Simultaneous adsorption of malachite green, methyl orange, and rhodamine B with TiO2/macadamia nutshells-derived activated carbon composite

In this study, TiO2/activated carbon (TiO2/MAC) composite was synthesized from activated carbon prepared from macadamia nutshells and a water-soluble titanium complex, and it was used to simultaneously adsorb malachite green (MG), methyl orange (MO), and rhodamine B (RhB) from aqueous solutions. The kinetic studies show that the pseudo-second-order kinetic model describes the adsorption experimental data the best. The equilibrium data of the trinary system were analyzed via the ideal adsorption solution theory (IAST) and the Langmuir and P-factor-Langmuir extended models that combine the three single-component isotherms (Langmuir, Freundlich, and Sips). The AICs (Akaike Information Criterion) values indicate that IAST incorporating the Langmuir model is the most suitable to describe the removal of the dyes in the trinary solution. The TiO2/MAC composite exhibits a high dye adsorption capacity compared with those of the published adsorbents. The thermodynamic analysis reveals that the adsorption is spontaneous and endothermic. The high adsorption capacity and the recyclability through photocatalytic self-cleaning show that TiO2/MAC can be utilized as a sustainable alternative for the simultaneous elimination of textile dyes from effluents.


Introduction
Organic dyes are dangerous environmental pollutants commonly found in wastewater discharged from numerous industries, including textiles, dyeing, printing, pharmaceutical and food processing, cosmetics, and papers [1][2][3][4].Because of their highly stable molecular structures, these dyes are resistant to light and oxidants and are non-biodegradable [5].Several studies have found that dyes are highly toxic substances containing carcinogenic compounds that can cause permanent health damage to both animals and humans [5][6][7].
According to Sudova et al, malachite green, C 23 H 25 ClN 2 (MG), a dangerous carcinogenic agent, has been shown to have serious effects on the human reproductive system and the immune system [8].Methyl Orange, C 14 H 14 N 3 NaO 3 S (MO), is a water-soluble anionic dye and one of the most common dyes used in the textile industry [9].Rhodamine B, C 28 H 31 ClN 2 O 3 (RhB), is also a dye that can cause cancer and neurotoxins, as well as respiratory diseases, kidney failure, liver and thyroid damage, skin irritation, gastrointestinal tract infections, and eye infections [9][10][11].These dyes have numerous negative impacts on the ecosystem, aquatic flora and fauna, and the quality of human life.Therefore, completely removing organic dyes from wastewater is a challenging task, and the issue has increasingly become a topic of interest to scientists.Numerous technologies have been developed for eliminating textile dyes from effluents, such as photochemical and biological degradation, coagulation, flotation, chemical oxidation, reverse osmosis, and adsorption [12][13][14][15][16].Among these methods, adsorption is widely used to remove the colour from wastewater because of its simplicity, high efficiency, recyclability, easy adsorbent recycling, and low residue production.
Activated carbon is known as one of the excellent adsorbents for different types of waste.Particularly, it is widely used in dye wastewater treatment [17].At present, macadamia trees are planted in numerous places in Vietnam, stretching from the North to the South, particularly in the Central Highlands.According to studies, one ton of macadamia nuts produces approximately 0.7-0.77tons of nutshell [18].The macadamia nutshell has a high percentage of carbon and a low ash content, and it is a cheap and readily available agricultural by-product.The macadamia nutshells are primarily discarded, with only a few percentages used as fuel [19].This indicates that macadamia nutshell is an available source of raw materials for producing activated carbon.In adsorption, the low cost of the adsorbent and its high capacity and simple recycling are very advantageous.Although active carbon is a common adsorbent used to remove coloured products, its regeneration requires organic solvents, leading to secondary and more dangerous pollutants.
Titanium dioxide (TiO 2 ) is a typical semiconductor and is able to oxidise organic compounds completely.Hence, the combination of TiO 2 with activated carbon is expected to obtain a type of self-cleaning adsorbent.However, controlling the synthesis of TiO 2 -based materials becomes more difficult if TiO 2 is in the form of alkoxide or TiCl 4 because these precursors are easily hydrolysable in a humid environment.It is found that using alternative precursors, such as a water-soluble titanium complex, is an effective way to address this shortcoming.
Most previous investigations on dye adsorption were carried out in single-component systems [4,10], or binary systems [20].In practice, different dyes coexist in the effluent systems.Thus, the removal of colorants becomes a major challenge compared with single or binary component adsorption because of the increased complexity brought about by the increasing number of parameters involved in the treatment.
Although the ideal adsorbed solution theory (IAST) resulted from pure-component adsorption equations such as those of Langmuir, Freudlich, and Sips models of single-component isotherms has often been used for predicting the gas mixture adsorption equilibrium [21], its application to trinary or more-component dye adsorption is still limited.
In a previous paper, we reported the adsorption of a three-component system containing methylene blue, methyl orange, and methyl red from an aqueous solution using TiO 2 /activated carbon prepared from rice husks [22].Expanding this idea, we conducted the adsorption of MG, MO, and RhB dyes from aqueous solutions in a three-component system with TiO 2 /MAC as an adsorbent.TiO 2 /MAC was prepared from a water-soluble titanium complex and activated carbon derived from macadamia nutshells.Several isotherm models were applied to interpret the equilibrium data in the MG, MO and RhB trinary system.The OriginPro19/ deconvolution Software-assisted UV-vis spectroscopy was used to develop the simultaneous analysis of MG, MO and RhB.The Akaike information criterion (AIC) was applied to compare the multi-parameter isotherm models instead of the determination coefficient (R 2 ).

Preparation of activated carbon from macadamia nutshells (MAC)
The raw materials were washed with deionized water and dried in an oven at 110 °C for 24 h to remove moisture.The dried shells were crushed to a size of 5 mm and calcined at 350 °C for 2 h to obtain biochar.Then, the biochar was activated with sodium hydroxide at a mass ratio of 1:1 for 3 h.After being dried, the product was calcined in the absence of oxygen at 350, 500, and 700 °C for 2 h, followed by washing with distilled water to completely remove the excess NaOH.Finally, the solid was dried at 100 °C for 6 h to obtain MAC.The sample was denoted as MAC350, MAC500 and MAC700, in which the number part stands for the calcination temperature.
TiO 2 complex was prepared in two steps.First, 2 g of TiO 2 powder and 100 ml of a 15 M NaOH solution were exposed to ultrasound for 30 min in an autoclave, followed by thermal hydrolysis at 130 °C for 10 h.The white solids were separated and washed several times with 0.1 M HCl and distilled water to remove the excess NaOH and HCl.Second, 1.75 g of the dry product was dissolved in 245 ml of H 2 O 2 at 70 °C under stirring for one hour to form a yellow, transparent peroxo-hydroxo-titanium complex solution (1.75 g TiO 2 /245 ml).
The detailed experimental design is listed in table 1.

Instruments
The material was characterised via the XRD patterns obtained on a D8 Advanced Bruker with Cu-Kα radiation.Its textural properties were studied by using the nitrogen adsorption/desorption isotherms recorded on a Micromeriti CMS 2020 volumetric adsorption analyzer system.The morphology of the samples was studied from the scanning electron microscopy (SEM) images taken with an SEM-JEOL-JSM 5410 lV.The elemental dispersion was detected by EDX mapping (JSM-IT200 InTouchScope).Raman scattering measurements were conducted on a Raman microscope XPLORATMPLUS, Horiba, with a diode laser (785 nm).Ultraviolet-visible  diffuse reflection spectrographs (UV-vis DRS) were obtained on a UV-2600 Shimadzu, and the UV-vis spectroscopy study was carried out with a Lambda 25 Spectrophotometer-Perkin Elmer.

Kinetic study
Adsorption kinetics was studied by performing an adsorption process with one litre of the mixed dye solution containing MG, MO, and RhB with different concentrations and 0.1 g of TiO 2 /MAC in a two-litre beaker.The samples were placed in a bath at 300 K and with a stirrer rotating at 150 rpm.Five millilitres of the solution were drawn at predetermined intervals and centrifuged at 4400 rpm.The dye concentration in the supernatant was determined by using UV-visible spectroscopy at a wavelength range of 350-750 nm (UV-vis 6850 JENWAY).

Effect of pH
The pH influence on the adsorption process was studied by varying the solution's initial pH from 3 to 11.The pH was adjusted with a 0.01 M HCl or 0.01 M NaOH solution.The initial dye concentrations were fixed at 0.109, 0.122, and 0.125 mM for MG, MO and RhB, respectively, with a TiO 2 /MAC dosage of 0.01 g/200 ml.The samples were shaken in the dark for four hours to ensure adsorption equilibrium.The solid was removed by centrifuging, and the dye concentration in the supernatant was determined.

Isotherm studies
In the single-component system, the isothermal adsorption of each dye was performed in a batch at 290, 300, 310, and 320 K.An amount of MAC from 0.01 to 0.11 g (0.010, 0.024, 0.037, 0.049, 0.062, 0.075, 0.088, and 0.11) was added to a series of flasks containing 200 ml of MG (0.109 mM), MO (0.122 mM), or RhB (0.125 mM).All 24 flasks were shaken in the dark for 24 h to ensure adsorption/desorption equilibrium.The dye concentration in the solutions before and after adsorption was determined with a UV-vis spectrophotometer at maximum absorbance (λ = 446 nm for MO, 560 nm for RhB, and 617 nm for MG).In the ternary system, the adsorption isotherm was investigated as in the single-component system.The only difference is that all three dyes (MG (0.109 mM), MO (0.122 mM), and RhB (0.125 mM) are in the same system.

Regeneration of adsorbent
The recyclability was investigated via the adsorbent's photocatalytic self-cleaning.The adsorption was conducted in a 0.6 L solution containing MG, MO, and RhB at a concentration of 0.109, 0.122, and 0.125 mM, respectively, and 0.2 g of adsorbent.The used adsorbent was separated and desorbed by stirring under halogen lamp illumination for 3 h, followed by washing with deionized water and drying at 100 °C for subsequent use.The kinetics and isotherms models in this work are presented in detail in the supplementary data.

Characteristics of adsorbents
Figures 1(a)-(c) present the SEM images of MAC calcinated at 350, 500, and 700 °C, respectively.It seems that the higher the calcination temperature, the more likely it is that the porous system collapses.The SEM image of MAC calcinated at less than 500 °C displays a typical layered structure of carbon with numerous mesopores and macropores, indicating the development of a pore structure after activation.However, the MAC calcined at 700 °C shows blocks, and the porous system disappears.Figure 1(d) presents the XRD patterns of the MAC calcined at 350, 500, and 700 °C.At 350 °C, the diffractions are observed nearly at 2θ of 24.2 and 45.3°, assigned to the (002) and (101) reflections of the graphite structure [19].At higher calcinating temperatures, the broad diffraction peak at 10-30°can be assigned to the amorphous carbon structures.FT-IR of the obtained MACs are similar in which mainly stretching vibrations of hydroxyl group (-OH) may be observed at 3440 cm -1 and 1630 cm −1 [23] (figure 1(e)).The textural properties of the obtained MAC were studied from the nitrogen adsorption/desorption isotherms (figure 1(f)).The specific surface area of the biochar decreases with calcination temperature.The sample calcined at 350 °C has the largest specific area of 620 m 2 •g -1 , while the ones calcined at 500 and 700 °C have a specific surface area of 354 and 109 m 2 •g -1 .As calcined temperature increases, both the mesoporous area (S mes ) and microporous area (S mic ) also decrease significantly, possibly because the porous system collapses (table 2).Hence, the MAC calcined at 350 °C was used to prepare TiO 2 /MAC.
Figure 2(a) presents the XRD patterns of TiO 2 and TiO 2 /MAC.The anatase phase according to the JCPDS Card no.21-1272 is formed when the peroxo-titanium complexes are pyrolysed.On the XRD patterns of TiO 2 /MAC, the characteristic peaks of anatase are observed, but their magnitude decreases with the MAC amount, indicating the formation of TiO 2 /MAC composites.The nitrogen adsorption/desorption isotherms of the obtained materials are presented in figure 2(b).The specific surface area of TiO 2 prepared from peroxotitanium complexes of around 124 m 2 •g -1 is relatively high compared with that of TiO 2 prepared with other methods [24].By contrast, the specific surface area of MAC is smaller than that in the published literature.In fact, the surface area of biochar depends on the source of biomass and the manufacturing conditions [23].The specific surface area of TiO 2 /MAC(1/5), TiO 2 /MAC(2/5), TiO 2 /MAC(5/5), and TiO 2 /MAC(7.5/5),calculated from the BET model, is 412, 363, 349, and 119 m 2 •g -1 , respectively (figure 1(b) and table S1).It is obvious that the surface area decreases steadily with the amount of TiO 2 , and this fall could be explained by the fact that very fine particles of TiO 2 partially block the pores of MAC.
Figures 3(a) and (b) present the structural information from the Raman spectroscopic data.In figure 3(a), four Raman active modes of anatase TiO 2 with symmetries E g , B 1g , A 1g , and E g observed at 144, 398, 517, and 638 cm −1 , respectively, are assigned to the characteristic vibrational frequencies of the pure anatase TiO 2 phase [25].The inset presents two broad peaks at about 1352 and 1589 cm −1 on the MAC spectrum, attributed to the D-band and the G-band that correspond to the characteristics of graphite structure [26].The former is

Activated carbon
Surface area BET (m 2 •g -1 ) The porous volume MAC  Figure 4(a) shows that TiO 2 consists of fine particles of about 10 nm in size, agglomerating into larger particles in the peroxo-titanium complex.Figures 4(b)-(e) show that the TiO 2 /MAC surface is covered by TiO 2 particles that disperse on the MAC surface with a layered structure of numerous carbon-containing mesopores and macropores.These pores increase light absorption and, thus, improve the photocatalytic activity of the material.
The elementary dispersion in the TiO 2 /MAC composite was examined by using EDX energy-dispersive spectroscopy (figure 5).The spectra reveal the presence of a large amount of the C, Ti, and O elements.They are the main components of the composite.A small amount of Na (1.3%) is also present, possibly because it remains after washing.This demonstrates that TiO 2 /MAC of relatively high purity was successfully synthesized.
UV-vis diffuse reflectance spectroscopy spectra were used to determine the band-gap energy of the composites (figure 6(a)).As shown in the figure, TiO 2 /MACs reduce absorption in the ultraviolet region.The enhanced absorption of TiO 2 /MAC in the visible region from 400 to 650 nm is possibly attributed to the interaction of MAC and TiO 2 .The band-gap energy is calculated from Tauc's equation [28] , where α is the absorption coefficient; A is the constant; h × ν is the photon energy, and E g the optical band gap energy of the material.Figure 6(b) shows the plot of (α × h × ν) 2 versus the photon energy (h × ν) for the TiO 2 , MAC, and TiO 2 /MAC samples.Extrapolating the linear region of these plots to the abscissa provides the energy of the optical band gap of the obtained material.It was found that the band-gap energy of  TiO 2 /MAC(1/5), TiO 2 /MAC (2/5), TiO 2 /MAC (5/5), TiO 2 /MAC (7.5/5), and TiO 2 , calculated according to the Tauc's equation, is 1.7, 1.8, 2.0, 2.8, and 3.1 eV, respectively (figure 6(b)).
The combination of TiO 2 and MAC shows that the absorption region of TiO 2 /MAC has a shift towards the red-light region compared with TiO 2 .This shift indicates that adding MAC to TiO 2 enhances visible-light absorption of the resulting TiO 2 /MAC composite.The results show that the large contact surface between activated carbon and TiO 2 can narrow the band-gap energy and thus significantly enhance the absorption efficiency of TiO 2 in the visible-light region.

Simultaneous determination of MG, MO and RhB with UV-vis spectroscopy
The maximum wavelength used for MG, MO, and RhB in UV-vis spectroscopic measurements is 617, 464, and 560 nm, respectively (Figure S1(a)).However, the spectra of the three dyes overlap, resulting in a bigger error in the simultaneous determination of the dyes' concentration.In order to solve this problem, the absorption band is deconvulated by using OriginPro19/deconvolution Software, as illustrated in figure S1(b).The divided spectra were employed to determine the dyes' concentration with the standard addition method.It was found that the absorbance of each component increases linearly with the increase in the dyes' concentration.The linear range is 0.0055-0.1233mM for MG, 0.0104-0.1357mM for MO, and 0.0104-0.1357mM for RhB.The limit of detection (LOD) for MG, RhB, and MO is 0.0058, 0.0086, and 0.012 mM, respectively, which is suitable for detecting these dyes in the adsorption study.The recovery is 90%-96% for MG, 90%-93% for MO, and  90%-93% for RhB.These acceptable recovery values manifest that MG, MR, and RhB could be determined simultaneously with the proposed UV-vis method (Table S2).

Adsorption capacity of TiO 2 , MAC, and TiO 2 /MACs
The equilibrium adsorption capacities of MAC, TiO 2 , and TiO 2 /MACs composites and their reuse in the simultaneous removal of MG, MO, and RhB are depicted in figure 7.Although activated carbon presents the highest adsorption capacity for the studied dyes, the adsorption capacities dropped significantly to 0.9%-4.35%compared with the initial ones because the adsorbent cannot be regenerated.Pure TiO 2 also shows a really low adsorption capacity.For the first adsorption, the composites tend to exhibit decreased adsorption capacity with an increasing amount of TiO 2 .After recycling, the adsorption capacity increases with an increase in the TiO 2 amount and peaks for TiO 2 /MAC(2/5).The adsorption capacity of TiO 2 /MAC(2/5) is lower than that of TiO 2 /MAC(1/5), but TiO 2 /MAC(2/5) exhibits the highest recyclable capacity or the best self-cleaning catalyst; hence the TiO 2 /MAC(2/5) composite was selected for further investigation.

Effect of pH on the adsorption of TiO 2 /MAC
The pH of the solution can either enhance or hinder dye adsorption because it can affect the surface charge of the adsorbent and the adsorbate species in the solution.In the ternary system containing MG, MO, and RhB, the pH of the solution significantly affects the adsorption capacity (figure 8).It is obvious that RhB and MG dyes tend to have an increased adsorption capacity as pH rises to a certain extent, while MO exhibits a decrease.The RhB removal from aqueous solutions increases significantly with pH and peaks at around 7.5; then, it decreases slightly at higher pHs.This decrease may be due to the repulsion between the negatively charged TiO 2 /MAC surface (the pH at the point of zero charge (pH PZC ) is around 6.1, as shown in figure 8(a)) and the anionic form of RhB in the base medium (pK a = 3.7).The adsorption efficiency of this dye is also low at pHs 2-4.This low efficiency can be related to the electrostatic repulsion between the positively charged surface of the material and the positive charge of RhB at low pHs (below PZC).When pH > 8, the slight change in RhB adsorption capacity may be due to the saturation adsorption on the TiO 2 /MAC surface.As for MO, its adsorption on TiO 2 /MAC decreases with increasing pH.The rapid decrease occurs at lower pHs (pH PZC < 6.1), and after that, it decreases at a lower rate (up to pH 11).The higher adsorption capacity of MO at lower pHs can be attributed to the dye's opposite electrostatic interaction with the positively charged surface of the material at pHs less than pH PZC .At pHs greater than pH PZC , the repulsion of the negatively charged surface and the MO anion further reduces the equilibrium adsorption capacity.Similarly, the MG cations at pHs lower than 7 are attracted to the gradually decreasing positive surface charge, leading to an increase in q e at pHs less than pH PZC .In the base medium (pH > 6.1), both the adsorbate and adsorbent (TiO 2 /MAC) are negatively charged, leading to a rapid decrease in adsorption capacity.The mechanism of simultaneous adsorption of MG, MO, and RhB on TiO 2 /MAC becomes more complicated because the adsorption process can be affected by the interactions between each dye and the adsorbent in different ways.These interactions could be supportive or competitive, resulting in an increase/decrease in the adsorption efficiency of the components.It is possible that π-π interactions could be created between the aromatic rings of the RhB, MG, and MO components with the graphitic rings in TiO 2 /MAC, and hydrogen bonds between RhB and TiO 2 /MAC and between MG and TiO 2 /MAC can also form.Furthermore, electrostatic interactions between RhB and MG cations and MO anions can take place in the system, affecting the adsorption efficiency of these dyes during adsorption.A possible interaction between the dyes and TiO 2 /MAC is illustrated in figure 9.

Kinetic studies
To find the necessary time for the MG, MO, and RhB adsorption onto TiO 2 /MAC to reach equilibrium, we study the adsorption kinetics in a trinary system at various initial concentrations with 0.1 g of adsorbent at pHs 5-5.5. Figure 10 reveals that the adsorption is rapid in the first 40 min, followed by a gradual decrease in the rate as adsorption progresses until the equilibrium is reached after around 100 to 140 min.
As shown in figure 10 and table 3, an increase in the initial dye concentration results in an increased dye uptake.The equilibrium adsorption capacity of MG increases from 0.308 to 0.531 mmol•g -1 with the concentration rising from 0.0548 to 0.109 mM; the equilibrium adsorption capacity of MO rises from 0.179 to 0.325 mmol•g -1 with the concentration increasing from 0.061 to 0.122 mM.For RhB, the equilibrium adsorption capacity increases from 0.2040 to 0.3270 mmol•g -1 with an increase in the initial concentrations from 0.0626 to 0.1250 mM.This increase is explained by the fact that a higher initial dye concentration increases the mass transfer driving force, leading to a higher adsorption efficiency.
The data in table 3 show that the determination coefficient (R 2 ) values calculated according to the second-order kinetics of all dyes are substantially higher than those of the first-order kinetics at all three investigated dye concentrations.Furthermore, the adsorption capacity values calculated according to the pseudo-second-order kinetic model are also closer to the experimental values.Therefore, it can be concluded that the adsorption of MG, RhB, and MO onto TiO 2 /MAC in a three-component system follows the pseudo-second-order kinetic model.

Single component system
The adsorption isotherms in single-component systems were studied at 290, 300, 310, and 320 K, and the experimental data were analyzed by using the Langmuir, Freundlich, and Sips isotherm models (Figures S2, S3,  S4).The sum of squared deviations, SSE S , is described in equation (1)   where y exp is the experimental response, and y est is the response from the model.The coefficient of determination (R 2 ) is widely used to decide which model fits the data best.It is well-known that the closer the R 2 value to 1, the more compatible the model is.However, the model with more parameters always fits the experimental data better than the model with fewer parameters.Hence, R 2 could not be employed to compare the models with different parameters (The Langmuir and Freundlich isotherm models with two parameters and the Sips isotherm model with three parameters), the Akaike's information criterion (AIC) is often used to compare two models with different parameters to decide which one is better.The Akaike's information criterion estimates how well the data fit each model.The model with the lowest AIC s value is considered the most likely correct [27].The Akaike's information criterion is calculated according to equation (2) where N is the number of data points; N p is the number of parameters fit by the regression.The model's parameters are obtained with the least square method by minimizing the value of SSE S by using the Solver function in Microsoft Excel.
According to table 4, there are twelve AIC s values for MG, MO, and RhB at four temperatures (290-310 K).Therefore, it is difficult to estimate how small is an AIC s value for each model.To simplify the comparison of the AICs values, we use the sum of AIC s of the three dyes for each model.The order is as follows: Langmuir (-157.9)< Freundlich (-151.9)< Sips (-134.7).This indicates that the experimental data best fit the Langmuir model because its AIC s are the lowest at -157.9.Although the sum of AIC s of the Freundlich isotherms is not the smallest but is closer to that of the Langmuir model, their similarity (-152) shows that the compatibility level of these two models is relatively high.This means this is monolayer adsorption with a heterogeneous surface on the adsorbents.The maximum adsorption capacity of TiO 2 /MAC materials for the studied dyes tends to increase with the increase in temperature from 290 to 310 K (q e,MG is from 0.303 to 0.332 mmol•g -1 ; q e,RhB from 0.205 to 0.221 mmol•g -1 ; q e,MO from 0.287 to 0.302 mmol•g -1 ).It proves that MG, MO, and RhB adsorption favour thermodynamically in the studied temperature range.

Trinary system
Although well-known isotherms models such as Langmuir and Freunlich models are widely used for singlecomponent systems, they are unsuitable for multi-component systems.Therefore, other models have been developed for multi-component systems [5,25,[28][29][30].
Five isotherm models are utilized for evaluating equilibrium data in trinary systems, namely Extended Langmuir (EL), Langmuir model with P-factor (P-factor-L) and three IAST Langmuir, Freundlich, and Sips isotherm models (IAST-L, IAST-F, and IAST-S, respectively) (see Supporting Information).The results are presented in table 5.
As in the single-component system, the total AIC values of the models of the trinary system are selected to compare the goodness-of-fit of the model.The IAST-L (-174.07) is the lowest, indicating that this model best fits the experimental data of the three dyes.The compatibility of the isotherm models decreases as follows: IAST-L (-174.07)> P-factor-L (-170.73)> IAST-F (-163.49)> IAST-S (-138.98)> EL (-124.16).In these systems, MO exhibits the strongest goodness-of-fit with the models, while RhB always shows the weakest fit with the models in ternary systems.The maximum adsorption capacity of MG and MO in the ternary system, according Table 3. Parameters of pseudo-first and pseudo-second-order kinetic models for simultaneous adsorption of MG, MO, and RhB onto TiO 2 /MAC at 300 K.

Pseudo-first-order model
Pseudo-second-order model Con.(×10 2 mM) q e,Exp (mmol•g -1 ) Langmuir isotherm model   to the IAST-L model, is much lower than that in the single-component system, whereas RhB has a significant increase in adsorption efficiency in the trinary system compared with the single-component system at all investigated temperatures.This can be explained by the competition between the components/adsorption sites on the adsorbent in a multi-system.To estimate the competitive influence of the components on the adsorption efficiency of these three dyes, we use the P-factor model according to Langmuir, as presented in the experimental section.For MG, MO, and RhB, the P-factor can be set as follows: P 1.27 1; This proves that the adsorption of MG and MO has a competitive interaction that makes the adsorption process hindered by other components in the system, so the adsorption efficiency decreases.In which the adsorption efficiency of MO (an anionic dye) in the three-component system is much lower than that of MG (a cationic dye), possibly because of the greater influence of the adsorption affinity of the hindered TiO 2 /MAC composite surface and the interaction and competition between the MG/MO components that occupy the adsorption sites on the surface as well as the interaction between the MO anion and the MG and RhB cations.By contrast, the adsorption of RhB (a cationic dye) in a ternary system has < indicating that there is a supportive interaction in the system, causing the adsorption capacity of RhB to increase.Unfortunately, the adsorption information on the trinary system of MG, MO, and RhB is not found in the literature; therefore, only the single-component system is referenced for comparison.Table 6 presents the adsorption capacity of some adsorbents based on the Langmuir model.We can see that the present TiO 2 /MAC exhibits excellent adsorption for three dyes compared with those reported in the literature.

Adsorption thermodynamics
The thermodynamic parameters of MG, MO, and RhB adsorption onto TiO 2 /MAC in the trinary system were investigated at 290, 300, 310, and 320 K.As shown in figure S5, the equilibrium adsorption capacity, q e , of all dyes tends to increase with temperature, indicating that the process is endothermic.The pseudo-second-order kinetic model fits the kinetic data well in the temperature range of 290-320 K because this model has a higher R 2 than the pseudo-first-order kinetic model.Hence, the adsorption rate constant, k 2 , is used to calculate the thermodynamic parameters (Table S3).The standard Gibbs free energy, ΔG°, is more negative when the temperature is higher, suggesting the adsorption of MG, MO, and RhB on TiO 2 /MAC is spontaneous.A positive value of ΔH°(4.836kJ•mol -1 for MG, 8.1838 kJ•mol -1 for MO, and 10.624 kJ•mol -1 for RhB) suggests that the adsorption process in the system is endothermic (table 7).This means that the total energy absorbed by the system for bond-breaking is greater than the total energy released due to bond-forming [43].The positive value of ΔS°suggests the increase in randomness in the system as the dye molecules get adsorbed onto TiO 2 /MAC.By plotting lnk 2 versus 1/T (figure 11 and table S3), we can calculate the activation energy of dye adsorption according to k 2 .The values of activation energy are 6.94, 10.66, and 9.1 kJ•mol -1 for MG, MO, and RhB, respectively.The magnitude of activation energy helps to decide whether the adsorption is a physical or chemical process.Low activation energies (5-50 kJ•mol −1 ) suggest physical adsorption, while higher activation energies (60-800 kJ•mol −1 ) are evidence for chemical adsorption [44].This is because the pore diffusivity loosely depends on the temperature.Here, the diffusion process is related to the solute movement to the adsorbent external surface but not the diffusivity of the solute along micropore wall surfaces in the particles [45].The results obtained in the adsorption of the dyes onto TiO 2 /MAC manifest that the adsorption processes are physisorption.Hence, the affinity of MG, MO, and RhB to TiO 2 /MAC may be attributed to van der Waals forces and electrostatic attractions between the dyes and the surface of the adsorbents.A low value of E a generally implies that the adsorption process is controlled by diffusion, and a higher value represents a chemically controlled process.We can, therefore, conclude that the E a value obtained from the data suggests a diffusioncontrolled process, which is a physical step in adsorption.

Recyclability
In fact, the adsorbent reuse is always a crucial issue in adsorption.The most commonly utilized regeneration technique is calcination in the atmosphere without oxygen, ion exchange, or the use of organic solvents (e.g., methanol, ethanol, and benzene) to remove dyes from the adsorbent.These methods can lead to secondary pollutants and become more expensive.The TiO 2 /MAC composite is a photo-catalytically active material, so under visible-light radiation, it decomposes MG, MO, and RhB dyes without further intervention.The recyclability is presented in figure 12(a).It can be seen that the adsorption capacity of the material decreases after three self-cleaning cycles.The dye removal efficiency remains 58%-63% of its initial value (61% for MG, 58% for MO, and 63% for RhB).After four recycles, its XRD patterns retain the original characteristic peaks with a slight decrease in intensity (figure 12(b)).This means the TiO 2 /MAC composite is stable during adsorption and is suitable as a potential adsorbent.

Conclusion
TiO 2 /MAC composites with a large surface area and self-cleaning characteristics were successfully synthesized from peroxo-titanium complexes and activated carbon prepared from macadamia nutshells.The TiO 2 /MAC composites were utilized to efficiently remove MG, MO, and RhB dyes in a ternary adsorption system.The adsorption is exothermic and spontaneous and follows the Langmuir isotherms according to the IAST theory.The TiO 2 /MAC composites exhibit a high dye adsorption capacity.At relevant conditions of pHs = 5-5.5;and TiO 2 /MAC with mass ratio of 2/5 and 27 °C the high maximum monolayer adsorption capacities, q e , in the single/trinary solution for MG, RhB, and MO are 0.318/0.241,0.212/0.314,and 0.291/0.145mM•g -1 , respectively.TiO 2 /MAC with narrow bandgaps (1.7-2.8 eV) acts as a visible-driven photocatalyst for selfcleaning surfaces.The dye removal efficiency decreases only around 58%-63% of its initial value after four reuses.The TiO 2 /MAC composites are a promising adsorbent for the simultaneous removal of MG, MO, and RhB dyes from aqueous solutions over a wide concentration range.

Scheme 1 .
Scheme 1. Schematic diagram for TiO 2 /MAC synthesis and simultaneous adsorption of MG, MO and RhB.
associated with a disordered carbon structure and indicates polycrystalline graphite, and the latter is attributed to single crystalline graphite[27].The TiO 2 /MAC composite possesses the characteristic vibrational frequencies of both TiO 2 and activated carbon.The Raman data reconfirm the formation of the TiO 2 /MAC composite (figure 3(b)).

Figure 3 .
Figure 3. Raman spectra of MAC, TiO 2 ; the inset shows the Raman spectrum of TiO 2 and MAC (a) and TiO 2 /MAC (b).

Figure 9 .
Figure 9. Plausible mechanism for simultaneous removal of MG, MO, and RhB dye using TiO 2 /MAC.

Table 1 .
The experimental design of the synthetic materials.

Table 2 .
The textural properties of activated carbon activated at different temperatures.

Table 4 .
The parameters of isotherm models for MG, MO, and RhB adsorption onto TiO 2 /MAC in single component system at different temperatures.

Table 5 .
The parameters of Langmuir models and IAST models for simultaneous adsorption of MG, MO, and RhB onto TiO 2 /MAC at 300 K.

Table 6 .
Comparison of maximum monolayer adsorption capacity calculated by Langmuir model of TiO 2 /MAC with those of literature.