Wastewater biotreatment and bioaugmentation for remediation of contaminated sites at an oil recycling plant

AHs

aliphatic hydrocarbons

BA

bioaugmentation

BOD₅

biological oxygen demand

CFU

colony forming units

COD

chemical oxygen demand

Cr

chromium

CSTR

continuous stirred tank reactor

Cu

copper

DGGE

denaturing gradient gel electrophoresis

EC

electrical conductivity

Fe

iron

FTIR-ATR

Fourier transform infrared spectroscopy – attenuated total reflectance

GC/MS

gas chromatography-mass spectrometry

GI

germination index

HRT

hydraulic retention time

LC-MS/MS

liquid chromatography-mass spectrometry

NA

natural attenuation

Ni

nickel

OLR

organic load rates

PAHs

polycyclic aromatic hydrocarbons

Pb

lead

PCR

polymerase chain reaction

Ph

phytane

Pr

pristane

rRNA

amplified 16S ribosomal ribonucleic acid

TF

trickling filter

TKN

total Kjeldahl nitrogen

TN

total nitrogen

TP

total phosphorus

TPH

total petroleum hydrocarbons

TSS

total suspended solids

UCM

unresolved complex mixture

Zn

zinc

Used oils are all oils obtained from crude oil or blends, including synthetic oils: used motor, gear, compressor, hydraulic, turbine, transformer, and heat transfer oils (Xiaojie et al. 2022). In Tunisia, the main source is used motor oil. It is collected from petrol stations, company fleets and specialized waste oil collection facilities. Used oil contains products of combustion, corrosion and wear, residual fuels, additives, and other chemicals (Lam et al. 2016; Adetunji & Olaniran 2021). Used oil recycling facilities provide a second life for the recycled oil, reduce the amount of hazardous waste, and improve the overall environmental condition (Kajdas 2000; Li et al. 2021). However, the collection, transport and especially the refining of used oil generate a wide range of waste. These include oily tank sludge and highly contaminated water. In the daily process of refining used oil, the petrochemical waste generated poses significant risks to human health and causes significant environmental problems and economic impacts such as ground and surface water contamination, soil degradation, seed germination inhibition, plant growth disruption, and air quality degradation (Wake 2005; Varjani & Upasani 2017; Zeneli et al. 2019; Ossai et al. 2020; Gao et al. 2022). Varieties of physicochemical techniques are available for the treatment of petrochemical effluents (adsorption, coagulation, ultrafiltration, advanced oxidation, and membrane separation) and for the remediation of contaminated soils (chemical oxidation, stabilization, disposal, containment, and thermal desorption). Some of these processes treat a wide range of pollutants and provide relatively quick solutions, but they are expensive, energy-demanding, technically challenging and have high environmental and social impacts (Kumar et al. 2022).

As part of sustainable development, many industries prefer simple, cost-effective, and environmentally friendly remediation methods (Banerjee et al. 2016; Kasztelan 2017; Velvizhi et al. 2020). The bioremediation technologies appear to be an attractive solution for used oil refineries in Tunisia, as they are easy to implement, require little infrastructure and equipment, and effectively reduce hydrocarbon concentrations (Gargouri et al. 2011; Jemli et al. 2017; Karray et al. 2020). Natural attenuation (NA) is the simplest bioremediation technique. It relies on indigenous microbial metabolism (Roldan et al. 2010). Bioaugmentation (BA), also known as seeding or inoculation, consists of introducing exogenous microorganisms capable of enhancing the degradation of pollutants in contaminated environments (Moussavi et al. 2015; Velvizhi et al. 2020; Martínez-Rabelo et al. 2023). There are several parameters that can hinder the efficiency of BA processes, especially in the field, such as operating conditions, competition with indigenous microorganisms, unpredictability of treatment, and nutrient deficiency (Nzila et al. 2016; Marchand et al. 2017; Curiel-Alegre et al. 2022). Hence, the present study was carried out in a waste oil recycling plant in Bizerte, Tunisia. Processed water from the refinery unit, referred to as stripped wastewater, was tested in a continuous stirred tank reactor (CSTR), which was operated continuously for 260 days. The wastewater collected from the whole industry, known as the collected wastewater, was treated in a trickling filter (TF). An integrated approach was then developed to valorize the TF effluent by using it as a feedstock for BA of contaminated soil obtained from the used oil recycling plant. The microorganisms in the TF effluent have no natural habitat in soil, but it is quite possible that they continue to exert their purifying power when first introduced into soil. The abundance and activity of the dominant soil microbial populations were assessed using a molecular technique, denaturing gradient gel electrophoresis (DGGE) of polymerase chain reaction (PCR)-amplified 16S ribosomal ribonucleic acid (rRNA). The efficiency of bioremediation in reducing hydrocarbon contaminants was investigated by gas chromatography-mass spectrometry (GC/MS).

The entire experimental study lasted for 260 days and involved the biotreatment of industrial wastewater. The initial 90 days were devoted to the start-up and stability of the CSTR, which was conducted to reduce toxicity of the influent. For the next 170 days, the CSTR and TF were operated simultaneously to treat the stripped and collected wastewater, respectively. They were inoculated with an appropriate bacterial consortium and operated under aerobic conditions. The stripped wastewater from the refinery process first flows from the 40 m³ primary settler to the CSTR with a usable volume of 20 m³ and then flows to a secondary settler. The CSTR was run continuously at a constant temperature of 30 °C with a dissolved oxygen level of about 4–6 mg L⁻¹, and with a pump to ensure return of sludge from the secondary settler.

A 105 m³ separator (APIS) receives CSTR effluent, steam condensate, lubricants, rinsing water, and rainwater. After the gravity decantation in the APIS, the collected wastewater is treated in the 40 m³ TF (14 m³ liquid volume + filter bed material consisting of plastic material having a specific surface of 130 m²/m³), coupled with a static decanter for potential effluent storage.

The soil bioremediation took place for 24 weeks using two different water sources, tap water and TF effluent with the NA and BA strategies, respectively. The TF effluent was used to introduce the exogenous well-adapted microorganisms. Their concentration was estimated to be 2 × 10⁶ colony forming units (CFU) mL⁻¹ (Table 1). Almost 10,000 kg of soil contaminated by the used oil recycling plant was excavated and divided into two similar plots of 3 × 2 m and 0.4 m depth. Plot NA, which contained only the indigenous bacteria, was considered the control plot. The second plot, BA, contained TF exogenous bacteria in addition to the indigenous ones. The soil was mechanically rototilled twice a week to improve homogenization and aeration. Meanwhile, humidity was controlled and maintained at 18–20% and daily temperatures were maintained around 22–36 °C.

Wastewater sampling was carried out at selected points of the used oil recycling plant as the CSTR and TF influent and effluent; manual grab samples were used. Soils were analysed using a pooled sample. This consisted of five samples of 200 g each, selected at random. The frequency of the wastewater and soil sampling (weekly, daily, or less) depended on the type of analysis required (chemical or biological). Samples were kept in triplicate at 4 °C for the assay. Only the post-centrifugation pellets were maintained at −80 °C for further use in PCR.

The effluent characteristics of total suspended solids (TSS), chemical oxygen demand (COD), biological oxygen demand (BOD₅), pH, electrical conductivity (EC), phenolic content, and biomass were determined as reported earlier by Jemli et al. (2017). The BOD₅ was determined by the manometric method with a respirometer without inhibition of nitrification according to the standard method (Clesceri et al. 1998). Total petroleum hydrocarbons (TPHs) were determined using weight loss gravimetry. The Mishra et al. (2001) method was applied to the effluent. Total phosphorus (TP), total nitrogen (TN), total Kjeldahl nitrogen (TKN), and organic matter (OM) were analysed according to standard methods (Schinner 1996). The concentrations of heavy metals were measured by flame atomic absorption spectrometry in accordance with US EPA 3050B (U. S. EPA 1996).

The phenolic composition of soil extracts was determined following the Keskes et al. (2017) method. Liquid chromatography-mass spectrometry (LC-MS/MS) analysis of the soil methanolic extract was performed using an Agilent 1100 LC system. Spectra between m/z 50 and 1,200 were taken in negative and positive ionization modes. Fourier transform infrared spectroscopy – attenuated total reflectance (FTIR-ATR) soil analysis was carried out with a Bruker Vector 22 mid-infra red (IR) spectrometer. The wave numbers were scanned from 4,000 to 400 cm⁻¹ (Gargouri et al. 2011). Soil phytotoxicity was evaluated by measuring the germination index (GI) with the method of Zucconi et al. (1981) on seeds of Lepidium sativum. Soil heterotrophic bacterial and yeast viability was determined using the dilution plate count technique following Gargouri et al. (2014), and results were reported as CFU per gram of dry soil.

The extraction, purification and GC/MS characterization of the aliphatic hydrocarbons (AHs) and polycyclic aromatic hydrocarbons (PAHs) fractions were detailed by Jemli et al. (2017) and Zaghden et al. (2017). GC/MS analysis was done on an Agilent model 6890N chromatograph fitted with a Hewlett-Packard HP-5 capillary column (internal diameter, 250 μm; layer thickness, 0.25 μm; length, 30 m). The traversing gas was helium. The flow rate was 1 mL/min. Calibration standards for AHs and PAHs were used for peak identification and quantification. The chromatographic area was correlated with the molar concentrations of these compounds according to their response factors and retention times.

Total genomic desoxyribose nucleic acid (DNA) was isolated with the use of the EZ-10 Spin Column Soil DNA Mini-Preps Kit (BIO BASIC INC.). PCR products were revealed on 1% agarose gels stained with ethidium bromide (EtBr) after amplification of the V3 hypervariable regions of 16S rRNA sequences using a bacterial-specific primer set (341FGC/518R). The predominant DGGE bands were excised from the gels. After their elution for one night in 35 mL MilliQ water, they were reamplified by PCR employing primers without the GC clamp. The PCR products were ligated into the pGEMT-Easy vector (Invitrogen) as recommended by the manufacturer. The EZ-10 Spin Column Plasmid DNA Kit was used to isolate recombinant plasmids. Positive transformants were selected for sequencing using the Big Dye^® terminator cycle organic loading rate (OLR) Sequencing Kit and an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems). Initially, the 16S rRNA gene sequence was searched against GenBank and RDP using Basic Local Alignment Search Tool (Altschul et al. 1997) and Seqmatch (Ribosomal Database project II; version 10) (Cole et al. 2009). Potential chimeric structures were identified using the DECIPHER databank (http://decipher.cee.wisc.edu/FindChimeras.html) (Wright et al. 2012).

The CSTR and the TF were operated simultaneously for 170 days at OLRs of about 1,200 and 3,000 g COD m⁻³ day⁻¹, respectively. The hydraulic retention time (HRT) of the TF was about 17 h, which is seven times shorter than that of the CSTR. The average performance of CSTR and TF is summarized in Table 2. The stripped wastewater from the refinery process was more polluted than the collected effluent due to higher COD, TPH, phenol, and metal contents. The BOD5 was approximately 28% of the COD indicating a low biodegradability of the stripped and collected wastewater. The treatment with the TF resulted in a COD and BOD removal of almost 89.5 and 81% to levels of 220 and 90 mg L⁻¹, respectively, and in a phenol and hydrocarbons (HCs) removal of 91.25 and 94%, respectively. EC and pH were still within the discharge limits and TKN, TSS, and P were removed above 60.41%. Iron, lead, and zinc (Fe, Pb, and Zn) were also detected at low concentrations ranging from 0.6 to 5.48 and 0.44 to 0.63 mg L⁻¹, in the stripped and collected wastewater, respectively. It should be highlighted that the effluent concentrations and removal efficiencies for the CSTR and TF, respectively, are not directly comparable as the wastewater influent sources are different. Many researchers have found that bacteria are able to eliminate heavy metals from wastewater via their cell wall functional groups, including ketone, carboxyl groups and aldehyde (Qu et al. 2014). The treatment by CSTR reduced Fe, Zn, and Pb by about 31.75, 75.64, and 49.43%, respectively, while the effluent from the TF was free of metals. According to the results (Table 2), both processes were efficient in the removal of AHs. Although the TF seems to perform better. The residual concentration of AHs in the effluents of the TF and CSTR was about 4 and 21mg L⁻¹, respectively. Similarly, the removal efficiency of aliphatic compounds from C20 to C29 was in the range of 90.98–97.80% compared to 83.71–89.98% in the CSTR. In fact, the biomass at the level of the TF is tightly immobilized on the support surface. This protects it from shocks caused by sudden changes in the organic load and allows it to adapt better (Maurício et al. 2013; Zhao et al. 2013).

The long-chain aliphatic compounds were removed to a lesser extent than the short-chain compounds, showing an opposite relationship between degradation fraction and molecular weight. The degradation of the C26–C30 alkanes was about 85% and reached 100% for the C13 and C14. In addition, the C30 triacontane (the longest chain) showed the lowest removal rate of 76.78% with the CSTR. In line with the results of earlier studies (Sabaté et al. 2004; Kiamarsi et al. 2019), total n-alkanes (TNAs) are easily biodegraded, especially those with low molecular weight.

In addition, nutrient concentrations, TKN and TP, decreased similarly by CSTR and TF, probably due to microorganisms consuming nutrients for their active growth and metabolism (Table 2).

The efficiency of removing hydrocarbons from the soil was tested by measuring the total concentration of hydrocarbons, including their fractions, throughout the bioremediation process. The findings show that the BA strategy resulted in a reduction of 82.23% in TPH at week 24, reaching 9.70 g kg⁻¹ against 34.90% and 36.08 g kg⁻¹ for the NA strategy (Table 3). The TNA removal rates at weeks 8 and 24 were 58.17 and 81.74%, respectively, compared to 27.98 and 51.5% for the NA strategy. Thus, almost 24 weeks of NA were required to achieve a similar level of TNA removal as that achieved after 8 weeks of BA.

Metal concentrations in the soil were higher than those measured in the stripped wastewater, especially Zn and Pb, which were two and four times higher. Copper, nickel, and chromium (Cu, Ni, and Cr) were also present in the soil, but at lower concentrations than those of Fe, Zn, and Pb. Notably, the initial levels of heavy metals are low, without reaching the risk threshold concentrations in soils (Bisone et al. 2013). The metal concentration remained almost unchanged in the bioaugmented soil (Table 3), except for Cu and Zn, whose concentrations decreased by 31.13 and 17.21%, respectively.

Overall, the biodegradation of TPH in soil suggests a two-step evolution, as described by Yan et al. (2023). Nearly 60–70% of the degradation rate occurred in the first 8 weeks. The next 14 weeks showed a slower degradation, which is explained by the reduced degradability of the residual petroleum hydrocarbons as well as the lower metabolic capacities of the microorganism population (Wu et al. 2016; Abena et al. 2019). In addition, adsorption of hydrocarbons at inaccessible sites reduced their bioavailability for microbial degradation (Amir et al. 2005). On another note, the use of TF effluent improved the HC removal rate in soil, confirming the efficacy of the site indigenous microorganisms and supporting the closed-loop BA strategy. Pelz et al. (1999) and Curiel-Alegre et al. (2022) demonstrated that soil bacterial biodiversity is important for complex enzymatic responses in the decomposition of contaminants and their various intermediates. In the absence of nutrient limitation, NA is reported to be an effective bioremediation strategy (Makadia et al. 2011). For soil aromatics (Figure 3(b)), five of the 16 USEPA focus PAHs (Manoli et al. 2000; Yan et al. 2009) were less than the 0.01 μg kg⁻¹ detection level. The remaining 11 PAHs, which are an amalgam of high molecular weight (HMW) and low molecular weight (LMW), were assessed: benzo(g,h,i)perylene, six rings; benzo(b)fluoranthene (B[b]F), benzo(k)fluoranthene (B[k]F) and benzo(a)pyrene, five rings; fluoranthene (FLTH) and pyrene (PYR), four rings; acenaphthene, fluorine, anthracene (and phenanthrene (PHEN), three rings and naphthalene, two rings. It appears that the LMW PAHs (two and three rings) were more easily degraded than the HMW PAHs (five and six rings), especially at the start of treatment. In fact, at week 8, LMW PAHs depletion was much higher (61.62%) than HMW PAHs (38.75%) with the NA strategy. In parallel, the depletion rates of LMW and HMW PAHs with the BA strategy were 92.82 and 79.55%, respectively. The increase in the duration of treatment from 8 to 24 weeks improved the removal rates of total PAHs from 47.55 to 85.26% and from 81.38 to 96.81% for the NA and BA strategies, respectively.

HMW PAHs are known to exhibit greater carcinogenicity, mutagenicity and teratogenicity than LMW PAHs (Tiwari et al. 2016). Benzo[a]pyrene appears to be a dangerous pollutant as it is highly carcinogenic (IARC 2012; U. S. EPA 2017). Phenanthrene was known to be a photosensitiser to human skin and a bacterial mutagen (Samanta et al. 2002). The ultimate concentration of these organic contaminations in the soil depended on the treatment strategy used. At week 24, the NA strategy resulted in the partial removal of PHEN (84.25%) and benzo[a]pyrene (70.46%). This was enhanced by the BA strategy, with removal rates reaching 100 and 94.75%, respectively. The results regarding the higher removal of PAHs compared to AHs (Table 3) were similar to those of Varjani et al. (2017) and contrasted with previous studies by Oudot et al. (1998). Liu et al. (2011) highlighted that soil microbial communities prefer to remove some hydrocarbons. In addition, some AHs and PAHs can be evaporated or adsorbed on soil particles and lost through leaching or degradation (Okere et al. 2012; Zafra & Cortés-Espinosa 2015). Highly degradable PAHs are positively associated with decreased molecular weight, hydrophobicity, and lipophilicity, as well as increased water solubility and vapour pressure (Lu et al. 2011; Louvado et al. 2015; Patel et al. 2020). Indeed, a wide range of abiotic parameters (climate, substrate, amount, molecular structure, and type of hydrocarbon) can influence the behaviour of PAHs in soil. The critical role of solubility and dissolution rates in the bioavailability of aromatic compounds has been highlighted by Varjani et al. (2020). For example, the degradation efficiency of some hydrocarbons was lower in 20% soil organic matter (SOM) than in 2% SOM when SOM was dissolved in the soil water system (Chen et al. 2019). For comparison, the decomposition rates of some hydrocarbons were lower with 20% SOM than with 2% SOM when SOM was dissolved in soil solution (Chen et al. 2019). Of the 11 PAHs monitored in the soil in this study, B[k]F was the most recalcitrant. In addition, PAHs with the same molecular weight showed different degradation rates. Anthracene, PYR and B[k]F appeared more recalcitrant than PHEN, FLTH, and B[b]F, with degradation reduction estimated at 10 ± 2% at week 8.

The monitoring of soil physicochemical characteristics with the NA strategy showed that OM and TN decreased to 11.04 and 0.078%, indicating about 10.17 and 21.55% reduction, respectively (Table 5).

There is global evidence that pollutant biodegradation requires nutrients (Jiang et al. 2016). With the BA strategy, OM and TN were stable or slightly decreased (Table 5). This is probably due to the TN, TP, and BOD concentrations in the TF effluent, estimated at about 49, 5, and 90 mg L⁻¹, respectively. Chen et al. (2020) indicated that increasing SOM content from 2 to 10% improved soil biomass from around 10 to 102 CFUg⁻¹ and hence TPH degradation efficiency. The evolution of T, pH, and EC were similar for the NA and BA strategies. The pH was relatively alkaline at the beginning of the experiment, about 8.4, and then decreased to about 7.6 in week 24. This change is probably related to the release of protons formed when hydrocarbons decompose in the soil, as mentioned in a preceding study (Tang et al. 2012).

The increase in soil nutrient and OM content enhanced the growth of indigenous soil microbes and improved microbial mineralization (Liu et al. 2013; Suja et al. 2014). With the NA strategy, the number of bacteria increased from 2 × 10⁶ to 3.1 × 10⁶ at week 8 and then decreased to 1.3 × 10⁶ CFU g⁻¹ at week 24. The increase in the number of native soil microorganisms, without any amendments, underlines the value of tillage watering and mixing. In contrast, the BA strategy resulted in the soil bacterial community recovering by week 16, likely by creating suitable new conditions. Regular microbial inoculation helped to maintain appropriate levels of soil microbial activity throughout the treatment period. By the end of the study, the bioaugmented soil contained 3.7 times more bacteria than were present at the beginning of the experiment.

The monitoring of soil biomass revealed that:

• (a) A positive correlation between biomass abundance and efficiency degradation. This was consistent with the results of prior researchers (Mishra et al. 2001; Varjani et al. 2017).

• (b) A successful colonization of the soil by bacteria in the TF effluent. The introduced bacteria were not only able to withstand the stress of competing with native soil bacteria but also cooperated with them in removing hydrocarbons.

• (c) An increase in the native bacterial abundance is due to improved environmental conditions.

Most DGGE bands were closely associated with the genera Mycobacterium, Proteiniphilum, Nocardioides, Luteimicrobium, Azospirillum, Bradyrhizobium, Sphingomonas, Algoriphagus, Rhodococcus, Paraoerskovia, Sphingobium, Reyranella, Lacibacterium, Rhizobium, Porphyrobacter, Rodococcus, and Defluviimonas. In the phylum of Actinobacteria, the sequences corresponded to the genera Mycobacterium, Nocardioides, Uteimicrobium, Rhodococcus, and Paraoerskovia. Four bands were closely related to the Mycobacterium (100%) and five to Nocardioides (99%) and Luteimicrobium (100%). The sequences of DGGE bands NA1.C2 and NA3.C2 were highly similar to Rhodococcus and Paraoerskovia, respectively. Nocardioides, Rhodococcus, and Mycobacterium from Gram-positive bacteria have played a significant role in degrading persistent HMW PAHs (Liang 2006; Khan et al. 2009a). Due to their enzymes, mainly the dioxigenases, dehydrogenases, and hydrolases, these bacteria were effective in degrading toxic substances such as PYR. Rhodoccoccus and Mycobacterium have also been described as biosurfactant-producing bacteria and enhancers of hydrocarbon bioavailability (Shekhar et al. 2015). The second largest phylum after the Actinobacteria was the Bacteroidetes, with sequences clustered in the Proteiniphilum and Algoriphagus. Eight sequences were related to the Proteiniphilum (97–98%), while only one sequence represented the Algoriphagus. Proteiniphilum acetatigenes was successfully employed in the anaerobic bioremediation of hydrocarbon-laden sludge. Larsen et al. (2009) showed it to be a very effective bacterium for removing up to 80% of PAHs. The bacterium of the genus Algoriphagus has also been considered as an oil degrader (Yang et al. 2013; Wang et al. 2014). Among the Proteobacteria, the sequences were grouped as members of the classes alphaproteobacteria, gammaproteobacteria, and deltaproteobacteria. Within the alphaproteobacteria, eight sequences belonged to the genera Sphingomonas, Sphingobium, Bradyrhizobium, Reyranella, Lacibacterium, Rhizobium, Porphyrobacter, and Defluviimonas. The Proteobacteria and Actinobacteria phyla were identified as efficient PAHs removal bacteria (Singleton et al. 2011). Defluviimonas alba has been isolated from an oil field, Lacibacterium aquatile from a freshwater lake and Reyranella soli from soil (Khan et al. 2009b; Kim et al. 2013; Sheu et al. 2013). The genus Sphingomonas was distinguished from Pseudomonas by Yabuuchi et al. (1990) and was widely used in biodegradation assays, particularly for chlorophenols, due to its remarkable ability to break the bonds of hydrocarbons (Bosso & Cristinzio 2014). According to Lu et al. (2019), Sphingomonas and Sphingobium may be essential, especially in the early removal of LMW PAHs, whereas Nocardioides seem to be particularly important for the further breakdown of HMW PAHs. The uncultured bacteria extracted from hydrocarbon-contaminated soils were all associated with gamma-proteobacteria, delta-proteobacteria, and acidobacteria (Table 6). Moreover, the DGGE patterns (Figure 7) show some differences in the number and abundance of the bands as a function of treatment time, suggesting a different behaviour of the bacterial species. For example, some of the DGGE bands remain strongly present while others disappear or appear over time.

The NA strategy was effective due to the naturally occurring bacteria in the native consortium. Early in the study, N7, N2, and N1 bands associated with Sphingobacterium, Sphingomonas, and Rhodococcus were detected. However, a dramatic decrease in the abundance of N1 and N2 bands was observed at weeks 12 and 16. In addition, bands N3 and N4, related to Mycobacterium and Nocardioides, appeared in week 4. Bands 10 and 11, related to Mycobacterium szulgai and Mycobacterium doricum, became clearly visible at week 16. Thus, monitoring of bacterial dynamics and analysis of PAHs content showed that bacteria involved in the reduction of HMW PAHs, such as Mycobacterium and Nocardioides, were labelled at the mid and end of the experiment. However, Sphingobacterium and Sphingomonas, which are implicated in the destruction of LMW PAHs, were present at both baseline and mid-treatment. Thus, a more than 50% reduction of LMW PAHs was achieved at week 8 compared with 24 weeks for HMW PAHs. For the BA strategy, the DGGE patterns were similar but relatively more pronounced. For example, the bands associated with the phylum Actinobacteria were strongly marked at week 12, whereas most of the DGGE bands corresponding to Proteiniphilum acetatigenes had decreased by week 24.

Industrial influent treatment began by stripping water through a CSTR with OLR between 360 and 1,440 g COD m⁻³ day⁻¹. The OLR was maintained at 1,200 g COD m⁻³ day⁻¹ and the HRT was still high in order to avoid the organic overload of the CSTR. This is principally due to the high toxicity of the stripping wastewater. Consequently, the operating procedure seems to need improvement. The CSTR treatment resulted in a COD removal rate of more than 88.16% at about 710 mg COD L⁻¹ day⁻¹. The reduction of C13 and C14 alkanes was complete, while the removal of C26−C30 was about 85%. In addition, the wastewater collected in the APIS was treated by the TF. The COD, BOD, TPH, AHs, phenol, TKN, and TP values were significantly reduced to about 220, 90, 9, 4, 0.49, 49.35, and 4.9 mg L⁻¹, respectively. For soil BA, the effluent from the TF was recovered. This provided a continuous supply of water and well-adapted microorganisms in a cost-effective environmental approach. The results show that the BA strategy significantly improved the abundance of biomass in the soil, the removal of pollutants and the reduction of phytotoxicity. In fact, the high microbial activity was related to the growth of indigenous bacteria and the successful colonization of exogenous bacteria, which cooperated to remove hydrocarbons and phenolic compounds from the soil. The dominant bacterial phyla in the soil were primarily Actinobacteria, followed by Bacteroidetes, Alphaproteobacteria, and then Gammaproteobacteria. The efficiency of soil bioremediation was affected by biotic and abiotic parameters, including treatment time, soil mixing and wetting, nutrient availability, abundance and biodiversity of specific microorganisms, nature, and molecular composition of the contaminant.

Recovery of TF effluent and restoration of specific bacteria through a soil BA strategy successfully support environmental sustainability. Future work will extend the soil aspect, including SOM and nitrogen/phosphorus ratio on growth and microbial activity to improve contaminated soil treatment. In addition, it could be suggested to investigate the use of different RedOx zones (aerobic, anoxic, and anaerobic) in combination; it might provide improved wastewater treatment and/or improved BA in the contaminated soils, where different RedOx environmental conditions may be relevant.

This study was supported by the Ministry of Higher Education and Scientific Research, Tunisia. The authors are grateful to the Tunisian Lubricant company, SOTULUB for their kind help in the field. The authors extend their appreciation to the Researchers Supporting Project number (RSP 2024R75), King Saudi University, Riyadh, Saudi Arabia.

M. J. conceptualized the work, rendered support in data curation, developed the methodology, wrote the original draft, wrote the review and edited the article. F. K. rendered support in data curation, developed the methodology, validated the process; visualized the resources; wrote the original draft, developed the software. L. M. rendered support in data curation and data curation, validated the process; visualized the resources; wrote the review and edited the article. S. L. rendered support in data curation, developed the methodology. R. B. developed the methodology and software. K. K. Y. rendered support in data curation, reviewed and edited the article. S. S. conceptualized and supervised the work, rendered support in project administration and data curation, wrote the review and edited the article.

All authors and the institute where the analysis was done have the consent for the publication and have no objections.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.