**3. Results**

### *3.1. Physicochemical Characteristics of the Soil and Dust Samples*

The soil texture in Lagos and Aba is sandy loam and in Ibadan is sandy clay [42]. The physiochemical characteristics of the soil and dust samples such as pH, total organic matter content (TOC), are presented on Supplementary Table S1a, b. The pH at the burning sites ranged from 7.2 to 8.4, dismantling sites (1.9 to 9), repair sites (7.9 to 9.24), and control sites (7.4 to 9) while the TOC for the burning sites ranged from 8 to 36.3%, dismantling sites (1.6 to 24%), repair sites (1.14 to 24%), and control sites (0.98 to 5.3%). There were significant differences in the pH (*p* = 0.004) and TOC (*p* = 0.000) levels between all the sites. Also, there were significant differences in the pH (*p* = 0.006) and TOC (*p* = 0.00) between the various e-waste recycling sites.

### *3.2. PBDE Concentrations at the Various Activity Sites*

PBDE concentrations at e-waste sites and control sites are presented in Supplementary Tables S2–S4 and in Figure 3. PBDE congeners were detected in all sampling sites, indicating that PBDEs were widespread pollutants in this research area. The total concentration (∑13PBDE) ranged from 1.702 to 149,770.560 ng/g. The most abundant PBDE congener in all the sites and samples was BDE-209, with concentrations ranging from 0.850 to 147,091.400 ng/g. The maximum ∑13PBDE was found in direct dust from TV repair shops in Ibadan. This high variability in PBDE concentrations is a reflection of the activities on the sites. The abundance of the PBDE congeners (considering the median of all the samples) in all the locations are generally in this order: BDE-209 > BDE-207 > BDE-206 > BDE-183 > BDE-208, BDE-99 > BDE-153 > BDE-47 > BDE-190 > BDE-154 > BDE-100 >BDE-28 > BDE-138 >. The PBDE concentrations at the e-waste exceeded the control sites by many folds, see Figure 3 and Supplementary Tables S2–S4 for details of the exceedance (EX) levels. The general pattern of the PBDEs distribution at the e-waste sites showed concentrations in this decreasing order: burning sites > dismantling sites > repair sites > control sites. This shows that burning activities contribute most to the PBDE concentrations in the environmental matrices, as is also clear from the patterns in ∑13PBDE (Figure 3). In all the sample types, the highest concentrations were found in Lagos.

**Figure 3.** Exceedances of ∑13PBDE concentrations in the samples from the e-waste recycling sitescompared to the control sites across locations.

### *3.3. Assessment of the Top Soil Samples*

The one-way ANOVAs showed a significant difference in the concentration of ∑13PBDE and all the PBDE congeners, except BDE-28, in top soils between the activity sites, (*p* = 0.05). Post-hoc comparisons indicated the main differences is between control and burning sites, and control and dismantling sites for all the PBDE congeners and ∑13PBDE, see Figure 4 and Supplementary Table S5.

### *Int. J. Environ. Res. Public Health* **2019**, *16*, 360

**Figure 4.** PBDE concentrations in top soils across the sites. The PBDE congeners influenced by activity at the sites are indicated with \*, and an additional \* for those influenced by location, and # for interaction between activity and location.

While the type of activities on the sites influences the PBDE concentrations on the sites, that effect might differ across locations. To test this assumption, a two-way ANOVA was performed and this ANOVA confirmed a significant difference in all the PBDE concentrations between the activity sites, with the burning sites having the highest concentrations, followed by dismantling sites, then control sites. Forty-one to 100% of the variability in PBDE concentrations was accounted for by the activities at the sites. However, there was no significant difference in the concentration of any of the PBDE congeners between locations, except for BDE-28, 208, and 207. The PBDE concentrations were generally highest in Aba, followed by Lagos, then Ibadan. This excludes BDE-28, which is highest in Lagos. Also, the interaction of activities at the sites and location showed no significant difference on any of the PBDE congeners and ∑13PBDE concentrations, except for BDE-28 (see Supplementary Table S6).

### 3.3.1. Assessment of the Floor Dust Samples

A one-way ANOVA showed a significant difference (*p* = 0.05) in the concentration of some of the PBDE congeners (BDE-99, 47,100, 154, 190, 138) in floor dust between dismantling and repair sites (Figure 5).

**Figure 5.** PBDEs concentration in floor dust across the sites. The PBDE congeners influenced by activity at the sites are indicated with \*.

The two-way ANOVAs testing individual PBDE congeners and total PBDE ( ∑13PBDE) concentrations in floor dust from the activity sites (dismantling and repair sites) and across the locations (Lagos, Ibadan, and Aba), showed significant differences in the concentration of some PBDE congeners (except BDE-47, 100, 99, 154, 138, 190) between the activity sites, with dismantling sites having higher concentrations than repair sites. Also, there were significant differences in the concentration of some PBDE congeners (BDE-47, 100, 99) between locations, with Ibadan having generally the highest PBDE concentrations, followed by Lagos, then Aba. The interactions of activities at the sites and location showed no significant difference on any of the PBDE congeners and ∑13PBDE concentrations, except for BDE-154 (see Supplementary Table S7).

### 3.3.2. Assessment of the Roadside Dust Samples

Roadside dust samples were collected only from Lagos and Aba. One-way ANOVA showed a significant difference (*p* = 0.05) in the concentration of almost all the PBDE congeners in roadside dust between dismantling and control sites (Figure 6).

**Figure 6.** PBDE concentration in roadside dust across the sites. The PBDE congeners influenced by activity at the sites are indicated with \*.

### 3.3.3. Assessment of the Direct Dust Samples

Direct dust samples from electronics were collected from dismantling and repair sites from Ibadan only. A one-way ANOVA showed no significant difference in the concentration of any of the PBDE congeners in direct dust between dismantling and repair sites, except BDE-100, (Figure 7).

**Figure 7.** PBDE concentration in direct dust from electronics at the sites. Generally, there were no significant differences in the PBDE concentrations, except for BDE-100.

### *3.4. Patterns in PBDEs Contamination*

The principal component analysis (PCA) using direct oblimin rotation was performed on the correlation matrix of the PBDE concentrations to establish whether the contaminants were actually arising from the same source or not. The analysis revealed one common axis of variation in PBDE concentrations, which accounted for 84% of the total variance. All PBDEs varied in the same direction (Figure 8). All the PBDE congeners had high positive loadings of ≥0.803. These findings indicate that PBDEs contamination has one common driver, which might sugges<sup>t</sup> on common source. Pearson correlation confirmed that all PBDE congeners at the e-waste sites strongly correlated positively with each other. There was also correlation between the TOC and all the individual PBDEs and ∑13PBDE (Supplementary Table S8).

**Figure 8.** PCA plot of PBDE concentrations.
