergistic effects between GO and the MOF and can cover some of the active sites, blocking the fast charge transfer channel (Chen et al., 2022).In addition, Tabel 3 shows the comparison of CR degradation effieciency between GO(10)/ZIF-8/HKUST-1 with other photocatalysts used in previous studies.It can be seen in Table 3 that GO (10)/ZIF-8/HKUST-1 shows great potential as a photocatalyst for CR degradation due to a high degradation efficiency that is obtained in a short amount of time.

Effect of pH
Modifying the pH of the CR solution may alter the surface charge of the photocatalyst.Therefore, pH affects the interaction between CR and the photocatalyst (Cai et al., 2023).The effect of pH on CR degradation was investigated for pH values of 5, 7, 9 and 11 (Fig. 7a).It was shown that by decreasing the pH of the CR solution from 11 to 5, the degradation performance was increased up to 92.14%.The addition of HCl to the CR solution increases the amount of H + ions which results in the surface of GO(10)/ZIF-8/HKUST-1 becoming positively charged.The electrostatic attraction between CR (anionic dye) with the positively charged surface of GO(10)/ZIF-8/HKUST-1 caused a higher degradation of CR in lower pH values.Meanwhile, the addition of NaOH to the CR solution increases the amount of OH − ions which results in the surface of GO(10)/ZIF-8/HKUST-1 becoming negatively charged.The electrostatic repulsion between the CR dye and GO(10)/ZIF-8/HKUST-1 lead to the decrease in photocatalytic performance (Gangwar and Jeevanandam, 2023).

Effect of scavenger
The main active species in the photodegradation of CR dye were further investigated through trapping experiments (Wang et al., 2023), in which methanol, ascorbic acid, IPA and DMSO were employed as the trapping agents of holes (h + ), superoxide radical (•O 2− ), hydroxyl radical (•OH), and electrons, respectively (Zulfa et al., 2023).As shown in Fig. 7b, the addition of DMSO and methanol caused a slight inhibition towards the photocatalytic degradation, indicating that electrons and holes are not the main active species in the photodegradation of CR dye.However, it can be observed that the presence of ascorbic acid and IPA induces a higher inhibition towards the photocatalytic degradation, indicating that superoxide radicals and hydroxide radicals are the main active species in the photodegradation of CR dye.Notably, IPA causes a more obvious inhibition towards the degradation of CR dye, which suggests that hydroxyl radicals play an important role in the degradation of CR dye.This was indicated by a valence band edge potential value of HKUST-1 (3.37 eV) that was more positive than the H 2 O/•OH potential (2.34 eV).The conversion of O 2 to •O 2− was made simpler by the higher valence band edge potential value of HKUST-1.The amount of •OH required for oxidation decreased when IPA was added with the intention of collecting •OH, reducing the photocatalytic degradation performance.

Table 2
Parameters of pseudo first order and pseudo second order kinetic of the CR photocatalytic removal.

Adsorbent
Pseudo First Order Pseudo Second Order Qe cal (mg/g) R 2 Qe cal (mg/g) k 1 (min

Adsorption-photocatalytic mechanism
As shown in Fig. 8a, the possible mechanisms between GO/ZIF-8/ HKUST-1 and CR can be studied as follows: the porous structure and large surface area of GO/ZIF-8/HKUST-1 in the initial step allows CR dye to be transported from the solution to the surface and interior of GO/ ZIF-8/HKUST-1.Pore filling has a significant impact.CR dye can easily pass through some of the larger pores in the composite.The pores of GO/ ZIF-8/HKUST-1 are larger than the particle size of CR dye (1.90, 0.71.15,0.48 nm) based on the calculations using the Avogadro application.Therefore, CR dye can easily fill the GO/ZIF-8/HKUST-1 pores.In the second step, the active sites on GO/ZIF-8/HKUST-1 contribute to the binding of CR dye through π-π stacking, electrostatic interactions, and hydrogen bonding.The addition of GO increased the amount of π-π stacking between GO/ZIF-8/HKUST-1 and CR dye, which is correlated with the abundance of π bonding (benzene ring) in GO (Chen et al., 2019).
Based on the results that were obtained in this study, a possible photocatalytic mechanism was proposed to verify the higher photocatalytic performance of the GO(10)/ZIF-8/HKUST composite compared to the other composites, particularly compared to pristine ZIF-8 and HKUST-1.The main barrier impeding the efficiency of the photocatalyst is the high rate of charge carrier recombination and lower charge separation due to the charge carriers' short mean free paths.As a result, most of the charge carriers are lost to recombine before they can perform redox reactions on the semiconductor surface to produce radicals and degrade the pollutant (Fakhri and Bagheri, 2020).Under UV LED irradiation, photogenerated electron-hole pairs are generated in the valence band of HKUST-1 and ZIF-8.The electrons are then excited from the valence band to the conduction band, producing holes in the valence band (Fig. 8b).The use of GO as an electron mediator serves to easily transfer electrons from the conduction band of HKUST-1 to the valence band of ZIF-8, enabling the retention of the highly reducing electrons in the conduction band of ZIF-8 and the highly oxidizing holes in the valence band of HKUST-1 to produce significant amounts of hydroxyl and superoxide radicals.In the conduction band of HKUST-1, O 2 is easily reduced to •O 2− due to the reduction potential of the conduction band of ZIF-8 (− 1.137 eV) that is more negative than that of O 2 /•O 2− (− 0.33 eV).Meanwhile, in the valence band of HKUST-1, H 2 O is easily oxidized to •OH due to the reduction potential of the valence band of HKUST-1 (3.37 eV) that is more positive than that of H 2 O/•OH (2.34 eV).In addition, the presence of GO as an electron mediator serves to increase the charge and to prevent recombination of electrons in each semiconductor.The following describes the precise reaction pathway for the photocatalytic degradation of CR dye (Khan et al., 2020).

Reusability and stability
To confirm the stability of GO(10)/ZIF-8/HKUST-1 for the degradation of CR dye under UV-LED irradiation, a photodegradation experiment with three consecutive runs was carried out.It can be seen in Fig. 9b that GO(10)/ZIF-8/HKUST-1was successfully reused after the third reaction cycle and retained its composition and functional group.However, the efficiency of GO(10)/ZIF-8/HKUST-1 in the photodegradation of CR decreased after three consecutive cycles, which was indicated by the low transmittance in the FTIR spectra.The stability of GO(10)/ZIF-8/HKUST-1 was evaluated using FTIR characterization after the photocatalytic degradation of CR, which is shown in Fig. 9a.After photodegradation under UV LED irradiation, the FTIR spectra showed that the characteristic peak of GO(10)/ZIF-8/HKSUT-1 was maintained with a slight shift but still within the characteristic peak range, indicating the stability of GO(10)/ZIF-8/HKUST-1.The characteristic peaks of the Cu− O, C− O and C=O groups are 727, 1367 and 1640 cm − 1 , respectively, indicating the characteristic spectra of HKUST-1.Similar peaks were also observed after the photocatalytic degradation, shifting slightly to 722, 1357 and 1617 cm -1 .This indicates the interaction of the active site from GO(10)/ZIF-8/HKUST-1 with the CR molecule via HKUST-1 site.Furthermore, characteristic attenuation of the peaks at 3461, 1581, 1445, and 1176 cm − 1 can be observed, which occurred from the interaction between the composite and the -NH 2 , − N=N− , C=C, and − SO 3 groups in the CR molecule (Martak et al., 2021).

Conclusion
Novel ternary composites comprised of three different MOFs as photocatalysts, namely graphene oxide (GO), ZIF-8, and HKUST-1, with varying amounts of GO were successfully synthesized via the solvothermal method.The adsorption and photocatalytic degradation test results for the removal of CR dye showed that all the synthesized materials followed the pseudo-second order adsorption.The GO(10)/ZIF-8/HKUST-1 composite exhibited the best performance for the removal of CR dye (91.81%,50 mg/L CR in 60 min) by means of adsorption and photocatalysis compared to the other materials.The addition of GO R. Ediati et al. resulted in an increase in the number of active sites and also suppressed charge recombination, which are beneficial for the removal of CR dye.However, excessive amounts of GO can result in active sites being covered and in turn decrease the degradation performance of the composite.Therefore, the optimal loading amount of GO was found to be 10% in this study.Future research can be focused on developing composite comprised of MOFs that are enhanced through post-synthetic modification such as MOF-derived semiconductors.Furthermore, the economic advantage of using MOFs as a semiconductor precursor is that the use of additional chemicals is reduced in instances where metal doping is needed.Thus, extensive research can be carried on fabricating MOF-derived semiconductors and their application as photocatalysts in wastewater treatment in future works.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Table 3
Comparison of CR degradation onto different photocatalyst.