Subsurface Detection of Organic Compounds by Thermal Extraction Cone Penetrometry

Subsurface Detection of Organic Compounds

by Thermal Extraction Cone Penetrometry (TECP)

and Ultrafast Gas Chromatography/Mass Spectrometry (GC/MS)

The most often used tools to rapidly map the subsurface and to collect soil and water samples at depth are the Geoprobe or Cone Penetrometer (CP). CP has evolved into an effective means for detecting organics in situ by laser induced fluorescence (LIF), Raman, and IR spectroscopy. Increasingly, and direct sampling mass spectrometry has been used in conjunction with CP to speciate and provide quantitative analysis of organics. Cone penetrometry has also been used to detect metals by laser induced breakdown, gamma and x-ray fluorescence spectroscopy.

Typically, 3 ft pipe sections are threaded together and pushed underground by truck weights of up to 40 tons. Our challenge was three-fold: 1) to design a flexibly heated, 300 ⁰C, transfer line that could be woven through each pipe section safely, 2) to develop a programmable thermal extraction sampling probe capable of heating soil to about 350 ⁰C, and 3) to analyze organics in the field quick enough to support dynamic hazardous waste site investigations and cleanups. These target temperatures were based on past studies we made aimed at developing a direct measuring thermal desorption (TD) gas chromatography sample inlet system. Results of the thermal extraction cone penetrometer (TECP) GC/MS system we developed for subsurface sampling of soil-bound organics from depths of up to 25 m are described. Construction details for the TECP heated transfer line and probe can be found in references provided below.

The dependence of the TECP transfer line on length (3-30 m) and diameter (0.75-2.16 mm) as well as the carrier gas linear velocity (0.4-5 m/sec) and collection volume was determined. These experiments were carried out by directly desorbing known concentrations of organics from soil using the TD GC inlet connected to the transfer line, with organics freeze-trapped in the TECP sample collector. The TD/transfer line simulates closed cell operating conditions, namely, no loss of organics into the environment. The data showed that the TECP transfer line and organics collection efficiency were independent of length and diameter but dependent on flow rate and collection gas volume, see Figure 1. Because of its wide range in volatility, a mixture of 50 chlorinated biphenyls containing Cl-2 to Cl-7, Aroclor 1248, was used to test the system. PCB extraction efficiency from dry soil as a function of the modified Reynolds number is shown in the figure:

Re_m = Re * V₀= w² d r t/m L (3)

where, V₀ is the ratio of the carrier gas collection volume to system dead volume, w is the linear velocity of the carrier gas, r is the carrier gas density and m the viscosity, L is the length of the transfer line and d the diameter, and t is the time it takes for the vapor to be transported to the sample collector. Maximum recovery was obtained when Re_m was > 5700 and the collection volume was thirty times the dead volume.

To further test the system, closed cell (TD) extraction efficiencies were determined for di- and trinitrotoluene, 16 polycyclic aromatic hydrocarbons (PAH), and 17 chlorinated pesticides (OCP) at Re_m = 6000. Although the recovery for each compound was different, see Table 1, the shape of the modified Reynolds curve was the same as that found for PCBs. These experiments were carried out under dry soil conditions, with the organics collected during a 5 min heating period. Backflushing the transfer line and probe for a few minutes reduced any adsorbed organics to non measurable amounts.

The TECP in situ collection process was simulated by fortifying dry soil around the collection probe. Thermocouples were placed 1-3 mm out from the center of the probe. When the probe temperature was 450 ⁰C, soil temperatures reached 350 ⁰C 1 mm from the window and 265 ⁰C at 2 mm. Table 1 also lists the TECP extraction efficiency as well as the percent difference between in situ and closed cell results. The closed cell data indicate the maximum amount of organics that potentially can be thermally extracted from soil by the TECP in an open environment, see adjacent figure. Impressive is the fact that TECP recoveries for all compounds were within 20% of the ideal, direct soil thermal desorption data.

As soil moisture increases, PCB recoveries decrease for both the TD and TECP systems, see Table 2. When the soil-water content was 20% and a 5 min heating cycle was applied, both the TD and TECP extraction efficiencies dropped from 95% (dry soil) to 45% and 35%, respectively. The temperature-time profile is shown in Figure 2 for this sample at distances of 1 mm to 3 mm from the TECP probe collection window. After 5 min of heating, the soil within 1 mm of the window was 180 ⁰C. No material change in temperature was observed from 100 ⁰C at distances beyond 1 mm. Table 3 lists the recoveries under these conditions.

If, on the other hand, the TECP is backflushed during the 5 min heating cycle after which the valve is switched back to the sample collection position, extraction efficiencies increase and are within statistical agreement of the 5 min dry soil results. The exception are the thermally unstable pesticides and the more volatile PAH. After 10-15 min of heating, soil 1 mm from the window reaches 300 ⁰C and as high as 200 ⁰C at 2 mm. Presumably the volatile PAH are lost to the surroundings as vapor is swept from the TECP backflush period. In contrast, 10% soil moisture experiments showed no difference in the measured recoveries between wet and dry soils, with soil temperatures maximizing at 350 ⁰C after 12 min. Results indicate that analyte concentrations should be calculated on a dry weight basis for soils containing more than 20% water, as is the normal practice. Note that maximum SVOC recoveries cannot be achieved unless soil temperatures reach 300 ⁰C, see PAHs in Table 4. On the other hand, measurement precision was excellent.

Field experiments were performed to evaluate TECP ruggedness under typical site conditions. The TECP was pushed 10 m into the ground. The TECP passed both the mechanical and safety standard tests. The TECP - TDGC/MS was tested by analyzing soil collected from a manufactured gas plant (MGP), which burned coal to produce electrical energy. The site contained extremely high levels of PAH, with concentrations ranging between 10 ppm and 2,500 ppm. Table 5 shows the results of a soil sample that was solvent extracted, thermally desorbed directly into the GC/MS, and TECP extracted into the GC/MS. All extracts were analyzed with no sample cleanup. Table 6 presents the results for a much lower level coal tar contaminated soil sample. Results are in remarkable agreement. They were produced in < 15 min per sample, certainly quick enough to support the on-site decision making process as opposed to the typical EPA SW 846 method 8270 GC/MS, which can take as long as 35 min/sample. Although only PAH were found in the MGP samples, the same technique can be used to quantify PAH, PCB, pesticides and explosives in the same mixture simultaneously. The total ion current (TIC) and reconstructed ion current (RIC) chromatograms are shown in Figure 3. Although TIC signals were 10⁶, individual PAH signals were between 10² and 10⁶for the highly contaminated sample.

The IFD software detected low concentration PAHs in the presence of high concentration interferents present in the coal tar extract. By "seeing through" the matrix, the algorithms increase confidence in surrogate, internal standard, and target compound identification and quantitation without the need for extensive sample cleanup, see Ultrafast GC/MS.

Recently, the TECP has been mated to a field GC/MS to provide on-line, "real-time measurements." Although additional field testing of this system is needed, preliminary results are remarkable. Figure 4 shows 8 repetitive measurements of a Joliet Army Ammunition Plant soil contaminated with TNT and DNT that was also spiked with other organics. Note that the TECP collection and GC/MS analysis cycle time was ~ 45 sec. Table 7 compares the results of a Joliet sample that was solvent, TD, and TECP extracted, with all extracts analyzed by GC/MS. Figure 5 reveals that compound speciation is possible even under these extremely fast analysis times. For example, 50 VOCs were detected in situ by sniffing vapors for 10 sec, trapping, then thermally desorbing into the MS, Figures 6 and 7.

We proposed that the real-time TECP - GC/MS system be used to make depth profile contamination maps by simultaneously detecting VOCs and SVOCs, where the geology and soil moisture content is amenable. We have shown that both contaminant classes can be extracted from soil quickly, transferred to the instrument quickly, and analyzed quickly. Should there be interest in collaborating on a test of this tool, please contact arobbat@tufts.edu