Method for Analysis of Environmental Lead Contamination in Soils

Abstract

A method for lead (Pb) detection in soil is presented. Pb is a dangerous environmental pollutant that is present in soils, posing a health risk to millions of people worldwide, and regular monitoring of Pb contamination in soils is essential to public health. Many sensitive methods for detection of heavy metals in solid matrices exist, but they cannot be performed on-site because they are costly (>$30/sample), require trained personnel, and many classical sample preparation methods are not safe to bring into the field. We describe an alternative process, combining a safer sample preparation method with electrochemical analysis. The process requires minimal training, making it an attractive overall method for regular environmental screening of Pb in soils. Extract obtained from the soil is pH adjusted and analyzed using a stencil-printed carbon electrode and square wave anodic stripping voltammetry. In this work, a study of 15 neighborhood soils examining the concentration of Pb present post-extraction was performed to demonstrate the method. The limit of detection for the electrochemical analysis was calculated to be 16 ppb—well below the United States Environmental Protection Agency’s action limit for Pb in soils (400 mg/kg or 4,000 ppb)—and third party inductively coupled plasma-optical emission spectroscopy analysis validated the results obtained in this study to within ± 17% on average.

1. INTRODUCTION

It is commonly known that lead (Pb) is a dangerous and toxic pollutant, causing both ecological damage and impacting human health worldwide.¹ Exposure to Pb is associated with multiple health problems (anemia, brain and kidney damage, and death in severe cases), and the World Health Organization has reported that over one million deaths globally in 2017 could be attributed to Pb exposure.^2,3 Ingestion of soil and/or dust are primary pathways of exposure to Pb, and urban soils can have high concentrations of multiple toxic metals including arsenic (As), cadmium (Cd), chromium (Cr), mercury (Hg), lead (Pb), etc.⁴

Typical sources of high levels of Pb pollution in soil are anthropogenic: deposition and runoff downwind/downstream of smelting operations; exhaust emissions along highways; paint in residential areas; and land application of treated sewage sludge on farmlands.^4–6 Pb in paint is an especially ubiquitous problem for homes across the United States (US)—the Centers for Disease Control estimates that 24 million housing units have significant lead-based paint hazards.⁷ The Pb can mobilize via acid rain, aerosolization, etc. and eventually end up in the soil surrounding these homes, posing a health-risk to the people living inside.⁸

Due to the health threats associated with Pb, the US Environmental Protection Agency (EPA) has set an action limit for soil Pb contamination at 400 mg/kg (4,000 ppb).^9,10 Consistent monitoring of soils to ascertain pollutant concentration poses many challenges though due to the large areas (km²), long times required in sampling and preparation, preservation of samples from preparation to analysis, and the expenses and timeline delays associated with laboratory analysis of samples.¹¹ Not only are these barriers to analysis for large scale testing agencies, but this also eliminates the opportunity for citizen scientists to be able to test their environment, which is still a cause of concern in many affected areas.^12,13 Greater access to low-cost, easy-to-use, and in-field screening tools is highly desirable to overcome testing barriers and move towards environmental equity.

Many sensitive analysis methods for environmental Pb are available (e.g., inductively coupled plasma-optical emission spectroscopy/mass spectrometry (ICP-OES/MS), atomic absorption spectroscopy (AAS)); nonetheless, these methods are typically costly (>$30 per sample) and cannot be performed in situ.^14,15 Yet another barrier to the use of laboratory methods for environmental analysis is the requirement for transformation of the solid matrix into solution. Several types of acid digestions that use mineral acids (e.g., nitric (HNO₃), hydrochloric (HCl)) and suitable digestion equipment exist for this purpose.^16–19 Unfortunately, these procedures require stringent safety measures, and their hazardous nature also precludes bringing them into the field.

Alternatively, electrochemical methods including both potentiometry and square wave anodic stripping voltammetry (SWASV) can be used for rapid, on-site determination of Pb.^20–23 Portable, miniaturized instrumentation has made SWASV a more accessible technique in the field in recent years as well.²⁴ For these reasons, along with fast analysis times (several minutes) and the need for minimal training, electrochemistry is an attractive alternative for point of use analyses.^25–27 The use of bismuth (Bi) plated in-situ on the surface of an electrode, as an alternative to mercury, has shown much promise for trace metal detection.^28–37 Screen-printed carbon electrodes (SPCEs) have also been widely adopted for heavy metals analysis. SPCEs are low-cost, relatively simple to fabricate, amenable to mass production, disposable, and require small sample volumes (<100 μL).^25,38

Pairing SPCEs with SWASV makes for an accessible fieldable technique for the monitoring of environmental Pb in soil because it is non-invasive, quick, inexpensive, and easy to complete; however, the use of SWASV also requires transformation of the solid matrix into solution prior to analysis. Less hazardous methods of sample transformation do exist as compared to the strong acid digestions.^39–43 Even so, they use chelating agents (e.g. ethylenediaminetetraacetic acid and diethylenetriamine pentaacetate) with high binding coefficients for metals. The presence of the strong chelating agents, though helpful for the extraction of heavy metals from the solid matrix, lowers the electrochemical activity of the desired analyte because it often times can no longer be plated on the surface of the electrode for detection once bound to the ligand.^44–48

Here we report a method for Pb detection in soil. Soil is first extracted using a solution of dilute strong acid with modest chelating agents compatible for use with an electrochemical method of detection. The extract is analyzed using a Bi/Nafion modified SPCE and SWASV to determine the concentration of Pb via a standard addition analysis. To demonstrate utility of this method, a study of 15 neighborhood soils was conducted to examine the concentration of Pb suspected to be present from paint mobilization.

2. MATERIALS AND METHODS

2.1. Chemicals and Materials

All solutions were prepared using pure deionized (DI) water. Ammonium fluoride (NH₄F), ammonium chloride (NH₄Cl), potassium acetate (KC₂H₃O₂), nitric acid (HNO₃), and certified reference material soil: Metals in Soil (Product #: SQC001–30G) were all purchased from Sigma-Aldrich (Missouri, USA). Glacial acetic acid was purchased from Fischer-Scientific (New Hampshire, USA), and potassium hydroxide (KOH) was purchased from Macro Fine Chemicals (Pennsylvania, USA). Extractions in the laboratory were carried out in 4 mL cryovials purchased from Simport (Beloeil, QC). pH strips from MACHEREY-NAGEL GmbH & Co. KG (Düren, Germany) were used to test the pH of the extracts. Standard solutions of 1,000 ppm concentration Pb and Bi were purchased from Inorganic Ventures (Virgina, USA) and diluted using pure DI water for desired concentrations. The SPCEs were prepared on transparency film (polyethylene terephthalate/PET) from 3M (Minnesota, USA) with carbon ink (E3178) from Ercon Incorporated (Massachusetts, USA) and glassy carbon from Alfa Aesar (Massachusetts, USA). Silver/Silver chloride (Ag/AgCl) ink (Cl-4025) was purchased from Engineered Material Systems (Delaware, USA). A solution of 5% Nafion, purchased from Alfa Aesar (Massachusetts, USA), was diluted to a concentration of 0.5% and prepared in a 1:1 solvent of pure DI water and isopropyl alcohol from Sigma-Aldrich (Missouri, USA) for modification of the working electrode. An Excalibur dehydrator (RES10) was used to dry all soil samples before extractions were performed. All electrochemistry was performed using a PalmSens4 Bi-potentiostat and PSTrace 5.8 software from PalmSens. The Soil, Water, and Plant Testing Laboratory at Colorado State University (Colorado, USA) was used for ICP-OES validation of the results.

2.2. Collection of the Soil Samples

Soil samples used in this study were collected by study participants from residents in the states of Colorado, Oregon, Massachusetts, and Florida. Participants were instructed to sample from directly next to their houses in a place where it had been painted. They were free to use their own garden tools to collect soil from the surface of the ground. If there was any mulch or ground covering present at their location, participants were asked that they pull this back and sample from the soil beneath it. A total of 23 soil samples were obtained and tested within this study.

2.3. Sample Extractions

All soils were first dried for at least 12 hours at 74°C in the dehydrator, then homogenized using a ceramic mortar and pestle. Soils were measured into 200 mg aliquots and extracted in 2 mL of the extraction cocktail: 250 mM NH₄Cl, 15 mM NH₄F, and 50 mM HNO₃ for 5 minutes after shaking to combine initially. This extraction (cocktail and process) was modified from the previously reported Mehlich 3 protocol for use with an electrochemical method of analysis.⁴¹ All extractions for this study were carried out in triplicate on each sample. Post-extraction, samples were centrifuged for 10 minutes at 10,000 rpm, the supernatant was collected, and its pH adjusted to be above pH 2 using 3 M KOH. The pH of the supernatants was measured using pH strips. The volume of KOH used for each soil varied based on the neutralization of the extraction cocktail that occurred during the extraction due to varying soil compositions.

2.4. Electrochemical Analysis

Bi/Nafion-modified SPCEs with a Ag/AgCl reference electrode were fabricated according to previously reported methods.^32,34,49 Briefly, two stencils were made by laser cutting transparency film: one with all three electrodes and another for adding the Ag/AgCl to the reference electrode. The first stencil was placed over another transparency film and aligned using metal pins and the fixturing holes. Ercon carbon ink (0.9 g) was thoroughly mixed with 1.0 g of glassy carbon. A rubber squeegee was used to spread the carbon paste onto the transparency sheet. Once completed, the sheets were separated, and then the carbon paste was dried at 60°C for 30 min. After drying, the sheet with the dried electrodes was placed back on the fixturing pins, and the stencil for the Ag/AgCl was placed on top. A rubber squeegee was used to spread the Ag/AgCl ink onto the transparency sheet containing the electrodes. This was again dried at 60°C for 30 min. Once dried, the sheet of electrodes was laminated while also allowing the working area to be exposed to sample, as well as connect the electrodes to the potentiostat. Before use, the electrodes were coated with 1.0 μL of 0.5 % Nafion solution.

A single electrode was used for each soil sample, cleaning the electrode between anodic depositions with an oxidizing potential (0.4 V). Before the first deposition, electrodes were cleaned twice at 0.4 V to remove any contaminants, first for 100 s then for 50 s. Depositions were performed at −1.0 V for 360 s, and the electrodes were cleaned twice again between as detailed. A standard addition was used for analysis of all samples, and the extract was diluted appropriately in 100 mM pH 4 acetate buffer, 2 parts per million (ppm) Bi, and additions of Pb between 0 and 200 parts per billion (ppb).

3. RESULTS AND DISCUSSION

3.1. Optimization of the SWASV Method

The analysis method this work is based on was modified from previous work to optimize it for the quantification of Pb in complex matrices.³² A more positive deposition and starting sweep potential (−1.0 V selected for this work, as compared to −1.4 V used previously) reduced the background current and produced more resolved stripping peaks [Fig S1]. Based on other published articles in this field, a smaller amplitude of 0.02 V in the square wave was chosen.^29,30,33 The square wave frequency was also increased to 44 Hz and the sampling increment decreased to 0.01 V. The higher frequency increased the size of the stripping peaks. SWASV is a frequency dependent technique, and increasing the frequency decreases the time constant, which in turn increases the amount of faradaic current that can be measured [Fig. S2].^50,51 This adjustment resulted in fewer data points being collected for each sweep, so the increment was decreased to regain lost data density. These settings were used for all further data presented.

A tetrafluoroethylene-based fluoropolymer-copolymer film (Nafion) was coated onto the working electrode of the SPCE.^52,53 This creates a semipermeable layer of negatively charged hydrophilic groups that facilitates the transfer of positively charged species and water across it.^52,53 The sulfonate groups also have been shown to aid in preconcentrating metal cations.⁵⁴ The preconcentration of metal cations was supported in this work, as the size of metal stripping peaks increased with the addition of Nafion by ~64% [Fig. S3].

The use of stirring increases mass transport rates, and therefore the effects of agitation (a stirring substitute) applied to the solution during the deposition step using a piezoelectric vibrator situated underneath the electrode was studied here to determine if this could aid in decreasing limits of detection (LOD). ^55–57 While agitation led to larger stripping peaks (55% increase), the variability between repeated measurements was larger [Fig. S4B]. We hypothesize that the variability between measurements comes from disruption of the electrode material during the deposition process when agitation is applied. Once dried during the fabrication, the electrodes become slightly brittle, and the agitation could be causing the electrode surface to be damaged in ways that it would otherwise not without agitation. For this reason, the use of the motor was removed from the method; eight of the collected soil samples measured with agitation were excluded from the results below because of the high variability [Fig. S5].

3.2. Method Development and Validation in the Lab

In total, 15 soil samples were extracted and the concentration of Pb in each extract was quantified using the optimized conditions [Figure 1]. Example voltammograms from soil sample #1 can be seen in Fig. 2A (voltammograms for all other soil samples (#2–15) can be found in the supporting information [Fig. S5–S18]). The stripping peaks for Pb and Bi are present at −0.63 V and −0.10 V, respectively, which is consistent with previous work.³² Bi is a non-toxic, and cost-effective metal that aids the Pb in plating onto the electrode by forming an amalgam on the surface of the working electrode.⁵⁸ This allows for more Pb to be plated onto the electrode, which in turn improves the LOD.

Figure 1.

Depiction of the optimized process for the method to quantify Pb in soils. Samples were collected followed by drying and homogenization. Then 200 mg of the soil was extracted in 2 mL of extraction fluid. Samples were centrifuged for 10 min at 10,000 rpm, the supernatant was collected and pH adjusted. Analysis was performed using SWASV, and a five-point standard addition analysis was used to determine the concentration of Pb.

Figure 2.

(A) Example voltammograms for soil sample #1 with the peak for Pb present at −0.63 V and the peak for Bi present at −0.10 V (B) Plot for the standard addition analysis of soil sample #1. The extraction was performed in triplicate and the calculated concentration of Pb based on this plot was found to be 9.1 (± 8.6%) ppm. In total, 15 soils were extracted using the optimized conditions for this method, and all other voltammograms and standard addition plots can be found in the supporting information.

To perform the standard addition analysis, baseline adjusted peak areas were collected and a linear least squares regression line was fit to each set of data. An example of this plot for soil sample #1 is shown in Fig. 2B. Concentrations of Pb were calculated from each regression line equation. The slopes from each soil’s regression line can be seen in Table 1. There is a slight variance (15% coefficient of variation) when the slopes are compared showing that the solution matrix is not statistically the same across the soil samples. These matrix variations support the need for a robust method, like standard addition, as opposed to using a simple calibration curve analysis.

Table 1.

Values collected for all 15 standard additions. Samples with Pb concentrations measured below the LOD are reported as ≤0.2 ppm, accounting for a dilution factor, and without an associated uncertainty. Values for the concentration of Pb obtained by third-party ICP-OES analysis are also shown.

Sample Number	Year of House	Dilution Factor	Lead Concentration (ppm)	Lead Concentration Uncertainty (%)	Lead Concentration Determined by Third Party (ppm)
1	1890	50	5.4	14.3	8.440
2	1910	10	0.18	23.5	0.460
3	1893	10	0.16	22.0	0.320
4	1923	50	8.3	3.4	10.535
5	1928	10	0.44	7.3	0.935
6	1991	10	≤0.2	n/a	0
7	1996	10	≤0.2	n/a	0
8	1987	10	≤0.2	n/a	0.010
9	1959	10	≤0.2	n/a	0
10	1978	10	≤0.2	n/a	0.045
11	1956	10	0.33	11.8	0.640
12	1892	250	11	20.0	18.670
13	1984	10	≤0.2	n/a	0
14	1900	50	0.8	33.1	1.780
15	1895	100	0.9	8.6	12.080

Table 1 shows Pb concentrations in the soil extracts calculated using this method, as well as Pb determined by third party ICP-OES analysis. The method presented here has a LOD for Pb calculated to be 16 ppb: Six of the 15 measured samples had Pb concentrations measured to be less than the LOD, so they are presented in Table 1 as ≤0.2 ppm, accounting for a dilution factor between the extract and the fluid analyzed. Uncertainties for the measured Pb concentrations are presented as percentages of the reported ppm concentrations, and it is relevant to note that four of the 15 samples had uncertainties calculated to be ≥20.0% of the associated reported concentrations. Prior to measurement, samples were dried and mixed thoroughly to mitigate challenges associated the homogeneity of solid matrices, like soil. Although efforts to homogenize the soil were made, it remains true that solid matrices are less homogenous by nature, and, specifically in the case of soil, concentrations of pollutants can vary even across the same sampling sites. It is for these reasons that having a method that can be performed quickly and easily on site is vital to characterizing the magnitude of contamination across large sampling areas made up of a highly non-homogenous matrix.

This method underestimates the Pb concentration in some cases compared to the ICP-OES analysis [Figure 3]. Though the method sometimes underestimated the Pb concentration in the extract when compared to ICP-OES analysis, the two methods were calculated to be aligned within ± 17% on average across all 15 soil samples. Interestingly, the standard addition used in our methods is not accounting for what could be complexities within the matrix, making it more difficult to measure the Pb in the fluids. For this reason, it is hypothesized that there is some sort of Pb “sink” present on the surface of the electrode. By “sink” we mean that some of the Pb is plated on the electrode and never stripped off during the cathodic sweep that is supposed to remove all metals from the electrode. The Bi film is used to increase the sensitivity for detecting heavy metals on SPCEs; however, it is plated in-situ with the Pb, and we hypothesize that a small portion of the Pb stays trapped within the Bi amalgam that is also not removed during the stripping process. To our current knowledge, an in-depth study of the in-situ Bi film used for SPCE detection of heavy metals has not been carried out and would serve as interesting research to better understand these systems in the future.

Figure 3.

Measured Pb concentration obtained using the method in this work vs. values obtained via ICP-OES analysis for validation with inset of values between 0 and 2 ppm (top left). This work shows relatively good agreement with the ICP-OES analysis, though some of the values are underestimated, possibly due to the varying thickness of the Bi-film on the surface of the electrode. Uncertainties are shown for the method present in this work. Values that are being reported as the LOD do not have associated uncertainties reported.

Since the method verification investigated soil samples from neighborhood houses, we hypothesized the age of the home would correlate with the presence of Pb. Use of Pb in paint was banned in the US in 1978; however, houses built before this year often still have paint containing this toxic element at high levels.⁷ After all the soil samples were analyzed, the Pb concentration from each soil was plotted against the year the house was built [Figure 4]. This analysis shows that the five soil samples collected from houses built after the 1978 ban had no detectable Pb extracted from the soil based on our methods. In houses built before the ban, nine out of 10 contained a measurable amount of Pb, and the amount of Pb tended to increase with the age of the house. As described, participants in this study were asked to sample soil from directly next to their houses so that this research could begin to investigate Pb that might be present within soil from an anthropogenic contamination source, such as paint. It should also be stated that levels of Pb concentration are likely higher immediately next to areas that have been painted with paint containing Pb, and (without physical removal of the paint due to damage, remodeling, etc. on these homes) it is likely that concentrations of Pb would decrease as soil was sampled farther from the houses. This data still supports the positive impact the Pb paint ban has had in reducing human exposure and further supports the need for an analysis method that is accessible, low-cost, easy to use.

Figure 4.

Plot showing the measured Pb concentration for samples measured using the method presented in this work vs. the age of the house associated with the sample. High levels of Pb in paint were banned in the United Sates in 1978, and a decrease in the amount of Pb present over the years in the samples measured in this work was noted. Uncertainties are shown, and values that are being reported as the LOD do not have associated uncertainties depicted.

It is important to note that when working with transformation of a solid matrix into solution, one must consider the extraction efficiency of the method, which has been long reported on over the years across multiple matrices.^59,60 We define extraction efficiency as the amount of analyte extracted from the solid matrix (soil in this case) into solution for analysis divided by the total amount of analyte actually present in the solid.^61,62 For example: if a sample of soil actually contained 100 ppm Pb, but the method of analysis measured only 70 ppm Pb in the soil post-extraction, then the extraction efficiency would be determined to be 70%. Extraction efficiency of soils can have multiple factors affecting it: fluid used for the extraction, composition of soil being extracted, etc. The extraction efficiency of this method was not investigated for this work, so the reported concentrations for Pb extracted from the soil are most likely an underestimate of the total concentration of Pb in the soils themselves. With further investigation, the extraction efficiency of this method could be determined. Overall, although the total concentration of Pb was not determined for the soil samples investigated herein, it can still be definitively concluded that the extraction fluid used in this investigation does in fact extract Pb—as shown in the detection of Pb post-extraction for all soil samples measured in this study—and serves as a helpful method of screening samples if needing to know the total Pb concentration is not pertinent. Finally, also of note, although this technique was not tested in the field, we believe that many of the methods presented herein could easily translate to use in-situ and further investigation through a field study would be helpful in this determination. Specifically, though the samples in this study were dried prior to preparation and analysis, we hypothesize that if analyses were to be performed in the field without a drying step, this would not cause issue for the effectiveness of the presented method, except to slightly dilute the concentration of Pb within the samples. And while certain samples measured within this study were found to have concentrations of Pb below the LOD of this electrochemical method, the LOD of the method itself (16 ppb) is still well below the action limit set by the US EPA for Pb in soil (4,000 ppb), and, therefore, slight dilutions due to differences in water content across soils could be handled effectively.

4. CONCLUSIONS

A method for Pb detection in soil was presented. Historically, heating of a mixture of concentrated strong acids is required for transforming a solid matrix, making this unsafe for use outside of a laboratory. This work showed an alternative process which could be safer for use in the field compared to current soil transformation procedures because it combines a much milder concentration of acid with carefully selected chelating agents as opposed to using heated mixtures of concentrated acids to break down the matrix. To demonstrate this method, a study of 15 neighborhood soils examining the concentration of Pb suspected to be present from paint leaching was performed. Pb was detected electrochemically in all samples analyzed, and comparison to ICP-OES analysis showed agreement to within ± 17% on average. Combining this safer and easy-to-use sample preparation for soils with electrochemical detection makes this overall method a sensitive and attractive alternative for environmental screening of Pb in soils.