Methods for preparation of aqueous extracts of Talinum triangulare have been previously reported [39]. Matured dry fruits of Hura crepitans were harvested at Uyo Metropolis, Akwa Ibom State, Nigeria. The dry fruits were mechanically broken to remove the seeds. The seeds were air-dried and ground using a manual grinder. The Hura crepitans seed oil was extracted from the ground seeds through solvent extraction using petroleum ether as solvent at a temperature range of 60 to 80°C. Exactly, 500 g of the sample was soaked in petroleum ether, squashed in a borosilicate container, and filtered in order to separate the petroleum ether fraction. The filtrate was kept overnight in a fume cupboard, the settled debris at the bottom was removed by decantation. The filtrate was kept in an oven at the temperature of 70°C to get rid of petroleum ether leaving behind yellow-coloured oil with a pH of 5. Plantain peels were locally sourced and washed with running water to remove dirt, sun-dried for three (3) weeks and burnt to ash using a muffle furnace. The ash formed was cooled and stored in a sealed plastic container. About 500 g of the ash was added to a clean plastic container and 1,000 mL of deionised water added, stirred for 30 min and kept overnight. The resultant solution was filtered using Pall water filtration apparatus with 47 mm, 5-micron glass micro-filter membrane into a clean bottle using a vacuum pump. A pale brown filtrate with pH of 13 was collected and stored in a 250 mL volumetric flask. About 250 mL of the ash solution was added to a stainless-steel pan and boiled to 100°C and then 62 g of the Hura Crepitans seed oil was added and stirred thoroughly. The mixture foamed signifying saponification; the temperature was reduced to 60°C to control the foaming. Stirring continued until the foaming reduced leaving behind a grey soapy viscous liquid.

The soil washing experiments were done according to modified methods from the literature [4, 23, 40]. Soil samples were collected from an urban agricultural soil in Uyo, Nigeria. The air-dried homogenized soil samples (approx. 20 g) were added into 30-mL borosilicate round-bottom glass centrifuge tubes. Naphthalene and phenanthrene (AR, 99.9%) were purchased from British Drug House (now Merck, UK), while Tween 80 (extra pure grade) was acquired from Loba Chemie PVT. Ltd (Mumbai, India). Stock solutions of PAHs (naphthalene and phenanthrene) were prepared in volumetric flasks by dissolving appropriate amounts in methanol and stirred using a magnetic stirrer. A measured volume (15 mL) of the PAH stock solution was then added into the centrifuge tubes containing the thoroughly mixed soil samples such that the medium would be saturated. The spiking of the stock solution (concentrations: 0, 25, 50, 75, 100, and 125 mg/L) was done such that on evaporation, it would give a theoretical concentration in g/Kg of soil (0, 0.5, 1.0, 1.5, 2.0, 2.5 mg PAH/g soil). The mixture was then kept for 72 h to allow complete evaporation of methanol. Exactly, 20 mL of surfactant solutions prepared at appropriate concentrations were then added to the soil mixture. For aqueous extracts of waterleaf, a volume-by-volume aqueous concentrations were used (10, 20, 30, 40, and 50% vol/vol). Surfactant solution from saponified Hura Crepitans seed oil prepared using deionised water at different concentrations (5, 10, 15, 20, 30 mg/L). To compare the performance of bio-derived surfactants with a commercial surfactant, Tween 80 was used. Different concentrations (10, 20, 30, 40, 50 mg/L) of Tween 80 was prepared with deionised water from around its critical micellar concentration (CMC) which has been previously determined to be 13.10 mg/L according to Zhang and Zhu [41]. The amounts of soil and the surfactants were chosen so that the glass tubes would be filled up, and then sealed with Teflon-lined screw caps to avoid any loss. The mixtures were then agitated for required period in a thermostated shaker bath maintained at a speed of 120 rpm at room temperature (25°C). Thereafter, the samples were centrifuged at 12,000 g for 15 min. Aliquots (0.5 mL) of the supernatant were withdrawn using syringes, dissolved in 1.5 mL methanol, filtered through 0.22 μm PTFE membrane filters and placed in 1.5 mL vials for subsequent analysis.

The concentration of naphthalene and phenanthrene were determined using high performance liquid chromatograph (HPLC) (Shimadzu LC-20AT model) with an ultraviolet (UV) detector and a 5 μm 4.6 × 250 mm Agilent C18 reversed phase column. The following conditions were applied: flow rate of 1.0 mL/min, injection volume of 15 μL, under isocratic flow condition with acetonitrile and water (80:20) as mobile phase. The temperature was maintained at 30°C. The UV wavelength was fixed at 254 nm. The HPLC-UV method was a modification of a suitable method in the literature [42]. Blanks were prepared for all experiments. All aqueous solutions were prepared using deionised water. Experiments were carried out in triplicates.

The HPLC-UV analysis of the standards, the retention times recorded were 4.8 min and 8.2 min for naphthalene and phenanthrene, respectively (Fig. 1). The R² from the standard calibration curves for both naphthalene and phenanthrene were above 0.99 (Fig. S1).

The efficiency of the removal of PAHs (R) was estimated by determining the ratio of the concentration of PAHs in the supernatant solution (C) and the initial concentration of PAHs in the soil (C_i) according to Eq. (1) [43]:

The prepared surfactant solutions without further purification except filtration, were characterised for functional group determination using Fourier-transform infrared (FTIR) spectroscopy analysis. The FTIR spectra were recorded using Thermo Scientific Nicolet iS5 spectrometer (Thermo Scientific, Waltham, USA).

Physicochemical parameters of soils and trace metals analyses were carried out using standard analytical protocols according to the American Public Health Association [44]. Physicochemical parameters determined include pH, electrical conductivity, total organic carbon (TOC), cation exchange capacity (CEC), and particle size. Trace metals (Cd, Pb, Zn) were determined after nitric/hydrochloric acid digestion using an atomic absorption spectrometer (AA500 Atomic Absorption, PG Instruments Ltd, USA). The trace metals standards were purchased from Accustandard (New Haven, USA). The R² for the standard calibration curves of the trace metals were greater than or equal to 0.995.

The Physicochemical characteristics of the soil used for the present study are presented in Table 1. As expected, the soil is generally classed as ultisols which are characterised by low nutrient and common in tropical regions in both Nigeria and China [3, 45, 46]. The soil belongs to the sandy textural class.

Functional group characterization of the prepared bio-derived surfactants is essential to provide insight into their possible chemical properties and behavior. The FTIR spectra of the aqueous extract of Talinum triangulare and saponified Hura crepitans seed oil are presented in Fig. 2. For the aqueous extract of Talinum triangulare (AET), the following key absorption peaks recorded were: 3362.1 cm⁻¹ (O-H bond from alcohols), 1,621.4 cm⁻¹ (C=C from alkenes), 1438.8 cm⁻¹ (C=C from aromatic compounds), 1,304.6 cm⁻¹ (O=C-O-C from aromatic esters), 1084.7 cm⁻¹ (C-O from alcohols or esters), 1,036.2 cm⁻¹ (=C-O-C from aromatic ethers), 898.3 cm⁻¹ (C-H from geminal disubstituted alkenes), and 790.2 cm⁻¹ (C-H from trisubstituted alkenes). On the other hand, the soap prepared from saponification of oil from seeds of Hura crepitans (HCS) recorded the following major absorption peaks: 3,473.9 cm⁻¹ (N-H from amides), 3,011.7 cm⁻¹ (C-H from aromatic compounds), 2,922.2 cm⁻¹, 2,855.1 cm-1, 1,461.1 cm⁻¹, 1,379.1 cm⁻¹, 723.1 cm⁻¹ (C-H from alkanes), 1,744.4 cm⁻¹ (C=O from esters or ketones), 1,237.5 cm⁻¹ (=C-O-C from aromatic ethers), 1,159.2 cm⁻¹ (C-O-C from ethers), 1,095.8 cm⁻¹ (C-O from ethers or alcohols) and 913.2 cm⁻¹ (=C-H from alkenes). Generally, based on the FTIR spectra of the two surfactants, AET recorded more peaks assigned to heteroatomic functional groups than HCS. HCS contained more saturated functional moieties than AET. One major explanation for the differences is that AET is aqueous-based, while HCS is oil-based. Surfactants with multiple heteroatom functionalities usually show superior surface activity [47–50]. In addition, previous reports recorded similar absorption peaks for aqueous extracts of Talinum triangulare [14, 51].

The effect of washing time was determined and the results are presented in Fig. 3. Generally, removal efficiencies tend to increase with increase in washing time. HCS recorded very low removal efficiencies (< 30%), while AET and Tween 80 recorded approximately 70–96% (Fig. 3). For HCS, as it can be observed in Fig. 3, there was no clear difference between the removal efficiencies of naphthalene and phenanthrene. However, AET tends to remove phenanthrene faster, although naphthalene recorded higher removal efficiency. Given the plateau nature of the graph for Tween 80 data, it seems that equilibrium was achieved earlier due to the fact that naturally derived surfactants possess lower CMCs compared to synthetic surfactants [52]. On the other hand, little changes are observed for the removal efficiencies of HCS over time. This implies that the performance of HCS is not influenced by time. As observed in a previous study, the efficiency of removing contaminants by soil washing increases with increase in washing time, but is later stabilized [7]. According to the study, the time-dependent relationship is caused by rate-limited dissolution and desorption of contaminants.

In an actual contaminated site, it may be necessary to determine the amount of remediant materials required for a certain level of contamination. We measured the removal efficiencies of different concentrations of naphthalene and phenanthrene in soils and the results are presented in Fig. 4. Generally, more contaminants led to higher removal efficiencies. Although beyond 100 mg/Kg the graphs flattened, the result is unusual since the same concentration was used. As explained in a previous study, certain amounts of contaminants could partition into the soil matrix, such that the amount removed depends on the partitioning of the overall concentration of the contaminant at any given time [22]. More so, the present soil used in this study recorded TOC > 0.30 (Table 1). Usually, for soils with organic content > 0.1, PAHs get partitioned into the soil organic matrix [53]. This observation was pronounced in AET and Tween 80 which showed high reactivity during the study. Therefore, to ensure complete removal, it is recommended to carry out soil washing more than once [18]. Mostly, naphthalene, which is of a lower molecular weight, was removed with higher removal efficiencies compared to phenanthrene. This may be due to easier diffusion. The adsorption data was fitted with first and second order pseudo-kinetic models and the results are presented in Fig. S2. From the results, we observed that the pseudo second order show better fitting with R² > 0.98 for both naphthalene and phenanthrene in all applied surface-active materials. This implies that PAHs-soil-surfactant interactions were beyond physisorption and could be better classified as chemisorption process.

Different surfactant concentrations were used to test the removal efficiencies after 15 min washing time. We observed that the removal efficiencies increased with surfactant concentrations (Fig. 5). In this case, phenanthrene recorded higher removal efficiencies compared to naphthalene. One explanation to this observation could be that phenanthrene with higher octanol-water partition coefficient (log K_ow = 4.46) than that of naphthalene (log K_ow 3.36) will likely dissolve more in organic phase of the surfactant solutions and gets easily dissolved while some part of naphthalene remains in the aqueous phase [54]. Another explanation is that appreciable amounts of naphthalene could have been lost by evaporation since it is lighter than phenanthrene such that the amounts available for removal was less [55]. Again, higher removal efficiencies were recorded by Tween 80, followed by AET, and lastly HCS. The results demonstrate the superior performance of AET which has been characterized to contain saponins. This implies that saponin-based surfactants may be more suitable for removal of PAHs from contaminated soils compared to oil-based soaps. In a previous study, saponin-based surfactant has been shown to compete favorably with Tween 80 in soil washing of phenanthrene [18]. Other researchers have reported the use of oil alone from plant substrates to achieve removal efficiencies above 80% [53, 56]. When soybean oil was used for removal of anthracene from contaminated soil, it was reported that removal efficiencies decreased with increase in initial anthracene concentration [57]. It seems therefore, that preparation of the soap further introduces additional materials that may lead to high viscosity of the materials and consequent clogging. Furthermore, previous studies have established that light permeability, hydrogen donor capability of soil washing agents significantly influenced their performance for removal of organic pollutants from soils [58]. In this study, AET is more transparent than HCS, which may permit photo-induced attenuation of the contaminants.

The trace metals characteristics of the soil before and after surfactant washing are presented in Table 2. The choice of the three trace metals was based on their usual occurrence in urban agricultural soils. With the exception of zinc, it can be observed from the results that cadmium and lead exceeded the World Health Organization (WHO) allowable levels in agricultural soils (Table 2). Since it is an urban area, the source of contamination may be attributed to anthropogenic activities such as transport-related emissions and atmospheric deposition from industrial processes [59–61]. The soil washing experiments significantly affected the distribution of trace metals levels. A cursory visualization of the data reveals that after soil washing, the cadmium levels significantly reduced in all treatments. In the case of lead (Pb) and zinc levels, only treatments with AET and Tween 80 recorded significantly lower values. Generally, the levels of trace metals were lowest in soils treated with AET, followed by Tween 80 and HCS. Tween 80 is a hydrophilic nonionic surfactant that has been used mostly for remediation of organic pollutants [62]. The nonionic nature of Tween 80 may have limited its reactivity with the negatively charged soil matrix [63]. As such, the acidic nature of the soil (Table 1) may not have changed significantly which allowed a significant amount of the trace metals to remain mobile in the soil. On the other hand, AET with a lot of heteroatomic moieties as established from the FTIR spectra (Fig. 2) provides a lot of adsorption sites for metal adsorption and subsequent extraction [64]. Also, it has been established from previous research that AET demonstrates acidic soil pH moderation which may result in immobilization of the trace metals [39]. For HCS, clogging of the soil matrix and weak reactivity may have accounted for the low metal extraction.