United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-97/055
October 1997
&EPA
Field Validation of a
Penetrometer- Based
Fiber-Optic Petroleum,
Oil, and Lubricant (POL)
Sensor
-------
-------
EPA/600/R-97/055
October 1997
Field Validation of a Penefrometer-Based
Fiber-Optic Petroleum, Oil, and
Lubricant (POL) Sensor
by
William C. McGinnis and Stephen H. Lieberman
Research, Development, Test and Evaluation Division
Naval Command, Control and Ocean Surveillance Center
San Diego, CA 92152
Interagency Agreement #DW17936217
Project Officer
Charlita G. Rosal
Characterization and Monitoring Branch
Environmental Sciences Division
Las Vegas, NV 89193-3478
This study was conducted
in cooperation with
U.S. Department of Defense
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Printed on Recycled Paper
-------
Notice
The U.S. Environmental Protection Agency, through its Office of Research and Development, partially
funded and collaborated in the research described here under Interagency Agreement # DW17936217
with the U.S. Navy. It has been subjected to the Agency's peer and administrative review and has been
approved for publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
Abstract
This report details comprehensive comparisons of in situ measurements from a cone penetrometer-
deployed laser induced fluorescence (LIF) petroleum, oil, and lubricant (POL) sensor with traditional field
screening methods. Operational procedures were developed to facilitate comparison between methods
and across multiple sites. Using a field screening detect/non-detect criterion, agreement between sensor
measurements corresponding to the sampled interval and the laboratory analytical measurements on
those samples was better than 85 percent. Comparison between measurements from the two accepted
analytical techniques, on splits of the same sample, was only slightly better. We conclude that the LIF-
POL sensor, deployed from a cone penetrometer, provides significant advantages for subsurface field
screening of POL-contaminated sites. The LIF technique offers the advantages of rapid, in situ, real-time
measurements, coupled with increased data density, not possible with traditional screening methods.
This report was submitted in fulfillment of Interagency Agreement # DW17936217 by the Naval Command,
Control, and Ocean Surveillance Center (NCCOSC), Research, Development, Test and Evaluation
(RDT&E) Division, Code 521 under the partial sponsorship of the U.S. EPA. This report covers the project
period of 07/01/93 to 06/30/96
11
-------
Preface
Success of initial research efforts in laser induced fluorescence (LIF) combined with the
need for effective field screening led to development of the Site Characterization and
Analysis Penetrometer System (SCAPS), an on-site LIF cone penetrometer (CPT) for in
situ chemical measurement. To meet restoration needs of the Navy/DOD environmental
community, the Navy embarked on an aggressive program pursuing LIF sensor
validation/certification and commercialization through technology transfer.
The validation/certification . process was aimed at expeditiously achieving regulatory
acceptance of LIF CPT at all levels, ranging frohrthe local to the national level. The
coordination of validation/certification efforts has resulted in supporting multiple
objectives. In collaboration with the Consortium for Site Characterization Technology, a
partnership between EPA, DOD, and DOE, a SCAPS technology demonstration was
performed at two sites: the Hydrocarbon National Test Site, Naval Construction Battalion
Center, Pt. Hueneme, CA, and a spill site at Sandia National Laboratory, Albuquerque,
NM. An Innovative Technology Evaluation Report, EPA/540/R-95/520, was generated by
the EPA as a result of these demonstrations. The Western Governor's Association
Demonstration of Innovative Technology (DOIT) committee participated as observers and
evaluators in the SCAPS U.S. EPA Consortium demonstration. Western Governor's
Association cooperation on regulatory acceptance at the state level is extended to several
western states including: Idaho, Utah, Nebraska, and Texas. Regulatory acceptance at
the state level is being pursued through the Interstate Technology and Regulatory
Cooperation Technology Specific Task Group.
A detailed description of the LIF-CPT technology, capabilities and limitations, and
historical site data was compiled and submitted to the State of California - Environmental
Protection Agency for certification under Hazardous Waste Environmental Technologies
Certification Program (AB2060). SCAPS was awarded California EPA certified
technology certificate number 96-01-021. In addition to the formal technology
demonstration, the Navy's SCAPS has been employed for field screening investigations at
16 sites. A second/validation phase to each of these operational outings consisted of
confirmatory sampling and analysis. Funding provided through this interagency
agreement between the EPA and the Navy was used to perform confirmatory sampling
work at two sites, Naval Air Station, Alameda, CA, and Guadalupe Oil Field, Guadalupe,
CA.
Commercialization through technology transfer has been pursued to provide the Navy
with a means to contract for LIF-CPT services. Two companies hold licenses to the
patented technology: Unisys Corporation provides LIF-CPT services using the Rapid
Optical Screening Tool (ROST); while Vertek/ARA, Inc. markets a line of CPT products
including a Fuel Fluorescence Detector (FFD) product line.
111
-------
Table of Contents
Notice ii
Abstract ii
Preface in
List of Figures v
List of Tables v
Acronyms and Abbreviations vi
Acknowledgments viii
Chapter 1 introduction 1
System Description 1
Technology Constraints 4
Chapter 2 Conclusions 7
Chapters Recommendations 9
Chapter4 Methods and Materials 10
Calibration Procedures 10
Sampling Procedures 11
Analytical Methods 11
Data Reduction and Analysis Methods 12
Fluorescence Threshold and Detection Threshold Calculation 13
Site Description: Naval Air Station, Alameda 14
Site Description: Guadalupe Oil Field 15
Chapters Results and Discussion 17
Summary Results for Sixteen Sites 18
Naval Air Station, Alameda Results 19
Guadalupe Oil Field Results 22
References 27
IV
-------
List of Figures
Figure 1. Schematic of fiber-optic LIF POL sensor system for petroleum hydrocarbons 2
Figure 2. SCAPS in situ LIF Spectra from Alameda push P44 3
Figure 3. Real-time data display for Alameda push P44 3
Figure 4. Validation sampling scheme 12
Figures. Cumulative scatter/contingency plot of LIF vs. TRPH (n = 552) 19
Figure 6. Cumulative scatter/contingency plot of LIF vs. TPH (n = 552) 20
Figure 7. Cumulative scatter/contingency plot of TPH vs. TRPH (n = 552) 20
Figures. MAS Alameda scatter/contingency plot of LIF vs. TRPH (n = 45) 22
Figure 9. MAS Alameda scatter/contingency plot of LIF vs. TPH (n = 45) 23
Figure 10. NAS Alameda scatter/contingency plot of TPH vs. TRPH (n = 45) 23
Figure 11. Guadalupe scatter/contingency plot of LIF vs. TRPH (n = 16) 25
Figure 12. Guadalupe scatter/contingency plot of LIF vs. TPH (n = 16) 26
Figure 13. Guadalupe scatter/contingency plot of TPH vs. TRPH (n = 16) 26
List of Tables
Table 1. Soil Classification Cross Reference 4
Table 2. Detection Limits for the LIF POL Sensor 4
Table 3. Sixteen SCAPS Validation Sites 18
Table 4. Sixteen Site Cumulative Contingency Analysis Results Summary (n=552) 19
Table 5. Summary of SCAPS Validation Sampling at NAS, Alameda 21
Table 6. NAS, Alameda Contingency Analysis Results Summary (n = 45) 24
Table 7. Summary of SCAPS Validation Sampling at Guadalupe Oil Field 24
TableS. Guadalupe Oil Field Contingency Analysis Results Summary ...: 25
-------
Acronyms and Abbreviations
AEC
ASTM
ATI
bbl
bgs
cm
CAS
CPT
CSCT
DFM
DHS
DOD
DOE
DOT
DQO
EDM
EMMC
EnTICE
EPA
ESD-LV
ETI
ft
ft/ft
FVD
FY
gal
GC/FID
HASP
HNTS
HSA
Hz
1AG
IDW
in
!R
IRP
ITER
kg
km
L
Ib
LIF
m
urn
mg
mg/kg
mg/L
mi
min
mJ
mL
Army Environmental Center
American Society for Testing and Materials
Analytical Technologies, Inc.
Barrel (equivalent to 42 U.S. Gallons)
Below Ground Surface
Centimeter
Chemical Abstract Service
Cone Penetrometer Testing
Consortium for Site Characterization Technology
Diesel Fuel Marine
Department of Health Services (California)
Department of Defense
Department of Energy
Department of Transportation
Data Quality Objective
Engineering Development Model
Environmental Monitoring Management Council
Environmental Technology Innovation, Commercialization, and Enhancement
(Program)
Environmental Protection Agency
Environmental Sciences Division-Las Vegas
Environmental Technology Initiative
Foot
Feet per Foot
Fluorescence Versus Depth
Fiscal Year
Gallon
Gas Chromatography/Flame lonization Detector
Health and Safety Plan
Hydrocarbon National Test Site
Hollow Stem Auger
Hertz
Interagency Agreement
Investigation Derived Waste
inch
Installation Restoration
Installation Restoration Program
Innovative Technology Evaluation Report
Kilogram
Kilometer
Liter
Pound
Laser-Induced Fluorescence
Meter
Micrometer
Milligrams
Milligrams per Kilogram
Milligrams per Liter
Mile
Minute
Millijoule
Milliliter
VI
-------
Acronyms and Abbreviations (continued)
mm
m/min
ms
msl
NCCOSC RDT&E
NERL
NFESC
nm
NRaD
ns
PAH
PDA
PE
PM
POL
PPE
ppm
PRO
QA
QAPP
QC
R2
RAB
RI/FS
ROST
s
SCAPS
SOP
SPT
TER
TPH
TPM
TRPH
TSF
U.S.
uses
UV
WES
WTM
Millimeter
Meters per Minute
Millisecond
Mean Sea Level
Naval Command, Control, and Ocean Surveillance Center Research,
Development, Test, and Evaluation (Division)
National Exposure Research Laboratory
Naval Facilities Engineering Service Center
Nanometer
Unofficial Abbreviation for NCCOSC RDT&E Division
Nanosecond
Polycyclic Aromatic Hydrocarbon
Photodiode Array
Performance Evaluation
Program Manager
Petroleum, Oil, and Lubricant
Personal Protective Equipment
Parts per Million
PRC Environmental Management, Inc.
Quality Assurance
Quality Assurance Project Plan
Quality Control
Correlation Coefficient
Restoration Advisory Board
Remedial Investigation/Feasibility Studies
Rapid Optical Screening Tool
Second
Site Characterization and Analysis Penetrometer System
Standard Operating Procedure
Standard Penetrometer Testing
Technology Evaluation Report
Total Petroleum Hydrocarbons
Technical Project Manager
Total Recoverable Petroleum Hydrocarbons
Tons per Square Foot
United States
Unified Soil Classification System
Ultraviolet
Waterways Experimental Station (Army Corps of Engineers)
Wavelength Time Matrix
vn
-------
Acknowledgments
The SCAPS-LIF technology was the result of a Tri-Service (Army, Navy, Air Force) collaborative
development effort with major support from the Strategic Environmental Research and Development
Program (SERDP). This report reflects the efforts of many people including the SCAPS field crew and
post processing/data analysis personnel - Robert Cook, BJ King, and James Melega of NCCOSC RDT&E
Division; Peter Stang and Donald McHugh of PRC, Environmental Management, Inc.; Lora Kear and
Michele Davey of Computer Sciences Corporation; and Karina Wu and Amy Walker of San Diego State
University Foundation.
Vlll
-------
Chapter 1
Introduction
The U.S. Environmental Protection Agency (EPA),
Characterization Research Division-Las Vegas (CRD-
LV) evaluated field screening techniques to expedite site
characterization and monitor corrective actions. In
collaboration with the Naval Command, Control and
Ocean Surveillance Center, RDT&E Division (NRaD), the
effort described here was undertaken to evaluate the use
of a cone penetrometer system equipped with a fiber
optic-based laser-induced fluorescence (LIF) petroleum,
oil, and lubricant (POL) sensor for real-time field
screening of subsurface POL contamination.
The feasibility of using a truck-mounted cone
penetrometer system to push chemical sensors into the
ground to delineate subsurface contaminant plumes was
first demonstrated through the Department of Defense
(DOD) Tri-Service Site Characterization and Analysis
Penetrometer System (SCAPS) program. The LIF cone
penetrometer test (CPT) technology was developed
through a collaborative effort of the Army, Navy, and Air
Force under the Tri-Service SCAPS program. To satisfy
the objective of this Inter-Agency Agreement (IAG), a
comprehensive inter-comparison effort was established
to directly compare sensor results with conventional
sampling and laboratory analyses. This effort was
proposed as a jointly funded collaborative effort between
the U.S. EPA and the U.S. Navy and as such leverages
funding provided by Naval Facilities Engineering
Command. This IAG enabled confirmatory sampling
work to be performed at two sites: Naval Air Station,
Alameda, CA, and Guadalupe Oil Field, Guadalupe, CA.
To date, confirmatory sampling has been performed at
16 sites in addition to the two above listed sites,
therefore summary results will be presented.
System Description
The SCAPS system uses a truck-mounted cone
penetrometer testing (CPT) platform to advance its
chemical and geotechnical sensing probe into
subsurface soils. CPT has been widely used in the
geotechnical industry for determining soil strength and
soil type from measurements of tip resistance and sleeve
friction by an instrumented probe. The CPT platform
provides a 20-ton static reaction force associated with
the weight of the truck. The forward portion of the truck-
mounted laboratory is the push room. It contains the
rods, hydraulic rams, and associated system controllers.
Underneath the SCAPS CPT push room is the steam
cleaning manifold for the rod and probe decontamination
system. The rear portion of the truck-mounted
laboratory is the isolatable data collection room in which
components of the LIF system and onboard computers
are located. The combination of reaction mass and
hydraulics can advance a 1-m long by 3.57-cm diameter
threaded-end rod into the ground at a rate of 1 m/min in
accordance with American Society of Testing and
Materials (ASTM) Standard D3441, the standard for
CPT. the rods, various sensing probes, or sampling
tools can be advanced to depths in excess of 50 m in
naturally occurring soils. As the rods are withdrawn,
grout can be injected through %-in-diameter tubing within
the interior of the probe's umbilical cable, hydraulically
sealing the push hole. The platform is fitted with a self-
contained decontamination system that allows the rods
and probe to be steam cleaned as they are withdrawn
from the push hole, through the steam cleaning manifold,
and back into the CPT push room. Subsurface
investigation in this manner produces rinsate but no soil
cuttings as investigation derived waste. A schematic of
the LIF system as deployed in the cone penetrometer is
shown in Figure 1.
LIF sensors rely on impinging ultraviolet (UV) light to
excite molecular electrons to excited/higher energy
states. As the electrons return to lower energy states,
the transition produces UV fluorescence photons of
longer wavelength than the UV excitation. The LIF probe
consists of a standard penetrometer probe modified with
a %-in-diameter, flush-mounted sapphire window which
is 24 in behind the probe tip. Two 500-nm silica clad
silica optical fibers, one for laser excitation and one for
fluorescence emission, are included in the 300-ft
umbilical cable and are internally mounted in the probe
terminating at the sapphire window. Excitation light at
337 nm, generated from a pulsed nitrogen laser (0.8 ns
pulse width, 1.4 mJ pulse energy), travels down the
optical fiber and excites fluorescence from polycyclic
aromatic hydrocarbons (PAHs) in the soil (Apitz et. al.,
1992a). The method detects PAHs in the bulk soil matrix
throughout the vadose, capillary fringe, and saturated
-------
Figure 1. Schematic of fiber-optic LIF POL sensor system for petroleum hydrocarbons.
zones. The emission fiber collects the laser-induced
fluorescence and returns it to the surface.
At the surface, the fiber is coupled to a spectrograph
where the light is spectrally dispersed. The dispersed
light then impinges on an intensified linear photo diode
array detector (1024 pixels) which is gated on for 100 ns
at the time of signal return. An optical trigger from the
pulsed laser via a pulse delay generator is used to gate
the detector. The laser-induced fluorescence signal is
emitted over a broad range of wavelengths longer than
the excitation light. Approximately 16 ms is required to
read the fluorescent signal from a single laser shot. The
maximum spectral resolution is approximately 0.5 nm.
The detector is set to measure the wavelength range
from 350 nm to 720 nm. In practice, the system usually
integrates the emission from 20 laser shots with detector
pixels grouped by four. Since the laser repetition rate is
10 Hz, the total time to collect a fluorescence emission
spectrum is 2 s. This represents approximately a 2.5-in
depth resolution. The spectral resolution for pixels
grouped by four is approximately 2 nm. An optical
multichannel analyzer accumulates the detector readings
and reports the sum as a single measurement to the
data acquisition computer. Figure 2 presents several in
situ spectral curves from Alameda push P44.
Data acquisition is automated under software control
using a 486 host computer. The computer sets and
controls the sensor system, stores fluorescent emission
spectra and strain gauge data, and generates the real-
time depth plots shown in Figure 3. From the spectral
curve at each depth, the SCAPS software extracts the
maximum intensity and associated peak wavelength for
real-time depth display. The Raw Fluorescence and
Wavelength at Peak strips of Figure 3 contain this data.
SCAPS standard electrical cone penetrometer
instrumentation consists of strain gauges measuring tip
resistance and sleeve friction in accordance with ASTM
Standard D3441. An empirical relationship between tip
resistance and sleeve friction provides a soil type
classification relating to grain size, Table 1 (Robertson
and Campanella, 1989). This data is contained in the
real-time display strips as Cone Pressure, Sleeve
Friction, and Soil Classification. As the probe is forced
into the ground, the real-time display presents a 10-ft
interval on a scrolling basis. Data acquisition time for the
20-ft push displayed in Figure 3 was about 12 min.
Focusing on the data, the fine scale depth resolution is
readily apparent. Within the Raw Fluorescence profile,
intensity can be seen to increase significantly above a
-------
SPECTRAL PLOT(S)
350.0 396.3
Time: 07:03:32
Date: 04-06-1994
535.0 581.3 627.5
WAVELENGTH IK NM
Main: C:\BASIC71\DATA\ALA13P«.PSH
Probe: C:\BASIC71\DATA\PROBEL1.PRB
Calibration: C:BAS1C71\DATAU36APRDFM.CAL
673.8 720.0
Figure 2. SCAPS in situ LIF spectra from Alameda push P44.
Cone
Pressure
Sleeve
Friction
Soil
Classification
Raw
Fluorescence
Wavetengtfi
at Peak
OX» 117S 2SS.1 O.
Figure 3. Real-time data display for Alameda push P44.
-------
Table 1. Soil Classification Cross Reference
(Robertson and Campanella, 1989)
Soil Classification
Number
1
2
3
4
5
6
7
8
9
10
11
12
N/A
Soil Behavior
Type
sensitive fine grained
organic material
clay
silty clay to clay
clayey silt to silty clay
sandy silt to clayey silt
silty sand to sandy silt
sand to silty sand
sand
gravely sand to sand
very stiff fine grained (overcohsolidated)
sand to clayey sand (overconsolidated)
mixed soil types across 6 in sample
background value for an interval and then return to the
background value with continued penetration. This
observation supports the conclusion that the probe
window ts self cleansing.
SCAPS LIF is a nonspecific field screening technique
which detects PAH compounds with at least two
aromatic rings but is most effective for three and more
aromatic rings. To date, LIF measurements over optical
fibers have not been used extensively for detection of
BTEX (benzene, toluene, ethylbenzene, and xylenes)
compounds. Greater attenuation of shorter wavelength
UV radiation in optical fibers is a technological barrier for
transmitting the excitation pulse over long fiber lengths.
Table 2 presents detection limits for common fuel
products found as soil contaminants. Measurements
were made in the laboratory on spiked soils over a 50-m
fiber and are reported at the 95 percent confidence level.
Detection limits vary with fuel type depending on
constituent compound abundance. Detection limits also
vary with soil type due to particle size and mineralogy.
Most importantly, these limits fall well within the range of
utility considering regulatory action limits.
Technology Constraints
The SCAPS CPT support platform is a 20-ton all-wheel
drive diesel powered truck. The dimensions of the truck
require a minimum access width of 10 ft and a height
clearance of 15 ft. It is conceivable that some sites, or
certain areas of sites, might not be accessible to a
vehicle the size of the SCAPS CPT truck. The access
limits for the SCAPS CPT vehicle are similar to those for
conventional drill rigs and heavy excavation equipment.
The CPT sensors and sampling tools may be difficult to
advance in subsurface lithologies containing cemented
sands and clays, buried debris, gravel units, cobbles,
boulders, and shallow bedrock. As with all intrusive site
characterization methods, it is extremely important that
all underground utilities and structures be located before
undertaking activities at a site. This should be done
geophysically even if subsurface utility plans for the site
are available for reference.
Table 2. Detection Limits for the LIF POL Sensor
Soil Type
Soil A - Sand
Soil B - China Lake
Soil C - Columbus
Unleaded
Gasoline
17 ppm
36 ppm
121 ppm
Fuel Type
Diesel Fuel #2
329 ppm
25 ppm
83 ppm
Diesel Fuel, Marine
(DFM)
14 ppm
4 ppm
5 ppm
-------
The LIF sensor response may be sensitive to
hydrocarbon variability, matrix effects, and non-
hydrocarbon fluorescent interferents. The relative
response of the LIF POL sensor depends on the specific
analyte being measured. The instrument's sensitivity to
different hydrocarbon compounds can vary by as much
as two orders of magnitude (Lieberman et al., 1992;
Apitz et al., 1992a; Apitz et al., 1992b; Davey et al.,
1994a). These variations in sensitivity are primarily a
reflection of the variations in the PAH distribution found
within petroleum hydrocarbon products. Other
contributing factors such as optical density, self
absorption, and quenching are less important. The LIF
POL sensor responds only to PAHs that fluoresce when
excited at 337 nm. This wavelength will excite aromatic
compounds with three or more rings as well as some
two-ring compounds (Lieberman et al., 1993). Aliphatic
species and single-ring aromatics do not contribute to
the LIF POL sensor signal. The total observable
fluorescence produced by any given petroleum
hydrocarbon sample depends on the mole fraction of
fluorescing PAHs along with the relative quantum
efficiency of each of the fluorescing species. The
fluorescence properties of a hydrocarbon contaminant in
soil may also change after long-term exposure to and
interaction with the environment. A contaminant that has
been in the ground for any period of time will undergo
changes in chemical composition due to weathering,
biodegradation, and volatilization. In terms of
degradation and transport, the BTEX compounds
volatilize and biodegrade quickest, the lighter PAHs
biodegrade slowly, and the heavier PAHs are persistent.
The heavier PAHs are preferentially excited by the 337
nm laser source used in the LIF POL sensor.
The in situ fluorescence response of the LIF sensor to
hydrocarbon compounds is sensitive to variations in the
soil matrix. Matrix properties that affect LIF sensitivity
include surface area, grain size, mineralogy, and
moisture content. Each of these factors influences the
relative amount of analyte that is adsorbed on, or
absorbed into, the soil. Only the relative fraction of
analyte that is optically accessible at the window of the
probe can contribute to the fluorescence signal. Of the
four influencing factors mentioned above, the dominant
variable appears to be soil surface area (Apitz et al.,
1992b). LIF sensitivity to petroleum hydrocarbons in soil
has been shown to be inversely proportional to the
available surface area of the soil substrate (Apitz et al.,
1992c). Sandy soils have a much lower total available
surface area than clayey soils. Hydrocarbon compounds
in sandy soils generally yield a correspondingly higher
fluorescence response than they do in clay-rich soils
(Apitz et al., 1992c). The relative LIF response to diesel
fuel marine (DFM) in each soil is essentially identical
once the response curves were normalized to the
available surface area of each of the soils. The moisture
content of the soil matrix is another influencing factor.
The LIF sensitivity to petroleum hydrocarbons generally
increases with greater soil moisture content, although the
effect appears to be small. It was shown that increasing
the amount of water in a soil tends to narrow the
sensitivity difference between sandy and clayey soils
(Apitz et al., 1992b). It is thought that water physically
displaces the hydrocarbons from within the pore spaces
of the matrix, effectively reducing the surface area
available to contaminants. The effects of soil grain size
has also been examined in laboratory studies. LIF
sensitivity generally increases with increased grain size.
The measured fluorescence was shown to be
substantially greater in the clean coarse sands.
The LIF POL sensor is sensitive to any material that
fluoresces when excited with ultraviolet wavelength light.
Although intended to specifically target petroleum
hydrocarbons, the excitation energy produced by the LIF
system's laser may cause naturally occurring or
introduced substances to fluoresce as well. Many
common fluorescent minerals can produce a measurable
LIF signal. Other non-hydrocarbon fluorescent material
introduced through human activity may be found in the
subsurface environment including: deicing agents,
antifreeze additives, and many detergent products. The
potential presence of fluorescence emission from non-
target (non-hydrocarbon) analytes within the soil matrix
can lead to reduced sensitivity and false positives.
However, the LIF sensor collects full spectral information
which permits discrimination between hydrocarbon and
non-hydrocarbon fluorescence in most cases.
The LIF POL sensor system uses a multichannel
detection scheme to capture a complete fluorescence
emission spectrum at each point along the push. An
advantage of this approach is that spectral features are
obtained that can be used to associate the signal with a
specific petroleum class, mineral substance, or other
material. The spectral patterns collected in situ provide
the means to uniquely distinguish hydrocarbon
fluorescence from potential interferents. The sensor's
ability to recognize non-hydrocarbon fluorescence has
been tested in several laboratory experiments. In one
study (Andrews and Lieberman, 1994), the spectra of
eight fluorescent minerals and five fluorescent chemicals
were obtained with the LIF sensor. These spectra were
compared with the LIF spectra obtained from multiple
samples of jet fuel, gasoline, diesel fuel, and lubrication
oil. In all cases, the hydrocarbon spectra could easily be
recognized (by both computer algorithm and human
analysts) as being different from the non-hydrocarbon
spectra. The specific substances used in the experiment
were chosen because they fluoresced in the same
spectral region as the fuel products. Many other
fluorescent chemicals and minerals fluoresce in a
-------
spectral region which is far removed from the
hydrocarbon spectra. The materials used included:
Calcium carbonate
Tide® detergent
Aragonite
Prestone® antifreeze
Simple Green® detergent
Turritella agate
Resinous coal
Norbergite
Fluorite
Fossil algae
Scapolite
Quinine sulfate
In addition, the organic component of some soils
contains humus (humic acid). This naturally occurring
residue of plant decay often contains small amounts of
fluorescent PAHs. Laboratory tests (Davey et al.,
1994b) have demonstrated that humic acid does not
interfere with LIF POL sensor detection of hydrocarbon
in soil. This is because humic acid fluorescence is
minimal at concentrations found in even the most
organic-rich soils.
-------
Chapter 2
Conclusions
The SCAPS technology was developed to provide rapid,
in situ, real-time field screening of the physical and
chemical characteristics of subsurface soil at hazardous
waste sites. The current configuration is designed to
quickly, and cost effectively distinguish hydrocarbon-
contaminated areas from unimpacted areas. Although
LIF induces only the PAH portion of the petroleum
hydrocarbons to fluoresce, petroleum hydrocarbons are
the general target analytes. This capability allows further
investigation and remediation decisions to be made more
efficiently, on site, and reduces the number of samples
that need to be submitted to the laboratory for costly
confirmatory analyses. A site can then be further
characterized with reduced numbers of borings or wells
placed on a plume specific sampling pattern rather than
a grid. Remediation efforts can be directed on an
expedited basis as a result of the immediate availability
of the LIF and soil matrix data. Further, the SCAPS CPT
platform: (1) allows for the characterization of
contaminated sites with minimal exposure of site
personnel and the community to toxic contaminants, and
(2) minimizes the volume of investigation-derived waste
(IDW) generated during typical site characterization
activities.
The following is a list of specific conclusions derived from
validation efforts and general operations.
• Near-continuous measurements generated by
the sensor provide detailed mapping of the
distribution of subsurface petroleum
contamination. At standard push rates of 1
m/min, fluorescence data is typically collected at
intervals of 6 cm.
• The distribution of contamination provided by the
LIF POL sensor push data shows qualitative
agreement with the pattern of contamination
levels derived from laboratory analytical
measurements (EPA Method 418.1 and EPA
Method 8015-Modified) of semicontinuous soil
samples.
« The method provides a "detect/non-detect" field
screening capability relative to a specified
detection threshold derived for a specific fuel
product on a site-specific soil matrix. Calibration
procedures have been developed to provide the
detection threshold. This procedure is used to
report the detection capability of the LIF POL
sensor, specified in both intensity units and in
concentration units common to traditional
analytical methods.
• Direct comparison of sensor data with samples
collected using a split spoon sampler by
overboring the push hole with a conventional
auger, using the "detect/non-detect" criteria,
shows better than an 85 percent agreement with
conventional laboratory methods (EPA Method
418.1 and EPA Method 8015-Modified).
• The LIF POL sensor uses a detector system
comprised of a spectrograph coupled to a linear
photodiode array detector to collect the spectral
signature of the induced fluorescence emission
response. The entire fluorescence spectrum is
collected and stored for each depth point during
advancement of the probe.
• Qualitative use of spectral data provides a
means of distinguishing different classes of
hydrocarbon products and can be used to
minimize potential false positives from non-POL
fluorescence. Different contaminants often have
a different PAH distribution, resulting in a
distinctive fluorescence spectrum for each class
of contaminants. When dissimilar spectra are
encountered during a site characterization, this
can be indicative of more than one contaminant.
Differences in spectral signatures can also be
used to discriminate between contaminant
products and non-POL fluorescent species
present in the soil.
• Data from the LIF POL sensor is available in real
time as the sensor is advanced into the ground.
This allows real-time decisions on how deep to
sample the site.
• The location of future pushes can also be
decided in real time at the site using the
-------
information available from all previous pushes.
This can greatly speed location of the
contaminant plume edge.
* The LIF method can detect the presence of
hydrocarbons in the bulk soil matrix throughout
the vadose, capillary fringe, and saturated
zones.
• Measurements can be made to depths up to 50
m (150 ft) when the LIF POL sensor is used in
conjunction with an industry-standard 20-ton
penetrometer push vehicle.
• Geotechnical sensors (cone pressure, sleeve
friction) are integrated with the LIF POL sensor
to provide simultaneous continuous geotechnical
and stratigraphic information to aide in
interpreting contaminant distributions and
providing guidance for effective remediation
designs.
• The in situ nature of the LIF POL sensor
minimizes possibilities for contaminating or
altering soil samples that are inherent with
traditional collection, transport, and analysis
procedures.
• The SCAPS CPT provides more accurate
measurement of the depth of the contaminant,
especially for sites where the contaminant is
found in the saturated zone. During typical
operations, the uncertainty in depth with the
SCAPS CPT is approximately 7.5 cm (3 in).
Whereas, uncertainty in conventional methods
are 15 cm (6 in) or greater when flowing sands
are encountered or sample retention is poor.
• The LIF POL sensor produces minimal IDW. A
typical 6-m (20-ft) push with the sensor produces
approximately 38 L (10 gal) of IDW water (used
to clean the push rods). A typical 6-m (20-ft)
auger boring produces 210-235 L (55-75 gal) of
IDW soil as well as 76 L (20 gal) of water used to
clean the flights of the auger. Furthermore, the
penetrometer rods are steam cleaned directly
upon removal from the ground, reducing
potential contamination hazards to site
personnel.
Validation summary results, including examples of
contaminant discrimination by spectral differences from
the Alameda site and a compelling plume boundary
delineation from the Guadalupe site, were presented in a
paper at the Field Screening Methods for Hazardous
Wastes and Toxic Chemicals Symposium (Lieberman et
al., 1995). Additionally, a comprehensive report and
data packages from field screening and validation data of
sites investigated with SCAPS, including the Alameda
and Guadalupe sites, were submitted to the State of
California Environmental Protection Agency (EPA)
Department of Toxic Substances Control and the U.S.
EPA Consortium for Site Characterization Technology
(CSCT). The California EPA evaluation resulted in
SCAPS becoming certified technology number 96-01-
021. CSCT review generated "The Site Characterization
and Analysis Penetrometer System (SCAPS) Laser-
Induced Fluorescence Sensor and Support System
Innovative Technology Evaluation Report," EPA/540/R-
95/520, verifying system performance.
-------
Chapter 3
Recommendations
As a result of field experience including efforts
undertaken to validate the LIF POL sensor technology,
the following items are recommended.
• Further development should be pursued to refine
the LIF measurement technique for expanded
use in additional applications such as monitoring
in situ remediation.
• Continue research efforts to develop better
quantitative aspects by defining the dominant
chemical source of the fluorescence. Improve
contaminant discrimination by spectral signature
using neural network pattern recognition
techniques and developing a database of
fluorescent signatures.
• Develop methods to compensate for matrix
effects using additional sensors and algorithms
accounting for grain size distribution and
volumetric moisture content.
-------
Chapter 4
Methods and Materials
This chapter specifies the SCAPS calibration
procedures, sampling procedures, and data analysis
methods to ensure data quality, integrity, and
comparability. Consistent methods and materials have
been used, including the same analytical lab, for each of
the 16 sites investigated.
Calibration Procedures
Initial system setup requires calibration of a number of
components in the LIF POL sensor system. A time-
delay calibration is performed to set the delay time
parameter which enables gating the detector for the
duration of fluorescence emission return. An automated
software procedure is run to determine the optimum time
delay between laser firing and enabling the detector. A
plot of intensity versus time delay is acquired and
determines the optimum delay. The time delay varies
solely as a function of the optical path length between
the laser and the detector, which changes only with the
length of fiber in the probe umbilical cable. This
procedure is required whenever a different probe
assembly is used and will be carried out prior to the
demonstration testing.
A wavelength calibration is performed for the LIF POL
sensor system to determine the parameters AO and A1,
the intercept and slope of the line that converts detector
pixel number into wavelength. A micrometer on the
spectrograph adjusts the angle of the grating so that the
wavelength impinging on the center of the detector is
500 nm. The center 700 pixels of the 1024 in the
detector are intensified; therefore, the starting pixel is set
to 162 and the pixels to read parameter is set to 700. A
mercury lamp is used to provide known wavelengths for
calibration. A helium-neon (HeNe) laser is used to verify
the calibration. This procedure is required after the
spectrograph, the fiber input to the spectrograph, or the
detector are changed. Recalibration is required when
the wavelength of the fluorescent standard is greater
than 5 nm from the standard value.
Strain gauge calibration is performed in accordance with
ASTM Standard D3441. A load cell device and an
automated software procedure are used to determine the
scale and offset, which converts strain gauge output in
millivolts to tons per square foot (TSF), for both the
sleeve and cone tip strain gauges. This procedure is
required each time a different probe assembly is used or
when strain gauge zero load checks (performed after
each push) differ from zero by more than 1 TSF for the
sleeve and 10 TSF for the cone tip.
Concentration calibration is performed using a set of five
to seven calibration standards (spiked site-specific soil
samples). The standards are prepared by the serial
addition method with typical concentrations of 0, 100,
200, 500, 1000, and 2000 ppm. The calibration
standards are run in triplicate at the beginning of each
day and again when equipment is changed. These
samples are sequentially presented to the sapphire
window for measurement. After measurement, the
average and standard deviation is computed for each
sample. If the standard deviation exceeds 20 percent for
replicate analyses of any single sample, that sample is
rerun. If deviation remains excessive, the quinine sulfate
system check standard (described below) is measured.
If the check standard is out of compliance, system
checkout and debugging is required. Since
measurements are made on native soil, sample
heterogeneity may be the cause of a large deviation. If
this is the case, the number of measurements is doubled
until the deviation falls within 20'percent. A calibration
curve is generated by plotting the average of maximum
fluorescence peak intensity versus the concentration of
fuel product added to the calibration soil sample. A
linear fit is performed yielding slope, intercept, and
correlation coefficient, R2. The R2 value must be greater
than 0.90. The calibration curve will be regenerated if R2
is less than 0.90.
System performance is monitored before and after each
push by measuring fluorescence from a prepared
fluorescent standard of 10 mg/L quinine sulfate in a
dilute nitric acid solution. This measurement also
provides a means for normalizing measurements. Both
wavelength and intensity of the standard are monitored.
If the wavelength of the quinine sulfate standard
measurement deviates by more than 5 nm from the
accepted wavelength of 458 nm, detector calibration is
performed. If the fluorescent intensity changes by more
than 20 percent of the initial value determined during
10
-------
pre-push calibration, system trouble shooting procedures
are initiated.
Sampling Procedures
SCAPS site operations typically consisted of two phases:
site investigation and validation. In the investigation
phase, pushes were performed to delineate the plume
boundaries. During the validation phase, areas of
interest were selected from the first phase and revisited.
At the selected locations, a validation push was
performed followed immediately by collection of
confirmatory soil samples.
During validation, the SCAPS CPT pushed the LIF probe
into the ground and acquired fluorescence and
geotechnical data. After the probe was pushed to the
total depth anticipated or was blocked from further
penetration, the probe was retracted. The CPT rig
moved away from the location and a hollow stem auger
(HSA) drill rig was positioned approximately 20 cm (8 in)
from the push hole. The HSA rig drilled a hole such that
the advancing auger flights destroyed the push hole
while allowing for the collection of split spoon soil
samples within approximately 7.5 cm (3 in) (horizontally)
of the push cavity. This offset between the push hole
and the auger boring permitted sampling far enough
apart so that the soil samples were not affected by
possible cross contamination due to sloughing down the
penetrometer hole, yet near enough to minimize
variability due to small scale spatial heterogenities of the
soil and the contaminant distribution. This sampling
strategy ensured that samples were representative of the
region sampled by the LIF sensor. Each borehole was
logged by a geologist. Soil samples were collected with
a split spoon sampler lined with 15-cm (6-in) long
stainless steel tubes. The sampler was driven in
advance of the lead auger using a 63.5-kg (140-lb) slide
hammer falling over a 75-cm (30-in) distance, in
accordance with the ASTM 1586 Standard Penetration
Test.
A California modified split spoon sampler was used for
sample collection. The split spoon sampler is a 75-cm
(30-in) long, 7.5-cm (3-in) diameter steel tool. The
sampler consisted of a 10-cm (4-in) long (reduced to 5
cm (2 in) when fully threaded) cutting head or shoe
section, followed by either a 45-cm (18-in) or 60-cm (24-
in) long sample barrel containing three or four 15-cm (6-
in) long stainless steel soil sampling tubes, and ending in
a waste soils catch barrel section.
Soil samples were collected at depth intervals to confirm
the LIF POL sensor depth profile in both background and
elevated fluorescence intervals. The sampler was
overdrilled approximately 15 cm (6 in) prior to retrieval to
reduce the amount of sloughed soils typically in the
bottom of the borehole. Only tubes containing sample
soils that appeared relatively undisturbed were used.
Samples for confirmatory analysis were collected from
the lower (deepest) and middle 15-cm (6-in) soil tubes in
the 45-cm (18-in) sampler. The sample was Teflon™-
sealed, capped, taped, labeled, logged, and placed into a
chilled ice chest. Each confirmatory sample was
analyzed by EPA Method 418.1 (TRPH), a water
analysis method modified for soil, and EPA Method
8015-Modified (TPH). Samples for geotechnical analysis
(soil moisture, grain size, and density) were sealed and
shipped in the stainless steel tubes retrieved from the
split spoon sampler. Those samples chosen for
geotechnical analysis were generally the uppermost
(shallowest) tube of the three from the split spoon
sampler, but only if the tube appeared full as a result of
complete sample recovery by the split spoon sampler.
Figure 4 depicts the validation sampling scheme. In
each boring, these sampling procedures were usually
repeated three to eight times to gather samples for
traditional laboratory analytical measurements. From
three to eight validation borings were performed at a site.
Analytical Methods
EPA Method 418.1, Total Recoverable Petroleum
Hydrocarbon by infrared absorption (TRPH), and EPA
Method 8015-Modified, Total Petroleum Hydrocarbon by
gas chromatography with flame ionization detection
(TPH by GC/FID), represent two of the most frequently
used methods employed for delineating non-volatile POL
contamination. It is important to note that these
analytical methods do not measure exactly the same
constituents that are targeted by the LIF POL sensor but
were selected because they represent the technology
that is currently being used on a day-to-day basis to
make decisions about the distribution of subsurface POL
contamination. This data is then compared with the in
situ fluorescence data gathered with the sensor..
It is recognized that Methods 418.1 and 8015-Modified
are subject to systematic biases related to the
composition of the POL contamination and, therefore, it
is anticipated that there will be some deviations between
results from the sensor and the different methods. In
order to account for these differences and to .better
understand the nature of observed anomalies, splits of
selected validation samples have also been analyzed by
EPA Method 8270 Modified, a full scan gas
chromotography/mass spectrometer (GC/MS) that
includes semi-quantitative data for the alkyiated PAH
compounds and diagnostic dibenzothiophenes and
triterpanes (Douglas et al., 1992). The modified EPA
Method 8270 described by Douglas involves the
identification and straight baseline integration of each
alkyiated PAH homologous series and quantification of
each series (C-1 naphthalenes, C-2 naphthalenes, etc.).
11
-------
In Situ Push
Fluorescence
Drilling Rig
(Split Spoon Sampler)
Waste
Geotech
1 hydrometer
• moisture
• density
1 grain size
Chemistry
~ • TRPH 418.1
•TPH8Q15
• SVOC 8270 enhanced
• Discrete fluorescence
-top
- bottom
- integrated
Visual Log
Figure 4. Validation sampling scheme.
EPA Method 418.1 uses a solvent (Freon™-113) to
extract the hydrocarbons from the soil, and then
measures the strength of the infrared absorption at 2930
cm-1, corresponding to a CH2 stretch vibration. The
absorption is compared to that measured on a standard
mixture of hydrocarbons. This method has the
advantage that it is quick and relatively easy to perform,
even in the field, and is generally not very expensive.
However, it has several drawbacks (Douglas et al.,
1992) including the fact that volatile compounds are
usually lost in the extraction procedure, extraction
efficiency for high-molecular-weight hydrocarbons is
poor, and all soluble materials within the soil, including
contaminants and other benign materials, are extracted,
possibly affecting the measurement. This method relies
on a hydrocarbon mixture such as isooctane, n-
hexadecane, and chlorobenzene as the standard for
quantitative comparison, which is generally not the same
as the petroleum product found at the site. Since all
hydrocarbons do not respond equally to infrared
absorption, the difference between the in situ product
and the standards may result in artificially high or low
readings. Note that the same problem is encountered in
choosing a standard fuel with which to calibrate the LIF
POL measurements.
Method 8015 (TPH) utilizes a gas chromatograph
coupled with a flame ionization detector (GC/FID) to
separate the components of the contaminant by
molecular weight. The hydrocarbon extract is mixed with
a surrogate internal standard (SIS) for quality control,
and a quantitative internal standard (QIS) for
quantification. The chromatogram produced by this
analysis covers the carbon range from C7 through C36
and can help to identify the product type ("fingerprint")
(Douglas et al., 1992) using the n-alkane pattern
distribution; pristane ratio, phytane ratio, and the width of
the unresolved complex mixture.
Data Reduction and Analysis Methods
The LIF POL sensor records fluorescence intensity as a
function of depth as the probe is pushed into the ground.
In addition to this raw data, quinine sulfate is measured
as a system check standard before and after each push,
and a series of calibration samples are measured on a
daily basis during the site operations. Data reduction
and analysis methods are discussed below.
LIF POL sensor data is evaluated on a detect/non-detect
basis to determine percentage agreement between
sensor data above or below a fluorescence threshold
and both TRPH and TPH results above or below a
sensor detection threshold. SCAPS independently
provides detect/non-detect data relative to a specific
detection limit derived for a specific fuel product on a
site-specific soil matrix. The detection limit is determined
for the site by generating a concentration calibration
12
-------
response curve for a set of calibration standards (spiked
site-specific soil samples) prepared by standard addition.
Following the conclusion of site operations, a site-
average quinine sulfate value is calculated by averaging
all the pre-push measurements of the quinine sulfate
standard. For each push, and for the daily calibration
measurement, a dimensionless normalization factor,
equal to the pre-push quinine sulfate measurement
divided by the site average quinine sulfate value, is
calculated. The LIF data from each push are normalized
by dividing the fluorescence intensity by the
normalization factor. The fluorescence intensity values
for the calibration samples are also normalized by
dividing by the normalization factor. The fluorescence
threshold and detection threshold values for each day
are normalized by dividing them by the normalization
factor, which is equivalent to regressing the normalized
calibration data. The normalized threshold values are
averaged to provide an overall site fluorescence
threshold and detection threshold. These average
threshold values are used to determine detect and non-
detect for the verification phase of the site
characterization.
To compare the in situ data with the soil sample analysis
results, the normalized fluorescence intensity
measurements taken at the depths from which the soil
samples were gathered are tabulated. Because the
spacing between LIF data points is less than 15 cm (6
in), the fluorescence data from all points corresponding
to the 15-cm (6-in) interval of soil sample are averaged
to produce a single fluorescence intensity for a given
sample.
Fluorescence data are reduced to a "detect/non-detect"
reading using the fluorescence threshold and associated
detection limit as determined from the calibration
samples. For compiling multiple sites into a summary,
the fluorescence reading for each soil sample is divided
by the fluorescence threshold. A sample with a ratio
greater than or equal to 1 is considered a "fluorescence
detect," and that with a ratio less than 1 is considered a
"fluorescence non-detect." Similarly, the laboratory
analytic result (TPH and TRPH) for each soil sample is
divided by the TPH or TRPH concentration
corresponding to the fluorescence detection threshold. A
sample with a ratio greater than or equal to 1 is
considered a "TPH detect" or "TRPH detect," and a ratio
less than 1 is conside'red a "TPH non-detect" or "TRPH
non-detect."
Fluorescence Threshold
Threshold Calculation
Three quantities are needed
fluorescence threshold and the
and Detection
to determine the
detection threshold:
determined using the calibration samples prepared
immediately prior to the site visit using soil from the site
and standard analytical techniques.
The fluorescence intensity for each calibration sample is
measured daily, in triplicate, at the start of operations.
The three measurements are averaged to provide a
single measured intensity for each concentration. The
data is regressed to establish a slope and intercept. The
intercept is given by the intensity of the unspiked
calibration standard (0 ppm). The slope is found from
the least squares fit using:
Intercept: b = y0 = intensity measured on
0 ppm calibration sample
slope: m
:th
where: y.{ is the fluorescent intensity measured on the i
calibration sample and x. is the concentration of the ith
calibration sample. The variance in the regression is
given by:
V= (n - I/I,(mxi + b- y/
where the variance, V, is the biased estimator of the
residual mean square of the fit and the data, n is the
number of calibration samples, and m, x.t, b, and y. are
defined as above.
The standard deviation, a, of the fit is:
noise, background, and sensitivity. These quantities are
a =
The sensitivity and background are defined as follows:
Sensitivity = slope of fitted data = m
Background = intercept of fitted data = b
The noise is defined as:
Noise = standard deviation of the fit • 1 .00 = a • 1 .00
The noise is defined as the standard deviation multiplied
by 1 .00 in order to establish a conservative fluorescence
threshold. The fluorescence threshold is given as the
sum of the background and the noise values. Using the
standard assumption of a normal "student's t" distribution
statistics, and the number of points used in these fits
(typically 4-5), this corresponds to an 80 percent
confidence limit. This was chosen because the sensor is
used as a field screening tool and it was considered
important to reduce the possibility for false negatives.
13
-------
This may raise the occurrence of false positives but this
Is a conservative error and these positives can be further
investigated.
This procedure is carried out using only the lower
concentration calibration standards. For example, when
using diesel fuel marine (DFM) as the target fuel, the
standards will typically consist of samples with
concentrations of 0, 500, 1000, 1500, and 2000 mg/kg.
Experiments have shown that for the full range of
calibration standards (up to 100,000 mg/kg), the
calibration data is not fit well by linear regression. This is
not surprising because of the complicated interaction
between the fuel and soil type. By restricting the data
set to the low concentration samples, the data fits well
using the linear regression and this approach gives much
more confidence in the sensitivity near the detection
threshold.
The quantities needed to calculate the LIF POL sensor
fluorescence threshold and the detection threshold are
now known. These are determined from:
fluorescence threshold = background + noise
detection threshold = noise / sensitivity
= aim
The fluorescence threshold is the quantitative limit that
the fluorescence intensity must exceed in order to qualify
as a "detect." If the fluorescence intensity is less than
the fluorescence threshold, the sensor indicates "non-
detect."
The detection threshold is the amount of contaminant
(based on the target fuel used to prepare the spiked
calibration samples) that corresponds to the
fluorescence threshold. This is the practical detection
level in milligrams per kilogram as determined from the
calibration standards for a given site. It is found by
taking the fluorescence threshold intensity and working
back to the concentration needed to produce this
intensity.
When the average in situ fluorescence result exceeds
the fluorescence threshold, the data are considered to be
"detect'." Because the soil samples are 15-cm (6-in)
long, the fluorescence for the 15-cm (6-in) interval
associated with each sample is averaged and this
average is compared to the fluorescence threshold. For
a consistent comparison between the fluorescence data
and the reference method data, the TRPH and TPH
measurements are considered to show a "detect" when
the value also exceeds the LIF POL sensor detection
threshold.
A plot of "detect" versus "non-detect" results from each
reference method (TRPH and TPH) and the LIF
fluorescence data using the criteria above will generate
data in four categories: (1) those where both methods
indicate no petroleum hydrocarbons ("non-detect"); (2)
those where both methods indicate the presence of
petroleum hydrocarbons ("detect"); (3) those where the
LlF method indicates petroleum hydrocarbon
contamination but the reference method indicates
petroleum hydrocarbon contamination below the
corresponding site-specific LIF detection threshold
("false positive"); and (4) those where the fluorescence
indicates no detection but the reference method
indicates petroleum hydrocarbons above the
corresponding site-specific LIF detection threshold
("false negative"). The percentage false negatives and
comparability data correlating with the reference method
results were calculated using the following equations:
% False negatives =
x 100
XT
where:
X.+ = Number of samples where fluorescence
is less than the detection threshold and the
corresponding TRPH (or TPH) result is greater than the
corresponding detection limit.
XT = Total number of samples collected for
comparison.
% Comparability =
X"
X++
100
XT
where:
x_ _ = Number of samples where fluorescence
is less than the detection threshold and the
corresponding TRPH (or TPH) result is also less than the
corresponding detection limit.
x++ = Number of samples where fluorescence
is greater than the detection threshold and the
corresponding TRPH (or TPH) result is also greater than
the corresponding detection limit.
xr = Total number of samples collected for
comparison.
Site Description: Naval Air Station,
Alameda
SCAPS field operations were undertaken at Naval Air
Station (NAS), Alameda, CA, Site 13 - Former Oil
Refinery, from the 17th of March through the 6th of April
14
-------
1994. Validation operations were performed from the 4th
through the 6th of April 1994.
NAS, Alameda is located at the western end of Alameda
Island, in Alameda and San Francisco Counties,
California. Alameda Island lies along the eastern side of
San Francisco Bay, adjacent to the City of Oakland. The
air station occupies 1,070 hectares (2,643 acres) and is
approximately 1.24-km (2 mi) long and 0.62-km (1 mi)
wide. Most of the eastern portion of the air station is
developed with offices and industrial facilities; runways
and support facilities occupy the western portion of the
station.
Alameda Island is located within the San Francisco Bay
basin, which lies within the Coast Range physiographic
province of California. The island lies at the foot of a
gently westward-sloping plain that extends from the
. Oakland/Berkeley Hills on the east, to the shore of San
Francisco Bay on the west. Alameda Island is underlain
by approximately 400 to 500 ft of unconsolidated
sediments unconformably overlying consolidated*
Jurassic/Cretaceous Franciscan Formation bedrock.
The unconsolidated units from oldest to youngest, are
Pliocene to late Pleistocene terrestrial and estuarine
deposits, late Pleistocene estuarine deposits, late
Pleistocene/ Holocene alluvial and eolian deposits, and
Holocene estuarine deposits. These units are roughly
equivalent to the Alameda, San Antonio, and Posey
formations; the Merritt Sand; and the Young Bay Mud.
NAS, Alameda is a topographically flat base comprising
approximately two square miles, At Site 13, hydraulic fill
thickness is approximately 3 m (10 ft) at the western
edge and thins easterly to approximately 1.5 m (5 ft).
The hydraulic fill predominantly consists of dark-brown to
brown, silty fine sand (SM), and clayey fine sand (SC)
with minor amounts of clay and gravel. The Holocene
Bay Mud Unit underlies the fill in the western portion and
at the southeast corner of Site 13. Where present, the
Holocene Bay Mud Unit is typically encountered between
2.7 and 3.3 m (9 and 11 ft) below ground surface (bgs)
and consists of a dark gray silty clay with iron oxide
stains near the top. Merritt Sand deposits underlie the
Holocene Bay Mud Unit at approximately 3.6 m (12 ft)
bgs, where the Holocene Bay Mud Unit is present. In
areas where the Merritt Sand is directly overlain by
hydraulic fill, it occurs at depths between 1.5 and 3.3 m
(5 and 11 ft) bgs. The Merritt Sand consists
predominantly of orange-brown silty to clayey fine sand.
Groundwater is encountered beneath the site at
approximately 1.8 m (6 ft) bgs and is most likely under
tidal influences.
Originally a peninsula, the land that is now Alameda
Island was isolated from the mainland in 1876, when a
channel was created through the tip of the peninsula,
linking San Leandro Bay with the main portion of San
Francisco Bay. Dredging was conducted to deepen the
canal and allow commercial and industrial traffic to and
from the island's early industrial sites. These sites
included a borax processing plant and an oil refinery
(Pacific Coast Oil Refinery) which is now known as "Site
13, Former Oil Refinery."
The U.S. Army acquired the land from the City of
Alameda in 1930 and began construction activities in
1931. In 1936, the U.S. Navy acquired title to the facility
and began construction of the air station in response to
the military buildup in Europe prior to World War II. After
entry of the U.S. into the war in 1941, more land was
acquired adjacent to the air station. Following the end of
the war, the Navy returned NAS, Alameda to the mission
of providing support for fleet aviation activities.
Site 13 lies within an area formerly occupied by the
Pacific Coast Oil Refinery. The refinery operated from
1879 to 1903. Refinery wastes and asphaltic residues
were reportedly disposed of at the site.
Site 13, the former Pacific Coast Oil Refinery site has
been contaminated with refinery waste. Previous
investigation has indicated that soils at Site 13 have
elevated levels of TPH, volatile organic compounds
(VOC), semi-volatile organic compounds (SVOC), and
phenolic compounds in soils from the surface to the
groundwater. The petroleum hydrocarbons and SVOCs
are compounds and groups of compounds that are
appropriate for detection with the SCAPS. Additionally,
the soil at Site 13 is considered to be sufficiently
unconsolidated for penetration with the system.
Site Description: Guadalupe Oil Field
SCAPS field operations were undertaken at Guadalupe
Oil Field, Guadalupe, CA, from the 23rd of August
through the 8th of September 1994. Validation
operations were performed on the 7th and 8th of
September 1994.
The Guadalupe Oil Field is located in San Luis Obispo
County, approximately two miles northwest of the town of
Guadalupe, CA. The oil field is bounded by the Pacific
Ocean to the west, the Santa Maria River to the south,
State Route 1 to the east, and unincorporated San Luis
Obispo county to the north.
The Guadalupe Oil Field is situated in a beach dune
environment in the Coastal Range Geomorphic Province.
The soils consist of beach dune sand deposits of
Quaternary Age. Topography varies from sea level
adjacent to the Pacific Ocean to approximately 120 ft
above mean sea level in the eastern portion of the oil
field. Groundwater is present at approximately sea level
beneath the oil field and is subject to minor fluctuations in
15
-------
elevation due to tidal forces in the western portion of the
Held.
Guadalupe Oil Field was operated as a production oil
field between approximately 1949 and early 1994. As
primary production in the field slowed, UNOCAL, the field
operator, elected to implement secondary recovery
techniques to improve recovery in the field. Diluent
injection was one secondary recove'ry technique
implemented by UNOCAL at the Guadalupe Oil Field.
This process involves the injection of diluent into the
producing formation, which causes crude oil in the
formation to become less viscous, thus aiding the
process of pumping the crude oil to the surface. Diluent
is a hydrocarbon with a chemical composition similar to
kerosene. During the secondary recovery process,
diluent releases occurred at multiple locations throughout
the oil field. Potential release points include the two
above ground storage tanks (ASTs) which were used to
store the diluent, the on-grade product distribution lines
which were used to distribute diluent from the ASTs to
the injection wells, and the injection wells. Diluent
release impacts are now known to be wide spread
throughout the UNOCAL Guadalupe Oil Field. In
previous investigations by UNOCAL , light non-aqueous
phase liquid, analyzed as diluent has been detected
above the potentiometric surface at numerous areas
throughout the field.
16
-------
Chapter 5
Results and Discussion
For an in situ field screening measurement technique,
such as LIF, determining the accuracy of the technique
presents a particular challenge. This is because it is not
a simple matter to confidently assign a "true" value to a
subsurface contaminant distribution. With conventional
laboratory-based measurements, the accuracy of the
method is a function of both the sampling errors and
errors associated with the measurement method. To
evaluate the accuracy of a laboratory method, the
conventional approach is to compare the results
obtained from analysis of a spiked sample of known
concentration. It should be recognized, however, that
this approach does not address the issue of whether the
result is an accurate representation of the true value of
the contaminant in the ground. In other words, errors
related to sampling are not addressed. Because there is
no independent measure of the subsurface value of
contaminant concentration, it will be necessary to
evaluate the accuracy of the in situ measurement by
comparing in situ results with results from conventional
methods that may not provide a true value of the
subsurface contaminant distribution because of errors
associated with the sampling process.
It should be noted that the three methods for quantifying
hydrocarbon contamination discussed in this document
(namely the analytic EPA Methods 418.1 and 8015-
Modified, and the LIF method) all measure and quantify
the amount of contaminant using a different physical
property of the contaminant. The EPA Method 418.1
measures the infrared absorption of the extract from the
soil sample. The EPA Method 8015-Modified passes the
extract from the soil sample through a gas
chromatograph and uses a flame ionization detector to
measure the contaminant according to the retention time
of the constituents. The LIF method measures the
fluorescence (under laser excitation) from the PAHs
present in the contaminant.
The two EPA Method measurement techniques (418.1
and 8015-Modified) require comparison to a similar
measurement of a target fuel in order to quantify the
contaminant. Note that it is not possible to ensure that
the target fuel is identical in composition to the
contaminant extract. The EPA Method 418.1 uses a
single standard hydrocarbon mixture for quantification,
while EPA Method 8015-Modified quantifies using a
target fuel that produces a similar chromatogram. The
LIF method does not use an extract from the soil sample,
but it measures the contaminant in situ as it is presented
to the window of the probe. For this reason, the LIF
sensor is more sensitive to matrix effects. Because of
this matrix sensitivity, the LIF sensor does not employ a
target fuel for quantification but only to set a detection
threshold for the site.
Another difference between in situ and conventional
laboratory-based measurements is that laboratory
measurements usually employ extraction or matrix
simplification procedures, whereas in situ measurements
offer limited opportunities for controlling matrix effects.
For the LIF sensors, studies have shown that variability
in sensor response results from changes in the sample
matrix and from variations in fluorescence response
related to fuel product type, age, and origin of the
hydrocarbon contaminant. Since it is not possible to
account for all sources of variability that affect sensor
response at this time, the sensor is intended to operate
as a field screening method. It provides only qualitative
data on the distribution of petroleum hydrocarbon
contamination.
The approach for evaluating accuracy presented here
depends on the direct comparison of in situ sensor data
with the analysis of discrete samples collected as close
as possible to the soil sample measured by the in situ
sensor. Although it is believed that this approach
provides the best opportunity for evaluating the accuracy
of the in situ measurement, it should be noted that it will
not be possible to account for all variability associated
with the uncertainty in depth from which the discrete
samples are collected. It is possible that the depth of the
discrete sample may be in error by up to 15 cm (6 in) in
the vadose zone. Due to sloughing and flowing sand
conditions in the water saturated zone, depth
measurement uncertainty during discrete sampling may
be greater. In stratified soils, sharp vertical boundaries
of the contamination plume may exist. This sampling
error could, therefore, lead to poor comparisons between
in situ data and laboratory data. For example, due to an
error of 15 cm (6 in) in the sample depth, contaminant
concentration can change from strongly impacted
17
-------
(greater than 10,000 ppm) to not impacted (less than 100
ppm). For this reason, the depth of the sample must be
known for the comparability of the samples to be firmly
demonstrated. In addition, because there will be several
inches of horizontal offset between the push location and
the location of the split-spoon sampler, there may also
be some small-scale horizontal variability that will not be
accounted for. Both the vertical uncertainty and the
small-scale horizontal variability will not be a factor when
comparing the two laboratory methods because splits of
a homogenized sample will be measured.
Summary Results for Sixteen Sites
To date, validation efforts at 16 sites have been
completed. A list of sites, dates, and sampling statistics
is provided in Table 3. These sites presented varied
conditions in hydrogeology including: (1) arid with deep
groundwater and (2) coastal with tidally-influenced
shallow groundwater. These sites also presented
various contaminant source products including old
refinery waste, heating oil, diesel fuel marine, and JP-5
jet fuel.
Based on results calculated for the sites to date, the LIF
detection threshold varies somewhat from site to site,
Table 3. Sixteen SCAPS Validation Sites
Site
NAS North Island, Fueling Facility
(N)
NS 32nd Street, Fire Fighting
Traininq Facility (F)
Naval Amphibious Base,
Coronado, Abandoned Tank (C)
NAS, Alameda, Site 13 -Former
Oil Refinery** (A)
MCAS, Yuma, Site 7 - Fire School
Area (Y)
MCB, Camp Pendleton, Air Station
Ground Control Approach Site (M) •
NAS North Island UST (U)
Guadatuoe Oil Field** (G)
Naval Training Center, San Diego,
Gas Station (X)
Naval Training Center, San Diego
Old Auto Hobby ShopJT)
Imperial Beach, Fuel Farm (I)
Marine Corps Recruit Depot, San
Diego, Exchange (D)
Marine Corps Recruit Depot, San
Dieqo. Site13(B)
Naval Radio Receiving Facility,
Imperial Beach (R)
National Hydrocarbon Test Site
Pre-Demo, Port Hueneme (P)
National Hydrocarbon Test Site
Demonstration, Port Hueneme (H)
TOTALS
Dates
14 Jul-31 Augand
5-6 Oct '93
11 Jan-8Feb'94
15Feb-1 Mar '94
17 Mar -6 Apr '94-
17 May -9 Jun '94
27 Jun - 6 Jul '94
25 Jul - 4 Aug '94
23 Aug - 8 Sep '94
24 Oct- 16 Nov '94
2-16Nov'94
30 Nov -15 Dec '94
30 Jan - 9 Feb '95
21 Feb-1 Mar '95
6 - 22 Mar '95
4 - 1 1 Apr '95
17 -22 May '95
Total
Pushes
40
22
22
45
29
25
25
+m -A
33
16
38
25
23
36
14
8
437
Valid*
Pushes
6
3
3
8
4
4
4
'•••'- "'": '4l£X .|''J
4
4
6
4
4
5
7
8
78
Valid*
Samples
30
12
9
45
24
14
26
•-~i-:' A:R?.. ••' -
' ':- ;::w*$Vx^ti: .„:-::.-
24
19
30
20
22
29
102
130
552
* Valid = Validation
** Shaded sites supported by this IAG.
18
-------
but it is approximately 100 to 300 mg/kg as TRPH by
EPA Method 418.1. Data from the 16 sites were
normalized to permit compilation of cumulative plots.
Figure 5 presents the cumulative scatter plot of in situ
LIF versus TRPH with fluorescence threshold and
detection threshold lines plotted. The plot shows the
largest number of points in the Non-Detect quadrant. A
slight trend of increased fluorescence with increased
TRPH is apparent. Figure 6 presents the cumulative
scatter plot of in situ LIF versus TPH and shows features
similar to Figure 5. Figure 7 presents the cumulative
scatter plot of TPH versus TRPH. As expected, TPH vs.
TRPH shows better correlation than the LIF plots. A
significant number of false positive and false negative
points are present. Table 4 contains the cumulative
contingency analysis results on a percentage basis
showing better than 85 percent agreement between LIF
and analytical measurements. As expected, comparison
between analytical methods is slightly better since these
measurements were made on splits of the same sample.
Naval Air Station, Alameda Results
SCAPS field operations were undertaken at Naval Air
Station (NAS), Alameda, CA, Site 13 - Former Oil
Refinery, from the 17th of March through the 6th of April
1994. Validation operations were performed from the 4th
through the 6th of April 1994. A total of 37 pushes were
performed during phase one SCAPS investigation. After
review of phase one data, eight validation pushes were
performed each directly followed by a hollow stem auger
boring and sample collection. Forty-five samples were
collected from the eight validation borings.
Table 5 presents the laboratory results for the 45
validation samples collected at NAS, Alameda. The LIF
Table 4. Sixteen Site Cumulative Contingency Analysis Results Summary
(n=552)
Comparison
LIF vs. TRPH
LIF vs. TPH
TPH vs. TRPH
% Correct
87
86
95
% False Positive
5
7
2
% False Negative
7
7
3
le+03
le-02
Normalized TRPH
Figure 5. Cumulative scatter/contingency plot of LIF vs. TRPH (n = 552).
19
-------
lo-HB •
le-KU —
le-HW •
le-01 -
le-02 •
Y
0
N
C
M
G
F
X
I
A
T
P
D
B
R
H
Yuma
UST
NIFS
NABC
MCBCP
GOF
FFTF
NEX
IBFF
ALA
HS
PRE
MCRD
MRB
NRRF
PHD
False
Positive
H
Non-Detect
Detect
T
fT T G
T p
N
Detection
Threshold
False
Negative
T
T
le-03
le-02
le-01 le-WO le-KIl le-M)2 le-K)3 le+04
Normalized TPH
Figure 6. Cumulative scatter/contingency plot of LIF vs. TPH (n = 552).
1«4£J4
W
10400
le-01 -
le-02 -
lc-03
le-04 le-03 le-02 le-01 le«0 le-tOl le-K>2 le-H)3
Normalized TRPH
Figure 7. Cumulative scatter/contingency plot of TPH vs. TRPH (n = 552).
20
-------
Table 5. Summary of SCAPS Validation Sampling at NAS, Alameda
NRaD SCAPS
Sample #
ALA13B38-1
ALA13B38-2
ALA13B38-3
ALA13B38-4
ALA13B39-1
ALA13B39-2
ALA13B39-3
ALA13B39-4
ALA13B39-5
ALA13B40-1
ALA13B40-2
ALA13B40-3
ALA13B40-4
ALA13B40-5
ALA13B41-1
ALA13B41-2
ALA13B41-3
ALA13B41-4
ALA13B41-5
ALA13B41-6-
ALA13B41-7
ALA13B41-8
ALA13B42-1
ALA13B42-2
ALA13B42-3
ALA13B42-4
ALA13B42-5
ALA13B42-6
ALA13B42-7
ALA13B42-8
ALA13B43-1
ALA13B43-2
ALA13B43-3
ALA13B43-4
ALA13B44-1
ALA13B44-2
ALA13B44-3
ALA13B44-4
ALA13B44-5
ALA13B44-6
ALA13B45-1
ALA13B45-2
ALA13B45-3
ALA13B45-4
ALA13B45-5
Depth
Interval
1-1.5'
5-5.5'
10-10.5'
15-15.5'
3-3.5'
8-8.5'
11.5-12'
14-14.5'
16-16.5'
0.5-1 '
5-5.5'
10-10.5'
13.5-14'
16.5-17'
0.5-1 '
4.5-5'
7-7.51
8-8.5'
8.5-9'
9-9.5'
11-11.5'
15-15.5'
3-3.5'
5.5-6'
6.5-7'
7.5-8'
9.5-10'
12.5-13'
16-16.5'
19.5-20'
4-4.5'
9.5-10'
12-12.5'
15-15.5'
2-2.5'
3.5-4'
6-6.5'
8.5-9'
11.5-12'
13.5-14'
6.5-7'
7.5-8'
11-11.5'
14.14.5'
17-17.5'
TRPH
cone.
310
17
2
ND
7
4
58
69
ND
520
ND
3
200
ND
1700
ND
170000
24000
410
38000
2
ND
ND
26000
590
2100
2600
9300
110
5
ND
310
1100
ND
47
3700
170
1500
ND
ND
13000
550
9300
6500
ND
TPH
cone.
30
ND
ND
ND
ND
ND
23
15
ND
ND
ND
ND
19
ND
270
ND
31000
1900
220
12000
ND
ND
ND
4500
490
620
130
1300
30
12
ND
430
430
ND
49
2500
200
2100
ND
ND
8400
440
6000
530
ND
in situ LIF
Avg. of
Interval
5466.50
5621 .33
5425.33
5259.67
18354.00
5822.67
7530.33
8954.33
5773.67
1363.33
620.33
650.00
938.50
617.00
12918.67
6052.00
246101.00
92273.57
70992.33
86169.00
6123.67
5459.00
6513.00
2032.32
84627.33
174340.00
14782.50
79970.00
18116.33
8963.00
5817.00
9205.67
66423.67
5242.00
6872.50
23488.00
11047.50
44022.50
6060.00
5824.67
256909.50
130729.00
87220.00
10626.00
5791.67
LIF
SPT-Top
6715
9733
4598
4581
6517
6765
6864
6189
4323
5802
10142
7877
5888
3852
16319
5997
245135
33079
33079
33079
4886
5358
4839
201210
34258
19494
8708
104225
76614
51673
8370
13379
10018
4691
6562
20742
14446
56711
. 10890
7574
135365
52391
130983
177593
5337
LIF
SPT-
Bottom
5490
6082
5251
4590
6376
4907
5353
6300
3899
18474
5775
8311
7287
4265
12026
7387
73543
109909
109909
109909
4673
4697
4952
96928
124222
11013
18032
48297
15592
9033
6867
249702
13131
5456
8339
79767
13068
9608
9596
8454
219217
111215
44260.5
8876
5551
LIF
SPT-
Homog.
6139
5697
5730
4393
6754
6853
6320
6016
4016
7378
4634
8660
5871
4478
11552
6923
256801
56416
56416
56416
5109
5198
4700
179930
34160
20197
20891
196958
29175
28229
7684
143510
32377
4632
5710
44981
12433
28762
9147
7660
189639
54133
95088
124761
5203
Concentrations in mg/kg.
ND = not detected, TPH<12, TRPH<1.
Fluorescence reported as detector counts.
21
-------
SPT-Top, -Bottom, and -Homog. columns contain results
of LIF POL sensor benchtop measurements of scrapings
from the top and bottom of the split spoon soil sample
and the homogenized sample. The data in this table is
used for direct comparison of the LIF measurements and
the laboratory methods.
The calculated fluorescence threshold was 10,620
relative fluorescent counts, with a corresponding
detection threshold of 137 ppm. Figures 8-10 present
scatter plots annotated with contingency threshold lines
for LIF versus TRPH, LIF versus TPH, and TPH versus
TRPH. The plots show a general trend of increased
fluorescent intensity with increased TRPH and TPH.
Contingency analysis was performed on the data as
discussed under the methods section. The percentage
of LIF false negatives was 7 percent versus TRPH and 4
percent versus TPH. The percentage comparability or
percentage correct of LIF versus TRPH was 91 percent
and versus TPH was 87 percent. These results are very
favorable considering comparison between the
laboratory analytical methods with 11 percent false
negative and 87 percent comparability. Table 6
summarizes these contingency results. A data report for
the NAS, Alameda site, which includes all LIF POL
sensor data and laboratory results from confirmation
sampling, is available from Dr. Stephen Lieberman,
NCCOSC RDT&E Division Code 521.
Guadalupe Oil Field Results
The LIF POL sensor was employed at the Guadalupe Oil
Field in San Luis Obispo County, CA, for subsurface
investigation of diluent contaminated soils from the 23rd
of August through the 8th of September 1994. In
accordance with the NRaD Work Plan for Guadalupe Oil
Field, validation pushes with overborings and sampling
were performed on the 7th and 8th of September 1994.
The diluent was a kerosene-like hydrocarbon pumped
into the oil producing formations to enhance mobility and
recovery of the viscous crude oil. Both shallow (<6 m
(20 ft)) and deep (to maximum push depth and below)
contamination was expected due to surface pipeline and
tank leakage and subsurface pipeline leakage. A total of
32 SCAPS push holes were advanced in the initial
investigation. Four validation holes were pushed,
subsequently overbored, and a total of 23 soil samples
collected. LIF POL sensor benchtop measurements
were performed on scrapings from the top, middle, and
bottom of 15-cm (6-in) soil sections submitted for
analysis (two 7.5-cm (3-in) samplers). Soil samples
were sent to an analytical laboratory for analysis by EPA
Methods 418.1 (TRPH by IR) and 8015-Modified (TPH
by GC/FID).
Table 7 is a compilation of corresponding measurements
for each of the 23 samples collected at Guadalupe Oil
Field. Samples B35-3 through B35-7, B36-5 and B36-6
were collected at depths greater than the bottom of the
le+02
t
le-HJO
o
55
le-01
False Positive
Non-Detect
Detect
AAA
AA
1 A
A A
AA
Fluorescence
$&.
Detection
Threshold
Threshold
False Negative
I ' I ' I 'I 'I 'I
le-02 le-01 le+00 le-HH le+02 le-HJS
Normalized TRPH
Figure 8. NAS Alameda scatter/contingency plot of LIF vs. TRPH (n = 45).
22
-------
le+02-
_
le+01 -
1 p+fin
-
1 f m
False Positive
A
A A
/ancXr"*
Wi •"•
Detection
T,T T-, ^ A Threshold
Non-Detect
ie-UJ. ! , !
Detect
'A
A
AA
A A
Fluorescence
A Threshold
A
False Negative
1 1 '1 '1
le-02 le-01 le+00 le+01 le+02 le+03
Normalized TPH
Figure 9. MAS Alameda scatter/contingency plot of LIF vs. TPH (n = 45).
le+03 -
le+02 -
£ le+01 -
13
° le+OQ
~
le-01 -
-
False Positive
A
A
A
f\
A AAAAA A A
Non-Detect
1 i | i
Detect A
A A
A A A A
A A
^ A Detection
A Threshold
A
A
A
^1
A
Detection
Threshold
False Negative
1 1 '1 '1
le-02 le-01 le+00 le+01 le-KJ2 le+03
Normalized TRPH
Figure 10. NAS Alameda scatter/contingency plot of TPH vs. TRPH (n = 45).
23
-------
Table 6. NAS, Alameda Contingency Analysis Results Summary (n = 45)
Comparison
LIF vs. TRPH
LIF vs. TPH
TPH vs. TRPH
% Correct
91
87
87
% False Positive
2
9
2
% False Negative
7
4
11
Table 7. Summary of SCAPS Validation Sampling at Guadalupe Oil Field
Sample #
GOF
B33-1
B33-2
B33-3
B33-4
B33-5
B33-6
B34-1
B34-2
B34-3
B34-4
B35-1
B35-2
B35-3
B35-4
B35-5
B35-6
B35-7
B36-1
B36-2
B36-3
B36-4
B36-5
B36-6
Depth
Interval
0.5-1 '
5.5-6'
8.75-9.25'
9.25-9.75'
10.25-10.75'
11.25-11.75'
5-5.5'
8.5-9'
9-9.5'
10.5-11'
20.5-21'
41.5-42'
48-48.5'
50.5-51'
51-51.5'
53.5-54'
54-54.5'
59-59.5'
72.5-73'
75-75.5'
79-79.5'
89.5-90'
90-90.5'
in situ LIF
Avg. of
Interval
600
262
113841
265463
40464
1392
183
120
199
199
129
177
N\A
N\A
N\A
N\A
N\A
174
54051
104245
91678
N\A
N\A
LIF
SPT-
Top
493
266
511
7290
260772
8268
392
278
302
153
294
465
69382
17481
58522
20443
4694
188
48241
N\A
48829
89168
1105S
LIF
SPT-
Middle
523
317
295
36995
221877
7101
370
161
237
165
486
288
76136
32363
86535
25425
53118
196
20853
327522
81933
157111
27622
LIF
SPT-
Bottom
385
369
1450
240127
82004
2433
324
290
227
172
203
361
120454
60964
103388
4523
60392
201
114602
315021
90324
10960
5497
LIF
SPT-Avg.
467
317
752
94804
188218
5934
362
243
255
163
328
371
88657
36936
82815
16797
39401
195
61232
321271
73695
85746
14726
TRPH
Cone.
_ 24
1
8
40000
62000
800
1
1
1
1
84
41
8800
8400
10000
9800
10000
17
9300
76000
9500
7200
3400
TPH
Cone.
310
5
5
45000
54000
310
5
5
5
5
110
44
8800
12000
18000
11000
12000
17
10000
67000
7200
8600
1900
N\A = Not applicable, measurement not performed.
Avg. = Average
Cone. = Concentration
24
-------
LIF POL sensor push hole. Thus there is an N/A for
these samples in the Fluorescence Avg. of Interval
column and they were not used in the validation. The
values presented in the Fluorescence Avg. of Interval
column are calculated from the in situ fluorescence
values corresponding to the 15-cm (6-in) sampling
interval (typically three points). The SPT-Top, -Middle,
and -Bottom are LIF POL sensor benchtop
measurements of scrapings from the soil samples. The
data in this table is used for direct comparison of the LIF
measurements and the laboratory methods.
The calculated fluorescence threshold was 350 relative
fluorescent counts, with a corresponding detection
threshold of 77 ppm, Figures 11-13 present the scatter
plots annotated with contingency threshold lines for LIF
versus TRPH, LIF versus TPH, and TPH versus TRPH.
The plots show a general trend of increased fluorescent
intensity with increased TRPH and TPH. Contingency
analysis was performed on the data as discussed under
the methods section. There were no LIF false negatives
versus TRPH and 6 percent versus TPH. The
percentage comparability or percentage correct of LIF
versus both TRPH and versus TPH was 88 percent.
These results are very favorable considering 88 percent
comparability between the two laboratory analytical
methods. Table 8 summarizes these contingency
results. A data report for the Guadalupe Oil Field site,
which includes all LIF POL sensor data and laboratory
results from confirmation sampling, is available from Dr.
Lieberman, NCCOSC RDT&E Division Code 361.
le-HB-
le+02 -
le-KJl -
n.
TS
le-HJO
le-01 -
False Positive
G
G
Non-Detect
G G
G
G
G
G
Detect
G
G
G
Fluorescence
Detection
Threshold
Threshold
False Negative
I •—I r
le-02 le-01 le-H)0 le-HH le-H)2 le-K)3
Normalized TRPH
Figure 11. Guadalupe scatter/contingency plot of LIF vs. TRPH (n = 16).
Table 8. Guadalupe Oil Field Contingency Analysis Results Summary
Comparison
LIF vs. TRPH
LIF vs. TPH
TPH vs. TRPH
% Correct
88
88
88
% False Positive
12
6
12
0
6
0
25
-------
Ie403-
t
g le-H)2 -
g le-KJl -|
le-WO
o
a
le-01 -
-
-
-
—
False Positive
G
Non-Detect
1 '1
Detect
G
G G
G
G
^ Fluorescence
Threshold
' Detection
Threshold ...
False Negative
'1 '1 '1
le-02 le-01 le-K)0 le-K)l le-K12 le-K)3
Normalized TPH
Figure 12. Guadalupe scatter/contingency plot of LIF vs. TPH (n = 16).
ffi
PI
o
55
le-HB-
le+02-
le-HDl -
-
le-01 -
1 - no
False Positive
G
G
G
G G
Non-Detect
Detect
3
G
Detection
Threshold
Detection
Threshold _ . „
False Negative
'I 'I 'I
le-02 le-01
le-K)2 le-HD3
Normalized TRPH
Figure 13. Guadalupe scatter/contingency plot of TPH vs. TRPH (n = 16).
26
-------
References
Andrews, J. M., and S. H. Lieberman. A Neural Network
Approach to Qualitative Identification of Fuels and Oils
from Laser Induced Fluorescence Spectra. Anal.
Chim. Acta, 285: 237-246, 1994.
Apitz, S. E., G. A. Theriault, and S. H. Lieberman.
. "Optimization of the Optical Characteristics of a Fiber-
Optic Guided Laser Fluorescence Technique for the in-
situ Evaluation of Fuels in Soils." In: Proceedings of
the OE/LASE '92 International Conference on
Environmental Process and Treatment Technologies,
Los Angeles, CA, 1992a.
Apitz, S. E., L. M. Borbridge, S. H. Lieberman, and G. A.
Theriault. Remote In-Situ Determination of Fuel
Products in Soils: Field Results and Laboratory
Investigations. Analusis, 20: 461-474, 1992b.
Apitz, S. E., L. M. Borbridge, K. Bracchi, and S. H.
Lieberman. "The Fluorescent Response of Fuels in
Soils: Insights Into Fuel-Soil Interaction." In:
Proceedings of the International Conference on
Monitoring Toxic Chemicals and Biomarkers, Berlin,
Germany, 1992c.
American Society for Testing and Materials Standard D
3441-94, Standard Test Method for Deep, Quasi-
Static, Cone, and Friction-Cone Penetration Tests of
Soil. 1994.
Davey, M., J. M. Andrews, S. H. Lieberman, and K. D.
Wu. "DOD Tri-Service Test of Laser Induced
Fluorescence Sensors for SCAPS." NRaD TD. San
Diego, CA. U.S. Navy, 1994a.
Davey, M., L. M. Borbridge, S. H. Lieberman, and K. D.
Wu. The Effect of Humic Material on Fluorescence.
NRaD Unpublished report. San Diego, CA. U.S.
Navy, 1994b.
Douglas, G. S., S. H. Lieberman, W. C. McGinnis, D. S.
Knowles, and C. Peven. "The Influence of PAH
Concentration and Distribution on Real-Time In-Situ
Measurements of Petroleum Products in Soils Using
Laser Induced Fluorescence." In: Proceedings of the
Fourth International Symposium - Field Screening
Methods for Hazardous Wastes and Toxic Chemicals
Las Vegas, NV, 1995.
Douglas, G. S., K. J. McCarthy, D. T. Dahlen, J. A.
Seavey, W. G. Steinhauer, R. C. Prince, and D. L.
Elmendorf. The Use of Hydrocarbon Analysis for
Environmental Assessment and Remediation. Journal
of Soil Contamination, 1 (3): 197-216(1992).
Lieberman, S. H., W. C. McGinnis, P. Stang, and D.
McHugh. "Intercomparison of In-situ Measurements of
Petroleum Hydrocarbons Using a Cone Penetrometer
Deployed Laser-Induced Fluorescence (LIF) Sensor
with Conventional Laboratory-Based Measurements."
In Proceedings of the Fourth International Symposium
- Field Screening Methods for Hazardous Wastes and
Toxic Chemicals, Las Vegas, NV, 1995.
Lieberman, S. H., and S. E. Apitz. "Real-Time In-situ
Measurements of Fuels in Soil: Comparison of
Fluorescence and Soil Gas Measurements." In:
Proceedings of the Third International Symposium -
Field Screening Methods for Hazardous Wastes and
Toxic Chemicals, Las Vegas, NV, 1993.
Lieberman, S. H., S. E. Apitz, L. M. Borbridge, and G. A.
Theriault. "Subsurface Screening of Petroleum
Hydrocarbons in Soils via Laser Induced Fluorometry
Over Optical Fibers With a Cone Penetrometer
System." In: Proceedings of the International
Conference on. Monitoring Toxic Chemicals and
Biomarkers, Berlin ..Germany, 1992.
McGinnis, W. C., M. Davey, K. D. Wu, and S. H.
Lieberman. "Capabilities and Limitations of a Cone
Penetrometer Deployed Fiber Optic Laser Induced
Fluorescence (LIF) Petroleum, Oil, and Lubricant
(POL) Sensor." In: Proceedings of the International
Symposium on Optical Sensing for Environmental and
Process Monitoring, McLean, VA, 1994.
Robertson, P. K., and R. G. Campanella. "Guidelines for
Geotechnical Design Using CPT and CPTU," Soil
Mechanics Series No. 120. Department of Civil
Engineering, University of British Columbia, 1989.
27
-------
U. S. EPA, The Site Characterization and Analysis
Penetrometer System (SCAPS) Laser-Induced
Fluorescence Sensor and Support System Innovative
Technology Evaluation Report." EPA/540/R-95/520,
1995
28
•&U.S. GOVERNMENT PRINTING OfrFICE: i997 ^50-001/80166
-------
-------
m
I
o
o
1
-vl
01
-? m
O c
-> en
•05'
3- CD
< co
S- w
CD
cr
CO
CD
o •&
B'S
CD
O m c
CD 9 3
'
^ CD 3 a.
B =J- n C/3
ill Is
||| s
OO 3 CD o
3 cr.
CD
I
T)
O
•o
m
30
O
&
------- |