Remediation of contaminated soil is an environmental priority for existing utility companies. Contaminated soil was the result of the operation of coal fired power generators dating back to the mid-1800s.
This soil is excavated and transferred to a remote site, where hydrocarbons are removed and the resultant clean soil is returned for use as landfill. Due to on-site excavation, coal tars release hydrocarbons into the air. Some hydrocarbons are noxious and their concentration is regulated by the US EPA, while others are simply perceived by humans as toxic odors.
Since these emissions have a negative impact on the surrounding community, site managers have to track and minimize the emission of volatile organic compounds (VOCs) and odors into the air.
Figure 1. Excavation at remediation site involves removal of coal tar contaminated soil
Soil analysis by an independent laboratory can give an estimation of odor chemical signature, and this will reveal the name and concentration of the individual compounds contaminating the soil. The compounds’ physical properties such as boiling point or molecular weight provide an estimate of the odor chemistry.
For a site contaminated with coal tar, naphthalene is the main VOC with an odor threshold of 27 ppbv.
Figure 2. Laboratory analysis of contaminated soil provides a first order approximation of odor signature based upon molecular weight
Odor Measurement Method
zNose® is a new type of portable electronic nose that enables on-site chemical measurements of odors and VOC emissions in near real time. The zNose® separates and quantifies the odors’ hydrocarbon chemistry in less than a minute. A new solid-state detector helps achieve universal selectivity and parts per trillion sensitivity.
The US EPA’s Environmental Technology Verification (ETV) program has validated the performance of the technology. Quality control of odor measurement methods follows the same procedure used in laboratory testing. In addition, the instrument employs an optional GPS receiver, which allows linking an odor measurement to a particular time and location.
Both off-site and on-site measurements of odors and VOC emissions provide real-time data to site managers and are a useful tool for controlling and monitoring the effect of such emissions on the surrounding community.
In order to characterize only the odors from the soil, a 10 gallon bucket was taken which was half filled with soil and sealed with aluminum foil. After 5 minutes, sampling and measurement were made on headspace vapors. A side-ported GC needle was then connected to the inlet of the zNose® and inserted via the aluminum foil.
It took 2 seconds to remove 1 ml of headspace vapor, and it took less than 1 minute to measure the concentrations of the individual chemicals within the odor . While 27 different compounds were separated, the primary hydrocarbons and their concentrations were Toluene (5.7 ppm), Benzene (9.5 ppm), m,p-Xylene (12.6 ppm), methyl-naphthalene (2.5 ppm), and Naphthalene (17 ppm). Due to the high concentration of VOCs in the soil headspace vapors, background or interference odors were not a major issue.
Figure 3. Model 4100 zNose® ultra-high speed gas chromatograph
Figure 4. Soil odors were tested in foil covered buckets. A characteristic odor signature for the site was obtained by measuring the headspace vapors.
On-Site Odor Measurements
A patented solid-state detector was used to directly measure odor intensity versus elution time from a GC column, which was temperature programmed between 40 °C and 200 °C at rates as high as 20 °C per second. Sensitivity was regulated by two factors (1) the detector’s temperature and (2) the amount of the vapor sampled.
Contaminated soil in a closed environment had high concentration of chemical vapors and odors can be easily evaluated using only a relatively hot 80 °C detector and a 1-milliliter vapor sample. At such high odor concentrations, background odors from ambient air were not a major factor.
Olfactory images, referred to as VaporPrints®, are two-dimensional, high-resolution images that are fully based on the relative concentrations of the individual chemicals constituting an odor. The image is a polar plot of the retention time (volatility) and odor intensity (radial direction = sensor signal).
Complex odors can be detected by their characteristic shapes, depending on the unique chemistry of the odor. In fact, the olfactory image allows transferring olfactory response to a visual response. Computers and humans are well suited to the analysis and recognition of visual patterns. Computer processing of olfactory images also enables identification, quantification, and comparison of separate chemicals within the odor.
Figure 5. Linear odor intensity (radial direction) vs elution time from GC column (angle) with start and stop time at 12 o’clock position.
Figure 6. Logarithmic odor intensity (radial direction- 100 to 1 span) vs elution time from GC column (angle) with start and stop time at 12 o’clock position.
The solid-state detector directly measures the concentration of the odor chemicals, and peaks in the GC column flux are identified to determine the retention times for each of the chemicals detected. Column flux is measured in real time by mathematically doing the time derivative of the detector signal.
This produces a chromatogram spanning <1 minute and representing the speed of adsorption and de-sorption of vapors onto the detector. The compounds have unique retention times, which enabled their separation and identification. Tabulating the unique retention times along with the individual and total concentration counts (cts) offers a quantitative measure of the chemicals within an odor.
Defining alarm bands centered on the individual retention times of each chemical peak helped achieve automatic quantification and tabulation of individual chemicals within an odor. A narrow time range is defined for each compound to be quantified.
Moreover, defining bands and alarm levels for particular chemicals within an odor leads to a virtual array of chemical sensors that are specific to that odor. Through alarm bands, each peak is automatically detected, quantified, and compared with a user-defined alarm concentration level. Identified peaks are shown in a peak list in RED along with their concentration counts and retention time.
Figure 7. The derivative of odor intensity is a chromatogram used to determine chemical retention times. Using a <1 minute analysis of the soil odor, 27 different compounds, their individual intensity, and the total of all intensities is tabulated.
Figure 8. Peak identification table listing identified compounds in RED together with their retention time and concentration counts.
Figure 9. Top trace shows alarm bands (in RED) which are used to identify individual chemicals and compare their concentration (lower trace) to a user-defined alarm level.
The detector response is calibrated using a known vapor concentration of target chemicals (standard vapor). A standard vapor is produced by injecting a container of a known volume with a known amount of volatile chemicals. Calibration response factors can be either multi-point or single point and are connected to specific instrument sensitivity settings.
Variable sensitivity is obtained by changing the size of the vapor sample (sampling time) or the detector’s temperature. When a 1-milliliter vapor sample of naphthalene standard vapor is used, the response factor is 0.5 counts per ppbv with an 80 °C detector. Increasing the size of the sample to 15 ml gave a minimum detection level of 100 parts per trillion and a response factor of 300 cts/ppbv. Cooling the detector to 20 °C increases the response factor to 7.5 cts/ppbv.
Figure 10. Naphthalene sensitivity Vs detector temperature with a 1-milliliter vapor sample.
Figure 11. Calibrating with standard vapor concentration.
Usually, vapor standards for all chemicals at a location are not available. There are many instances where the exact chemical name of a detected compound is not known, but despite this fact the compound can still be identified by indexing its retention time to that of a known chemical. Most often, identification is made by indexing the retention time of unknown compound to that of the nalkanes and subsequently searching a library of indices for a match.
Part of zNose® software is an expandable library of chemical smells and indices, known as Kovats indices, and is based on the measured retention times from a n-alkane vapor standard. This provides an easy way to calibrate and tentatively detect unknown odors in the field because it needs only a single calibration standard for all compounds present within the user library of smells and Kovats indices.
Figure 12. System response to a vapor standard containing nalkane vapors C7 to C14. Indexed compound retention times relative to that of an n-alkane is called Kovats Indices.
BTXX and BTEX Calibration
Calibration vapor standards for ethylbenzene, toluene, benzene, and the m,p, and o-xylene were produced by filling tedlar bags from gas canisters with certified concentrations of these compounds. BTEX - one standard Vapor - contained 1 ppm of benzene, o-xylene, ethylbenzene, and toluene. Another standard vapor (BTXX) contained 1 ppm of each of the three xylenes, 1 ppm benzene, and toluene.
Both m and p-xylenes co-elute together, and therefore cannot be easily separated from each other or from ethylbenzene. Identical retention times were showed by two other compounds at this site, thiopene and benzene. However, since co-eluting compounds possess similar response factors, total concentrations were determined and calibrated as total m-, p-xylene.
With the help of software, users can graphically select alarm bands, alarm levels, retention time¸ or odor thresholds for chosen chemicals within an odor. Response factor and peak identification data are stored in files that contain all pertinent calibration information for particular odors.
Figure 13. Tedlar bags make constant concentration vapor standards for calibration of the zNose®.
Sensitivity to the BTEX and BTXX standard vapors was characterized by response factors in counts (cts) per ppm. For benzene, which is the lightest compound, the response factor was about 100 cts/ppm using a 20 °C detector and a 15 milliliter vapor sample (30 second sample time). When replicate measurement methods were used, the minimum detection level was approximately 300 ppb.
Reducing the detector temperature to 0 °C resulted in increased sensitivity and reduced the minimum detection level for benzene to 40 ppbv. There were proportionally larger response factors for higher molecular weight compounds such as 30,000 cts/ppm for naphthalene and 3000 Cts/ppm for o-Xylene. Retention times were expressed either in seconds or as Kovats indices referenced to a file that contained the system response to n-alkane vapors.
Figure 14. Expanded response to BTXX and BTX calibration vapors.
Figure 15. Response and alarm window settings for 1 ppm BTEX.
Figure 16. Peak identification file for BTEX standard vapors. Retention times are listed as Kovats indices and response factors are per ppm.
Outside Air Measurements
At several locations, odors and their intensity within and surrounding the remediation site were quantified in real time. One particular location, downwind from the site, was adjacent to an entrance gate roughly 100 feet from where active excavation was being performed.
To sample ambient air, the zNose® was placed on top of a 3 foot high concrete wall facing into the site.
Figure 17. Real time monitoring of site odors located at street entrance (downwind).
Figure 18. Downwind location (arrow) near active excavation of contaminated soil.
Every 80 seconds, repetitive measurements of the site odors were taken using a 30 second vapor preconcentration (15 ml) and then by a <1 minute analysis time and 30 second recovery. In Figure 19, offset chromatograms reveal a series of 10 analysis runs which started at approximately 9 am, soon after active work was began on the site. 35 measurements were taken over a period of 50 minutes.
Since methyl naphthalene and naphthalene dominate site odor chemistry, both these compounds were employed to track the odors being released from the site.
Figure 19. Consecutive measurements were taken every 80 seconds using a 30 second sample time, <1 minute analysis, and 30 second recovery time. The prominent peak at 5.7 seconds is naphthalene.
It is assumed that other trace elements within the odor will differ in proportion to the concentration of these compounds.
The intensity of site odors demonstrated significant short term variation in odor concentration; in the warm afternoon, concentrations were higher than during the cooler morning hours. Naphthalene concentration, during a one hour period in the morning, differed from 15 to 10 ppbv with an average of 11.4 ppbv. This difference of the odor concentration was reflected in a standard deviation of 43% for 35 successive measurements.
As the odor threshold is 27 ppbv for naphthalene, morning odors at this location and time may not be detected.
Measurements taken at the same downwind location in the late afternoon and over a period of 10 minutes showed a sizeable increase in the concentration of naphthalene odors. After an upward trend, the concentration of naphthalene as high as 60 ppbv, much above the odor threshold of 27 ppb, were determined.
As expected, ambient air within 10 feet of contaminated soil indicated high odor concentration. Soil piles that were arranged into open bins and had to be treated with biochemical and chemical reducing agents was tested over a period of 1 hour in the late morning. The location was somewhat protected from winds, but the concentration of both methyl naphthalene and naphthalene still had large short-term variations and routinely surpassed the odor threshold.
Figure 20. Concentration of Naphthalene and Methyl Naphthalene at downwind location.
Figure 21. Concentration of Naphthalene at downwind location in afternoon
Figure 22. Ambient odors were measured near soil being treated to reduce odors
Figure 23. Naphthalene odor concentration near soil bins
Summary and Results
Based on ultra high-speed gas chromatography, a new, advanced electronic nose now enables precise and accurate quantification of odor chemistry in near real time. Over a period of 3 days, more than 800 odor measurements were carried out at various locations in and around a soil remediation site that was contaminated with coal tar.
Based on chemical measurements, a visual olfactory image clearly demonstrated that naphthalene was the most dominant chemical compound in the site odor; however, other hydrocarbon elements were also present but at lower concentrations.
The instrument’s sensitivity allowed quick and easy measurement of odor chemical concentrations at low ppt levels. Headspace vapors, in foil-covered bucket samples of contaminated soil, exhibited vapor concentrations at part per million levels. The concentrations of chemical vapor from toluene (5.7 ppm), benzene (9.5 ppm ), m,pXylene (12.6 ppm ), methyl naphthalene (2.5 ppm ), naphthalene (17 ppm ), and various trace elements were determined and their relative concentrations were able to define the odor signature (VaporPrint®) of the site.
Concentrations of ambient air vapor close to the contaminated soil (less than 1 foot) were found to be in the low 1-10 ppm concentration range. Odor concentrations at a downwind location adjacent to the site (about 200 feet from active excavation) were found to be in the 10 to 50 ppb range. Odor concentrations at an upwind location were much lower, usually in the part per trillion range.
Considerable short term variability was shown by replicate odor samples (30 second) taken every 80 seconds, for example 43% standard deviation for 35 samples. While afternoon levels were considerably higher (typically 60 ppbv downwind adjacent to the site), morning levels of naphthalene were somewhat below the odor threshold levels (27 ppbv).
Since the electronic nose is based on the science of gas chromatography, odor measurements can now be validated through independent laboratory measurements taken on quality control samples acquired from the site. Real-time, on-site analytical measurements of odors and VOCs provide site managers with a new cost effective tool to monitor these compounds and thus reduce the effect of site odors on the surrounding community.
This information has been sourced, reviewed and adapted from materials provided by Electronic Sensor Technology.
For more information on this source, please visit Electronic Sensor Technology.