Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 8.
Published in final edited form as: Sci Total Environ. 2022 Aug 2;850:157705. doi: 10.1016/j.scitotenv.2022.157705

Occurrence and spatial distribution of individual polychlorinated biphenyl congeners in residential soils from East Chicago, southwest Lake Michigan

Andres Martinez a, Jason BX Hua a, Ezazul Haque b, Keri C Hornbuckle a, Peter S Thorne b,c
PMCID: PMC9907466  NIHMSID: NIHMS1859447  PMID: 35931157

Abstract

We report individual polychlorinated biphenyl congeners and the sum of all congeners (ΣPCB) in residential soils of East Chicago, Indiana. ΣPCB in soils ranged from 20 to 1700 ng/g dry weight (DW), with a geometric mean of 120 ng/g DW. These values are significantly higher than other locations, but similar or lower to locations nearby well-known PCB contamination sites. No PCB spatial distribution pattern was observed. PCB concentrations increase with total organic carbon in the soils and proximity to Indiana Harbor and Ship Canal (IHSC), where sediments are contaminated with PCBs. Most samples are similar in their PCB distribution and Aroclor 1254 yielded the highest similarity to all the samples. A fifth of the samples highly resemble other PCB profiles such as EPA background and Cedar Rapids Iowa soils, and volatilization from Lake Michigan, whereas volatilization from IHSC could not explain the PCBs found in soils. IHSC was expected to be the main source of PCBs in the nearby soils. It is possible that soils are impacted by variety of known and unknown sources, including volatilization from Lake Michigan, resulting in a regional PCB signal. Although PCB concentrations are higher than other locations, samples were below the current US EPA non-cancer residential soil level remediation goal for dioxin TEQ.

1. Introduction

Although polychlorinated biphenyls (PCBs) were banned > 40 years ago and control regulations have been implemented in most industrialized countries, PCBs are still found in the environment and in humans at measurement levels. It is well known that soils are one of the largest reservoirs of PCBs. Meijer et al. (2003) reported that the global burden of PCBs in soils is at least 21,000 t, although the study only focused on background surface soils. Due to their resistance to biological, chemical and thermal degradation, as well as continuing releases from diffusive and new sources (e.g., pigment and silicone use and manufacturing (Jahnke and Hornbuckle 2019; Hombrecher et al. 2021)), PCBs persist in the soils for a long period of time. Indeed, PCB half-life in soils ranges from years to decades (Jones and de Voogt 1999).

Residual levels of PCBs in urban soils, especially soils located in residential areas and places where children and other vulnerable populations live, are of concern due to human exposure on a regular basis (Li et al. 2018). Levels and distributions of PCBs in urban areas are very difficult to generalize. Variables such as industrial and land use history, age of contamination, chemical properties of the soils, meteorological conditions and soil movement/replacement can greatly affect the levels and distribution of PCBs in urban soils. A few studies have shown that in an urban setting, heterogeneity is common: nearby samples can show differences of > 2 orders of magnitude and different PCB congener distributions (Martinez et al. 2012; Vane et al. 2014).

We previously investigated PCBs in soils in Cedar Rapids, Iowa, where no records of production, heavy use, or concentrated disposal (i.e., Superfund Sites) of PCBs are known (Martinez et al. 2012). Soil PCB levels were significantly higher than background values, with an average of total PCBs of 4.0 ± 5.5 ng/g DW for the US (USEPA 2007) and similar or higher than many cities around the world. As expected, soils located nearby a known PCB contamination, such as a PCB Superfund sites, are much higher than background levels, but available soil data are limited.

East Chicago, Indiana, has a long history of heavy industrial operations, including steel mills, refineries, foundries, and lead smelters (Haque et al. 2021b). These activities are historical sources of heavy metals, polycyclic aromatic hydrocarbons and PCBs. Further, the IHSC, which is located in the center of East Chicago and connects the Grand Calumet River to Lake Michigan, is undergoing a navigational dredge. The PCB contaminated sediments are being deposited into a confined disposal facility (CDF) located in East Chicago just 750 m from East Chicago Central High School. Through the Airborne Exposure to Semi-volatile Organic Pollutants (AESOP) Study, a longitudinal cohort study of adolescent children and their mothers, we have extensively studied the presence of PCBs in the indoor and outdoor air at homes and schools (Ampleman et al. 2015; Marek et al. 2017), in water and sediment from the IHSC (Martinez et al. 2010; Martinez et al. 2010; Martinez et al. 2015), as well as in blood from participants living in East Chicago (Koh et al. 2015). Recently, we have also investigated the concentration of toxic metals such as lead, arsenic and manganese in AESOP Study participants’ blood (Haque et al. 2021a; Haque et al. 2022) and in East Chicago soil samples (Haque et al. 2021b). It is unclear if PCBs from sediments in the IHSC or CDF can impact residential soils in East Chicago.

We hypothesized that concentrations of PCBs in East Chicago soils were entirely, or primarily, due to deposition of PCBs emitted from the nearby Indiana Harbor and Ship Canal. We also hypothesize that PCB soil levels in East Chicago are higher than in other locations without history of PCB contamination, but similar to places with known PCB contamination. To address these hypotheses, the following aims were accomplished: (i) determined the occurrence and distribution of PCBs in soils collected in residential areas of East Chicago; (ii) investigated if distance to the IHSC contributes to the levels and distribution of PCBs in soils; and (iii) analyzed the PCB congener distribution and compared it with potential PCB source signatures, including the PCB distribution found in IHSC.

2. Material and method

2.1. Soil sampling method.

Thirty-three surficial soil samples were collected in East Chicago between 2017 (July) and 2018 (August) following the method described elsewhere (Haque et al. 2021b). The sites sampled were yards in residential areas throughout a 26 km2 area of the city (Fig. S1). The sampling locations were selected via a random points generator. Some sites were substituted with other points close in proximity when locations were not accessible or did not contain soil. The top 3 cm of soils were collected with a hand trowel, homogenized and stored in Ziploc bags. We chose the top 3 cm since we were interested in looking at the most likely portion of soil that can become wind-blown dust and impact children (Haque et al. 2021b). The samples were brought to The University of Iowa and kept in a freezer at −10 °C until extraction.

2.2. Analytical method

2.2.1. PCB Analysis.

The analytical methods employed here are described elsewhere (Martinez et al. 2010; Martinez et al. 2012; Martinez et al. 2016). Homogenized soil samples were weighed (~2.5 g) and mixed with (~5 g) combusted diatomaceous earth (DE, Thermo Fisher Scientific). Each sample was spiked with 50 ng surrogate standard chlorobiphenyls (CB), PCB 14 (3,5-di-CB), d-PCB 65 (2,3,5,6-tetra-CB-d5, deuterated), and PCB 166 (2,3,4,4’,5,6-hexa-CB). The mixed soils were extracted via pressurized fluid extraction (Accelerated Solvent Extractor, Dionex ASE-300), with equal parts hexane and acetone. The solution was concentrated using a Caliper TurboVap II to 1 mL. Each sample was liquid-liquid extracted with a mixture of sulfuric acid and ethanol (2:1 v/v) and left overnight. The hexane extracts were passed through a Pasteur pipette filled with 0.1 g of combusted silica gel and 1 g of acidified silica gel (2:1 silica gel:sulfuric acid by weight) and eluted with 10 mL of hexane. The solution was concentrated to 0.5 mL by TurboVap. Fifty ng of d-PCB 30 (2,4,6-tri-CB-2′,3′,4′,5′,6′-d5, deuterated) and PCB 204 (2,2’,3,4,4’,5,6,6’-octa-CB) were spiked as internal standards. Table S1 shows the individual or coeluting PCB congeners target in this analysis, including their corresponding homolog groups.

Gas chromatography (GC) in tandem with Mass Spectrometry GC-MS/MS (Agilent 7000) in multiple reaction monitoring (MRM) mode was used to quantify all 209 congeners in 171 individual or coeluting congener peaks (see Table S1). The GC was equipped with a Supelco SBP-Octyl capillary column (Poly(50% n-octyl/50% methyl siloxane, 30 m × 0.25 mm ID, 0.25 μm film thicknesses)) with UHP helium as the carrier gas (0.8 mL/min) and UHP nitrogen as the collision gas (1.5 mL/min). The GC operated in solvent vent injection mode at the following injection conditions: initial temperature 45°C, initial time 0.06 min, ramp 600°C/min to inlet temperature 325°C at 4.4 psi. The GC oven temperature program was 45°C for 2 min, 45 to 75°C at 100°C/min and hold for 5 min, 75 to 150°C at 15°C/min and hold for 1 min, 150 to 280 at 2.5°C/min and final hold 5 min (total run time 70.86 min). The triple quadrupole MS electron ionization source was set to 260°C.

2.2.2. TOC analysis.

The total organic carbon (TOC) soil content was analyzed for 28 of the 33 soil samples, using a Shimadzu Total Organic Carbon Analyzer (TOC-V Series) and Solid Sample Module (SSM-5000A). A ceramic sample boat was filled with about 0.2 g of soil and combusted at 900 °C to produce CO2, which was detected by a non-dispersive infrared gas analyzer. Reagent grade glucose was used to calibrate the instrument. TOC measurements followed the standard operation procedure described by the manufactured. The measured TOC was divided by the initial sample mass and reported as a percentage.

2.2.3. Water content.

The soil water content was determined gravimetrically from a separate soil aliquot of the 33 samples. Around 1–2 g of soil was placed in an aluminum weigh boat and dried for 12 h at 104 °C. The samples were allowed to cool in a desiccator for at least 1.5 days, and then their masses were measured. The percent dry weight was calculated using the final mass divided by the initial mass times 100.

2.3. Quality assurance and control.

The quality of the data was assessed using method blanks, surrogate PCB standards, method or laboratory blanks, replicates, and a standard reference material (SRM, 1944, National Institutes of Standards and Testing). The mean and standard deviation of PCB 14, d-PCB 65, and PCB 166 recoveries were 76.0 ± 14.3%, 82.4 ± 12.0% and 106.2 ± 19.0%, respectively. PCB congener masses were corrected using the fractional recovery of PCB 14 (congener 1–39), d-PCB 65 (congeners 40–128), and PCB 166 (128–209). A total of 8 method blanks, i.e., combusted DE (1 blank per batch), were included during extraction, clean up and quantification of the samples. The masses from the blanks were consistent with a log-normal distribution. Therefore, the individual congener limit of quantification (LOQ) was calculated as the upper limit of the 95% confidence interval of the log10 transformed method blanks. The raw (i.e., untransformed) congener-specific LOQ ranged from 1.0 pg/g to 0.9 ng/g of DE with an average of 0.03 ng/g of DE. Any congener mass value below the LOQ was replaced by LOQ/2. A sample (EC3-S-1253) was extracted and quantified three times yielding a relative standard deviation (RSD) of 4%. SRM was sampled and analyzed to test the accuracy of our method. Five SRM were extracted and quantified using the same method for soils. The percent recovery of our measured values against the certified values for the 28 PCB congeners reported yielded a mean of 80 ± 1% (Table S2). Although we quantified 171 individual or coeluting congener peaks, further evaluation focused on 128 individual congeners with ≥ 75% detection frequency (Table S1). TOC measurement of a sample (EC3-S-1257) was performed five times yielding an RSD of 76%.

2.4. Data and Statistical Analysis.

PCB soil measurements were consistent with a log-normal distribution, and thus they were log10-transformed for future analysis. ANOVA and Tukey test for multiple pairwise comparisons were utilized to determine statistical significance between sampling sites. To evaluate potential covariances that can explain the levels of PCBs found in the soils, we performed linear and multiple linear regressions, using the PCB log10-transformed data, and TOC and distance to the IHSC Fork as independent variables. To evaluate similarity between PCB conger profiles, cluster analysis such as Principal Component Analysis (PCA) and cosine theta (cos θ) were used. Cos θ is the cosine of the angle between two multivariable vectors (PCB congener profiles); a value of 0.0 indicates that the two vectors are completely different (perpendicular) and 1.0 describes two identical vectors (DeCaprio et al. 2005; Martinez et al. 2012; Martinez et al. 2016). All statistical analyses were performed using R (Version 3.6.0). Concentration of individual PCB congener in soils, mases of individual PCB congeners detected in blanks, concentrations of individual PCB congeners measured in SRM, distance from samples to the Fork, and TOC measurements are published open access at https://doi.org/10.1594/PANGAEA.941894 (Martinez et al., 2022). Further, all the R codes created here to fabricate all the plots and maps, and perform the statistical analyses are freely available at: doi: https://doi.org/10.5281/zenodo.6984449 (Martinez, 2022).

3. Results and discussion

3.1. Concentration and spatial distribution of PCBs in soils.

The ΣPCB (sum of 128 congeners) concentration in East Chicago soils range from 20 to 1700 ng/g DW, and followed a log-normal distribution (GM=120 ng/g DW, median=93 ng/g DW) (Fig. 1, Fig. 2, bottom panel). Most of the samples fell in the range of 50–100 ng/g DW, where only three samples were above 500 ng/g DW (Fig. 1, Fig. 2, bottom panel). We observed no clear evidence of spatial correlation of ΣPCB, where samples nearby (~ 50 m) show a large difference in concentration, such as locations s8 and s10 (Fig. 1, Fig. 2, bottom panel).

Figure 1.

Figure 1

Open in a new tab

ΣPCB concentration in East Chicago soils in ng/g DW. The size of the circle indicates the concentration in the soil and the numbers the concentration in ng/g DW.

Figure 2.

Figure 2

Open in a new tab

Concentration of PCBs 8, 11, 52, 136, 206, 209 and ΣPCB in East Chicago soils in ng/g DW, where left to right sites in the x-axis are approximately organized from north to south and from east to west. Grey vertical lines separate clusters of sites less than 500 m of distance. Note the differences in the y-axis scale.

Individual congeners ranged from non-detected to 180 ng/g DW, with GMs varying from 0.01 to 6.4 ng/g DW (Fig. 3). Likewise, both the ΣPCB and individual congeners were log-normally distributed. We observed no relationship between concentration level and congener chlorination. For example, PCBs 8 and 209 yielded similar GM of 0.5 and 0.8 ng/g DW, respectively. Non-Aroclor PCBs 11, 35, 55, 67, 68, 73, 89, 94, 131, 133, 152 and 175 were also detected (< 0.2% wt% in original Aroclor mixtures) (Rushneck et al. 2004; Koh et al. 2015).

Figure 3.

Figure 3

Open in a new tab

Summary of soil concentration of individual PCB congeners from East Chicago. A total of 33 samples and 128 individual or coeluting PCB congeners are plotted. The box plots include the maximum, 75th percentile, median, 25th percentile and the minimum value.

Spatial distribution of congeners from the same homolog group were similar in spatial distribution. For example, most of the mono-CB to tetra-CB congeners were similar in spatial distribution to ΣPCB, such as di-PCB 8 and tetra-PCB 52 shown in Fig. 2. Penta-CB to hexa-CB were similar to each other as were hepta-CB to nona-CB, but not to ΣPCB (e.g., hexa-PCB 136 and nona- PCB 206 shown in Fig. 2). There were a few congeners that did not resemble any other spatial distribution, including non-Aroclor PCB 11 and PCB 209 as shown in Fig. 2. These disparities in the spatial distribution can result from how each congener partitions between soil and air, where the octanol to air partition coefficients vary from > 4 log10 units, and also from different PCB sources impacting the soil.

3.2. Comparison of PCB concentrations with locations around the world.

Comparing PCB soil levels reported by others (Carey et al. 1979; Vorhees et al. 1999; Meijer et al. 2003; Blankenship et al. 2005; USEPA 2007; Tang et al. 2010; Vane et al. 2014; Yu et al. 2020) is not straightforward due to the difference in the sampling methods carried out, e.g., random vs. selective sampling, rural vs. urban vs. industrial, analytical methods, total number of PCB congeners quantified, and the sampling year. Nevertheless, we compared the East Chicago ΣPCB data with Cedar Rapids, IA (Martinez et al. 2012), London, UK (Vane et al. 2014), 29 provinces and cities of China (Yu et al. 2020), and background levels from the US (USEPA 2007) and globally (Meijer et al. 2003). For all comparisons, we use the log10-transformed ΣPCB concentrations. First, we found significant differences among the locations (ANOVA, p < 0.05) and second, through a multiple comparison corrected test for differences among locations, we found East Chicago to be significantly higher than the rest of the cities, including Cedar Rapids, London, and Chinese cities with the exception of Wenling, and both background levels (Tukey test, p < 0.05) (Fig. 4). The relative high levels found in East Chicago were as expected due to the nearby sediment PCB contamination from the IHSC (Martinez et al. 2010; Martinez et al. 2010). The East Chicago soils are not as high as sites near known PCB contamination sites, such as Gadsden, New Bedford, Wenling and Kalamazoo, where our values are in the same range or even 1000-fold lower than those sites (Fig. 4). Although there are dissimilarities among the sampling methods, analytical methods and year of sampling, these results suggest that soils near contaminated sites generally present higher levels of PCBs. Interestingly, PCBs from three of these highly contaminated sites (i.e., East Chicago, Kalamazoo and New Bedford) are currently present in submerged sediments.

Figure 4.

Figure 4

Open in a new tab

Comparison of East Chicago ΣPCB soil levels against other locations around the world, background levels and known contaminated PCB soil locations. Wenling, Gadsden, New Bedford and Kalamazoo are displayed as arithmetic means due to available published data. A significant difference was found between the log10 concentrations and the first 3 locations (a to c), and both background levels (ANOVA, p < 0.05) to East Chicago. The asterisk (*) indicates a significant difference among East Chicago and the first three locations (a to c), and both background levels (Tukey test, p < 0.05, see text). The box plots include the maximum, 75th percentile, median, 25th percentile and the minimum value. Refs: Cedar Rapids (Martinez et al. 2012), China (Yu et al. 2020), London, UK (Vane et al. 2014), Global background (Meijer et al. 2003), USA background (USEPA 2007), Gadsden (Carey et al. 1979), New Bedford (Vorhees et al. 1999), Wenling (Tang et al. 2010) and Kalamazoo (Blankenship et al. 2005).

3.3. TOC and distance to the IHSC Fork regressions.

Total organic carbon or TOC is a well-known surrogate for organic matter present in soils samples, and it is frequently one of the main drivers in the occurrence and levels of PCBs found in soils (Meijer et al. 2002; Meijer et al. 2003). TOC levels in East Chicago soils range from 1.3 to 27%, with a GM of 6.1%. We suspect that those few high value (> 15%, n = 4) reflect accumulation of high carbon dust from coal storage and use at a large steelmaking facility in East Chicago. Further, US EPA has been excavating and replacing soils in East Chicago as a lead remediation strategy, which could also alter the levels of TOC found in these soils. We found that log10 concentration of 95% of all congeners and ΣPCB are positively correlated with TOC (p < 0.05), resolving in an average of 27 ± 8 % of the variability in the soil concentration (See Table S2 for individual and ΣPCB results). Fig. 5 shows selected congeners from different homolog groups (di-, tetra-, hexa- and deca-), but also a few non-Aroclor congeners, and ΣPCB regressions against TOC.

Figure 5.

Figure 5

Open in a new tab

Plots of log10 concentration against TOC of selected congeners and ΣPCB (n = 28). m and p correspond to the slope and p-value of the linear regression, respectively. All correlations are significant (p < 0.05). Note the difference in the y-axis scale.

Distance from the soil sampling location to the IHSC Fork was also evaluated as a covariate for the PCB concentration in soils (Fig. 1). Only 46% of congeners and ΣPCB yielded a negative correlation (p < 0.05), averaging only 14 ± 4% of the total variability (See Table S3 for individual and ΣPCB results). We determined that most of the hepta-CB to deca-CB were significantly correlated with distance to The Fork, whereas only five tetra-CB were significantly correlated (PCBs 56, 60, 63, 66 and 67). Fig. 6 shows selected congeners and ΣPCB regressions against distance from the Fork. For these 59 congeners and ΣPCB, this finding indicates that soils nearby the IHSC present with higher values of PCBs.

Figure 6.

Figure 6

Open in a new tab

Plots of log10 concentration against distance to the Fork of selected congeners and ΣPCB (n = 33). Although the correlations are low, they are all statistically significant (p < 0.05). Note the difference in the y-axis scale. m and p correspond to the slope and p-value of the linear regression, respectively.

Combining both covariates TOC and distance to the Fork into a log10 PCB concentration model (multiple linear regression, MLR), yielded stronger correlations than the individual TOC and distance models (See Table S4 for individual and ΣPCB results). Here, 95% of congeners and ΣPCB yielded a significant correlation (p < 0.05), averaging 30 ± 8% of the total variability, where all the congeners and ΣPCB increased their R2 in relation to the other two linear models. The MLR model indicates that as more TOC in the soil and a location closer to the Fork yields higher PCB concentrations in the soils.

3.4. PCBs and heavy metals in soils.

We have previously measured calcium, chromium, copper, iron, potassium, manganese, lead, rubidium, sulfur, strontium, titanium, zinc and zirconium in the same soil samples in East Chicago (Haque et al. 2021b). In general, no significant correlations among PCBs and these metals were found, with a few exceptions such as Cr, Fe, Ti and Zr. Iron was positively correlated with concentration of ΣPCB and 29% of congeners (R2 ranged from 0.1 to 0.3, p < 0.05). Chromium was positively correlated with log10 concentration of 60% of congeners (R2 ranged from 0.12 to 0.3, p < 0.05). Titanium was positively correlated with log10 concentration of 35% of congeners (R2 ranged from 0.1 to 0.24, p < 0.05). Zirconium was positively correlated with log10 concentration of 30% of congeners (R2 ranged from 0.1 to 0.2, p < 0.05). From all congeners, PCBs 105, 107, 108+124, 118, 170 and 177 were significantly correlated with Fe, Ti and Zr. Overall, these results show that the association between both PCBs and metals in East Chicago soils is metal-congener-specific, where different sources or their chemical properties differently affect their fate in the environment.

3.5. PCB source analysis.

The soil congener profiles are mostly dominated by congeners with four to six chlorines, including PCBs 52, 61+70+74+76, 90+101+113, 95, 110, 118, 129+138+163 and 153+168, with an average fraction of at least 0.03. But also, there are a few tri-CB, nona-CB and deca-CB. Most congeners only contribute a relatively small amount to the variability of total PCB, with the exception of PCBs 52, 61+70+74+76, 95 and 129+138+163 (standard deviation (std dev) ≈ 0.024). This is shown in Fig. 7 by the relatively large error bars for PCBs 52, 61+70+74+76, 95 and 129+138+163. Further, non-Aroclor congeners 11, 35, 55, 67, 68, 73, 89, 94, 131, 133, 152 and 175 (Rushneck et al. 2004; Koh et al. 2015) contribute very little to the total PCB, with an average fraction for these congeners of only 0.0006, with the exception of PCB 11, with an average fraction of 0.003.

Figure 7.

Figure 7

Open in a new tab

Average PCB congener profile in mass fraction of ƩPCB (n = 33) using untransformed data in the soil. Error bars represent one standard deviation above the average.

An initial cluster analysis (PCA) of the untransformed data, which also included Aroclors mixtures (Frame et al. 1996; Koh et al. 2015), EPA soil PCB background (USEPA 2007), Cedar Rapids, IA (Martinez et al. 2012), PCB emission from IHSC (Martinez et al. 2018; Martinez et al. 2019), East Chicago PCB airborne (Marek et al. 2017) and PCB volatilization from Lake Michigan (Boesen et al. 2020) profiles, resulted in one cluster (elliptic) containing all the soil samples, together with Aroclors 1254, 1248 and 1260, EPA soil background, PCB emission from IHSC, Cedar Rapids soil, volatilization from Lake Michigan, and at the edges of the cluster, Aroclors 1016, 1242 and 1262 (Fig. 8), and East Chicago PCB airborne. If Aroclors 1221 and 1232 are removed and this analysis is repeated, very similar results were obtained (Fig. S4). Although the first two principal components explained 67% of the total variance, this preliminary analysis suggest that all the soil samples are potentially similar in their PCB distribution and that Aroclors 1248, 1254 and volatilization from Lake Michigan, and to a lesser extent Aroclors 1260 and 1262, and the emissions from IHSC are potential sources of the PCB congeners found in the soils. Further, the soil samples seem to show similarities with EPA background soil samples and our previous samples from Cedar Rapids soils.

Figure 8.

Figure 8.

Open in a new tab

Biplot of PC1 and PC2 for the PCB profiles of the East Chicago soil samples, commercial Aroclor mixtures, EPA background level (EPA soil, USEPA (2007)), Cedar Rapids, IA (CR soil, Martinez et al. (2012)), East Chicago PCB airborne (EC air, Marek et al. (2017)), PCB emissions from IHSC (Emission IHSC, Martinez et al. (2018) & Martinez et al. (2019)) and volatilization from lake Michigan (Volatilization LM, Boesen et al. (2020)). Soils samples from East Chicago are not labeled. Asterisk (*) in the Aroclor number means data obtained from Frame et al. (1996), whereas the other Aroclors were analyzed in our laboratory (Koh et al., 2015). A 67 % of total variance is explained by the first two PC.

We further evaluated the statistical similarity of the PCB profiles using cos θ. Firstly, we investigated the average profile of the East Chicago soil samples (Fig. 7). The East Chicago soil congener average profile closely resembles Aroclor 1254 with a cos θ of 0.90. There is also some resemblance to Aroclor 1248 (cos θ = 0.70) and, to a lesser extent, to Aroclors 1242 (cos θ = 0.46) and 1260 (cos θ = 0.57). Although these last two Aroclors do not show much similarity with the soil average profile, these Aroclors are potential sources of the low and high chlorinated congeners present in the soils. Further, the average profile is similar to the EPA background measurement with a cos θ = 0.81 and to the Cedar Rapids soils, with a cos θ = 0.77, but also to the Lake Michigan volatilization profile (cos θ = 0.74). The air profile in East Chicago is slightly similar to the soil profile, with a cos θ = 0.71, but not similar to the emission’s profile from IHSC (cos θ = 0.52).

Secondly, even though the concentration levels are different for nearby samples, the PCB distribution is very similar among nearby samples, with an average of cos θ of 0.81 ± 0.15, but ranging from 0.29 to 0.99. Mostly, nearby samples yielded higher average cos θ than the overall average, such as samples s26 to s29 (< 48 m between the samples) and s30 to s32 (90 m) (see Fig. 1), with an average cos θ of 0.97 ± 0.01 and 0.94 ± 0.02, respectively.

Comparisons of individual samples against Aroclor mixtures resulted in a 40% of the samples very similar to Aroclor profiles (cos θ > 0.90). For example, sample s9 and s2 yielded a cos θ of 0.99 and 0.98 against Aroclors 1248 and 1254, respectively. The highest concentration samples yielded the highest similarity with Aroclor 1248. That is, samples s18 (1700 ng/g DW) and s10 (1300 ng/g DW) yielded a cos θ of 0.90 and 0.95, respectively. Interestingly, surficial sediment from the IHSC resembles very well Aroclor 1248 (Martinez et al. 2010). Overall, 70% of the samples most closely resembled Aroclor 1254 (cos θ > 0.71), 21% of samples best resembled Aroclor 1248 (cos θ > 0.78) and only 9% of the samples resembled Aroclor 1260 (cos θ > 0.78).

Twenty percent of the samples most closely resembled the other selected profiles: that is EPA background, Cedar Rapids, PCB emission from IHSC, East Chicago PCB airborne and volatilization from Lake Michigan (cos θ > 0.90). For example, samples s14 and s15 yielded a cos θ of 0.96 and 0.95 against Cedar Rapids and EPA background profiles, respectively. The two highest concentration samples show a strong similarity to the Lake Michigan volatilization profile (cos θ ≈ 0.75). Overall, 21% of the samples yielded highest similarity to Cedar Rapids profile (cos θ > 0.77), 45% of samples best resembled EPA background profile (cos θ > 0.72) and 33% of samples best resembled the profile from the Lake Michigan volatilization (cos θ > 0.70). These similarity results against the EPA and Cedar Rapids profiles were not expected, due to the relatively high levels of PCBs found in the soils and also due to the presence of a contaminated PCB site nearby (IHSC). Further, the similarity of the soil samples to the volatilization profile from Lake Michigan and not to the emissions from IHSC suggest that Lake Michigan could be a more important source to soils nearby the lake and overwhelms the signal from IHSC.

It is also interesting to notice that the airborne profile from East Chicago is more similar to the Lake Michigan volatilization profile, with a cos θ = 0.87, than the emissions from IHSC (cos θ = 0.77). One potential source of non-Aroclor PCB 11 found in the soil can be Lake Michigan, where PCB 11 contribution to the total PCB volatilization profile is 5.8%. Similarly, the East Chicago air profile contains 2.5% of PCB 11. For the rest of the non-Aroclor PCBs detected in the soils, the Lake Michigan volatilization or the East Chicago air profiles are much less likely as potential sources, and these congeners are < 0.2% of the total. Another probable source for these non-Aroclor congeners could be residues from industrial and house paint applications, especially for the more highly chlorinated PCB congeners.

3.6. Toxic equivalents (TEQs).

We calculated the TEQs using the WHO-human toxic equivalents factors (TEFs) for 11 of the 12 dioxin-like PCBs (Van den Berg et al. 2006) that are individually quantifiable with our analytical method (PCB 156+157 coelute). See details in the Supplementary Data. Soil samples ranged from 0.03 to 5.7 pg TCDD TEQ/g DW, yielding an average of 0.5 ± 1 pg TCDD TEQ/g DW. PCB 118 contributed ~ 50% of total TEQ in the soils. Because PCB 118 is an Aroclor congener, a strong correlation was obtained between TEQ and ΣPCB (R2 = 0.89, p < 0.05). Importantly, all East Chicago samples were below the current US EPA non-cancer residential soil regional screening level remediation goal for dioxin TEQ of 51 pg TCDD TEQ/g DW (USEPA 2021). We notice that currently there is no available cancer screening level for PCBs in soils. Note that no measurements of dioxins (polychlorinated dibenzo-p-dioxins and dibenzofurans) were conducted on East Chicago soils, which could significantly increase the values of TEQ calculated by roughly 70% (USEPA 2002).

4. Conclusions

Our findings do not support our initial hypothesis. The spatial distribution, magnitude, and PCB congener patterns indicate a variety of sources of PCBs impacting the soils in East Chicago. Our findings are consistent with our previous studies of airborne PCBs in the region (Martinez et al. 2015). PCB concentrations in soils from East Chicago are higher than other urban industrial locations around the world, but at the same level or lower than other locations adjacent to known PCB contaminated sites. TOC and proximity to the IHSC were determined to be predictors of the levels of individual congeners and ΣPCB found in the soils, but no spatial correlation between sites was observed. In general, PCBs did not correlate with metals, with a few exceptions, such as Cr, Fe and Ti. Non-Aroclors PCBs 11, 35, 55, 67, 68, 73, 89, 94, 131, 133, 152 and 175 were detected, where one possible explanation of PCB 11 could be atmospheric deposition from PCB volatilization from Lake Michigan. Concentration of PCB 11 in the soil samples were too low to consider a paint leakage or paint residues at the sampling site. Interestingly, a few samples resembled a USA background profile, even though the levels from East Chicago are at least 2 orders of magnitude higher. We speculate that soils are impacted by multiple sources and from atmospheric regional transport, resulting in a similar distribution of PCB congeners, independent of proximity to a known PCB source.

Supplementary Material

Appendix A. Supplementary data.

Acknowledgments

We thank the Superfund Research Program of the National Institute of Environmental Health Sciences, Grant NIH P42ES013661 and the University of Iowa’s Environmental Health Sciences Research Center, Grant NIH P30ES005605 for funding. The authors thank the members of the East Chicago Community Advisory Group (CAG), Debbie Chizewer from Northwestern University’s Environmental Advocacy Clinic, Denise Abdul-Rahman from NAACP, and Sarah Rolfes from US EPA Region 5 for their collaborative assistance in working within the East Chicago community. We also thank Drs. Benjamin Bostick and Andrea Adamcakova-Dodd for assistance with soil sample collection, Deborah Williard for managing the analytical laboratory, and Dr. Panithi Saktrakulkla for laboratory assistance. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Supplementary data

Supplementary data to this article can be found online at https://

References

  1. Ampleman MD, Martinez A, DeWall J, Rawn DFK, Hornbuckle KC and Thorne PS (2015). “Inhalation and Dietary Exposure to PCBs in Urban and Rural Cohorts via Congener-Specific Measurements.” Environmental Science & Technology 49(2): 1156–1164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Blankenship AL, Zwiernik MJ, Coady KK, Kay DP, Newsted JL, Strause K, Park C, Bradley PW, Neigh AM, Millsap SD, Jones PD and Giesy JP (2005). “Differential accumulation of polychlorinated biphenyl congeners in the terrestrial food web of the Kalamazoo River superfund site, Michigan.” Environmental Science & Technology 39(16): 5954–5963. [DOI] [PubMed] [Google Scholar]
  3. Boesen AC, Martinez A and Hornbuckle KC (2020). “Air-water PCB fluxes from southwestern Lake Michigan revisited.” Environmental Science and Pollution Research 27(9): 8826–8834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Carey AE, Douglas P, Tai H, Mitchell WG and Wiersma GB (1979). “Pesticide-residue concentrations in soils of 5 United States cities, 1971 - Urban Soils Monitoring Program.” Pesticides Monitoring Journal 13(1): 17–22. [PubMed] [Google Scholar]
  5. DeCaprio AP, Johnson GW, Tarbell AM, Carpenter DO, Chiarenzelli JR, Morse GS, Santiago-Rivera AL, Schymura MJ and Akwesasne Task Force E (2005). “Polychlorinated biphenyl (PCB) exposure assessment by multivariate statistical analysis of serum congener profiles in an adult Native American population.” Environmental Research 98(3): 284–302. [DOI] [PubMed] [Google Scholar]
  6. Frame GM, Cochran JW and Bowadt SS (1996). “Complete PCB congener distributions for 17 aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congener-specific analysis.” Hrc-Journal of High Resolution Chromatography 19(12): 657–668. [Google Scholar]
  7. Haque E, Moran ME and Thorne PS (2021a). “Retrospective blood lead assessment from archived clotted erythrocyte fraction in a cohort of lead-exposed mother-child dyads.” Sci Total Environ 754: 142166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Haque E, Moran ME, Wang H, Adamcakova-Dodd A and Thorne PS (2022). “Validation of blood arsenic and manganese assessment from archived clotted erythrocyte fraction in an urban cohort of mother-child dyads.” Science of the Total Environment 810. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Haque E, Thorne PS, Nghiem AA, Yip CS and Bostick BC (2021b). “Lead (Pb) concentrations and speciation in residential soils from an urban community impacted by multiple legacy sources.” Journal of Hazardous Materials 416: 125886. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hombrecher K, Quass U, Leisner J and Wichert M (2021). “Significant release of unintentionally produced non-Aroclor polychlorinated biphenyl (PCB) congeners PCB 47, PCB 51 and PCB 68 from a silicone rubber production site in North Rhine-Westphalia, Germany.” Chemosphere 285: 131449. [DOI] [PubMed] [Google Scholar]
  11. Jahnke JC and Hornbuckle KC (2019). “PCB Emissions from Paint Colorants.” Environmental Science & Technology 53(9): 5187–5194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Jones KC and de Voogt P (1999). “Persistent organic pollutants (POPs): state of the science.” Environmental Pollution 100(1–3): 209–221. [DOI] [PubMed] [Google Scholar]
  13. Koh WX, Hornbuckle KC and Thorne PS (2015). “Human Serum from Urban and Rural Adolescents and Their Mothers Shows Exposure to Polychlorinated Biphenyls Not Found in Commercial Mixtures.” Environmental Science & Technology 49(13): 8105–8112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Li G, Sun GX, Ren Y, Luo XS and Zhu YG (2018). “Urban soil and human health: a review.” European Journal of Soil Science 69(1): 196–215. [Google Scholar]
  15. Marek RF, Thome PS, Herkert NJ, Awad AM and Hornbuckle KC (2017). “Airborne PCBs and OH-PCBs Inside and Outside Urban and Rural US Schools.” Environmental Science & Technology 51(14): 7853–7860. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Martinez A, (2022). valdiman/R-codes-for-East-Chicago-soil-PCB-analysis: (v1.0.0). Zenodo 10.5281/zenodo.6984449. [DOI] [Google Scholar]
  17. Martinez A, Awad AM, Herkert NJ and Hornbuckle KC (2018). “PCB congener data of gas-phase, freely-dissolved water, air-water fugacity ratios and air-water fluxes in Indiana Harbor and Ship Canal, IN, USA.” PANGAEA . DOI: 10.1594/PANGAEA.894908. [DOI] [Google Scholar]
  18. Martinez A, Awad AM, Herkert NJ and Hornbuckle KC (2019). “Determination of PCB fluxes from Indiana Harbor and Ship Canal using dual-deployed air and water passive samplers.” Environmental Pollution 244: 469–476. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Martinez A, Erdman NR, Rodenburg ZL, Eastling PM and Hornbuckle KC (2012). “Spatial distribution of chlordanes and PCB congeners in soil in Cedar Rapids, Iowa, USA.” Environmental Pollution 161: 222–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Martinez A, Hua J, Haque E, Thome PS and Hornbuckle KC (2022). “Dataset of concentrations of individual polychlorinated biphenyl congeners and total organic carbon in soils from East Chicago, Indiana, USA in 2017/2018.” PANGAEA . DOI: 10.1594/PANGAEA.941894. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Martinez A, Norström K, Wang K and Hornbuckle KC (2010). “Polychlorinated biphenyls in the surficial sediment of Indiana Harbor and Ship Canal, Lake Michigan.” Environment International 36(8): 849–854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Martinez A, Schnoebelen DJ and Hornbuckle KC (2016). “Polychlorinated biphenyl congeners in sediment cores from the Upper Mississippi River.” Chemosphere 144: 1943–1949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Martinez A, Spak SN, Petrich NT, Hu D, Carmichael GR and Hornbuckle KC (2015). “Atmospheric dispersion of PCB from a contaminated Lake Michigan harbor.” Atmospheric Environment 122: 791–798. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Martinez A, Wang K and Hornbuckle KC (2010). “Fate of PCB Congeners in an Industrial Harbor of Lake Michigan.” Environmental Science & Technology 44(8): 2803–2808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Meijer SN, Ockenden WA, Sweetman A, Breivik K, Grimalt JO and Jones KC (2003). “Global distribution and budget of PCBs and HCB in background surface soils: Implications or sources and environmental processes.” Environmental Science & Technology 37(4): 667–672. [DOI] [PubMed] [Google Scholar]
  26. Meijer SN, Steinnes E, Ockenden WA and Jones KC (2002). “Influence of environmental variables on the spatial distribution of PCBs in Norwegian and UK soils: Implications for global cycling.” Environmental Science & Technology 36(10): 2146–2153. [DOI] [PubMed] [Google Scholar]
  27. Rushneck DR, Beliveau A, Fowler B, Hamilton C, Hoover D, Kaye K, Berg M, Smith T, Telliard WA, Roman H, Ruder E and Ryan L (2004). “Concentrations of dioxin-like PCB congeners in unweathered Aroclors by HRGC/HRMS using EPA method 1668A.” Chemosphere 54(1): 79–87. [DOI] [PubMed] [Google Scholar]
  28. Tang X, Shen C, Shi D, Cheema SA, Khan MI, Zhang C and Chen Y (2010). “Heavy metal and persistent organic compound contamination in soil from Wenling: An emerging e-waste recycling city in Taizhou area, China.” Journal of Hazardous Materials 173(1): 653–660. [DOI] [PubMed] [Google Scholar]
  29. USEPA (2002). Denver Front Range Study of Dioxins in Surface Soil. Summary Report. Region 8, Denver, CO.
  30. USEPA (2007). Pilot Survey of Levels of Polychlorinated Dibenzo-P-Dioxins (PCDDs), Polychlorinated Dibenzofurans (PCDFs), Polychlorinated Biphenyls (PCB) and Mercury in Rural Soils of the U.S. Washington, DC.
  31. USEPA. (2021). “Regional Screening Levels (RSLs) - Generic Tables. Tables as of: May 2021. https://www.epa.gov/risk/regional-screening-levels-rsls-generic-tables. https://semspub.epa.gov/work/HQ/400754.pdf.”
  32. Van den Berg M, Birnbaum LS, Denison M, De Vito M, Farland W, Feeley M, Fiedler H, Hakansson H, Hanberg A, Haws L, Rose M, Safe S, Schrenk D, Tohyama C, Tritscher A, Tuomisto J, Tysklind M, Walker N and Peterson RE (2006). “The 2005 World Health Organization reevaluation of human and mammalian toxic equivalency factors for dioxins and dioxin-like compounds.” Toxicological Sciences 93(2): 223–241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Vane CH, Kim AW, Beriro DJ, Cave MR, Knights K, Moss-Hayes V and Nathanail PC (2014). “Polycyclic aromatic hydrocarbons (PAH) and polychlorinated biphenyls (PCB) in urban soils of Greater London, UK.” Applied Geochemistry 51: 303–314. [Google Scholar]
  34. Vorhees DJ, Cullen AC and Altshul LM (1999). “Polychlorinated Biphenyls in House Dust and Yard Soil near a Superfund Site.” Environmental Science & Technology 33(13): 2151–2156. [Google Scholar]
  35. Yu HY, Liu YF, Shu XQ, Ma LM and Pan YW (2020). “Assessment of the spatial distribution of organochlorine pesticides (OCPs) and polychlorinated biphenyls (PCBs) in urban soil of China.” Chemosphere 243. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix A. Supplementary data.
NIHMS1859447-supplement-Appendix_A__Supplementary_data_.docx (1.5MB, docx)

RESOURCES