Introduction

Technological advancements during the last century have led to the use of synthetic carbon-based polymers for everyday household and office items such as furniture, fabrics, automotive parts, housings for electronic equipment and surface coatings for other materials (de Wit 2002; Guerra et al. 2011). The high tendency of these materials to act as a source of fuel is an indication that their very presence could portend danger, particularly in instances where a high risk of ignition is associated with the item’s use. Fire outbreak is a major cause of damage to both private and public properties. In most cases, fire outbreak could involve loss of lives and public expenses. For instance, in the USA in 2007, over 1.5 million fires were reported, which resulted in 17,675 injuries, 3430 deaths and direct losses of over $14 billion (Karter 2008). Similarly, in 2013, 1.24 million fires were reported, which resulted in 15, 925 injuries, 3240 deaths and direct loss of $11.5 billion (Karter 2014). In Lagos State of Nigeria, 2342, 1774 and 1499 fire cases were reported in 2012, 2013, and 2014, respectively, and direct property losses worth N54bn during the 3 years were reported (Vanguard, January 20, 2015). Also, 262 people were reportedly lost in 368 fire incidents in 2011, while 185 lives were lost in 470 fire incidents in 2012 (Premium Times 2013). In Nigeria, about N50bn were lost annually due to fire disasters (The Punch Newspaper 2013).

Flame retardants (FRs) play an important role in safeguarding life and property. Today, in order to prevent fire outbreak, FRs are incorporated into combustible materials such as plastics, wood, textiles, electronic products and paper materials. Combustion involves four stages: preheating, decomposition and evolution of volatiles, ignition and propagation (Troitzch 1990). Prevention of any of these four stages should lead to the suppression of fire (Zhang et al. 2016). Halogens are very effective in trapping free radicals produced during the combustion process (Guerra et al. 2011). Approximately 25 % of all FRs contain bromine as the active ingredient, which is very effective in trapping free radicals, hence removing the capability of the flame to propagate. Organobromine compounds have become more popular due to their stability, higher trapping efficiency, and lower decomposing temperature (Guerra et al. 2011. More than 80 different aliphatic, cyclo-aliphatic, aromatic, and polymeric compounds are used as brominated flame retardants (BFRs). BFRs, such as polybrominated biphenyls (PBBs), polybrominated diphenyl ethers (PBDEs), hexabromocyclododecane (HBCD), and tetrabromobisphenol A (TBBPA), have been used in different consumer products in large quantities, and consequently, they have been detected in environmental matrices (Guerra et al. 2011). Of all the four common BFRs, PBDEs are the most commonly reported. Plastics in electrical and electronic equipment and related wastes are considered to contain the largest share of PBDEs followed by polyurethane foams found in car, furniture, mattress or baby products (Stockholm Convention 2011).

PBDEs are used as additives in combustible materials; they are capable of being released into the environment, thereby settling in dust, water, sediment and biota (de Wit 2002). PBDEs are toxic and persistent environmental xenobiotics that may undergo long range transport and are capable of bioaccumulating and biomagnifying along the food chain (Kuriyama et al. 2005; Moon et al. 2007). These properties and their diffusion patterns in air have triggered concern about the possible general environmental effects of this group of substances. In Africa, several studies on PBDEs have been carried out and have confirmed the presence of PBDEs at significant levels in environmental matrices such as the eggs of African ibis Polder et al. (2008), crocodile eggs (Bouwman et al. 2014), leachate (Odusanya et al. 2009; Daso et al. 2013a), human serum (Linderholm et al. 2010), cow milk (Asante et al. 2010), human milk (Darnerud et al. 2010; Hassine et al. 2012), sludge (Daso et al. 2011b), bivalves (Bodin et al. 2011), fish (Wepener et al. 2012), dust (Kefeni and Okonkwo 2012; 2014), altitudinal soil (Parolini et al. 2013), water (Daso et al. 2013b), air and precipitation (Arinaitwe et al. 2014). Olukunle et al. 2012b and Daso et al. 2013c carried out the evaluation of various extraction methods for the analysis of PBDEs and their application to aqueous environmental samples. The presence of PBDEs in river sediments has been reported by many researchers (Bodin et al. 2011; Daso et al. 2011a; Olukunle et al. 2012a; La Guardia et al. 2013; Adewuyi and Adeleye 2013),

Research concerning the presence of PBDEs as new, emerging, and ubiquitous contaminants released into the Nigerian environment from the breakdown of combustible materials is still in its infancy, and information on production and usage of PBDEs in Africa is scarce generally. The data on cathode ray tubes (CRT) in the inventory of electrical and electronic equipment and related wastes in Nigeria is reported as 791,325 t, out of which, 23,7000 t was plastics with 216, 000 t obtained from TV casings and 21,000 t from computer CRT casings (Ogunbiyi et al. 2012). From these figures, average concentrations of c-Octa-BDE, deca-BDE, and TBBPA were estimated by Sindiku et al. (2014) to be 594, 1880, and 1450 t, respectively. Sindiku et al. (2014) reported that average PBDE levels of c-Octa- and deca-BDE in Nigerian stockpiled CRT casings were 1.1 % for television and 0.13 % for computer, and these values were above the hazardous substance limit. There is virtually no environmental data regarding PBDE levels in stream water and sediment with the exception of the study conducted by Adeleye and Adewuyi (2013) on sediment of Lagos Lagoon. This study was therefore aimed at determining the occurrence and distribution of some PBDEs in the sediments of Asunle stream of the Obafemi Awolowo University, Ile-Ife, Nigeria.

Materials and methods

Study area

Obafemi Awolowo University dumpsite has been in operation since the inception of the university in 1971. The dumpsite serves as the final destination where all refuse collected within the university community are dumped and subjected to open-air incineration. Among the wastes commonly dumped are medical wastes, domestic and household wastes, television and computer monitors, X-ray cathode lamps, florescent tubes, discarded refrigerators, metallic materials from engineering and mechanic workshops, waste papers, cosmetic products, expired drugs, and chemicals. The university dumpsite lacks appropriate solid waste treatment facility, which leads to dumping and subsequent open-air burning of hazardous wastes, thereby causing environmental pollution. The destruction of these wastes via incineration is capable of adding potentially toxic metals and persistent organic pollutants to the surrounding and the adjoining stream. This can happen through leaching, as residues of incinerated waste materials are washed down into the adjoining stream, thereby adversely impacting the aquatic environment. Very close to the dumpsite is Asunle stream, a perennial stream whose source is located about 0.25 km uphill from the Obafemi Awolowo University Ile-Ife refuse dumpsite (Fig. 1). The stream serves as the water source for several farmlands of the study area and runs a stretch of more than 10 km, cutting across human communities such as Abagbooro, Agbogbo and Amuta (Ogunfowokan et al. 2013). Commonly planted crops on the farmlands around the stream include cash crops such as cocoa, cola nut, palm trees, etc. and food crops such as cassava, maize, yam, cocoyam, plantain, banana, pineapple, oranges, pepper, and various types of vegetables. This makes human activities around the stream vigorous. The rural dwellers living along this stream rely on the water from the stream for household purposes, irrigation and agricultural applications, palm oil processing, and mixing and dilution of pesticides used for spraying cocoa and other crops. These factors were considered in deciding to monitor the level and distribution of PBDEs in the sediments of Asunle stream.

Fig. 1
figure 1

Map showing the sample location

Full size image

Sample collection, preparation and extraction

Samples were collected on a monthly basis over 8 months spanning November, 2012, through February, 2013, representing the dry season, and May through August, 2013, representing the wet season. Sediment samples were collected at five locations along the course of the stream. The bottom sediment samples were collected in aluminium foil previously cleaned with pure acetone and wrapped in a black cellophane nylon. The samples were air dried to constant weight in an aerated cupboard to prevent cross-contamination. The dried samples were ground and sieved with 500-μm stainless steel sieve. The sieved samples were stored in air-tight cellophane bags kept in clean 250-mL-capacity amber-coloured bottles and preserved in the refrigerator at 4 °C until further analysis was required.

Approximately 10 g of the prepared soil sample was weighed and transferred into a pre-extracted cellulose thimble for Soxhlet extraction. The soil sample was spiked with 100 μL each of 100 ng/mL of two surrogate standards, pentachloronitrobenzene (PCNB) and BDE77, to monitor the analytical recovery of the target analyte. The sample was then extracted with 150 mL of n-hexane/dichloromethane (2:1 v/v) for 3 h. Prior to extraction, about 2 g of copper powder was added to each sample to remove traces of elemental sulphur that could possibly interfere with the analyte’s determination.

Sample cleanup

Before extract cleanup, sample extracts were concentrated to approximately 1 ml using a rotary evaporator. Purification of extract in this study was done using a modified multilayer silica gel column technique of Yun et al. (2008) and Kupper et al. (2008). The concentrated extracts were then cleaned on a preconditioned multi-layer silica gel column using n-hexane as the eluting solvent. The glass column was packed from the bottom with 1 g activated silica gel, 4 g of basic silica gel (30 % NaOH, w/w), 1 g of activated silica gel, 8 g of acidic silica gel (44 % H2SO4 w/w), 2 g activated silica gel, and 4 g of anhydrous sodium sulphate. The packed column was preconditioned with 10 mL of doubly distilled n-hexane to remove trapped air and background contaminants within the column. The n-hexane layer over the uppermost layer in the column was maintained at 2 mm to prevent further infiltration of air into the column.

The concentrated extract was then quantitatively transferred into the column and eluted with 100 mL of n-hexane. The eluate was concentrated to approximately 1 mL using a rotary evaporator. This was followed by addition of 1 mL of isooctane serving as keeper before final concentration to a suitable volume in an amber sample vial under a gentle stream of nitrogen. The prepared samples were kept in the refrigerator prior to final instrumental analysis.

Instrumental analysis

Analysis of target compounds was performed on a Trace 1300 Series GC coupled to a TSQ8000 Mass Spectrometer equipped with a TriPlus RSH Auto Sampler available at Stellenbosch University, South Africa. The chromatographic separation of analyte was performed using a ZB 274305 semi-volatile column (30-m length, 0.25-mm id, 0.25-mm film thickness). Helium gas was employed as a carrier gas with a flow rate of 1.0 mL/min using a constant flow mode. High-purity nitrogen gas was used as make-up gas with a flow rate of 20 mL/min. The injector and detector temperatures were set at 280 and 320 °C, respectively. The oven temperature was programmed as follows: 150 °C held for 2 min, ramped at 8 °C/min to 320 °C and held for 2 min. One microlitre each of the mixed standard solutions containing the target compounds and extracts of the experimental samples was injected using a splitless injection mode.

Quantification was based on examination of the peaks of target compounds using external calibration curve technique. The calibration plots were composed of six levels of 5, 10, 20, 50, 75, and 100 ng/mL for all the target compounds. The linearity (r 2) of the calibration plots for each target compound was greater than 0.9914. Identification of analyte was done by comparing the retention times with those of reference standards.

The quality assurance and quality control (QA/QC) samples included the use of amber-coloured bottles for sampling storage, regular injection of solvent blank and standard solutions of the target compounds. Matrix spiked experiments were also performed. The results of the recoveries of the target compounds ranged from 92 % (BDE100) to 105 % (BDE28), while recovery of surrogate standards ranged from 72 % (BDE77) to 87 % (PCNB). A calibration check of 5 ng/mL standard was run after every five sample runs to ensure that less than 20 % variation was found from initial calibration standards.

TOC determination

The total organic carbon (TOC) content of sediment samples was determined by weight loss on ignition method described by Dean (1974). Prior to the gravimetric determination, the sediment samples were dried to constant weight at 105 °C to ensure total removal of moisture. In each case, about 5 g of sediment samples was weighed into crucible and ignited in a Vescar muffle furnace model ECF 3 at 440 °C for 4 h. The setup was allowed to cool and later transferred into a dessiccator, followed by weighing of the ignited sediment samples. The final weight of the ignited sediment was then determined to estimate the total organic content of each sample.

Chemicals used, their sources and purification

The chemicals and reagent used included dicholoromethane, n-hexane, isooctane, acetone (Merck, South Africa); silica gel 60–200 mm, anhydrous sodium sulphate, and copper powder (Sigma-Aldrich, South Africa); PBDEs and PCNB standards (Cambridge Isotope Laboratories, Andover, MA, USA, via Industrial Analytical (Pty), Midrand, Gauteng, South Africa); helium; and nitrogen (99.99 %) were supplied by Afrox (Pty) Ltd., Cape Town, South Africa.

All the organic solvents were triply distilled to obtain pure solvent that precluded all trace organic contaminants. Other materials such as glass wool, anhydrous sodium sulphate, silica gel, and copper powder were all heated in a muffle furnace model ECF 3 at 450 °C for 4 h prior use.

Statistical analysis

All data were managed with Microsoft Excel software (2007 version). The various data obtained were analyzed using descriptive statistics and one-way analysis of variance along with the Duncan multiple range test to determine significant difference between means. The two-tailed Pearson correlation coefficient was used to determine the strength of association between PBDE congeners. Principal component analysis was used to predict and identify possible sources of these contaminants. Statistical analysis of the data was performed using the Statistical Package for the Social Science (SPSS) software 15.0 for Windows Evaluation version.

Results and discussion

Monthly levels of PBDEs in bed sediment

The result obtained from the analysis of the bed sediment for the selected PBDE congeners, as presented in Table 1, indicated that all the congeners investigated in this study were present in the sediments. Polybrominated diphenyl ethers (PBDEs), like most persistent organic pollutants, bind to suspended materials in the aqueous phase and settle along with the suspended materials into the bed sediments, thus leading to accumulation of these contaminants in the bed sediments. Organic matter in the bed sediments thus acts as a repository for persistent organic contaminants.

Table 1 Monthly variation of PBDEs (ng/g) in bed sediment of Asunle stream
Full size table

The overall mean concentrations of the total PBDEs ranged from 1.80 to 9.46 ng/g. The highest concentration was recorded in August. This period was characterized by heavy rainfall. Thus, high precipitation during the rainy season was probably responsible for increased levels of these contaminants during this period as a result of their mobilization from dumpsite soil into the stream through surface runoff. With respect to the variation of these contaminants at different sampling periods, only BDE47 showed statistically different (p < 0.05) values from the others with analysis of variance (ANOVA). The overall mean concentrations of the total Ʃ6PBDEs throughout the period of study did not show any significant difference. Stream hydrodynamics such as flow rate and the degree of turbulence might have affected the settling rates and deposition of suspended matter in the river (Forstner 2004).

Levels of PBDEs in bed sediment of Asunle stream at various locations

The distributions of PBDE congeners at various locations in the bed sediments were investigated and are presented in Table 2. Total PBDEs ranged from 0.73 to 10.43 ng/g. The lowest mean value was recorded at the upstream (control) site (L0), while location 5 had the highest recorded mean concentrations. Location 1 was the site closest to the dumpsite where open incineration takes place from time to time. The concentration of PBDEs showed an increasing trend with the distance from L0 to L5. It appeared that the PBDE congeners were retained more by river sediments as the distance increased from the dumpsite. This could be as a result of larger volume of water with corresponding slower transportation rate of the river body, which resulted in deposition of larger PBDE-bearing bottom sediments over time. Dumping of e-wastes and their open-air combustion is likely responsible for the addition of PBDE contamination to the soil and river (Zhang et al. 2016). Leung et al. (2006) and Leung et al. (2007) affirmed that PBDEs are present in soil and river sediments as a result of insufficient destruction resulting from dumping, open burning, and acid leaching. In some instances, atmospheric deposition could become an important route for PBDE transportation to the sediments (Hale et al. 2003; Song et al. 2006). The presence and levels of PBDEs recorded upstream, about 0.25 km away from the dumpsite, could be explained in keeping with such previous findings, while prevailing wind and stormy weather during the rainy season could enhance the transport and re-distribution of PBDEs to other regions far away from the point of generation. This observation was in agreement with Zhao et al. (2009) who found that PBDEs from e-waste recycling area diffused into ambient regions and resulted in a halo pattern of PBDE contamination of at least 74 km in radius. The elevated mean concentration value obtained at location 5 could be attributed to the pronounced sedimentation that occurred at this site since these contaminants could bind to humic substances in the sediment.

Table 2 Distribution of PBDEs (ng/g) in bed sediment at various locations of Asunle stream
Full size table

To further investigate the behaviour of the target compounds in the stream sediments investigated, ANOVA was conducted. Results indicated that BDEs 47 and 154 did not exhibit significant difference (p < 0.05) in all the locations unlike other PBDE congeners in the bed sediments although the flow rate as well as the pattern of stream flow may affect the settling of suspended matter and sediment deposition in the river bed.

Total organic carbon (TOC) is a measure of the organic carbon in the sediment sample. Since particles rich in organic carbon have the greatest potential to bind PBDEs and since PBDEs preferentially bind to particulates in water (Environment Canada 2009), measuring the TOC concentration in the sediment samples was used to establish whether there was any correlation between the measured PBDE and TOC. The TOC measured in this study ranged from 0.87 ± 0.28 % at L1 to 1.80 ± 1.67 % at L5. The TOC showed an increasing trend with the distance from L0 to L5.

Seasonal variations of PBDE in bed sediment of Asunle stream

Extensive monitoring of contaminants in different environmental matrices for a period of 2 years would be most appropriate to predict seasonal variability of the target compounds (Chapman 1996). However, assessment of the total PBDEs in different seasons could also be used to predict seasonal characteristics of most contaminants in a given environmental matrix. Seasonal variations of PBDE congeners in sediment samples collected from Asunle stream are presented in Fig. 2. The results show that the concentrations of these contaminants were slightly higher in the sediment during the wet season. However, BDE153 demonstrated an exceptionally higher concentration in the wet season. During the wet season, the leaching of these contaminants into the aquatic system could have been a contributive factor to their high levels as detected and recorded in the sediments during the wet season. The lower values of these contaminants during the dry season might be due to decreased rate of precipitation and consequently lower amount of the contaminants transported via erosion during this period.

Fig. 2
figure 2

Seasonal variations of PBDE congeners in bed sediment of Asunle stream

Full size image

Principal component analysis of PBDEs in bed sediment of Asunle stream

To predict and identify the possible sources of the PBDE contaminants, the results obtained were subjected to principal component analysis to establish whether the contaminants were actually arising from the same source or not. The principal component analysis of PBDEs in the bed sediment of Asunle stream is presented in Fig. 3. The analysis revealed that there were varied sources of these contaminants. The hexa-BDE technical formulation did not cluster with the other congeners. The total variance for these two components came up to 57.87 %. The first component accounted for 36.89 % of the recorded variance. Congeners BDE99 and BDE28 had high positive loadings of 0.845 and 0.742, respectively. The second component accounted for 20.98 % of the recorded variance, in which case, BDE47 had high positive loadings of 0.786.

Fig. 3
figure 3

PCA analysis of PBDEs in bed sediments of Asunle stream

Full size image

In addition, looking at the association pattern of the various congeners, one could suggest that various materials that contain different proportion of PBDEs may be the attributive factor that made different congeners behave differently given that they came from the same source-dump. Studies have shown that the higher the vapour pressure and the lower the octanol-air partition coefficient, the more likely the BDE congeners will volatilize (USEPA 2010). Consequently, lower brominated congener BDEs have the greatest tendency to volatilize from PBDE-containing consumer products. Coated plastics used in electrical products and other electronic wastes, polymer products such as polyurethane foam in car/transport accessories, furniture, construction, mattress and baby care products could also be sources from where PBDEs are derived (Ogungbuyi et al. 2012; Stockholm Convention 2011). Generally, the main sources of PBDEs were probably from leachates emanating from the PBDE-containing consumer products and other domestic wastes from the dumpsite that found their way into the stream.

Correlation analysis of PBDEs in bed sediment of Asunle stream

Correlation analysis of PBDE congeners in the bed sediment of Asunle stream is presented in Table 3. The results showed that there were positive correlations among the lower congeners in particular. BDE28 was significantly positively correlated with BDE47 and BDE99 at p < 0.01 and with BDE153 at p < 0.05. BDE47 correlated positively with BDE99 at p < 0.05, while BDE99 correlated positively with BDEs 100 and 153 at p < 0.01. Those PBDEs with positive correlation values are probably traceable to the same source, such as BFR-containing products in the dumpsite.

Table 3 Pearson correlation coefficient of PBDEs in bed sediment of Asunle stream and corresponding TOC levels
Full size table

The Pearson correlation carried out to establish possible relationship between target PBDE congeners and total organic carbon (TOC) in the bottom sediment of Asunle stream showed that BDE28, BDE47, BDE99, and BDE100 congeners were positively correlated with TOC, but the correlation was weak. BDE153 and BDE154 were negatively correlated with TOC, thus suggesting that the presence of organic matter might not influence their levels in the stream. However, none of the congeners had a significant correlation at either 0.01 or 0.05 levels (two tailed) with the TOC.

Comparison of PBDE levels in sediment of the study area with levels in other countries

Polybrominated diphenyl ether (PBDE) concentrations in river sediments have been measured and reported in several regions of the world. A comparison of the results obtained in this study with similar studies conducted around the world is shown in Table 4. The levels of the target compounds found in the bed sediment of Asunle stream were lower than those reported elsewhere. This may be attributed to the more pronounced industrial activities in the catchments of waste water treatment plant and high population density of the areas where the previous studies were conducted. For instance, PBDE demand figures for 2001 suggested that North America accounted for 95 % of the global penta-BDE consumption (Law et al. 2006). The two rivers of the Pearl River delta, the Zhujiang and Dongjiang, flow through the world’s most densely urbanized region of about 120 million people and a major electronic manufacturing centre. Durban Bay and most of the rivers studied in South Africa were associated with residential and industrialized areas, while Lagos Lagoon is highly polluted as a result of population density with a high percentage (over 75 %) of Nigeria’s industry located in Lagos (Friends of the Environment 2006; La Guardia et al. 2013). However, it is important to note that accurate and direct comparisons of PBDE concentration levels are impossible as a result of difference in test parameters and congeners from one study to another.

Table 4 Comparison of PBDE concentrations in the sediment (ng/g) with other studies across the world
Full size table

Findings in the present study suggested that as local demands of polymer products and electronics are increasing within the university community, there is the tendency for used polymer and electrical and electronic wastes at the dumpsite to accumulate, and hence, levels of BFR concentrations may continue to increase. This will inadvertently lead to increased contamination of Asunle water body with BFR. Hence, the rural dwellers using the water and the aquatic habitat within the stream vicinity might be negatively impacted. This underlines the need to further investigate the environmental burdens on plants and animals, as well as risks associated with BFR exposure among the rural dwellers and farmers depending on this water body for farm produce processing and drinking purposes. Although the current levels of PBDEs detected in the sediments of Asunle stream were below the levels detected in sediments of rivers located within highly urbanized regions, it is pertinent to put in place measures that could prevent heightened presence of PBDEs in the aquatic ecosystem of the study area in the future. This will go a long way to protecting those who rely on water from Asunle stream for various day-to-day uses. Trash should be properly disposed, and conventional ways of getting rid of waste should be employed. Proper leachate management is an important factor for any dumpsite; hence, a typical landfill/dumpsite should be constructed to contain leachate collection and treatment, either internally or externally.

Conclusion

The inappropriate disposal and open incineration of solid waste and e-waste have become important sources of environmental contamination and pollution by a number of toxic chemicals including PBDEs. This study is one of the few studies in Nigeria conducted on the contamination of PBDEs in stream sediment adjoining the Obafemi Awolowo University dumpsite. The results suggest that past uncontrolled waste disposal most likely resulted in the occurrence and migration of PBDEs into the surrounding environment. The leaching of PBDEs along with high levels of suspended solids into the aquatic system could have contributed to the higher levels detected and recorded in the sediment during the wet season.