The sampling procedure was performed as described in previous works [13]. Sampling activities were carried out in six campaigns, for a total of 72 samples. The selected study area is located along the coastline which is 204 km long from Toukouzou Hozalem to Assinie. Twelve (12) sampling sites were selected and numbered as follows: Toukouzou (TKZ), Addah (ADA), Adjué (ADJU), Jacqueville (JAC), Adjouffou (ADJ), Bassam-Modeste Plage (BMP), Bassam-Modeste Lagune (BML), Azurety (AZT), Assinie-Plage (AP), Assinie-Lagune (AL), Assinie Canal Droite (ACD), Assinie Canal Gauche (ACG).

Samples were soxhlet extracted (Grosseron, Gerhardt; France) over 24 h with 400 mL of dichloromethane (99.8% of purity)/methanol (99.9% of purity) (3/1, v/v). Then, the sediment was mixed with 5 g of anhydrous sodium sulfate (Na₂SO₄). The extracts were concentrated using a rotary evaporator (IKA, Germany) evaporated under a gentle stream of nitrogen (N50; 99.999% of purity). These extracts were further purified for analysis using 5 g of silica gel columns (silica 60; 0.063 - 0.200 mm; 5% deactivated), 2 g of alumina (aluminium oxide 90 active, 0.063 - 0.200 mm; 5% deactivated) and 2 g of Na₂SO₄, with 120 mL of hexane/dichloromethane (3:1 v/v). A solvent mixture (10 mL) hexane (99.1% purity)/dichloromethane (DCM) (4/1, v/v) was used for elution and evaporated under a stream of nitrogen to incipient dryness.

PAHs in sediments were determined by GC - MS (Perkin Elmer Clarus 680 - Perkin Elmer Clarus 600). Calibration and quantification of PAHs were performed in SIM (Selected Ion Monitoring) mode. The identification of PAHs is carried out by acquiring the total ion current (TIC) and complete mass spectra, as well as by comparison with the chromatographic retention times determined for the standards of the pure compounds. All the operating conditions used in GC-MS are presented in Table 1.

Pearson’s correlation analysis was used to analyze the connections between PAH concentrations and external factors to identify the probable PAH sources. Statistical analyses were performed with XLSTAT (version 2017) to identify the relationships between the variables. The excel tables were performed using the office 365 proplus version.

All sediments analyzed by GC-MS had detectable concentrations of PAHs. Thus, 16 PAHs were identified (Figure 1). The average concentrations range from 0.11 to 332.72 mg/kg with an average concentration of (25.63 ± 16.10) mg/kg. Benzo (k) fluoranthene has the lowest concentration in the sediment, ranging from 0 to 0.85 mg/kg (average 0.24 mg/kg). Naphthalene (Nap) has the highest relative percentage (332.72%) of total PAHs in sediment, followed by phenanthrene (Phe = 41.75%), pyrene (Pyr = 9.42%), anthracene (Ant = 7.68%) and fluoranthene (Fla = 5.63%). The least dominant compounds are benzo (k) fluoranthene (B(k)F), benzo (a) anthracene (B(a)A) and acenaphthene (Ace). They represent 0.11%, 0.22%, and 0.28% of the sediment, respectively.

This abundance of naphthalene and phenanthrene could be explained by activities such as petroleum exploitation, buffer storage of imported petroleum products, and tanker navigation. In contrast to this finding, [14] with concentrations ranging from 13 to 780 μg/kg, linked the presence of PAHs at the atlantic coast of the democratic republic of Congo to other anthropogenic activities. Indeed, these authors believe that these pollutants (PAHs) come from untreated urban, industrial and hospital effluents, agricultural runoff, solid waste, and landfills. Based on the classification adopted by [15], the level of PAH contamination is classified as low to highly contaminated. Accordingly, 16.67% of the samples were presented as low contamination (<100 ng/g); 33.33% as moderately contaminated (100 - 1000 ng/g); 41.67% as highly contaminated (1000 - 5000 ng/g), and 8.33% as very highly contaminated (>5000 ng/g). However, the level of PAH contamination in the coastal sediments of the Persian Gulf was quite low in the sediments and ranged from 3.63 to 21.75 μg/kg with an average of 6.35 μg/kg with respect to this study due to various sources such as cement complexes, fish and shrimp farming, and urban and industrial activity [16].

PAHs are generally classified into two main groups: low molecular weight (LMW) PAHs (PAHs with 2 to 3 aromatic rings) and high molecular weight (HMW) PAHs (PAHs with 4 to 6 aromatic rings) [17]. Based on the benzene ring number, sixteen PAHs can be separated into three types: 2 - 3 rings (Naphthalene (Nap), acenaphtylene (Acy), acenaphtene (Ace), anthracene (Ant), phenanthrene (Phe), fluoranthene (Fla), fluorene (Fle)); 4 rings Pyrene (Pyr), benzo (a) anthracene (B(a)A, and chrysene (Chr)) and 5 - 6 rings PAHs Benzo (k) Fluoranthene (B(k)F), Benzo (b) Fluoranthene (B(b)F), Benzo(a) Pyrene (B(a)P), Dibenz (a, h) anthracene (DB(ah)A), Indeno (1, 2, 3, cd) Pyrene (IND), Benzo (g, h, i) Perylene (B(ghi)P). PAHs are mainly divided into two sources: diagenetic source and pyrogenic source. PAHs from the first source were mainly low molecular weight (LMW) PAHs, mainly exist in gaseous form and are volatile and easily degraded, while those from the second source were mainly heavy molecular weight (HMW) PAHs [4]. In this research, Naphthalene which is 2 aromatic rings is more abundant with 90.26% of total PAHs followed by 3-rings (acenaphthylene, acenaphthene, anthracene, phenanthrene, fluoranthene, fluorene) with 5.48% abundance of total PAHs, which suggest petrogenic sources [18] and also that the LMW-PAHs tend to be more conserved in the sediments [1] (Figure 2). 4 - 6 ring PAHs had a lower composition ratio (approximately 4.26%) than 2 - 3 ring. High ring number PAHs have a large molecular weight and are primarily found in the form of particles, which are stable and not easy to volatilize and thus can be retained in the environment for a long time [7]. Therefore, the HMW 5 - 6 ring PAHs were primarily from a pyrogenic source. This variation can be explained by the action of industrial activities, oil, gas, oil spills, pipelines and shipping off the coastal zone [17]. In addition, the highest PAH level for this study is

approximately 45.5 times lower than the maximum PAH concentration (78,523 μg/kg) for Guanabara Bay in Brazil, likely due to industrial activities and high population density [19]. The TKZ, ADJU, JAC, ACG, and AL stations have the highest PAH levels. This is because these are subject to oil activities, intense shipping activities, and oil spills. These high concentrations can be attributed to industrial and household wastes and oil activities on platforms upstream of these stations. The highest percentage of PAHs in sediment was found at the estuary BML station. Estuaries are characterized by receiving large amounts of organic matter and minerals from the drainage basin. This station is located not far from a main thoroughfare confirming the increased traffic intensity which is a significant source of PAH contamination [20]. It is the most urbanized and populated. Also, due to industrial, petroleum, and domestic activities related to offshore drilling activities upstream, it receives inputs of runoff from nearby landfills, atmospheric fallout from exhaust fumes from high traffic areas. There is also a wide range of economic and tourism activities along the roadside. Similar trends have been observed for river sediments in other studies. In addition, there is significant PAH pollution attributed to urban runoff and industrial activities in an industrialized region with a large anthropogenic influence [21]. The lowest concentrations at ADDA (90.67 mg/kg), AZT (50.81), ACD (40.76) stations could be attributed to the confined nature of these locations and the lesser road traffic to which they are exposed. It is shown that areas away (downstream) from industrial activities and heavy traffic have relatively low PAH concentrations. Higher concentrations are also observed near petroleum-related activities such as the oil refinery, oil terminal, and oil and gas production areas [15]. Also, residential sites affected by heavy traffic emissions have fairly high concentrations (AL: 473.72 mg/g and AP: 838.01 mg/g) [21].

PAH sources were predicted using conventional diagnostic ratios reported in several studies [10]. Different ratios between low and high molecular weight PAHs were used to analyze the main sources of emission of this compound in soils and sediments (Table 2). The diagnostic ratio of PAH components can be used to determine the contributions from different PAHs sources. Different diagnosis ratios indicate specific emission sources.

All sediment samples analyzed show ∑LMW/∑HMW ratios greater than 1 (Table 2) with a predominance of low molecular weight species over high molecular weight PAHs. This suggesting similar source for all samples and that the low weight species tend to be more conserved in the sediments [1]. Thus, the primary source of PAHs is petrogenic [1]. This is confirmed by the value of the Fla/Pyr ratio of 0.60 less than 1 [21]. The main source of low molecular weight (LMW) PAHs is mainly due to anthropogenic activities involving low temperature combustion process of coal and biomass [21]. Other sources include industrial activities, urban wastewater, chemicals, spills, oil leaks from drilling near oil installations, ships, discharges from oil exploration and exploitation, discharges from fuels (diesel, super), lubricating oils or tire debris and shipping [10]. However, high molecular weight PAHs (HMW; 4 - 6 rings) were low, this may reflect the presence of high combustion from pyrogenic sources [20]. The BaA/(BaA + Chr) ratio is more appropriate for identifying pyrolytic sources [22]. All nearshore samples have BaA/(BaA + Chr) ratios below 0.2, suggesting petroleum contamination (Yunker et al., 2002). 25% of the samples have BaA/(BaA + Chr) ratios greater than 0.1, indicating inputs from pyrogenic sources [18]. The value of IND/(IND + B(ghi)P) ratio [23] of value 0.5 between (0.2 - 0.5) corresponds to a petroleum combustion source [8]. The same is true for the Ant/(Ant + Phe) ratio with an average value of 0.2 greater than 0.1 which therefore confirms the source of oil combustion. Fla/(Fla + Pyr) ratio values less than 0.4 indicate petrogenic sources [24] subject to liquid petroleum, and chemical contamination. These are the JAC, ADJ, BML, AZT, and ACD sites, which are major shipping lanes with the presence of a BML estuarine area [25]. For TKZ, ADDA, ADJU, BMP and AP where oil exploration is undertaken, the Fla/(Fla + Pyr) values are between 0.4 and 0.5. This implies that these sites are polluted by fossil fuels [24]. There are large freight and tourist ships along the coastline that could account for the source of fossil fuels. These stations are located where there is frequent traffic from oil and shipping activities. Fla/(Fla + Pyr) values above 0.5 are obtained at the ACG and AL sites. The samples from these sites are far from the action of wastewater discharges, industrial, oil and shipping activities, which explains their low PAH content. Mixed sources of PAHs in these areas may be due to transport of pollutants from other areas via the tidal current to local pollution sources. It is clearer that these sources are a mixture of oil and combustion sources (fossil fuels and biomass) but suggest that oil sources are the primary and fossil fuel sources and biomass combustion are the secondary sources of PAHs. The ratios of the different PAH isomers are usually plotted in a cross plot to determine the types and source of PAHs in the environment. Thus, the graphs of the diagnostic ratio (LMW/HMW and Fla/Pyr); IND/(IND + B(g h i)P) and Fla/(Fla + Pyr) (Figure 2 and Figure 3) consolidate the results obtained by the previous table and other integrated diagnostic ratios and indicate the sources.

The statistical analysis carried out on the data of the 16 detected hydrocarbons allowed to obtain the table of the eigenvalues (Table 3). Of the 11 factors obtained only 6 were retained to do the principal component analysis (Table 2) [20]. Indeed, a factor is chosen when its eigenvalue is ≥1. Of the 6 factors, F1, F2, F3 and F4 were selected because the correlations between certain parameters and these factors are significant. These factors combined two by two resulted in the correlation circles. According to Table 4, the hydrocarbons Ace, Phe, Ant, Chr, B(a)P and IND have significant correlations with the factor F1. These correlations are 0.744, 0.878, 0.817, 0.700, 0.731 and 0.726, respectively. As for the Fle and Fla hydrocarbons, they admit significant correlations of 0.795 and 0.763 respectively with the F2 factor. Like the F2 factor, the F3 factor is correlated with two hydrocarbons (Pyr; r = 0.758) and (B(k)F; r = 0.700). The F4 factor is correlated only with the hydrocarbon B(ghi)P.

The hydrocarbons differ on the one hand by the number of carbon they contain. The statistical study carried out allows us to see a link between certain hydrocarbons by observing the correlation table (Table 5). Ant is significantly correlated to Nap and Phe with respective correlations of r = 0.76 and 0.73. B(a)P and Ace with a correlation r = 0.70 are significantly correlated. Like B(a)P, B(b)F is moderately correlated with Ace with a correlation of r = 0.61. Also, Fle is correlated to Fla with a significant correlation of r = 0.79. With a highly significant correlation of r = 0.94, B(k)F is related to Pyr. However, if Ant is significantly correlated to Phe, it is not so for Chr which is moderately correlated with a correlation of r = 0.62.

According to [18] a correlation is considered strong when it is equal or higher than 0.70. The principal component analysis carried out according to the factors F1 and F2 resulted in the distribution circle in Figure 4. From this figure, it can be seen that the detected hydrocarbons are divided into 4 groups. Fle, B(ghi)P and B(a)A form one group. Also, Fla, B(k)F, Pyr, Acy, Phe, Ace, Chr and Acr formed another group. A third group is formed by B(b)F, B(a)P, Ant and Nap. The fourth group is formed by IND and DB (ab)A. According to [18], these groups formed could indicate a common origin for the elements belonging to the same group.

However, considering the factors F1 and F3 (Figure 5), which total 46% variance, the distribution circle (Figure 5) shows that some hydrocarbons although considered to belong to a group are not. Instead of four groups for the factors F1 and F2 considered, we obtain three groups for the factors F1 and F3 considered. The IND and DB(ab)A which formed the fourth group are associated with DB(ab)A to form a group. The B(ghi)P, B(k)F, Pyr, Acy, B(b)F, B(a)P, Ace and Chr form a second group that is full of the most hydrocarbons. More than 90% of the elements belonging to this second group belonged to the second group when considering the F1 and F2 factors and in the same quadrant (Figure 5). This allows us,

in accordance with [26] to affirm that these hydrocarbons are indeed of common origin. The third group formed in this figure (Figure 5) contains Fle, Fla, Acy, Nap, Ant and phe. Only Nap and Ant did not leave the quarter circle. This suggests that these two elements could indeed be of common origin.