Task 2c – Review of Road Core Analysis from Commercial Laboratories and UVF Site Based Tool
Scope and Objectives
Task 1 identified the potential for the use of a field-based screening approach to rapidly identify the presence of coal tar in road core samples, as an alternative to the more time consuming and expensive laboratory analysis currently accepted by regulators and receiving waste facilities. A commercially available UVF site based tool was identified as a field-based method with the potential to meet these needs.
Following discussions with the UVF site based tool supplier, SEPA and SRRB stakeholders, concerns were identified that a field-based approach may not be perceived as sufficiently robust or accepted by regulators. While contractors are aware of UV fluorescence approaches, there are indications that parts of industry are currently hesitant to use these approaches without a better understanding of the approach and regulator acceptance as well as up-front costs.
Following discussions with the commercially available UVF site based tool supplier and noting the concerns regarding regulatory acceptance of field-based analysis, SRRB/TS arranged for chemical analysis of road cores with two laboratories and the UVF site based tool supplier.
The brief for Task 2C was for Jacobs to provide technical support for the comparison of laboratory analysis and the UVF site based tool results, and provide a review of the resulting data.
Summary of Approach and Data Management Challenges
Jacobs prepared and issued a recommended scope and testing protocol to the SRRB who requested TS who actioned the testing with support of their supply chain on three schemes (Site A, B and C) where coal tar was known to be present. The intention was for duplicate cores to be collected and sent to two laboratories for the same analysis in addition to screening by the UVF site based tool supplier. The sampling and laboratory analysis was completed between November 2019 and March 2020.
Upon initial review, it was apparent that there were inconsistencies in sample numbering and descriptions between the three sites where testing was completed. Discussions with the supporting laboratories to understand the issues and uncertainties associated with the provided analysis, flagged that the detail on the chain of custody records was, at times, limited and the quality of instruction was variable. This resulted in it being difficult to ascertain how/ if/ when the sample was sub-sampled and prepared for analysis and hence the comparability of the results between laboratories. For example, one laboratory reported that they took specific targeted sub samples from a core whereas the other laboratory noted the sample was received pre-crushed in bags so targeted subsampling wasn’t appropriate, and therefore crushed the full sample to homogenise it prior to analysis. It is not entirely clear if there is a single aspect in the communication chain that resulted in the variations noted in the end approaches adopted and the results reported, but this clearly highlights some of the challenges around the correct and comparable testing of asphalts for coal tar presence and the associated variation in report conclusions between individual laboratories.
Each laboratory completed the analysis using in-house methods/ approaches which not only varied between laboratories but also between projects within the same laboratory when requested by different supply chain Contractors.
This clearly illustrates one aspect of laboratory testing which is often overlooked when reviewing results against a standard value, such as the 50 mg/kg threshold for Benzo(a)pyrene in the WM3 guidance. The result of any analysis for any sample of heterogeneous source material such as road core or soil is highly dependent on many factors which are not generally standardised across testing facilities. These include (but are not limited to): sample condition, sample preparation by the laboratory, analysis of whole sample vs. sub-sample, choice of extraction solvents and choice of analytical instruments.
The UVF site based tool supplier noted that samples were received as full cores and they then interpreted which was the sample top and then targeted three swab analyses each on specific layers that they deemed appropriate based on the specification provided. The UVF site based tool supplier also noted however that free product/ tar was notable on some of the cores and that this had smeared across the length of the core which could result in cross contamination across asphalt layers.
In this case it appears that, with the exception of the UVF site based tool supplier who understood the intention of the project, the testing facilities were not aware of the special requirements for comparative duplicate analysis. It is anticipated that this apparent critical gap in testing instruction is common in the submission of most routine road core samples for testing. This usefully illustrates the inherent variability in laboratory approaches to sample preparation and testing methods.
The variation between data sets provided by the individual laboratories and the UVF site based tool ( resulting from their interpretation of the specification, e.g. the samples to be tested and the testing requirements/ methods, and in-house interpretation approaches) means that none of the results are directly comparable between the different testing facilities. Consequently, no clear picture can be drawn for comparison between the methods used for the three data sets.
This highlights the need for an industry standard/ protocol for road tar laboratory analysis to allow consistent and comparable results to be obtained. Such a protocol would need to be developed by laboratory industry groups. Past experience suggests attaining agreement across laboratories is often challenging noting that aligning to a particular analytical method would likely be of variable commercial detriment to individual laboratories in the short term due to changes in accredited methodologies, staff training and potential changes to equipment set up. Maintaining close communication with the analyst during scheduling and clear instructions on the purpose of the testing would also be beneficial.
Review of Analytical Approaches
Coal tar is formed as a by-product of coal gas manufacturing and coke making and is a complex mixture of organic and inorganic compounds, many of which can be toxic or carcinogenic. There is no definitive coal tar composition noting it comprises a wide range of organic compounds depending upon its origin, age and purity e.g., if it has been mixed with other products such as bitumen binders within asphalt. As there is no definitive test for the identification of coal tar, laboratories have developed in-house methodologies utilising published information on characteristics of coal tar such as the relative proportions of PAH compounds, phenols and other indicators.
As discussed in the summary report for Task 1, coal tar does however contain elevated Polycyclic Aromatic Hydrocarbons (PAHs) compared to bitumen and hence WM3 advocates the use of PAHs such as benzo(a)pyrene (BaP) as marker compounds when assessing coal tar for waste classification purposes if the coal tar content cannot be quantified. BaP is also recognised to be one of the most toxic and potent PAH compounds and is therefore also used as an indicator for assessment of risk to human health and the environment.
As discussed in Section 1.3, the 50mg/kg BaP criterion set out in WM3 (and referenced by ADEPT) has been routinely used as an extrapolated marker indicator that the quantity of coal tar present in an asphalt is sufficient (0.1% or more) for the material to be classified Hazardous, which triggers further consideration of the reuse and disposal options available for the asphalt. However, as there can be other explanations for elevated BaP in asphalt (including vehicle emissions) that may not be a definitive indication of coal tar presence, hence the use of BaP alone can result in false positives. Some types of coal tar can also contain lower BaP concentrations and hence simple reliance on BaP as a marker can also mean there is a potential for false negatives.
Furthermore, whilst WM3 is explicit that asphalt containing greater than 50mg/kg BaP is Hazardous, it is only by implication that materials with less than this threshold are therefore not Hazardous. The WM3 guidance causes ambiguity by stating the polyaromatic hydrocarbons cannot be solely used to determine if AWCCT is Hazardous. As such, the interpreted approaches recommended by ADEPT and adopted by laboratories do not align with the position adopted by the UVF site based tool supplier who rely on chemical profiles and fingerprints to identify the presence of coal tar.
The commercial testing laboratories have recognised the above issues so in addition to routine analysis for BaP (and other PAHs), laboratories also use a combination of various interpretive approaches following in-house methods to provide an indication of coal tar presence. Interpretive approaches used by laboratories within this study included the use of:
- Comparative reviews of the presence and concentrations of TPH, Phenol and PAHs from Gas Chromatography Mass Spectrometry (GC-MS) and/or Gas Chromatography Flame Ionisation Detector (GC-FID) against thresholds provided in WM3 guidance.
- Double ratio plots for specific PAH compounds collected by GC-MS or GC-FID to assess the relative proportions of certain PAH compounds and indicate empirically whether the materials analysed are likely to be coal derived, petroleum derived, combustion products or plant derived.
- Interpretation of chromatograms from TPH analysis collected by GC-MS or GC-FID to identify composition using fingerprints against various hydrocarbon compounds from databases.
- Screening for presence of other biomarker compounds expected in Coal Tars from Semi-Volatile Organic Compounds (SVOC) suites such as Dibenzothiophene.
- “In-house method of comparison of various data by senior analyst”, approach unknown.
It is highlighted that none of these in-house interpretive approaches are UKAS accredited for the asphalt matrix and given each laboratory has differing approaches there is the potential for individual laboratories (and analysts) to provide different conclusions on duplicate samples regarding coal tar presence. In addition, testing methodologies can also vary between laboratories.
The UVF site based tool adopts a different approach to the derivation of base data. As noted above, laboratories typically select and crush a core sample comprising the aggregate and binder and complete a solvent extraction to provide an interpretation for the full sample and use GC-MS or GC-FID analytical methods. The UVF site based tool approach relies on a direct solvent extraction from the binder and analysis using Ultraviolet Fluorescence (UVF), using swabs taken directly from the core. Three swabs are generally taken per layer. The spectra produced by the tool are then compared to a bespoke in-house fingerprint library and matched to provide an interpretation of the products present e.g., coal tar and bitumen and their relative percentages (noting the level of uncertainty). The concentration of BaP in the swab is also estimated based on the spectrum and library. The reported relative percentage of coal tar in the binder is then extrapolated to report the coal tar concentration in the full material based on an assumed ratio of binder to aggregate.
This approach therefore primarily distinguishes between a materials Hazardous classification based on whether extrapolated concentration of coal tar in the sample is greater than 0.1%. A BaP concentration in the coal tar is also provided to demonstrate it is greater than 0.005% of the coal tar content. This is then adjusted based on an assumed ratio of binder to aggregate to provide a BaP concentration range for the full sample. This BaP range for the full sample would therefore be expected to be broadly comparable to the laboratory analysis considering method and extrapolation variations between the approaches.
Review of Reported Data for All Study Sites
An initial comparative review of the results and the laboratories’ individual interpretations of coal tar presence is presented in Tables 2, 3 and 4 for Sites A, B and C respectively. For each site, the supply chain was requested to provide two samples each from five separate road cores, which were believed to contain layers including coal tar and layers expected to be without coal tar, based on site observations and site records. As detailed above, there were some limitations with regard to sample labelling and chains of custody and sample preparation therefore it is not entirely certain what can be compared.
UVF site based tool supplier provided a separate interpretive summary report providing their interpretation of when they would assess coal tar to be present warranting characterisation as Hazardous in accordance with WM3, but this level of specialist interpretation would not be available for each site hence it has not been included in the comparisons presented in Tables 2, 3 and 4.
Key to tables 2, 3 and 4
* - Worst case interpretation from spot samples taken
NT - Not Tested
| Core Ref. | Core Description | UVF Tool | Laboratory A | Jacobs' Interpretation of Laboratory A PAH Result | Laboratory B |
|---|---|---|---|---|---|
| Core 1 | Coal Tar Sample | 1C Sample 1 | 1A Tar 1375612 | BaP >50mg/kg | 1B 934361 |
| Core 1 | No Coal Tar Sample | 1C Sample 2 | 1A No Tar 1375606 | BaP <50mg/kg | NT |
| Core 2 | Coal Tar Sample | 2C Sample 1 | 2A Tar 1375609 | BaP >50mg/kg | 2B 934362 |
| Core 2 | No Coal Tar Sample | 2C Sample 2 * | 2A No Tar 1375607 | BaP <50mg/kg | NT |
| Core 3 | Coal Tar Sample | 3C Sample 1 | 3A Tar 1375610 | BaP >50mg/kg | 3B 934363 |
| Core 3 | No Coal Tar Sample | 3C Sample 2 | 3A No Tar 1375605 | BaP <50mg/kg | NT |
| Core 4 | Coal Tar Sample | 4C Sample 1 | 4A Tar 1375611 | BaP >50mg/kg | 4B 934364 |
| Core 4 | No Coal Tar Sample | 4C Sample 2 | 4A No Tar 1375614 | BaP <50mg/kg | NT |
| Core 5 | Coal Tar Sample | 5C Sample 1 | 5A Tar 1375613 | BaP >50mg/kg | 5B 934365 |
| Core 5 | No Coal Tar Sample | 5C Sample 2 * | 5A No Tar 1375608 | BaP <50mg/kg | NT |
Note: see text for description of uncertainties associated with sample collection and analysis; this table presents one interpretation of the data received; the uncertainties mean that other interpretations of the data might be made.
| Core Ref. | Core Description | UVF Tool | Laboratory A | Jacobs' Interpretation of Laboratory A PAH Result | Laboratory B |
|---|---|---|---|---|---|
| Core 1 | Coal Tar Sample | 1C Sample 1 * | 1A Tar 1375572 | BaP <50mg/kg | 1B 934343 |
| Core 1 | No Coal Tar Sample | 1C Sample 2 | 1A No Tar 1375576 | BaP <50mg/kg | NT |
| Core 2 | Coal Tar Sample | 2C Sample 1 * | 2A Tar 1375571 | BaP <50mg/kg | 2B 934344 |
| Core 2 | No Coal Tar Sample | 2C Sample 2 | 2A No Tar 1375569 | BaP <50mg/kg | NT |
| Core 3 | Coal Tar Sample | 3C Sample 1 | 3A Tar 1375575 | BaP >50mg/kg | 3B 934345 |
| Core 3 | No Coal Tar Sample | 3C Sample 2 | 3A No Tar 1375577 | BaP <50mg/kg | NT |
| Core 4 | Coal Tar Sample | 4C Sample 1 | 4D Tar 1375574 | BaP <50mg/kg | 4B 934346 |
| Core 4 | No Coal Tar Sample | 4C Sample 2 * | 4D No Tar 1375568 | BaP <50mg/kg | NT |
| Core 5 | Coal Tar Sample | 5C Sample 1 * | 5D Tar 1375570 | BaP >50mg/kg | 5B 934347 |
| Core 5 | No Coal Tar Sample | 5C Sample 2 | 5D No Tar 1375573 | BaP <50mg/kg | NT |
| Core Ref. | UVF tool | Laboratory A | Laboratory B |
|---|---|---|---|
| Core 1 | B1 Layer 1 | S1-1 Layer 1 0-0.08m | D1 Layer 1 (920952) |
| Core 1 | B1 Layer 2 | S1-3 Layer 3 0.12-0.21m | D1 Layer 3 (920953) |
| Core 1 | B1 Layer 3 | NT | NT |
| Core 2 | B2 Layer 1 | S2-1 Layer 1 0-0.03m | D2 Layer 1 (920954) |
| Core 2 | B2 Layer 2 | NT | D2 Layer 3 (920955) |
| Core 2 | B2 Layer 3 | S2-3 Layer 3 0.09-0.16m | NT |
| Core 3 | B3 Layer 1 | NT | NT |
| Core 3 | B2 Layer 2 | S3-2 Layer 2 0.03-0.08m | D3 Layer 2 (920956) |
| Core 3 | B2 Layer 3 | S3-3 Layer 3 0.08-0.13m | D3 Layer 3 (920957) |
| Core 3 | B2 Layer 4 | NT | NT |
| Core 4 | B4 Layer 1 * | S4-1 Layer 1 0-0.10m | D4 Layer 1 (920958) |
| Core 4 | B4 Layer 2 | S4-2 Layer 2 0.10-0.16m | D4 Layer 2 (920959) |
| Core 4 | B4 Layer 3 * | S4-3 Layer 3 0.16-0.23m | D4 Layer 3 (920960) |
| Core 5 | B5 Layer 1 * | NT | NT |
| Core 5 | B5 Layer 2 * | S5-2 Layer 2 0.06-0.10m | D5 Layer 2 (920961) |
| Core 5 | B5 Layer 3 * | S5-3 Layer 3 0.10-0.16m | D5 Layer 3 (920962) |
| Core 5 | B5 Layer 4 * | NT | NT |
From Tables 2, 3 and 4, it is apparent that the interpretations provided by the individual laboratories and the UVF site based tool supplier rarely directly correlate, and laboratory interpretation often results in inconclusive results that require further interpretation to allow a judgement to be made on whether coal tar is present or not. Whilst challenges were noted in the sample handling/ delivery, the variation in results highlights the challenges that highways contractors encounter during maintenance and renewals schemes. The UVF site based tool supplier noted that free product/ tar was notable on some of the cores and that this had smeared across the length of the core which could result in cross contamination across asphalt layers.
Other concerns include the lack of transparency between inhouse interpretations most notable in the testing certificate extract from Site C (Figure 1 below) where samples with significantly elevated PAH concentrations have not been noted as containing coal tar and vice-versa. This potential for false negatives - non-identification of elevated contamination - is potentially of more concern and could pose a risk to the workforce and, in addition, present liability to waste producers through misclassification of wastes and inappropriate reuse or disposal. These aspects are discussed further in the conclusions.
Noting the above, a further review of the Site A and Site C data was not possible or considered of value. The Site B data set was slightly better in terms of correlation of sample assumed to be comparable for the laboratories and UVF site based tool. A separate comparison of Site B is therefore provided below.
Site B Data Review
Approach
A comparison including a review of the following has been completed for the Site B data set and is presented in Table 5:
- reported benzo(a)pyrene concentrations
- UVF site based tool and laboratory Interpretations
- Jacobs’ assessments of the data provided in accordance with WM3 (this is, the interpretation that might be reached by a professional assessment of the information presented in the laboratory reports, in accordance with WM3, without further context); anda summary of correlation between the laboratories and UVF site based tool.
Table 5
Comparative Review of Site B Data sets - Reported Benzo(a)pyrene Concentrations, UVF site based tool Supplier and Laboratory Interpretations, Independent WM3 Assessments and Summary of Correlation Review.
Key
- Core 1 - No evidence of coal tar in photo
- Core 2 - Possible indication of coal tar in layer 1 photo
- Core 3 - Apparent free phase coal tar in layer 2 and 3 photo
- Core 4 - Apparent free phase coal tar in layer 2 and 3 photo
- Core 5 - Apparent free phase coal tar in layer photo
- Deg.Bit.Road.Binder – Degraded Bitumen Road Binder
- V.Deg.Bit.Road.Binder – Very Degraded Bitumen Road Binder
- UVF – UltraViolet Fluorescence
- GC-FID – Gas Chromatography - Flame Ionisation Detection
- GC-MS – Gas Chromatography – Mass Spectroscopy
- CT - Coal Tar
- NT - Not Tested
- ND - Not Detected
- Haz - Hazardous Waste
- Non- Haz - Non-Hazardous Waste
Note: see text for description of uncertainties associated with sample collection and analysis; this table presents one interpretation of the data received; the uncertainties mean that other interpretations of the data might be made. The purpose of this table is to illustrate the challenges involved in data assessment and no reliance should be made based on the interpretation presented here.
| Core | UVF | Laboratory A | Laboratory B |
|---|---|---|---|
| Core 1 | B1 Layer 1 spots a, b & c | S1-1 Layer 1 0.00-0.08m | D1 Layer 1 (920952) |
| Core 1 | B1 Layer 2 spots a, b & c | NT | NT |
| Core 1 | B1 Layer 3 spots a, b & c | S1-3 Layer 3 0.12-0.21m | D1 Layer 3 (920953) |
| Core 2 | B2 Layer 1 spots a, b & c | S2-1 Layer 1 0.00-0.03m | D2 Layer 1 (920954) |
| Core 2 | B2 Layer 2 spots a, b & c | NT | NT |
| Core 2 | B2 Layer 3 spots a, b & c | S2-3 Layer 3 0.09-0.16m | D2 Layer 3 (920955) |
| Core 3 | B3 Layer 1 Spots a, b & c | NT | NT |
| Core 3 | B3 Layer 2 Spots a, b & c | S3-2 Layer 2 0.03-0.08m | D3 Layer 2 (920956) |
| Core 3 | B3 Layer 3 spots b and c | S3-3 Layer 3 0.08-0.13m | D3 Layer 3 (920957) |
| Core 3 | B3 Layer 3 spot a | S3-3 Layer 3 0.08-0.13m | D3 Layer 3 (920957) |
| Core 3 | B3 Layer 4 / NT / NT | NT | NT |
| Core 4 | B4 Layer 1 spot a | S4-1 Layer 1 0.00-0.10m | D4 Layer 1 (920958) |
| Core 4 | B4 Layer 1 spot b | S4-1 Layer 1 0.00-0.10m | D4 Layer 1 (920958) |
| Core 4 | B4 Layer 1 spot c | S4-1 Layer 1 0.00-0.10m | D4 Layer 1 (920958) |
| Core 4 | B4 Layer 2 spots a, b & c | S4-2 Layer 2 0.10-0.16m | D4 Layer 2 (920959) |
| Core 4 | B4 Layer 3 spot a | S4-3 Layer 3 0.16-0.23m | D4 Layer 3 (920960) |
| Core 4 | B4 Layer 3 spots b & c | S4-3 Layer 3 0.16-0.23m | D4 Layer 3 (920960) |
| Core 5 | B5 Layer 1 spot a | NT | NT |
| Core 5 | B5 Layer 1 spot b & c | NT | NT |
| Core 5 | B5 Layer 2 spot a and c | S5-2 Layer 2 0.06-0.10m | D5 Layer 2 (920961) |
| Core 5 | B5 Layer 2 spot b | S5-2 Layer 2 0.06-0.10m | D5 Layer 2 (920961) |
| Core 5 | B5 Layer 3 spot a & b | S5-3 Layer 3 0.10-0.16m | D5 Layer 3 (920962) |
| Core 5 | B5 Layer 3 spot c | S5-3 Layer 3 0.10-0.16m | D5 Layer 3 (920962) |
| Core 5 | B5 Layer 4 spot a & c | NT | NT |
| Core 5 | B5 Layer 4 spot b | NT | NT |
| Core | BaP Conc. (mg/kg) UVF | Interpretation (fingerprint) - UFV considers 3% coal tar to be potentially Hazardous |
|---|---|---|
| Core 1 | 0.46-3.7 | Deg.Bit.Road.Binder |
| Core 1 | 1.4-4.4 | Bit.Road.Binder |
| Core 1 | 1.0-3.6 | Bit.Road.Binder |
| Core 2 | 51.3-244 | Deg.Bit.Road.Binder (+ ~36% Coal Tar) |
| Core 2 | 1.7-27 | Deg.Bit.Road.Binder |
| Core 2 | 1.8-8.8 | Deg.Bit.Road.Binder |
| Core 3 | 3.5-37.1 | Bit.Road.Binder |
| Core 3 | 23.1-96 | Deg.Bit.Road.Binder (+ ~26% Mobile Coal Tar) |
| Core 3 | 15.8-42.8 | Deg.Bit.Road.Binder |
| Core 3 | 76.3-178 | Deg.Bit.Road.Binder (+ ~19% Mobile Coal Tar) |
| Core 3 | 13.3-33.9 | Deg.Bit.Road.Binder |
| Core 4 | 12.7 - 29.5 | Deg.Bit.Road.Binder (+ ~21% Mobile Coal Tar) |
| Core 4 | 2.7 - 6.3 | Bit.Road.Binder |
| Core 4 | 34.4 - 80.4 | Bit.Road.Binder (+ ~3.8% Mobile Coal Tar) |
| Core 4 | 118-285 | Mobile Coal Tar |
| Core 4 | 331-773 | Coal Tar |
| Core 4 | 2.1-5.5 | Deg.Bit.Road.Binder |
| Core 5 | 33.6-78.5 | Deg.Bit.Road Binder (+ ~12% Mobile Coal Tar) |
| Core 5 | 3.8-9.8 | Deg.Bit.Road.Binder |
| Core 5 | 6.2-17 | Deg.Bit.Road Binder (+ ~1.6% Mobile Coal Tar) |
| Core 5 | 176-410 | Coal Tar |
| Core 5 | 193-918 | Coal Tar |
| Core 5 | 39.2-91.5 | V.Deg.Bit.Road.Binder |
| Core 5 | 11-31.3 | Deg.Bit.Road.Binder |
| Core 5 | 24.4-56.9 | Deg.Bit.Road.Binder (+ ~8.2% Mobile Coal Tar) |
| Core | BaP Conc. (mg/kg) GC-FID | BaP Conc. (mg/kg) GC-MS | Interpretation from Chromato-grams |
|---|---|---|---|
| Core 1 | NT | <0.05 | Not fuel but PAHs |
| Core 1 | NT | NT | NT |
| Core 1 | NT | <0.05 | Not fuel but PAHs |
| Core 2 | NT | <0.05 | PAHs |
| Core 2 | NT | NT | NT |
| Core 2 | NT | 180 | Not fuel but PAHs |
| Core 3 | NT | NT | NT |
| Core 3 | NT | 10 | Not fuel but PAHs |
| Core 3 | NT | 55 | Not fuel but PAHs |
| Core 3 | NT | 55 | Not fuel but PAHs |
| Core 3 | NT | NT | NT |
| Core 4 | NT | 0.49 | PAHs |
| Core 4 | NT | 0.49 | PAHs |
| Core 4 | NT | 0.49 | PAHs |
| Core 4 | NT | 120 | PAHs |
| Core 4 | NT | 34 | Not fuel but PAHs |
| Core 4 | NT | 34 | Not fuel but PAHs |
| Core 5 | NT | NT | NT |
| Core 5 | NT | NT | NT |
| Core 5 | NT | 1.9 | Not fuel but PAHs |
| Core 5 | NT | 1.9 | Not fuel but PAHs |
| Core 5 | NT | 5.5 | Not fuel but PAHs |
| Core 5 | NT | 5.5 | Not fuel but PAHs |
| Core 5 | NT | NT | NT |
| Core 5 | NT | NT | NT |
| Core | BaP Conc. (mg/kg) GC-FID | BaP Conc. (mg/kg) GC-MS | Ratio Plot Hydrocarbon Source Interpretation | Interpretation noting Chromatograms and Biosensor |
|---|---|---|---|---|
| Core 1 | <0.1 | 4.5 | ND | No Coal Tar |
| Core 1 | NT | NT | NT | NT |
| Core 1 | <0.1 | 1.6 | ND | No Coal Tar |
| Core 2 | 88 | 36 | Combustion | Inconclusive |
| Core 2 | NT | NT | NT | NT |
| Core 2 | <0.1 | 32 | ND | No Coal Tar |
| Core 3 | NT | NT | NT | NT |
| Core 3 | <0.1 | 11 | ND | No Coal Tar |
| Core 3 | <0.1 | 12 | ND | No Coal Tar |
| Core 3 | <0.1 | 12 | ND | No Coal Tar |
| Core 3 | NT | NT | NT | NT |
| Core 4 | <0.1 | 7.1 | ND | No Coal Tar |
| Core 4 | <0.1 | 7.1 | ND | No Coal Tar |
| Core 4 | <0.1 | 7.1 | ND | No Coal Tar |
| Core 4 | 31 | 8.3 | Combustion | Inconclusive |
| Core 4 | <0.1 | 8.7 | ND | Inconclusive |
| Core 4 | <0.1 | 8.7 | ND | Inconclusive |
| Core 5 | NT | NT | NT | NT |
| Core 5 | NT | NT | NT | NT |
| Core 5 | <0.1 | 27 | ND | No Coal Tar |
| Core 5 | <0.1 | 27 | ND | No Coal Tar |
| Core 5 | 23 | 29 | Combustion | Inconclusive |
| Core 5 | 23 | 29 | Combustion | Inconclusive |
| Core 5 | NT | NT | NT | NT |
| Core 5 | NT | NT | NT | NT |
| Core | UVF Data set | Laboratory A Data Set | Laboratory B Data set |
|---|---|---|---|
| Core 1 | Bitumen - Non-Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 1 | Bitumen - Non-Haz | NT | NT |
| Core 1 | Bitumen - Non-Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 2 | CT present - Haz | Bitumen - Non-Haz | CT present - Haz |
| Core 2 | Bitumen - Non-Haz | NT | NT |
| Core 2 | Bitumen - Non-Haz | CT present - Haz | Bitumen - Non-Haz |
| Core 3 | Bitumen - Non-Haz | NT | NT |
| Core 3 | CT present - Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 3 | Bitumen - Non-Haz | CT present - Haz | Bitumen - Non-Haz |
| Core 3 | CT present - Haz | CT present - Haz | Bitumen - Non-Haz |
| Core 3 | Bitumen - Non-Haz | NT | NT |
| Core 4 | CT present - Non Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 4 | Bitumen - Non-Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 4 | CT present - Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 4 | CT present - Haz | CT present - Haz | Bitumen - Non-Haz |
| Core 4 | CT present - Haz | CT present - Non Haz | Bitumen - Non-Haz |
| Core 4 | Bitumen - Non-Haz | CT present - Non Haz | Bitumen - Non-Haz |
| Core 5 | CT present - Haz | NT | NT |
| Core 5 | Bitumen - Non-Haz | NT | NT |
| Core 5 | CT present - Non Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 5 | CT present - Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 5 | CT present - Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 5 | CT may be present - Haz | Bitumen - Non-Haz | Bitumen - Non-Haz |
| Core 5 | Bitumen - Non-Haz | NT | NT |
| Core 5 | CT present - Haz | NT | NT |
| Core | Interpretation Agreement between labs? | Interpretation Agreement between UVF and lab/labs? |
|---|---|---|
| Core 1 | Broadly. | Broadly. |
| Core 1 | N/A | N/A |
| Core 1 | Broadly. | Broadly |
| Core 2 | No - unless core inverted for Laboratory A analysis then weak. | Broadly for both if Laboratory A core inverted. |
| Core 2 | N/A | N/A |
| Core 2 | No - unless core inverted for Laboratory A analysis then weak. | Broadly for both if Laboratory A core inverted. |
| Core 3 | N/A | N/A |
| Core 3 | Yes | No - depending on sensitivity of UVF range. |
| Core 3 | No | No but potentially if UVF spots averaged. |
| Core 3 | No | No but potentially if UVF spots averaged. |
| Core 3 | N/A | N/A |
| Core 4 | Weak. | No - UVF shows high variability in layer. May align if spots averaged. |
| Core 4 | Weak. | No - UVF shows high variability in layer. May align if spots averaged. |
| Core 4 | Weak. | No - UVF shows high variability in layer. May align if spots averaged. |
| Core 4 | No - wide variation. | Yes (with Laboratory A). |
| Core 4 | Weak. | No but potentially if UVF spots averaged. |
| Core 4 | Weak. | No but potentially if UVF spots averaged. |
| Core 5 | N/A | N/A |
| Core 5 | N/A | N/A |
| Core 5 | No. | No unlikely even if UVF spots averaged. |
| Core 5 | No. | No unlikely even if UVF spots averaged. |
| Core 5 | Weak. | No unlikely even if UVF spots averaged. |
| Core 5 | Weak. | No unlikely even if UVF spots averaged. |
| Core 5 | N/A | N/A |
| Core 5 | N/A | N/A |
Discussion of UVF site based tool Results
To allow for sample variation, the supplier took three individual swabs / spot samples for each layer analysed. For simplicity, the data summary in Table 5 presents the highest result for each layer. It is notable that, in some layers, significant variations have been noted between spot samples in the same layer (e.g., Core 4) to the extent that different waste classifications would apply. The supplier suggests that this could be attributable to cross-contamination during sampling/ transportation (as noted by free phase smearing) despite efforts to clean up the samples before testing. The supplier also postulates that some coal tars are “mobile” and that this impact could occur “in-situ” as well as during coring/ transportation. This seems a reasonable explanation and ADEPT recognise the miscible nature of coal tar and the potential for cross-contamination between asphalt layers. The variation of coal tar presence within individual layers could also be attributable to gravimetric “pooling” of tar on one side of the core when stored horizontally.
These issues raise questions over the applicability of the both the UVF site based tool and laboratory analysis and the inherent risks incurred during transportation of samples if not segregated and appropriately wrapped on site. A site-based testing strategy would reduce the risk of cross-contamination during transportation (but not in-situ migration). Conversely, the ability to adequately clean any coring-induced cross-contamination using site best techniques would be reduced compared to a laboratory setting. This illustrates the inherent sampling, transportation and sample preparation issues that apply to all current testing approaches. On-site separation of layers before transportation would help reduce ex-situ cross contamination issues.
The UVF site based tool suppliers’ interpretation notes that in some instances where coal tar is detected at relatively low concentrations (e.g., 3.8% in Core 4 and 1.6% in Core 5) this could be attributable to in-situ migration within between bitumen and coal tar layers. Other explanations could include localised mixing during road surface replacement. Instances such as these are of potential overall concern to site-based attempts to separate the obviously coal tar impacted layers from overlying layers with targeted planing depths if localised migration could have occurred. If such areas are inadvertently planed, this could cross-contaminate an otherwise Non-Hazardous load.
The fingerprint library used by the UVF site based tool is understandably hidden within their software however, this means that there is limited opportunity to audit the UVF site base tool supplier’s interpretation. The spectra included in the UVF site based tool report are not overly easy to validate and in some cases variations between a bitumen interpretation and bitumen with coal tar interpretation are not easy to spot manually.
The UVF site base tool supplier was the only testing facility to provide photos of the cores prior to analysis for Site B (see above). These images clearly are extremely informative and show areas of discolouration attributable to coal tar binders. In some of the photos, a sheen of free phase / miscible coal tar is observable. The UVF Site based tool results generally correlate well with the observable coal tar areas which reiterates the importance of good quality logging, field observations and photos as supporting lines of evidence in determine the presence of coal tar.
Discussion of Laboratory Results
The review of presented laboratory data in Table 5 highlights further detail on areas of uncertainty with notable variations between recorded BaP concentrations and even variations in results, dependent upon the testing methodology used, within a single laboratory. Given the noted variations in results, there is the potential to misclassify a material depending on which analytical method is used. The GC-MS analytical technique is deemed a more accurate analysis for samples where there could be interference from other compounds. ADEPT provides guidance on this aspect.
It is notable that a number of the samples tested by the laboratories which were expected to contain coal tar based on site observations, and where the UVF site based tool supplier photos of the duplicate cores suggest coal tar is present, were not reported as having BaP concentrations that would classify the sample as Hazardous. The reasons for this are not entirely clear but could be associated with sample preparation and inherent sample variability noting the similar variability noted in individual layers by the spot samples. There is no photographic record of the samples received by the labs, or detailed description of sample preparation.
There also appear to be instances where sample labelling may be the cause of contradictory results between laboratories e.g., results suggest that Core 2 may have been inverted by Laboratory A. This reiterates the importance of sample logging, labelling and potentially sub-sampling on site.
The chromatogram interpretations provided by Laboratory A and Laboratory B are often not helpful or conclusive and it would appear, given the range of product types referenced, that the fingerprint libraries are not comprehensive with regards to coal tars. The majority of interpretations suggest PAH presence where coal tars were suspected however, this is not clear in the information provided by the laboratory and requires subsequent independent interpretation.
The Laboratory B interpretations typically report an absence of coal tar or an inconclusive outcome. This leaves the client/contractor having to either assume the materials are Hazardous or conduct additional interpretation.
In summary, the laboratory testing does not always produce the outcome that would be anticipated given the visual observations from the site photographs. It is clear that a single method of PAH analysis should be recommended/ requested to allow comparable results to be produced. The difference in interpretation methods between laboratories is visible form the difference in noted outcomes provided although the provided outcomes are often inconclusive and leave clients/ contractors with uncertainty.
Comparison of Laboratory Testing with UVF site based Tool Approach
The issues regarding sample handling, labelling and requested analysis, have somewhat undermined the intended purpose of the overall study and hence it is not possible to directly correlate the laboratory and UVF site based tool results.
Whilst there are indications of comparability in some instances, overall the UVF site based tool typically returns more positive coal tar results compared to the laboratory methods. The reasons for this are likely to be largely attributable to adoption of different interpretive methods defined in WM3. Whereas laboratories adopt the WM3 approach that assumes coal car cannot be quantified and focuses on the simplified BaP marker approach (supported by other assessments), the UVF site based tool uses spectral signatures to not only identify coal tar, but to also provide a semiquantitative concentration. The use of BaP as a marker compound for coal tar relies upon an empirical relationship between BaP and total PAH concentrations, and largely ignores the cumulative concentrations of other potentially carcinogenic coal tar components included in the UVF site based tool output. This could theoretically result in the laboratory outputs providing a comparatively higher threshold concentration for coal tar than the UVF site based tool fingerprint method. Both the UVF site based tool and laboratory methods seem accurate for determining the presence of 100% bitumen or >50% coal tar, but provide less confidence within mixed materials where cross-contamination is possible. Given the limited accuracy of road planing depths, and inherent mixing of materials during planing, there remains concern that the misclassification of blended bitumen and coal tar horizons at the pre-construction stage could lead to subsequent unintentional intermixing of Hazardous and Non-Hazardous materials. This could result in a higher proportion of planings being cross-contaminated and classified as Hazardous Waste.
The UVF site based tool outputs do however tend to correlate relatively well with the field observations which does flag the important of field observations but also the training of the analyst and understanding of the purpose of the analysis. It is arguable that UVF site based tool supplier had a clear understanding of project objectives compared to the commercial laboratories where requirements were interpreted from a schedule.
It is highlighted that, the UVF site based tool supplier has provided their own assessment of the correlation/ compatibility of the laboratory data vs the UVF site based tool outputs which has not been fully reviewed as part of this phase. The UVF site based tool UVF site based tool UVF site based tool supplier correlation assessment suggests a number of potentially erroneous aspects within the laboratory data that supports the technical quality of the UVF analysis. In addition, it concludes that the UVF site based tool does align with the laboratory analysis (where the laboratory data are deemed valid) but this assertion has not been reviewed in detail with the required level of subjectivity to comment further.
Task 2C Conclusions and Recommendations
In summary, it is unlikely that the UVF site based tool approach and current laboratory approaches are going to directly correlate given the alternative interpretive approaches followed. Further detail on this is provided in Attachment 1. Correlation could probably be increased if improvements were made to standardise sampling, handling and specification of analysis as discussed below.
Areas for Improvement to Analytical Methods
There are some valuable observations that can be drawn from the study with regards to sample collection, handling and transportation, sample preparation and laboratory instruction which would both improve routine coal tar testing and any future studies.
It is essential that good quality site records are kept to support the interpretation of the data. This includes photographs, scaled core logs, observations of free phase, odours etc, and records of sub-sample depths and correct core orientation. These records would be beneficial to the receiving laboratory or analyst to support their ability to provide a conclusive interpretation of the coal tar presence.
Both UVF site based tool analysis and the laboratories were affected by potential cross contamination of core samples, due either to coring induced impacts or migration of miscible coal tar fractions post sampling/ during transit. Appropriate sub-sampling of cores on site to allow targeted analysis and minimise cross contamination between layers will be critical if targeted planing is to be considered. Similarly, appropriate sample preservation and packaging is also required to ensure the as-received sample is as close to field conditions as possible.
The party responsible for the scheduling of the analysis needs to provide clearly defined instructions to the laboratory specifically identifying that the analysis is for the classification of coal tar. This will allow the laboratory to schedule the appropriate analysis and interpretation.
It is recommended that the testing of AWCCT should be better standardised between laboratories. Whilst this is unlikely to result in exactly the same interpretation between laboratories, it would have the benefit of producing a more standardised output that would be more readily accessible to clients/ contractors. This would need to be led by the laboratory testing community. Recommended methods for inclusion within a standardised approach include (but are not limited to):
- TPH, Phenol and PAHs from either Gas Chromatography Mass Spectrometry (GC-MS) Gas Chromatography Flame Ionisation Detector (GC-FID), Double ratio plots for specific PAH compounds collected by GC-MS or GC-FID to indicate empirically whether the materials analysed are likely to be coal derived, petroleum derived, combustion products or plant derived.
- Interpretation of chromatograms from TPH analysis collected by GC-MS or GC-FID to identify composition using fingerprints against various hydrocarbon compounds from databases which should include fingerprints for coal tars.
- Screening for presence of other biomarker compounds expected in Coal Tars from Semi-Volatile Organic Compounds (SVOC) suites such as Dibenzothiophene.
- Overall comparison of various data sources by a senior analyst to provide a conclusive output.