Reuse of waste cooking oil (WCO) as diluent in green emulsion liquid membrane (GELM) for zinc extraction

Zinc (Zn) was identified as one of the most toxic heavy metals and often found contaminating the water sources as a result of inefficient treatment of industrial effluent. A green emulsion liquid membrane (GELM) was proposed in this study as a method to minimize the concentration of Zn ions in an aqueous solution. Instead of the common petroleum-based diluent, the emulsion is reformulated with untreated waste cooking oil (WCO) collected from the food industry as a sustainable and cheaper diluent. It also includes Bis(2-ethylhexyl) phosphate (D2EHPA) as a carrier, Span 80 as a surfactant, sulfuric acid (H2SO4) as an internal phase, and ZnSO4 solution as an external phase. Such formulation requires a thorough understanding of the oil characteristics as well as the interaction of the components in the membrane phase. The compatibility of WCO and D2EHPA, as well as the external phase pH, was confirmed via a liquid–liquid extraction (LLE) method. To obtain the best operating conditions for Zn extraction using GELM, the extraction time and speed, carrier, surfactant and internal phase concentrations, and W/O ratio were varied. 95.17% of Zn ions were removed under the following conditions; 0.001 M of H2SO4 in external phase, 700 rpm extraction speed for 10 min, 8 wt% of carrier and 4 wt% of surfactant concentrations, 1:4 of W/O ratio, and 1 M of internal phase concentration.


Introduction
Worldwide concerns on the heavy metal contaminant in water sources are reflected by the escalating numbers of works published by the scientific community as various contaminants such as heavy metals and pharmaceutical wastes were found in industrial effluents. Zinc (Zn) was identified as one of the most toxic metals by the World Health Organization (World Health Organization (WHO), 2011) and often found in effluents from industries such as mining, photo etching, metal plating, smelting, tanneries, petroleum refining, pesticides as well as paint, battery, and pigment manufacture (Asli et al. 2013;Fouad 2008). Zn was once reported to exist in the effluent of printed circuit board manufacturing units at a concentration up to 1700 mg/L (Sengupta et al. 2009). On top of that, areas with high anthropogenic activities such as residential areas and agriculture were suspected as the major factors for a considerable risk of Zn contamination, and it was predicted to be amplified into higher trophic levels due to the bioaccumulation factor (Razak et al. 2021). Globally, uncontrolled anthropogenic activities could lead to severe Zn contamination in the open water ecosystem (Obasi and Akudinobi 2020;Tonhá et al. 2020), consequently impacting the aquatic lives negatively (Lidman et al. 2020). The Department of Environment (DOE) Malaysia has set the permitted Zn discharge limit at 2 mg/L, while the threshold limit for Zn in water outlined by WHO was 4 mg/L (as zinc sulfate). Although Zn is essential in humans, its excessive intake will lead to system dysfunctions, causing growth and reproduction impairment. Acute Zn toxicosis would lead to manifestations of clinical signs including diarrhea, vomiting, jaundice, bloody urine, anemia, kidney failure, and liver failure (Oghenerobor Benjamin Akpor et al. 2014). Other than the direct discharge of effluent, this alarming situation is caused by an ineffective method of effluent treatment. Either situation constitutes a greater risk to the aquatic ecosystem (Sengupta et al. 2009).
Finding an effective, low-cost, and sustainable method to minimize the content of heavy metal to the permissible limit is a great challenge. Several methods have been reported to remove Zn from an aqueous solution, such as liquid-liquid extraction (Łukomska et al. 2020), solvent extraction (Mohammadzadeh et al. 2020), chemical precipitation (Chen et al. 2018, ion exchange (Abdelwahab et al. 2013), biosorption (Jagaba et al. 2020), adsorption (Astuti et al. 2020;Lim et al. 2012), electro-driven process (Tang et al. 2020), and pressure-driven membrane process (Jellouli Ennigrou et al. 2014). Unfortunately, each of these methods possesses drawbacks. On the other hand, liquid membrane is regarded as a green alternative (Kumar et al. 2018b) due to its low-cost feature in terms of energy utilization, compared to a pressure-driven membrane, and it works relatively fast, with less equipment and chemical consumptions (Björkegren et al. 2015;Othman et al. 2019;Raja Sulaiman et al. 2014). There are several configurations of liquid membrane reported to date, but the focus of this work is the emulsion type. It is a promising technique in the separation process of heavy metal contaminants due to its ability to combine extraction and stripping processes in a single step simultaneously. Emulsion liquid membrane (ELM) is able to recover a very low concentration of heavy metals from an aqueous solution, even at a concentration of lesser than 10 ppm (Ahmad et al. 2011;Raja Sulaiman et al. 2014). This is possible due to the high interfacial area per unit volume provided by the large number of small droplets formed in the double emulsion (Valenzuela et al. 2010). ELM is a multiple water-in-oil-in-water (W/O/W) emulsion system which was developed by homogenizing the membrane phase which consists of carrier and surfactant dissolved in a diluent, with the internal phase consists of stripping agent to form a primary water-in-oil (W/O) emulsion. The primary emulsion was then dispersed in the external phase containing heavy metals, resulting in the formation of W/O/W emulsion system where the mass transfer of solute from the external to the internal phase takes place (Ahmad et al. 2017;Othman et al. 2016). Continuous improvements of the configuration and formulation of the system were made since its introduction by Li (1968). Among them is the replacement of diluent with a more sustainable source.
The formulation of ELM is crucial as it dictates the stability and eventually the performance of the system. This includes the selection of carrier, strip agent in internal phase, surfactant, diluent, and preparation method (Shah Buddin et al. 2022). Diluent in ELM is a liquid or homogeneous mixture of liquids in which carrier and possible modifier may be dissolved to form a solvent. Since diluent is typically the largest portion in the emulsion, it must exhibit several general properties such as a high solute capacity for extraction of the desired species, inert toward the extraction system and miscible in organic but immiscible in the aqueous phase, low viscosity, nontoxic, noncorrosive, and low cost Shokri et al. 2020). Otherwise, the system will be unable to selectively remove the targeted contaminants. At the moment, petroleum-based diluents such as kerosene (Ma et al. 2016;Rosly et al. 2020a), heptane (Ahmad et al. 2019;Venkatesan and Begum 2008), and hexane (Dâas and Hamdaoui 2010) are widely used as the diluent. They are nonrenewable, flammable, naturally volatile, and most importantly not environmentally friendly and nonbiodegradable Shokri et al. 2020). The movement towards the utilization of a more environmentally benign vegetable oil as diluents such as corn oil (Ahmad et al. 2017(Ahmad et al. , 2015, rice bran oil (Kumar et al. 2018a(Kumar et al. , 2019a(Kumar et al. , 2019b, sunflower oil (Daraei et al. 2019), sesame oil (Kazemi et al. 2013), and palm oil has recently been studied (summarized in Table 1). The use of such food commodities would jeopardize our food sources where the competition with the food industry will exacerbate the existing global food crisis, turning it to be a less attractive option and should be avoided for the sake of global food security (Chang et al. 2015). Hence, waste vegetable oil is a viable option as a more sustainable organic solvent for wastewater treatment in ELM. Despite the numerous studies of applying vegetable oil as the diluent in ELM technique, so far the work from Shokri et al. (2020) had attempted to use WCO (sunflower oil) as diluent which has recorded the maximum extraction efficiency of dye methyl violet 2B as high as 99.1%. However, to the best of our knowledge, there is no study reported on the usage of WCO from palm oil as diluents in ELM formulation for heavy metal extraction.
Malaysia is among the world's largest producers of palm oil contributing part of the 85% of global palm oil supply, accounting for 34% of world vegetable oil consumption in 2018 (Kushairi et al. 2019). Thus, Malaysia with a current population of approximately 30 million people could be generating 0.12 Mt/y of waste cooking oil (Chuah et al. 2017); equivalent to 10.96 g/day per person. Although knowingly some of this waste vegetable oil is recycled into household items like soaps and candles or functioning biodiesel (Chuah et al. 2015), not all wastes are lucky enough to be recycled as a major part of it is illegally dumped into rivers and landfills. It worsens the environmental issues (Chang 2014); thus, finding a better way to reuse the waste could reduce the environmental issues. WCO is mainly converted into biodiesel. However, as green as it seems, the biodiesel industries are still grappling with the high free fatty acid (FFA) content in WCO that is constantly affecting the biodiesel quality (Athar and Zaidi 2020). Though with great potential, WCO as feedstock for biodiesel production is plagued by insecurities including higher FFA amount causing high production cost and low yield, vast changes in the oil properties as a waste, and other production constraints of long reaction times, high alcohol, and catalysis amount (Singh et al. 2021).
WCO is expected to serve as diluent similar to the fresh refined cooking oil in transporting solute across liquid membrane based on their analogous fatty acid composition and mean molecular weight. Nevertheless, higher free fatty acid, glycerol, and water contents in waste vegetable oil are the results of the oxidation; hydrolytic and cracking reaction that occurred during frying (Yaakob et al. 2013) might affect its efficiency in extracting solute from wastewater. Therefore, this work will look at the feasibility of using WCO as diluents in ELM system for the removal of Zn ions from an aqueous solution. The properties of the oil will be investigated, and the interaction between the components in the membrane phase will be looked at. As the emulsion is produced, the ideal condition to remove Zn from the aqueous solution will be identified.

Materials
The chemicals used for the ELM formulation and their functions are listed in Table 2. Deionized water was used for all solution preparations. The WCO obtained from a banana chips factory was filtered to separate physical impurities before its usage.

Liquid-liquid extraction (LLE)
LLE technique was used to determine the extraction equilibria, carrier compatibility, and the suitable pH for the extraction process. The process requires the mixing of an equal volume (25 mL) of WCO containing varying carrier concentration (0 to 20 wt%), and external phase where the pH of the latter was manipulated. The mixture of the phases was stirred at 400 rpm for a duration of time, and sample collection was made hourly until equilibrium was achieved   (Noah et al. 2018a). At each fixed interval, the content of the vessel was let to settle via gravity settling (Rosly et al. 2020b) before the treated external phase was collected to measure the concentration of Zn.

Water-oil-water (W/OW) preparation
Span 80 and D2EHPA were dissolved in the WCO and stirred at 300 rpm for 5 min to produce the membrane phase (Ahmad et al. 2014). Then, H 2 SO 4 was mixed with the membrane phase solution at a volume ratio of membrane to internal phase of 3:1, otherwise mentioned (Kusumastuti et al. 2018). The mixture was homogenized for 3 min at 3500 rpm using a homogenizer (WITEG HG-15D). The homogenizer probe was immersed at the interface of the membrane-internal phase which then produced water-oil (W/O) emulsion. Figure 1 shows the emulsion produced in this study, as observed using a microscope under 10 × magnification. The external phase was prepared before mixing with the W/O emulsion by dissolving ZnSO 4 of 50 ppm in deionized water with specified molarity of H 2 SO 4 for pH adjustment. A lower external phase concentration was decided by referring to the concentration of zinc pollutant in acidic mine drainage reported by Valenzuela et al. (2007). This is in order to verify the ability of the liquid membrane produced in this work to extract a lower concentration than that found in the wastewater. The mixing process of W/O emulsion in the external phase solution produced W/O/W emulsion.

Zinc extraction by ELM
The extraction process was executed with the assistance of a mechanical agitator at different operating conditions. The parameters investigated in this study were external phase pH, stirring speed, extraction time, concentration of the components of membrane phase, and internal phase as well as the W/O ratio. They were varied as specified in Table 3. After the predetermined extraction time was completed, the solution was left for 5 min for the emulsion and external phase to be completely separated by the act of gravity, before the sampling of the external phase can be made.

Analysis of samples
The metal content in the external phase was determined spectrophotometrically using Atomic Absorption Spectrophotometer (AAS) model HITACHI Z-2000 at a wavelength of 213.9 nm to determine the efficiency of extraction, Eff(%) (Ahmad et al. 2014): where C 0 e is the initial concentration of Zn in the external phase (ppm), while C e is the final concentration of Zn in the solute (ppm).

Sample characterizations
Fourier transform infrared (FTIR) spectroscopy (Spectrum One, FT-IR Spectrometer, Perkin Elmer) with wavenumbers spanning between 400 cm −1 and 4000 cm −1 was operated to recognize the infrared spectra of the functional groups in components. A programmable Electronic Rheometer (Anton Paar MCR 301) operated at room temperature was utilized to determine the viscosity. The density of waste cooking oil was manually calculated for every membrane, and the average value was reported.

Characteristics of the membrane phase components
The density and viscosity of WCO were found to be 798 kg/ m 3 and 64.4 mPa·s, respectively, at room temperature. As compared to the properties of WCO reported by Sujatha and Rajasimman (2021), the one used in this work has lower density and higher viscosity. Generally, vegetable  oil has higher viscosity due to the intermolecular attractions of the long-fatty acids in vegetable oils (Chang 2014 It is evident that some peaks are shifted, broadened, and intensified as carrier and surfactant were dissolved in the membrane phase. In Fig. 2(a), peak at the wavelength of 1743.77 cm −1 , 1161.11 cm −1 and 721.87 cm −1 represents C = O stretching, C-O stretching, and C = C bending of triglyceride in palm oil, respectively (Rosly et al. 2020b), while Fig. 2(b,c) shows considerable changes in the peak configurations with the inclusion of D2EHPA and Span 80 in WCO. The peak observed at around 1230 cm −1 and 1030 cm −1 is representing P = O and P-O-H vibration bands in D2EHPA (Azizitorghabeh et al. 2016), while the intense peak at 1465 cm −1 indicates the presence of C-H scissoring vibration in Span 80. The stretching vibration of the ester group function C-O-C was detected at 1161 cm −1 (Bharti et al. 2021;Farooq et al. 2019). All these peaks prove the presence of D2EHPA and Span 80 in the WCO.
Prominent changes are detected with the addition of D2EHPA and Span 80 in WCO forming emulsion as compared to when both are added individually into the WCO. The slight stretching of the absorption bands at around 1230 cm −1 and 1030 cm −1 in Fig. 2 respectively, proving the presence of D2EHPA in the WCO and emulsion (Senthilnathan et al. 2005;Shi et al. 2017). It is noted that the peaks for the phosphoryl group of P = O and P-O-H bonds shifted when D2EHPA is added to the emulsion confirming the existence of both components. The peak around 1600 cm −1 exhibits the presence of cyclic compound which is due to the Span 80 structure. Strong peaks between 1650 and 1800 cm −1 represent C = O stretch, while as shown in Fig. 3, C-H stretching was in Span 80 peaks between 2850 and 2925 cm −1 (Kohli et al. 2018). There are not many changes observed in the peak configurations with the addition of Span 80 in the emulsion since the surfactant is nonionic and has little interaction in complex formation. As all components are present, the extraction of Zn ions is highly likely at this point, given that the carrier is able to form a complex with the Zn.

Extraction equilibria
The extraction of Zn ions at a varying concentration of D2EHPA in WCO was tested. The extraction of Zn by D2EHPA as the carrier can be represented by the general equation as below (Svendsen et al. 1990): where M represents metal Zn, RH denotes the carrier species D2EHPA that acts as a liquid cationic ion exchanger, m metal valence, and n stoichiometric coefficient, while subscripts (aq) and (org) refer to the aqueous and organic phases, respectively. Hence, the equation above is expressed as Considering A + nB ⟺ C + D and assuming ZnR 2 .n(RH) 2 = MR 2 , the extraction equilibrium constant of K eq is as follows (Ahmad et al. 2013): Zn m+ (aq) + (1 + n)(RH) 2(org) ⟺ ZnR 2 .n(RH) 2(org) + mH + (aq) Zn 2+ (aq) + (1 + n)(RH) 2 (org) ⇔ ZnR 2 .n(RH) 2 (org) + 2H + (aq) Rearranging, According to the chemical reaction presented, the organophosphorus compounds denoted as HR in D2EHPA are the liquid ion-exchangers in the ELM system to exchange the Zn ions for protons (Fouad and Bart 2008). The carrier is dissolved in the membrane phase to selectively react with the Zn ions. The extraction process for this system is governed by a couple-transport, where the Zn ions are coupled with D2EHPA. The oil-soluble carriers that exist in the membrane form complexes with the Zn ions at the outer oilmembrane interface. The complex then diffuses through the membrane where the Zn ions exchange the counter ion with a suitable ion present inside the internal phase, in this case, the H + ions (Venkatesan and Begum 2008). The process is illustrated in Fig. 4.
The nature of the extraction process was evaluated by plotting a graph of log K D versus log [C B ] based on the Eq. (10) as shown in Fig. 5. The LLE molar ratio of the Zn reacting with D2EHPA is described by a straight-line gradient of the graph. The analysis suggested that the extraction mirrored as the first order with R 2 value of 0.9882. Based on Fig. 5, the value of n is 0.9364, a value close to 1. According to Eq. (4), two molecules of D2EHPA form a complex with one mole of Zn 2+ ion which is then released in the internal phase, exchanging each Zn ion with two protons to maintain electroneutrality (Medina et al. 2005). On the other hand, two moles of H + ions are relieved into the aqueous phase since the stoichiometric slope of line log K D versus log [C B ] is near to 1. Therefore, the Zn extraction model by D2EHPA can be suggested in ZnR 2 complexes with a molar ratio of Zn to D2EHPA to be 1:2. The mechanism of complexation, in this case, is translated in Eqs. (11) and (12):

Determination of external phase pH
LLE is a technique to separate compounds based on their relative solubility in two distinct immiscible liquids (Noah et al. 2018a), and it was applied in this study to determine the suitable pH for the reaction. The direct influence of  acidity of the external phase can be observed in Fig. 6. At the highest molarity of H 2 SO 4 used, minimal extraction efficiency was recorded. At 0.001 M (pH 3) of H 2 SO 4 in the external phase, the extraction peaked (85.49%) at 2 h but steadily decreases afterward. The experiment was halted at 7 h as it was concluded that the period is adequate for the system to reach equilibrium. Slightly lower efficiency was also found by adjusting the external phase with 0.01 M (pH 2) of H 2 SO 4 in a shorter duration of extraction. The results are in line with Fouad (2008) and Urtiaga et al. (2010) who achieved the highest Zn removal at pH 3. The H + ion activity in the external phase is closely related to the reaction rate of Zn extraction (Reis and Carvalho 2004). As illustrated in Fig. 4, the extraction is governed by the cation exchange reaction in which the protons are released (Fouad 2008). Therefore, the presence of high H + ion concentration in the external phase might compete with Zn ions for the available sites at the interface (Valenzuela et al. 2007), hence causing the kinetics and mass transfer resistance in the external phase to be less significant (Fouad and Bart 2008). Meanwhile, lower H + concentration induces the formation of metal complex with a carrier, which is favored for metal extraction in ELM (Venkatesan and Begum 2008). Medina et al. (2005) agreed that the carrier lost the ability to form a complex with the Zn ions at higher proton concentrations. Such behavior is found typical for metal cation extraction by cationic extractants where the increase in pH (lower concentration of proton) commonly results in enhanced metal extraction up to a threshold (Pereira et al. 2007). Other than that, Zn ion speciation at varying pH is another factor to be considered.
Apparently, increasing the pH from 1 to 3 has enhanced the extraction performance. Lower H + ion concentrations supposedly improve the extraction efficiency due to the formation of Zn-carrier complex at the interface between the external phase and liquid membrane. Although this condition is favored based on the equilibrium relationships, counterproductive results can be anticipated. In this case, too low H + ion concentration in the external phase will instigate the formation of zinc hydroxide precipitates, a species that is less extractable by D2EHPA (Fouad 2008;Fouad and Bart 2008). As the extraction takes place, there is a net transfer of protons to the external phase, diminishing the pH of external phase, although not to a great extent. As the pH lowers, the transport of Zn stops as the reaction in Eq. (4) proceeds to the left side of the equation (Medina et al. 2005). This situation explains the decreasing efficiency as a function of time. Nevertheless, the data available showcases the feasibility of WCO as the diluent and its compatibility with D2EHPA as the carrier for heavy metal extraction.

Effect of extraction time and speed
The effect of stirring speed and extraction time is demonstrated in Fig. 7. Data was collected at speeds between 300 and 700 rpm while the extraction time varies from 3 to 20 min. Interestingly, a similar trend can be observed across all speeds, steady increment at the early stage of the extraction and peaked at 10 min. The highest extraction was achieved by dispersing the system at 700 rpm for 10 min. 87.22% of Zn ions were removed at these conditions. It is evident that the extraction efficiency at all stirring speeds from 300 to 700 rpm was reduced when stirring time is longer than 10 min. Longer extraction duration resulted in either stagnant or worst and decrease in the efficiency. The latter is typically caused by the membrane breakage as the system is exposed to the shear by the impeller at longer time which may lead to coalescence of the emulsion. Evidently, the system requires a specific duration to allow the ions to be transported through the membrane phase. As the system obeys Type II facilitated transport, the formation complex between carrier and metal ions acted as the driving force (Malik et al. 2012). Additional extraction time is supposedly sufficient for the solute ions to form a complex with the carrier at the external-membrane interface. Since the stirring speed directly influences the emulsion droplet size, a sufficient time is important to generate higher numbers of small internal droplets which subsequently provides larger interfacial mass transfer for the solute extraction . The large emulsion globule sizes formed due to the short span of stirring time reduce the interfacial mass transfer area, resulting in a low complexation occurring among the Zn ions and carrier molecules at the externalmembrane phase interface (Kumar et al. 2018a).
Unfortunately, fine emulsion droplets can cause emulsion swelling due to the osmotic pressure gradient. The osmotic swelling was followed by large water molecule transfer from the external phase into the internal phase which will then lead to the leakage of the internal phase and expulsion of the extracted solute from the internal phase (Ahmad et al. 2017). A stable emulsion is crucial to minimize emulsion breakage that will result in secondary pollution of the treated water. Hence, the extraction efficiency is typically compromised in this situation. The hydration of surfactant also promotes water diffusion from the external phase into the internal phase, thereby resulting in emulsion breakage, as agreed by Othman et al. (2019). The rupture of internal phase droplets occurred due to the increasing spilling of stripping agents into the external phase as the extraction time increases. Increasing the stirring time increases the interfacial area for mass transfer but consequently increases the transportation of more water molecules (Kumar et al. 2018a). On top of that, the stirring speed has a direct influence on the size of the emulsion droplets.
It is essential to provide sufficient mechanical energy to uniformly disperse the emulsion droplets in the external phase. The agitation will provide shear that diminishes the membrane phase thickness and produces smaller emulsion globules. A thin membrane wall promotes better diffusion of Zn as it needs to travel a shorter distance to be stripped at the membrane-internal interface (Ahmad et al. 2017), resulting to high efficiency attained in a shorter time. Meanwhile, the synergistic effect between a thin membrane layer and small emulsion size contributed to high efficiency. At low stirring speed, the dispersion of emulsion droplets in the external phase is weak and slow, whereas higher agitation speed enhances the dispersion performance and facilitates the formation of smaller droplets. As a result, an increase in the contact surface area for mass transfer would then lead to high solute concentration . Mass transfer can be elevated with the increase in stirring speed, but unnecessary high speed will only increase the tendency of emulsion breakage. Beyond the optimum speed, extraction efficiency is undermined because the excessive shear stress results in a very thin membrane layer. Following that is the formation of unstable globules where the internal phase can easily leak into the external phase Rosly et al. 2019). The stability of emulsion generally is comprised of a high quantity of finer W/O emulsion droplets. However, these droplets can become bigger and imposes instability problem. Small emulsion globules tend to coalesce and causes emulsion instability (Kumar et al. 2018a;Sulaiman et al. 2020). It was observed in this study that the W/O/W emulsion turned cloudy as higher speed (> 700 rpm) was applied. Such stirring speed is relatively high for ELM system. Typically, the speed applied is between the range of 300 rpm to 500 rpm (Ammar et al. 2012;Venkatesan and Begum 2008). Undoubtedly, this is due to the influence of WCO's viscosity that is higher than the typical petroleumbased diluent.
As reported by Noah et al. (2018a), high viscosity palm oil resultantly produces an emulsion with higher viscosity and mass transfer resistance. To reduce the diluent viscosity, some studies mixed the palm oil with kerosene since a highly viscous and unstable solution (mayonnaise-like) mixture was formed by using palm oil Rosly et al. 2020b). However, such a condition was not encountered in this study even though the WCO viscosity recorded was pretty high too. However, due to the high viscosity of the emulsion, Othman et al. (2016) and Björkegren et al. (2015) claimed that 600 rpm is the minimum stirring speed for a pure palm oil-based ELM to be well dispersed in smaller sizes when in contact with the external phase which makes 700 rpm as the most suitable speed in this study. Nevertheless, the best speed identified is similar to Fouad [2]. Figure 8 shows the effect of varying carrier concentration on extraction efficiency. D2EHPA was used as a carrier in the formulation due to its low cost, excellent extraction selectivity, and high extraction efficiency . Particularly in the ELM system, the organophosphorus compounds in D2EHPA are the usual liquid ionexchangers in exchanging the Zn ions for protons (Fouad and Bart 2008). Available data suggested that extraction of the metal ions is impossible at 0 wt% of carrier. This is conspicuously caused by the insolubility of the ions in the WCO. Therefore, it is obvious that the carrier plays a critical role in the extraction of Zn ions. A sharp increment of the efficiency was noted at 2 wt% of D2EHPA. Furthermore, the emulsion became less viscous as the carrier concentration increases. This is a desired condition as reduction in viscosity allows better dispersion of the emulsion droplet. As a result, a thinner membrane layer is produced.

Effect of carrier concentration
A steady increment can be observed until 8 wt% of D2EHPA. Beyond this point, no significant improvement of extraction, indicating an excess of carrier in the membrane phase ). This finding is in line with Venkatesan and Begum (2008), in which they reported that the increase in carrier concentration decreases the stability of the membrane phase. Besides, an excessively high carrier content in the membrane phase will resultantly lead to swelling and breakage of the internal phase into the external phase and higher cost is incurred. From the perspective of the reaction equilibria, the number of moles of the carrier is sufficient at 8 wt% to completely extract the Zn ions from the external phase, even in the absence of ion exchange on the interfaces. The required carrier can be calculated based on the stoichiometry equation determined in "Extraction equilibria." Although such estimation can be made, the efficiency of the system is rather complex to be predicted as the emulsion is thermodynamically unstable and is always exposed to breakage and swelling.
Therefore, 8 wt% of D2EHPA concentration resulting in 86.22% of extraction efficiency was chosen to be applied for the next parameter. Identically, Fouad and Bart (2008) concluded that 8 wt% of D2EHPA was the best carrier concentration for Zn extraction using isododecane as a diluent. This proves that the usage of WCO as diluent does not influence the amount of carrier required in the system.

Effect of surfactant concentration
Surfactant is pivotal to maintain the emulsion stability and to initiate the formation of W/O emulsions through the lowering of interfacial tension between the two immiscible phases (Kiani and Mousavi 2013;Sulaiman et al. 2020). It reduces the energy required to break up droplets and elevates the emulsion stability. Surfactant, which is commonly comprised of amphipathic compounds that are soluble in both water and organic solvents (Ahmad et al. 2011), forms a thin layer between the immiscible interfaces and entraps the stripping agent in the internal phase (Shokri et al. 2020). Nonionic surfactant Span 80 has been widely used in various studies proving that it is specifically practical at low concentrations for emulsion stability and high extraction efficiency (Shokri et al. 2020). Selecting Span 80 as the surfactant will lessen not only the emulsion cost but also the use of toxic substances as it is one of the inexpensive nonionic commercial compounds (Valenzuela et al. 2010) and is environmentally friendly and biodegradable for its sugar base compound originated from renewable sources (Kassem et al. 2019;Kumar et al. 2019c) (Fig. 9).
It was observed that at 0 wt%, no emulsion was formed at all indicating the use of surfactant in the emulsion formation is indispensable. At 2 wt%, incomplete formation of the emulsion was spotted due to the insufficient surfactant to reduce the interfacial tension. The surfactant is not enough to effectively envelop the internal phase droplets of the membrane phase, thus making it highly likely to contribute to low emulsion stability (Daraei et al. 2019). Not only that, the insufficient Span 80 concentration reduces the coalescence rate of the dispersed phase droplets and thereby produces larger emulsion droplet sizes and consequently causes the emulsion droplets to easily break . As the concentration increases, more surfactant molecules adsorb at the interface between oil membrane and the aqueous internal phases which improves the strength of the adsorption layer and increases stability. Increasing Span 80 concentration to 4 wt% induces the formation of smaller droplets size due to the reduction of surface tension at the membrane phase interface and typically enhances the emulsion stability (Noah et al. 2018a). As a result, a clear incline of extraction efficiency was recorded. The formation of small emulsion droplets is highly desired until the number is too high. Beyond 4 wt%, the extraction efficiency started to plateau due to the destabilization of emulsion resulted from rapid coalescence between droplets. On the other hand, unnecessarily high surfactant concentration increases the emulsion viscosity which leads to the formation of a thicker membrane. Interfacial mass transfer resistance increases in this case and interfering with the carrier-solute reaction at the interface (Daraei et al. 2019;Jusoh and Othman 2017;Noah et al. 2018a). In fact, the formation of reverse micelle is made possible by exceeding the threshold limit of surfactant concentration, termed as critical micelle concentration (CMC). According to Valenzuela et al. (2003), the CMC value of Span 80 in the membrane phase was determined to be 0.135 mol/L (~ 5.87 wt%). Therefore, the obtained peak efficiency at 4 wt% of Span 80 concentrations in this study is well-matched with the cited CMC value. Apparently, beyond that concentration, the excess surfactant added to the system might turn into a reverse micelle that transports water from external phase into the internal phase due to the osmotic pressure, hence causing swelling. Increased volume of the internal phase caused by swelling will then trigger the breakdown of the globules (Ooi et al. 2015;Yan and Pal 2001). The ELM system developed in this study requires only 4 wt% of Span 80 to produce the W/O emulsion that can remove 93.41% of Zn ions from the external aqueous phase. Figure 10 illustrates the effect of W/O ratio on the extraction efficiency. To retain the effect of the stirring speed and time, the total volume of emulsion was kept constant while the volumes of internal and membrane phases were varied accordingly. Encapsulation of the internal phase is highly dependent on this parameter. A high W/O ratio provides insufficient value of membrane phase volume to efficiently encapsulate the internal phase droplets causing the internal phase to leak easily into the external phase (Daraei et al. 2019). This condition is highly likely to occur at 1:1 ratio. Consequently, the lowest extraction efficiency was identified, as shown in Fig. 10.

Effect of W/O ratio
As claimed by Kumar et al. (2019a), larger membrane phase volume enhances the emulsion stability due to the increment of interfacial area, but in turn, it will exert more resistance for the solute transfer process. The W/O ratio in the ELM process is instrumental as it affects the extraction rate and imposes changes in the emulsion structure and properties. The capacity of the emulsion to extract is also affected (Kumar et al. 2018a). Besides, the tendency of phase inversion can be hampered by controlling the ratio between water and oil phases.
According to Fig. 10, the extraction rate increases as the W/O ratio decreases from 1:1 to 1:4 due to the optimized amount of internal phase agent molecules per unit volume which contributes to the improved stripping of solute ions in internal phase at the inner interface. Upon decreasing the internal phase volume, the emulsion globules started to effectively surround the stripping phase fine droplets for the extraction of Zn ions into the internal phase (Kumar et al. 2019b). As the ratio reached 1:4, the amount of stripping agent in the internal phase is large enough for dense distribution of internal phase globules inside the emulsion globule. It will subsequently shorten the diffusion path length for the transportation of complexes (solute-carrier) within the emulsion globules (Kumar et al. 2018a).
However, not many changes in extraction efficiency were recorded at the ratio of 1:5. This is because the emulsion viscosity increased remarkably, resulting in the magnification of the internal phase droplets size (Daraei et al. 2019). Droplets with huge diameter preclude a larger contact area between the external and emulsion which consequently inhibit the Zn ions transfer into the internal phase. The inadequate volume of internal phase as the ratio decreases recorded lower extraction efficiency as there is less availability of stripping agent to cater for the high Zn concentration (Kumar et al. 2018a). Therefore, the W/O ratio of 1:4 was chosen as the best condition to attain the highest extraction efficiency (94.31%).

Effect of internal phase concentration
The use of stripping agent in the internal phase was mainly to maintain a sufficient driving force for the solute transportation by trapping the metal ion in the internal phase droplets through its conversion into a membrane insoluble compound (Björkegren et al. 2015;Khadivi and Javanbakht 2020). Figure 11 elucidates the effect of internal phase concentration on extraction efficiency, where the efficiency peaked at 1 M and decreases at a concentration higher than that. The trend is in agreement with the results reported by Othman et al. (2016). Note that internal phase concentration directly influences the osmotic pressure gradient between the external and internal phase, as suggested by Kumar et al. (2019c). Appropriate concentration allows more solute complexes to be transported at the inner interface between the membrane and internal phases (Kumar et al. 2018a;Othman et al. 2019). Sufficient internal phase concentration increases its stripping ability which then elevates the formation of a carrier-solute complex in the membrane, resulting in high extraction efficiency (Kumar et al. 2019b). On the contrary, excessively high osmotic pressure gradient triggered the transportation of water molecules from the external to the internal phase, hence triggering emulsion swelling (Kumar et al. 2018a). Based on our observation, apparent cloudiness of the aqueous solution was visible after the extraction process as 0.25 M was used as the external phase.
The sharp fall of the efficiency at 1.5 M might be due to the reaction between Span 80 with the acidic internal phase, resulting in the loss of surfactant properties and emulsion breakdown. There might be a competition for the limiting reagent between Span 80 and H 2 SO 4 as they are involved in the stability of emulsion and stripping reaction, respectively (Shokri et al. 2020). Span 80 was found to undergo hydrolysis reaction in acidic media leading to its decomposition (Abbassian and Kargari 2016;Daraei et al. 2019). Surfactant hydrolysis inflicts the loss of an effective number of surfactant molecules, causing emulsion instability and breakage (Othman et al. 2016).
Within the studied range, the extraction efficiencies are nearly identical, presenting no substantial impact as internal phase concentration increases. This outcome is also supported by Urtiaga et al. (2010) and Fouad and Bart (2008). Therefore, the optimum concentration of internal phase, in this case, is concluded to be 1 M where 95.17% of Zn was extracted.

Conclusions
The incorporation of WCO as the diluent in ELM system was proven in this study. Evidently, the formulation of the emulsion produced does not differ much from the other typically reported ELM using petroleum derivatives as diluent. This study tested few parameters and documented that the highest 95.17% of Zn ions were extracted at the conditions as listed; 0.001 M of H 2 SO 4 in the external phase (pH 3), 700 rpm of stirring speed at 10 min of extraction time, 8 wt% of carrier, 4 wt% of surfactant concentration, 1:4 of W/O ratio, and 1 M of internal phase concentration. More importantly, in kinetics, hence, the performance of the system is not compromised. WCO was proven as the alternative green diluent in the GELM method substituting the toxic conventional petroleum-based diluents in the current application. The use of waste as the major component in the formulation opens new opportunities for ensuring the availability and sustainable management of water treatment. The findings of this study shall, therefore, serve as another milestone in the development of ELM as we can now shift the attention towards the production of safer technology to treat heavy metal ions from an aqueous solution. Data availability Not applicable.

Declarations
Ethics approval and consent to participate Not applicable.

Competing interests
The authors declare no competing interests.