Screening Level Assessment of Arsenic and Lead Concentrations in Glass Beads Used in Pavement Markings

Results and Discussion

QA/QC

Table 6 reports the MDL in terms of the mass of metal per mass of glass bead (ppm) for the analysis of total, extractable, and bioaccessible metals in glass beads. Because the MDLs were lower than the lowest calibration standard (1 µg/L), the lowest calibration standard became the PQL. Analytes with a concentration between the PQL and the MDL are reported as below the quantitation limit (BQL). Analytes detected in the sample that were below the MDL but still had a measured value are reported as below the detection limit (BDL). Analytes with a no observable measured response are reported as non-detectable. Interferences were not observed for arsenic and lead within the experimentally derived samples.

For the total metal extraction, 0.25 g of glass beads was used in the KOH fusion method, and the final samples of extract were made up in 1 L of solution. For the extractable and bioaccessible extractions, 1 g of glass beads was used, and the final extract volume was 100 mL. Therefore, the MDLs for arsenic and lead in the glass beads for the extractable metal and bioaccessible metal extractions are different than the total metal extractions.

Table 6. MDL and PQL for arsenic and lead for total, extractable and bioaccessible metals.

ppm = parts per million

MDL = method detection limit

PQL = practical quantitation limit

Total Arsenic and Lead

The total arsenic and lead contents in the glass beads measured using the KOH fusion method are presented in table 7 and figure 9. Arsenic content in all the glass beads examined was less than 100 ppm, while lead content was less than 200 ppm. NIST reports a nominal arsenic content of 50 ppm and a certified lead content of 38.57 ± 0.2 ppm in the SRM 612 wafers. The arsenic and lead content obtained by the KOH fusion method for the SRM are shown in table 8. The results for total metal analysis in glass beads show large standard deviations for both arsenic and lead, indicating a high degree of variability within the replicates of each bead sample. However, less than 6-percent variability was observed between three SRM replicates for both arsenic and lead over six analyses of the material.

Because all QA/QC checks were met with the instrument, and acceptable results were obtained for the SRM, variability associated with instrument and methodology was ruled out. The variability in glass beads could be associated with varying sources of glass and varying amounts of heavy metals in the recycled glass and glass cullet used for manufacturing glass beads. The inconsistency of the reclaimed product used to make the glass beads could result in a very high concentration of heavy metals in some glass beads. This variability in different samples and subsamples of glass beads used in pavement markings has also been observed in a previous study by NJIT/RU and TTI.^{(5, 6)} In the future, extraction methods may need to be modified to process a larger subsample to reduce the chance of selecting a high metal content glass bead randomly.

Table 7. Total arsenic and lead content (ppm) in glass beads provided by State transportation department participants.

ppm = parts per million

Figure 9. Graph. Total mean arsenic and lead (ppm) content in the glass beads supplied by the State transportation department participants.

Table 8. Total arsenic and lead content (ppm) in SRM.

ppm = parts per million

^aNominal arsenic concentration in glass matrix

^bCertified lead concentration in glass matrix

Extractable Arsenic and Lead

The extractable arsenic and lead in the glass beads measured using EPA method 3050B is presented in table 9. The concentrations of arsenic were below the MDL (0.07 ppm) for all the glass beads. For lead, one sample was measured between the MDL (0.004 ppm) and PQL (0.1 ppm) and was reported as BQL. Using the PQL as a lower limit of calibration, lead was observed within the extractable metals extracts in 7 of the 15 samples at reportable concentrations. When observed, the levels of lead ranged from 0.21 ± 0.002 to 3.29 ± 1.00 µg extractable lead per gram of bead. Therefore, when present, lead within the extractable metals extracts was up to 2.5 percent of the total observed lead in the beads.

Table 9. Extractable and bioaccessible arsenic and lead content (ppm) in glass beads provided by State transportation participants.

ppm = parts per million

BQL = below quantification limits (<0.1 µg/g for arsenic and lead)

BDL = below detection limits (< 0.07 µg/g for arsenic, < 0.004 µg/g for lead)

Bioaccessible Arsenic and Lead

The bioaccessible arsenic and lead content in the glass beads measured is also presented in
table 9. Bioaccessible arsenic concentrations were not in the reportable range for all the glass beads because the observed value was below the MDL (0.07 ppm). For lead, several of the measured values fell between the MDL (0.004 ppm) and PQL (0.1 ppm). Using the PQL as a lower limit, lead in the bioaccessible extracts was less than 0.7 percent of the lowest observed total lead concentration.

Comparison of Total Metal Content in Glass Beads Evaluated from Different Methods

The intra-method comparison for the analysis of total arsenic and lead in the glass beads included the following methods: 1) KOH fusion digestion followed by ICP-MS analysis, 2) EPA Method 3052 (microwave assisted hydrofluoric acid digestion) followed by ICP or AAS analysis to evaluate metals in siliceous solids, 3) bench-top XRF, and 4) FP-XRF. NIST SRM 612 was also analyzed for total arsenic and lead using KOH fusion and EPA Method 3052.^{(2, 3)} KOH fusion digestions and ICP-MS analysis were completed at TAMU, EPA Method 3052 and ICP or AAS analysis was completed at the EPA’s NRMRL, bench-top analysis XRF was completed on a Panalytical system at FHWA’s TFHRC, and portable XRF analysis was completed within FDOT’s laboratories.

Table 10 and table 11 provide a summary of the mean total arsenic and lead content (respectively) for each of the analyzed samples using the four methods. Results of the intra-method comparison were not in agreement for the samples, although agreement was achieved between the KOH fusion method and EPA Method 3052 for the NIST SRM 612 standard. One possible explanation for the observed difference was a difference in glass bead preparation specified within the methods. KOH fusion digestion required that all glass beads and SRM samples be crushed and sieved prior to analysis. The EPA Method 3052 digestions included whole beads that were not crushed or sieved; however, because the SRM was received as a disk, crushing the SRM was required before digestion. The portable XRF samples were not altered, but the bench-top XRF samples were fused prior to analysis. Because the methods did not agree, but did meet their respective QA/QC specifications, additional work was carried out to determine the cause of the difference between KOH fusion (an alkaline digestion) and EPA Method 3052 (an acidic digestion). However, crushing the beads did not result in a substantial change in measured arsenic or lead content when digested according to EPA Method 3052.

Sieving the crushed beads was ruled out as a source of error (due to potential metal contamination from the sieve or due to size selectivity of crushed glass beads). However, processing the SRM samples by crushing and sieving them prior to KOH fusion digestion did not indicate an issue because the SRM recovery was acceptable. Further analysis of beads and SRM to evaluate the effects of crushing and sieving did not result in any explanation of the difference in measured concentrations between the two methods when the methods were followed as described. (However, running KOH fusion on whole beads greatly reduced both arsenic and lead content measured in the samples.)

Table 10. Comparison of arsenic content (ppm) in glass beads from intra-method evaluation.

ppm = parts per million

XRF = X-ray fluorescence

EPA = Environmental Protection Agency

KOH = potassium hydroxide

ND = not detected

— indicates not analyzed

Table 11. Comparison of lead content (ppm) in glass beads from intra-method evaluation.

ppm = parts per million

XRF = X-ray fluorescence

EPA = Environmental Protection Agency

KOH = potassium hydroxide

ND = not detected

— indicates not analyzed

Instrumental errors were also ruled out by analyzing extracts on more than one analytical platform. The cross-over analysis between platforms (which was also conducted in different laboratories) reproduced the extract concentrations for the digestions. Therefore, after ruling out sources of method and instrumental error, the observed difference is considered to be caused by either intra-replicate variability of arsenic and lead in the glass beads or a matrix interference/ enhancement present in the glass beads but absent in the SRM.

Despite using a systematic approach to evaluate the cause of the differences in total metals content measured using the four methods, no method or instrument errors arose that explained the observed difference. Overall, KOH fusion provided the result closest to the nominal arsenic and certified lead content of the SRM for glass. In addition, several literature reports indicated better digestion of metals from glass through alkali fusion methods compared with acid digestion.^{([12], [13])} However, the metals considered in these studies were trace metals such as rhenium, zirconium, hafnium, thallium, and uranium. Finally, because KOH fusion gave the overall most conservative (highest) estimate of arsenic and lead content observed in the beads while meeting QA/QC limits, the authors felt most comfortable performing subsequent digestions using this method.

Relationship Between Total Arsenic Content in Glass Beads and the Retroreflective Performance of the Beads

The relationship between arsenic and retroreflective performance was evaluated because of the historical use of arsenic within glass production as a high temperature oxidant to remove imperfections in glass. The researchers were curious to evaluate whether higher arsenic levels would also correlate to higher retroreflectivity measurements, which would have implications for the performance of the beads placed on the roadway surface. The retroreflective performance of the bead samples was determined by creating pavement-marking samples on metal slabs and measuring the resulting marking retroreflectivity using a retroreflectometer. The Pearson’s product moment correlation coefficient was determined and used to assess the direction and the strength of the correlation between the arsenic content of the beads and their retroreflective performance.

The total arsenic levels found through KOH fusion digestion followed by ICP-MS analysis were used to assess the relationship between arsenic content and the measured retroreflectivity of the beads applied within a pavement marking. Table 12 and figure 10 show the relationship between retroreflectivity and total mean arsenic content within the glass beads for all 15 samples.

Table 12. Mean ± standard deviation retroreflectivity and total arsenic content for each bead sample evaluated in this research.

ppm = parts per million

Figure 10. Graph. Relationship between mean arsenic content and mean retroreflectivity of each sample of glass beads evaluated within this research.

The calculated Pearson’s product moment correlation coefficient for all 15 samples shown in figure 10 is 0.564. The coefficient indicates a positive moderate correlation between the two factors, suggesting that a correlation may exist between the arsenic content and retroreflective performance. However, additional samples should be analyzed to determine whether a correlation actually exists. In this study, the range of observed arsenic contents was limited to between 15 and 88 ppm. Arsenic contents of previously analyzed beads were an order of magnitude greater than the beads analyzed from State transportation department samples, and these high arsenic beads and beads with lower arsenic content (below 15 ppm) should be included in additional studies that explore this correlation. The retroreflectivity data, however, also demonstrate that suitable retroreflectivity performance can be achieved at low levels of arsenic.

Speciation of Metals in Leachate From Glass Beads

Arsenic speciation in water plays an important role in determining its potential toxicity to exposed organisms. Because glass beads were determined to leach arsenic, understanding the speciation of arsenic within the leachate solutions became of interest. The speciation analysis of arsenic revealed the presence of both arsenite (As³⁺) and arsenate (As⁵⁺) as the two predominant species of arsenic in solution. Table 13 presents the results for the analysis. While total arsenic levels (As³⁺ + As⁵⁺) in the pure leachate exceeded the 10 µg/L drinking water maximum containment level for arsenic, under environmental conditions, the impact of leachate on existing groundwater or surface water reserves would be minimal for these samples owing to dilution. However, speciation data are considered during the human health risk assessment.

Table 13. Arsenic speciation observed in samples of bead leachate generated from an up‑flow cartridge system.

ND = not detected

Characterization of Arsenic and Lead Content in Mixed Glass Bead and Soil Samples Taken From a Bead Storage and Transfer Facility

Five soil samples were collected from a glass bead storage and transfer facility of a pavement-marking company to study the contribution of glass beads to the total metal content within bead-impacted media samples. The bead storage facility has been storing beads on site for more than 20 years. Because beads were evident upon visual examination, the site samples served as a real-world exposure scenario for bead-impacted environmental media. The MDL for the field sample analysis was determined separately because a new ICP-MS was installed in the later stages of this study, and a new matrix was used. The field site MDLs are given in table 14. Because the MDLs were lower than the lowest calibration standard (1 µg/L), the lowest calibration standard became the PQL.

Table 14. MDL and PQL for arsenic and lead for total, extractable, and bioaccessible metals for storage yard soil samples.

ppm = parts per million

MDL = method detection limit

PQL = practical quantitation limit

Table 15 lists the average glass bead content in site soil samples collected from a bead storage and transfer facility. The difference in the content of glass beads is owing to different sampling locations in the vicinity of the facility, including the storage zone and the transfer zone. The glass beads’ content varied from 19 percent to a maximum of 78 percent. Because most of the workers at the glass beads manufacturing facility do not wear protective equipment other than a hard hat, they are likely exposed to a high volume of glass beads through various routes of exposure, including direct contact and inhalation.

Table 15 presents the total metal content of site soil samples containing glass beads. While total arsenic was only reportable in sample locations 1 and 3, the concentrations of lead in the storage facility soil samples ranged between 11 and 122 ppm. The elevated arsenic and lead content in the soil samples over the control may be associated with the presence of glass beads. However, the glass bead content in field site soil sample does not correlate with the metal content in site soil samples. A detailed study, with more samples from a variety of facilities, is needed to assess the contribution of glass beads to the total metal content of the soil.

Table 15 presents the total arsenic content in the respirable portion (< 10 µm in size). The concentrations of total arsenic were below the MDL (0.70 µg/L in aqueous phase) for all samples and are reported as BDL. These preliminary findings indicate that studies involving more samples from multiple facilities are required to evaluate the contribution of glass beads to total arsenic content in the respirable fraction of the soil.

Table 15. Glass bead content (by weight), total arsenic and lead (ppm) in site soil samples, and total arsenic (ppm) in respirable fraction of site soil samples.

¹Only one reportable data point out of three replicates

ppm = parts per million

SRM = standard reference material

— indicates not applicable
BDL = below detection limit (< 2.8 µg/g for arsenic, < 0.44 µg/g for lead)