Preparation of Palm Oil Industry’s Biomass-Based Graphene Composite for the Adsorptive Removal of Methylene Blue

The ability of POME-based graphene shell composite (P-GSC), an adsorbent generated from oil palm wastes abundantly available in Malaysia such as POME and PKS, was examined in removing methylene blue (MB) dye by adsorption. Adsorption experiments, involving a batch column study and a batch equilibrium study, were conducted to investigate the efficiency of synthesized P-GSC from PKS as a base material in the removal of MB dye. The batch column study demonstrated that small-sized synthesized P-GSC from PKS as a base material could remove up to 98.5% for concentration. Therefore, the following batch equilibrium study was carried out on small-sized P-GSC only. Adsorption isotherms and kinetic isotherms were studied, from which the experimental data showed that the adsorption exhibited a good fit with the Freundlich model ( R 2 = 0.8923 ) and followed the pseudo-second order model ( R 2 > 0.98 ). FESEM, XPS, and XRD morphological and elemental analysis indicated the successful graphinization of POME on the P-GSC surface. The concept of deploying POME as the carbonaceous source to produce P-GSC, and then, deploying the resultant P-GSC as the adsorbent for MB dye removal has presented promising practical potential. Such cost-effective and environmentally friendly reuse of waste materials is envisioned to promote a ‘zero-waste industry.’


Introduction
Malaysia's palm oil industry has reported a remarkable and sustainable growth over the past 40 years in the global market as one of the world's largest oil palm producer; annual production in Malaysia has been reported to reach 15.4 million tonnes between 2016 and 2020 ( [1]).In this regard, each tonne of crude palm oil (CPO) production required 5 to 7.5 tonnes of water, of which 50% would end up as wastes in the form of palm oil mill effluent (POME) [2].POME is a brownish colloidal suspension containing water, oils, and total solids that pose environmental hazards [3,4].Among existing treatment methods of POME, the biological treatment with the ponding system is most commonly used because of low costs, though others are available such as membrane technology, coagulation-flocculation, and evaporation [5].However, the major drawback of this biological treatment system is that the continuous generation of POME poses serious environmental problems due to the handling and disposal that incur substantial operational costs [6,7].The palm oil industry in Malaysia has been identified as the factor contributing most substantially to river pollution throughout the country.The discharge of untreated effluent into the watercourse will compromise the aquatic ecology and source of drinking water, due to its high biochemical oxygen demand, chemical oxygen demand, oil and grease, total solids, and suspended solids [8,9].Despite its environmental hazards, POME is not only a renewable and inexpensive carbonaceous source but also a green alternative for industrial-scale production of graphene [10].Not only that, other biomass that been produced from palm oil industry such as oil palm kernel shells (PKS) also possess a major problem for disposal such as landfilling [11].
The water pollution due to industrilization is of great concern.Health problem is associated with water pollution leading to vomiting, nausea, diarrhea, abdominal pain, and methemoglobinemia [12,13].Various treatment techniques were applied for wastewater treatment such as adsorption, photodegradation, membrane filtration, and coagulation [14][15][16][17].The uses of adsorbents such as activated carbon (AC) is commonly known for adsorption in wastewater treatment [18].However, it is unsuitable for large-scale practice such as the palm oil industry due to its high operating cost and high regeneration cost [19].As an alternative, low-cost materials such as durian peel, rice husk, and nutshells have been produced as adsorbents.However, despite the reduced costs of such raw materials, they warrant preparation via high thermal treatments (700-900 °C) which likewise incur high costs [20,21].Thus, efforts have been underway to develop an adsorbent; the synthesis of which required the use of two wastes (POME and PKS) in producing POME-based graphene shell composite (P-GSC).This involved chemical activation in an adsorption treatment for the removal of methylene blue (MB) dye: this activation has been found to lower energy costs and also shorten the time needed, thus reducing the overall costs of production [22].
MB dye was used as an adsorbate in this study.A cationic dye used as a material for dyeing wood, cotton, and in textile industries, the MB dye has been one of the most significant environmental concerns in industrial wastewater [23].Some dyes used in textile industries such as reactive, acid, and basic dyes are considered toxic [24].Other dyes, while not classified as highly toxic, need to be treated before discharge, because some of their derivatives are nonbiodegradable and have carcinogenic properties or mutagenic efficacy, which could endanger humankind and aquatic ecology [25,26].
This paper is aimed at identifying the feasibility of synthesized P-GSC in MB dye removal and to study the mechanism of solute molecules of MB dye onto the synthesized P-GSC.The parameters affecting the adsorption such as the weight of P-GSC were investigated in a batch equilibrium study.The adsorption isotherms (Langmuir and Freundlich) and kinetic isotherms (pseudofirst-order and pseudosecondorder) were also investigated.The concept of deploying POME as the carbonaceous source to produce P-GSC, and then, deploying the resultant P-GSC as the adsorbent for MB dye removal has presented promising practical potential.This green technology can significantly contribute towards the body of knowledge on sustainability, economic growth, and mitigating environmental degradation; thus, the environmentally friendly reuse of waste materials is envisioned to promote a 'zero-waste industry.'

Experimental
2.1.Materials.PKS and raw POME were collected and used as raw materials for this study.PKS formed the base material in the synthesis of the adsorbent while raw POME was used as a carbonaceous source for coating of the base material.
Sulphuric acid (H 2 SO 4 ) (95-98% purity) supplied by R&M Chemicals, United Kingdom, was used to activate the synthesized adsorbent.MB dye powder (C 16 H 18 ClN 3 S2H 2 O) synthesized by Merck, Germany, was used in the preparation of the feed solution.
2.2.Preparation of PKS.Impurities from the PKS were first washed away with tap water [27].The PKS was then left to dry overnight in an oven at 110 °C before it was crushed with a universal cutting mill (FRISTCH, Germany), sieved, and separated into 3 different ranges of sizes: large (1.18-2.36mm), medium (0.60-1.18 mm), and small (0.30-0.60 mm).Upon completion of the heating cycle, the crucible dish was left to cool to 40 °C in the muffle furnace.The black material obtained after the heating cycle was termed "P-GSC" and was left to stand in a concentrated solution of H 2 SO 4 for 20 minutes to activate it.Excess H 2 SO 4 was then washed off with ultrapure (UP) water until the P-GSC reached pH 7. The P-GSC was then dried in an oven at 110 °C prior to analysis.The above steps were then repeated for the medium-sized and small-sized PSK particles.

Characterization of P-GSC.
Field emission scanning electron microscope (FESEM), X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD) characterizations were done on the synthesized P-GSC to determine its physiological behavior.Prior to the FESEM analysis, the outer surface of P-GSC was coated in a thin layer of platinum by the K550 sputter coater under a vacuum to reduce the charges on the P-GSC surface to ensure that contrastive images could be obtained.The analysis was then carried out with SUPRA 55VP-ZEISS (Merlin-Zeiss, Germany) to observe the surface morphology of the synthesized P-GSC.Following that, the P-GSC was examined under an electron microscope at a potential of 15 kV while the FESEM imaging was carried out at 100x, 500x, and 1.00 Kx magnifications.XPS analysis completed by AXIS ULTRA DLD (Kratos/   The batch equilibrium study determined the adsorption capacity of the synthesized P-GSC at the equilibrium stage. 2.6.1.Batch Column Study.The batch column study was carried out in a chromatography column held vertical by a retort stand.A 10 cm height of large-sized P-GSC adsorbent was packed into the column before 40 mL of 20 mg/L synthetic MB dye solution was allowed to flow through via the pull of gravity.The filtrate was collected at the bottom of the column.Concentration of the filtrate was measured by UV-vis spectrophotometer at room temperature.The concentrate of the filtrate was then measured with the UV-vis spectrophotometer at room temperature.The GENESYS 10 UV-vis spectrophotometer (Thermo Spectronic, USA) was set at 664 nm and calibrated with a blank and sample cell that contained an MB dye filtrate [28].The measured data were recorded in nanometres (nm).The percentage of MB ions removal (%) was calculated through Equation (1).The above procedures were then repeated for the medium-and small-sized P-GSC adsorbents.

Percentage of removal
where F is the initial concentration of synthetic MB dye solution and P is the final concentration of synthetic MB dye filtrate.

Batch Equilibrium Study
(1) Effect of P-GSC Weight on Adsorption Capacity.A set of 250 mL Erlenmeyer flasks were employed to carry out the batch equilibrium study.Different weights of large-sized P-GSC (5 g, 10 g, 15 g, and 20 g) were added individually into Erlenmeyer flasks, each filled with 250 mL of 20 mg/L synthetic MB dye solution.The samples were then put on the SHO-1D isothermal shaker (Wiseshake, Korea) and left to mix at of 230 rpm, 24 hours, room temperature.A sample was taken from each flask in the first 20 minutes of the study, 1 hour for the first 10 hours, and every 2 hours for the next 12 hours (stopped at 24 hours) of the study.The adsorption values of the samples were measured by the UV-vis spectrophotometer and their concentrations were read from the calibration curve.The equilibrium adsorption capacity of P-GSC, q e (mg/g), was calculated through Equation (2).The above procedure was then repeated for the medium-and small-sized particles.
where C o is the initial concentration of synthetic MB dye solution (mg/L), C e is the equilibrium concentration of syn-thetic MB dye solution (mg/L), V is the volume of MB dye solution (L), and W is the weight of P-GSC (g).(1) Langmuir Model.The Langmuir model assumes that no further adsorption will take place after the formation of the first monolayer sorption of solutes Once the solutes adsorb to an active sites, there is no new transmigration of adsorbate occurs on the same sites [29].The Langmuir model is expressed as per Equation (3) [30]: where C e is the concentration of synthetic MB dye solution at the equilibrium stage (mg/L), q e is the amount of MB ions adsorbed onto P-GSC at the equilibrium stage (mg/g), Q o is the adsorption capacity (mg/g), and b is the Langmuir constant that related to energy of adsorption (L/mg).
(2) Freundlich Model.The Freundlich model assumes that the adsorption of adsorbates on the surface is multilayered, where several adsorption energies are involved at different active sites [31].The Freundlich model is as expressed in where q e is the amount of MB ions adsorbed onto P-GSC at the equilibrium stage (mg/g), C e is the concentration of synthetic MB dye solution at the equilibrium stage (mg/L), k f is the adsorption capacity constant, and n is the adsorption intensity constant.

Kinetic Isotherms.
The optimum operating conditions were selected by investigating the mechanism of sorption and rate-controlling steps and studying the adsorbate uptake rate on adsorbents.These investigations were carried out via kinetic isotherms and can be described as the pseudofirst order model and psedosecond order model.
(1) Pseudofirst-order Model.The pseudofirst-order model assumes that the adsorption process occurs via the physical attachment of solutes on the adsorbent with a diffusion through a boundary [32,33].The pseudofirst-order model of Lagergren is expressed in Equation ( 5): where q e is the sorption capacity at equilibrium (mg/g), q t is the sorption capacity at time (mg/g), k 1 is the rate constant of pseudofirst-order model (min -1 ), and t is the time taken (min).
(2) Pseudosecond-order Model.The pseudosecond-order model indicates that the chemisorption process involving the exchange of adsorbate ions and adsorbent by valence forces is the limiting step for the adsorptive rate [34].The pseudosecond-order model is expressed in where q e is the sorption capacity at equilibrium (mg/g), q t is the sorption capacity at time (mg/g), k 2 is the overall rate constant for pseudosecond-order model (g mg -1 min -1 ), and t is the time taken (min).Adsorption Science & Technology

Results and Discussion
3.1.Characterization of P-GSC.The P-GSC surface was rough, irregular, and covered with a thin sheet of graphene layers (Figure 1).This observation coincides with the study conducted by Dubey et al. [35].The graphene layers with the hexagonal lattice structure had a large surface area that could have contributed to the synthesized P-GSC's high adsorption capability [36,37].The presence of heterogeneous pores on the surface of the P-GSC (Figure 1) could have been formed by the concentrated H 2 SO 4 during the activation of the P-GSC [38].This is a positive finding that could enhance the adsorption capability of the adsorbent.The elements of C and O peaks were detected at 284.5 eV and at 286.8 eV, respectively (Figure 2), thus indicating signs of an extended carbon backbone.This might be due to the oxygenation of carbon in the C-O bonds responsible for adsorption [39][40][41].The three peaks of O1s spectrum represent three different types of components and binding energies; the centered peak at 533.1 eV represents C-O, the second peak at 531.6 eV represents C=O, and the third peak at 534.3 eV represents O-C=O [42,43].Two peaks of XRD were detected at 21.53 °and 26.47 °(Figure 3).They exist in the range of 20-27 °, the successful range that allows for the anchoring of graphene on the adsorbent surface as per a similar result obtained from the deposition of graphene from sugar [35].The analysis from FESEM, XPS, and XRD thus prove that the synthesizing of P-GSC with POME as carbonaceous source, graphenized to graphene, and anchored onto the PKS surface.The high carbon content from POME and PKS indicates that the material can be synthesized as a good quality adsorbent for the adsorption process [44].

Batch Column Study.
The percentage of MB dye removal by different sizes of P-GSC is shown in Figure 4. Small-sized P-GSC adsorbents have high removal capabilities of 98.5% as compared to their medium-sized (97.2% removal) and large-sized (96.4% removal) counterparts.This indicates that the small-sized P-GSC adsorbents have good adsorption capabilities and can remove almost all the MB dye adsorbates.Small-sized PKS reduce the spaces between them when put together, trapping the suspended POME solids easily.This allows the color of the MB dye to clear up.The efficiency of decolorization and concentration removal by small-sized P-GSC adsorbents were due to the sufficient amount of hydroxyl radicals available on the adsorbents [18] The MB dye used in this study produced cations when dissolved in water, leading to interactions with the negatively charged, activated surface of P-GSC [45].This provided a large driving force to overcome the mass transfer resistance between the MB dye solutions and the solid phase of P-GSC [6].These functional groups aid the adsorption process [46].The carbons on the small-sized PKS P-GSC also have a mesopore structure that adsorbs molecules such as the MB dye.The clear color of the filtrate after the adsorption process in the batch column study is shown in Figure 5.  7 Adsorption Science & Technology batch equilibrium study was conducted only on the smallsized P-GSC (0.30-0.60 mm).The potential performance of adsorption was studied by manipulating the weights of P-GSC (5 g, 10 g, 15 g, and 20 g) while keeping the concentration of MB dye constant at 20.0 mg/L and running it for 24 hours.The trend of the graph indicates that a rapid uptake stage occurred at the initial stages of the process.This maybe due to available, empty active sites that bounded the MB dye ions on the P-GSC surface until an equilibrium, or assumed saturated has been reached [6].The adsorption rate slows down after equilibrium has been reached and will continue until the end of the process [47,48].
After 24 hours, the q e (mg/g) value at each time taken was calculated through Equation ( 2), and a graph of log q e vs. (C e /q e ) was plotted as per Figure 6.The equilibrium points for each weight of P-GSC were different (Figure 6).The time taken for P-GSC at 5 g, 10 g, 15 g, and 20 g was 20 hours, 14 hours, 8 hours, and 7 hours, respectively.At this stage, nearly all the solutions were on the verge of achieving adsorption equilibrium.Adsorption equilibrium refers to the point where the amount of MB dye adsorbed by the P-GSC remains constant.The 20 g weight of P-GSC indicated the best trend in MB dye removal.As the weight of P-GSC increases, the time taken for the adsorption process to reach equilibrium becomes shorter because the high amount of P-GSC provides more active sites for adsorption to occur [49].In addition, the activation by concentrated H 2 SO 4 , small particle size of P-GSC (0.30-0.60 mm), and high porosity structure of PKS contributed to the formation of more active sites on the P-GSC surface which allowed for more attachment of MB dye solutes [40].No more MB dye solutes were adsorbed onto the P-GSC past the equilibrium point thus indicating the nonavailability of active or adsorption sites.

Adsorption Isotherms
3.4.1.Equilibrium Isotherms.Equilibrium isotherms, Langmuir model and Freundlich model, were studied to deter-mine the best fit for the adsorption mechanism.A graph was plotted and linear regression analysis was applied to calculate the constants and for comparison of the correlation coefficient (R 2 ).The isotherm that had an R 2 value that approached 1 was selected as the best fit model.The equilibrium isotherms parameters for small-sized P-GSC are tabulated in Table    the first layer of solutes would have formed on the P-GSC surface during adsorption.A second layer will then form upon the first layer.This phenomenon is known as the heterogeneous adsorption process.Freundlich constants, k f and n are indicators of adsorption capacity and intensity (strength), respectively [53].
The n value for MB dye adsorption at n = 1:2129 indicates that the adsorption process is favorable [53], where the constant n range between 1 and 10 showing the impact of heterogeneity surface more significant [54,55].Active sites that have higher energy levels tend to form multilayers of MB dye adsorption due to strong chemical bonding (Freundlich isotherm), while sites with lower energy levels stimulate monolayer MB dye adsorption due to electrostatic forces [56].
3.4.2.Kinetic Isotherms.The parameters for the pseudofirstorder model and pseudosecond-order model are tabulated in Table 2.These isotherms were used to study the solute uptake rate in adsorption of MB dye for small-sized P-GSC.The correlation coefficient of the pseudosecond-order model for 5 g (R 2 = 0:9963), 10 g (R 2 = 0:9981), 15 g (R 2 = 0:9942), and 20 g (R 2 = 0:9899) were higher compared to the pseudofirst-order model (R 2 = 0:0208-0.1425).This indicates that the pseudosecond-order model was the most fitted kinetic isotherm for adsorption.The values of q e and k 2 calculated from the graph of (t/q t ) against t.q e indicate the sorption capacity at equilibrium (mg/g) with k 2 as the rate constant of the pseudosecond-order model.The pseudosecond-order model further predicted that the rate-controlling step in this study was chemical sorption (chemisorption) [57].

Conclusion
The removal of MB dye by a low-cost, natural adsorbent (P-GSC) synthesized from POME and PKS has been successful.Small-sized particles of P-GSC performed exceptionally with a performance of 98.5% in the batch column study.The high porosity structure of PKS, large surface area of the smallsized PKS, and the presence of graphene sheet layers on the P-GSC surface indicated a higher adsorption capacity when treating the MB dye solution.The adsorption capability of the small-sized P-GSC was further studied in the batch equilibrium study.It was found that the adsorption mechanism was well fitted to the Freundlich model and the pseudosecond-order model.The Freundlich model assumed that the uptake of adsorbates occurred by multilayer adsorption followed the pseudosecond-order by chemisorption process.These 'green waste' adsorbents of POME and PKS can be utilized and processed into a value-added product that contributes to the 'zero-waste' concept.This can thus eliminate issues with the environment and pollution.

2. 6 .
Adsorption Study.The adsorption study was divided into batch column and batch equilibrium studies.The batch column study evaluated the efficiency of MB ion removal from the synthetic MB dye solution by the synthesized P-GSC.

3. 3 .Figure 4 :
Figure 4: Percentage of MB dye removal by using different ranges of the particle size of P-GSC.

Figure 5 :
Figure 5: Comparison of colour between MB dye feed and filtrate water after adsorption process in batch column study.

Figure 6 :
Figure 6: Adsorption capacity graph with different weights of small-sized P-GSC (volume = 250 mL, concentration of MB dye solution = 20 mg/L, mixing speed = 230 rpm, and room temperature).

Table 2 :
Kinetic isotherms for different weights of small-sized P-GSC.

Table 1 :
[50][51][52]isotherms for small-sized P-GSC.that the Freundlich model was the best fit equilibrium isotherm.Since the Freundlich model assumes that different sites with several adsorption energies are involved between the adsorption of solutes from adsorbate solution (MB dye) onto the P-GSC surface[50][51][52],