Fabrication and response optimization of Moringa oleifera-functionalized nanosorbents for the removal of diesel range organics from contaminated water

The purpose of this research is to synthesize environmentally friendly nanosorbents for the novel adsorption of diesel range organics (DRO) from contaminated water. Central composite design (CCD) analysis of response surface methodology (RSM) was employed in a model fitting of the variables predicting the adsorption efficiency of Moringa oleifera-functionalized zerovalent iron particles (ZINPs) for the removal of DRO. The effects of the reaction parameters on the response were screened using 24 factorial designs to determine the statistically significant independent variables. A quadratic model predicting the DRO adsorption efficiency of ZINPs with an F value of 276.84 (p value < 0.0001) was developed. Diagnostic plots show that the predicted values were in excellent agreement with actual experimental values (R2 = 0.99). The maximum percentage removal of DRO of 92.6% was achieved after optimization, using the synthesized ZINPs after 8 h of contact between DRO substrates and ZINPs at pH of 8, the temperature of 25 °C, with an adsorbent dosage of 2 g/L and at composite desirability of 1. Characterization of ZINPs revealed the formation of quasi nanospheres and nanocubes with an average particle diameter of 50.9 ± 9.7, a crystallite size of 15.31 nm, a crystallinity index of 32.47% and a pore width of 75.69–88.59 nm. The adsorption equilibrium data modelling of ZINPs for adsorption of DRO was best described by Langmuir isotherm with the maximum monolayer coverage capacity of 7.194 mg/g. The separation factor RL=0.472\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{L}=0.472$$\end{document}, indicated favourable adsorption. The adsorption kinetic data were consistent with pseudo-second-order kinetics indicating probable chemisorption.


Introduction
Petroleum products and their derivatives are one of the major sources of pollution in today's world of growing industrialisation (Bandura et al. 2017;Mohanakrishna et al. 2019). Oil spills impose serious damage to the environment. Spilt crude oil or its products affect aquatic flora and fauna (Paulauskienė et al. 2014). However, pollutants such as petroleum hydrocarbons (PHCs), which are widely applied in the areas of transportation, electricity, production of plastics, etc., are the major cause of the pollution of surface and groundwater through accidental and anthropogenic activities such as accidental spills, leakages and industrial activities (Al-hawash et al. 2018). PHCs consist of majorly four categories of compounds namely; aliphatics, aromatics, resins and asphaltenes (Adeniji et al. 2017;Ahmed et al. 2020). The short-chain hydrocarbons are usually degraded by microorganisms, but large branched aliphatic chains and aromatic hydrocarbons usually persist in the environment (Ahmed et al. 2020), which highlights the importance of PHCs remediation from water bodies. PHCs are toxic compounds classified as priority contaminants because of their nature and versatility in use (Varjani 2017). The release of these compounds into the environment accidentally or by anthropogenic activities leads to environmental pollution and consequent destabilization of the ecosystem (Ahmed et al. 2020). For instance, PHCs in water forms a film that could prevent oxygen entry, leading to the death of aquatic organisms (Faustorilla et al. 2017). This has inevitably been of major concern over some decades and has raised the need for PHCs remediation of surface and groundwater in which physical and chemical methods such as the dissolved air flotation, use of skimmers, barriers, booms, dispersant or surfactant spray have been employed, whereas the biological and thermal techniques were based on bioaugmentation, and burning of crude oil, respectively (Barthlott et al. 2020;Liu et al. 2021;Navarathna et al. 2020;Tran et al. 2021). Although these methods have proven effective, major challenges including the formation of toxic polycyclic aromatic hydrocarbons (PAHs) through oxidation, the limitation of microorganisms to perform due to PAHs and high cost have limited their wide application (Bhattacharya et al. 2018;Bianco et al. 2021;Fuentes et al. 2020).
Nanoparticles (NPs) remediation has presented a cheaper, broader and more efficient way of removing major water contaminants from water bodies due to their large surface areas and wide applications (Khan et al. 2019a, b;Xue et al. 2017). NPs can be effectively used to treat contaminated water in situ by sequestering contaminants (via adsorption or complexation) or by degrading harmful contaminants into less harmful compounds. NPs are, however, mostly toxic due to the chemical nature of their precursors. NPs also require capping agents to prevent agglomeration of the particles. These capping agents are mostly toxic and expensive ; hence, the need for eco-friendly, non-toxic and cheaper nanoparticles such as those synthesized from plant materials (green synthesis).
The green synthesis of NPs has been considered as a rapid, clean, non-toxic, cost-effective and environmentally friendly method compared to other conventional techniques (Gautam et al. 2019). The large surface area of these nanoparticles, coupled with the antioxidant contents of plant extracts used in the formulation of these biogenic NPs (BNPs), plays a very important role as biodegradation agents and nanosorbents. The antioxidants and phytochemicals from plant extracts act as nontoxic bio-reductants and capping agents for the metal-based nanoparticles (Devatha et al. 2016;Murgueitio et al. 2018). This dual function of plant extracts to act as reducing and capping agents due to their antioxidant contents makes them even more beneficial for the synthesis of NPs.
Iron nanoparticles (INPs) modified with various plant extracts have been extensively studied and employed in the biosynthesis of nanoparticles for water remediation (Ercan 2019;Ghidan et al. 2016;Pan et al. 2019;Vitta et al. 2020). Iron is a transition metal with empty s-and d-orbitals in its oxidized state. The electron species from the antioxidants of plant extracts can be donated to these empty orbitals causing the reduction of Fe 2+ or Fe 3+ to Fe 0 which is the basis for the formation of zerovalent iron nanoparticles (ZINPs). This phenomenon is capable of creating unique surfaces for the adsorption of organic pollutants. Although the physicochemical properties of INPs synthesized from different plant extracts under various conditions differ (Ebrahiminezhad et al. 2018), the INPs generally possess a unique core-shell structure and large surface area capable of organic pollutant degradation and remediation through oxidative and reductive pathways. Iron is often preferred as a precursor for the production of nanoparticles due to its availability, nontoxicity, and cheapness.
The Moringa oleifera plant belongs to the Moringaceae family. M. oleifera is one of the best known medicinal plants used in traditional medicine in most countries and contains antioxidants such as quercetin, gallic acid, catechin, β-carotene, ascorbic acid, phenolics and flavonoids in various parts of the plant such as leaves, seed, bark and pod (Kerdsomboon et al. 2021). M. oleifera plant extract has been used for the biosynthesis of INPs which has been successfully applied in the remediation of surface and groundwater contaminants such as nitrate (Katata-Seru et al. 2018;Velu et al. 2021), nitroaromatic explosive compounds (Kgatitsoe et al. 2019), fluoride (Silveira et al. 2018), chromium (Bahador et al. 2021) and in antibacterial applications (Aisida et al. 2020;Jegadeesan et al. 2019). Moringa oleifera leaves being very rich in antioxidants and phytochemicals makes them very suitable for the reduction and stabilization or capping of iron by deposition of the electronic species contained in the antioxidants, in the nanoscale (Oladeji et al. 2020;Osamede and Kaewmanee 2019), leading to the fabrication of zerovalent metal nanoparticles. The physicochemical characteristics of the synthesized INPs by using different plant extracts are not the same. Reaction conditions such as temperature, concentration of iron precursor, amount of plant extract, retention time and chemical specifications of the plant extract (i.e. preparation temperature, concentration, pH and phytochemical molecules) have significant effects on the physicochemical properties of the resultant nanoparticles (Ebrahiminezhad et al. 2018).
Although researchers have studied the application of Moringa oleifera functionalized ZINPs for remediation of various water contaminants, the application of ZINPs for the removal of diesel range organics (DRO) from water has not been studied before now. DRO, which are major water pollutants, are compounds typically found in diesel fuel, which comes from crude oil. DRO are grouped as total petroleum hydrocarbons (TPH), which are compounds containing hydrogen and carbons. They typically range anywhere from 8 to 28 carbon atoms. Few studies have reported the adsorption of DRO using different adsorbents. Aziz et al. (2020) reported that biochar made with fruit/vegetable and sewage showed 82.86% removal efficiency when applied to DRO, while Agamuthu and co-workers (2013) reported that diesel contaminated soil amended with cow dung recorded oil biodegradation of 26%. This present study aims to synthesize and characterize ZINPs as nanosorbent using M. oleifera leaf extracts (MOL) as the bio-reductant and capping agent for the iron precursor. In this study, ZINPs was applied for the first time for the novel biodegradation and adsorption of DRO from contaminated water while varying the reaction parameters such as temperature, pH, adsorbent dosage and contact time. The sorption properties and reaction mechanism of the ZINPs were elucidated by employing adsorption isotherms and adsorption kinetic models. A regular 2 4 factorial design was applied to screen, evaluate and determine the main and interaction factor effects on the adsorption efficiency (response) of ZINPs. The central composite design of the response surface methodology was applied for the surface analysis and optimization of the statistically significant independent variables affecting the response.

Chemicals and reagents
The TPH Standard Mix (Certified Reference Material) was obtained from Sigma-Aldrich USA, Catalogue number: 861424-U Supelco. This stock standard solution which is DRO, contained 10 components (n-C 10 − n-C 28 ) each at 2000 µg/mL in dichloromethane:hexane (1:1) solvent. The analyte values of DRO are shown in Table S1. The internal standard was 5-alpha-androstane (Certified Reference Material), obtained from Sigma-Aldrich USA, Catalogue number: A0887 Sigma-Aldrich. Reagent-grade n-hexane (PN: MER-107288, Merck, Germany) was used as blank, for standard solution preparations and for dilutions. Dichloromethane for the injection syringe wash was obtained from BG Chemicals, Malaysia. Organic-free reagent water was used throughout. Ultra-high pure helium (99.999%) while ultra-high pure grade hydrogen (99.999%) and high pure (99.99) grade were used as the detector gases. Purified air of high purity grade was employed as the oxidant gas for the detector. Other chemicals included Folin Ciocalteu reagent, quercetin, iron sulphate heptahydrate, aluminium chloride, sodium hydroxide and hydrochloric acid, all obtained from BG Chemicals, Penang, Malaysia. The Moringa oleifera plant leaves were collected from Universiti Sains Malaysia (USM) campus premises and confirmed by the authorized staff of School of Biological Sciences, USM. The DRO samples were obtained by spiking organic reagent-free water using the TPH standards. All chemicals were of analytical grade or reagent grade and were used without further purification.

Extract preparation
The Moringa oleifera leaf extracts (MOL) were prepared according to the method used in literature (Ademiluyi et al. 2018), with slight modifications. The obtained fresh Moringa oleifera leaves were washed with 30% ethanol to remove organic contaminants, washed under running tap water, and drained in a plastic sieve. Thereafter, the leaves were sun-dried and pulverized using mortar and pestle. The leaf broth solution was prepared by weighing 1 g of powdered leaves and extracted in 100 mL of 95% ethanol at 30 °C for 4 h using a Soxhlet extractor. The resulting extract mixtures were filtered using Whatman filter paper 42 (pore size: 2.5 µm). The filtrates were centrifuged (Kubota 5920, Japan) at 3500 rpm for 20 min and the obtained supernatants were concentrated under reduced pressure and controlled temperature (40-45 °C) using a rotary evaporator (Heidolph, Laborta 4011 digital, Germany) to obtain the ethanolic extract of MOL. The extracts were stored at 4 °C prior to analysis. All analyses of the leaf extracts were performed within 10 days of extraction.

Determination of total phenolic content and total flavonoid content of MOL
The total phenolic content (TPC) of MOL was determined using the Folin-Ciocalteau method as described in the literature (Lamuela-Raventós 2017; Mahdi et al. 2016). Briefly, a mixture of 0.4 mL of MOL in methanol, 2 mL of Folin-Ciocalteau reagent and 1.6 mL of Na 2 CO 3 was vortexed and kept for 2 h and the absorbance was measured at 765 nm using a spectrophotometer (Shimadzu-UV-2600, 00,491, Malaysia).
Using gallic acid monohydrate (1-100 ppm), a calibration curve was prepared. Using the curve, the TPC was calculated and expressed as gallic acid equivalent (GAE) (mg of GAE/g of dried extract).
The total flavonoid content (TFC) of MOL was determined by the well-known aluminium chloride method as described by Mathur and Vijayvergia (2017) with quercetin as standard, and the flavonoid content of the extracts, expressed in mg quercetin equivalent (QE)/g of dried extract. The absorption of the solutions was measured at 415 nm against blank using a spectrophotometer (Shimadzu-UV-2600). The amount of flavonoid was calculated from the linear regression equation obtained from the quercetin calibration curve. The flavonoid content was calculated as mean ± SD (n = 3) and expressed as mg of QE/g of dried extract.

Fabrication of zerovalent iron nanoparticles using Moringa oleifera leaf extracts
The zerovalent iron particles (ZINPs) were prepared according to the method used in literature ) with some modifications. Figure 1 describes the basic steps for the fabrication of ZINPs. MOL was added dropwise to different concentrations (0.05-0.5 M) of ferrous sulphate with moderate stirring using a magnetic stirrer (Wisestir SMHS-3, Korea) at room temperature and ambient air for 30-45 min to achieve the reduction of iron (II) ions to zerovalent ions evident by the colour change of the solutions from pale green to intense black colour. Different volume ratios (1:1, 1:2, 2:1, 3:1 and 1:3) of MOL:ferrous sulphate were used for the preparation of ZINPs. The best percentage yields of ZINPs were obtained with a volume ratio of 1:2 (MOL:ferrous sulphate). The percentage yields of ZINPs using the volume ratio of 1:2 (MOL:ferrous sulphate) for ZINPs 0.05 , ZINPs 0.1 , ZINPs 0.25 and ZINPs 0.5 were 19.1, 17.4, 23.2 and 29.5%, respectively. Different concentrations of ZINPs were denoted by ZINPs 0.05 , ZINPs 0.1 , ZINPs 0.25 and ZINPs 0.5 .

Instrumental methods of analysis and characterization
The formation of ZINPs was investigated by UV analysis (UV-2600 SHIMADZU S. No. A116650). in the range of 185-800 nm. The morphology and elemental analysis of ZINPs were investigated on a scanning electron microscope (Quanta FEG 650 SEM) equipped with an energy-dispersive X-ray spectroscopy system (Oxford X-Max 50 mm 2 EDX). Fourier transform infrared (FTIR-KBr) (Perkin Elmer System 2000 FTIR-KBr spectrometer) and FTIR-ATR (Perkin Elmer FT-NIR spectrometer) spectrum analyses were performed to identify the presence of functional groups on the surface of the MOL and ZINPs which indicated the possible antioxidants as bio-reductant and capping agents for the ZINPs. The liquid samples of MOL were prepared and introduced into the diamond sample holder of the FTIR-ATR spectrometer for analysis, with spectra range of 600-4000 per cm while the powdered ZINPs samples were mixed with KBr in the ratio of 1:100, and introduced into the sample holder of the FTIR-KBr spectrometer with spectra range of 400-4000 per cm. The baseline-corrected FTIR spectra were obtained. The distribution, size and morphology of the NPs were further investigated using energy-filtered transmission electron microscopy (EFTEM) (Zeiss Libra 120, Germany). Before the EFTEM analysis, ZINPs samples were prepared by dispersion in ethanol under sonication (Brason 1510) for 5 min and stained with uranyl acetate. The dispersed samples were dropped on copper grid surfaces before introduction into the EFTEM microscope for analysis. The XRD analysis was done for ZINPs on a fully automated Bruker's D8 Advance X-ray diffractometer armed with Cu-K α radiation source at 45 kV, a current-voltage of 40 mA, λ = 1.54060 Å, 2θ scanning range of 10-70°, scanning speed of 0.04°/s at 25 °C for 25 min, available at Makmal Pencirian Bahan Bumi of the Centre of Global Archaeology Research, Universiti Sains Malaysia. The samples for XRD analysis were prepared by hydraulic pressing of the powdered samples in a sample holder (25-mm diameter circular disc). Mean crystalline sizes were determined via Scherrer approximation (Sani et al. 2021). The specific surface area, pore size and volume distribution were studied using Brunauer-Emmett-Teller (BET) nitrogen adsorption-desorption isotherm analysis.

Preparation of DRO working standards
A secondary standard solution of DRO was made from the 2000 μg/mL stock solution. Working standard solutions from for calibration was prepared from the primary and secondary standard solutions in the range of 0.1-100 μg/mL (for each component). These were made for linearity, precision, and recovery determinations.

GC-FID analysis
DRO analysis was performed on an Agilent 6890 gas chromatography system equipped with a flame ionization detector (FID). Separation was carried out using an Equity 5 capillary column (30 m × 0.25 mm i.d. × 0.25 µm film thickness) obtained from Supelco. Nitrogen (purity 99.99%) was employed as carrier gas at a flow rate of 1.5 mL/min. Hydrogen and purified air were used as a fuel and an oxidant to support the combustion in FID, respectively. The injector temperature was held at 290 °C and the detector temperature at 305 °C. The oven temperature was programmed from 40 °C (held 2 min) to 305 °C at 30 °C/min. Then, 1 μL was injected in the split mode with a split injection ratio of 10:1. The total run time was 15 min.

Method of DRO extraction and method validation
Extraction of DRO from water samples was performed using US EPA 3510C method (Akpaden and Enin 2016). Sample analytes were transferred into a separatory funnel and n-hexane was used for liquid-liquid extraction of DRO. After 10 min of vigorously vortexing, the mixture was allowed to separate and the organic phase was collected through anhydrous sulphate. The concentrated extract was stored at 4 °C. Method validation was evaluated in terms of linearity, precision and accuracy. The linearity of the calibration curve of DRO was determined by injecting seven standard solutions with different concentrations ranging from 0.1 to 100 μg/mL. The linear regression was obtained by plotting the calibration curve of peak area against the concentration of the standard solutions. The n-hexane acted as the solvent blank. The response of the solvent blank was determined by injecting the n-hexane onto the GC-FID to ensure a contamination or interferences-free system. Organic-free reagent water was subjected to extraction and analysis to ensure that there were no interferences by the method and apparatuses. Accuracy was assessed by testing the recovery of spiked samples. System precision was evaluated by assessing the reproducibility of results using the method of relative percent difference (RPD) as shown in Eq. 3. As part of quality control, the calibration relationship initially established by the DRO calibration curve was periodically confirmed by periodic injection and assessment of the 20 μg/mL standard which served as the continuing calibration verification (CCV) solution during sample analysis. Replicate injection of the CCV solution was performed and analysed at the start, middle and end of each batch analysis, or after an analysis of a relatively concentrated DRO sample (> 50 μg/mL). A batch is a group of samples analysed as a unit. Percent relative standard deviation (RSD%) of the method was calculated. EPA method requires that the RSD must be 20% or less (U.S. EPA. 2003).

Adsorption batch experiments
In each adsorption experiment, 1 g of ZINPs was added to 1000 mL of each DRO sample and were agitated at 180 rpm using a rotary orbital shaker (Stuart SSL1) at room temperature and pH of 6.5 for 1 h. The shaking step started after recording the initial pH of the solutions. The initial pH solutions were adjusted to the desired value by adding either 1 M of HCl or 1 M of NaOH solution. The effect of pH, temperature, adsorbent dosage and contact time were all investigated by varying the pH (4, 6, 8 and 10), the temperature (25, 35, 45 and 55 °C), the adsorbent dosage (0.1, 0.5, 1, 2 and 3 g) and the contact time (10,20,30,40,50,60,120,240,420 and 720 min) of the solutions. Other reaction conditions were left constant during the investigation of each reaction parameter. Thereafter, the samples were filtered, centrifuged and subjected to DRO extraction using the method described earlier ("Method of DRO extraction and method validation" section). The extracted analytes were concentrated to 2 mL and stored in 5 mL GC vials at 4 °C pending analysis. GC-FID analysis of all analytes was done within 7 days after the extraction.

Experimental design
The regular two-level, 4 factors factorial design (2 4 ) was employed in the simultaneous screening of four independent variables (contact time, adsorbent dosage, pH and temperature) to determine the statistically significant effects of the variables on the adsorption efficiency of Moringa oleifera functionalized zerovalent iron nanoparticles. A central composite design (CCD) of the response surface methodology (RSM) on Design-Expert software, version 11.0 (Stat-Ease, Minneapolis, MN) was used to explore and optimize the relationships between the statistically significant explanatory variables (initially screened with the 2 4 factorial design), and the response variable. The central composite design was preferred due to its ability to generate response surfaces with comparatively fewer experimental runs (Laid et al. 2021). A suggested reduced quadratic model was used to analyze and predict the variables for optimization purposes, and to define the main and interaction statistically significant effects of the variables. The experimental design matrix where 14 experiments at different low-high factor levels and 6 experiments at the central point of the selected factor levels were proposed. To characterize the interactions and effects of different factors on the responses, the general form of the second-order polynomial quadratic model (Eq. 1) was employed (Fereidonian Dashti et al. 2021): where Y denotes predicted response, X i and X j represent the coded independent variables, k is the number of factors, e shows the random error, β 0 defines the constant-coefficient and β i , β ii , β ij specify linear, quadratic and interaction coefficient, respectively. Analysis of variance (ANOVA) was also carried out to analyze interactive effects between independent variables and dependent responses. Confirmatory experimental runs using the predicted optimum values were used to validate the suggested model.

Proposed mechanism
The mechanism pathway for the synthesis of ZINPs has not been intensively and extensively studied. However, the fabrication of ZINPs requires the Fe 2+ ions to be reduced to Fe 0 atoms with a core-shell structure possessing unique reactive surfaces capable of adsorption and transformation of pollutants via oxidation (Li et al. 2021). The antioxidants and phytochemicals in the plant extracts are responsible for the donation of electron species or radicals to the Fe 2+ ions leading to the reduction of Fe 2+ ions to Fe 0 atoms, thereby stabilizing the synthesized ZINPs. In the presence of dissolved oxygen, ZINPs are capable of oxidation of a wide (1) range of organic substrates/pollutants with high activity, activated by Fe 2+ ions generating radical species in solution (Galdames et al. 2020):

Determination of antioxidant capacity in Moringa oleifera leaf extracts
The total phenolic content (TPC) of MOL expressed as gallic acid equivalent (GAE) in mg/g of dry plant material and the total flavonoid content (TFC) of MOL expressed as quercetin equivalent (QE) in mg/g of dry plant material are presented in Table 1. The TPC of MOL was determined from the gallic acid calibration curve (R 2 = 0.991) while the TFC of MOL was determined using the linear regression method to calculate the quercetin equivalent concentration from the quercetin calibration curve. The TPC and TFC results of the three replicate experiments were pooled and expressed as mean ± standard deviation (SD). The TPC of MOL in this study was found to be 55.97 ± 3.10 and the TFC of MOL was found to be 11.3 ± 0.06. This indicates that the MOL contained antioxidants and phytochemicals responsible for the reduction of Fe 2+ to Fe 0 and the stabilization of ZINPs. Similar results have previously been obtained by previous researchers (Mahdi et al. 2016;Zullaikah et al. 2019).

UV analysis
The UV-visible spectrophotometer was used to assess the formation of the zerovalent iron nanoparticles (ZINPs). First, there was a visual confirmation of ZINPs formation by a colour change from pale-green of the Fe 2+ solution to translucent yellow after stirring (due to oxidation), then to greenish-black on the addition of MOL to aqueous Fe 2+ solution (Fig. 2b). This colour change is due to the excitation of the surface plasmon resonance (λ SPR ) in the metal nanoparticles proving the reduction of Fe 2+ to Fe 0 by the flavonoids and polyphenols (Da'na et al. 2018). UV peaks were observed at 207, 265, 319-329, 403, 461 and 660 nm for the MOL spectrum due to the presence of polyphenols and flavonoids (Fig. 2a(i)). However, from the UV spectra of ZINPs, as shown in Fig. 2a(ii), absorption peaks of ZINPs were only observed at 195-210, 225 and 272-275 nm which are characteristic λ SPR values of ZINPs due to the formation of ZINPs and subsequent homogenous nucleation of iron nuclei (Katata-Seru et al. 2018;Devatha et al. 2016;Khajelakzay et al. 2015;Murgueitio et al. 2018). The disappearance of other peaks from the UV spectrum of MOL to just a few peaks at 195-210, 225 and 272-275 nm for the ZINPs is related to the utilization of the antioxidant species in the MOL for the reduction of Fe 2+ and stabilization of ZINPs (Murgueitio et al. 2018). The hypsochromic shift to shorter wavelengths after ZINPs formation could be attributed to electron transitions.

Fourier transform infrared spectroscopy
The Fourier transform infrared (FTIR) analysis investigated the presence of phytochemicals in MOL and the functional groups of ZINPs as shown in Fig. 3. More intense peaks are noticed with the FTIR spectrum of MOL as shown in Fig. 3a than with the ZINPs as shown in Fig. 3b-e. The intense FTIR peaks of MOL are due to the antioxidant and phytochemical contents of MOL (Madubuonu et al. 2019;Zullaikah et al. 2019). These antioxidants and phytochemicals are responsible for the reduction of the iron (II) ions of ferrous sulphate to zerovalent ions leading to a more stable and lower energy state of the ZINPs which is translated to a lower vibrational frequency, leading to the lower peak intensities of ZINPs (Katata-Seru et al. 2018). Therefore, the more intense FTIR peaks of MOL in Fig. 3a than the FTIR peaks of ZINPs as shown in Fig. 3b-e and the shift in the wavenumbers of the absorption bands indicate the interaction of the antioxidants or bioactive species of the MOL with the iron precursor (Fe 2+ ) (Laid et al. 2021). This depicts that functionalization by the MOL was successful. Furthermore, the peaks of MOL and ZINPs at 3200-3400 cm −1 are attributed to the -OH stretching due to the presence of polyphenols and water moisture (Aksu Demirezen et al. 2019). The bands at 1624-1668 cm −1 are attributed to C = O stretching which indicates the presence of ketones, aldehydes and carboxylic acids (Mahdavi et al. 2013). The band at 2888 cm −1 characterizes the band of the C-H bond of the alkane group (Katata-Seru et al. 2018).
Similar results on extraction of Moringa oleifera leaves using ethanol and water as solvents have been previously reported by researchers (Welch and Tietje 2017).
In Fig. 3a-e, it can be seen that generally, a few peaks corresponding to MOL disappear after the addition of FeSO 4 .7H 2 O solutions, which shows that the flavonoids and antioxidants from MOL were utilized for the reduction of Fe 2+ and for the stabilization of ZINPs. This lowering of peak intensities and disappearance of peaks were observed only after the interaction of MOL. This indicates that the antioxidants in MOL might have been utilized in the capping of the synthesized ZINPs. Chemical shifts were also observed after the formation of BINPs from MOL, but the major chemical shifts were observed in stretching frequencies of 1643 to 1617 cm −1 . These chemical shifts are due to the involvement of the antioxidants in the capping and stabilization effects of the MOL on the ZINPs (Raut et al. 2014).
It is worth noting that some of the functional groups in the MOL could be as a result of different substrates such as oil or sugar as will be explained in the XRD analysis result . However, FTIR analysis only identifies the functional groups in a given sample. Since different substrates can have common functional groups, FTIR analysis alone cannot be used to determine the type of substrate in a given sample except for the functional groups. The formation of ZINPs was further confirmed by a characteristic band at 610 cm −1 as shown in Fig. 3

Scanning electron microscopy/electron dispersive X-ray analysis
The scanning electron microscopy/electron dispersive X-ray analysis (SEM/EDX) was an important tool for the assessment of the morphology and elemental analysis of ZINPs. The micrographs of ZINPs in Fig. 4 confirm the formation of nano-sized particles with quasi nanocube shapes. Figure 4a and b show the micrographs of MOL in different resolutions while Fig. 4c shows the elemental composition of MOL. From the figure, it is obvious that MOL is amorphous. Previously, Attah et al. (2020) have also previously reported the morphology of MOL to be amorphous with no discreet shape. However, it can be observed that the morphology of ZINPs with the lowest iron precursor concentration of 0.05 and 0.1 M (ZINPs 0.05 and ZINPs 0.05 , respectively) as shown in Fig. 4a and b is partly spherical and partly quasi-nanocube shaped. The quasi nanocube morphology of the ZINPs kept improving with an increase in iron precursor concentration (ZINPs 0.05 < ZINPs 0.1 < ZINPs 0.25 < ZINPs 0.5 ) while the nanosphere morphology kept decreasing until it completely disappeared. It will later be observed that the ZINPs of higher iron precursor concentration, which in turn are dominantly quasi nanocubes in shape possess higher adsorption efficiency reflecting higher surface areas. Hence, a probable relationship between the morphology and the adsorption efficiency of the ZINPs exists. This can be explained by the fact that cubes possess higher surface areas than spheres of the same volume and consequently, higher surface energies than spheres. Particles with higher surface energies have a higher tendency to form agglomerates (Seipenbusch et al. 2010). This is consistent with a study where it was found that nanoparticles with cubic morphology formed oriented aggregates, unlike their spherical counterparts ). These oriented agglomerates may also possess more active sites on the surface of the particles for the adsorption of contaminant species but will also be more prone to destabilization. The dominant quasi nanocube-shaped particles and the porous morphology of ZINPS are conspicuously shown in Fig. 4d, e, g, h, j, k, m and n. The EDX results show a progressive increase in iron content with precursor concentration as expected. However, the EDX elemental analysis of the ZINPs shows that the stoichiometric composition of Fe and O was 12.70 ± 4.26 and 37.30 ± 3.87 (by weight percent, respectively). The value is lower than the theoretical value (Fe: ~ 36%, O: ~ 42%) and this could be because theoretical values are obtained under the assumption that there are no crystalline or structural defects or loss of mass whereas such defects and loss of mass are possible during the process of synthesis, and because EDX is a semi-quantitative analysis (Sani et al. 2021). Also, the antioxidants in the plant extracts as reducing/capping agent may have contributed to the lower oxygen content of the ZINPs.

Energy-filtered transmission electron microscopy
The energy-filtered transmission electron microscopy (EFTEM) analysis further elucidated the morphology, size and distribution of the ZINPs. The morphology of the nanoparticles as shown by the EFTEM analysis results in Fig. 5 confirmed that the nanoparticles dominantly assumed a quasi nanocube in shape. This shape became more obvious with increasing metal precursor concentration. Using the Java-based image processing program (ImageJ software program), data were obtained from the edge length and diameter measurements of 125-143 particles in the size range of 19.09-97.88 nm. The average size of the particles was determined as 42.3 ± 7.2. Similar findings have been reported elsewhere (Iqbal et al. 2020). It can also be seen that the particles were distinctly departmentalized as against agglomeration (Fig. 5 b, e, h and k) due to the capping effect of the antioxidants and phytochemicals of MOL on the ZINPs (Kumar et al. 2020). The particle size distributions of the nanoparticles were fitted with a Gaussian distribution function over histograms constructed on the micrograph of each sample as shown in Fig. 5c, f, i and l. The white crystals and surroundings seen in Fig. 5d, g and j are a result of beam contrast known as "diffraction contrast" or "amplitude contrast." It is generated as long as the (parallel) electron beam is scattered at the Bragg angles by the atoms displaced in the lattice. This phenomenon can also occur when there are variations in the mass or thickness of the NPs because of the interaction with the electron beam with more materials. Similar findings were also reported by Iqbal et al. (2020).

X-ray diffraction analysis
The XRD diffractometry was used to assess the crystallinity of ZINPs, which indicates the degree of the structural order of ZINPs. The crystallinity index (CI) is used as a quantitative indicator of crystallinity. The XRD pattern shown in Fig. 6 indicates the crystalline nature of MOLmediated ZINPs. First, the XRD peak patterns were analysed, smoothened and the baseline was subtracted using Origin 8 software programmes (OriginPro 2018 SR1 version 9.5.1.195) and indexed with hkl values using X'pert Highscore software (X'pert Highscore Plus).
The diffraction peaks at 2θ value ≈31° reveal the existence of iron oxide (FeO, Fe 2 O 3 or Fe 3 O 4 ), while that at 2θ value ≈37° indicates the presence of of FeOOH (Ebrahiminezhad et al. 2018). Worth mentioning that other minor peaks in the diffractograms are also attributed to iron oxide phases (Kozma et al. 2016). The average crystallite sizes of ZINPs are subsequently calculated using the peak of highest intensity, both indexed at 2θ value of ≈20° and obtained as 15.31 nm using Scherer's approximation (Sani et al. 2021).
where K is the Scherrer constant, is the X-ray wavelength, is the full width at maximum height (FWHM) and θ is the Bragg angle. The crystallinity indexes (CI) of ZINPs of various concentrations were obtained from the average peak areas of the ZINPs and are presented in Table 2.

Brunauer-Emmett-Teller nitrogen adsorption-desorption analysis
The specific surface area, pore size, and volume distribution were studied using Brunauer-Emmett-Teller (BET) nitrogen adsorption-desorption analysis as shown in Fig. S1 Table 3. The pore sizes decreased with increasing iron precursor concentration leading to a progressive increase in the surface areas of the ZINPs. The increase in surface areas of the ZINPs with increasing iron precursor concentration is beneficial for improving the permeation of DRO species into the pores of ZINPs. The average BET surface area of ZINPs in this study was found to be 8.62 m 2 /g.

Method validation and quantitation
Linearity for DRO determination was established by plotting the peak areas of the DRO against the nominal concentration of the standards using the least-squares method (R 2 = 0.998). Quantification of peak area was performed with manual integration of the DRO as a single group area of the portion of the chromatogram with an established retention time of 15 min. The recovery results for the accuracy of the method are shown in Table S2. The excellent mean recovery for all DRO of four batches was obtained in the range of 99.75 to 101.83% with % RSD lower than 3.00 (RSD < < 20%). The precision was assessed by the reproducibility of results and determined by calculating the RPD of duplicate samples with an obtained value of 6% using Eq. (3). The acceptable RPD limit is 20% (the lower the RPD, the more precise the method is) (U.S. EPA. 2003). The response factor (RF) is a measure of the slope of the calibration relationship for a given standard curve (Fig. S2). Under ideal conditions, RF will not vary with different standard curves. However, response factors (RFs) vary with different standards in practice. When this variation, measured as the %RSD of the RFs is ≤ 20%, then the slope of the standard curves is sufficiently close and linearity is established. The mean value of the RF was found to be 1215.8 as shown in Table S3. The RF was used for quantitative determinations of sample analytes using Eq. (4). Figure S3 shows the chromatograms of standard solutions of DRO (a, b and c), and DRO after treatment with BINPs (d and e).
where C s is the calculated/predicted concentration of the DRO analyte in the sample aliquot introduced into the instrument after treatment with ZINPs, A s is the peak area of the sample analyte, A is is the peak area of ISTD and C is is the concentration of ISTD (µg/mL).
where C s is the calculated/predicted concentration of the DRO analyte in the sample aliquot introduced into the instrument after treatment with ZINPs, A s is the peak area of sample analyte, A is is the peak area of ISTD and C is is the concentration of ISTD (µg/mL). (3)

Effect of contact time
The adsorption efficiency of ZINPs was investigated under the effect of contact time between the DRO contaminant species and ZINPs nanosorbents at a constant pH and temperature of 6.5 and 30 °C, respectively using 1 g of ZINPs. Figure 7a(i) shows the DRO adsorption pathway using ZINPs. It can be seen that the uptake of DRO increased with time from 10 to 480 min and remained constant or slightly decreased (depending on the metal precursor concentration of ZINPs) after 480 min. The effect of contact time on the adsorption of DRO using ZINPs for 10, 20, 30, 40, 50, 60 and 120 min is depicted more clearly in Fig. 7a. From Fig. 7a, it can be seen that the graph gradually started to plateau after 120 min and remain so until 480 min. With contact time variations of 10,20,30,40,50,60,120,240,480 (Khan et al. 2019a, b). The DRO species move from the contaminated water (area of higher concentration) to the active pores on the surface of the adsorbent (area of lower concentration) leading to an increase in the adsorption efficiency of ZINPs. However, as time increases, the number of active sites gets saturated and adsorption efficiency remains constant at equilibrium or slightly diminishes which is in agreement with previous studies by Alkhatib et al. (2015) and Dashti et al. (2020).

Effect of dosage
The effect of ZINPs dosage on the removal of DRO from contaminated water samples was investigated using ZINPs of different dosages (0.1, 0.5, 1, 2 and 3 g) and 20 µg/mL DRO solutions contained in 150 mL of contaminated water, at a constant pH and temperature of 6.5 and 30 °C, respectively. Figure 7b shows the effect of dosage on remediation efficiency. It can be seen that the percentage of DRO removal by ZINPs increased with an increase in dosage. With ZINPs dosages of 0.1, 0.5, 1 and 2 g/L, the adsorption efficiency increased from 26.29 to 83.54%, 32.64 to 86.44%, 50.20 to 90.71% and 73.05 to 92.21% using ZINPs 0.05 ZINPs 0.1 , ZINPs 0.25 , ZINPs 0.5 , respectively. The removal efficiency increased with dosages of 0.1 to 3 g/L. However, there was either a decrease or no significant increase in adsorption of DRO by ZINPs after the application of 2 g of ZINPs although adsorption slightly increased with ZINPs 0.25 when 3 g of the ZINPs 0.25 was used. This adsorption increase is insignificant because it is less than 7%. It can be drawn that the optimum dosage of ZINPs for the removal of DRO from contaminated water was 2 g with the dosage of 2 g of ZINPs 0.5 showing the highest percentage removal of 92.21%. Increasing the dosage of ZINPs provides more active sites for the DRO to be adsorbed onto. This explains the increase in adsorption efficiency with the dosage of ZINPs. The higher adsorption capacity via using a higher ZINPs dosage could be elucidated by the increase in the number of accessible active sites for adsorption and the pattern was expected (Fundneider et al. 2021;Park et al. 2020). The situation where a further increase of the adsorbent dosage results in almost constant removal is related to the saturation of the binding sites (Feng et al. 2020).

Effect of Initial pH
The initial pH of solutions is an important factor in the removal of DRO from contaminated water using NPs because it affects the surface of the adsorbents and the diffusion of substrates in solutions. The effect of pH on the removal of DRO from contaminated water samples was investigated for 1 h within an acceptable pH range (4-10). The dosage and temperature were fixed at 1 g/L and 30 °C, respectively. Figure 7c shows the effect of pH on removal efficiency. The percentage removal of DRO by the ZINPs sharply increased from pH 4 to pH6 and thereafter remained almost constant. It can be seen that the adsorption efficiency increased from 43.75 to 58.39%, 82.55 to 88.29%, 83.12 to 89.76% and 83.27 to 91.55% for ZINPs 0.05 , ZINPs 0.1 , ZINPs 0.25 , ZINPs 0.5 , respectively at pH 4-8. The adsorption efficiencies of ZINPs were seen to either remain constant or slightly change at pH values greater than 6. The change in adsorption efficiency of ZINPs at pH values greater than 6 was less than 7%. This shows that the dissolution of metal, as well as the activity of the functional groups such as amino, carboxyl, sulphate and phosphate, present on the adsorbent, are affected by pH (Uzunoǧlu et al. 2014).
The achieved results also conform to the findings in some literature. For example, Santos et al. (2021) observed that the adsorption capacity of activated carbon modified with iron nanoparticles for adsorption of oil and grease was practically maintained with the pH increase, up to 7.0, but an additional increase in pH led to a lower adsorption capacity (Santos et al. 2021). Ercan (2019) studied the effect of pH on the adsorption of Cu 2+ using iron oxide nanoparticles synthesized from Enteromorpha spp. extract. The results showed that the adsorption capacity of the iron oxide NPs increased from pH 3 to pH 5. Kgatitsoe et al. (2019) observed that the adsorption capacity of magnetite nanoparticles functionalized with Moringa oleifera plant leaf extracts for extraction of nitroaromatic explosive compounds from an aqueous solution increased with increasing pH.

Effect of temperature
The pattern for the effect of temperature on adsorptive removal of DRO using ZINPs is shown in Fig. 7d. DRO remediation of contaminated water by ZINPs was investigated at constant pH of 6.5 for 1 h while varying the solution temperature (25, 35, 45 and 55 °C). From the diagram, the removal efficiency of ZINPs decreased with the solution temperature up to 45 °C, and thereafter, remained almost constant except for ZINPs 0.5 which showed a further slight decrease in adsorption capacity from 45 to 55 °C. The reason for this further slight decrease by ZINPs 0.5 could be due to saturation of the surface of the adsorbent with DRO substrates after 1 h, thereby causing an equilibrium shift to the area of lower concentration of DRO. The decrease in the adsorption capacity of ZINPs in the adsorption of DRO denotes that the process could be exothermic (Santos et al. 2021;Zou et al. 2017). A possible reason for the slight decrease in adsorption capacity of ZINPs for adsorption of DRO with increasing temperature is that, although temperature increase can lead to lower viscosity of the environmental media and accelerate the diffusion of substrates (DRO) to the surface of the adsorbent (ZINPs), the adsorption equilibrium possibly moved backwards as the temperature increased (Acosta et al. 2018;Hu et al. 2017

Equilibrium modelling
The well-established adsorption isotherms were used to study the relationship between the adsorbate (DRO) and adsorbent (ZINPS) at adsorption equilibrium. The linearized form of Langmuir and Freundlich isotherms as given in Eqs. (5) and (6), respectively, was applied to the experimental data at equilibrium for the determination of adsorption constants. The monolayer adsorption on the homogenous surface of the adsorbent with weak interaction between the adsorbate and adsorbent is assumed for Langmuir adsorption isotherm with a maximum monolayer coverage capacity of 7.194 mg/g. Multilayer adsorption was assumed for the application of experimental data using Freundlich isotherm Fig. 8a. Figure 8b clearly shows that the adsorption of DRO equilibrium data was described by the Langmuir adsorption model (R 2 = 0.999). Table S4 shows the absorption parameters of ZINPs while Tables S5 and S6 show the Langmuir and Freundlich isotherm parameters respectively. Figure S4 shows a graph representing the adsorption capacity of ZINPs 0.5 for DRO adsorption with contact time. The separation factor R L was assessed using Eq. (7) which indicates the adsorption possibility either as (i) favourable; (0 < R L > 1), unfavourable; ( R L > 1), linear; ( R L = 1) or irreversible; ( R L =0) (Ayawei et al. 2017;Ayub et al. 2020).
where q e is the equilibrium adsorption capacity (mg/g), q m is the maximum adsorption, C e is the equilibrium concentration (mg/L), C i is the initial concentration (mg/L), V is the working volume (L), W is the adsorbent weight (g), K L is the Langmuir constant relating to the energy of adsorption between the adsorbent and the adsorbate (L/mg) and K f is the Freundlich constant (mg/g)/(L/mg) 1/n .

Kinetic modelling
To assess the adsorption rate and to facilitate process modelling, the pseudo-first-order kinetic and pseudo-second-order kinetic models were used to correlate the adsorption kinetic data using the linearized form of the equations as given in Eqs. (9) and (10) (Unuabonah et al. 2018): 9) In(q e − q t ) = Inq e − k 1 t Dynamic equilibrium was achieved in 9 h. The adsorption kinetic parameters are presented in Table S7. From Fig. 9a and b, it can be seen that the kinetic model is well described by pseudo-second-order kinetics due to the high regression coefficient (R 2 = 0.999) and the consistency of the experimental data with the calculated data as shown in Table S7. The above-mentioned results show that the concentrations of both the nanosorbent and the DRO species were responsible for the rate-determining step during binding on the ZINPs. Although this may suggest that adsorption of DRO species on the ZINPs might be by chemisorption mechanism (Soltani et al. (10) , conclusions cannot be drawn since the adsorption mechanism cannot be based on simple fitting of pseudosecond-order model, rather, diffusion models should be investigated before such conclusions are drawn. Furthermore, the SEM images of ZINPs before and after DRO adsorption as presented in Fig. 10a, b are useful to further confirmation of DRO adsorption onto the surface of ZINPs. The surface of the ZINPs from SEM imaging before adsorption as shown in Fig. 10a appears to be smooth. However, after DRO adsorption by the ZINPs as shown in Fig. 10b, the surface of ZINPs can be seen to be rough and more agglomerated with closed pores due to the tiny particles (DRO species) adsorbed onto the surface, indicating successful adsorption of DRO onto the surface of ZINPs. Factorial design for the assessment of factor effects on the response Multifactorial analysis in a regular two-level factorial design was employed to investigate and characterize the most significant main effects and interaction effects of the independent variables on the response (adsorption efficiency of ZINPs). A 16-run experiment at different low-high factor levels (Table 4), with a signal-to-noise ratio of 2 (power = 95.3%), was conducted to study the main, interaction and higher-order interaction effects of four predictor factors; contact time, adsorbent dosage, pH and temperature, represented as A, B, C and D, respectively, on the response. The half-normal probability plot, the normal plot and the Pareto chart as shown in Fig. S5 (a-c) show the factors with statistical significance in the response. From the half-normal plot shown in Fig. S5 (c), it can be seen that the adsorbent dosage, B; contact time, A; pH; C; and the main interaction between the three factors, ABC, are on the right side of the red line, depicting that these were the chosen factors with statistically significant effects on the response with a p value of less than 0.05 (p < 0.0001) and at 95% confidence level. This is confirmed by the Pareto charts where the statistically significant explanatory factors are seen to be above the Bonferroni and t value limits. Statistical significance of the factors followed the sequence; B > A > C > ABC > D > BC > AB > AC. The adsorbent dosage (A), contact time (B) and pH (C) were obtained as the three most statistically significant main factors affecting the response for optimization and modelling. Table 5 reports the results of 20 experimental runs of the central composite design (CCD) analysis for DRO removal process. The significance of the fitted model and the variance of the predictor factors with experimental results were tested by employing a one-way analysis of variance (ANOVA). A suggested quadratic model with an effective p value (p < 0.0001 and correlation coefficient, R 2 = 0.9951) expressed the best relationship between the factors and response as shown in a coded equation represented in Eq. (11): Table 6 shows the prediction adequacy of the model while Table 7 summarizes the analysis results of each model term of the adsorption process. According to the reported results by ANOVA (Tables 6 and 7), the quadratic model and model terms are statistically significant with 95% confidence and p values less than 0.05. The coefficient of variance (CV) ratio of less than 10% is also reported for the model. ANOVA's results (R 2 ) for DRO adsorption is 0.9951 for absorption using ZINPs. It is required to have an R 2 value of more than 0.75 to guarantee the competency of a model for anticipating the experimental results (Jaafari and Yaghmaeian 2019). From Table 7, the model f value of 276.84 was obtained which implies the quadratic model obtained in Eq. (11) is significant.  The "lack of fit" f value of 2.73 obtained implies the "lack of fit" is not significant relative to the pure error which is a desirable outcome. Linear regression was used to assess the linear relationship between the predicted and actual values. The predicted correlation coefficient (R 2 ) and the adjusted correlation coefficient (Adj-R 2 ) was used to measure the model adequacy of the response, while the adequate precision (Adeq Precision) was used to measure the signal to noise ratio. The predicted R 2 of 0.9697 obtained is in reasonable agreement with the Adj-R 2 of 0.9915 (d < 0.2). Adeq-Precision of 55.866 was obtained which indicates an adequate signal and implies that the model can be used to navigate the design space.

Analysis of variance of the reduced quadratic model
Therefore, we can conclude that the reduced quadratic model fit was good and can be used to predict the response at a 95% confidence interval. The reliability of the model under the optimal conditions and experiments were performed. The experimental values were in good agreement with the predicted values under the optimal working conditions as shown in Fig. S6, which indicates that the response could be accurately predicted by the quadratic regression model obtained by RSM.
Optimization design experiment using RSM and modelling fitting Figure 11a, b shows the plots for the response surface of DRO removal by ZINPs from contaminated water. As it can be seen in Fig. 11a, DRO removal is low at low pH, low ZINPs dosage and short contact time. Improving the pH, dosage and contact time consequently improved the DRO removal by ZINPs. A higher dosage creates more surface area and hence, more active sites for the adsorption of DRO while increase in contact time allows for the interactions of the substrate with the active site of the adsorbent (Rasheed et al. 2016). A relevant study has shown that adsorption process initially increases rapidly because of the availability of the adsorbent active sites. However, this slows down after some time because of the saturation of the active sites on the adsorbent (Dehghani et al. 2020). The ramps in Fig. 11c-f further depict the optimum conditions which were found at a pH of 8, contact time of 8 h and adsorbent dosage of 2 g, giving a maximum adsorption efficiency of 92.6%. The desirability of the model is equal to 1 ( Table 7). The best experiment results and the most optimum condition of each solution could be customized for further validation when desirability is 1 or close to 1 (Bandara et al. 2019). Figure 11 also shows that the prediction of model responses is in close agreement with the results from the experiment. Experiments were conducted at optimum conditions to validate the model and the response was in agreement with the proposed model.

Adsorbent reusability study
Recycling of the adsorbent for reusability is paramount as a cost-effective process in wastewater treatment. A reusability study was conducted in a five-cycle adsorption/ desorption experiments using 1 g of ZINPs 0.5 at pH and temperature of 6.5 and 30 °C, respectively, to demonstrate the reusability and stability of ZINPs for the removal of DRO. The eluent for the regeneration of the adsorbent was HCl. ZINPs 0.5 was selected for the study because it has the best adsorption efficiency of the synthesized nanosorbents, and HCl was used as the preferred eluent due to the smaller size of chloride ions (Cl − ) compared to other ions reported in literature such as NO 3 − and SO 4 2− (Naushad et al. 2019). Figure 12 shows the cycles of adsorption/ desorption of ZINPs 0.5 . The adsorption efficiency or the percentage pf adsorption remained fairly at 74% after the fourth and fifth cycles. The slight reduction in the adsorption efficiency could be due to the loss of the ZINPs during the washing and the recovery process. This shows that the

Conclusion
In this study, ZINPs were synthesized using Moringa oleifera extract and was used sorbent material for diesel range organics (DRO). The fabrication of the ZINPS is both costeffective and environmentally friendly as Moringa oleifera leaf extracts used as the capping/stabilization agent and the iron precursor utilized are cheap, non-toxic and readily available. The characterization revealed the formation of nanospheres and quasi nanocubes with an average particle diameter of 50.9 ± 9.7 nm and the morphology become more obvious with increasing iron precursor concentration. The crystallite size as revealed by XRD analysis showed ZINPs of an average size of 15.31 nm with a crystallinity index of 32.47%. The adsorption of DRO by ZINPs was best described by Langmuir isotherm with a maximum monolayer coverage capacity of 7.194 mg/g. The separation factor R L =0.472 are indicating favourable DRO adsorption and the adsorption kinetic data were best fitted to the pseudosecond-order kinetic model. The adsorption of DRO by the ZINPs increased with contact time, adsorbent dosage and pH. The synthesized ZINPs showed more than 92.6% DRO removal at optimum conditions, thus making them excellent alternative nanosorbents for the removal of DRO from contaminated water.