FIFTH ANNUAL
WASTE TESTING
AND
QUALITY ASSURANCE
SYMPOSIUM
July 24-28, 1989
OMNI SHOREHAM HOTEL
WASHINGTON, D.C.
PROCEEDINGS
VOLUME I
-------
FIFTH ANNUAL
WASTE TESTING
AND
QUALITY ASSURANCE
SYMPOSIUM
July 24-28, 1989
OMNI SHOREHAM HOTEL
WASHINGTON, D.C.
PROCEEDINGS
Symposium Managed by American Chemical Society
-------
TABLE OF CONTENTS
Volume I
PaPer Page
Number* Number
AIR AND GROUND WATER
1 Influence of Well Casing Materials on Chemical Species in Ground Water. K.T. 1-1
Lang, M.H. Stutz, L.V. Parker, A.D. Hewitt, T.F Jenkins
2 The Efficacy of Indicator Parameters to Detecting Incidents of Ground Water I-2
Contamination. P.H. Friedman
3 An Intel-laboratory Study of Volatile Organic Compounds in Ground Water by I-3
Capillary Column GC/MS. R.W. Slater, K.W. Edgell, R.J. Wesselman
4 The Determination of Total Hydrocarbons, Methane, Carbon Dioxide, Oxygen and I-4
Nitrogen with an Automated Gas Chromatographic System. N. Kirshen, E.
Almasi
5 A Case Study of the Use of the "Summa Canister" for Passive "Off Gas" Vent I-8
Sampling and Analysis. H. Syvarth, P. Campagna, M. Solecki, W. Batz
6 Development of a Highly Reliable Field Deployable Analyzer for VOC. E.B. 1-19
Overton, R.W. Sherman, T.H. Backhouse, C.B. Henry, E.G. Collard, C.F. Steele,
B.S. Shane, T.R. Lrvin
76 An Evaluation and Comparison of the Photovac TIP H and HNU PI 101 Total 1-20
Organic Vapor Analyzers. L. Accra, A. Hafferty
77 The Automated Determination of Volatile Organic Contaminants in Ambient Air 1-36
and/or Soil Gas by Gas Chromatography with Selective Detectors. N. Kirshen ,
E. Almasi
78 The Determination of Fixed Gases and Non-Methane Organic Compounds in 1-59
Landfill Gas or Air. N. Kirshen, E. Almasi
*Refer to Final Program and Abstract Book.
-------
BIOLOGICAL TEST METHODS
7 Effect of Chemicals on Soil Nitrifying Populations Using a Continuous-Flow 1-65
Culture Technique. C.W. Hendricks, A.N. Rhodes
8 Toxicity Evaluations for Hazardous Waste Sites: An Ecological Assessment 1-77
Perspective. G. Linder, M. Bollman, W. Baune, K. Dewhitt, J. Miller, J. Nwosu,
S. Smith, D. Wilborn, C. Bartels
9 Application of Microbial Toxicity and Mutagenicity Assays for the Identification 1-94
and Evaluation of Toxic Constituents in Fractionated Hazardous Wastes. B.S.
Shane, K.C. Donnely, E.B. Overton, T.R. Irvin, L. Butler, J. Norcerino, J. Petty
10 Application of Mammalian Cell Culture Systems to Evaluate and Monitor I-95
Hazardous Wastes and Waste Sites. T.R. Irvin, J.E. Martin, B.S. Shane, L.
Butler, N. Norceringo, J. Petty, E.B. Overton
11 Screening for PCDD and PCDF by Immunoassay. M. Vanderloan, L. Stanker, I-96
B. Watkins
92 The Use of Screening Protocols to Evaluate Bioremediation Technology for Site |-97
Cleanup. J.A. Glaser
ENFORCEMENT
12 Computers in the Decision Process: Legal Implications of Electronic Data |-99
Management Systems. J.C. Worthington, R.P. Haney
13 A Planning Tool for Site Managers: Historical Perspective on Litigation Uses of 1-100
Sample Data for EPA CERCLA/SARA Cases. C. Miller, J. Worthington
14 Examples of the Use of an Advanced Mass Spectrometric Data Processing 1-101
Environment for the Determination of Sources of Waste. B.M. Hughes, D.E.
McKenzie, C.K. Trang, L.R. Minor
15 California's Proposition 65: A Voter Approved Environmental Law. P. Marsden 1-115
16 Enforcement of RCRA at Radioactive Mixed Waste Facilities. M.S. Barger 1-130
17 Hazardous Ground Water Task Force Data Base/Implementation of Field QA/QC. 1-131
T. LaCosta, K. Jennings
-------
INORGANICS
18 Validation of a Method for Determining Elements in Solid Waste by Microwave 1-133
Digestion. D.A. Binstock, P.M. Grohse, A. Gaskill, Jr.
19 Microwave Digestion for ICP Analysis: Region V Alternate Test Procedures. M. 1-146
Shannon, G. Payton, P. Howard
20 A Comparison Study of Quality Control Performance Between ICP/MS Method 1-150
6020 and the ICP-AES and GFAA Spectroscopy Methods. K.A. Aleckson, F.C.
Garner, L.C. Butler, M.L. Hurd
21 Performance of ICP-MS Method 6020. T.A. Hinners, E.M. Heithmar, L.C. 1-151
Butler, M.L. Hurd, D.E. Dobb, G.A. Laing
22 ICP-MS Method 200.8. The Determination of Trace Elements in Waters and 1-161
Wastes.S. Long, T.D. Martin
23 Selected Comparisons of Low Concentration Measurement Compatibility Estimates 1-165
in Trace Analyses: Method Detection Limit and Certified Reporting Limit. K.T.
Lang, M.H. Stutz, C.L. Grant, A.D. Hewitt, T.F. Jenkins
93 Report of an Interlaboratory Study Comparing EPA SW-846 Method 3050 and an 1-166
Alternative Method from the California Department of Health Services. D.E.
Kimbrough, J. Wakakuwa
94 A Performance Evaluation of the Inorganic Methods Used in the Contract 1-180
Laboratory Program. K.A. Aleckson, Y.J. Lee, E.J. Kantor
95 Studies of Intelligent Automation for Water Analysis by ICP-AES with CLP 1-181
Protocol. S.F. Zhu, A.K. Merrick, FA. Glodas
LABORATORY INFORMATION MANAGEMENT
24 Customized LIMS Data Treatment Through Interaction with a User-Definable 1-183
Spreadsheet. R.D. Beaty
25 Early Warning Report: Automated Checking of QC Data. R. Peak, P. Duerksen, 1-195
K. Wong, E. Szeto
26 A Quality Assurance and Management System for Large Environmental Projects. 1-210
J. Karmazyn, C. Schrenkel, S. Nordstrom, R.F. Weston
27 A Smart Data Base System for Selecting Analytical Methods for Environmental 1-217
Analysis. R.A. Olivero, M.T. Homsher, J.L. Boyd, D.W. Bottrell
96 Interfacing of an HP GC-MS (5970) with a 1000A Computer System to a VAX |_227
Computer and DEC LIMS. J.T. Bychowski, D. Couch, D. Hockman, A.
O'Donnell, S. Srivastava, M. Demorotski, M. Hartwig, M. Rank
-------
MOBILITY METHODS
28 Migration of Chlorinated Pheno, Dibenzo-P-Dioxins, and Dibenzofurans in Soils
Contaminated with Wood-Treatment Oils. D.R. Jackson, D.L. Bisson, DA.
Stewart
29 Leach Testing of Stabilized Contaminated Soils. B.J. Mason, J.J. Barich, G.L. I-232
Rupp, K.W. Brown
30 Geochemical Basis for Predicting Leaching of Inorganic Constituents from Coal- I-246
Combustion Residues. I.P. Murarka, D Rai, C.C. Ainsworth
31 Evaluation of Methods 1311 for Determining the Release Potential of Oily Wastes 1-257
(part I). R.S. Truesdale, J.J. Pierce, G.A. Hansen
32 Evaluation of Methods 1311 for Determining the Release Potential of Oily Wastes 1-257
(part n). R.S. Truesdale, J.J. Pierce, G.A. Hansen
33 The Liquid Release Test (LRT). G. Kingsbury, P Hoffman, B. Lesnik 1-258
97 Residual Fuel Oil as Potential Source of Groundwater Contamination. B. Davani, 1-259
B. Sanders, G. Jungclaus
98 Leachability of Chemicals from Hazardous Waste Land Treatment Site Soils. D. 1-274
Erickson, L. Rogers, R.C. Loehr
99 Evaluation of Leachability of Radium Contaminated Soils. T.F. McNevin 1-283
100 Interlaboratory Comparison of Methods 1310, 1311, and 1312 for Lead in Soil. A. 1-298
Gaskill, Jr., G.A. Hansen
101 A Comparison of the TCLP and a Modified TCLP in an Evaluation of Stabilised I-299
Oil Sludge. K.B. Toner, E.D. Keithan, S. Pancoski
102 TCLP Extraction of Reference Waste Samples Stored Over Time. S.S. Sorini I-308
103 Modification of the TCLP Procedure to Accommodate Monolithic Wastes. L. I-323
Bone, M. Bricka, P Hannak, S.I. Shah, N. Prange, P.J. Marsden, J.E. Waggener,
M. Miller, E. Johnson, S. J. Robuck
104 Precision and Ruggedness Evaluation of Method 1312. D. Miller, P Marsden 1-336
105 The Pacific Basin Consortium for Hazardous Waste Research Hazardous Materials I-337
Leachate Database. E.A. Burns, L.E. Michalec, G.A. Hansen
VI
-------
AJDR, AND GROUND WATER
-------
INFLUENCE OF WELL CASING MATERIALS ON CHEMICAL
SPECIES IN GROUND WATER
LANG, KENNETH T. AND SIUTZ, MARTIN H., U.S. ARMY TOXIC AND HAZARDOUS MATERIALS
AGENCY, ABERDEEN PROVING GROUND, MARYLAND 21010-5401; PARKER, LOUISE V., HEWITT,
AIAN D., JENKINS, THOMAS F., U.S. COLD REGIONS RESEARCH AND ENGINEERING
IABORATORY, HANOVER, NEW HAMPSHIRE 03755-1290
ABSTRACT. In this study four well casing materials were examined:
polyvinyl chloride (PVC) , polytetrafluoroethylene (PTFE) , stainless steel 304 (SS
304) and stainless steel 316 (SS 316) to determine their suitability for monitoring
both inorganic and organic constituents in ground water. Analyte solutions exposed
to the well casing materials were compared to controls that consisted of an
identical solution and container, but were devoid of the well casing material. For
the inorganic studies, solutions containing two concentrations of As, Cr, Pb, and
Cd, at two pHs, with and without added organic carbon, were tested. Samples were
taken after 0.5, 4, 8, 24, and 72 hours of exposure. Results showed that PTFE well
casings had no significant effect onxthe concentration of any of the aqueous metals
monitored. Both stainless steels were susceptible to surface oxidation in ground
water solutions. Rusting of the metal casings appears to create both active sites
for sorption and a mechanism for the release of impurities and major constituents.
The sporadic occurrence of surface oxidation makes the solutions prone to randan
error. PVC showed release of Cd and sorption of Pb after 72 hours of exposure.
The magnitude of these effects is most likely not a major concern for purging tines
of less than 24 hours. Overall PTFE was the best-suited material for monitoring
these trace inorganic species in ground water, followed by PVC, SS 304, and SS
316. The well casings were also tested for sorption of the following organic
substances: RDX, trinitrobenzene (TNB), cis-and trans-l,2-dichloroethylene (c-DCE,
t-DCE) , m-nitrotoluene (m-NT), trichloroethylene (TCE) , chlorobenzene (CB) , and 0-,
p-, and m-dichlorobenzene (0-DCB, p-DCB, and m-DCB). The two sets of isomers were
selected to examine the effect of the structure on sorption. Samples were taken
after 0 hours, 1 hour, 8 hours, 24 hours, 72 hours, 7 days, and approximately 6
weeks. Even after 6 weeks there was no loss by sorption for either stainless
steel, although they did show signs of rusting. The greatest losses were due to
PTFE with the chlorinated organics. While there was some slight sorption by PVC,
after 72 hours the concentrations of analytes were still not significantly
different from those of the glass controls. The results of these studies indicate
that the choice of a well casing material for ground water monitoring of both
inorganic and organic constituents is a compromise, but that PVC is probably the
best overall choice.
1-1
-------
THE EFFICACY OF INDICATOR PARAMETERS IN DETECTING
INCIDENTS OF GROUND WATER CONTAMINATION
PAUL H. FRIEDMAN, TECHNICAL DIRECTOR, BCM ENGINEERS, INC. 1850 GRAVERS ROAD,
NORRISTOWN, PA 19401
ABSTRACT: The detection of ground water contamination requires the application
of extensive and expensive testing. While, a numer of indicator tests are
available to detect the presence or absence of chemcial groups, compounds or
substances in ground water, there has been no systematic comparison of the
application of indicator analytes or parameters to the detection and
quantification of analytes on the Hazardous Substance List and other compunds as
identified as a result of automated library searches.
Indicator parameter analyses and hazardous substance compound analyses for a
group of 58 waste management sites are examined. The correlation between the
results of indicator parameter analyses and the detection of specific hazardous
substances at hazardous waste sites is assessed.
The efficacy of indicator parameters as predicators for specific classes of
compounds such as organic volatile compounds, phenolics and halogenated species
is also studied.
1-2
-------
AN INTERLABORATORY STUDY OF VOLATILE ORGANIC COMPOUNDS
IN GROUND WATER BY CAPILLARY COLUMN GC/MS
KENNETH W. EDGELL, BIONETICS CORPORATION, CINCINNATI, OHIO; ROBERT W. SLATER
AND RAYMOND 0. WESSELMAN, ENVIRONMENTAL MONITORING SYSTEMS LABORATORY, U.S.
ENVIRONMENTAL PROTECTION AGENCY, CINCINNATI, OHIO
ABSTRACT
The Environmental Monitoring Systems Laboratory - Cincinnati, (EMSL-
Cincinnati) develops analytical methods and provides quality assurance (QA)
support for U.S. Environmental Protection Agency (USEPA) programs involving
water regulations. One of these QA support activities is to conduct inter-
laboratory method validation studies to evaluate analytical methology selected
for the Agency's operating programs. These studies establish the reliability and
legal defensibility of the data collected by the Agency, state regulatory
authorities and commercial laboratories performing compliance analyses.
EMSL-Cincinnati has completed a method validation study (MVS) for Method
524.2, "Volatile Organic Compounds in Water by Purge and Trap Capillary Column
Gas Chromatography/Mass Spectrometry" using reagent water and well water at a
Superfund hazardous waste site as the relevant matrices. Nine laboratories were
selected to participate in the study based upon laboratory experience, quality
control practices, and satisfactory completion of a performance evaluation
sample.
Analysts from the participating laboratories performed analyses of reagent
water and ground water which were spiked with known concentrations of 60
volatile analytes. Six concentration levels, as three Youden pairs were
examined for each matrix. After elimination of outliers, approximately 5500
data points were used to develop regression equations for the recovery, overall
precision and single laboratory precision estimates for each of the sixty
analytes. Also, any matrix effects between the water types were identified.
Percent recoveries for all compounds indicate a general high bias for the
method. The pooled mean recovery for 59 compounds in reagent water, excluding
dichlorodifluoromethane, was 111% with a range of 99 to 133% for the individual
components; whereas, the recovery in groundwater samples was 110% with a range of
89% to 130%. Several of the gases exhibited quite high recoveries for the study.
Overall precision expressed as percent relative standard deviation (%RSD)
for reagent water samples was 15.4% for all compounds in a range of 6.8% to
40%. Only 6 compounds had RSD of greater than 20%. For groundwater samples,
the pooled %RSD was 16.4%, range 6.3% to 34.7%; indicating no major differences
due to the matrix.
Single analyst precision for the pooled data in reagent water was 9.4% RSD
with a range of 3.2% to 25.7%. Seven of the analytes exhibited RSD greater than
15% in reagent water. In groundwater, the pooled %RSD was 10.9% over a range of
5.8 to 32.3%.
Statistically significant matrix effects were identified for 14 of the
volatile analytes; however, these effects were considered of practical
significance for only 8 cases.
1-3
-------
THE DETERMINATION OF FIXED GASES AND TOTAL HYDROCARBONS IN
SOIL GAS AND AIR
Norman Kirshen, Senior Chemist, and Elizabeth Almasi, Chemist,
Varian Instrument Group, 2700 Mitchell Drive, Walnut Creek,
California 94598
The growing number of emission sources in the industrial,
municipal, and transportation areas makes source gas and
ambient air monitoring necessary. The safety of existing
landfill sites requires the monitoring of soil gas and/or
ambient air both inside and outside the perimeter of the
disposal area. The presence and amounts of 02 and CO2 can
provide important information about aerobic and anaerobic
reactions inside the landfill. The monitoring of methane to
follow its migration can eliminate explosion hazards by
preventing its collection in airpockets in residences or in
other populated areas.
To attain the requirements of the Clean Air Act, total
hydrocarbon monitoring is necessary since many organics behave
as precursors in ozone formation.
The system described here is intended for measuring CO2 N2,
CH4 and total hydrocarbons (THCs) in aerial matrices. ' (See
Figure 1).
The two-loop gas sample valve delivers a one milliliter sample
directly to the FID for the detection of the THC's and one
milliliter to a three column set for the separation of the
fixed gases, the heavies being backflushed to vent.
The FID is capable of detecting THCs and CH4 from ppm to low
percentage levels while the TCD can detect CO2, O2, N2, and
CH4 from low ppm to high percentage levels.
A propane standard is used as a calibration compound for the
quantitation of the THC response as ppm Carbon providing
excellent results for all hydrocarbons. Substituted organic
compounds may give a lower Carbon response resulting in
somewhat low biased results. In most cases this bias can be
neglected. -1
Figure 2 shows the chromatogram of a standard gas mixture.
This system exhibits excellent reproducibility (% standard
deviation, n=8) . In addition a wide linear range may be
covered from low ppm to % levels for the fixed gases and total
hydrocarbons.
1-4
-------
The instrument can be used in the manual or automated mode.
With a Stream Selector Valve, 16 samples, blanks or
calibration mixtures can be analyzed unattended. The
automation can be GC based or directed by the Varian DS-650
data system.
-'-Absolute values of Non-Methane Organic Carbon as methane may
be obtained as described in Application Note #12.
1-5
-------
Instrument: Varian 3400 GC
Detector: FID; 220° Range 12
TCD; 120° Filament 160°
Columns: Dedicated Column Set
Carrier: Helium, 30 ml/min
Air, 30 ml/min
Column Temp: 90° isothermal
Concentration: 300 ppm CO2
300 ppm CH4
800 ppm THC in air
THC
FID
,CH4
CO2
TCD
Figure 2
Chromatogram of Standards
I-6
-------
Sample 1 . Sample 16
02, N2, CH4,
CO2 Separation
Vent Heavies
TCD
"Total Hydrocarbons"
FID
Figure 1
Total Hydrocarbons - Fixed Gases
I-7
-------
A CASE STUDY OF THE USE OF THE "SUMMA CANISTER" FOR PASSIVE
"OFF GAS" VENT SAMPLING AND ANALYSIS
MICHAEL F. SOLECKI, Atmospheric/Marine' Physical Scientist,
National Oceanic and Atmospheric Administration (NQAA) Liason
to USEPA Environmental Response Team, Edison, New Jersey;
HOWARD M. SYVARTH, Environmental Scientist, International
Technologies Corporation, Weston/REAC Project, Edison, New
Jersey. PHILIP CAMPAGNA, Chemist, United States
Environmental Protection Agency, Environmental Response Team,
Edison, New Jersey; Wilma Batz, Environmental Scientist, Roy
F. Weston, Weston/REAC Project, Edison, New Jersey.
ABSTRACT
The "SUMMA CANISTER" with a restricted flow orifice was used
to collect and preserve "off-gas" vent samples from a land-
fill in Illinois. Since the method is rarely used for vent
sampling, standard accepted methods were used to confirm the
integrity of the technique. The technique presented some
problems in analytical procedures. They were overcome
however, and the analysis was continued. The problems
encountered are presented here with possible solutions. The
method appears, to have the potential, to be one of the more
efficient ways currently available for collecting and
preserving gas samples.
INTRODUCTION
This paper describes the use of the "SUMMA CANISTER" with a
restricted flow orifice for the collection of gases from a
passive "off gas" venting system, in an actual case study.
This method has rarely been used or documented in the past
for this type of vent sampling. Charcoal tubes and Tedlar
bags were used as a back-up in the event the "SUMMA CANISTER"
method was deemed invalid, since they are methods commonly
used for this type of sampling.
The site is in north central Illinois. It had been a sand
and gravel pit for thirty years prior to its conversion to a
sanitary landfill in 1941. The landfill was in operation
until .1978. At the request of the State of Illinois a
"Remedial Investigation/Feasibility Study" (RI/FS) was
performed between 1983 and 1985. The site was placed on the
National Priority List (MPL) for hazard remediation, as the
result of that study. Based on the RI/FS interim remedial
measures were proposed and accepted. The proposed action was
to repair the cap, install a leachate collection system,
which included a large leachate storage building, and fence
the perimeter. This was completed by 1987.
1-8
-------
In 1988, the United States Environmental Protection Agency
(USEPA) Region V office requested the activation of the
USEPA/Evironmental Response Team (ERT). The ERT was asked to
sample the off-gases of the passive venting system on the
site. There are five vents.
In May of 1988 an on-site evaluation was made. The initial
site entry was performed by the ERT and Remedial Engineering
Analytical Contract (REAC) support personnel. A "site-safety
plan." was established prior to entry based on a previous
study performed by the regional Technical Assistance Team
(TAT). This study indicated that "dichloromethane" was the
main contaminant of concern to on—site personnel. In
addition to the results of the TAT study, the entry team
found quantities of hydrogen cyanide, vinyl chloride and
methane near the base the vents. At this time the "site
safety plan" was changed to reflect the new findings and the
appropriate safety precautions were established and enforced.
Logistical support requirements, equipment staging and
perimeter sampling locations were also established.
Since the venting system is passive, flow rates were
monitored systematically throughout the sample period. This
was done in order to establish an example of the daily trend
of gas flow out of each vent. The data can also be used for
modeling efforts that may be undertaken in the future.
The chemical compounds of concern at the landfill were
methylene chloride, 2 butanone (methyl ethyl ketone), 2
propanone (acetone), 1,1 dichloroethene (vinylidene chloride)
and 1,2 dichloropropane (propylene dichloride). These
compounds were established as those of concern by the
remedial response manager (RPM).
METHODOLOGY
The sampling occurred over a four day period. The first day
"grab" samples were taken to establish an analysis criterion
and to observe if certain compounds are more prevalent during
certain times of the day,
The samples, from each vent, were collected in 1-liter Tedlar
bags. This was accomplished by inserting a Teflon tube
approximately 2 feet down into the vent using a vacuum pump
and desiccator to fill the bag. A sample was to be taken
every two hours for the first twenty-four hours. A temporary
laboratory was set-up at a local hotel. The samples
collected were taken to the temporary laboratory for
analysis. The laboratory was equipped to do flame ionization
and photo ionization gas chromatography analysis.
1-9
-------
Due to diffusion, the bags could not hold the high volatiles
in the gas causing them to escape the bags. This forced the
lab personnel to dispose of the bags for safety reasons
before ,analysis could be completed.' Afternoon and evening
samples that were brought to the lab were immediatly
extracted on to "Tenax" tubes and shipped back to the ERT-TAT
lab in Edison, Hew Jersey for analysis via gas chromat-
ography/mass spectroscopy. The Tedlar bags were, obviously,
not suitable for our purpose at the site. Since the bags had
a high potential for leaking, they should not be shipped. It
was not feasible to make the temporary lab capable of
extracting the samples from tubes, therefore the temporary
lab was eliminated and the technique was discontinued.
Activated charcoal tubes rated at 150mg were used for the
second two days of sampling. A vent extension was
constructed of sheet metal and attached to each vent. It
extended two feet above the top of the vent. This was to
reduce the influence of the wind on the actual vent. A hole
was drilled into the extensions so that the intakes of the
tubes were adjacent to the tops of the original vent pipes.
The off-gases were drawn through the tubes with a vacuum type
sampling pump. The pump was set at a flow rate of 350 cc/min
for 180 minutes. An air sample was also taken from the
leachate storage building using this method, at the same time
period as the "SUMMA CANISTER" and "Anderson" pump described
below. The tubes were analyzed in accordance with NIOSH 1500
for" Hydrocarbons, NIOSH 1501 BP 36°-126°C for Aromatic
Hydrocarbons and NIOSH 1003 Halogenated Hydrocarbons.
In-tandem with the charcoal tubes, two-stage silica-gel tubes
were used to determine concentrations of inorganic acids in
the vent o,ff-gasses. The silica-gel tubes were set at a flow
rate of 200 cc/min for 180 minutes. These tubes , were
analyzed in accordance with NIOSH 7903 for inorganic acids.
The sampling procedure was repeated four times each day for
the two days. The tubes were sent to the ERT-REAC laboratory
in Edison via overnight delivery at the end of each of the
sampling days. This was to insure that the sample tubes
would be stored properly until the analysis could be
completed.
The "SUMMA CANISTER" sampling occurred on the fourth day.
The canisters were evacuated and verified to 0.1 psi pressure
within each canister, prior to sampling. Two methods of
injecting the sample into the canisters were used, a preset
critical orifice and "Anderson" type sample pumps.
Two sampling periods were used one began at approximately at
1030 and the other began at approximately a 2100. This time
span allowed for optimum daily changes in atmospheric
stability. During these same periods samples were taken
1-10
-------
upwind, downwind and in the leachate storage building. The
upwind and downwind samplers were placed no closer than 150
feet to the nearest vent. The "Anderson" type pumps were
used to sample these locations, since it was assumed that the
concentrations of the chemical constituency would be
relatively low. The flow rate for the pumps were set at 27
ml/'min for 360 minutes. This method pressurises the
canisters relative to atmospheric pressure.
It was determined the high concentrations coming out of the
vents would ruin the Anderson pumps. The "Anderson" type
pumps therefore, could not be used for these vents. The
critical orifice was used to sample the off-gas vents. Each
orifice was preset at 10 cc/min. This allowed the canisters
to fill over a six hour period to between 10 psi and 14,7
psi. Making them near neutral to the atmospheric pressure
for the period. The orifice, being a clear orifice (no
traps) allowed pure off-gas to enter the canister. The
canister was then attached to the vent and the intake tube
was inserted into one of the holes used for the flow
measurements. The intake tube was made of "Teflon."
The canisters were sent back to the ERT-TAT laboratory in
Edison for analysis by gas chromatography/mass spectroscopy
<'GC/MS). Some of the canisters filled with critical orifice
were not sufficiently pressurized for the sampling train.
Pressure had to be added via "ultra zero air." Figure 1
shows a schematic of the dilution-pressurization train used.
Upon attempting to analyze the vent samples it was discovered
that the samples had higher concentrations of methane and
carbon dioxide than expected. This caused the cryogenic
traps to condensate and freeze. This rendered them useless
and they had to be completely replaced. After a delay of
approximately eleven weeks the analysis was continued. To
overcome the problem aliquots of sample were adsorbed on to
TENAX/CARBOK" MOLECULAR SIEVE (TENAX/CMS) cartridges by way of
the following method: The TEBTAX/CMS cartridge were placed in
a desorb oven CMS side down to allow a helium purge through
the TEUAX portion first. An aliquot of sample of 15 or 20 cc
was injected into a 20 to 30 cc/min gas stream. Following a
five minute purge, the sample cartridge was reversed and
replaced in the desorb oven for analysis. This allowed for an
accurately measured dilution of the sample, thus methane and
carbon dioxide was reduced to a safe level. The GC/MS was
able to analyze these samples with no difficulty.
The canisters filled by the "Anderson" type samplers were
already pressurized. this allowed them to be attached
directly to the analysis train, see Figure 2. The train
extracts an accurately measured aliquot of sample using a
mass flow controller, dries the aliquot and cryogenically
traps it for subsequent GC/MS analysis.
1-11
-------
P-Bromofluorobenzene and Bromochloromethane were added as
surrogates to all samples and standards prior to analysis.
The TENAX/CMS cartridges were spiked concurrently with the
thermal desorb procedure described1 above. The standards
analyzed contained eighteen components. The sample component
identification/quantitation were done using the Aquarius
software available on the RTE—6 data system.
Proper quality assurance procedures were adhered to. A
sufficient amount of blanks and duplicate samples were
provided as well as a chain of custody.
RESULTS
The laboratory attempted to analyze the "SUMMA" vent samples
approximately five weeks after they arrived from the site.
However, as stated in the methods section, the concentrations
from the samples were so high the cryogenic traps were ruined
and new traps had to be ordered, postponing the analysis.
The Tenax tubes, which contained the off-gas taken from the
Tedlar bags, were then analyzed via the methodology stated
above. The lower detection limits were in the 10 part per
billion range. Table 1 shows the chemical compounds found in
those tubes from an early afternoon and late evening's
sampling for each vent.
The data from the charcoal tubes are contained on Table 2.
The table includes data acquired from a morning and evening
sample periods. This data was done by the ERT laboratory
under the REAC contract as opposed to the TAT laboratory
that performed the SUMMA and Tenax tube analysis. The
samples were analyzed \by gas chromatography only. The high
volatiles, such as vinyl chloride and trichloroflouromethane,
eluted too rapidly and were undetected.
As explained above the off-gas samples taken directly from
the vent by the "SUMMA CANISTERS" with the critical orifice
were adsorbed on to a Tenax/CMS cartridge just prior to
analysis. The samples were stored for eleven weeks in the
"SUMMA CANISTERS" under no special conditions, awaiting the
arrival of the replacement cryogenic traps. Table 3 contains
a list of the compounds found in those samples.
The data from the leachate storage building samples collected
with the "SUMMA CANISTER" connected to the Anderson sampler
compared very closely with the charcoal tube data. As
expected of the compounds tested for, trace amounts of
benzene and toluene appeared. META and Ortho xylene only
appeared on the SUMMA data. All other compounds were
indicated in the raw data as not detected.
1-12
-------
DISCLAIMER: The data recovered from this project is immense.
We therefore did not list all of the compounds found in the
samples. The compounds listed are those of concern for the
project and those usually of concern for this type of
landfill. We feel we have shown a sufficient representative
sample of the data for the reader to understand the point of
this paper.
DISCUSSIOff
Of the compounds of concern only 1,1 dichloroethene appeared.
Many other common landfill compounds also appeared and are
listed in the tables. Although the samples were collected on
different days the results were fairly close in compound
content, especially between the Tedlar bag samples and the
"SUMMA CANISTERS." The compounds from the charcoal tubes
that coelute on the gas chromatograph were identified by
correlating the data with that found in the gc/ms data from
the Tedlar bag and "SUMMA CANISTER" samples. This is with
the exception of ortho-xylene and styrene which coelute and
are both compounds of concern. As can be seen in the tables
the quantities measured varied between the three methods.
Several factors can be attributed to the differences found in
the quantities. The first is that the sampling was performed
over several days. This means that if the quantities of the
compounds varies periodically a true comparison of
quantitation can not be accomplished. The second factor is
that the size and purity of each sample was different. The
Tedlar bag sample was taken over a few seconds to fill a 1-
liter bag of pure off—gas then evacuated on to a TENAX tube.
Evacuating on to a tube is not the standard practice in the
field normally the bags are either analyzed in the field or
shipped as they are. The charcoal tubes were taken over a
three hour period allowing 63-liters of off-gas to pass
through the trap. The "SUMMA CANISTER" collection method
extended over a six hour period allowing approximately 3.6-
liters of pure off-gas to be trapped in the canister. Just
prior to the analysis a 15 or 20 cc sample was removed from
the canister. This aliquot was then purged through a
TENAX/CMS trap prior to analysis (as described in the METHODS
section). The same questions arise with all three of these
methods: How many of the compounds of concern and how much of
each were actually trapped, and does each tube contain a
representative sample of each parcel of off-gas from the
vent? The third factor is the durability of the samples
during shipping and storage. The important factor here is
time, how long will the compounds of concern remain trapped
in each sample container. Currently with these questions,
certain assumptions are made, giving the methods the benefit
of the doubt.
1-13
-------
COHCLUSIOBr
This was the first time that the ERT used the "SUMMA
CANISTER" to collect and store off-gas samples from a. passive
venting system of a landfill. As shown in this report
several problems were encountered that 'had to be overcome to
make the method work. Solutions to those problems
encountered were also presented.
The study compared some of the problems with the current
methods being used for this type of sampling to the "SUMMA
CANISTER" method. It is the feeling of the group that the
"SUMMA CANISTER" method, although does not eliminate, reduces
the impact of the normal problems encountered with the
standard methods. The largest impact, due to its
construction, is on the preservation of the sample during
shipping and storage. Since each canister is shipped under
slightly positive or at atmospheric pressure, and in its own
shipping crate there is little chance of infiltration from
other samples or surrounding contaminated atmospheres. The
valving is well protected so there is little chance of
breakage which could contaminate the surrounding atmospheres.
The canister shell is made of metal so there is little chance
of cracking or shattering as with the glass tubes of the
different types of traps.
The "SUMMA CANISTER" can hold several liters of pure sample.
This allows the pure sample to be transported to a laboratory
for more ideal conditions should the transfer on to tubes
become - necessary. Also, depending on the analysis scheme,
the large volume makes it passible to da more than one type
of analysis on each sample. At the time this paper was
written critical orifices for the canister have been improved
to collect sample over an eight hour period. For passive
venting . this is ideal since the flow of off-gas is not
consistent throughout the day. This type of vent releases
off-gas when a build-up occurs within the laiidfill. The long
sampling time span means that there is more of a chance of
collecting the release when it occurs. By using the critical
orifice type valving it is not necessary to have power at the
vents which cuts down on the logistical problems often
encountered at a landfill. Due to less moving parts, such as
with powered pumps, there is less chance of malfunction
during the sampling period.
These are only some of the advantages we found in using the
"SUMMA . CANISTER" method of sampling. The problems we
encountered due to inexperience with the system were easily
overcome. There is no question that we will use the system
again for this type of project and expand and try it in other
situations.
1-14
-------
SUMMA CANISTER
SAMPLE DILUTION LNE
-------
TABLE 1: TEDLAR BAGS TO TESTAX TUBES IN PPB
VENT1A VENT1E VENT2A VENT2E VENT3A VEIT3E VENT4A VEFT5A VEIT5E
VINYL CHLORIDE
TRICHLORO-
FLOUROKETHANE
1,1 DICHLOROETHENE
METHYLENE CHLORIDE
TRANS 1,2
DICHLOROETHENE
1,1,1
TRICHLOROETHANE
BENZENE
TRICHLOROETHYLEIE
TOLUENE
TETRACHLOROETHYLENE
ETHYL BENZENE
M-XYLENE
0-XYLEIE
STYRENE
META ETHYLTOLUENE
LIMIT OF
QUANTITATION (PPB)
BLOQ=BELO¥ LIMIT OF
QUAITITATION
ND= TOW DETECT
VENT tf/A= AFTERNOON
VENT ¥/E= EVENING
536
1530
ND
ND
ND
53,5
71.3
ND
928
BLOQ
1240
1810
400
ND
49. 4
40
698
2290
BLOQ
608
ND
69.5
146
37.5
1810
44.3
3110
4520
1780
340
706
20
22400
9510
1170
5970
ND
1290
1400
1260
22500
2940
3260
6520
1790
2060
1030
200
20000
4510
263
12400
ND
1310
1410
ND
25000
3440
4010
9680
2440
3690
1680
200
3530
6540
565
17200
ND
637
1050
ND
5650
ND
8060
13500
4320
859
1970
200
2730
1700
420
11100
ND
880
771
ND
5260
ND
5580
9540
2660
510
1330
200
JTD
78
ND
ID
ID
ID
ND
ID
42.2
ID
BLOQ
BLOQ
BLOQ
ND
ND
10
ND
1420
549
1140
ND
2860
384
ID
278
ID
BLOQ
324
BLOQ
214
ID
200
ID
2,790
312
663
ND
580
110
ND
263
ID
BLOQ
219
ND
279
BLOQ
114
1-16
-------
TABLE 2: CHARCOAL TUBES ANALYZED BY GC OILY II PPB
VENT1M VEFT1E VENT2H VEIT2E VEFT3I VENT3E VENT4M VENT5M VENT5E
VINYL CHLORIDE
TRICHLORQ-
FLOURQMETHANE
1,1 DICHLORQETHENE
METHYLENE CHLORIDE
TRANS 1,2
DICHLOROETHENE
1,1,1
TRICHLQROETHANE
BENZENE
TRICHLOROETHYLENE
TOLUENE
TETRACHLOROETHYLENE
ETHYL BENZENE
M-XYLENE
0-XYLENE @
STYRENE @
1ETA ETHYLTOLUENE
NOTES:
BLOQ=BELO¥ LIMIT OF
QUANTITATION
ND= ION DETECT
VEIT V/M= MORNING
VENT ¥/E= EVENING
@=COELUTE CAN NOT
DESCRIMINATE
ND
ND
ND
ND
ND
ND
ND 26.2 99.4
ND ND ND
1442.2 1881.4 508.4
ND
ND
ND
ND
4751
ND
ND
ND
ND
25.7 ND
ND ND
997.4 2289
195.9 202.7
478 2175.9 513.5 1146.7
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
58.6
ND
ND
ND
ND
ND
ND
ND
333,4
ND
772
ND
1558
2966
113
ND
164.1 453.25
ND ND
1414 20634
ND ND
135 4440
5900 7528
13.4
ND
!8802.
ND
6605
6864
447.8
ND
368
ND
6929
6780
569.3
ND
2806
ID
7943
6873
100.3
ND
ND
ND
ND
ND
112.3
ND
ND
ND
75.48
70.8
100.6
ND
ND
ND
ND
2379
ND
3034
ND
ND
ND
2725
ND
1434
ND
ND
ND
3505
ND
775
ND
1-17
-------
TABLE 3: TENAX/CMS CRTRDGS FM SUMMA CNSTRS IN PPB
VEFT1M VENT1E VENT2M VENT2E VENT3M VENT3E VENT4M VENT5M VENT5B
VINYL CHLORIDE
TRICHLORO-
FLOUROMETHANE
1,1 DICHLOROETHEIE
KETHYLENE CHLORIDE
TRANS 1,2
DICHLOROETHENE
1,1,1
TRICHLOROETHANE
BENZENE
TRICHLOROETHYLENE
TOLUENE
TETRACHLOROETHYLENE
ETHYLBEIZENE
M-XYLENE
Q-XYLENE
STYRENE
META ETHYLTOLUENE
LIMIT OF
QUANT ITAT ION (PPB)
ELOQ=BELOW LIMIT OF
QUANT I TATION
ND= NON DETECT
VENT V/M= MORNING
VEIT W/E= EVENING
477
ND
ND
ND
ND
ND
263
63.9
1970
73. 1
3360
5560
1740
ND
621
20
1180
88,5
ND
110
ND
ND
556
155
4790
155
7220
11400
3750
ND
1260
20
6920
474
ND
2350
108
ND
1380
1360
13100
2270
2290
4840
1450
245
814
15
6240
57,7
ND
1770
118
66.4
1430
1400
12550
2180
2230
4680
1540
ND
760
20
1460
156
ID
5170
ND
ND
638
83. 5
5004
96, 7
3170
5200
1590
ND
776
20
1360
202
ND
7000
ND
ND
796
185
6040
106
3620
6140
1830
ND
902
15
ND
ND
ND
ND
ND
ND
650
ND
107
ND
32
83
25
ND
ND
20
ND
ND
ND
ND
ND
ND
62.9
ND
49. 1
ND
179
474
238
ND
170
20
ND
ND
ND
ND
ND
ND
93
ND
134
ND
807
757
180
FD
ND
20
1-18
-------
ABSTRACT: DEVELOPMENT OF A HIGHLY RELIABLE FIELD DEPLOYABLE
ANALYZER FOR VOC
E.B. OVERTON, R.W. SHERMAN, T.H. BACKHOUSE, C.B. HENRY, E.G.
COLLARD, C.F. STEELE, B.S. SHANE, AND T.R. IRVIN, LOUISIANA STATE
UNIVERSITY, BATON ROUGE, LA (EBO, RWS, THE, CBH ECC, CFS, BSS),
AND TEXAS A&M UNIVERSITY, COLLEGE STATION, TX (TR).
Release of toxic volatile organic compounds can cause a particularly hazardous
situation because of the compounds mobility and routes of exposure. Effective chemical
hazard assessment of a given situation requires a knowledge of the identities and toxicity
of airborne contaminants that can be obtained within the time frame before exposures
occur. This requirement translates into the need for a highly reliable qualitative identifier
with rapid response times and sensitivities well below the IDLH concentrations. We
have investigated a microchip gas chromatographic system that can be field deployed
and will provide rapid, highly reliable qualitative and quantitative analyses of volatile
organic compounds (VOC). The microchip GC has essential components etched into a
silicon wafer providing a small, rugged and rapid analyzer for VOC. The unit uses,
simultaneously, narrow-bore high resolution gas chromatographic separation on two
different liquid phases to achieve highly reliable qualitative and quantitative analyses.
Alternatively, we have used the microchip GC as the front end to a small mass spectral
detector. Both units have analysis times in the one to two minute range with detection
limits around 1 ppm for most VOC analytes. If lower detection limits are needed for
ambient air monitoring, a modular concentration device can be used to pretreat air and/or
water samples prior to analysis with the microchip GC systems. We have used the
devices for a variety of applications including analyses of ambient air, industrial
effluents, hazardous waste, and groundwater. Additionally, the data is readily tied into
a database, CAMEO, to provide information on the physical, chemical and toxic
properties of compounds detected at the scene of an incident.
1-19
-------
AN EVALUATION AND COMPARISON OF THE PHOTOVAC TIP II AND
HNU PI 101 TOTAL ORGANIC VAPOR ANALYZERS
Lila Accra, Chemist
Andrew Hafferty, Assistant FIT Office Manager
Ecology and Environment, Inc.
101 Yesler Way, Suite 600
Seattle, Washington 98104
ABSTRACT. Since the beginning of EPA's national hazardous waste site
investigation program in the late 1970s, one of the objectives of
on-site screening for contaminants at hazardous waste sites has been the
health and safety monitoring for on-site personnel. While there are
several hand-held air monitoring instruments currently available,
Ecology and Environment, Inc. (E & E) has primarily used the HNU PI 101
photoionization detector (PID) based instrument for field investigations
and site safety determinations. Photovac, Inc. has more recently intro-
duced the TIP II, which operates on the same principles and serves the
same purpose as the HNU PI 101. Both instruments can provide nonquali-
tative, semiquantitative data regarding the presence of ionizable
species in ambient air. However, there is little published information
which documents the effectiveness of the TIP II in hazardous waste site
investigative work. E & E, under its Zone II Field Investigation Team
(FIT) contract with the EPA, has completed a comprehensive side-by-side
evaluation of the TIP II and PI 101 to determine the utility of the TIP
II for the EPA pre-remedial program.
Evaluation of the TIP II and the PI 101 involved collection and inter-
pretation of both objective data and subjective evaluations. The study
conducted at E & E included a discussion of instrument theory and appli-
cations in order to provide background information. Manufacturer's
literature on operating parameters were considered, including instrument
descriptions, and design specifications, as well as a comparative evalu-
ation of instrument user manuals. Laboratory test results of instrument
operational parameters were evaluated, including battery tests, evalua-
tion of controls, zero stability, flow rates, instrument response time,
and response fluctuation. Instrument performance with selected standard
gases was compared in a controlled environment. As the instruments are
used under a wide variety of ambient conditions, an examination of field
instrument performance was necessary to complete the evaluation. Obser-
vations of experienced PI 101 users are also presented as they apply
specifically to EPA pre-remedial waste site investigations. Test re-
sults indicate that there are differences in the performance and opera-
tion of the PI 101 and TIP II. These differences should be considered
by field personnel when the use of an air monitoring instrument is re-
quired for field operations.
1-20
-------
DISCLAIMER. This report has been prepared by Ecology and Environment,
Inc. under EPA Contract 68-01-7347, and reviewed and approved for public
release by the U.S. Environmental Protection Agency (EPA). Mention of
commercial products does not constitute endorsement by the U.S. Govern-
ment. Editing and technical content of this report are the responsibi-
lity of Ecology and Environment, Inc., Seattle, Washington, and do not
necessarily reflect the views or policies of the EPA.
1.0 INTRODUCTION
Air monitoring survey instruments are routinely used by E&E to assess
site conditions relative to health and safety protocols during the per-
formance of hazardous waste investigative work. Under EPA's Remedial
Planning/Field Investigation Team (REM/FIT), Field Investigation Team
(FIT), and Technical Assistance Team (TAT) programs, the primary air
survey instruments used by E&E have been the Foxboro OVA-128, a flame
ionization detector- (FID-) based instrument, and the HNU System, Inc.'s
PI 101, a photoionization detector- (PID-) based instrument.
Recently, Photovac, Inc. introduced a PID-based air survey instrument,
the TIP II. The TIP II is similar in operation to the PI 101, but is
contained in a single unit (as opposed to two distinct components for
the PI 101). The TIP II also utilizes a liquid crystal display (LCD)
readout (versus an analog meter readout for the PI 101). Unlike the
PI 101, which has been extensively field-tested and described in several
articles, little published information is available to document the ef-
fectiveness of the TIP II in hazardous waste site investigative work.
To facilitate an assessment of the TIP II's utility to FIT Preliminary
Assessment/Site Inspection (PA/SI) activities, E&E's Seattle-based FIT
purchased a TIP II and initiated a comparative evaluation of the TIP II
and the PI 101. A single instrument representing each manufacturer's
model was employed for all tests. The objectives of the study were to:
o Evaluate and compare the specifications and operating parameters
of both instruments in a controlled environment; and
o Evaluate and compare the field performance of each instrument.
2.0 THEORY AND APPLICATIONS
2.1 Photoionization Detector Theory
The PID, which is used in both the Photovac TIP II and the HNU PI 101,
is a non-specific organic and inorganic vapor/gas detector. Ambient air
is drawn through a probe and into a sensor chamber. Inside the sensor
chamber, an ultraviolet lamp emits photons with a prescribed energy.
The photons emitted from the lamp pass through an ultraviolet trans-
mitting window and enter the ionization chamber. Molecules with a lower
1-21
-------
ionization potential (IP) than the energy of the photons will absorb a
photon and become ionized as described in the equation:
RH + hv -> RH+ + e~
where RH = molecule
hv = photon with energy greater than IP of RH
RH = ionized molecule
e = electron
The ionization chamber contains a pair of oppositely charged electrodes.
Ionized molecules migrate to the negative electrode, called the collect-
or electrode. The resulting current is proportional to the concentra-
tion of ionizable species in the air.
Clean air constituents (e.g., 0^, N2, CO, C02> H^O) have IPs which are
higher than the energy of any ultraviolet lamp source commercially
available, and therefore are not detected by the instrument. Although
most volatile organic contaminants in air have a sufficiently low IP,
certain contaminants with very high IP may not be detected by the PID,
even with the highest energy lamp available. For example, CH,, HCN, and
acetonitrile all have IPs greater than 12 eV and are not detected by the
PID.
The detector is calibrated by adjusting the span control so that the
instrument readout matches the concentration of a calibration gas.
Benzene is the usual calibration gas for the PI 101. Isobutylene is
used for the TIP II.
The PID does not detect all contaminants with the same sensitivity. The
only gas or vapor which will be detected quantitatively is the cali-
bration gas. All other vapors are measured as calibration gas equiva-
lents whose "true" concentrations will depend on individual response
factors relative to the calibration gas.
2.2 Applications and Limitations of the TIP II and PI 101
The TIP II and PI 101 are both direct reading instruments designed for
real-time monitoring of contaminants in air. Both instruments measure
total ionizable gases and vapors present in air samples. Neither is
compound-specific within the range of the IP of the instrument lamp. As
stated in Section 2.1, all compounds have unique IP values. In addi-
tion, not all compounds respond with quantitative equivalency. Response
factors for both instruments are compound-specific. For example, equiv-
alent concentrations of toluene and methyl ethyl ketone (MEK) do not
yield the same response on the PI 101 after calibration with benzene.
The response of a 500 ppm MEK sample is approximately 850 while the
response for an equal concentration of toluene is over 1,500.
1-22
-------
For a known contaminant, the actual concentration of the vapor or gas
can be calculated from the instrument reading based on its response
factor relative to the response factor of the calibration gas. However,
if the contaminant(s) are unknown, the instrument readout can yield only
semi-quantitative estimates of the actual levels detected.
These limitations do not preclude use of these instruments when the
objective is to obtain real-time, non-specific "scoping" assessments of
environmental conditions of a site. Applications include: general air
monitoring, leak detection, spill control, site characterization, envi-
ronmental surveys, and emergency response.
3.0 LITERATURE COMPARISONS OF OPERATING PARAMETERS
3.1 Instrument Descriptions
3.1.1 Photovac TIP II
The TIP II weighs approximately 3.5 pounds and is contained in a single,
hand-held unit.
The instrument readout is a lighted digital LCD with a range of 0 to
1,999 ppm. The display update frequency is greater than one per second.
The LCD is visible through a clear curved plastic cover which is flush
with the instrument casing.
TIP II standard equipment includes a 10.6 eV lamp with four other inter-
changeable lamps available: 8.4, 9.5, 10.2, and 11.7 eV. Lamps have a
one-year warranty, except for the 11.7 eV lamp, which has a 90-day war-
ranty because of its inherent instability. The standard 10.6 eV lamp
was used throughout this study.
The TIP II field kit includes a carrying case, span kit containing
standard gases, adjustable headset for audio monitoring, power cord for
battery pack, power cord for portable recorder, three-meter probe exten-
sion, and a wrist strap. Reported instrument specifications are sum-
marized in Table 1.
3.1.2 HNU PI 101
The PI 101 weighs approximately 9 pounds and is comprised of a readout
unit and a probe. The readout unit weighs 7 pounds and is usually
carried with a shoulder strap; the probe weighs 1.5 pounds and is
hand-held.
The instrument has an unlit analog meter readout with a needle that
deflects proportional to the concentration of contaminants. There are
three range settings: 0-20 ppm, 0-200 ppm, and 0-2,000 ppm. The scale
is divided into one-hundred 0.2 ppm increments (0-20 ppm range). The
1-23
-------
operator should use the 0-20 ppm range when in the field, unless a high-
er range is indicated by the meter response.
TABLE 1
REPORTED PHOTOVAC TIP II AND HNU PI 101
INSTRUMENT SPECIFICATIONS
Specifications
Calibration Gas
Safety Class
Detection Limit
Range
Linear Range
Response Time
Zero Drift
Zero type
Weight
Dimensions
Number of Controls
Warm-Up Time
Battery Type
Battery Life
Charge Time
Operation Using A/C
Different Energy
Light Source
Operating
Temperature
Operating Humidity
Sample Injection
Service
TIP II
PI 101
Isobutylene
Division I
0.1 ppm
0-1,999 ppm
0-100 ppm, ±10%
100-1,000 ppm, ±15%
3 seconds
1% Precision
Need zero gas
3 Ibs
45 x 6.3 cm diameter
2 minutes
NiCd
4 hours
16 hours
No
Change Lamp
Not reported
Not reported
Pump
Charged by the repair,
average 8-day turn-
around, free loaners
available, no extended
warranty available
Benzene
Division II
0.1 ppm
0.2-2,000 ppm
0.1-600 ppm
Less than 3 seconds
Less than 1% over 10 hrs
Electronic
Readout unit: 7 Ibs
Probe: 1.5 Ibs
Readout unit:
21 x 13 x 24 cm
Probe: 28.5 x 6.3 cm
diameter
3
20 seconds
Gel Cell
10 hours
3 hrs, 90%; 14 hrs, 100%
Yes
Use different probe
Ambient to 40°C
0 to 95% R.H.
Fan
With Super Warranty,
24-hour service, free
loaners after 72 hours
The PI 101 comes with one of three probes: 9.5, 10.2, or 11.7 eV.
Additional probes may be purchased separately. Different energy lamps
are not interchangeable within a single probe. For example, a 10.2 eV
1-24
-------
probe could not be converted to 11.7 eV by changing the lamp. The
10.2 eV probe is used for most fieldwork, and was used for this study.
No field kit is available for the PI 101. All accessories must be pur-
chased separately. Table 1 lists reported instrument specifications.
3.2 Comparative Evaluation of Instrument User Manuals
Both instruments have extensive user manuals which include sections on
unpacking the instrument, operating instructions, PID theory, detailed
diagrams of the instrument, maintenance and servicing instructions,
troubleshooting guides, lists of accessories, and replacement parts
lists. The manuals do differ somewhat in their organization and com-
pleteness.
3.2.1 Photovac TIP II Manual
The TIP II manual includes an explicit table of contents and an int-
roductory section with a brief overview of the contents of each follo-
wing section. Step-by-step operating instructions are detailed, with
separate sections for qualitative and quantitative operation. A table
of specific span settings for various compounds is presented for quanti-
tative use. The maintenance section includes brief procedures for
troubleshooting and servicing the instrument, as well as normal mainten-
ance procedures. Also included in the manual is an appendix which de-
scribes how to prepare span gas.
3.2.2 HNU PI 101 Manual
The PI 101 manual contains a more general table of contents than the
TIP II manual. A detailed instrument specification table follows the
introductory section. Operating and calibrating instructions are in
narrative form. The manual contains an extensive table of ionization
potentials for different compounds and a reference table of instrument
sensitivities for selected compounds. The PI 101 troubleshooting sec-
tion is more extensive than the TIP II manual. Maintenance and ser-
vicing of the PI 101 are not explicitly described. However, complete
disassembly instructions and detailed electrical diagrams are included.
4.0 LABORATORY TEST RESULTS FOR OPERATIONAL PARAMETERS
4.1 Batteries
Batteries are an important factor in field instrumentation and must
satisfy certain criteria for use in the field. Battery life must meet
or exceed the expected period of active use anticipated between time
periods allotted for battery recharge. Time required for recharge must
be sufficiently brief to avoid delays in fieldwork due to equipment
down-time. The battery must be versatile in order to accommodate a
variety of field applications, such as intermittent use or prolonged
1-25
-------
monitoring. Some type of low battery warning system is also desirable.
The battery must also be convenient to carry, since weight may become a
significant factor.
Each battery was fully charged according to manufacturer's specifica-
tions prior to the tests. Battery life was measured by allowing each
unit to operate without interruption or intermittently. Each battery
was considered expended when the low battery indicator was internally
activated.
4.1.1 Photovac TIP II
The TIP II is equipped with a four-hour NiCd battery. An indicator in
the upper left-hand corner of the LCD will read "LOBAT" when the battery
is low and needs recharging. The instrument should not be used after
the "LOBAT" indicator appears. When the NiCd battery was tested, it was
found to have a three-hour life. Also, NiCd batteries may develop a
"memory." For example, if it is repeatedly used for only one-hour in-
tervals between charges, the battery will eventually have only -.a one-
hour maximum life. Therefore, it is preferable to run the TIP II until
the "LOBAT" indicator appears before complete recharging of the battery.
Recharge time for the TIP II NiCd battery is mandated by the manufac-
turer at a minimum of 16 hours to prevent battery damage. Continuous
overcharging may also reduce battery life. The test results showed that
the NiCd battery life was approximately 3 hours.
The shelf life of the charged battery is less than one month. After
approximately 1.5 hours of use, the unit was stored for 34 days. When
the instrument was turned on, no LCD reading appeared, indicating that
the battery was dead.
For extended use of the TIP II, gel cell batteries are available with
12- or 36-hour operating lives. These batteries are worn over the
shoulder and connected by an electrical cord to the hand-held probe.
The power cord for the battery fits loosely into the main unit and has a
tendency to become disconnected even during limited activity.
4.1.2 HNU PI 101
The PI 101 uses a lead gel cell battery contained in the readout unit.
A low battery is indicated by a red LED light. Also, a battery pro-
tection circuit in the PI 101 shuts down the instrument when the battery
is discharged to below 11 volts. This prevents deep discharging and
extends the battery life. The manufacturer's literature states that the
life of this battery is greater than 10 hours. Tests showed that the
battery life was approximately 16 hours. Partial or incomplete dis-
charge and/or recharge of the gel cell battery will not affect battery
life. PI 101 battery recharge follows an exponential curve in which a
three-hour recharge will bring the battery to 90% capacity, but 14 hours
are required to achieve 100% recharge.
1-26
-------
Shelf life tests were not conducted since the PI 101 is continuously
charged when not in use. Overcharging and operation of the unit during
recharge is not detrimental to the battery.
4.2 Controls
The controls on both instruments were examined in detail to determine
ease of use for fieldwork. The power ON/OFF, zero, and span controls
are the most frequently used and are critical for proper instrument
operation. The zero control is used to set instrument baseline at 0 or
no concentration of ionizable species in a sample of clean air. Instru-
ment sensitivity is controlled by the span setting. Increased sensiti-
vity will result in higher concentration readout even if actual sample
concentration remains unaltered. Span control is employed to match the
instrument's concentration readout to that of a known standard.
4.2.1 Photovac TIP II
The TIP II has an ON/OFF button, and zero and span control knobs with
locking rings. Readout is from 0-1,999 ppm. Both control knobs are
marked from 0 to 9 in 0.5-unit increments over approximately 300 de-
grees. Controls are coarse compared to the PI 101. Turning the zero
control knob from 2 to 3, a 30° rotation, altered the readout from -11.4
to 3.3 ppm when the unit was calibrated with isobutylene. When sampling
100 ppm isobutylene, the readout changed from 65.8 to 122.5 ppm with a
rotation of the span control from a setting of 7 to 8, a 30° turn.
4.2.2 HNU PI 101
The PI 101 controls consist of a function knob, an electronic zeroing
knob and a multi-turn rotations venier span control knob (Figure 7).
The function knob is used to turn the unit on and off and set the read-
out range (0-20 ppm, 0-200 ppm, or 0-2,000 ppm). Since the zero control
is electronic, no "zero air" is needed to calibrate the instrument be-
fore sampling ambient air. The zero control is finer than the control
on the TIP II — one 360° rotation of the zero control knob covered a
range of 0.5 to 6.2 ppm on the 0-20 ppm scale. The span control is also
finer than the TIP II's; a full 360° rotation, from 7 to 6, altered the
readout from 96 to 110 ppm on the 0-200 ppm scale when sampling 100 ppm
benzene.
TIP II sensitivity increases as the span is increased from 0 (lowest
sensitivity) to 9 (maximum sensitivity). PI 101 sensitivity increases
as the span setting is decreased from 10 (lowest sensitivity) to 0 (max-
imum sensitivity).
4.3 Zero Stability
Because action levels for site safety are often in the 1 to 5 ppm range,
it is critical that low end instrument sensitivity be subject to minimal
fluctuation; that is, the instrument should exhibit a high degree of
1-27
-------
precision over time. Drift of the zero setting was measured after each
instrument was charged, zeroed, and calibrated according to manufacturer
specifications.
4.3.1 Photovac TIP II
Zero was set at 0.0 ppm according to the manufacturer's calibration
instructions. The instrument was then used for 0.5 hour. The unit was
turned off for approximately 16 hours. Power was then restored to the
unit and resampling of "clean" zero air continued for another two hours.
The zero calibration held steady at 0.0 during the initial period of
operation. Upon restart, the reading was 0.5 ppm. Zero readings ranged
from -0.4 to +1.0 ppm during the second operating period. The unit was
recharged for 16 hours, after which the zero reading was -0.7 ppm.
4.3.2 HNU PI 101
The PI 101 was used to sample zero air for six hours. The zero readout
remained constant at 0.0 during the entire period. The unit was then
recharged for approximately 60 hours at which time zero air analyses
yielded a reading of +0.4 ppm.
4.4 Flow Rates
The operating flow rate should not affect the quantitative output of
these PIDs, if the concentration of available sample does not change.
The combination of flow rate and response time (Section 4.5) together
dictates the minimum sample volume required for accurate analyses. To
compare flow rates of the two instruments, a Buck Calibrator was used to
collect readings over 10-minute intervals. Ten tests were run on each
instrument, after fully recharging the batteries. Percent relative
standard deviation (%RSD) statistics were calculated and are necessary
to accurately compare flow rate variability of two such different flows.
4.4.1 Photovac TIP II
The average flow rate of the TIP II was 765 ml/min. Flows ranged from
715 to 822 ml/min. The standard deviation for ten readings was 765
±28 ml/min, which was used to calculate a %RSD of 3.7%.
4.4.2 HNU PI 101
The average flow rate of the PI 101 was 168 ml/min. Flows ranged from
163 to 173 ml/min. The standard deviation for ten readings was 168
±3.2 ml/min, which yielded a %RSD of 1.9%. The impact, if any, of these
fluctuations on instrument performances is unknown.
4.5 Response Times/Fluctuation
Response time is the time required by an instrument from the onset of
sampling until a defined percentage or maximum (100%) response is
1-28
-------
achieved. The times required to achieve both 90% and 100% response of
100 ppra calibration gases were measured after both the TIP II and PI 101
were prepared for use according to manufacturer specifications. The
time required for return to a zero reading from 100 ppm was determined
by measuring the time from removal of the 100 ppm standard and immediate
onset of zero air sampling until instrument readout returned to zero.
Fluctuation of the response at 100 ppm was measured to provide an esti-
mate of instrument precision above the zero or low end range.
4.5.1 Photovac TIP II
The TIP II had a 10-second 100% response time. When analyzing 100 ppm
isobutylene, the TIP II display read 98.5 after four seconds. The 90%
response time was less than four seconds. The final reading of
100.2 ppm required 10 seconds. The display fluctuated between 99.5 and
100.2 ppm over a two-minute period. It took approximately 20 seconds
for the instrument to return to zero after changing the sample source to
zero (clean) air.
4.5.2 HNU PI 101
The PI 101 100% response time was 30 seconds. When calibrated with
100 ppm benzene, the 100 ppm isobutylene standard was analyzed. After
10 seconds, the meter read 54 ppm, approximately 87% response. At 30
seconds, the reading was stable at 62 ppm (100% response) and during a
three-minute period the level varied between 62 and 62.5 ppm. Fifteen
seconds after resumption of clean air sampling, the meter read 3.4 ppm,
and after three minutes the level stood at 1.2 ppm. Approximately 15
minutes of zero air sampling were required for the meter to return to
zero.
5.0 LABORATORY STANDARD GAS ANALYSIS TEST RESULTS
5.1 Choice of Gas Standards
In order to test the analyzing capabilities of these instruments, it was
necessary to choose standards with varying IPs, varying sensitivities,
and at least one standard with varying concentrations to verify the
linearity of the instruments.
Compounds on the Target Compound List (TCL) and/or those most frequently
encountered at National Priority List (NPL) sites were chosen as stand-
ards because they would be the most applicable for field use of the
instruments. The limiting factor was availability of the preferred
gases. Benzene, xylenes, and vinyl chloride are all on the TCL and also
are ranked in the top 25 most frequently detected compounds on NPL
sites. Isobutylene was chosen because it is the normal calibration gas
for the TIP II. It was also used for the linearity test for both in-
struments since certified benzene gas standards were not available in
concentrations greater than 350 ppm. Isobutane was used because its
1-29
-------
high IP tested the upper energy limits of the lamps. See Table 2 for a
complete listing of gas standards used.
TABLE 2
GASES USED TO TEST
PHOTOVAC TIP II AND HNU PI 101 RESPONSE
lonization
Standard Concentration* Potential (eV)
Isobutylene, in air
Benzene, in air
p-Xylene, in air
Vinyl chloride, in air
Isobutane, in air
5.8 ppm
22 ppm
100 ppm
539 ppm
100 ppm
39 ppm
205 ppm
512 ppm
9.24
9.24
8.44
9.99
10.57
* Gas concentrations were certified to ±2%.
5.2 Method for Standard Analysis
Standards were purchased from Byrne Specialty Gases, Inc. in Maxicyls
containing 8 cubic feet of gas at 240 psi. One 0-15 psi regulator was
shared for all cylinders. Gases were transferred from the Maxicyls into
tedlar bags through inert tubing. Each tedlar bag and connecting tube
was filled with only one type of gas to minimize potential cross con-
tamination.
Tedlar bags were rinsed before use by partially filling them with stand-
ard gas and then emptying. The bags were then filled with standard gas
just prior to analysis. For analysis, the full bag was connected to the
instrument probe with inert tubing. After a stable reading was achiev-
ed, the bag was emptied until immediately before the next analysis.
5.3 Standard Gas Analyses/Response Factors
The instruments were calibrated and standards were analyzed with both
instruments in three separate tests. The PI 101 uses benzene as a cali-
bration gas while the TIP II uses isobutylene. The PID is more sensi-
tive to benzene than isobutylene. An instrument calibrated to a 100 ppm
benzene standard reading would display only 70 ppm for a 100 ppm iso-
butylene standard. In order to directly compare the actual readouts of
the two instruments, standards were analyzed on the PI 101 twice: first
with the instrument calibrated to benzene, the usual calibration gas,
1-30
-------
and second with the instrument calibrated to isobutylene. Only the
TIP II isobutylene calibration and PI 101 isobutylene calibration data
are directly comparable. The PI 101 benzene calibration data must be
adjusted for varying response factors before they may be compared to
either of the other data sets.
For isobutylene, the TIP II had a greater response to the high (539 ppm)
level standard, while the PI 101 had a greater response to the low
(5.8 ppm) level standard. The PI 101 had a greater response to vinyl
chloride at 205 ppm. The TIP II had a greater response to isobutane at
512 ppm. The results for the p-xylene analyses may not be reliable
because it gave unstable readouts on both instruments, with the observed
concentration constantly rising.
Standard gas analysis results indicate that, in general, the instruments
perform comparably when sampling air. The TIP II, however, yields a
slightly larger and more linear response than the PI 101. Both instru-
ments show some variance in results between calibrations.
As previously stated, the PID does not detect all contaminants with the
same sensitivity. Table 3 presents the calculated response factors
(normalized to a benzene value of 10.0) for the data.
Isobutylene response factors were more consistent on the TIP II than on
the PI 101. Vinyl chloride response factors were approximately half the
expected value on either instrument. As noted earlier, the p-xylene
data were suspect. The measured response factor was one quarter the
expected value on both instruments. No data were available for an ex-
pected isobutane response factor.
Despite the difference in reported lamp energies (TIP II - 10.6 eV,
PI 101 - 10.2 eV), no significant difference in instrument sensitivity
was noted during the analysis of isobutane (IP = 10.57 eV).
5.4 Calibration Hold and Response vs. Battery Life (Charge Level)
In this test, both instruments were calibrated according to the manu-
facturers' recommended protocols immediately after charging the bat-
teries to maximum capacity. Samples were analyzed at regular intervals
over the life of the batteries to determine the effect of charge level
on instrument response. Following the manufacturers' recommendations,
the TIP II was calibrated with isobutylene while the PI 101 was cali-
brated using benzene.
The PI 101 appears to maintain calibration, that is, respond more con-
sistently than the TIP II throughout the life (charge level) of the
battery. Both instruments' responses varied from reading to reading,
but the TIP II appeared to indicate a more definite downward trend with
time.
1-31
-------
TABLE 3
PHOTOVAC TIP II AND HNU PI 101
MEASURED STANDARD RESPONSE FACTORS
(Based on Benzene Response Factor of 10.0)
Compound
Certi-
fied
Cone.
(ppm)
Reported
Response
Factors
Average Response Factors
TIP II
Isobutylene
Calibration
PI 101
Benzene
Calibration
PI 101
Isobutylene
Calibration
Isobutylene 5.8
Isobutylene 22
Isobutylene 100
Isobutylene 539
Benzene 100
p-Xylene 39
Vinyl
Chloride 205
Isobutane 512
7.
7.
7.
7.
10.0
11.4
.0
.0
.0
.0
5.
*
0
6.6
6.1
6.3
5.9
10.0
3.4
2.4
5.2
7.2
6.1
5.5
3.9
10.0
2.9
2.8
4.1
7.
6.
6.
4.
10.0
2.7
,7
,1
.0
.8
3.0
4.9
* None found.
6.0 FIELD MEASUREMENTS
All preceding tests were performed in a controlled environment (the E&E
Region X Mobile Support Base). Since the TIP II and PI 101 are used
under a wide variety of ambient conditions, a field comparison of in-
strument performance was necessary to complete the study.
Before the field comparisons were made, both instruments were fully
charged, zeroed, calibrated, and checked for operational readiness. For
field testing, air from various outdoor sources was monitored with the
two instrument probes held side-by-side (approximately six inches be-
tween sampling probe inlets). Simultaneous readouts for both units were
recorded. Field comparisons were conducted on two separate days. Mete-
orological conditions were similar during both field tests: tempera-
ture, approximately 60°F; winds, light; relative humidity, low with
bright sun.
1-32
-------
An overview of the field test results reveals differences in instrument
performance in an uncontrolled environment. In general, the PI 101
meter response was lower than the TIP II meter response. Since the
PI 101 was calibrated to benzene and the TIP II was calibrated to iso-
butylene, this difference in response was expected. The TIP II has a
significantly wider range in responses than the PI 101, especially when
sampling highly contaminated air. After the first field test was com-
pleted, the ionization chambers were rechecked, and the PI 101 chamber
appeared to be contaminated. Dust from the drums is the suspected cause
of this contamination. After the second round of field sample analyses
was completed, both instruments were again checked and no problems were
noted.
7.0 OTHER CONSIDERATIONS
7.1 Maintenance
Both the TIP II and PI 101 require routine maintenance to insure optimum
instrument performance. The TIP II ultraviolet lamp window requires
regular cleaning. The dust filter, which protects the ionization
chamber from contamination, must be changed when it becomes clogged, or
a drop in instrument sensitivity will result.
The PI 101 also requires periodic cleaning of the ultraviolet light
source window. In addition, the ionization chamber must be checked
regularly for contamination.
Instructions for maintenance are provided in both instruments' operating
manuals. HNU and Photovac will provide on-site training for normal
field maintenance. HNU also offers an in-depth repair course.
Routine maintenance on the PI 101 was completed without incident. A
problem was encountered during routine maintenance of the TIP II. Dur-
ing removal of the lamp, the lamp holder was allowed to rotate slightly,
which caused an electrical connection between the bulkhead and the PC
board to disconnect. Repairs could not be accomplished in the field and
the unit was returned to the manufacturer for service.
7.2 Manufacturer Service
Photovac does not offer field service for the TIP II. However, their
representatives will talk users through troubleshooting and repairs over
the phone whenever possible. If the unit must be returned to the manu-
facturer for repairs, a loaner unit is provided free of charge. Manu-
facturer repair costs are based on the type of repair needed. Average
turn-around for repair is eight days. No extended warranty is available
beyond the initial one-year warranty.
1-33
-------
HNU offers a "super warranty" which includes a toll-free 24-hour phone
line for technical assistance, priority turn-around (usually 72 hours)
for repair with a free leaner if repairs take longer than 72 hours, and
all parts and labor except lamp and battery. Annual inspection and
calibration for the instrument are also included. HNU offers mainte-
nance and support training for the PI 101 at negotiated rates. Addi-
tionally, certification training is available for servicing units.
7.3 Photovac TIP II Upgrade Notice
Photovac has introduced an upgrade program for the TIP II Analyzer.
Changes include: adding a coarse zero control which allows for 10 times
finer adjustment, changing the span control to a logarithmic scale which
would give three times finer adjustment in the most used range, and re-
ducing the display update frequency to one second.
8.0 SUMMARY
Like the HNU PI 101, the Photovac TIP II can provide non-qualitative,
semi-quantitative data regarding the presence of ionizable species in
ambient air. Results for both instruments are dependent upon several
factors, including lamp energy and age, sample flow rate, battery charge
level, specific compound(s) detected, calibration gas used, instrumen-
tation operating procedures, and ionization chamber cleanliness. Test
results for this study indicate that there are certain differences in
the performance of these two instruments (Table 4). The PI 101 has some
drawbacks when compared to the TIP II, such as an unlit meter readout,
greater size, greater weight, and longer purge time between samples.
However, the HNU PI 101 tends to be more suited to the overall needs of
E&E FIT for PA/SI work. Battery life, battery type, ease of control
knob manipulation, readout stability, and greater low range sensitivity
are the major advantages the PI 101 offers over the TIP II.
TABLE 4
SUMMARY OF SELECTED PHOTOVAC TIP II AND HNU PI 101
COMPARISON TEST RESULTS
Feature Tested TIP II PI 101
Response Time, 90% 4 sec. 10 sec.
Response Time, 1002 10 sec. 30 sec.
Sample Flushing Time 20 sec. 15 min.
Fluctuation (60 sec.) <1% <1%
Zero Drift (1 day) + or - 1 ppm + or - Q.4 ppm
Warm-Up Time 10 min. 5 min.
1-34
-------
(cont.)
Feature Tested
TIP II
PI 101
Battery Life
Charge Time
Mean Flow Rate
Flow Rate Range
Average % Response
after prolonged use
(near end of bat-
tery life), com-
pared with original
response
Response Drift %RSD
for all compounds
tested
Readout
Readout Design
Controls
Total Weight
3 hrs., may develop
memory
16 hrs., 100%
765 ml/min.
715-822 ml/min.
82%
12.1%
Digital
Difficult to read
without sample probe
extension or in
bright light
Difficult to adjust,
unprotected against
accidental changes
3 Ibs.
16 hrs.
3 hrs., 90%
14 hrs., 100%
168 ml/min.
163-173 ml/min.
102%
5.6%
Analog
Acceptable
Acceptable
8.5 Ibs.
GENERAL REFERENCES
1. Photovac Inc., TIP II User's Manual, Version 2.1, October 1986.
2. HNU Systems Inc., Instruction Manual for Model PI 101 Photoioniza-
tion Analyzer, 1975.
3. Ecology and Environment, Inc., FIT Operation and Field Manual,
March 1982.
A. Driscoll, J.N. and M. Duffy, Photoionization Detector: A Versatile
Tool for Environmental Analysis, Chromatography, pp. 21-27, May
1987.
1-35
-------
THE AUTOMATED DETERMINATION OF VOLATILE
ORGANIC CONTAMINANTS IN AMBIENT AIR AND/OR SOIL
GAS BY GAS CHROMATOGRAPHY WITH SELECTIVE DETECTORS
Norman Kirshen, Senior Chemist, and Elizabeth Almasi, Chemist,
Varian Instrument Group, 2700 Mitchell Drive, Walnut Creek,
California 94598
INTRODUCTION
VOCs and light gases are major air pollutants in our
environment. They enter the ambient air from industries such
as refineries, dry cleaners, power plants and even bakeries.
But the major source of ambient air pollution is from mobile
sources, i.e., automobiles and trucks. Some of these
pollutants can contribute to the ozone formation by acting as
precursors. The indoor air we breathe at home may be polluted
with VOCs originating from disinfected water, cigarettes or
from household products such as air fresheners, deodorants and
moth balls. The use of industrial solvents or paints and
finishing products in the workplace is another source of
exposure to VOCs.
The discovery of toxic air contaminants in homes adjacent to
hazardous landfill sites and in parks or other facilities
which are built on former landfills (active and inactive) has
led to recently enacted legislation in California, AB3525 and
AB3374, authored by Charles M. Calderon. This legislation
requires testing at all landfills to determine the chemical
composition of air contaminants above, within, and adjacent to
the site to determine the possibility of contaminant migration
beyond the solid waste disposal site's perimeter.
The contaminants that must be screened are shown below:
VOCs Light Gases
Vinyl Chloride Total Hydrocarbons
Benzene
1,2-Dibromoethane Methane
1,2-Dichloroethane Oxygen
Dichloromethane Nitrogen
Tetrachloroethene Carbon Dioxide
Carbon Tetrachloride
1,1,1-Trichloroethane
Trichloroethene
Chloroform
Several analytical methodologies exist for the analysis of
VOCs in ambient air and soil gas. Solid phase adsorption
followed by either solvent or thermal desorption has been used
1-36
-------
for many years in industrial hygiene applications and in stack
gas monitoring. This technique suffers from problems such as
adsorbent contamination, imprecision, and low recoveries for
low boiling VOCs.
Whole air sampling has become the primary method for sampling
air with the introduction of EPA method TO-14. With this
technique a whole air sample is drawn through a cryogenically
cooled trap to freeze out and concentrate VOC contaminants.
The trap is then quickly heated and the VOCs are transferred
to a cryogenically cooled capillary column. The column is
temperature programmed and the VOCs chromatographed to either
multiple selective GC detectors or to a mass spectrometer.
While this technique provides sub part per billion detection
limits, several improvements are possible: (1) automation for
multiple samples, (2) use of a variable temperature adsorption
trap to eliminate the necessity for water removal and (3)
elimination of the requirement for a cryogenically cooled
oven.
Two automated analytical systems have been developed to
analyze sub part per billion levels of VOCs: one a fixed
volume system and the other a variable volume system. Both
systems use a variable temperature adsorption trap (VTAT), a
0.53 mm capillary column (DB-624) and the photoionization
(PID) and Electrolytic Conductivity Detectors (ELCD) or
Electron Capture Detector (BCD) in series.
Herein is provided a description of these two systems, their
operation, and the general analytical procedures followed. In
addition, the results of the following studies are reported:
1) chromatographic retention time and peak area precision, (2)
method detection limits, (3) system linearity, and (4) run-to-
run carryover. Finally, ambient, dry-cleaner, and indoor air
samples are analyzed using the techniques described.
EXPERIMENTAL
Two types of gas chromatographic systems have been developed
for the determination of volatile organic chemicals in soil
gas and ambient air, one a fixed volume and the other a
variable volume system. A description of these two systems
follows.
Fixed Volume Mode
The fixed volume system shown in Figure 1 includes the
following: (1) a two loop gas sample valve (2-ml and 100-ml),
to handle high and low concentration samples, respectively,
(2) a surrogate standard valve for introduction of a surrogate
sample with each run, (3) a variable temperature adsorption
1-37
-------
trap (VTAT) packed with Tenax/activated charcoal and suitable
for trapping all VOCs studied, (4) a packed or 0.53 mm column
for compound resolution, (5) a Photoionization and either an
Electrolytic Conductivity Detector or Electron Capture
Detector plumbed in series and an optional (6) 16-port stream
selector valve (SSV) for automatic selection of up to 16
samples, calibration mixtures, or blanks.
The two-loop gas sample valve (CSV) provides the option of
introducing a small (2-ml) or large (100-ml) volume air
sample, standard or blank into the system. Therefore, soil
gas samples which might have high VOC concentrations or
ambient air samples which usually have low concentrations may
be analyzed over a wide concentration range. Samples are
drawn into the loop with a slight vacuum provided by a small
diaphragm pump or are allowed to purge the loop from a
pressurized canister.
A surrogate standard CSV with a loop <1 ml is filled from a
pressurized tank of a known concentration of a surrogate
standard. An 8 PPM concentration of 1,3-bromochloropropane
was used in this work.
The VTAT (Figure 2) is coiled about a 1 inch grooved aluminum
mandrel which contains a heating cartridge and temperature
probe. This is housed in a small insulated chamber into which
cryogenic fluid is delivered. The trap may be cooled to -
190°C and temperature programmed at rates up to 180°/min to a
maximum temperature of 400° with all temperature parameters
selectable from the GC keyboard. A 1/8" o.d. stainless steel
trap containing 14 cm of Tenax TA and 8 cm of activated
charcoal was used in this system. Glass beads may also be
used reguiring cryogenic (-180°C) temperatures for trapping
VOCs.
While a 1% SP-1000 on Carbopack B (60/80) packed column was
used in earlier work, a 30M X 0.53 mm DB-624 column provides
improved chromatography.
Fixed Volume Procedure
Samples of ambient air, soil gas, calibration mixtures or
blanks are collected in either Tedlar™ bags or canisters and
connected to the SSV. if automation is operative, the
appropriate GC method is automatically activated and the
sample is drawn into or allowed to purge the 2- or 100-ml loop
depending upon sample concentration. The sample loop is then
flushed for 5 minutes (see Table 1) to the adsorbent trap
(20 C) where VOCs are trapped, the air and moisture being
vented. Simultaneously the surrogate sample is flushed into
the trap. When trapping is completed the trap is isolated,
1-38
-------
preheated to 250°C, and the VOCs are backflushed to the
analytical column for separation. Subsequently the trap is
baked out for seven minutes and then cooled to 20°C with LN2
or LCO2 in preparation for the next analysis.
Variable Volume Mode
The variable volume system shown in Figure 3 consists of the
following: (1) a multi-port valve for switching sample to
VTAT, introducing surrogate sample, and allowing direct
injection to column, (2) the VTAT with trap isolation valve,
(3) mass flow controllers, (4) an optional SSV and (5) column
and multi-detectors as described above.
Unlike the fixed volume system, the variable volume system
allows volumes from 50 ml to 800 ml or greater to be sampled,
depending upon the mass flow controller setting, trapping
time, and trap temperature. The air or soil gas sample is
first drawn through the multiport valve from a Tedlar bag or
canister while the surrogate loop is flushed. Then a fixed
volume of surrogate standard along with the sample flow is
diverted to the trap for a set time and mass flow rate. When
the trapping is complete, the trap is isolated and preheated
to 250° and then backflushed to the column for chromatography
and detection. Subsequently trap baking and cooling occurs
prior to the next run.
RESULTS AND DISCUSSION
Two gas standards (Table 2) were purchased for use in the
evaluation and development of the fixed and variable volume
systems. These high and low concentration standards possess
concentration levels possible in soil gases and ambient air,
respectively. The low concentration standard was used for
approximately one year only because of degradation.
Ten VOCs plus a surrogate standard, 1,3-bromochloropropane,
are shown well resolved in Figures 4 and 5 on packed and
capillary columns.
The analyses times for these two columns are very similar.
But the capillary column provides several advantages: (1) a
more efficient column for resolving many additional VOCs of
interest, (2) narrower peak shapes to improve sensitivity, and
(3) lower bleed allowing for the screening for higher
molecular weight VOCs.
1-39
-------
To evaluate the VOC systems the following studies have been
performed:
1) Retention time and peak area precision
2) Run-to-run carryover
3) System linearity
4) Method detection limits
5) Variable volume sampling
All parameters were evaluated on the fixed volume system while
Nos. 1, 4 and 5 were evaluated on the variable volume system.
A series of ten low-concentration and high-concentration
standard runs were made on the system. Relative standard
deviation (RSD) of retention times and responses were
determined. Results are shown in Table 3. With the exception
of vinyl chloride, the RSDs of the retention times are
generally less than 0.03%. The response RSDs are less than 5%
for low- and less than 2% for high-concentration VOC gas
standards. Surrogate standard response RSD is 1.5%.
Retention time and response RSDs were similar with the
variable volume system.
When analyzing standards, blanks and samples of varying
concentrations, it is necessary to be aware of run-to-run
carryover. This has been investigated in the fixed volume
system by running the high-concentration standard in the 100-
ml loop followed by a blank. Results are shown in Figures 6.
Only EDB is seen as a carryover peak and this is due only to
its high sensitivity on the ECD. Its estimated carryover is
<0.1%.
The linearity of response of the system to varying
concentrations of VOCs is not necessarily determined by
detector linearity. Bag permeation, if using Tedlar™ bags,
adsorption, and decomposition are mechanisms that may be
operative and difficult to gauge in a complex gas analysis
system. Nevertheless, the linearity of several VOCs has been
studied over a wide range of concentrations using the PID, ECD
and ELCD detectors. Results are shown in Figures 7 and 8.
The data shown in Figure 7 was generated using a 100 ml sample
for the most dilute standards and 2 ml for the most
concentrated standards. Only 100 ml standards were used to
obtain data shown in Figure 8. The system linearity is
demonstrated by the constancy of the calibration factors over
a wide concentration range.
The flexibility of choosing either the 2- and 100-ml loop in
the case of the fixed volume system or the virtually unlimited
choice of sample volume in the variable volume system allows
-------
one to choose the sample volume that will cover the
concentration range of interest.
Method detection limits (MDL) were determined using 100 ml of
gas mixtures prepared from primary standards. VOC
concentrations of approximately five to ten times estimated
detection limits were prepared and analyzed at least seven
times after which standard deviations of the results were
determined. Standard deviations were multiplied by a T factor
3.14 (99% confidence level) giving the MDLs shown in Table 4.
Limits obtained using the PID, BCD and ELCD are shown compared
with those required under the California Calderon Bill.
The experimental MDLs meet these requirements, the BCD
providing its greatest sensitivity for those VOCs having the
highest electronegativity and the ELCD in the halogen mode
providing relatively uniform response for all the listed VOCs
with better response for the mono- and dichloronated solutes.
Therefore, the requirements of a particular application would
dictate the selection of either the BCD or ELCD for the
halogenated components: (1) sensitivity, (2) selectivity, and
(3) response uniformity.
APPLICATIONS
Several grab samples of air were collected in evacuated
canisters after which the canisters were pressured to 20 psi.
These canisters were then attached to the stream selector
valve on the fixed volume system and 100-ml aliquots flushed
through the trap and the trap subsequently desorbed to the
0.53 mm DB-624 column. The ambient air sample (Figure 9)
exhibited only a trace of tetrachloroethene on the ELCD and a
unidentified trace VOC on the PID. On the other hand the grab
sample of air taken from the dry cleaner establishment (Figure
10) showed significant levels of tetrachloroethene,
ethylbenzene, and toluene and traces of methylene chloride,
carbon tetrachloride and trichloroethene. Figure 11 shows an
air sample taken from the moist room air following a shower.
As expected from chlorinated water, traces of trihalomethanes,
methylene chloride, and tetrachloroethene are found. Finally,
a trace level standard was collected in a canister and sample
volumes of 100 and 300 ml were drawn through the trap of the
variable volume system. As expected the 300 ml sample
provides three times the VOC response of the 100 ml sample
(See Figure 12) . This system is capable of extremely high
sensitivity limited only by the breakthrough volume of the
trap which is approximately 700 - 800 ml at 20°C.
1-41
-------
CONCLUSIONS
Two gas chromatographic techniques are applicable to th<
multilevel determination of VOCs in soil gas and ambient air,
One is a fixed volume and the other a variable volume
technique. The fixed volume system allows either twc
milliliter or one hundred milliliter samples while the
variable volume system allows sample volumes of from fifty tc
about eight hundred milliliters. Central to these systems is
a variable temperature adsorption trap. This trap adsorbs anc
concentrates the VOCs from the sample while allowing moisture
to pass through thereby eliminating the need for a dryer.
Excellent precision is obtained both for retention times and
peak areas. And carryover from highly concentrated samples is
minimized. The method detection limits possible are in most
cases under 0.5 ppb and depend upon the detector and the
sample size chosen. Selectivity, sensitivity, and response
uniformity must be considered in detector choice. Finally,
the flexibility of these systems is paramount with
applications ranging from soil gases in hazardous waste sites
to indoor atmospheres to ambient air and source emissions.
1-42
-------
Table 1
Timing of the Air Analysis
Time
Analysis
Sequence
Trap
Temperature
Column
Temperature
VOC's
0.
SAMPLING
GC/DS to
Ready
Cooling to
20°
Cooling to
35°
01 5
.0 7.
CONCENTRATION
VOC Trapping
20°
35°
r
Trap
Isolation
Preheat
/Xf800/min
35°
I
0 10.0 17
.0 25.
SEPARATION ON ANALYTICAL COLUMN
Trap
Desorbtion
to Column
250°
35°
Trap Bake
250°
35° -
CH2=CHCI;CH2CI2
Cooling to 20°
^______—————~~W&t
— — — ~~~T6°/min
CHCI3; 1,2 DCE; CCI4; 1,1,1 TCE ....
-------
Table 2
Air/Soil Gas Analysis
Standard VOC Mixtures
Compounds Concentrations (PPB)
High Low
Dichloromethane 4000 5.0
Chloroform 500 0.2
1,2-Dichloroethane * 2.0
1,1,1-Trichloroethane 400 0.4
Carbon Tetrachloride 50 0.5
Trichloroethene 300 0.5
1,2-Dibromoethane 5 0.2
Tetrachloroethene 50 0.5
Vinyl Chloride 5000 6
Benzene 2000 5
* Not included in this mixture
I-44
-------
Table 3
Retention Time and Response Reproducibility
% Standard Deviations
Response
Components Retention Times Low Concentrations High Concentrations
Vinyl Chloride 0.50 2.2 0.38
Dichloromethane .035 3.8 0.21
Chloroform .017 3.1 0.55
1,2-Dichloroethane .025 3.4 —
1,1,1-Trichloroethane .020 0.61 0.82
Carbon Tetrachloride .018 2.41 4.0
Trichloroethylene .014 0.77 0.90
Benzene .022 3.90 0.35
1,2-Dibromethane .017 5.80 0.48
Perchloroethylene .013 3.80 1.14
-------
Table 4
Air/Soil Gas Analysis
Method Detection Limits in PPB (V/V)
ARB ELCD/0.53 Cap ECD/Packed
Compound Guideline Column Column
Vinyl Chloride 2.0 .20 .80
Benzene 2.0 .20 .25
1,2-Dibromoethane 0.5 .10 <.01
1,2-Dichloroethane 0.2 .05 .19
Dichloromethane 1.0 .05 .60
Tetrachloroethane 0.2 .05 <.01
Carbon Tetrachloride .02 .04 <.01
1,1,1-Trichloroethane 0.5 .03 <.01
Trichloroethane 0.6 .02 .01
Chloroform 0.8 .03 <.01
-------
Figure 1
Air/Soil Gas Analyzer
VOCs
Sample 1 ... Sample 16
Surrogate Sample
j PID
f
L
Vinyl Chloride
Benzene
2 ml/100 ml loop
Gas Sample
Valve
Trap Selection
& Backf lush
Function
Sampling
Trapping
Trap Preheat
Desorption
Trap Bake
Trap Cooldown
System Operation
Duration (min)
7
5
2
3
7
10
ECD
Dichloromethane
Chloroform
1,2-Dichloroethane
1,1,1 -Trichloroethane
Carbon Tetrachloride
Trlchloroethene
1,2-Dibromoethane
Perchloroethylene
1,3 Bromochloropropane
Trap Temperature
30°
30°
30° to 250°
250°
250°
250° to 30°
-------
Figure 2
VTAT
Coolant Input
Trap Column
Insulation
Heater Block
1-48
-------
Figure 3
Air/Soil Gas Analysis - VOCs
Variable Volume Mode
Surrogate
Standard
Flow Controller
Carrier Gas
Sample 1 — Sample 16
Sample
Selector Valve
| MFC
Multiport
Switching
Valve
[MFC_
Vent or Vacuum
Auxiliary
Gas
Column
Variable Temperature
Adsorption Trap
BCD or ELCD
-------
Figure 5
Landfill Gas/Air Analysis
Low Levels of VOC's on DB-624 Column
voc
PPB
N
Ch
VINYL CHL
CH2CL2
CHCL3
111TCE
CCL4
1.2DCE
TCE=
PCE=
EDB
BENZENE
X
U
•J
U
(N
n
J •*
U W J
SOU
O E-" U
U
EH
W O
U n
(Q
D
U
U
as
CQ
10.0
8.20
0.20
0.60
0.10
0.65
0.20
0.30
1.00
4.30
a,
oi
m
a
PID
s
HALL
-------
Figure 4
Air/Soil Gas Analysis
Column: 1% SP-1000/Carbopack B (60/80)
z
IU
z
T:
N
JC
u
o
Di
U
N
f f
11
co o
ro o
in o
— N
O
N ^w
N nn
N
'jj
r\
in
in
'JJ Dl
in o
N
co
01
PID
^
n
CO
tj-
'-*
N
N
N
co nrrj N in
*) NOW 10 CM
co corcn 01 o
CO
01
0
N
01
N
o
A/
111
z
UJ
N
Z
LU
m
ECD
lJ
D
D
!•'-
N
N
(vl •
—
ID
01
O -•
-------
Figure 6
Carryover, VOC's
100-ml Loop
High Concentration Standard
f j
PID
BCD
b 1 I
fc 3
Blank
1-52
-------
Figure 7
Air/Soil Gas Analysis
System Linearity
nt/Area
i
s ^
Fa
o
i
o
01
i
rati
b
2
Ca
o
01
i
o
2
i
Sample Sizes: 2 ml
100mL
Chloroform
A
Tetrachloroethene
-0—e-
• 0
Carbon Tetrachloride
H
A A
Detector: BCD
PID
-B-
A
1 10
Concentration (PPB)
100
A
0
500
-------
3.0-
1 1-5
o
1-
•C
I
I
§
0.3-
0.2
0.1
0.0
Figure 8
Air/Soil Gas Analysis
System Linearity
Sample Sizes: 100mL
Benzene
Chloroform
A
•e-
A
Dibromomethane
Vinyl Chloride
Detector: ELCD
PID
-B-
A
0
I
10
I
100
Concentration (PPB)
1000
-------
Figure 9
Ambient Air Sample
100ml
5
X
PID
I I I
i i
L»
H
H
X
ELCD
H «
Sin
B..
U«
td
Cu
u <
,A
C2CI4
O.3 ppb
I I I I I
-------
Figure 10
Dry Cleaner Air Sample
100ml
en
CD
2
M
z
o
p£
M
N
§
fj
I I I
Toluene
32ppb
PCE=
Ethylbenzene PID
130 ppb
x:
o
fi
z
H *
6..
oS
W
Ou
PCE=
1300 ppb
ELCD
I I I I I I I
-------
Figure 11
Shower Room Sample
100ml
u
H
•v
PID
S
en
u
H
2
H
as m
>s.
08
•«
-------
Figure 12
Variable Volume Sampling
100ml
V.
u
3
X
I
V.
u
s
u
S
300ml
ppb
1. Vinyl Chloride 5
2. Methylene Chloride 6
3. Chloroform .2
4. 1,1,1-Trichloroethane .4
5. Carbon tetrachloride .5
6. 1,2-Dichloroethane 2
7. Benzene 5
8. Trichloroethene .5
9. Tetrachloroethene .5
H
X
S >r>
ON
8-
H
Cu
w
1 2
U
H
H
H »
X in
"I
O N
H
Oi
U <
gg
O <
-------
THE DETERMINATION OF FIXED GASES AND NON-METHANE ORGANIC
COMPOUNDS IN SOIL GAS AND AIR
Norman Kirshen. Senior Chemist, and Elizabeth Almasi, Chemist,
Varian Instrument Group, 2700 Mitchell Drive, Walnut Creek,
California 94598
The growing number of emission sources in the industrial,
municipal, and transportation areas makes source gas and
ambient air monitoring necessary. The safety of existing
landfill sites also requires the monitoring of soil gas and/or
ambient air both inside and outside the perimeter of the
disposal area. The presence and amounts of CO and CO2 can
provide important information about aerobic and anaerobic
reactions inside the landfill. The monitoring of methane to
follow its migration can eliminate explosion hazards by
preventing its collection in airpockets in residences or in
other populated areas.
To attain the requirements of the Clean Air Act, total
hydrocarbon monitoring is necessary since many organics behave
as precursors in ozone formation. It is also necessary to
measure CO levels since they reflect the general ambient air
quality -
The system described here measures CO, CO2, CH4, and non-
methane organic compounds (NMOCs) in aerial matrices. The
light gases CO and CO2 are converted to CH4 enhancing their
detectability (
-------
NMO
HeavtesBaddlush
Hopcalite
CMdaUon
Reduction
CO,
Figure 1
NMOC Analyzer
I-60
-------
Column: Dedicated column set, 80° isothermal
Flow: 30 ml/min N2
FID: Range 12
Concentration: 50 ppm each
CH4
CO CO2
C2H6
I i £5 i 2| i i i i i i i I
Figure 2
Chromatogram of Standards
a i i
15 min
1-61
-------
Column: Dedicated column set, 80° isothermal
Flow: 30 ml/min N2
RD: Range 10
CH4
2.8%
NMOC
205 ppm
I I !
I u I u I
I I I" I I I I I
Figure 3
Chromatogram of a Typical Sample
15min
I-62
-------
2,0
1.5
• "8 •" T "*
5 1.0^- CF ± s %s
Methane 1.68 ± .10 (6.2%)
g U • Butane* 1.72 * -096 (6.2%)
I _
_ 'Normalized as methane
-------
BIOLOGICAL TEST METHODS
-------
EFFECT OF CHEMICALS ON SOIL NITRIFYING POPULATIONS
USING A CONTINUOUS-FLOW CULTURE TECHNIQUE
Charles W. Hendricks and Albert N. Rhodes1, U.S. Environmental Protection
Agency, 200 S.W. 35th Street, Corvallis, Oregon 97333
ABSTRACT
This study examines the effects of Roundup [N-(phosphonomethyl)glycine]
and N-Serve [2-chloro-6-(trichloromethyl )pyridine] on nitrifying organisms
in static batch, perfusion soil columns, and a new continuous-flow soil
column system. The continuous-flow method is new to nitrification studies
and was shown to produce greater nitrifier activity than either static
batch or perfusion techniques. Both N-Serve and Roundup were shown to
significantly inhibit nitrification in treated soils over untreated
controls. N-Serve completely inhibited nitrification at concentrations
greater than 42 ug nitrapyrin g'1 dry soil, and Roundup significantly
reduced nitrification at 6.8 and 68 mg glyphosate g"1 dry soil. Heterotro-
phic bacterial populations increased significantly in continuous-flow
columns treated with 42 mg nitrapyrin and 68 mg glyphosate g'1 dry soil.
Numbers of heterotrophs were not significantly different from controls in
soils at lower concentrations. Numbers of nitrifying bacteria did not
appear to change following treatment, although nitrification was
inhibited. Fluorescent antibody analysis of nitifiers revealed that
Nitrosolobus was more numerous than Nitrosospira and Nitrosomonas.
Nitrosolobus increased in number, whereas the other two genera remained
unchanged.
In this study, the continuous-flow system proved to be both reliable and
useful in the culture of nitrifying bacteria. This method is an
alternative to traditional static and perfusion culture techniques for
evaluation of the effects of chemicals on microbial biogeochemical cycles,
and can benefit soil toxicity assessment at hazardous waste sites.
INTRODUCTION
Traditional laboratory culturing of nitrifiers involves the use of static
soil cultures or perfusion columns (Lees and Quastel, 1946). Although
useful, both techniques suffer from continual changes in substrate
concentration and soil chemistry. A continuous-flow method has recently
been developed for the culture of heterotrophic soil microorganisms
(Hendricks et al.. 1987). This technique provides a fixed concentration
of nutrients continuously to a soil column and alleviates the limitations
mentioned above. Recently, Rhodes and Hendricks (1989) have modified the
technique and have shown that the continuous culture of nitrifying
bacteria in soil has promise as an alternative to traditional methods.
1 Present address: Department of Biology, U.S. Air Force Academy, Colorado
Springs, Colorado 80840-5701.
1-65
-------
This study was designed to evaluate the new continuous-flow method by
comparing it with the static and perfusion techniques for culturing
nitrifying bacteria in soil, and to determine the response of nitrifying
and heterotrophic bacteria to treatments of N-Serve and Roundup. N-Serve
is a commercial nitrification inhibitor and Roundup is one of the commonly
used herbicides in agriculture today. Preliminary experiments conducted
as part of this study determined that, of the three cultural methods, the
continuous-flow method supported high nitrifier activity and was the most
sensitive in measuring nitrification. For this reason, the continuous-
flow method was used to examine the effects of treatment on specific
genera of ammonium oxidizing bacteria, denitrifying microorganisms, and
heterotrophic microbial populations within the soil columns.
MATERIALS AND METHODS
Soil
The soil selected for this study was obtained from plots in Benton County,
Oregon. The soil is a dark brown fine-silty mixed mesic Argiaquic Xeric
Argialboll classified in the Amity series. Amity silty clay loams are
typically deep, poorly drained, and were formed in mixed alluvium
terraces. Soils of the Willamette Valley of Oregon receive approximately
60 cm of annual rainfall, but are dry for most of the summer months
(Knezevich, 1975). Soil used in this study was randomly collected from
the surface 4 cm, air dried at room temperature, sieved through a 2 mm
pore screen, and stored in plastic bags at 4°C.
Culture Technique
The continuous-flow method of Hendricks et a]_. (1987) was modified (Rhodes
and Hendricks, 1989) to culture nitrifying bacteria. The nitrification
medium was based upon Schmidt and Belser (1982) and amended with 250 mg
NH4-N 1" . Each column received 15 g dry weight Amity soil, and the medium
was metered into the column at a rate of 10 ml day"1 for 16 days. Effluent
was removed at the same rate and samples were collected every two days and
analyzed for NH4, N03, and N02. Each experiment was conducted for 16 days.
Treatment Chemicals
Roundup [N-(phosphonomethyl)glycine] (Monsanto Agricultural Products
Company, St. Louis, MO)] and N-Serve [(2-chloro-(trichloromethyl)pyridine
(A. T. Talcott, Dow Chemical Company, Midland, MI)] were obtained as
commercial formulations. Both compounds were well-mixed and added
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
1-66
-------
directly to the soil on day 8 at applications of 0, 0.042, 0.42, and 4.2
mg active ingredient (AI) g"1 dry soil for N-Serve and 0, 0.68, 6.7, and
68 mg AI g dry soil for Roundup.
Microbiological Analysis
Heterotrophic bacteria were enumerated by plate counts using soil extract
agar amended with 1% glucose and starch casein agar (Wollum, 1982).
Denitrifying bacteria were enumerated using the MPN technique of Focht and
Joseph (1973). Nitrifying bacteria were enumerated by the Most Probable
Number (MPN) technique using a 96 well microtiter plate (G. Stotzky,
personal communication) and the medium of Schmidt and Belser (1982). This
medium is an ammonium sulfate based, chemically defined substrate for
nitrifying bacteria. All cultures were incubated at 25° ± 0.1°C. Agar
plates were incubated for 7 days, denitrifier MPN tubes for 14 days, and
nitrifier MPNs for 6 weeks.
Direct Counts
Nitrifying bacteria were also enumerated by epifluorescent direct counts.
Fluorescent antibody cocktails specific to Nitrosomonas. Nitrosospira. and
Nitrosolobus were used to stain soil extract samples (Schmidt, 1974;
Demezas and Bottomley, 1986). All samples were viewed using a standard
Zeiss microscope (Carl Zeiss, Inc., New York, NY) configured for
epifluorescence. For the purpose of enumeration, 25 fields per slide
were counted and bacterial numbers calculated per gram of dry soil.
RESULTS
Nitrification rates using the static, perfusion, and continuous-flow
techniques are shown in Figure 1. The effect of application of the
various concentrations of the two chemicals on continuous-flow cultures
is shown in Figure 2 (N-Serve) and Figure 3 (Roundup). Inhibition of
nitrification by Roundup appears to be transient and suggests recovery of
the bacterial of activity towards the end of the experiments.
Table I summarizes results for heterotrophic and denitrifying bacterial
populations in both N-Serve and Roundup treatments. Numbers of organisms
cultured were statistically significant from controls in columns treated
with 4.2 mg AI g"1 dry soil and 68 mg AI g"1 dry soil for N-Serve and
Roundup, respectively.
MPN analysis for both ammonium and nitrite oxidizing bacteria are shown
in Table 2. No change in nitrifier numbers is evident over the course of
these experiments.
Table 3 lists the number of Nitrosomonas. Nitrosospira. and Nitrosolobus
enumerated by the Fluorescent Antibody procedure (FA). Nitrosolobus was
the most numerous at all treatment levels over the course of the
1-67
-------
experiments. However, no statistical change in populations was deter-
mined.
DISCUSSION
Roundup has been shown to be an inhibitor of soil nitrifying capability
in continuous-flow soil columns (Rhodes and Hendricks, 1989). The
mechanism of toxicity in other bacteria is thought to involve disruption
of protein synthesis (Fisher et al_, 1986). In our studies, Roundup does
not appear to inhibit protein synthesis in soil based upon the signifi-
cant increase in heterotrophic and denitrifying populations in contin-
uous-flow columns by day 16.
The mechanism of nitrification inhibition cannot be determined from our
studies. However, the MPN and FA data indicate that nitrifying bacterial
densities do not change between untreated controls and treated columns.
Our data suggest that of inhibition of the soil nitrification process is
probably the result of an effect on cellular metabolism rather than
inactivation of the nitrifier population.
It has been reported that glyphosate may not be inhibitory to all
microorganisms. Pipke et a]_. (1987) have shown that an Arthrobacter sp.
can utilize glyphosate as its sole source of phosphorus. They found that
Arthrobacter does possess a glyphosate transport system, but this has not
been reported for other bacteria. The uptake of glyphosate was inhibited
by organophosphates, organophosphonates, and orthophosphates in their
study.
Currently, only one study has examined the effects of glyphosate on soil
nitrogen processes. Carlisle and Trevors (1986) demonstrated that
glyphosate and Roundup induced inhibition of indigenous nitrifiers in
perfusion columns containing a Canadian sandy loam soil. They found that
pure glyphosate was more inhibitory than Roundup when applied at the same
concentrations based upon active ingredient g"1 dry soil. Inhibition by
Roundup was observed only at levels greater than 76.7 ug of glyphosate
g~ dry soil. Field studies conducted by Ana'yeva et aj_. (1986) indicate
that high concentrations of glyphosate and other commercial pesticides
disrupted the soil microbial population for several months following
treatment. Since nitrifying bacteria have been found to be sensitive to
a number of chemical compounds, and many of these compounds do not produce
the same response in other physiological groups of organisms, nitrifying
bacteria may prove to be unique and useful organisms to study the impacts
of xenobiotic chemical compounds on soil microbial processes.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to Dr. E. Schmidt,
University of Minnesota, for graciously providing the antisera for the
fluorescent antibody direct counts.
1-68
-------
LITERATURE CITED
Ana'yeva, N. D., B. P., Strekozov, and G. K. Tyuryukanova. 1986. Change
in the Microbial Biomass of Soils Caused by Pesticide. Agrokhim.
5:84-90.
Carlisle, S. M., and J. T. Trevors. 1986. Effects of the Herbicide
Glyphosate on Nitrification, Denitrification, and Acetyl Reduction
in Soil. Water, Air, Soil Pollut.
Demezas, D. H., and P- J. Bottomley. 1986. Autecology in Rhizospheres
and Modulating Behavior in Indigenous Rhizobium trifolii. Appl.
Environ. Microbiol. 52:1014-1019.
Fisher, R. S., A. Berry, C. G. Gaines, and R. A. Jenson. 1986.
Comparative Action of Glyphosate as a Trigger of Energy Drain in
Enbacteria. J. Bacteriol. 168:1147-1154.
Focht, D. D., and H. Joseph. 1973. An Improved Method for the
Enumeration of Denitrifying Bacteria. Soil Sci. Soc. Am. Proc.
37:698-699.
Hendricks, C. W., E. A. Paul, and P. D. Brooks. 1987. Growth Measure-
ments of Terrestrial Microbial Species by a Continuous-Flow
Technique. Plant Soil 101:189-195.
Knezevich, C. A. 1975. Soil Survey of Benton County Area, Oregon. US
Department of Agriculture, Soil Conservation Service, Washington D.C.
Lees, H., and J. H. Quastel. 1946. Biochemistry of Nitrification in
Soil. 1. The Site of Nitrification. Biochem. J. 40:803-815.
Pipke, R., A. Schultz, and N. Amphein. 1987. Uptake of Glyphosate by an
Arthrobacter sp. Appl. Environ. Microbial. 53:974-978.
Rhodes, A. N., and C. W. Hendricks. 1989. A Continuous-Flow Method for
Measuring Effect sof Chemicals on Soil Nitrification. Toxicity
Assess, (in press).
Schmidt, E. L. 1974. Quantitative Autecology Study of Microorganisms is
Soil by Immunofluorescence. Soil Sci. 118:141-149.
Schmidt, E. L., and L. W. Belser. 1982. Nitrifying Bacteria. In. Page,
A. L., R. H. Miller, and D. R. Keeny (eds.). Methods of Soil
Analysis — Part 2. Chemical and Microbiological Properties, 2nd
Edition. American Society of Agronomy, Inc. and Soil Science Society
of America, Inc.
Wollum, A. G. 1982. Culture Methods for Soil Microorganisms. In Page,
A. L., R. H. Miller, and D. R. Keeny (eds.). Methods of Soil
1-69
-------
Analysis — Part 2. Chemical and Microbiological Properties, 2nd
Edition. American Society of Agronomy, Inc. and Soil Science Society
of America, Inc.
1-70
-------
Table 1. Effect of Various Concentrations of N-Serve and Roundup on Heterotrophic Bacteria Growing on Soil
Extract Agar (SEA), Starch Casein Agar (SCA), and Nitrate Broth (DN) in Continuous-Flow Cultures
DAY
Chemical Q gi
DncenLrdLion
(mg/g Dry QN $EA $CA QN SEA SCA D
Soil)
N-Serve
0.00 8.012 8.36 8.22 7.04 8.26 8.23 7
(0.15)3 (0.15) (0.15) (0.08) (0.11) (0.16) (0
0.042 7
(0
0.42 7
(0
12
N SEA
.15
.58)
.40
.41)
.86
.41)
8
(0
8
(0
8
(0
.17
.09)
.12
.09)
.38
.09)
SCA
7
(0
8
(0
8
(0
.17
.21)
.90
.21)
.07
.21)
8
(0
7
(0
7
(0
DN
.10
.34)
.34
.34)
.94
.34)
8
(0
8
(0
8
(0
16
SEA
.07
• 17)
.42
.17)
.23
.17)
8.
(0.
8.
(0.
SCA
05
07)
20
07)
8.32
(0.07)
4 2 8.51 8.51* 8.20 7.89 8.65* 8.64**
(0.41) (0.09) (0.21) (0.34) (0.17) (0.07)
Roundup
0.00 8.02 8.36 8.22 7.04 8.26 8.23 6.88 7.98 7.96 7.01 8.21 8.14
(0.15) (0.15) (0.15) (0.08) (0.11) (0.16) (0.28) (0.18) (0.20) (0.17) (0.16) (0.16)
0.68 7.34 8.36 8.36 7.40 8.56 8.64*
(0.22) (0.16) (0.18) (0.17) (0.16) (0.16)
6.8 7.95* 8.82** 9.07** 7.44 8.68* 8.68*
(0.28) (0.18) (0.20) (0.17) (0.16) (0.16)
68 7.48 7.71 8.38 7.94** 9.02** 8.99**
(0.40) (0.26) (0.23) (0.17) (0.16) (0.16)
1 Cultures were treated with the chemicals on Day 8.
2 Data represents mean Log (CPU or MPN)/g dry soil.
3 Numbers in parentheses indicate standard error.
* = Significant difference from control (0.05 > p > 0.01).
** = Highly significant difference from control (p < 0.01).
-------
Table 2. Effect of N-Serve and Roundup on Ammonium (AO) and Nitrate Oxidizing
Bacteria (NO) in Continuous-Flow Culture
Chemical
oncentratioi
(mg/g Dry
Soil)
N-Serve
0.00
0.042
0.42
4.2
Roundup
0.00
0.68
6.8
68
DAY
i 0 81
AO NO AO NO AO
6.282 6.46 6.23 6.21 5.80
(0.15)3 (0.32) (0.21) (0.33) (0.56)
5.81
(0.49)
5.10
(0.49)
4.04
(0.97)
6.28 6.46 6.23 6.21 5.80
(0.15) (0.32) (0.21) (0.33) (0.18)
5.93
(0.18)
5.61
(0.18)
6.13
(0.21)
12
NO
6.01
(0.34)
6.15
(0.37)
5.82
(0.37)
5.57
(0.42)
6.01
(0.08)
6.18
(0.08)
6.08
(0.08)
5.94
(0.10)
AO
5.87
(0.43)
5.10
(0.43)
5.19
(0.37)
6.86
(0.75)
5.87
(0.26)
5.74
(0.26)
5.77
(0.30)
5.45
(0.26)
16
NO
6.37
(0.15)
6.33
(0.15)
6.07
(0.15)
5.74
(0.15)
6.37
(0.13)
6.30
(0.13)
5.86
(0.15)
5.85
(0.13)
1 Cultures were treated with the chemicals on Day 8.
2 Data represents mean Log (MPN)/g dry soil.
3 Numbers in parentheses are standard error.
1-72
-------
Table 3. Effect of Various Concentrations of Roundup on the Growth of Nitrosolobus (Ml), Nitrosospira (Np),
and Nitrosomonas (Ns) in Continuous-Flow Culture
DAY
Chemical Q oi •>•-,
Concentration
(mcj/cf Drv m >i >i >n n .. m »
\J":3/ ^^ uf- y ^^ ^_ ^^ ^j^ ^jp ^^ ^j^ j^_
Soil)
Roundup
0.00 8.022 8.36 8.22 5.55 5.08 4.27 5.40 5.01
(0.15)3 (0.15) (0.15) (0.16) (0.23) (0.25) (0.18) (0.1)
16
Ns Nl Np
3.97 5.59 5.02
(0.15) (0.0) (0.21)
Ns
4.36
(0.10)
0.68 5.37 5.26 4.37 4.99" 4.88 4.08
(0.13) (0.11) (0.21) (0.10) (0.15) (0.07)
6.8 5.71 5.21 4.34 5.50 5.11 4.14
(0.10) (0.09) (0.12) (0.10) (0.15) (0.07)
68 5.44 4.88 4.84* 5.40 5.17 4.44
(0.13) (0.11) (0.21) (0.10) (0.15) (0.07)
Cultures were treated with Roundup on Day 8.
Data represents mean Log (cells/g) dry soil.
Numbers in parentheses indicate standard error.
= Significant difference from control (0.05 > p > 0.01).
= Highly significant difference from control (p < 0.01).
-------
240
_ 200 -
o
CO
160 -
Q
ro
O
en
Days
Figure 1. Nitrification in soil using static (A) , perfusion
and continuous-flow (O) culture techniques.
-------
400
300
'a, 200
I I I I I I I I I I I I I I I I 1 I
0
10
15
Days
Figure 2. Effects of N-Serve on nitrification in continuous-flow
culture. N-Serve concentrations: 0.0 (O), 0.042 (X),
0.42 O , and 4.2 (A) mg nitrapyrin g dry soil.
-------
400
Days
Figure 3. Effects of Roundup on nitrification in continuous-flow
culture. Roundup concentrations: 0.0 (O), 0.68 (X),
6.8 (Q) , and 68 (A) mg glyphosate g"1 dry soil.
CD
-------
TOXICITY EVALUATIONS FOR HAZARDOUS WASTE SITES:
AN ECOLOGICAL ASSESSMENT PERSPECTIVE
GREG LINDER, MICHAEL BOLLMAN, WANDA BAUNE, KEVIN DEWHITT,
JENNIFER MILLER, JULIUS NWOSU, SHEILA SMITH, DAVID WILBORN,
AND CATHY BARTELS, NSI TECHNOLOGY SERVICES CORPORATION; JOSEPH
C. GREENE AND LAWRENCE A. KAPUSTKA, U.S. ENVIRONMENTAL
PROTECTION AGENCY, ENVIRONMENTAL RESEARCH LABORATORY, 200 S.W.
35th STREET, CORVALLIS, OREGON.
ABSTRACT
Ecological assessments for hazardous waste sites should
include acute toxicity tests as well as short-term tests which
measure biological endpoints other than death. Toxicity and
field assessment methods may be assembled into "tool boxes"
which reflect not only the site-specific demands made by the
ecological assessment process, but the continuing progress in
methods development. Toxicity assessment tools may yield
information regarding acute biological responses elicited by
site-samples as well as suggest longer-term biological effects
(e.g., genotoxicity or teratogenicity) potentially associated
with subacute and chronic exposures to complex chemical
mixtures characteristic of hazardous waste sites. Toxicity
tests, however, are but one component of an ecological
assessment for a hazardous waste site; field components must
be given equal regard during the early phases of site
evaluation. This becomes particularly important when field
sampling is considered, since integration of toxicity
assessments (be those in situ or laboratory-generated) and
field assessments requires a well-designed sample plan to
establish linkages among toxicity, site-sample chemistry and
adverse ecological effects, if apparent. Spatial statistic
techniques like kriging are finding increased applications in
linking toxicity with other elements of site-evaluation (e.g.,
field-sample chemistry). Through kriging, for example, areal
distributions for site-specific toxicity and chemistry data
sets may be derived, then "maps" of site-sample toxicity and
chemistries overlaid. Patterns of coincidence apparent in
these distributions may then suggest linkages among toxicity,
site-contaminants, and adverse ecological effects.
INTRODUCTION: Approaches to Ecological Assessment
Ecological assessments for hazardous waste sites have recently
gained increased attention after the passage of the Superfund
Amendments and Reauthorization Act of 1986 (SARA). As a
result, US EPA has drafted numerous guidance documents which
suggest approaches to the evaluation of adverse ecological
effects which may exist at hazardous waste sites (US EPA
1988a; US EPA 1989; Warren-Hicks, et al. 1989); the
application of ecological assessment techniques has become an
integral part of the remedial investigation/feasibility study
1-77
-------
process, and in general ecological assessments have assumed
a greater role in site assessment.
Ecological assessments at hazardous waste sites may require
various methodologies which reflect the site-specific demands
required by waste sites. Ecologically based hazard
assessments may be considered as integrated evaluations of
effects which are attained through site measurements of
toxicity and exposure completed in both laboratory and field.
As complex functions, toxicity and exposure integrations may
yield an estimate of hazard associated with any particular
waste site (Figure 1).
Figure 1. Toxicity and exposure assessments within the hazard evaluation process.
Toxicity assessments are derived from acute tests as well as
subacute and chronic tests which measure biological endpoints
other than death. Generally, these toxicity assessments are
derived from laboratory-generated data. Exposure assessments
within an ecological context frequently rely upon field
methods which measure ecological endpoints, either on-site or
at reference sites, and yield survey data relevant to
estimates of adverse ecological effects associated with a
waste site.
Both toxicity and field assessment methods may be compiled
into "tool boxes" (Kapustka and Linder 1989). Depending upon
the environmental matrix being tested, site-specific toxicity
assessments may be derived using tests selected from these
"tool boxes" (Horning and Weber 1985; Peltier and Weber 1985;
Weber, et ai. 1988; Greene, et al. 1988), which may include
invertebrate and vertebrate, algal, plant, and microbial test
systems (Figure 2). These toxicity assessment tools may yield
information regarding acute biological responses elicited by
site-samples or their derivatives, and may suggest longer-term
biological effects (e.g., genotoxicity or teratogenicity)
potentially associated with subacute and chronic exposures to
complex chemical mixtures characteristic of hazardous waste
sites. Potential contaminant migration as well as soil
attenuation of contaminant effects may also be evaluated using
these toxicity assessment tools, if eluates are prepared and
tested in the laboratory. In situ toxicity assessments, while
not as well developed as laboratory toxicity tests, are
becoming more prominent in the ecological assessment process
1-78
-------
(Warren-Hicks, et al. 1989; Murphy and Kapustka 1989),
especially since in situ methods infer a linkage between
toxicity and exposure functions, and reduce the problems
associated with lab-to-field extrapolations of toxicity data.
Figure 2- Teat profiles useful in toxicity evaluations for ecological assessments may include
representative organisms amenable to direct and indirect tests on site-sanplea.
Toxicity tests, however, are but one component of an
ecological assessment for a hazardous waste site; field
components must be given equal regard during the early phases
of site evaluation. This becomes particularly important when
field sampling is considered, since integration of toxicity
assessments (be those in situ or laboratory-generated) and
field assessments requires a well-designed sample plan to
establish linkages among toxicity, site-sample chemistry and
adverse ecological effects, if apparent. Ecological methods
applicable to hazardous waste site evaluations are varied, and
reflect the diversity apparent in waste sites which occur in
a variety of geographic settings. Depending upon the
environmental setting (aquatic or terrestrial) and the site-
specific questions being asked in the evaluation process,
methods have been collected into ecological assessment "tool
boxes" (e.g., LaPoint and Fairchild 1989; Kapustka L989; McBee
1989; Bromenshenk 1989) which provide a source of techniques
available for ecological site assessments.
Spatial statistic techniques like kriging may be applied to
site assessment and help establish linkages between toxicity
and other components of site-evaluation (e.g., field-sample
chemistry or field surveys). Kriging is a tool borrowed from
geostatistics which may help integrate site-specific measures
of toxicity and chemistry derived from site-samples. For
example, site-soil samples may be evaluated for their toxicity
potential through bioassays which evaluate soil eluates (e.g.,
algal or invertebrate short-term toxicity tests), then eluate
chemistries (e.g., routine eluate chemistry) may be completed
for these same site-samples. Through kriging, areal
distributions for these site-specific toxicity and chemistry
data sets may be derived, then "maps" of site-sample toxicity
and chemistries developed. Coincidence patterns among these
spatial distributions may subsequently suggest linkages among
toxicity, contaminants, and adverse ecological effects.
1-79
-------
To illustrate these techniques available in the ecologies!
assessment "tool box," a hazardous waste site case study wil;
be outlined briefly and the integration of site-sampl<
chemistry and toxicity data will be summarized spatially.
WASTE SITE CASE STUDY
The Drake Chemical Superfund site1 in Pennsylvania covers
approximately eight acres, and consists of landfills, various
lined and unlined treatment lagoons, and a dry unlined "canal1
lagoon. Historic records indicate that chemical intermediates
for dyes, Pharmaceuticals, and cosmetics manufacturing were
produced at the site. Also, various intermediates for the
herbicide and pesticide industry were produced, includinc
2,3,6-trichlorophenylacetic acid (and its intermediates),
which was the major contaminant found within the site
boundaries and at some distance from the source.
Stratigraphically, much of the site consists of an overburden
comprised of sludge and contaminated soil lenses, with
unconsolidated materials ranging from near-surface sandy silts
and clays to near-bedrock coarse sands and gravels. Bedrock
is characteristically composed of fractured shales and varies
in depth depending upon site location. An "erosional channel"
may exist, but has not been clearly delineated. Groundwater
occurs in the unconsolidated materials, with the water table
being consistently maintained at a depth of 10 to 15 feet.
Perched water lenses are not infrequent, and occur in
associations with the clays and silts characteristic of near-
surface formations. Hydraulic conductivities and flow
velocities indicate that movement of groundwater through the
unconsolidated materials occurs at 3.5 to 20 feet per year.
Surface waters near the site consist of two rivers. The site,
however, is bermed by road and railroad embankments.
Historically, the site has been inundated by floodwater, but
under normal conditions the runoff is contained within the
site boundary where it percolates into the soil or directly
enters the on-site leachate lagoon as surface runoff.
Topographically, the leachate lagoon is the low point on-site.
Various landscape constructions have been completed during the
site's history, including a French drain and various catch
basins designed to direct surface drainage into one of the
bordering creeks.
METHODS AND MATERTAT.fi
Site-samples received. Site-samples were obtained through ar
on-site contractor from locations (surface and below surface
sample sites, 40 sites x 3 depths plus eight additional
samples) identified in Figure 3. All site-soils were storec
'case study based on waste site assessment completed
conjunction with US EPA regional staff.
1-80
-------
DRAKE CHEMICAL SITE. LOCK HAVEN. PA
-------
at 4°C until soil processing and in-lab toxicity tests were
initiated.
Eluate preparations from site-soils. No direct toxicity tests
on soils (see Figure 2) were completed for the samples
received from the site, but eluates were prepared from
untreated site-soils to evaluate the potential mobility of
chemical constituents which were present in the soil samples.
Upon receipt, soil samples were screened through a 1/4" soil
sieve, then eluted. Screened site-samples were mixed with
four volumes eluent (deionized water) per gram dry weight
site-soil or sediment. The slurry was then mixed in complete
darkness for 48 hours at 20 +/- 2°C. After mixing, the
resulting eluate was centrifuged, then filtered through a 0.45
urn cellulose acetate or glass fiber filter. Eluate
preparation incorporated original sample moisture content into
its preparation and hence, a constant "solute/solvent" ratio
was assured during the extraction of any site-sample. The
filtered eluates were subsequently evaluated with an
appropriate aquatic test system such as the algae (Selenastrum
capricornutum') or macroinvertebrate (Daphnia magna) toxicity
test.
Toxicity assessment tools: Indirect tests on soil eluates.
A short-term chronic, Selenastrum capricornutum 96-hour
toxicity test was conducted on soil eluates, and yielded
toxicity estimates for soil contaminants which were
potentially mobile owing to their water solubility. In the
algal bioassay, growth in test concentrations which was less
than that in controls indicated inhibitory effects during the
96-hour exposure; stimulation was indicated if growth in
exposed systems was greater than growth in concurrent
controls. Briefly, in the test algae were exposed to known
concentrations of soil eluate. Algal cells were inoculated
into test flasks which contained known concentrations of soil
eluate, then were incubated for 96 hours in environmental
chambers held at 24 +/- 2°C and 4304 +/- 430 lux (continuous
light). The typical algal bioassay included a range of test
concentrations capable of yielding EC50 data (effective
concentrations which yield 50% reduction in algal growth
relative to controls after 96 hours), as well as subculturing
information pertinent to the evaluation of lethality. Cell
counts performed on electronic particle counters yielded algal
biomass estimates based upon cell counts and mean cell
volumes. Evaluations of 96-hour EC50s were completed using
appropriate statistical methods (Stephan 1977; Stephan and
Rogers 1985; Greene, et al. 1988).
Site-soil eluates were also evaluated with a standard Daphnia,
magna 48-hour static acute toxicity test. Owing to its
geographic distribution and relative ease in laboratory
culture systems, D. magna was used in these bioassays. As
a representative aquatic macroinvertebrate (e.g., role in
aquatic food chains), D. maqna complemented the algal bioassay
1-82
-------
component completed as part of the toxicity assessment.
Briefly, the bioassay used less than 24-hour old D. magna
neonates which were exposed to test sample concentrations
diluted (volume:volume) to yield logarithmically spaced
exposure concentrations ranging from 100% site sample to 0%
site sample (dilution water alone). The bioassays were
conducted at 22 +/- 2°C (16:8, light:dark) ; ten D. maana
neonates each were placed into triplicate exposure
concentrations when the bioassay was started, and at the end
of the 48-hour exposure mortality was assessed. Survivorship
data was subsequently evaluated to yield LC50 estimates
(Hamilton, et al. 1977; Stephan 1977), which are reflected as
the percent site sample which is associated with 50%
mortality.
For these indirect tests, toxicity categories were applied to
the data listed in Appendix 1. As such, a class ranking of
short-term toxicity estimates was established for the site
samples. For purposes of this toxicity assessment, the
following four groups of test results were established on the
basis of sample dilutions associated with LC50 or EC50
estimates:
Group 1 - median effect less than or equal to 20% sample
dilution
Group 2 = median effect greater than 20%, but less than or
equal to 50%
Group 3 = median effect greater than 50%, but less than or
equal to 80%
Group 4 = median effect greater than 80% sample dilution.
Routine eluate chemistries. Hardness, salinity, conduc-
tivity, total organic carbon (TOG), and pH were determined for
each site-soil eluate (Appendix 2), and contributed to
toxicity interpretations derived from preliminary spatial
statistic analysis. On the work completed for the site
toxicity assessment, sample pH was measured prior to toxicity
bioassays, and determined whether adjustments were required
before toxicity testing was begun. A second pH was
subsequently taken upon completion of the bioassays, since pH
effects may be quite pertinent to the toxicity information
which was collected over the course of exposure.
Spatial statistic tools for ecological assessment. For the
preliminary mapping of site chemistry and toxicity data,
Geostatistical Environmental Assessment Software (GEOEAS, US
EPA 1988b) was used. Variogram analyses and kriging were
performed on sample eluate data (Appendix 1 and Appendix 2)
using programs VARIO and KRIGE included in the GEOEAS "tool
box" of geostatistical methods; by using program CONREC the
resulting data files were used to construct preliminary
1-83
-------
contour maps for both chemistry and toxicity data generate.
from soil eluates. Examples of these maps are illustrated ii
Fiqures 4 5 and 6 which summarized representative spatia:
distribution estimates for selected toxicity and chemica:
endpoints measured on surface samples.
RESULTS AND DISCUSSION
Toxicity Assessment. Appendix 1 tabulates the comparative
toxicity estimates (as LC50s or EC50s) determined for the site-
samples. Indirect measures of biological effects were
determined on soil eluates and were derived from aquatic
toxicity tests (48-hour acute test with D. magna and 96-houi
short-term chronic test with S. capricornutuml . For the alga]
and daphnid toxicity tests respectively, Figures 4 and E
illustrate mappings of site-specific toxicity following
variogram analyses and kriging with GEOEAS, while a mappinc
for a representative soil eluate chemistry (e.g., hardness)
is given in Figure 6. Toxicity estimates were based on the
toxicity classifications listed above (e.g., Groups 1, 2,
3, and 4) and were derived from the toxicity test results
listed in Appendix 1.
Routine Chemistry Support. Appendix 2 lists the results for
routine chemical analyses for eluates. In part, these were
used for interpreting the biological effects summarized in
Appendix 1 (see Toxicity Assessment), though the chemistry
data routinely included in these toxicity screening tests
precludes extensive comments regarding mechanisms of toxicity.
Toxicity evaluation: Correlative Analysis. The sheer number
of samples and the heterogeneity in contaminant loads which
potentially influence toxicity estimates precludes any
definitive statement regarding causality and the toxicity
measures derived for the site in these preliminary analyses.
The complementary features of the toxicity tests, however, are
apparent from a brief overview of the short-term toxicity
estimates generated from the work on the site-samples. The
96-hour short-term chronic algal bioassay and the 48-hour
acute daphnid toxicity test provided the data base upon which
toxicity assessments were developed. For example, when
categorized into the toxicity ranks outlined above, the algae
and daphnid tests jointly identify 40% of all the site-
samples (40 sites x 3 depths plus eight additional soils or
128 samples) as presenting Group 1 endpoints (median effect
estimates < 20% eluate), which should be considered as
potentially highly toxic. These two single-species toxicity
tests are quite complementary. Ten samples (10 of 128, or 8%)
were jointly flagged by these tests as being highly toxic; the
algal and daphnid bioassays respectively identified 27% (35
of 128) and 5% (6 of 128) of site-samples as being in Group
* ,T0 effect" samples were jointly identified by the algal
and daphnid tests in 60% (77 of 128) of the soil eluates
evaluated for short-term toxicity. Alone, the 96-hour algal
1-84
-------
bioassay identified 35% (44 of 128) of all the samples as
being in Group 1, while the daphnid test identified 10%, or
13 of 128 samples as being in this highly toxic category. If
one extends this clustering approach to toxicity assessment
and includes Groups 1 and 2 as particularly relevant toxicity
categories, the algal and daphnid tests respectively
identified 50% and 23% of all sample eluates as being highly
to moderately toxic. Both identified only 9% of the samples
as being in Group 3, while Group 4 (low toxicity, median
effect estimates >80%) samples were identified in algal and
daphnid tests at 41% and 68%, respectively.
Toxicity evaluation: Spatial distributions. Summary
correlative analyses could be drawn for the toxicity
assessments at various depths, but Figures 4 and 5 amply
illustrate a preliminary toxicity mapping which considers the
spatial distribution of toxicity within two surface dimensions
for both algae and daphnid. Figure 6 illustrates two
distribution maps for surface soil-eluates tested with algae
*>
Figure 4. Spatial distribution of toxicity indicated oy tne snort-term algal tes-c and ilius-crated on
toxicity contour map generated by COKREC program in GEOEAS for surface samples [Group 4 toxicity
category ( ECK[JB > *O%} T Group 3 toxicity category (SO* < EC^^a <_ 8O% ) ? Group 2 toxicity category (2O*
< EC50s <_ 5O% ) ; Group 1 toxicity category :O% < EC50s ^_ 2Cf* > ] . H - highest toxicity location; L -
lowest toxicity location. Reference area unepecifled.
northeast
1-85
-------
and daphnia. From routine eluate chemistry results (Appendix
2), mappings for each chemistry data set (e.g., eluate
hardness and salinity in Figure 6) could be plotted, again
using the programs of GEOEAS. The preliminary assessment for
spatial correlations between site-chemistry and toxicity data
could then be achieved through simple "overlay" techniques,
and the linkage between toxicity and eluate chemistry inferred
from the coincidence patterns noted between the two mappings.
The significance of overlay techniques, however, must be
viewed in the context of potential bias associated with the
mappings generated through GEOEAS. For example, variogram
analysis in GEOEAS relies heavily upon interpretation and
"trial and error" computations to derive a variogram model
used in kriging (US EPA 1988b) ; more generally, map overlay
techniques present a variety of problems when independently
derived maps are compared in a spatial context (Bailey 1988).
Pigxu-e 6. Spatial distributlon of eluate chenistry data, e.g., hardness and salinity, illustrated on
toxicity contour oap generated by COHREC program In GEOEAS for surface samples. Reference area
unspecified. r
Even without using overlay techniques, distribution maps may
be beneficial to toxicity data interpretation nonetheless.
For Drake Chemical, even a cursory review of the toxicity data
at the various depths presented patterns of toxicity as
identified by algae and daphnia toxicity maps. For example,
surface point estimators of highly toxic soils (EC50s < 20%)
were identified by nine algal short-term endpoints which,
depending upon closest-neighbor toxicity characteristics,
resolves into contour mappings (Figure 4) similar to those
illustrated for daphnid toxic responses in Figure 5. The
daphnid toxicity rankings derived from the acute tests suggest
fewer highly toxic point estimators but the mapping presents
a highly coincident toxicity distribution
1-86
-------
Though exploratory by design, these preliminary mappings and
toxicity interpretations may bear more relevance to site
evaluation than methods related to hypothesis testing in a
strictly statistical sense. More rigorous analytical
techniques could be applied if additional site information
were gathered, and if greater resolution were indicated beyond
these short-term screening tests. Reference sites should be
critically defined in order to weight toxicity endpoints
within the overall site evaluation. Considering the influence
that sample integrity bears on toxicity assessment, well-
designed field studies should be emphasized regardless of
whether laboratory-based or in situ toxicity assessments are
integral components in the ecological assessment of a site.
Through these correlative analyses, preliminary toxicity
evaluations may yield information pertinent to site
assessments and identify potential topics relevant to site
management, e.g., site characterization and effectiveness of
remediation efforts (Athey, et al. 1987).
SUMMARY: Toxicity evaluations for ecological assessments
As part of the overall ecological assessment for hazardous
waste sites, toxicity assessments contribute significantly to
site evaluation. These expectations should be clearly defined
in the development of the site-specific data quality
objectives. Both laboratory and field studies are critical,
if the ecological assessment is intended to be an integral
part of site evaluation. The following summary points
illustrate the value of toxicity assessments in the evaluation
process for site:
(1) the waste site presented a characteristically widespread
spatial toxicity distribution, though its significance must
be evaluated relative to a site-specific reference; the role
of well-defined reference sites becomes invaluable in
interpreting the extent to which the potential toxicity is
expressed in the field;
(2) the algal and daphnid toxicity tests indicated that
quantitatively different biological responses occurred at
various sampling points (surface and below surface) on the
site; the toxicity assessment clearly indicated the potential
for high toxicity to occur within ecological contexts, if
water soluble constituents from the chemical mixtures in soils
were mobilized;
(3) the algal 96-hour short-term chronic test presented EC50s
of less than 20% site-sample dilution (Group 1) in 27% of the
site-sample eluates; the 48-hour daphnid acute test presented
Group 1 responses to 5% of the site-samples; the algal and
daphnid tests were complementary and together identified 40%
of the site-samples as potentially highly toxic (EC50s of less
than 20% site-sample);
1-87
-------
(4) vertical migration of toxicity was apparent from the
preliminary mapping completed for each of the depths tested,
but the significance of the biological responses requires more
familiarity with the site reference;
(5) mechanisms of toxicity and causality should not be
inferred from preliminary correlative or spatial distribution
analyses, nor should toxicity distributions based upon
associations between site locations and screening test results
infer chemical specific migration without the supporting
analytical results.
REFERENCES
Athey, L.A., J.M. Thomas, J.R. Skalski, and W.E. Miller.
1987. Role of acute toxicity bioassays in the remedial
action process at hazardous waste sites. EPA/600/8-87/044.
U.S. Environmental Protection Agency, Environmental
Research Laboratory, Corvallis, OR.
Bailey, R.G. 1988. Problems with using overlay mapping for
planning and their implications for geographic information
systems. Environ. Manag. 12:11-17.
Bromenshenk, J. 1989. Terrestrial invertebrate surveys. In
W. Warren-Hicks, B. Parkhurst, and S. Baker, Jr. (eds.).
Ecological assessment of hazardous waste sites. EPA/600/3-
89/013. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR.
Greene, J.C., C.L. Bartels, W.J. Warren-Hicks, B.R. Parkhurst,
G.L. Linder, S.A. Peterson, and W.E. Miller. 1988.
Protocols for short-term toxicity screening of hazardous
waste sites. U.S. Environmental Protection Agency,
Corvallis, OR.
Hamilton, M.A., R.C. Russo, andR.V. Thurston. 1977. Trimmed
Spearman-Karber method for estimating median lethal
concentrations in toxicity bioassays. Environ. Sci. Tech.
11(7): 714-719.
Horning, W.B. and C.I. Weber. 1985. Short-term methods for
estimating the chronic toxicity of effluents and receiving
waters to freshwater organisms. EPA/600/4-85/014. U.S.
Environmental Protection Agency, Environmental Monitoring
and Support Laboratory, Cincinnati, OH.
Kapustka, L. 1989. Vegetation assessment. in W. Warren-
Hicks, B. Parkhurst, and S. Baker, Jr. (eds.). Ecological
assessment of hazardous waste sites. EPA/600/3-89/013.
U.S. Environmental Protection Agency, Environmental
Research Laboratory, Corvallis, OR.
Kapustka, L. and G. Linder. 1989. Hazardous waste site
1-88
-------
characterization utilizing in situ and laboratory
bioassessment methods. In Proceedings of "Midwest State
Pollution Control Biologists and Instream Biological
Monitoring & Criteria Workshop," 14-17 February 1989,
Chicago, Illinois.
LaPoint, T. and J. Fairchild. 1989. Aquatic surveys. In W.
Warren-Hicks, B. Parkhurst, and S. Baker, Jr. (eds.).
Ecological assessment of hazardous waste sites. EPA/600/3-
89/013. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR.
McBee, K. 1989. Field surveys: terrestrial vertebrates. In
W. Warren-Hicks, B. Parkhurst, and S. Baker, Jr. (eds.).
Ecological assessment of hazardous waste sites. EPA/600/3-
89/013. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR.
Murphy, T. and L. Kapustka. 1989. Capabilities and
limitations of approaches to in situ ecological evaluation.
In Proceedings of symposium on in situ evaluation of
biological hazards of environmental pollutants. Plenum
Press, New York.
Peltier, W.H., and C.I. Weber. 1985. Methods for measuring
the acute toxicity of effluents to freshwater and marine
organisms. Third Edition. EPA/600/4-85/013. US EPA,
Environmental Monitoring and Support Laboratory,
Cincinnati, Ohio.
'50 '
Stephan, C.E. 1977. Methods for calculating and LCE
Aquatic Toxicology and Hazard Evaluation, ASTM STP 634
(F.L. Mayer and M.L. Hamelink, eds.), American Society for
Testing and Materials, Philadelphia, PA.
Stephan, C.E., and J.W. Rogers. 1985. Advantages of using
regression analysis to calculate results of chronic
toxicity tests. In Aquatic Toxicology and Hazard
Assessment: Eighth Symposium. ASTM STP 891, R.C. Bohner
and D.J. Hansen, Eds., American Society for Testing and
Materials, Philadelphia, PA. pp. 328-338.
Warren-Hicks, W. , B. Parkhurst, and S. Baker, Jr. (eds.).
1989. Ecological assessment of hazardous waste sites.
EPA/600/3-89/013. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR.
Weber, C.I., W.B. Horning, D.J. Klemm, T.W. Neiheisel, P.A.
Lewis, E.L. Robinson, J. Menkedick, and F. Kessler. 1988.
Short-term methods for estimating the chronic toxicity of
effluents and receiving waters to marine and estuarine
organisms. EPA/600/4-87/028. US EPA, Environmental
Monitoring and Support Laboratory, Cincinnati, OH.
1-89
-------
US EPA. 1988a. Review of ecological risk assessment methods.
EPA/230/10-88/041. U.S. Environmental Protection Agency,
Office of Policy Planning and Evaluation, Washington, D.c.
US EPA. 1988b. GEO-EAS (Geostatistical Environmental
Assessment Software): User's Guide. EPA/600/4-88/033.
U.S. Environmental Protection Agency, Environmental
Monitoring and Systems Laboratory, Las Vegas, NV.
US EPA. 1989. Risk assessment guidance for Superfund—
environmental evaluation manual (Interim Final).
EPA/540/1-89/001A. U.S. Environmental Protection Agency,
Office of Emergency and Remedial Response, Washington, D.C.
1-90
-------
Appendix l. Short-term toxicity estimates (EC5Ds and LCcns) derived from tests (algal and daphnid)
completed on site-sample eluates (sorted by site locations, see also Figure 3).
SAMPLE SITE t
1
2
•*
4
5
6
7
8
9
1O
11
12
13
14
15
16
17
18
19
2O
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
NO SAMPLES
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
NO SAMPLES
surface
middle
deepest
surface
middle
deepest
NO SAMPLES
NO SAMPLES
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
NO SAMPLES
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
NO SAMPLES
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
NO SAMPLES
NO SAMPLES
surface
middle
deepest
surface
middle
deepest
surface
middle
deepest
ALGAL ECsn (95% C.I.) DAPHNIA
1.51
18. O
1.5
15.1
71.7
5.6
67.1
COLLECTEr
1.1
32.7
17.88
5.5
2O. 8
16.7
21.6
15.9
43.8
COLLECTED
6. 3
37. O
COLLECTED
COLLECTED
l.O
8.O
O. 4
32.9
18. O
66.2
COLLECTED
O.7
5O.4
4.5
31 .1
14.2
1.6
5.8
B.I
COLLECTED
33.9
41.9
7.6
10.2
11.4
2.4
6.6
O.7
72.1
e . e
COLLECTED
COLLECTED
71.7
4.3
11.1
43.8
7.2
7.9
NE
(O.7O-3.3)
(9.2-31.6)
NE
NE
NE
NE
NE
NE
(0.7-3.3)
(9.6-24.6)
(17.O-8O.O)
NE
NE
NE
NE
(4.2-7.6)
NE
NE
(64. 3-69. 8)
)
-4O.O
(0.3-3.9)
(22.9-47.9)
(O.O-74 .1)
NE
NE
NE
NE
(1.0-39.8)
NE
(4.5-80.0)
(6.3-8O.O)
NE
(11 .O-42.2)
(4.O-63 .0)
NE
(23 .1-8O.O)
NE
NE
(1.1-35.5)
NE
NE
(O.7-8O.O)
NE
NE
NE
NE
(O.O3-39. 8)
(1.6-39.8)
(O. O1-20. O)
(23. 4-46. O)
(1O.7-31.6)
(25.1-8O.O)
(0.2-3.2)
(16.2-8O.O)
NE
NE
NE
NE
(O.O1-8O.O)
(5.O-8O.O)
NE
(5.6-22.4 )
(O.6-4.4)
1.7-19.7)
NE
(2.6-25.1)
NE
(0.0-80.0)
(28 .5-61.7)
(2-9-18.2)
(3.9-26.3)
(3.2-39.8)
(O.3-2O-O)
NE
(2.3-19.1)
(O.O4-14.1)
(43.6-8O.O)
NE
(2.5-30.2)
NE
NE
NE
(66.8-76.9)
NE
(1 .5-12.6)
NE
(5.8-2O.9)
(3O.3-61.7)
NE
(1 .O-51 .9)
(2.5-21.9)
67.
91 .
42.
8.
55.
66.
84 .
49.
16.
43 .
SO.
95.
SO.
89.
92.
87.
72.
2.
22.
46.
91.
45.
39.
67 .
21.
22.
2.
6.
4 .
2.
93.
43.
8 .
97.
21.
5.
ECRn (95% C
1
2
8
3
2
2
3
0
6
9
5
9
7
8
8
4
9
2
O
6
4
O
3
9
7
0
2
8
4
9
9
O
6
1
4
4
NE
(58 .4-77.
(87. 3-95.
NE
NE
NE
NE
(38.0-48.
.1. )
1)
3)
1)
(3.7-18.9)
(46.7-65.
(62.6-70.
NE
NE
-5O.O
(8O.4-88.
NE
(37.2-64.
NE
NE
NE
(1O.4-26.
(36.5-52.
(76.O-85.
NE
NE
NE
NE
NE
-30.0
NE
-16.7
NE
NE
NE
2)
°)
4)
6>
4O
9)
3)
(9O.2-1OO.O)
NE
NE
NE
NE
-20.0
NE
NE
NE
NE
(45.3-55.
NE
NE
(88.6-91.
(89.6-96.
(82.7-92.
(63.8-83.
(1 .8-2. 8)
NE
(18.9-25.
-53.0
-4O.O
NE
NE
NE
NE
(41.1-52.
NE
(89.4-93.
( 3O. 6-66.
NE
NE
NE
NE
(33.9-45.
(64.4-71 .
(17.9-26.
(18.1-26.
(1.8-2.7)
(5.6-8.4)
NE
(3.4-5.7)
(2.0-4.2)
(9O.O-97.
NE
(39.9-46.
NE
NE
NE
NE
NE
3)
1)
1)
4)
2)
4)
8)
5)
1)
4)
7)
2)
7)
3)
4)
(7.2-1O. 3)
(97.1-1OO
(17.O-26.
(4.1-7.1)
NE
-1O.O
NE
)
8)
1-91
-------
SAMPLE SITE » ALGAL EC50 (95% C. J. . ) DAPHNIA EC50 (95% C.I.)
40 surface 31.2 fS.2-8O.O) NE
middle 74.8 (9.9-8O.O) NE
deepest NE NE
41 HO SAMPLES COLLECTED
42 NO SAMPLES COLLECTED
43 NO SAMPLES COLLECTED
44 surface NE NE
middle 77.9 (18.2-8O.O) -47.0
deepest 46.5 (O-O-8O.O) NE
45 surface 37.9 (3.7-8O.O) NE
middle 73.5 (6O.6-8O.O) 92.8 (86.4-99-7)
deepest 65.3 (2O.8-8O.O) 55. O (47.7-63.3)
46 surface NE NE
middle NE 79.6 (74.3-85.2)
deepest NE 75.3 (67.5-84.1)
47 surface 29.O (O.O-8O.O) NE
Diddle 12.0 (2.9-5O.O) 51.7 (42.6-62-7)
deepest 56.1 (1-9-8O.O) 72.8 (58.8-9O.1)
48 surface 8.6 (2.5-29.5) 71.2 (5O.O-1OO.O)
middle 19.9 (7.6-52.5) 33.7 (27.8-4O.8)
deepest -SO.O 9O.2 (84.9-95.8)
49 surface 7.5 (3.1-17.4) -4O.O
middle 50.3 (29.0-87.1) 71.1 (63.5-79.5)
deepest 5.O (2.5-1O.O) 87.8 (85.O-9O.7)
SO surface NE NE
middle 17.O (4.3-67.6) NE
deepest 11.7 (6.5-2O.4) NE
51 surface NE NE
middle -4O.O NE
deepest NE -2O. O
1 Peports listed as negative percents reflect mortality or extent of Inhibition observed at highest
concentration tested (algae, 8O%; Daphnia, 1OO%). 95 % C.I. = 95% confidence interval about point
estimate.
Tf EC5D estimates are extrapolations and not interpolations, projections exceed highest (or lowest)
concentrations tested and no precision estimates are possible; median effect estimates should be
evaluated as relative measures.
3 NE no effect.
1-92
-------
•i^J
4400C
44001
44002
44100
44101
44102
44103
44104
4410!
44106
44107
44108
45101
45113
45114
45119
46000
46001
46002
46003
46004
46005
44007
44008
44009
44010
44011
44012
46013
4*314
46015
4Wli
4oGl*
4*)!5
4601?
44C25
4602:
46C22
46023
44024
46026
46027
46029
46030
46100
47000
47002
47006
W
LE
LE
LE
DE
DE
DE
OE
OE
K
DE
DE
OE
DE
OE
OE
DE
LE
LE
LE
LE
LE
LE
LE
LE
LE
LE
if
LC
LE
LE
LE
L£
;_£
: c
t~i
L£
: r
LS
L£
LE
LE
LE
LE
LE
LE
LE
DE
LE
LE
LE
COM. i
(uil
1925.3
471.0
860.0
2475.0
2550.0
2470.0
3150.0
2900.0
2770.0
2650.0
2900.0
1350.0
978.0
630.0
1890.0
842.0
333.0
1115.0
448.0
800.0
372.0
448.0
130.5
207.0
103.5
124.5
705.0
272.0
538.3
520.0
3:4.3
473.3
145.3
630.3
375.3
1700.3
1370.0
219.3
144.0
376.0
1860.3
375.0
1285.0
1375.0
374.0
439.0
740.0
497.0
iALMITY
(pot)
1.271
0.314
0.568
1.634
1.683
l.oJO
2.079
1.914
1.328
1.749
1.914
2.211
0.645
0.414
1.247
0.549
0.223
0.734
0.296
0.523
0.246
0.296
0.086
0.137
0.068
0.382
0. 465
0.180
0.355
0.343
0.214
o.:::
Q.395
0.41o
0.248
i.:54
0.904
0.145
0.09!
0.243
1.228
0.578
0.848
0.908
0.247
0.290
0.488
0.328
TOC
(pptl
256.40
8.10
24.40
31.30
116.00
105. aO
19.70
88.20
53.40
49.90
135.70
136.80
81.20
51. '0
37.30
19.10
9.10
10.00
8.19
24.60
12.70
11.30
4.40
11.30
6.40
2.70
23.70
4.60
11.30
10.30
11.30
-
-------
APPLICATION OF MICROBIAL TOXICTTY AND MUTAGENICITY
ASSAYS FOR IDENTinCATION AND EVALUATION OF TOXIC
CONSTITUENTS IN FRACTIONATED HAZARDOUS WASTES.
B S SHANE, K.C. DONNELLY, E.B. OVERTON, T.R. IRVIN, L. BUTLER,
J NORCERINO AND J. PETTY. INSTITUTE FOR ENVIRONMENTAL STUDIES,
LOUISIANA STATE UNIVERSITY, BATON ROUGE, LA (BSS, EBO),
ENGINEERING TOXICOLOGY DIVISION, DEPARTMENTS OF CHEMICAL
ENGINEERING AND VETERINARY ANATOMY, TEXAS A&M UNIVERSITY,
COLLEGE STATION, TX (KCD, TRI) AND ENVIRONMENTAL MONITORING
LABORATORY, EPA, LAS VEGAS, NV (LB, JN, JP).
Current monitoring methods for hazardous waste site chemicals present
numerous analytical problems due to (1) matrix interferences and (2) the complexity of
the chemical mixtures present We have employed microbial toxicity and mutagenicity
assays to rapidly identify toxic fractions and components of hazardous wastes and
individual chemical constituents representing the predominant toxic species in each
mixture. In this approach, extracts of environmental samples are fractionated via
normal-phase or gel permeation chromatography and coincubated in microbial bioassays
(the Salmonel 1 a/microsome assay). The resulting toxic or mutagenic responses
(produced in 24-48 hours) are used to define the toxicity of whole or fractionated waste
samples. This approach was utilized for the analysis of polycyclic aromatic
hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), and polychlorinated dioxins
(PCDDs) in three waste environments: municipal fly ash, estuary effluents, and soil
amended with wood-preserving waste. Fly ash samples were extracted with
dichloromethane for 24 hours using Soxhlet extraction, and the isolated organics were
evaporated to dryness and reconstituted in dimethyl sulfoxide. Estuary samples were air
dried and extracted at room temperature with acetonitrile; exchanged into pentane and
fractionated on a silica gel column. Fractions were evaporated to dryness with N2 and
dissolved in DMSO. In all three case studies, close correlation was found between the
identification of known classes of toxic chemicals and the toxic and mutagenic potency
of these wastes. Microbes were found to be sensitive to all major classes of toxicants
when tester bacteria were exposed to mixed wastes in the presence of a rodent liver
preparation (S-9) competent to biotransform polycyclic compounds to toxic/mutagenic
intermediates. Fly ash-associated mutagens, predominantly benzo(a)pyrene,
fluoranthene and phenanthrene, elicited toxic and mutagenic effects in the presence and
absence of rodent S-9 fractions. Estuary and wood-preserving wastes, containing
polycyclic aromatic and chlorinated hydrocarbons, exhibited toxic effects only in
cultures supplemented with rodent S-9. Data to be presented supports this approach as
an improved method for rapid identification of toxic constituents in uncharacterized
wastes as well as prioritization of wastes sites for remediation.
1-94
-------
APPLICATION OF MAMMALIAN CELL CULTURE SYSTEMS
TO EVALUATE AND MONITOR HAZARDOUS WASTES AND WASTE SITES.
T.R. IRVIN, I.E. MARTIN, B.S. SHANE, L. BUTLER, N. NORCERINGO, J.
PETTY AND E.B. OVERTON. DIVISION OF ENGINEERING TOXICOLOGY,
TEXAS A&M UNIVERSITY (TRI, JEM), ENVIRONMENTAL MONITORING
LOABORATORY, EPA, LAS VEGAS, NV (LB, JN, JP) AND INSTITUTE FOR
ENVIRONMENTAL STUDIES, LSU, BATON ROUGE, LA (BSS, EBO).
Current strategies targeted to evaluate hazardous wastes and design hazardous
waste site remediation strategies have traditionally focussed on ecotoxicological effects
in the absence of methods to determine the fate and effects of waste site mixtures on
human populations. While microbial methods have been advanced to assess the
genotoxicity of waste site chemicals, methods have not been applied to monitor other
human toxic endpoints, such as prenatal toxicity, neurotoxicity, and reproductive
toxicity, characteristic of hazardous wastes constituents. We have adapted mammalian
cell culture systems to monitor these toxic effects for uncharacterized chemical waste
mixtures, direct hazardous wastes, and direct environmental samples (soil, water, air
particulates) from hazardous waste sites:- these include: (i) prenatal toxicity (via
postimplantation rodent embryo culture (ii) neurotoxicity (via continuous glial cell
cultures), (iii) reproductive toxicity (via rodent blastocyst culture), as well as (iv)
carcinogenicity (via clastogenic analysis of primary human and rodent cell cultures) of
samples from hazardous waste sites. Using this approach, we have characterized the
toxic effects of complex waste mixtures (petroleum creosote, phenolic wood-preserving
wastes, aromatic and chlorinated hydrocarbon solvent mixtures) to prioritize waste sites
in terms of acute and chronic health effects. In addition, we have also employed these
assays to evaluate various hazardous waste clean-up strategies (e.g. extraction,
biodegradation) to identify conditions which optimize the removal of the key toxic
constituents of waste site chemical mixtures.
1-95
-------
SCREENING FDR TOLYCHD3RINATED DIOXINS AND FURANS BY IMMUNOASSAY
MARTIN VANDERIAAN, IARRY STANKER, AND BRUCE WATKINS, BICMEDICAL SCIENCES
DIVISICN, IAWRENCE LTVERMDRE NATIONAL LABORATORY, LTVERMDRE, CA 94550
ABSTRACT. We have reported (Toxicology 4J5:229) a set of monoclonal ant
that bind preferentially to polychlorinated dioxins and furans having c
in the lateral positions. Mice were immunized with a l-adipamino-3,7,8
trichlorodibenzodioxin, and the resulting monoclonal antibodies recogni
highly toxic 2,3,7,8-tetrachloro-dioxin (TCDD) and 1,2,3,7,8-pentachlor
and furans, and a limited number of other intermediately chlorinated cc
With this binding selectivity the antibodies are suitable for screening
for the presence of a subset of the recognized congeners, but GC/MS ana
needed to confirm the exact congeners present in positive samples. An
linked immunosorbant assay (ELISA) incorporating one of these antibodie
been developed, with a detection limit of about 100 pg of 2,3,7,8-TCDD.
assay has been successfully applied to the analysis of industrial chemi
PCBs, fly ash, and soil samples contaminated with TCDD at 1 ppb and abo
While samples extraction and preparation was required for the immunoass
sample clean-up was less extensive than that needed for GC/MS analysis.
Immunoassay results correlated with GC/MS analysis of the same samples,
encouraging us to pursue the use of immunoassays as a screen for contarc
Work is currently in progress to expand the range of matrices tested to
eggs, milk, and animal fats.
There is regulatory concern about the presence of hexa- and hepta- chlo
dioxins, which are not recognized by the current set of antibodies. A
we are developing a new set of monoclonal antibodies directed to these
chlorinated congeners. We have synthesized 2-carboxymethyl-3,6(9) ,7,8-
tetrachlorodibenzo-dioxin for use as a hapten, mice have been immunized
production of new monoclonal antibodies is now in progress. It is anti
that these new antibodies will show preference for more highly chlorina
congeners. It is also expected that the new hapten should produce anti
with higher affinity for all laterally substituted dioxins. Data on th
binding selectivity of the new monoclonal antibodies should be availabl
the next few months to confirm or refute these expectations. When used
conjunction with the existing antibodies these new monoclonal antibodie
provide a screen that recognizes all of the most toxic congeners of die
furans.
1-96
-------
THE USE OF SCREENING EEOTOOOIS
TO EVALUATE
KEOREMEDIATTON TECHNOLOGY FOR SITE CLEANUP
JOHN A. GLASER, U.S. ENVIRONMENTAL PROTECTION AGENCY, RISK REDUCTION ENGINEERING
LABORATORY, 26 W. MARTIN LUTHER KING DRIVE, CINCINNATI, OH 45268
The selection of suitable technology for hazardous
waste site cleanup is complicated by the lack of
objective performance criteria for many treatment
options. Where there has been an extended period of
development activity, the technology is recognized for
its more advanced.state. Consequently, the treatment
technologies identified as having less development
activity are regarded as incomplete and do not receive
recognition as competitive technology.
Bioremediation technology has not been left unscathed
in this comparison. In spite of its promise and
recognized beneficial effects it has been generally
avoided as a remediation technology. A major issue for
the acceptance of bioremediation is the establishment
of an objective means whereby prospective users can
differentiate between competitive and conflicting
claims. This paper presents the first of a proposed
series of protocols devoted to the formulation of an
objective criterion to measure biological treatment.
The initial protocol is devoted to the assessment of
aerobic biological treatment of contaminated soils. The
protocol has been organized to permit the adaptation of
the testing to configurations closely mimicing large
scale operations.
1-97
-------
ENFORCEMENT
-------
COMPUTERS IN THE DECISION PROCESS:
LEGAL IMPLICATIONS OF ELECTRONIC
DATA TRANSFER AND DATA MANAGEMENT SYSTEMS
JEFFREY C. WDRTHINGTGN, R. PARK HANEY, TECHIAW, INC., 12600 W. COLFAX C-310,
LAKEWOOD, CO 80215
ABSTRACT.
The use of computers and electronic information poses a
complex problem for potential litigation. The problem
currently manifests itself in at least two ways.
First, the EPA enforcement of CERCLA/SARA statutes is moving
quickly towards site clean-up activities that will require
responsible decisions based on quick turn-around analytical
data, site managers increasingly rely on direct data
transfer or FAX information while making crucial decisions
concerning both the scope (and expense) of clean-up activity
and potential liability for the site owners or operators.
Second, many laboratories in both the public and private
sectors are either developing their own data management
tracking systems or purchasing data management software
packages. These trends are likely to continue.
Insufficient documentation of electronic transfers may
render any action or decision "non-defensible." While draft
data in an electronic system may enable the end-user (i.e.,
On-Scene Coordinator) to make a necessary decision quickly,
that information may later be lost as documented support for
the decision because it is deleted as a "final" version of
the data is prepared. Hard copies of some data transfers
may never be produced. Some laboratory personnel are using
their data management tracking systems as replacements for
hand-written records of laboratory activities.
Also at issue are the evidentiary considerations related to
computer-generated data. How does the best evidence rule
apply to electronically transferred data? How is the issue
of admissability, more specifically authenticity, dealt with
for computer-generated data?
A solution to these considerations requires constant
monitoring by quality assurance personnel for successful
implementation. Documentation of data transfers is
essential; hard-copies of all data should be produced and
filed by the data-generator. Computer usage in the
laboratory in lieu of hand-written documents requires that a
record be computer-generated in a timely manner (i.e., the
same day) and signed and dated by the individual involved.
All users of field and laboratory data systems should
develop and maintain accurate Standard Operating Procedures
(SOPs) for this quality assurance procedure.
1-99
-------
A PLANNING TOOL FOR SITE MANAGERS:
HISTORICAL PERSPECTIVE ON LITIGATION USES OF
SAMPLE DATA FOR EPA CERCIA/SARA CASES
CYNTHIA MILLER, JEFFREY C. WCRffllNSKN, TECHLAW, INC., 12600 W. COLFAX 0-310,
LAKEWOOD, CO 80215
ABSTRACT.
Data quality objectives are defined in terms of
completeness, precision, accuracy, representativeness, and
comparibility. The "usability" of data has been
traditionally expressed in terms of the end-users
responsible for conducting a site study. A second end-user
may be a local, state, or Federal litigation team. This
second end-user is often not considered in planning for
individual site studies or in the management of large
analytical programs.
The needs of this second end-user go beyond custody and
document control considerations. Data quality and
implementation of written protocols by both field and
laboratory personnel are being attacked more frequently
during settlement negotiations and in the courtroom. Site
owners and operators are employing highly sophisticated
defense stratagies, sometimes based on criticism of non-
conformance by government agencies when following their own
procedures.
A review of the uses of data in the litigation process
during the last eight years can be a valuable planning tool
for both site managers and program managers. An awareness
of this secondary use could facilitate future litigation and
substantially reduce government costs. The authors will
present a review of the historical uses of data in
CERCLA/SARA cases during the last 8 years. Each component
of the sampling and analytical process will be reviewed.
This review will enable managers to accurately assesss their
data quality objectives as well as determine the sampling
and analytical protocols that have been most useful to this
second end-user.
1-100
-------
EXAMPLES OF THE USE OF AN ADVANCED MASS SPEC-
TROMETRIC DATA PROCESSING ENVIRONMENT FOR THE
DETERMINATION OF SOURCES OF WASTES
B. Mason Hughes. David E. McKenzie, Chi K. Trang and La Shawn R. Minor,
Environmental Sciences Center, Monsanto Company, 800 N. Lindbergh Boulevard,
St. Louis, Missouri 63167
ABSTRACT SUMMARY OF THE APPROACH
One of the most important and complex One important element in the comparison
issues facing the cleanup of Superfund Sites of chromatograms and mass spectra is
is "who is responsible for site contamination that, while these two types of data are quite
and what percentage of the cleanup costs different, they can be treated in analogous
should be prorated among the Potentially ways. In the present paper, examples are
Responsible Parties (PRPs) ". This paper de- given which show how chromatograms can
scribes one approach which has recently be compared using the same software and
been used to answer these questions. This approaches that have been developed for
method uses a state-of-the-art mass spectra- the comparison of mass spectra. Just as
metric data processing environment to pro- histograms of mass intensities (mass spec-
duce visual and quantitative comparisons of tra) can be used to identify what molecule
chromatographic and mass spectrometric is present in a complex mixture, so can
patterns in an attempt to determine the origin histograms of chromatographic peak ar-
of wastes present that produced these pat- eas (reduced chromatograms) be used to
terns. identify the likely source from which the
mixture of compounds originated. The fol-
lowing sections show three examples
where this approach has been used to de-
ESfTRQDUCTION termine: 1) the types of wastes at a Super-
fund Site, 2) the types of wastes in environ-
Routine capillary GC/MS instrumentation mental samples adjacent to the site, and 3)
and analytical methods using this instru- the similarity of different gasolines which
mentation has resulted in the collection of may be present from leaking underground
orders of magnitude more data than can storage tanks.
easily be processed. The Priority Pollutant
analysis protocols are, in effect, efforts to This approach could often be used in the
simplify and summarize complex mixture place of Priority Pollutant analyses, when
information by searching for specific tar- samples of suspected sources of wastes are
get compounds. However, for many Su- available and when rigid protocols are not
perfund Sites, Priority Pollutant charac- required for the quantification and identi-
terization gives an incomplete picture of fication of certain target compounds. Sig-
the site, because Priority Pollutants may nificant cost savings can be gained using
not be present. The present paper de- this approach since the unequivocal identi-
scribes a general approach that requires: fication of all components is not required
1) the waste components are chromato- to identify a waste source with high cer-
graphable, and 2) suspected sources of tainty. Indirect cost savings are also real-
these wastes are available for analysis. ized, since the extraction and analysis
1-101
-------
Mass Spectral/Gas Chromatoqraphic Analogies
Analogue Analogue
Mass Spectrum < = > Chromatogram
Mass Spectrum < = > Reduced Chromatogram
Compound Producing Formulation Producing
the Mass Spectrum < = > Reduced Chromatogram
methods are rapid, and interpretation
depends upon computer generated gra-
phical output. Field operations can be
supported overnight with these extraction,
analysis and interpretation protocols.
AND ANA-
INSTRUMENTATION
LYTICAL SYSTEMS
All data shown in the present paper were
obtained using Hewlett-Packard's RTE-
VI mass spectrometric data processing en-
vironment. This data processing environ-
ment is in effect an array processing sys-
tem. This system allows for arrays of mass
spectrometric and/or gas chromatogra-
phic data to be added, subtracted, multi-
plied, and divided in much the same way as
single numbers are manipulated using a
hand calculator. One important advan-
tage of this environment is that individual
commands or programs which are to be
executed can be combined in an ASCII file
and executed in much the same way as
batch or macro files can be processed in
personal computers. Several batch proce-
dure files were developed to produce the
various displays shown in the present pa-
per. These procedure files were also writ-
ten to create HPGL-compatible files that
could be transferred to a PC-compatible
personal computer which was used to
produce the various graphic displays on a
300 dots-per-inch laser printer.
The compatibility of these mass spectro-
metric data reduction programs with the
file structure of Hewlett-Packard's Labo-
ratory Automation System GC/FID and
GC/EC raw data files, results in the use of
the same batch procedure files for com-
paring chromatographic data not obtained
using a mass spectrometer detector sys-
tem. This capability allows for PCB formu-
lations and other complex formulations to
be easily identified in much the same way
as sources of wastes are determined.
SAMPLE EXTRACTION
The present analysis protocol was de-
signed so that chromatographic patterns
could be used as important indications as
to the source of unknown wastes. There-
fore great care was used so that sample
manipulation would not greatly modify the
pattern of organics present. In addition,
these methods were developed for the
relatively high level analysis of organic
components present in wastes and envi-
ronmental samples adjacent to waste sites.
Samples in these studies contained total
organic concentrations from 0.01 to 100%.
Therefore a very simple extraction proto-
col was required. Sample extraction in-
volved placing a known amount of the
waste or contaminated soil in a 14-mL glass
vial with teflon-lined lid, adding 10 mLs
methylene chloride, and ultrasonically
agitating for one-half hour. In most cases,
the concentration of individual chromato-
graphable components were in the 1 - 500
Mg/mL concentration range of the result-
ing extract. Since very little sample ma-
nipulation was required, no surrogate
standards were used. This resulted in the
simplification of the chromatographic
patterns. Further simplification occurred
1-102
-------
with the use of only one internal standard
(anthracene-d10) for extractable compo-
nent analysis, and three volatile internal
standards for the purgeables analysis.
THE CROSS CORRELATION COEF-
FICIENT
One of the early ways of comparing two
mass spectra which may contain between
10 and > 100 mass/intensity pairs of infor-
mation, was to calculate a cross correlation
coefficient (XCC). This was one of the
early ways of comparing unknown spectra
with a library of spectra. However, this
calculation, by itself, did not contain
enough information nor could it reliably be
used for mass spectra generated on various
mass spectrometers. Figure 1 shows the
XCC formula and how it can be used to
compare two mass spectra of a n-C16-
alkane present in two different samples.
Xj and YJ are the individual intensities of
the i-th masses in the X- and Y-arrays, and
X and Y are the average of the intensities
in these two arrays. Note that when the
differences between the individual intensi-
ties and the average intensity of the two
arrays approach each other, the value of
the XCC approaches 1.000. Therefore, to
a first approximation, the deviation of the
XCC from 1.000 can be used as a simple
way of determining the deviation of one
spectrum from another.
This approach is particularly useful be-
cause both spectra are taken under almost
identical conditions. Furthermore, the re-
tention times of the two components pro-
ducing the mass spectra can also be used as
an additional parameter in evaluating
whether the spectra are being produced
B000~
4000~
0
41
1,
GC/
57
.1,
MS FILE
71 6
, nil
= >L0760 [Mixture Component Spectrum for 17.25 Minutes
5
' — — _ Y-ARRAY
___ -gg ~"~~ Y _,2B
/ 127 141 155 IBB 184 \.
1 1 l 1 ' ' \ 193
1 , ill , ,111. . , ll, ll 1 .III 1. 1 X. .'
60 80 100 120 140 160 ISO 200
Calculation c
XCl
aooo"
40OO"
0
41
57
, Jl
220
f Cross Correlation Coefficients (XCCs):
, SiX,- I XT, - Y)
VZ (Xt -X)2 VziY, - Y)2
71 6
...III
5
.--
. Ill
- — —(X.-X) X-ARRAY
" " X
39
' I*3 ,o,
/ 127 141 155 iBg 1B3 19^
L , iilll . . , ill . . . ill. . . .lli . ll. ii i
60 80 10O 120 140 160 180 2OO
GC/MS FILE = >L0750 [Reference Component Spectrum for 17.28 Minutes
226
220
MASS SPECTRAL COMPARISON
CROSS CORRELATION COEFFICIENT = .931 (ALL PEAKS)
CROSS CORRELATION COEFFICIENT = .994 (ONLY REFERENCE PEAKS)
Figure 1. The Cross Correlation Coefficient Calculation Applied to Mass Spectra.
1-103
-------
fmm the same compound In Figure 1, it trometric data. This is done in Hewlett-
canb*eenratTPYarraySpectrumwas Packard's data processing environment by
obtained for a component having a reten- converting the analogue chromatogram to
tion time of 17 25 minutes and th! X-array a reduced chromatogram. This reduced
pec mm was obtained from a component chromatogram is no more than a histo-
STretention time of 17.28 minutes, gram display of the chromatographic (re-
Therefore the similarity of compounds as tention tune)/(peak area) pairs which is
rted from the XCC being Q.991, and exactly analogous to the mass spectrum
from almost identical retention times, re- histogram display of mass/intensity pairs.
su°tT in the conclusion that the two com- Figure 2 shows how the XCC is calculated
nnmids are identical for reduced chromatograms produced
pounds are identical. from chromatographic data. One impor-
One important property of the XCC calcu- tant difference between the reduced chro-
lation is that the correlation can be calcu- matograms and mass spectra is that mass
lated in a number of ways. As is shown in spectra have built-in digitizing features
FiRure 1 the XCC is calculated for "all since the x-axis is composed of integer
peaks" and for "only reference peaks", masses. This feature simplifies identifying
This is done by being able to identify a the features in the X- and Y-arrays which
window that is applied to all the data in the are the same and which shouldbe included
X- and Y-array In the present case, a in the XCC calculation. However, as
window of ±0 3 amu is used to determine mentioned above, since a mass window can
whether the masses in the two arrays be defined which is used to determine
should be treated as the same feature. All which mass features are the same, so can
mass/intensity information is included in the retention time window be used tor
the XCC for "all peaks". However, just exactly the same purpose. For the present
those masses in the Y-array which are study, a retention time range of ±4.8 see-
within the windows of the masses in the X- onds was used. This is very important since
array are included in the XCC calculated for chromatographic data, there are no
for "only reference peaks". These two integral parameters such as mass for the x-
calculations are used to determine axis parameter. The continuous property
whether the reference component spec- of chromatographic retention times means
trum (shown in the X-array) may be im- that comparisons can be properly made
bedded in the mixture component spec- only when the mixture and referencechro-
trum (shown in the Y-array). In the pres- matograms are obtained under identical
ent example, since both XCCs are almost conditions. Approaches have been re-
identical and almost 1.000, it can be con- ported in the literature which involves
eluded that over 95% of the mixture com- converting chromatographic retention
ponent spectrum is produced from the times to indices, which may help when
reference component. Examples will be comparing data obtained under different
shown later where this distinction is impor- conditions or from different laboratories.
tant in searching for spectra of reference The strength of the present approach is
components which may be imbedded in that there is no need for this conversion
spectra in a mixture of components. since the XCC calculation of chromatogra-
phic data includes the windowing feature
APPLYING THE XCC TO CHROMA- for determining whether chromatographic
TOGRAPHIC DATA peaks in two different samples are the
same feature.
In order to use the XCC calculation as a
way of comparing chromatographic data, In Figure 2, the reduced chromatogram
the chromatographic data must be pre- features identified as X^ and Yt are the two
sented in much the same way as mass spec- features for which mass spectra are com-
1-104
-------
GC
8000"
4000"
/MS RLE - >L0760 ; TOTA1
. ION ( 35 - 500 omu) UNKNOWN MIXTURE CHROMATOGRAM 1
---
UfLA^^wifi
17.0 18. O 19. O 20.0 21. O
y
^V~A_/v |U-f»~A — ,jJ^-~-A|wJ_^^J__-_~J, 1 ^
22.0 23.0 24.0 25.0 26.0 27.0 2B.O 29.0 30 . 0
Calculation of Cross Correlation Coefficients (XCCs):
™ * ft - *X*i - ?)
8000^
4000"
0
17
GC
Vz ft
\
^-jJLiJ
— — -
.0 18.O 19.0 20.0 21.0
-X)2 Vz ft - Yf
— — - /y yi X~ ARRAY
— ' i '
I
22 . 0 23 . 0 24.0 25 . 0 26 . 0 27.0 28 . 0 29 . 0 30 . 0
AIS FILE - >L0750 ; TOTAL ION ( 35 - 500 omu) REFERENCE MIXTURE CHROMATOGRAM |
REDUCED CHROMATOGRAIS COMPARISON
CROSS CORRELATION COEFFICIENT = .868 (ALL PEAKS)
CROSS CORRELATION COEFFICIENT = .911 (ONLY REFERENCE PEAKS)
Figure 2. The Cross Correlation Coefficient Calculation Applied to Reduced
Chromatograms.
pared in Figure 1. The comparison of
chromatograms in combination with the
comparison of mass spectra of each of the
chromatographicfeatures, is the basis for a
very complete and highly reh'able compari-
son of two samples. The following sections
show how these comparisons can be used
to efficiently and cost effectively answer
some difficult questions concerning com-
plex mixtures.
IDENTIFYING SOURCES OF
WASTES AT A SUPERFUND SITE
The application of this approach to the
study of wastes present at a Superfund Site
can be seen in Figures 3-8. Figure 3 shows
the range of compounds eluting between
n-C9-alkanes and n-C26-alkanes. Figure 6
shows the n-C16-alkane through n-C18-
alkane region and Figure 8 shows the n-
Cll-alkane through n-C17-alkane region.
The reduced chromatogram XCC calcula-
tion for "all peaks" shows that a large
number of Crude Oil Sample components
are present in the Waste Sample. Detailed
comparisons of mass spectra of the major
and minor crude oil components give mass
spectral XCCs on the order of 1.000. Fig-
ure 8 shows an example where an imbed-
ded chromatographic pattern of the Sty-
rene Waste Sample is also present in the
Waste Sample chromatogram. The re-
duced chromatogram XCCs of 0.305 for
"all peaks" and 0.874 for "only reference
peaks" shows that the Styrene Waste
Sample pattern is present in minor quanti-
ties. XCC comparisons of mass spectra of
the major and minor Styrene Waste Com-
ponents are on the order of 1.000. Figures
1-105
-------
GCAIS FILE - >LD760 : TOTAL ION ( 35 - 500 omu)
Was
e Sample
a 10 12 14 16 18 20 22 24 26 28 30
4000'
0
-4000
-BOOO
Ay^>.~4*
4
10 12 14 16 IS 20 22 24 26 S3 30
BOOO:
Cruce
^JMW^^il^^
Sample
A 6 8 10 12 14 16 18 20 22 24 26 28 30
FIE - XJJ790 ; TOTAL ION ( 35 - 500 omu)
REFERENCE MIXTURE CHROMMDCRAM
----- IHH7CXD CHH01UTOCJUM COMPiflEON STORED IN HLK AD0750
CROSS CORRELATION COBTKIENT - .702 (ALL PEAKS)
CROSS CORRELATION COEFFICIENT - .814 (ONLY REFERENCE PEWS)
Figure 3. Comparison of Chromatographic Patterns of a Superfund Site Waste
Sample and a Crude Oil Sample.
Component
Crude Oil Component
<*r 16J1 tfiuta
• ftl H
•D (atr
*OOO'
JOD D'
OCAB Fl£ - >LD7BO Mdun Cwnponwl SpKbum (™- 30.11 Unuta
^ " 1 ./ Waste Component
_^l, l|l i.l . Ilil J|l. .,, 1 ^ T, ,(. ^ i
^' " « » ,„ 1 .., -i T T '- "i-
, ,j ,.. i..' 'y „/ ,iJj / ,i iu iiv * i
» «o .«. ,=o ,;„ ' ,i, ' ,i ' m »
„ r 1 j' y Crude Oil Component
LyJ.ll,! JT T r^Vr'
OCVUS FU - >UJ7SO ^tonnc* Cempnuri Spwtnrl ter 2UJ UnjtM
MIH *KT»U. comix* mm e* «ii ii0»«
CWM coMunx coorcfid - jot (*a «*ffl
aosa comoxnM COPTCOIT - .in (BCT KFBVWX HMO)
Figure 4. Comparison of Pristane Mass Figure 5. Comparison of Phytane
bpectra in Superfund Site Waste and in Mass Spectra in Superfund Site Waste
Cmde O'l- and in Crude Oil.
1-106
-------
GC/MS FILE - >LD760 ; TOTAL ION ( 35 - 500 omu)
UNKNOWN MIXTURE CHROMATOGRAM
Or
Waste Sample
17.a 17.6 la.O 18.4 18.B 19.2 19.6 20.0 20.4
4000
0
-4000
-8000
' ' I ' ' ' I ' ' ' I ' '
17.2 17.6
18.4
1B.B
' ' I ' ' ' I ' • ' I ' ' ' I • ' ' I
19.6 20.0 20.4
Or
17.2 17.6 18.0 18.4
CC/MS FIE - X0750 ; TOTAL KM ( 35 - 500 omu)
te Oil Sampl
REFERENCE MIXTURE CHROtMTOGRAM
SEDUCED CHROJUTOGHAK COKPARTSON STOKED Hi fILK AH) 760 •
CROSS CORRELATION COEFFICIENT •
CROSS CORRELATION COEFFICIENT-
.86) (ALL PEAKS)
.910 (ONLY REFERENCE PEAKS)
Figure 6. Comparison of a Narrow Region of Chromatographic Patterns of a
Superfund Site Waste Sample and a Crude Oil Sample.
4 and 5 show the comparison of mass spec-
tra of pristane and phytane which are
crude oil biomarkers and Figure 7 shows
the comparison of naphthalene which is
the major chromatographable component
present in the Styrene Waste Sample.
Conclusions: A major source of
contamination at this Superfund Site has
its origins from the petroleum industry. As
a result of this conclusion, extensive inter-
views with waste haulers in the region
confirmed this fact. Ultimately, Clean
Sites Incorporated reviewed this data and
confirmed our findings which resulted in a
portion of the remediation costs being paid
by Petroleum Industry PRPs.
MOD
40GO
o-
D
1 accr
F
F
E
B
e -
N
E "at™
BWXT
OVUS FIE - >L07«] *0dv. Con^onnl SpKlrwn far 9JT7 Umt«
Waste Component ia^
39 7* "\
'l' l'l ^ l'|!l •'- '' '
^jg 76^
" » « ™ »
« ' IM ' M» ' ,» '
' ' ?:' '"I
90 100 1,0 .«
Styrene Waste Component 'ai-
r - /:.../ 7 '" ,n
GCAB RLE - >UJ73» *Werw*» Cwnponent Spectnm for" BJs'uVwt-
uisaencau, a
CROSS OTMELOMH COf
unisaM srorai m TOM uarx
7KXHT - M6(jll. PVXS)
TOtm - 1.003 fOMY fKTCTEWCE PE«S?
Figure 7. Comparison of Naphthalene
Mass Spectra in Superfund Site Waste
and in Styrene Waste.
1-107
-------
aVWS RLE - >LD760 ; TOTAL DM ( 35 - 500 omu)
Waste Sample
Styrene Waste Sample
B.O 9.0 10.0 11.0 12.0 13.0
OWE FILE - >L0738 ; TOTAL ION ( 35 - 500 omu)
r 1 r [ 1 I I I | I I I I [ I . I . j I I I I | 'I I I I | I I I I | I I I I |
16.0 17.0 18.0 19.0
REFERENCE MIXTURE CHROUATOGRW
----- BZDUCKD CHStMATOCJUM COMPARISOW STORE) IN RLE 1B0738 -----
CROSS CORRELATION COEFFICIENT - .305 (ALL PEAKS)
CROSS CORRELATION COEFFICIENT - .874 (ONLY REFERENCE PEAKS)
Figure 8. Comparison of Chromatographic Patterns of a Superfund Site Waste
Sample and a Styrene Waste Sample.
IDENTIFICATION OF SOURCES OF
CONTAMINATION ADJACENT TO A
SUPERFUND SITE
Figures 9 and 12 summarize chromatogra-
phic data obtained from an Adjacent Soil
Sample Extract, a #6 Fuel Oil Sample, and
a Styrene Waste Sample obtained fromthe
Superfund Site. Comparison of the XCCs
calculated from these comparisons indi-
cates that the #6 Fuel Oil pattern is more
similar to the Soil Extract than the Styrene
Waste Sample. In the comparison with the
Styrene Waste Sample shown in Figure 12,
the XCC for "all peaks" is 0.286 and the
XCC for "only reference peaks" is 0.563.
In this case, the higher XCC for "only
reference peaks" does not indicate that the
Styrene Waste Sample pattern is imbed-
ded in the Soil Extract pattern. This hicher
value is due to the fact that some of the
major components in the Soil Extract are
also present in the Styrene Waste Sample.
However, none of the minor components
present in the Styrene Waste Sample are
detected in the Soil Extract, although they
should have been detected if the Styrene
Waste pattern had been present in the Soil
Extract pattern. Figures 10 and 11 com-
pare mass spectra of methyl and dimethyl
naphthalene isomers in the Soil Extract
and in #6 Fuel Oil.
Conclusions: The likely source of
contamination in this area adjacent to the
Superfund Site was from fuel spills rather
than waste migration from the site. The
actual site from which the contaminated
soil was obtained was from an area which
had been filled with off-site soil prior to
commercial development.
1-108
-------
SO/MS FIE - >LS374 ; TOTAL ION ( 35 - 600 omu)
UNKNOWN MIXTURE CHROMATOGRAM
Soil Extract
l I l l 1 1 'l I l l . |Vl l t |'l' |T| I I I'l I l'i'1 I'l I 1 I I I [ f'l'l I |'l Ml I Tl I l | l I'l I'] l ill [ l Vl V |
9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0
aooo
40OQ-
01
-4000
1 I " " I " ' ' I '
i i ' l ' i i' l ' ' ' ' l ' ' ' ' I '' ' ' l ' ' ' ' l ' ' ' ' l ' ' ' ' l ' ' ' ' l ' ' ' ' l ' ' ' ' l ' ' ' ' l ' ' ' ' l '' ' ' i ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I
9.O 10.0 11.0 12.0 13.0 14.0 1S.O 16.0 17.0 18.0 19.0 20.0
#6 Fuel Oil
O-YT-, .-, f... ,-|.! i. i',-. -TI- r,' i iViT-nyvrrrr.'
9.0 10.0 11.0 12.0 13.0 14.0 15.O 16.0 17.0 18.0 19.0 20.0
GC/MS RLE - >L5390 ; TOTAL ION ( 35 - 600 omu)
REFERENCE MIXTURE CHROMATOGRAM
REDUCED CHROMATOCKAM COlffAJBBON STORED IN HLI AA5390
CROSS CORRELATION COEFFICIENT - .865 (ALL PEAKS)
CROSS CORRELATION COEmOENT - .877 (ONLY REFERENCE PEAKS)
Figure 9. Comparison of Chromatographic Patterns of a Site Adjacent to a
Superfund Site and #6 Fuel Oil.
C&W5 FIE - >L5374 &Uam Component Sptctrum tor 13,
.«j Soil Extract Component
« .0 10 .00
^
JOOO"
,li, ... \ J,
« ' .i
«
i N/ .>>.
190 l» .HO no
• . r - ^ ^
'"' r"11
,20
#6 Fuel Oil Component
/. ,'.' ..1^1.. |i ^>... >153W jbtmtKl CompofMlTt S^ctrvm for 13.76 LGmrim
o»ss ccHBQjom coemaEwr - •**>(*"- fwS_
MOSS CORHEUnW OBTKtW - JM (CWT RfffrfflCE PEWSJ
HOD'
4000
D aMD.
F
H
E
4000-
CC/US FIE • >L5J74 7hCxtura Cdmporwil SpKtrum lor
Soil Extract Component
/ /, ,"., ,ii; •' •?' "^i '\/' ,
,0 ,00
" ". "L ° V ,< ',>, >" '?'
'] ir ••
,0.
#6 Fuel Oil Component
^fu-^ft^a^a^
MISS sncntu CWFUIKW snmu a
CKOSS COMBJOnH COffTXXHI ~ JBB
OBSS OMREUTCM OXTTKfNT - JH£
18.CH
"
IS
1 ,11,
or 16.
ULIiOSS
StrTS
Unuta
152 1TJ
1GO I6D
\>" /' >(»
IflD
06 Ubnitn
BCEPEWS)
Figure 10. Comparison of Methyl
Naphthalene Isomer Mass Spectra in
an Adjacent Site and in #6 Fuel Oil.
Figure 11. Comparison of Dimethyl
Naphthalene Isomer Mass Spectra in
an Adjacent Site and in #6 Fuel Oil.
1-109
-------
FILE - >L5374 ; TOTAL ION ( 15 - 600 amu)
UNKNOWN MIXTURE CHROMATOGRAM
Soil Extract
o ]0.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 1S.O 19.0 20.0
jO 10.0 11.0 12.0 13.0 14.0 15 0 16.0 17.0 18.0 19.0 20.0
u,
Styrene Waste Sample
9.0 10.0 11.0 12 0 13.0 14.0
GQVS RLE - >L5378 ; TOTAL ION ( X - 600 omu)
REFERENCE MIXTURE OHROMATOGRAU
SEDUCED CHROMATOGflAM COMPARBON STOKED IN fUE M537B
CROSS CORRHA7TON COEFFICIENT - .286 (ALL PEAKS)
CROSS CORRELATION COEFTTOENT - .563 (ONLY REFERENCE PEWS)
Figure 12. Comparison of Chromatographic Patterns of a Site Adjacent to a
Superfund Site and a Styrene Waste Sample.
COMPARISON OF GASOLINE
SAMPLES
One potentially important source of envi-
ronmental contamination is from leaking
underground storage tanks. When these
tanks contain petroleum distillates such as
different grades of gasoline or diesel fuel, it
is very difficult to determine the sources of
this contamination, using standard EPA
protocols. This difficulty arises since many
of these components are not present in the
volatile priority pollutant list and standard
EPA analysis protocols would overlook
some of the major features of these formu-
lations. Figures 13, 16, and 19 compare
Chromatographic patterns of Gasoline A
(Regular) with Gasoline B (Unleaded),
Gasoline C (Diesel), and a Volatile Prior-
ity Pollutant Standard, respectively. The
comparisons of Gasolines A and B show
that Regular and Unleaded are very simi-
lar, while Gasoline C (Diesel) and the
Volatile Priority Pollutant Standard are
very different.
Figures 14 and 15 show the high similarity
of mass spectra of toluene and a trimethyl
benzene isomer in Gasolines A and B.
Figures 17 and 18 compare the mass spec-
tra of 1,4-dichlorobutane internal stan-
dard and a trimethyl benzene isomer in
Gasolines A and C. Note in Figure 18 that
Gasoline C also contains a significant
amount of n-hydrocarbon spectrum in the
Gasoline C feature mass spectrum. This
fact is reflected in the low XCC (0.579) and
the large positive and negative mass inten-
sities shown in the difference spectrum.
1-110
-------
«VMS FILE - >U0551 ; TOTAL ION < H - 3SO ornlj)
UNKNOWN MIXTURE CHROMATOGRAM
BOOO
4000
Gasoline A (Regular)
B 10 12 14 16 IB 20 22 24 26 28 30
B 10 12 14 16 IB 20 22 24 26 28 30
4000
Gasoline B (Unleaded)
B 10 12 14 16 IB 20 22 24 26 2B 30
GtVKS FILE - >M0552 ; TOTAL ION ( 33 - 350 omu)
REFERENCE MIXTURE CHROMMDCRAM
SEDUCED CBSOMATOGROl COlfPASISON STOKED IN TOE 1*0552
CROSS CORRELATION COSTOEVT - .750 (ALL PEAKS)
CROSS CORRELATION COEFFICIENT - .826 (ONLY REFERENCE PEAKS)
Figure 13. Comparison of Chromatographic Patterns of Gasoline A (Regular)
and Gasoline B (Unleaded).
eoao'
4000-
1OOO
0'
eooo1
tr
HVW5 FIE" >U0551 ^£itm Can!»in.,m,l Sp-ictnfln for 1169 Uinutn
(feasoline A (Regular) Component
«, »«,<«> 120 l» ,.. UO !00 =0 ,40 !
„ a.
A J,_J-.a .j . .'" ;." '." ."N. \ •/' ™N
60 HBO
fasoline B (Unleaded) Component
,,, ,„ ,4, '^ "^ ^ "^
CC4JS Ft£ - >UOS5Z Arfcrmc* Component 5p0dnjm (or 13.64 Mrann
BO 2M
cfit-KCOfiHaAnowcoEmaOiT. JBI ftu.pey.g
OBJSS OMRQAnO* OXFFIQP.T " JHI [ONL1' H7DEMZ PCVtS)
,„«,
0 j
F zoo
F
E
E
E
..a.
OC/US FILE - >U0551 ^£din ComponMt SpKlnm tor 21.41 Ifinutn
Gasolinb A (Regular) Component
,0 „ ,0. ,!0 ,40 1=0 1,0 8.0
ILL.&2S.TS -S
„ 00 .0 ,00 ,JO ,40 1.0 1.0 ioo
Gasolinb B (Unleaded) Component
to ao oo ioo IM uo IBO too 200
GC/US FILE - XJ0532 Afffwmcc Companaflt Soad/un. for 21.44 Ubuta
ItLSS SPKTUi QJBMJttSJX STUMD Of mS 1C03U
CSOSS OMROAIUN COCFnCENT • JW {AU. PtMSj
cross CWWEWTK* coFFC-wr - Jta (o*r RETOSNCE re«s)
Figure 14. Comparison of Toluene Figure 15. Comparison of Trimethyl
Mass Spectra in Gasolines A (Regular) Benzene Isomer Mass Spectra in Gaso-
and B (Unleaded). lines A (Regular) and B (Unleaded).
1-111
-------
GC/VS RLE - >M0551 : TOTAL ION ( 33 - 350 omu)
UNKNOWN MIXTURE CHROMATOGRAM
Gasoline A (Regular)
10 12 14 16 IB 20 22 24 26 2B 30
5000 :
0"!
-5000
10 12 14 16 IB 20 22 24 26 SB 30
8000~
Gasoline Q (Diesel)
I I III
\
I I MllirnTTTTnlM l,|,l..| M. l[l M.|l..ir
0 2 4 6 B 10 12 14 16 IB 20 22 24 26 2B 30
SOUS Fl£ - >U0553 ; TOTAL ON ( 33 - 350 omu)
REFERENCE MIXTURE CHROMATOGRAM
REWCED CHROUATOG&Uf COMPAKEON STORED IN FILE Ad05S3
CROSS CORRELATION COEFFICIENT - .27'I (ALL PEAKS)
CROSS CORRELATION COEFFICIENT - .412 (ONLY REFERENCE PEAKS)
Figure 16. Comparison of Chromatographic Patterns of Gasoline A(Regular) and
Gasoline C (Diesel).
j . >-MC6S1 M(Un ComponcfTl SpBcWum tor 18.
Gasoline A (Regular) Component
Gasoline C (Diesel) Component
* CDOTCC" . 1 ODD UiJ. ft*C5J
. i -
Figure 17. Comparison of 1,4-Dichloro-
butane Internal Standard Spectra in
Gasolines A (Regular) and C (Diesel)
RLE - >UOJS1 W«tu™ Comporwri Ipartnan IDT 20.86 U
Gasoline A (Regular) Component
.J-J-Vi' V - r »
Gasoline C (Diesel) Component
CfUS HIE - >WSW :R^Mnc« CanfranM Spcdrum tor MJI Urnul-
ORBS COBNOUOi OJpTPPff - JT> (Id PWOl
raoss co«D>noN co&mwr - JTI (o«.r ansocc PI
Figure 18. Comparison of Trimethyl
Benzene Spectra in Gasolines A (Regu-
lar) and C (Diesel)
1-112
-------
SCfliS FILE - >M0551 : TOTAL ION ( 33 - 350 omu)
UNKNOWN MIXTURE CHROMATOGRAM
Gasoline A (Regular)
10 12 14 16 IB 20 22 24 26 28 30
BOOO
40OO
0
-4000~
8 10 12 14 IB 18 20 22 24 26 28 30
Volatile Priority Pollkant Standard
0 I" " I"'' I''" I" " I" " 1" " I" " I" " I" " I" " 1" " I"'' I" " I' "' I " " I" " I" '' I" " I'''" I' " ' I " " I" " I" " I" " I " " I" " I" " I " "I" "'I'
0 a 4 6 B 10 12 14 16 IS 20 22 2-4 26 28 30
GOVS RLE - >M0539 ; TOTAL ION ( 33 - 350 emu)
REFERENCE MIXTURE CHROMATOSRAM
SEDUCES CSBOUATOGRAU COMPAflfSON STORED IN FILE MOS39
CROSS CORRELATION COEFFICIENT - .229 (ALL PEAKS)
CROSS CORRELATION COEFFICIENT = .771 (ONLY REFERENCE PEAKS)
Figure 19. Comparison of Chromatographic Patterns of Gasoline A (Regular)
and an 80 j/g/l_ Volatile Priority Pollutant Standard.
woo
D
R
E -sooo
K
E
flOOO1
GtVWS FU - >y0551 ;kfahn Component Spectrum for 17.95 Unlta,
Gasoling A (Regular) Component
40 » » .00 >» MO teo i» aDo 3SD »o 26o SBo
eo 60 >OD lao i« lao IBO aoo MD aUO£» ^efertno Component Spectrum (or 16J7 kCnutv*
Kin SFKIUL aaaaaon CTOBD w nu ireow
CROSS COffiOAnON OX7FOEMT - J15 Cui P&K5)
CROSS CORRELADOH COE7FKCNT - J13 (DULY RETEREHCC PEAKS)
C&VS FILE - >«0551 Jfcbr>> Comporant Spactmm for li.BS Mnutn
Gasoline A (Regular) Component
220 £10 MO
Volatile
Priority Pollutant Component
GCyMS FILE - >UQS39 flnfonmco Componnrt Spectrum for 19.B2 Mbiutn
Figure 20. Comparison of Dimethl Ben-
zene and Brompform Spectra in Gaso-
line A and Priority Pollutant Standard.
Figure 21. Comparison of Hydrocar-
bon and CI4-ethane Spectra in Gasoline
A and Priority Pollutant Standard.
1-113
-------
Figures 20 and 21 compare mass spectra of
two features present in Gasoline A and an
80 /ig/L Volatile Priority Pollutant Stan-
dard. Low XCCs and large positive and
negative intensities in the difference spec-
tra show that these Gasoline A features are
not present in the Volatile Priority Pollut-
ant Standard.
The comparison of gasoline samples high-
lights an important limitation of this ap-
proach. While the first two examples of
this approach for Superfund Site wastes
were for fairly high molecular weight
compounds which were present in highly
contaminated organic wastes, contamina-
tion of water by highly volatile organic
compounds presents several problems
when chromatographic pattern compari-
sons are used. Patterns can be changed
due to component volatility and due to
partitioning between soil and rock with
which the water samples may come in
contact. Therefore it may be very difficult
to obtain a pattern of the suspected waste
source which has been subjected to exactly
the same weathering processes as the ac-
tual sample in question. However, the
presence of many of the same components
in an unknown mixture which are also pres-
ent in gasoline samples, may be a strong in-
dication of the source of these compo-
nents.
SUMMARY
The calculation of Cross Correlation Coef-
ficients for reduced chromatographic and
mass spectrometric data is a very powerful
way of comparing complex chemical data.
This approach allows for the very rapid
evaluation of differences and similarities
of samples which may contain a large
number of discrete chromatographable
components. The major strength of this
method lies in the fact that specific, un-
equivocal identification of unknown com-
pounds is not required in order to make
very important, fundamental decisions
concerning complex mixtures. In addition,
this method can be extended to any data
which has been acquired on compatible
computer systems. This latter strength
makes this processing approach available
to HPLC/MS, pyrolysis/GC/MS, GC/FID,
GC/EC, and any GC/detector data which
has been acquired on HP-1000 systems
which have compatible file structures with
Hewlett-Packard's present RTE-VI mass
spectrometer operating system.
1-114
-------
CALIFORNIA'S PROPOSITION 65, A VOTER APPROVED ENVIRONMENTAL
LAW
Dr. Paul Marsden. Chief, Methods Research Branch, EMSL-LV,
P.O. Box 93478, Las Vegas, NV 89193-3478
ABSTRACT. California's Proposition 65, the Safe Drinking
Water and Toxic Enforcement Act of 1986, became law through
the initiative process. It is voter-passed legislation
designed to reduce public exposure to toxic chemicals.
Under the Act, the Governor's Scientific Advisory Panel is
mandated to develop a list of chemicals "known to cause
cancer or reproductive toxicity"; the list included 269
materials in January, 1989. Individuals and businesses are
prohibited from knowingly discharging a significant amount
of these chemicals into drinking water or from exposing
individuals to these chemicals without prior warning. The
Act defines a "significant amount" as any detectable amount
of a chemical, as long as that concentration (1) conveys
with it an excess risk of cancer of 1 in 100,000 (10~5) , or
(2) exceeds 1/1000 the no-observable-effeet-level (NOEL) for
reproductive toxicants. Because standardized measurement
methods are not available for all 269 chemicals, the Air and
Industrial Hygiene Laboratory of the California Departments
of Health Services (CDHS) has been made responsible for the
evaluation of suitable collection and analysis methods.
While the state does not have a specific monitoring program
for Proposition 65 analytes, these methods may be required
when individuals attempt to recover damages from businesses
under the provisions of the Act. The authors will discuss
how California is developing the mandated list of chemicals,
is setting their lawful limits, and is evaluating methods
for sampling and analysis.
INTRODUCTION
In 1986 the voters of the State of California manifested
their concern for hazardous chemicals by passing into law
"the Safe Drinking Water and Toxic Enforcement Act" (Health
and Safety Code section 25249.5 et seq.). This act became
law by a direct vote of the people under the initiative
process of the California constitution. The Act prohibits
persons in the course of doing business from knowingly
contaminating drinking water with chemicals known to the
State to cause cancer or reproductive toxicity or from
1-115
-------
knowingly exposing individuals to such chemicals without
prior warnings. The Governor is required to name and
consult with a Scientific Advisory Panel (SAP) in order to
establish a list of such chemicals. Businesses are_required
to issue warnings before any listed chemicals are discharged
into the environment (water, air, food, or soil),
incorporated into consumer products, or used in a fashion
that will result in occupational exposure.
The Act is aimed at limiting public exposure to these toxic
substances by requiring companies to restrict discharges of
the listed chemicals and to alert workers and customers
whenever any of the listed chemicals are discharged. Under
the penalty provisions of the act, businesses may be brought
to trial by the California Attorney General or local
Districts Attorney's. for violating the Act. Individuals
may initiate such proceedings and may be awarded portions of
any fines if they intervene in a case on an ex-parte basis.
TARGET COMPOUNDS AND ACTION LEVELS
The first 29 chemicals were listed under the Act on February
27, 1987; those chemicals included the 26 suspect human
carcinogens listed by the International Agency for Research
on Cancer (IARC) and three reproductive toxicants. As of
January 1, 1989, the state had increased the number of
chemicals to 269 (238 carcinogens and 31 reproductive
toxicants). Roughly half (137 of 269) of the listed
material are included in U.S. EPA's Appendix VIII (47
Federal Register 32296 [July 26, 1982]). Annual updates of
the Proposition 65 list of chemicals are required under the
terms of the Act. Once a chemical is listed, warnings are
required 12 months after the listed date, and discharge
prohibitions are required 20 months after listing.
The state must establish what constitutes a significant
amount of each listed material, that is, the quantity that
may exist before a warning must be issued or an illegal
discharge reported. The Act (section 25249.11, subdivision
(c) of the Health and Safety Code) defines "significant
amount" as any detectable amount of a chemical, unless the
chemical concentration in question conveys with it no
significant risk of cancer or does not exceed one one-
thousandth of the no observable effect level (NOEL) for
reproductive toxicants (section 25249.10 (c)) . The
California State Health and Welfare Agency performs
quantitative risk assessments to determine those
1-116
-------
concentrations and to establish the lawful levels for the
materials listed under the Act. As of January 1, 1989,
lawful levels have been established for 50 carcinogens
(sections 12709 and 12711) and two reproductive toxicants
(section 12805) . One chemical, ethylene oxide, has lawful
levels for both carcinogenicity and reproductive toxicity.
The detection limits for measurement methods suitable for
this application are calulated using the lawful levels and
exposure scenarios for environmental, occupational, and
dietary sources. To date, the exposure scenarios for
drinking water and breathing air have been established.
Section 12721 of the Act assumes that a person consumes two
liters of water a day. The target detection limit (TDL,
jug/L) for drinking water methods are determined by dividing
the lawful level (/ig/day) by 2 L/day- Section 12707 of the
Act assumes that a person respires 20 m3 of air per day.
The target detection limit (TDL, jug/L) for air methods are
determined by dividing the lawful level (/ug/day) by
2 Om3/ day) .
In order to implement the Act, it is critical for businesses
and enforcement agencies to have standardized sampling and
analysis methods which are capable of measuring the listed
chemicals at their lawful levels. In most cases, standard
methods are not available to measure the target chemicals at
the low concentations required. CDHS, through the Air and
Industrial Hygiene Laboratory and its contractor S-Cubed,
initiated a project to conduct a comprehensive survey and
evaluation of sampling and analysis methods that can provide
low detection limits. The U.S. EPA, through the
Environmental Monitoring Systems Laboratory-Las Vegas (EMSL-
LV) and Region 9, is cooperating in this effort. This
article describes the progress in evaluating methods for the
analysis of the 51 compounds with established lawful levels.
METHODS
There are three possible scenarios to be expected in the
evaluation of methods of detection for the listed chemicals
as follows: For some chemicals, there exists a validated,
standardized method that includes a comprehensive quality
assurance (QA) program with detailed quality control (QC)
requirements; methods for these compounds need only be
compiled, evaluated and reported. For other chemicals, a
scientifically acceptable method has been developed but it
1-117
-------
lacks validation or comprehensive QC procedures; these will
require single-laboratory validation. For a third group of
chemicals, there is no acceptable method available, and
method development efforts will be needed. In the initial
phase of this project, methods for chemicals in the first
category are evaluated and compiled.
For listed chemicals in the first category, there are
standardized methods that have already been published by
federal or state agencies or professional organizations such
as the U.S. Environmental Protection Agency (U.S. EPA), U.S.
National Institute of Occupational Safety and Health
(NIOSH), U.S. Food and Drug Administration (FDA), CDHS,
California Department of Food and Agriculture (CDFA),
California Air Resources Board (CARB), California State
Water Resources Control Board (CWRCB), California Regional
Water Quality Control Board (CRWQCB), local air pollution
control districts (APCDs), Association of Official
Analytical Chemists (AOAC), and American Society of Testing
and Materials (ASTM). In the initial phase of this project,
the following publications have been used to compile and
evaluate the standardized methods. Several method compendia
published by EPA include the "Test Methods for Evaluating
Solid Waste (SW-846)" organized by the Office of Solid
Waste, and the "Analytical Methods for CERCLA Hazardous
Substances" prepared for the Office of Emergengy and
Remedial Response by the EMSL-LV. The "Recommended Methods
of Analyses for the Organic Components Required for AB 1803"
published by CDHS includes a number of EPA and CDHS water
methods. CDHS's Environmental Laboratory Accreditation
Program is in the process of assembling a collection of
validated methods for drinking water and hazardous waste.
Method detection limits (MDLs) and analytical methods are
available from these publications and other literature for a
number of the chemicals listed.
During the first phase of this project, federal and state
methods for these 51 chemicals are being evaluated to find
methods which can detect chemicals in water at levels as
close as possible to the lawful levels. A critical factor
in the evaluation of methods is the MDL. MDL was chosen as
a basis of comparison because the term is recognized within
the analytical community as the smallest reliably detectable
quantity of an analyte (40 CFR Part 136, Appendix B, Vol.49,
No. 209, October 26, 1984). Most authorities in the field
agree that this quantity is related to the standard
deviation (SD) of replicated standard analyses at near-zero
1-118
-------
analyte concentrations. The detection signals must be at
least three times larger than the noise of the system. The
MDL is a basic performance characteristic of an analytical
method and is matrix-sensitive. The following equation is
used to calculate the MDL from seven replicate analytical
determinations:
MDL = (t.w) (SD)
where SD is the standard deviation of the average analyte
recovery from the seven determinations, and t.w is the
Student's t value (one-tailed) for n-1 determinations at the
99% confidence level. Generally, it is possible to estimate
closely the MDL for matrices with which the analyst has had
some experience.
Methods examined under this project are intended for
regulatory purposes. In order to provide a routine
procedure for evaluating and comparing the methods, a risk-
detection ratio (RDR) is obtained by dividing the TDL by the
MDL. Values for RDR that are one or greater indicate that
the method will detect analytes a concentrations equal to
the TDL. In general, RDR values of five to ten or greater
are preferred for regulatory purposes. This value was
chosen because analytical methods should not be used to
produce regulatory quality data a the MDL; action levels
should be greater than the limit of quantitation (LOQ) of a
method. The U.S. EPA defines this concentration as the
practical quantitation limit (PQL) of a method and CDHS
defines it as "detection limits for the purpose of data
reporting" (DLR).
All methods that are acceptable for monitoring and
enforcement provisions of the act will be accumulated in a
database by CDHS. The database will be developed using the
program Q and A. This database will contain method
performance data (precision, accuracy, method detection
limit, linear concentration range, and suitability for
specific matrices), the level of method validation, specific
QC requirements, and compatible cleanup procedures.
RESULTS
Table I includes the lawful levels (^g/day) for chemicals
that have been established under the Act as of January 1,
1989; these chemicals have been given higher priority for
method evaluation. Tables II and III list CDHS and EPA
1-119
-------
methods in water and compare the MDL or DLR with the TDL.
Table II lists chemicals with at least one water test method
which has a RDR equal or greater than 1; Table III lists
those with a RDR less than 1. Table IV shows chemicals
which do not require water methods because they are listed
as inhalation hazards only. Table V lists chemicals for
which no appropriate methods are currently available.
The information presented in Table II indicates that there
are methods suitable for measuring 21 of the analytes. The
most sensitive organic methods (those with the lowest MDL or
DLR) use gas chromatography (GC) with selective detectors
(EPA methods 502.2, 603, 607, 8081 and 8141) or high
performance liquid chromatography (HPLC) (EPA methods 605
and 8310). Analytical methods that use GC/mass spectrometry
(GC/MS, EPA Methods 524.2, 8270, 8280 and 8290) provide
better identification of analytes but have higher MDL's. In
contrast, the most sensitive method for arsenic, the only
inorganic analyte listed in Table II, is an inductively
coupled plasma/MS technique. It is anticipated that routine
EPA water sampling methods will be suitable for all 21 of
these analytes. It is hoped that the same measurement
methods can be used for the analysis of water, soil, air,
and other matrices once appropriate sample collection and
preparation methods are identified.
At the present time there are no accepted EPA or CDHS
methods with the sensitivity to measure the lawful levels in
water for nine of the analytes (Table III). Methods that
may be suitable for hexachlorobenzene (draft 8080.1) and for
lead (preconcentration) are under development at EMSL-LV and
are expected to be submitted for review in October.
Potential HPLC method(s) for the six nitrogen containing
analytes (benzidine, dinitrotoluene, and the nitroso
compounds) are being investigated at EMSL-Cincinnati (EMSL-
Ci). While a more sensitive method for 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD, Method 8290) has been
submitted for U.S.EPA review by EMSL-LV, its RDR value is
still less than one. Once the most suitable method is
obtained for any of these analytes, it will be added to the
CDHS data base.
Among the 51 analytes listed in Table I, according to
section 12707 of the Regulations, six materials have been
designated as inhalation hazard only (Table IV). They are
asbestos, beryllium, beryllium oxide, beryllium sulfate,
cadmium, and hexavalent chromium compounds. These analytes
1-120
-------
may require no methods for analysis in water; the air matr
will be dealt with in a later report.
Those materials for which no appropriate method is availab
are listed in Table V. Methods for three of these analyte;
are being developed by the EMSL-Ci in support of the RCRA
program. The remaining 14 analytes pose particular
analytical problems in water or are not clearly defined.
a result, it is unlikely that suitable methods will be
located for this group. A brief paragraph describing the
problems and a recommended solution is provided for each o
the 14 analytes.
1,3-Butadiene. The initial literature search did not
provide any analytical methods with low MDL. Since it is .
gas at room temperature with low water solubility, it may 1
determined in air only. 1,3-Butadiene reacts with hydroxy
radicals in air to form acetaldehyde and acrolein.
Bis(chloromethyl)ether. Analytical measurement of this
analyte in water at low MDLs is extremely difficult. This
is primarily due to decomposition of the analyte by water
(half-life 10-38 seconds at pH 7). This analyte may be
determined in air only.
Chlordane. Technical chlordane consists primarily of
heptachlor and two chlordane isomers with the remainder
being hexa-, hepta-, and nonachlor and other related
dicyclopentadienes. Chlordane itself has two structural
isomers, alpha and gamma chlordane. The GC pattern of a
chlordane residue may differ considerably from that of the
technical standard. Depending on the sample matrix and it
history, almost any combination of the 11 major and 30 min
components in the technical mixture can be found. Because
of the inability to predict a chlordane residue pattern fo
GC, a simple method cannot be specified. If alpha and gam
chlordane were specified under the Act, the ambiguity from
the analytical determination could be removed. Method 808
is suitable for the determination of the major chlordane
isomers and heptachlor.
Coke oven emissions. This is a complex mixture which
requires definition. Coke oven emissions are known to
contain polyaromatic hydrocarbons (PAH's) which could be
analyzed by EPA Method 8270 (GC/MS) or Method 8310 (HPLC).
1-121
-------
Epichlorohydrin. This analyte does not purge well and
decomposes in water (half-life 29 days). It is currently
not on any EPA water or soil method list. It may be
determined in air only.
Ethylene oxide. Epoxide compounds in general are reactive
and water-sensitive. This analyte does not purge well and
decomposes in water (half-life 9-14 days). It may be
determined in air only-
Hexachlorocyclohexane (technical grade). This analyte is a
mixture of chemicals also known, and improperly named, as
benzene hexachlorides (BHC's). The name arises from the
industrial production of technical hexachlorocyclohexane by
the chlorination of benzene. Technical-grade BHC or
hexachlorocyclohexane consists of a mixture of six
hexachlorocyclohexane isomers and several isomers of
heptachloro- and octachlorocyclohexane. The most toxic
isomer, gamma-BHC, is also called lindane. EPA Method 8081
will measure the four major components (alpha-, beta-,
gamma- and delta-BHC) of the technical mixture.
Nickel Refinery Dust from the pyrometallurgical process.
Further definition is needed for this analyte. Its toxicity
may be caused by the fact either that it is a dust or that
it is a metal. If its toxicity is due to the physical
properties as a dust, then air is an appropriate matrix to
sample. If the toxicity is due to the chemical properties
as a metal, then Method 6020 should be suitable.
Nickel subsulfide. This analyte presents a particularly
difficult analytical problem. It is a complex mixture of
two oxidation states of nickel. Because it is a sulfide
chelate of a metal, it is virtually insoluble in water, and
thus is an inappropriate analyte for that matrix. An
exhaustive literature search is currently in progress to
locate an analytical method specific for nickel subsulfide.
N-Nitrosodiphenylamine. Determination of this analyte is
ambiguous because it decomposes in the inlet of GC and GC/MS
systems to diphenylamine and is then quantitated as such.
While the EMSL-Ci has recently published an HPLC/MS method
for N-nitrosodiphenylamine, it has a significantly higher
MDL than Method 8270.
N-Nitroso-N-ethylurea and N-Nitroso-N-methylurea. No
suitable methods have been found for these analytes yet.
1-122
-------
Methods for most of the N-nitroso compounds researched thus
far have fallen short of the TDL by significant margins.
Based on the chemistry of these analytes, it is probable
that validated methods for their determination at the low
lawful levels specified under the Act cannot be found.
Polychlorinated biphenyls (60% or greater chlorination).
Polychlorinated biphenyls are normally found in complex
mixtures(i.e., Aroclors). The most sensitive methods for
Aroclors employ GC/electron capture detector (GC/ECD), and
the identification is almost always done by recognition of
the characteristic Aroclor pattern. These methods are not
congener- specific so they are inappropriate due to the
Regulation's stipulation for biphenyls with 60% or greater
chlorination. More appropriate methodology (Method 680)
employs GC/MS and software that allows the analyst to
determine level of chlorination for the biphenyls.
Unfortunately, Method 680 may not be suitable for this
application because congeners with different levels of
chlorination are not always resolved by GC and the GC/MS
hardware responds better to biphenyls with less than 60%
chlorination. If low-level detection of polychlorinated
biphenyls is required, a GC/ECD pattern recognition method
for determining Aroclors may be required. Improvements in
Method 8081 being developed at the EMSL-LV should provide
better PCB isomer identification than existing methods.
Toxaphene. The analyte is a complex mixture of chlorinated
camphenes. A determination of toxaphene is normally done by
using GC/ECD and pattern recognition. Pattern recognition
is difficult, however, in the presence of interfering GC/ECD
peaks.
CONCLUSIONS
California's Safe Drinking Water and Toxic Enforcement Act
of 1986 is the first environmental law in which the risk
assessment values of health effect evaluation have
regulatory impact. The state has listed 269 chemicals as
carcinogens or reproductive toxicants; lawful levels for
discharge or exposure for 51 of these compounds have been
established. Preliminary evaluations of analytical methods
for the 51 analytes have been completed and methods are
available for 21 of the analytes with sufficient sensitivity
for regulatory purposes. The limitations of standard
methods to measure these analytes at toxicologically
relevant concentrations is apparent. Equally apparent is
1-123
-------
the need for those responsible for generating lists of new
toxicants to recognize the relevance of the environmental
stability of chemicals and the difficulty of measuring
poorly defined parameters. Successful interaction of
analytical chemists, toxicologists, and regulatory
specialists is required to craft laws that will effectively
reduce risks associated with exposure to chemicals.
1-124
-------
Table I. Lawful levels of Proposition 65 chemicals
as of January 1, 1989.
Acetaldehyde
Acrylonitrile
Aldrin
Arsenic (inorganic compounds)
Asbestos* (inhalation only)
Benzene
Benzidine [and its salt]
Benzo[a] pyrene
Beryllium (inhalation only)
Beryllium oxide (inhalation only)
Beryllium sulfate (inhalation only)
Bis(chloromethyl)ether (bis(
1,3-Butadiene
Cadmium (inhalation only)
Carbon tetrachloride
Chlordane
Chloroform
Chromium (hexavalent) (inhalation only)
Coke oven emissions
Dichloromethane (Methylene chloride)
Dieldrin
Di(2-ethylhexyl)phthalate (b
2,4-Dinitrotoluene
Epichlorohydrin
Ethylene dibromide (1,2-Dibr
Ethylene dichloride (1,2-Dic
Ethylene oxide
Formaldehyde (gas)
Heptachlor
Heptachlor epoxide
Hexachlorobenzene
Hexachlorocyclohexane (technical grade)
Nickel subsulfide
N-Nitrosodi-n-butytamine
N-Nitrosodiethylamine
N-Hitrosodimethylamine
N-Nitroso-diphenylamine
N-Nitroso-N-ethylurea
N-Nitroso-N-methylurea
N-Nitrosopyrrolidine
Polychlorinated biphenyls (>60% chlorine)
Tetrachloroethylene
Toxaphene (polychlorinated camphenes)
Trichloroethylene (Trichloroethene)
2,4,6-Trichlorophenol
Vinyl chloride
CAS Number
CHEMICALS KNOWN TO THE STATE TO CAUSE
f>
ily)
sromethyDether)
yn only)
iloropheyOethane)
:hlorobenzidine)
-ide)
2-Ethylhexyl)phthalate)
sthane) (EDB)
•oethane) (EDC)
. grade)
jrgical process
chlorine)
dioxin (TCDD)
lenes)
tne)
75070
107131
309002
...
1332214
71432
92875
50328
—
...
—
542881
106990
—
56235
57749
67663
...
—
50293
91941
75092
60571
117817
121142
106898
106934
107062
75218
50000
76448
1024573
118741
—
...
12035722
924163
55185
62759
86306
759739
684935
930552
...
1746016
127184
8001352
79016
88062
Listed Date
CANCER
April 1, 1988
July 1, 1987
July 1, 1988
February 27. 1987
February 27, 1987
February 27, 1987
February 27, 1987
July 1, 1987
October 1, 1987
October 1, 1987
October 1, 1987
February 27, 1987
April 1, 1988
October 1, 1987
October 1, 1987
July 1, 1988
October 1, 1987
February 27, 1987
February 27, 1987
October 1, 1987
October 1, 1987
April 1, 1988
July 1, 1988
January 1, 1988
July 1, 1988
October 1, 1987
July 1, 1987
October 1, 1987
July 1, 1987
January 1, 1988
July 1, 1988
July 1, 1988
October 1, 1987
October 1, 1987
October 1, 1987
October 1, 1987
October 1, 1987
October 1, 1987
October 1, 1987
April 1, 1988
October 1, 1987
October 1, 1987
October 1, 1987
January 1, 1988
January 1, 1988
April 1, 1988
January 1, 1988
April 1, 1988
January 1, 1988
(/tg/day)
Lawful Level
90
3
0.04
10
100Fibers/day
20
0.003
0.06
0.1
0.1
0.0002
0.6
0.4
1
5
0.5
9
0.001
0.3
2
0.4
50
0.04
80
2
70
3
9
2
15
0.2
0.08
0.4
0.4
0.8
0.4
0.1
0.02
0.03
140
0.02
0.002
0.3
0.09
0.000005
14
0.6
60
40
1-125
75014
February 27, 1987
0.3
-------
Table I. Lawful levels of Proposition 65 chemicals
as of January 1, 1989.
(continued)
CHEMICALS PK3UM TO THE STATE TO CAUSE REPRODUCTIVE TOX1C1TT
Developmental toxicity
Lead
Female reproductive toxicity
February 27, 1987
Ethylene oxide
Lead
Hale reproductive toxicity
Lead
75218 February 27, 1987
February 27, 1987
February 27, 1987
0.5
20
0.5
0.5
*Fibers equal to or greater than 5 micrometers in length and 0.3 micrometers in width, with a length/width
ratio of greater than or equal to 3:1 as measured by phase contrast microscopy.
NOTE:The lawful level under the Act is established by the California State Health and Welfare Agency, 1600
Ninth Street, Room 450, Sacramento, CA 958H.
Table II. Validated water methods
with risk-detection ratios* equal to or greater than l
Lawful Level
ug/day
TDL*
(ug/L)
Method
HDL/DLR
(tfg/L)
RDR*
VDUTILES IN WATER
Acrylonitrile 3
Benzene 20
Carbon tetrachlorfd* 5
Chloroform o
Dichloromethane 50
(Hethylene chloride)
Ethyleoe dibromide
(1,2-Dibromoethane) (EDB)
Ethylene dichlorid* 9
(1,2-Dichloroethane) (EDO
Tetrschloroethylene u
(Perchloroethylene) (Tetrachloroethene)
1.5
10
2.5
4.5
25
1.5
4.5
EPA 603
0.5
EPA 502.2
DHS/A81803
EPA 524.2
EPA 502.2
DHS/AB1803
EPA 524.2
EPA 502.2
DHS/AB1803
EPA 524.2
EPA 502.2
EPA 524.2
DHS/AB1803
DHS/AB1803
EPA 504
EPA 524.2
EPA 502.2
EPA 524.2
DHS/AB1803
EPA 502.2
EPA 524.2
DHS/AB1803
0.01
0.5
0.5
0.01
0.5
0.5
0.02
0.5
0.5
0.02
0.09
0.5
0.02
0.02
0.06
0.03
0.06
0.5
0.04
0.05
0.5
1,000
100
20
250
25
5
225
45
9
1,250
278
250
375
75
5
150
75
45
175
140
70
1-126
-------
Table II. Validated Water Methods with
risk-detection ratios* equal to or greater than 1
(continued)
Chemical
Lawful Level
(ug/day)
TDL*
(lfg/L)
Method
VOLATILES IM WATER (continued)
MDL*
(ug/L)
RPR*
Trichloroethylene 60 30
(Trichloroethene)
Vinyl chloride 0.3 0.15
EXTRACT ABIES
Benzo[a]pyrene 0.06 0.03
3,3'-Dichlorobenzidine 0.4 0.2
Di(2-ethylhexyl)phthalate 80 40
(bis(2-Ethylhexyl )phthatate)
N-Nitroso-diphenylamine 140 70
2,4,6-Trichlorophenol 40 20
EPA 502.2
EPA 524.2
DHS/AB1803
EPA 502.2
DHS/EPA 524
EPA 524.2
DHS/EPA 502
IM WATER
EPA 8310
DHS/AB1803
EPA 8270
EPA 605
EPA 8270
DHS/AB1803
DHS/AB1803
EPA 8270
DHS/AB1803
DHS/AB1803
EPA 8270
0.01
0.02
0.5
0.04
0.1
0.17
0.3
0.001
10
10
0.13
10
20
5
10
5
5
20
3,000
1,500
300
4
10
1
2.5
30
0.015
0.003
2
0.1
0.01
40
4
70
20
1
PESTICIDES IM UATER
Aldrin 0.04 0.02
DDT 2 1
(1,1,1-Trichloro-2,2-bis(p-chlorophenyl)ethane)
Dieldrin 0.04 0.02
Heptachlor 0.2 0.1
Heptachlor epoxide 0.08 0.04
HETALS IM
Arsenic 10 5
DHS/AB1803
EPA 8080.1
DHS/AB1803
EPA 8080.1
EPA 8080.1
DHS/AB1803
EPA 8080.1
DHS/AB1803
EPA 8080.1
DHS/AB1803
UATER
EPA 206.2
EPA 7061
d-EPA 6020
0.01
0.01
0.02
0.1
0.02
0.05
0.01
0.02
0.02
0.1
1
2
2
10
2
250
10
1
2
10
20
2
2
5
3
3
*Risk-detection ratio (RDR) is the ratio of the target detection limit (TDL) versus the method detection
limit (HDL). RDR = TDL/DLR X 5, RDR = TDL/MDL
1-127
-------
Table III.
Water methods with risk-detection ratios*
less than one (continued)
Chemical
Benzidine [and its salt]
2,4-Dinitrotoluene
N-Nitrosodi-n-butylamine
N-Nitrosodiethyl amine
H-Nitrosodimethyl aminc
N-Nitrosopyrrolidine
Lawful Level
(ug/day)
0.003
2
0.1
0.02
0.03
TDL*
(ug/D
EXTRACTABLES
0.0015
1
0.05
0.01
0.015
Method
IN UATER
EPA 605
DHS/AB1803
DHS/AB1803
EPA 625
EPA 8270
EPA 8270
EPA 607
DHS/AB1803
EPA 8270
HDL*
(ttg/L)
0.08
5
5
5
10
10
0.15
5
10
RPR*
0.02
0.0015
1
0.2
0.005
0.001
0.1
0.015
0.002
0.3
2,3.7,8-Tetrachloro-dibenzo- 0.000005
para-dioxin (TCOD)
Hexachlorobenzene
0.4
Lead
0.5
0.15
0.0000025
EPA 8270
EPA 8280
PESTICIDES 1H UATER
HETALS IM UATER
0.25
EPA 239.2
draft
10
0.00044
0.02
0.006
0.2
DHS/AB1803
EPA 625
EPA 8080.1
5
5
0.1
0.2
0.04
2**
0.25
5**
* Risk-detection ratio, (RDR) is the ratio of the target detection limit (TDL) versus the method detection
limit (HDL). RDR TDL/DLR X 5 and RDR TDL/HDL
** Draft method
Table IV. Chemicals that do not require methods in water
Chemical
(inhalation hazard only)
Asbestos
Beryllium
Beryllium oxide
Beryllium sulfate
Cadmium
Chromium (hexavalent)
Lawful Level
(ug/day)
100 Fibers/day
0.1
0.1
0.0002
1
0.001
1-128
-------
Table V. Problem Analytes
Lawful Level
Chemical (ug/day)
COMPLEX MIXTURES
Chlordane 0.5
Coke oven emissions 0.3
Hexachlorocyclohexane (technical grade) 0.4
Nickel refinery dust from the pyrometallurgical process 0.8
Nickel subsulfide 0.4
Polychlorinated biphenyls (containing 60 or more percent chlorine by molecular weight) 0.09
Toxaphene (polychlorinated camphenes) 0.6
DECOMPOSE IN UATER
Bis(chloromethyl)ether 0.6
Epichlorohydrin 70
Ethylene oxide (Cancer) 2
INSOLUBLE AT PH 8
Nickel subsulfide 0.4
DRAFT METHODS - EMSL-Ci
Acetaldehyde 90
Acrylonitrile 3
Formaldehyde 15
SENSITIVITY PROBLEMS
1,3-Butadiene 0.4
N-Nitroso-N-ethylurea 0.02
N-Nitroso-N-methylurea 0.002
1-129
-------
ENFORCEMENT OF RCRA AT RADIOACTIVE MIXED WASTE FACILITIES
MRTANTF S. BftRGER, U.S. ENVIRONMENTAL PROTECTION AGENCY, 401 "M" STREET, N.W.,
WASHINGTON, DC 20460
Radioactive mixed wastes (mixed wastes) are wastes that
contain hazardous wastes subject to RCRA and radioactive wastes
subject to the Atomic Energy Act (AEA). It is clear that source,
special nuclear and byproduct material are exempt from RCRA. On
July 3 1986 EPA published a Federal Register Notice which
clarifies that wastes containing both hazardous wastes and
radioactive wastes are subject to dual regulation. Due to unique
technical and regulatory aspects concerning mixed wastes, special
enforcement considerations and issues arise. This paper will
focus on EPA's enforcement involvement for mixed wastes. Special
mixed waste considerations that will be addressed include
sampling, testing and analysis, documentation and safety.
1-130
-------
HAZARDOUS GROUND-WATER TASK FORCE DATA BASE/
IME3!LEMENTATION OF FIEID QA/QC
T. IA COSTA, K. JENNINGS, U.S. ENVIRONMENTAL PROTECTION AGENCY, OFFICE OF
ENFORCEMENT, MAIL CODE 05 520 401 "M" STKfclLT, N.W., WASHINGTON, DC 20460
In the Fall of 1984, EPA established the Hazardous Waste
Ground-Water Task Force to evaluate the level of compliance and
identify the causes of poor compliance with ground-water monitoring
regulations at hazardous waste disposal facilities. As a part of
this effort an evaluation team was formed to determine the status
of ground-water monitoring programs at existing hazardous waste
treatment, storage and land disposal facilities. The inspection
team evaluated 58 commercial and private land disposal facilities.
To assist the members of the Task Force in executing thorough
investigations and collecting representative samples, a protocol
was developed to provide detailed guidance and procedures. This
protocol has been made available to the EPA regional and state
personnel. Throughout the conduct of the Task Force
investigations, an intensive field quality control program was
utilized for ground-water sample collection to ensure that the data
was of high qua!i ty .
Analytical data obtained from the field inspections were
compiled and input into a data base. EPA intends to utilize this
data base to gain a better understanding of ground-water
contamination problems identified at facilities subject to RCRA.
Efforts are currently ongoing to develop a second data base to
house pertinent field information such as well construction
materials, screen lengths, well diameter, sampling devices,
sampling procedures, and well development techniques. The EPA
hopes to use this data base to study the impacts of field
components on the quality of the analytical data generated and to
apply the findings to enforcement and permitting decisions.
1-131
-------
INORGANICS
-------
VALIDATION OF A METHOD FOR DETERMINING ELEMENTS IN
SOLID WASTE BY MICROWAVE DIGESTION
David A. Binstock, Peter M. Grohse, and Alvia Gaskill, Jr., Research
Triangle Institute, Research Triangle Park, North Carolina 27709; and
Charles Sellers, Office of Solid Waste, U.S. Environmental Protection
Agency, Washington, D.C. 20460.
ABSTRACT
The techniques which are typically used to prepare RCRA wastes for anal-
ysis for metals and other elements are generally relatively time consum-
ing, requiring several hours to several days to complete. They also of-
ten involve the use of acid digestions and thermal decomposition steps
which may result in analyte losses, incomplete recoveries, or sample
contamination. These limitations are well known to the analytical com-
munity and to the end users of this data in EPA, States, and industry.
The resulting inefficiency of these techniques reduces laboratory sample
throughput, drives up the cost of analytical testing and impedes deci-
sionmaking. Given these concerns, the OSW Methods Section is interested
in developing cost effective sample preparation techniques for metals
and other elements in environmental and process waste samples. Once de-
veloped, these techniques can then be written as methods for inclusion
in "Test Methods for Evaluation of Solid Waste SW-846" and made avail-
able to the user community.
A microwave assisted sample preparation method for determining elements
in solid waste has been developed (Method 30XX). This paper reports on
the validation of this method by a collaborative study to determine its
precision and accuracy. Both qualitative and quantitative aspects of
the method were assessed. Qualitative factors evaluated were ease of
use and time requirements. Quantitative factors evaluated were the
precision of the method both within a single laboratory and the total
method precision, and the bias of the method.
The method was compared with Method 3050, an open vessel hot plate di-
gestion method.
INTRODUCTION
Microwave assisted sample dissolution is now receiving considerable
attention and use in the laboratory. The procedure generally involves
placing a sample in an acid solution in a closed vessel equipped with a
pressure relief valve. The vessel is then subjected to microwave energy
in a modified microwave oven. The conditions of high pressure generated
in the container, coupled with the rapid heating of the sample via di-
rect microwave energization of the acid molecules, can result in signif-
icantly reduced preparation time; from several hours in a conventional
convection oven, hot plate, or steam bath, to several minutes in the
microwave oven.
1-133
-------
Based on in-vessel temperature and pressure profile studies conducted by
the National Institute of Standards and Technology (NIST), formerly
known as the National Bureau of Standards (NBS), microwave oven prepara-
tion conditions for oils and soils have been determined and written as a
draft method. This involves the use of concentrated nitric acid as the
digestion medium. The intent is not to completely solubilize all ele-
ments in the sample; rather, it is to solubilize those elements most
likely to be made environmentally available.
Previous work has reported on the evaluation in a single laboratory of
the draft method using NIST Standard Reference Material (SRM) represen-
tative of oils and soils1- It was reported that this method should
prove a suitable alternative for SW-846 Method 3050 with a substantial
time/cost savings. Based on these results, a collaborative study was
conducted for final method validation.
This paper reports on results of the collaborative study for validation
of the microwave method. Four NIST SRMs and one solvent recovery waste
were digested by 15 laboratories using the microwave draft method. An-
alysis of these digests was carried out by RTI using Inductively Coupled
Plasma (ICP) Spectrometry and Graphite Furnace Atomic Absorption (GFAA).
In addition, laboratories digested the samples by SW-846 Method 3050
with results using the microwave draft method compared to SW-846 Method
3050.
EXPERIMENTAL METHODS
Microwave Oven
The MDS-81D Microwave system (CEM Corporation, Indian Trail, NC) was
used for this study. The oven resembles a standard microwave oven, but
is equipped with additional features to facilitate sample preparation.
For example, the Teflon-coated microwave cavity has a variable speed
corrosion resistant exhaust system. The main element of the system
couples a precise microwave variable power system with a programmable
micro-processor digital computer. Other elements include a rotating
turntable, Teflon vessels with caps, a patented pressure relief valve,
and a capping system.
The Teflon sample vessels and caps are designed to withstand pressures
up to 100 psi and temperatures up to 200°C.
Collaborative Study Materials
The collaborative study was carried out using the following materials:
NIST SRM 2704 - Buffalo River Sediment
NIST SRM 4355 - Peruvian Soil
NIST SRM 1085 - Wear Metals in Oil
1-134
-------
NIST SRM 1634b - Trace Elements in Fuel Oil
• Solvent recovery waste
To simulate a contaminated soil, a 1:1 mixture of 1634b and 2704 was
prepared and analyzed.
Microwave Test Method
The method described below was developed for only two vessels placed in
the microwave oven and is optimized for temperatures and pressures that
would produce efficient chemical decomposition of the sample. Essen-
tially the same conditions with an increased power setting are utilized
for six vessels.
Two vessels, each containing up to 0.5 g of sample in 10 ml concentrated
HN03 are heated in the microwave oven for ten minutes at a power setting
of 344 watts. These vessels are placed in the microwave oven carousel
with accompanying vapor trap vessels. For six vessels, a power setting
of 574 watts is used.
To reduce the likelihood of analyte loss when decomposing samples pro-
ducing significant gas on decomposition, a configuration is employed
using a second vessel to trap the hot acid vapor and any aerosol ex-
pelled when the pressure relief valve of the first vessel opens. A PFA
Teflon tube connects the digestion vessel to a second vessel with a
double-ported cap. The second port on the catch vessel is connected to
the center well of the carousel to capture potential venting from this
overflow vessel. The acid and any sample condensed in the second vessel
is washed back into the sample digestion vessel at the end of the micro-
wave procedure. The vessel contents are filtered into an acid-cleaned
50 ml volumetric flask.
Collaborative Study Design
The objective of the collaborative study was to validate the draft mi-
crowave method. This involved determination of the precision of the
method both within a single laboratory and the total precision (within
and between laboratory) of the method. In addition, the bias of the
method was evaluated for those samples where the method results in a
sample digest such that a compositional analysis of the original sample
can be made.
A total of 15 laboratories participated in the study. Each laboratory
was sent aliquots of the four NIST SRMs and the Solvent Waste along with
instructions and a copy of the draft method.
A scheme was developed where each laboratory digested two replicates,
one replicate was digested under two vessel conditions and the other
replicate digested under six vessel conditions.
1-135
-------
In addition, each laboratory was requested to perform duplicate diges-
tions using SW-846 Method 3050, an open vessel hot plate acid digestion.
All digests were forwarded to Research Triangle Institute (RTI) where
ICP and GFAA analyses were performed.
RESULTS
ICP results for the microwave draft method 30XX and SW-846 Method 3050
are shown in Tables 1 and 2. Buffalo River Sediment, Peruvian Soil, the
1:1 mixture of Buffalo River Sediment/Trace Elements in Fuel Oil, and
the Solvent Recovery Waste were analyzed for 19 elements. SRM 1085,
Wear Metals in Oil, was analyzed for the nine elements that are certi-
fied by NIST. The number of observations varies from the ideal of 30
due to exclusion of outliers and non-digested samples.
GFAA results for the microwave draft method for arsenic and selenium in
Buffalo River Sediment and Peruvian Soil are shown in Table 3. Digests
from the first six laboratories to return samples were analyzed. Preci-
sion for arsenic is excellent whereas selenium is only fair, probably
reflecting the extremely low concentration of selenium present in the
two samples. A comparison of mean concentrations with the certified
values reveals generally poor recovery, ranging from a percent bias of
12 for selenium in Peruvian Soil to 39 for selenium in Buffalo River
Sediment.
A closer look at the ICP data is shown in Table 4. Values obtained us-
ing the draft microwave method for SRM 1085, Wear Metals in Oil, are
compared to the NIST certified levels. Of the nine certified elements,
seven exhibit excellent recovery with 0 to 9% bias. Silver and molybde-
num are low, but are generally regarded as "problem" elements. The pre-
cision ranges from 8 to 15% RSD, which is quite good.
A comparison of the draft microwave method versus SW-846 Method 3050 is
extremely interesting (Tables 5 and 6). Because HC1 and ^2 are used
in addition to HN03, it would be expected to see higher recoveries for
Method 3050 and this is generally true, but the differences between the
two methods are slight. In the case of SRM 2704, Buffalo River Sediment
(Table 5), with the exception of boron, recoveries are very similar.
Recovery differences range from a low of 2% for chromium and zinc to a
high of 18% for beryllium. In the case of the Solvent Recovery Waste
(Table 6), with the exception of silver, which disappears in the pres-
ence of HC1 and calcium, recoveries are again very similar. Recovery
differences range from 0 to 8%.
Of further interest is a comparison of method precision. For SRM 2704,
Buffalo River Sediment, the microwave method is more precise than Method
3050 in 15 out of 17 elements (Table 5). The two exceptions are calcium
and cadmium. For the Solvent Recovery Waste, the microwave method ex-
hibits better precision for 14 out of the 18 elements (Table 6).
1-136
-------
An additional observation is the variation of method precision with sam-
ple heterogeneity. An SRM such as Buffalo River Sediment provides much
better overall precision than a non-homogeneous "real" sample (Tables 5
and 6) such as the Solvent Recovery Waste.
CONCLUSIONS
Evaluation of the draft microwave digestion method by
study indicates that this method should prove a suitable
SW-846 Method 3050 with a substantial time/cost savings.
Comparison of the draft method
with overall better precision.
method can be evaluated, it is
REFERENCES
with Method
For the one
excellent.
3050 reveals
sample where
a collaborative
alternative for
similar numbers
the bias of the
1. D. A. Binstock, P. M. Grohse, A. Gaskill, Jr., K. K. Luk, P. L.
Swift, H. M. Kingston, and C. Sellers. Validation of Methods for
Determining Elements in Solid Waste by Microwave Digestion, Solid
Waste Testing and Quality Assurance, 4th Annual Symposium (1988).
1-137
-------
TABLE 1 -- ICP Analysis Using Method 30XX (/*g/g)
Element Mean + S.D.(n)a Mean+S.D.(n)
Sample
3
Mean + S.D.(n)
Mean + S.D.(n) Mean + S.D.(n)
Ag <1.0
Al 1.18 + 0.137%(30)
B 34.6 + 9.31(30)
Ba 77.7 + 5.90(30)
Be 0.562 + 0.068(30)
Ca 2.00 + 0.383%(30)
Cd 3.19 + 0.613(29)
Co 10.7 + 1.46(30)
Cr 81.7 + 5.33(30)
Cu 80.3 + 6.92(30)
Fe 2.96 + 0.214%(30)
Mg 0.810 + 0.047%(30)
Mn 460 + 25.7(30)
Mo <2.5
N1 36.4 + 2.52(27)
Pb 143 + 9.46(30)
Sr 33.0 + 2.05(30)
2.05 + 0.908(18)
1.92 + 0.223%(30)
35.5 + 7.47(30)
135 + 10.8(30)
0.493 + 0.069(30)
1.09 + 0.275%(30)
0.901 + 0.227(27)
10.4 + 1.24(30)
13.8 + 1.18(28)
53.4 + 5.74(30)
2.50 + 0.392%(30)
0.705 + 0.041%(30)
541 + 29.1(30)
<2.5
9.59 + 1.10(28)
121 + 8.28(30)
81.0 + 7.04(30)
234 + 35.9(22)
295 + 31.1(28)
293 + 26.6(30)
289 + 23.8(30)
311 + 34.7(27)
270 + 29.1(28)
238 + 30.3(30)
293 + 25.0(30)
279 + 22.1(30)
0.685 + 0.145%(21)
20.7 + 7.65(22)
43.5 + 7.86(22)
0.297 + 0.064(23)
1.13 + 0.260%(23)
1.50 + 0.219(21)
5.89 + 1.27(23)
43.1 + 4.90(21)
41.4 + 5.03(23)
1.58 + 0.180%(19)
0.410 + 0.064%(23)
238 + 30.0(21)
<2.5
30.5 + 5.08(23)
74.5 + 8.88(20)
17.5 + 2.05(19)
3.28 + 1.90(9)
0.148 + 0.027%(28)
37.4 + 7.01(25)
538 + 110 (28)
<0.25
0.219 + 0.093%(28)
4.90 + 1.04(27)
21.9 + 5.04(26)
161 + 23.8(26)
208 + 32.6(26)
0.316 + 0.052%(26)
0.038 + 0.009%(26)
31.7 + 5.05(26)
19.1 + 3.36(28)
50.9 + 9.96(26)
437 + 70.3(26)
71.1 + 13.5(26)
-------
TABLE 1 — ICP Analysis Using Method 30XX (/*g/g)
(Continued)
Sample
12345
Element Mean + S.D.(n)a Mean + S.D.(n) Mean + S.D.(n) Mean + S.D.(n) Mean + S.D.(n)
V
Zn
21.0 + 2.46(30)
383 + 26.5(30)
61.2 + 5.85(30)
366 + 26.8(30)
34.2 + 6.46(23)
195 + 29.9(23)
9.92 + 1.76(26)
748 + 108(28)
a n = number of observations.
Sample: 1 = NIST 2704, Buffalo River Sediment
2 = NIST 4355, Peruvian Soil
3 = NIST 1085, Wear Metals in Oil
4 = 1:1 mixture - 2704 and 1634b, Trace Elements in Fuel Oil
5 = Solvent Recovery Waste
-------
TABLE 2 -- ICP Analysis Using SW-846 Method 3050 (/
-------
TABLE 2 — ICP Analysis Using SW-846 Method 3050 (/tg/g)
(Continued)
Sample
12345
Element Mean + S.D.(n)a Mean+S.D.(n) Mean+S.D.(n) Mean+S.D.(n) Mean+S.D.(n)
V
Zn
24.2 + 7.21(25)
393 + 60.7(27)
81.4 + 17.3(27)
401 + 49.2(27)
37.4 + 10.2(19)
207 + 24.3(20)
9.73 + 1.86(26)
747 + 120(27)
a n = number of observations.
Sample: 1 = NIST 2704, Buffalo River Sediment
2 = NIST 4355, Peruvian Soil
3 = NIST 1085, Wear Metals in Oil
4 = 1:1 mixture - 2704 and 1634b, Trace Elements in Fuel Oil
5 = Solvent Recovery Waste
-------
.TABLE 3 — GFAA Analysis Using Method 30XX (/*g/g).
Sample
SRM 2704 SRM 4355
Element Mean + S.D.(n)* % RSD NIST Value % Bias Mean+S.D.(n) % RSD NIST Value % Bias
As
Se
18.3 +
0.668 +
1.04 (12)
0.127 (10)
6
19
23.4
(1.1)
-22
-39
63.9
1.12
+ 3.
+ 0.
19 (12)
192 (12)
5
17
90
1
-29
+12
( ) Not certified.
a n = number of observations.
-------
TABLE 4 — ICP Analysis of SRM 1085 Wear Metals in Oil
Using Method 30XX (/ig/g)
Element
Ag
Al
Cr
Cu
Fe
Mg
Mo
Ni
Pb
Mean + S.D.
234 + 35.9
295 + 31.1
293 + 26.6
289 + 23.8
311 + 34.7
270 + 29.1
238 + 30.3
293 + 25.0
279 + 22.1
% RSD
15
10
9
8
11
11
13
8
8
NIST Value
(291)
296
298
295
300
297
292
303
(305)
% Bias
-20
0
- 2
- 2
+ 4
- 9
-18
- 3
- 8
( ) Not certified.
1-143
-------
TABLE 5 -- ICP Analysis of SRM 2704 Buffalo River Sediment --
Comparison of Methods 30XX and 3050 (/ig/g)
Element
Al(%)
B
Ba
Be
Ca(%)
Cd
Co
Cr
Cu
Fe(%)
Mg(%)
Mn
N1
Pb
Sr
V
Zn
3050
Mean + S.D.
1.31 + 0.367
55.4 + 25.8
85.3 + 17.9
0.682 + 0.209
1.83 + 0.200
3.32 + 0.436
11.1 + 2.75
83.3 + 14.0
83.2 + 11.0
3.06 + 0.308
0.850 + 0.120
472 + 57.2
37.7 + 5.15
147 + 16.6
35.0 + 7.04
24.2 + 7.21
393 + 60.7
% RSD
28
46
21
31
11
13
25
17
13
10
14
12
14
11
20
30
15
30XX
Mean + S.D.
1.18 + 0.137
34.6 + 9.31
77.7 + 5.90
0.562 + 0.068
2.00 + 0.383
3.19 + 0.613
10.7 + 1.46
81.7 + 5.33
80.3 + 6.92
2.96 + 0.214
0.810 + 0.047
460 + 25.7
36.4 + 2.52
143 + 9.46
33.0 + 2.05
21.0 + 2.46
383 + 26.5
% RSD
12
27
8
12
19
19
14
7
9
7
6
6
7
7
6
12
7
% Diff.
-10
-38
- 9
-18
+ 9
- 4
- 4
- 2
- 3
- 3
- 5
- 2
- 3
- 3
- 6
-13
- 2
1-144
-------
TABLE 6 -- ICP Analysis of Solvent Recovery Waste --
Comparison of Methods 30XX and 3050 (/*g/g)
Element
Ag
Al
B
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Mg
Mn
Mo
N1
Pb
Sr
V
Zn
3050
Mean + S.D.
1.25 + 0.253
0.141 + 0.030
39.6 + 10.8
513 + 132
<0.25
0.190 + 0.077
4.96 + 0.861
21.6 + 8.30
157 + 34.7
206 + 35.1
0.345 + 0.074
0.038 + 0.007
31.5 + 4.95
20.0 + 4.18
51.2 + 11.7
463 + 82.9
71.0 + 16.8
9.73 + 1.86
747 + 120
% RSD
20
21
27
26
• • •
40
17
38
22
17
21
18
16
21
23
18
24
19
16
30XX
Mean + S.D.
3.28 + 1.90
0.148 + 0.027
37.4 + 7.01
538 + 110
<0.25
0.219 + 0.093
4.90 + 1.04
21.9 + 5.04
161 + 23.8
208 + 32.6
0.316 + 0.052
0.038 + 0.009
31.7 + 5.05
19.1 + 3.36
50.9 + 9.96
437 + 70.3
71.1 + 13.5
9.92 + 1.76
748 + 108
% RSD
58
18
19
20
* • •
42
21
23
15
16
16
24
16
18
20
16
19
18
14
% Diff.
162
+ 5
- 6
+ 5
• * •
+15
- 1
+ 1
+ 2
+ 1
- 8
0
+ 1
- 4
- 1
- 6
0
+ 2
0
1-145
-------
MICROWAVE DIGESTION FOR ICP ANALYSIS
REGION V ALTERNATE TEST PROCEDURE
Marilyn Shannon, Dr. Gerald Payton, Paula Howard
uTVrtecT States Environmental Protection Agency-
Region V Central Regional Laboratory
536 South Clark Street
Chicago, Illinois 60605
AJBSTRA_C_T
The microwave oven in conjunction with pressurized
te-flon digestion vessels has reduced our sample
preparation time significantly- The Alternate Test
Procedure process, or other verification depending upon
the program requirements, are necessary for new
technology to enter our laboratory. The application
process and data necessary are briefly described.
There are several factors to consider when choosing
operating conditions and methodology. The operation of
the pressure vessel has an effect on whether the volume
remains constant. Pressure and temperature profiles
have a pronounced effect on the effectiveness of the
digestion on various waters. The method currently in
use in Region V contains several assumptions about
sample retention and temperatures achieved. The
assumptions about temperature have been validated. Some
experiments were carried out concerning the retention of
metals in the walls of the teflon vessels. There is
some retention but it is well below our current
detection limits. The method detection limits have not
changed much even though there is now dilution from the
acid added. (Dilution factor 1.22) This seems to imply
better precision. The data gathered to support the
Alternate Test Procedure application will be available
for examination.
INTRODUCTION
There has been a great deal of interest in using
microwave ovens as a heating source for sample
preparation. In our experience this interest is
justified; a significant time reduction in sample
preparation accompanied our adoption of the microwave
oven method. The time savings did not cause any loss in
detection limits or accuracy.
1-146
-------
In the EPA Central Regional Laboratory of Region V,
where this work was done, the methodology must be
defensible in court. Under different EPA programs there
are different requirements for method validation. The
NPDES validation is more formal than some of the
others. Approval for use in NPDES work requires an
Alternate Test Procedure approval. This approval
requires statistical validation of the test. My
inquiries determined that sample preparation, while not
a testing method, has such a great effect on the
analysis results that the stringent statistical
requirements also applied to it.
The validation required that five samples from six
different industries be analysed. Each of these samples
was prepared for analysis in quadruplicate by each of
the methods of digestion. Each sample had eight
digestions done and analysed. This made a total of two
hundred and forty separate analyses.
In the planning stages of the validation study we
considered the possibility of validating the digestion
for atomic absorption analysis as well as the
inductively coupled plasma. An estimate of a man year
of time for atomic absorption validation alone made us
reject this option. Thus our Alternate Test Procedure
Approval is only for preparation of samples for ICP not
for AA. The ICP work took about three months to
complete.
A value less than detection limit would tell us little
about the similarities or differences of the two
methods. All samples were analysed prior to use in the
validation study. Any element not present at detectable
levels was spiked into the sample at a level about five
to ten times the detection limit. The sample was then
digested by both methods and analyzed. We tried to
analyse the digestions associated with one sample as
close together as possible to minimise the instrument
variation in the analysis. The analysis was accompanied
by digested blanks and our usual instrument check
samples. The controls on the analyses were as stringent
as those for our usual work.
The samples were analysed on a Jarrell Ash Model 1160
Inductively Coupled Plasma. Our instrument reports
twenty five elements. The data was sorted by element
and entered into a spreadsheet. A floppy disk with the
spreadsheet as well as printed copies of the data were
sent for statistical analysis at EMSL Cincinnati. They
1-147
-------
-found twelve outliers. Upon investigation six of these
were -found to be typing errors. That left six thousand
pieces of data with only six outliers. In April of 1988
the data was approved as supporting an Alternate Test
Procedure Approval and we began using the method on a
regular basis.
The conditions for the digestion are not as severe as
are now being recommended. The conditions used were ten
minutes at 1007. power. The term 10O7. power is somewhat
ambiguous. We have found that there has been
significant losses in the wattage in our microwave since
this study was done. The initial wattage found when
the oven was a year old was 560; the most recent wattage
was 540. The wattage varies slightly with the position
of the beaker in the microwave cavity. An average of
several values is recommended to get a reliable number.
The milder conditions were chosen to avoid the venting
of the vessel. The information on the pressure and
temperature profiles was not as extensive as it is now.
Through trial and error we determined that there was no
venting at these settings. Our preliminary experiments
indicated that the digestion was as complete as the hot
plate digestion. The data gathered to for the Alternate
Test Procedure Application supports this conclusion.
Venting was undesirable because of questions about the
composition of the vented material and the problem of
correcting for the lost volume if it were only steam.
Currently a specially designed vessel can be used to
catch and condense overflow, but they were not widely
publicized three years ago.
The temperature profile as shown in the graph plotted
from CEM data shows that the solution reaches about 95°
under the conditions of the digestion. This is about
the same as the tempertures recommended for the hot
plate digestion. The hot plate/beaker digestion does
not digest oily waters well and neither does this
microwave digestion. Other papers have demonstrated
that a temperature of 160 to 165° C must be attained to
digest oils. As you can see this digestion does not
approach this temperature.
Recently there was some discussion about retention of
metals in the walls of the Teflon digestion vessels. In
fact this retention probably occurs but not at a level
that interferes with our analysis. Some experiments
were conducted with well used vessels. The normally
1-148
-------
washed and rinsed vessels were heated with the digestion
acids only. Some were at 10 ml and some at 20ml „ The
20 ml volumes showed nothing upon 1CP analysis. The 10
ml volumes showed a little iron and copper. This
contamination was well below our detection limit if the
acid had been diluted with 50 or even 25 ml of sample.
There has been much less difficulty with contamination
in the digestion blanks since we started using the
Teflon microwave vessels instead of glass beakers.
Filter paper has been a much larger source of
contamination than the vessels or even acid washed
beakers.
Although there is a dilution factor of 1.22 of each
sample, remarkably the method detection limits for the
ICP elements have not changed. This indicates the
greater stability and reproduction of the microwave
diqesti on.
Unfortunately, the permission to use this method for
official analytical work is currently limited to our
laboratory. Other laboratories may use our data to
support their application for test procedure approval.
They will need to supply some data to confirm that the
method works as well in their laboratory as it did in
ours to the Reagional Quality Assurance Branch. The
local Quality Assurance Branch can supply you with
further information. The last I heard the microwave
digestion was to be included in the next Statement of
Work for the CLP program.
SUMMARY
The microwave digestion, has much to recommend it. The
microwave oven is quicker, more reproducable than the
hot plate and less easily contaminated. At the mild
conditions of this digestion there is no venting or loss
of sample. The data support the conclusion that the
digestion does as well as the beaker hotplate digestion.
There have been no contamination problems with the
repeated use of the vessels with only washing between
uses. We have certainly been pleased with the microwave
oven digestion in our laboratory.
1-149
-------
A COMPARISON STUDY OF QUALITY CONTROL PERFORMANCE BETWEEN
ICP-MS METHOD 6020 AND THE ICP-AES AND GFAA
SPECTROSCOPY METHODS
KEITH ALECKSON AND FOREST GARNER, LOCKHEED ENGINEERING AND SCIENCES CO.,
1050 E. FLAMINGO RD, LAS VEGAS, NV 89119; MICHAEL KURD, U.S.E.P.A.,
ANALYTICAL OPERATIONS BRANCH, 401 M STREET S.W., WASH. D.C. 20460; DR.
LARRY BUTLER, U.S.E.P.A., ENVIRONMENTAL MONITORING SYSTEMS LABORATORY, P.O.
BOX 93478, LAS VEGAS, NV 89193-3478.
ABSTRACT. A multi-laboratory study has been conducted to measure
the comparability of the Inductively Coupled Plasma Mass
Spectrometry (ICP/MS) method 6020 with the present inorganic
methods used by the Contract Laboratory Program (CLP) of the U.S.
Environmental Protection Agency. Performance Evaluation (PE)
samples made up of solid and water matrices were split and sent
to a group of CLP laboratories for analysis by routine methods
and to another group of laboratories for analysis of the same
elements-by ICP/MS. The results of this study were used to
determine the comparability of ICP/MS to the Inductively Coupled
Plasma Atomic Emission (ICP-AES) and Graphite Furnace Atomic
Absorption (GFAA) spectroscopy methods. The purpose of the
presentation is to compare the performance of the quality control
(QC) analyzed by the ICP/MS method with that of the routine
methods. The following QC parameters have been reviewed and
compared: spikes, duplicates, interference check sample results,
as well as the linear range and detection limit data. The spike
and duplicate QC comparison supply accuracy and precision
performance, respectively, for both methods. The QC results for
the Interference Check Solutions have been evaluated for each
method's potential for spectral interference. Comparison of the
instrument detection limits and linear ranges will provide an
evaluation of the overall sensitivity and operating range for
each method. The results of this QC study will then be compared
with similar QC data for the CLP from the past year. This
presentation will provide information on the expected QC
performance of the method with comparison to other well
characterized methods to the users of the ICP/MS method.
1-150
-------
PERFORMANCE OF INDUCTIVELY COUPLED PLASMA MASS
SPECTROMETRY METHOD 6020
Thomas A. Hinners. Edward M. Heithmar, Larry C. Butler,
Environmental Monitoring Systems Laboratory, P.O. Box 93478,
Las Vegas, Nevada 89193-3478; Michael L. Kurd, Office of
Emergency and Remedial Response, 401 M Street, S.W.,
Washington, District of Columbia 20460; Dave E. Dobb, Guy A.
Laing, Lockheed Engineering and Sciences Company, 1050 East
Flamingo Road, Suite 120, Las Vegas, Nevada 89119
ABSTRACT
Inductively coupled plasma mass spectrometry (ICP-MS) offers
detection limits below a part per billion for multi-elemental
analysis with the convenience and speed of nebulizer sample
introduction. Digestion using hydrobromic acid (HBr) instead
of hydrochloric acid (HC1) avoids adding chloride
interferences to the ICP-MS measurements for arsenic and
vanadium, but gives reduced recoveries for silver compared
with HCl. An interlaboratory study has been completed
comparing several digestion procedures as well as analyses by
atomic absorption spectroscopy (AAS), inductively coupled
plasma atomic emission spectroscopy (ICP-AES), and ICP-MS for
water, fly ash, sediment, industrial sludge and soil samples.
The study required nearly 36,000 analyses. Analytical
precision using ICP-MS is more variable than with ICP-AES for
several elements. This lack of precision may have resulted
from the need to perform additional dilutions of samples for
ICP-MS analysis. Since results of the analysis of standard
reference materials by ICP-MS for selenium exceeded the
reference values provided by the National Institute of
Standards and Technology [formerly the National Bureau of
Standards (NBS)], more appropriate corrections for this
element are still required. The preliminary results from this
study indicate that it is suitable for many elements in solid
wastes, but do not support using ICP-MS for the determination
of potassium, selenium, silver, sodium or vanadium in solid
wastes. Silver data from ICP-AES also showed wide
variability.
NOTICE: Although the research described in this article
has been supported by the U.S. Environmental Protection
Agency, it has not been subjected to Agency review and
therefore does not necessarily reflect the views of the
Agency. No official endorsement should be inferred.
1-151
-------
INTRODUCTION
To evaluate the performance of inductively coupled plasma mass
spectrometry (ICP-MS) Method 6020 for waste analysis, an
interlaboratory study was conducted by the Environmental
Monitoring Systems Laboratory in Las Vegas (EMSL-LV) for the
Analytical Operations Branch of the Office of Emergency and
Remedial Response. The participants in this interlaboratory
study included ICP-MS manufacturers, private laboratories, and
government facilities. The participating laboratories
received both digested and undigested samples. Conseguently,
performance data were obtained with and without sample
preparation as a variable. Variations in the sample
preparation were investigated because the Contract Laboratory
Program specifies different procedures for furnace AAS and
non-furnace AAS analyses and because the amount of
hydrochloric acid added in sample preparation affects the
amount of ICP-MS interference for arsenic and vanadium.
While it was not part of the interlaboratory study, an
alternate HBr digestion procedure was tested as a means to
minimize ICP-MS chloride interferences. Hydrochloric acid
(HC1) is used in many digestion procedures for its
solubilizing benefits. However, chloride contributes
interferences to the ICP-MS measurements for the
environmentally important elements arsenic and vanadium.
Substituting hydrobromic (HBr) or hydroiodic (HI) acid for
HC1 in the digestion of samples might provide the solubilizing
benefit of a halide without adding to the chloride
interference on arsenic and vanadium. Silver is one element
that is especially sensitive to the chloride level because of
its solubility (Figure 1). According to the literature, HBr
and HI solubilize more of the corresponding silver halides
than HC1 on an equal-molar basis (Table 1).
Figures 2 and 3 show that HBr and HI produce less interfering
halo-oxide and halo-argon species than does HC1. But, more
importantly, these bromide and iodide species do not interfere
with arsenic or vanadium measurements. In the case of
bromide, molybdenum isotopes at mass-to-charge ratio (m/z) 95
and 97 are affected by BrO+ ions, but the most abundant
isotope of molybdenum is measured at m/z 98. Bromide ions
combined with argon affect the antimony isotope at 121 m/z
but not the 123-m/z isotope. HI produces interferences on
only one neodymium and one erbium isotope. Unfortunately, in
testing the HI, it was found that iodide is converted to
iodine even by dilute nitric acid. Consequently, the use of
HI was discontinued and HBr was used in place of HC1 in the
digestion of 4 samples spiked with silver and antimony. The
1-152
-------
results of this feasibility testing with single portions
of the 4 samples are shown in Table 2. Although the HBr and
HCl recoveries are similar for antimony, silver recoveries
with HBr are much lower than with HCl for 3 of the 4 wastes
analyzed. Therefore, while HBr does provide lower background
levels than HCl for arsenic and vanadium, it does not seem to
provide equivalent performance for other elements. The use
of HBr instead of HCl could be useful for samples containing
arsenic or vanadium where silver is not present. When HBr is
used, selenium measurements at m/z 82 are affected by HBr*
ions, but selenium can be measured at m/z 77 or 78 without any
contribute from bromide-containing ions.
Preliminary examination of the interlaboratory data reveals
more variability in the ICP-MS data than in the ICP-AES data
for several of the major elements including aluminum, calcium,
potassium, magnesium, sodium, and zinc. This lack of
precision (by ICP-MS in comparison to ICP-AES) may have been,
in part, the indirest result of the high sensitivity of ICP-
MS; laboratories were required to dilute samples extensively
or to use signal suppression for concentrations above a couple
of mg/L. This additional sample handling or signal alteration
may have increased the measurement variability. Vanadium ICP-
MS results are particularly variable, possibly because the
correction for CLO+ ions is inconsistent. Most of the
selenium values by ICP-MS exceed those obtained using furnace
atomic absorption spectroscopy. ICP-MS analysis of River
Sediment (NBS 1645) and Estaurine Sediment (NBS 1646) gave
selenium values exceeding those provided by NBS. Clearly,
there is a need for appropriate corrections for selenium
determinations by ICP-MS.
The preliminary results from this interlaboratory study
indicate that ICP-MS Method 6020 is suitable for many
elements, but do not support the use of ICP-MS for the
determination of potassium, selenium, silver, sodium or
vanadium in solid wastes. The interlaboratory relative
standard deviation values for these 5 elements (Table 3)
exceed 30% by ICP-MS for one or more of the tested samples and
are freguently several times higher than the standard
deviations obtained by the reference techniques. With proper
corrections applied, this ICP-MS database may provide data
comparable to furnace AAS for selenium and to ICP-AES for
vanadium.
1-153
-------
TABLE 1. Silver halide solubilities8
solution silver concentration (mg/L)
compound water HX (IN)
AgCl 1.45 11
AgBr 0.077 115
Agl 0.0012 29,400
aPublished data(no standard deviations provided)by Linke
and Seidell, Solubilities of Inorganic and Metal Organic
Compounds. American Chemical Society, Washington, B.C., 1958.
TABLE 2. Comparison of HCL and HBr digests spiked with 500
ug/L silver and antimony
concentration (ug/L)
sample silver antimony
HC1 HBr HC1 HBr
electroplating sludge
EPA waste soil
mixed waste sludge
river sediment (NBS 1645)
415
447
445
453
51
31
78
442
162
105
91
189
150
100
80
205
1-154
-------
TABLE 3. Comparison of relative standard deviations between
the tested methods
percent relative standard deviation
potassium
Sample ICP-AES ICP-MS
QC Standard #1
QC Standard #2
electroplating sludge
EPA waste soil
mixed waste sludge
river sediment (NBS 1645)
estuarine sediment (NBS 1646)
fly ash (NBS 1633a)
spiked soil
13
11
4.7
5.7
13
17
6.2
5.7
7.7
8.0
7.6
36
23
53
24
25
37
33
selenium
GFAAS ICP-MS
12
9.6
78
31
40
38
31
17
19
16
15
130
36
240
200
260
29
110
ICP-AES is inductively coupled plasma atomic emission
spectroscopy.
ICP-MS is inductively coupled plasma mass spectrometry.
GFAAS is graphite furnace atomic absorption spectroscopy.
QC is quality control, and BD is below detection.
1-155
-------
TABLE 3. (Continued)
percent relative standard deviation
silver
sample ICP-AES
QC Standard #1
QC Standard #2
electroplating sludge
EPA waste soil
mixed waste sludge
river sediment (NBS 1645)
estuarine sediment (NBS 1646)
fly ash (NBS 1633a)
spiked soil
69
43
29
60
30
38
33
BD
33
ICP-MS
51
9.0
29
39
16
17
10
9.0
19
sodium
ICP-AES ICP-MS
12
5.1
7.7
14
9.5
9.2
3.5
8.6
5.1
20
20
23
28
31
26
24
22
27
ICP-AES is inductively coupled plasma atomic emission
spectroscopy.
ICP-MS is inductively coupled plasma mass spectrometry.
GFAAS is graphite furnace atomic absorption spectroscopy.
QC is quality control, and BD is below detection.
1-156
-------
TABLE 3. (Continued)
percent
sample
QC Standard #1
QC Standard #2
electroplating sludge
EPA waste soil
mixed waste sludge
river sediment (NBS 1645)
estuarine sediment (NBS 1646)
fly ash (NBS 1633a)
spiked soil
relative standard deviation
vanadium
ICP-AES ICP-MS
6.3
10
28
12
16
11
7.2
4.3
7.4
9.5
14
33
24
19
44
25
21
26
ICP-AES is inductively coupled plasma atomic emission
spectroscopy.
ICP-MS is inductively coupled plasma mass spectrometry.
GFAAS is graphite furnace atomic absorption spectroscopy.
QC is quality control, and BD is below detection.
1-157
-------
7.0^
1-7.55
1x10
-0.108
HCL(M)
Figure 1. Solubility of silver chloride
in HCL at 25°C.
-------
£ 80K-
(0
c
0)
£ 60K-
c
o
- 40K-
20K-
CIO+
BrOH
h
10 +
51V 95Mo 143Nd
53Cr 97Mo
en
m
Figure 2. ICP-MS halogen-oxide ion intensities
for 0.05 N HX.
-------
V)
c
c
o
10K-n
8K-
6K-
4K-
2K-
CIAr+
BrAr+
IAr+
1 1
75As 119Sn 167Er
77Se 121Sb
Figure 3. ICP-MS halogen-argon ion intensities
for 0.05 N HX.
-------
ICP-MS METHOD 200.8
THE DETERMINATION OF TRACE ELEMENTS IN WATERS AND WASTES
Stephen E. Long, Technology Applications Inc. 26W Martin Luther King
Dr., Cincinnati, Ohio, 45219;
Theodore D. Martin EMSL, USEPA, 26W Martin Luther King Dr. Cincinnati,
Ohio 45268.
ABSTRACT.
EMSL-Cincinnati proposed Method 200.8 has been developed for the
determination of trace elements in waters and wastes by inductively
coupled plasma - mass spectrometry (ICP-MS). Some of the components of
the method which have received special consideration are sample
preparation procedures, recommended analytical masses, procedures for
sample analysis and calibration, spectral interference corrections and
matrix interference correction by the use of internal standardization.
Development of the method has been supported by the production and
review of a draft version and the design and implementation of a single
laboratory validation study to assess the performance of the method for
a range of typical sample matrices.
INTRODUCTION
ICP-MS has several characteristics which make it well suited to rapid
multi-element determinations at the trace level. Among these are element
specificity, extensive element coverage and high sensitivity. In the
past few years therefore, ICP-MS has offered an alternative to graphite
furnace atomic absorption (GFAA) and inductively coupled plasma -
emission spectrometry (ICP-ES) for the analysis of environmental
samples. This has resulted in an incentive to develop ICP-MS methodology
applicable to the interests of the EPA, and to determine the performance
and comparability of the technique using this methodology.
The design of such methodology requires careful consideration. The main
goals in the development of this method were to provide enough
flexibility such that the method could be used equally successfully with
existing commercial instrumentation and to accommodate individual
analytical practices in the use of the technique, yet to build in
sufficient specified procedures to maintain consistent inter-laboratory
performance. Two standard benchmarks for measuring performance are
precision and accuracy. Method design will greatly affect the accuracy
attained, particularly on a consistent basis. There is a requirement
therefore, to address some of the critical factors which will influence
accuracy, such as calibration, matrix and spectral interference
correction and quality control. In contrast, precision will be
influenced more by the quality of the instrumentation concerned and the
implementation of good laboratory practices, than by method protocols,
although these will be important.
1-161
-------
DISCUSSION
Method 200.8 describes procedures for the determination of nineteen
elements in waters and wastes. These elements are aluminum, antimony,
arsenic, barium, beryllium, cadmium, chromium, cobalt, copper, lead
manganese, molybdenum, nickel, selenium, silver, thallium, uranium,
vanadium and zinc. The majority of these are on the EPA's priority
pollutant list. Molybdenum and uranium have been included because of the
increasing interest in these elements in drinking water and other
environmental materials. Instrument detection limits for these elements
are listed in Table 1 together with the analytical isotopes which are
recommended in the method. The isotopes were chosen to minimize isobaric
elemental interferences and known polyatomic ion interferences while
providing acceptable detection limits. Of these elements, only selenium
has a detection limit greater than 1 jug/1.
For the analysis of solid materials or for the determination of total
recoverable elements, sample digestion procedures were chosen to provide
compatibility with proposed GFAA and existing ICP procedures, ie using a
mixture of nitric and hydrochloric acids for solubilization. Samples
prepared by this procedure can be analyzed by any of the three
techniques. The concentration of hydrochloric acid in the final
preparation is 3$ (v/v). This solution is diluted by a further factor of
five (0.6% HCl) for ICP-MS analyses in order to minimize chloride
TABLE 1. ICP-MS Instrument Detection Limits.
Element
Aluminum
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Cobalt
Copper
Lead
Manganese
Molybdenum
Nickel
Selenium
Silver
Thallium
Uranium
Vanadium
Zinc
Recommended
Analytical Mass
27
121
75
137
9
114
52
59
63
208
55
98
60
82
107
205
238
51
66
IDL
(jig/D
0.05
0.08
0.9
0.5
0.1
0.07
0.07
0.03
0.03
0.08
0.1
0.1
0.2
5
0.05
0.09
0.02
0.02
0.2
1-162
-------
interferences on arsenic, vanadium and chromium. However, data
corrections for these interferences are still necessary. The sample
preparation procedures were also designed to produce a maximum dissolved
solids level of 0.2$ (w/v), to attempt to reduce instrument drift and
matrix effects from sampling effects at the ICP-mass spectrometer
interface.
Prior to running samples, demonstration of instrument stability and mass
calibration is required by the method. This is achieved by the running
of a tuning solution, with associated acceptance criteria. This tuning
solution consists of beryllium, magnesium, cobalt, indium and lead. The
isotopes of magnesium and lead are used for adjusting the resolution at
low mass and high mass respectively. Following this, the instrument is
calibrated for response prior to the running of samples. The method
requires that the calibration standard is run as a surrogate sample
after every ten analytical samples. An acceptance window of i 10$ of the
initial calibration is required. If this window is exceeded by any
element, a mandatory recalibration is carried out.
Internal standardization is used in the method to correct for instrument
drift or matrix effects. Internal standards accepted for use in the
method are lithium, scandium, yttrium, rhodium, indium, terbium,
holmium, lutetium and bismuth. For full mass scan determinations, a
minimum of three internal standards are required, although five are
recommended. The absolute response of these internal standards cannot
exceed a window of ± 50$ of the original response. If this limit is
exceeded, the analysis has to be terminated and the cause of the drift
investigated, before restarting any analyses.
Interference correction on acquired data is also a requirement of the
method. ¥ith a few exceptions, the analytical masses recommended for use
with the method are not subject to major spectral interferences.
Interferences of significant concern are those of chloride species on
arsenic, vanadium and chromium and those of molybdenum oxide on cadmium.
However, other interferences have to be monitored where possible by the
examination of isotope ratios, and a list of required masses which have
to be monitored are included in the method to accommodate this.
Standard quality control protocols are used in the method. In the
present version these include an initial demonstration of laboratory
performance followed by method quality control involving the use of a
laboratory reagent blank, laboratory fortified blank, spiked sample
duplicates and the analysis of a quality control sample. Control limits
for acceptance criteria are presently under consideration.
A single laboratory validation study is being completed to assess
the precision and accuracy of the proposed method. This utilizes a
representative set of five waters and three solids consisting of EPA
hazardous waste soil, EPA Electroplating Sludge (WP286) and NBS SRM 1645
(river sediment) . Most of the materials have an established database
which can be used to indicate the performance of the sample preparation
1-163
-------
procedures. These samples have been digested by the procedures described
in the method and sample splits analyzed in parallel by ICP-MS (VG
PlasmaQuad) using method 200.8 and by ICP-ES (jarrell Ash Model 1160
Atomcomp). By using sample splits, sample variability effects were
eliminated allowing a direct comparison of the performance of the two
techniques. Accuracy has been established in each matrix case by the
analysis of spiked duplicates at two concentration levels. Precision has
been assessed by analyzing five replicate digests of each matrix.
SUMMARY
EPA method 200.8 provides procedures for the determination of nineteen
elements in waters and wastes by ICP-MS. A draft of the method has been
prepared, reviewed and tested using a single laboratory validation
protocol. Data obtained from the validation study will be considered and
the method will be modified as necessary before release to external
review.
1-164
-------
SELECTED COMPARISONS OF LOW CmCENTRATION MEASUREMENT
CAPABILITY ESTIMATES IN TRACE ANALYSIS: METHOD
DETECTION LIMIT AND CERTIFIED REPORTING LIMIT
LfiNG, KENNETH T. AND STOTZ, MARTIN H. , U.S. ARMY TOXIC AND HAZARDOUS MATERIALS
AGENCY, ABERDEEN PROVING GROUND, MARYLAND 21010-5401; GRANT, C.L., CHEMISTRY
DEPARTMENT, UNIVERSITY OF NEW HAMPSHIRE, DURHAM, NEW HAMPSHIRE; HEWITT, ALAN D.
AND JENKINS, THOMAS F., GEOCHEMICAL SCIENCES BRANCH, U.S. ARMY COLD REGIONS
RESEARCH AND ENGINEERING LABORATORY, HANOVER, NEW HAMPSHIRE
ABSTRACT. Two large data sets were obtained over a four-day period for graphite
furnace atomic absorption spectroscopic measurement of copper (Cu) and
reversed-phase high performance liquid chromatographic determination of
dinitrobenzene (DNB) at a number of concentrations near the lower limit of
measurement. Low concentration measurement capability estimates for each analyte
were obtained using the U.S. Environmental Protection Agency's method detection
limit (MDL) protocol and the U.S. Army Toxic and Hazardous Materials Agency's
certified reporting limit (CRL) protocol. For DNB, analytical variance was found
to be homogeneous over the concentration range examined and MDL estimates were
independent of concentration over the range of concentration examined. MDL
estimates varied by as much as a factor of three from day-to-day emphasizing the
uncertainty in these estimates. CRL estimates varied to about the same extent and
were numerically quite similar to MDLs when equivalent alpha and beta risks were
used. For Cu, analytical variance was found to be proportional to concentration.
Thus CRL estimates were very dependent on the concentration range examined. MDLs
were less sensitive to this problem. Recommendations regarding the choice of
target reporting limits for the CRL protocol were made. The influence of risk
assumptions on both MDL and CRL estimates were examined and recommendations for
modifications to both procedures made to incorporate an operational beta-risk
appropriate to the problem at hand. A case was made for using outlier tests to
edit data used to estimate low concentration measurement capabilities.
1-165
-------
REPORT OF AN INTERLABORATORY STUDY COMPARING EPA SW 846
METHOD 30501 AND AN ALTERNATIVE METHOD FROM
THE CALIFORNIA DEPARTMENT OF HEALTH SERVICES
DAVID E. KIMBROUGH, PUBLIC HEALTH CHEMIST, AND JANICE WAKAKUWA,
SUPERVISING CHEMIST, CALIFORNIA DEPARTMENT OF HEALTH SERVICES, SOUTHERN
CALIFORNIA LABORATORY, 1449 W. TEMPLE STREET, LOS ANGELES CALIFORNIA
90026-5698.
ABSTRACT
The existing EPA method for the analysis for toxic elements in solid
matrices, SW 846 method 3050, cannot accurately or precisely determine
(i.e., within SW 846 quality control limits) the concentrations of either
antimony or silver and can do so for barium only at lower concentrations.
In response to this situation an alternative method was developed by the
Southern California Laboratory of the Department of Health Services which
could simultaneously solubilize all of the above elements over a broader
range of concentrations and can solubilize all other regulated elements
with equal or superior success.
As part of the process of validation of this alternative method, an
interlaboratory study was conducted with ten public and private
laboratories comparing existing methods with the alternative method. A
wide range of analytical instrumentation was used including Flame Atomic
Absorption Spectroscopy (FAA), Graphite Furnace Atomic Absorption
Spectroscopy (GFAA), Inductively Coupled Plasma -Atomic Emission
Spectroscopy (ICP-AES), and Inductively Coupled Plasma Mass
Spectroscopy (ICP-MS). The results of the study clearly show
statistically significant differences between method 3050 and the
alternative method proposed in this paper.
INTRODUCTION
0*5 /
Federal ' and State laws list antimony, arsenic, barium, beryllium,
cadmium, chromium, cobalt, copper, lead, molybdenum, nickel, selenium,
silver, thallium, vanadium, and zinc (hereafter referred to as the
"target elements") as toxic elements. At the present time there are no
validated EPA digestion methods for the preparation, by a single
digestion, of soils, sludges, sediments and solid waste samples for the
analysis of the target elements for determination by FAA, GFAA, or ICP.
EPA method 3050 in SW 846 is satisfactory for most of the elements listed
above but not for antimony or silver. Consequently, the EPA, in the
third edition of SW 846, removed silver and antimony from method 3050's
list of validated elements; a decision supported by research here and in
other laboratories5'6. The SW 846 establishes a 75% to 125% recovery for
spikes or reference materials and reproducibility of less than 20%
relative percent difference7. Silver and antimony cannot be recovered
within these limits using method 3050.
1-166
-------
Hazardous materials and environmental laboratories are routinely
requested to analyze solid phase samples for antimony and silver, usually
in conjunction with the other target elements, by legally recognized
methods. This leaves the laboratories in the position of using other
methods in addition to method 3050, which is time consuming and legally
ambiguous.
In response to this problem, a new method has been developed by the
Southern California Laboratory (SCL) of the California Department of
Health Services (DOHS), hereafter referred to as the SCL method. The
proposed method is capable of solubilizing both antimony and silver along
with the other target elements, using a single digestion procedure. Data
accumulated at SCL indicate this is an effective method for elemental
analysis of solid matrices.
Although results from the SCL have demonstrated this new method to be
superior method to 3050 , the possibility of analyst or laboratory bias
still exists. To explore these possibilities, an inter-laboratory study
was designed to compare the two methods with outside laboratories and
analysts.
EXPERIMENTAL SECTION
A) Analytical Methods. (1) EPA SW 846 method 3050 was used as
designated. It directs that 1.00-2.00 grams of sample be digested first
with 10 mL 1:1 (v/v) nitric acid at 95 C for 15 minutes, then 5 mL cone.
nitric acid is added and is refluxed for 30 minutes. This step is
repeated until the sample no longer changes in appearance. The digestate
is then concentrated to 5 mL. The sample is treated with no more than 10
mL of 30% hydrogen peroxide. Finally, 5 mL of cone. hydrochloric acid
and 10 mL of deionized water are added and the sample is refluxed for 15
minutes. The sample is then either filtered (Whatman 41 or equivalent)
or centrifuged (2-3,000 rpm for 10 min) and the filtrate (or supernatant)
is collected in 100 mL volumetric flask and analyzed by either FAA or
ICP.
(2) The SCL method calls for 1.00 4.00 grams of sample to be digested
in a mixture of 20 mL cone. hydrochloric acid and 5 mL of cone, nitric
acid at ambient temperature. The sample and reaction mixture are slowly
heated to 95 C to prevent an overly vigorous reaction. The digestion is
continued until the disappearance of N02 (reddish brown) fumes and no
more change in appearance. The digestate is then evaporated to near
dryness, washed with about 40 mL of l%nitric: l%hydrochloric acid (v/v)
solution and filtered (Whatman 41 or equivalent). The filter paper is
washed with no more than 5 mL hot (95 C) cone. hydrochloric acid, and
then 20 mL hot deionized water, all of which is collected into one flask
(this will be referred to as the primary filtrate). The filter paper and
residue are placed back in the digestion vessel, 5 mL cone, hydrochloric
acid are added and refluxed at 95 C until the filter paper disintegrates
(approx. 10-15 min.) The disintegrated paper is then washed with
deionized water and again filtered (this filtrate will be referred to as
1-167
-------
the secondary filtrate.) These filtrates are then analyzed by either ICP
or FAA. The results are combined as follows:
(X1 + X2) * V/w - Xt*V/w - C
(equation 1)
X i = concentration of element x in primary filtrate
X o = concentration of element x in secondary filtrate
X = total concentration of element x = X-^ + ^
V = volume of volumetric flask used for both primary and secondary
filtrate.
w = weight of sample taken
C = concentration of element x in sample.
B) Statistical Methods. Four statistical tests will be used for
analyzing the data. For comparing the variance between methods, SNEDECOR
criterion using the F-test8'^ will be used. For comparing means between
methods, the Student t-Test10 is used. Testing for outliers within a
method, a variation of the t-test was used . The term percent relative
standard deviation (%RSD) is defined as the standard deviation divided by
the mean times one hundred.
C) Laboratories. Eight commercial environmental laboratories were
selected to participate. Selection was based on their accreditation with
California DOHS, ability to perform hazardous materials analysis and
their willingness to participate.
D) Instrumentation. The following ICP-AESs were used, Perkin Elmer 5500,
Perkin Elmer Plasma II, Leeman ICP, and Thermo Jarrell Ash 61, A Vg
ICP-MS was used at one location. The following FAA and GFAA instruments
were used, Perkin Elmer 3030, IL 257, Varian 20, and a Thermo Jerrall Ash
V12.
E) Materials. Six solid phase samples and two liquid samples were
prepared for the study. The solid phase samples were designated from A
to F The liquid samples were designated "PQL solution" and "Unknown
solution". Samples A, B, and C are composites of samples from several
industrial sites in southern California with quantifiable levels of some
of the target elements in each. Since these materials were homogenized,
they represent actual field samples and as such have a greater inherent
variance than materials produced in the laboratory. Samples D, E, and F
were spiked samples with all of the target elements in each sample. Each
analyte was spiked at three different concentrations, designated low,
middle, and high. These materials were spiked and homogenized in the
laboratory and are very homogeneous. The concentrations in these samples
are referred to as the "true" value.
A liquid sample was prepared at a concentration of 1.00 ug/ml for all of
the target elements except beryllium, which has a concentration of 0.20
ug/ml in 5% nitric acid/deionized water. This was designated "PQL
Solution." This solution was analyzed by all participating laboratories
1-168
-------
to verify that the instruments used could, in fact, accurately quantify
at that level.
A second liquid sample was prepared for the target elements at
concentrations ranging from 10 to 90 ug/ml in a solution of approximately
20% acid (nitric and hydrochloric in water.) This solution was designated
as an "Unknown solution".
F) Analytical Lines. Each laboratory selected the analytical lines they
wished to use for ICP- AES, FAA, and GFAA. Table I lists all of the
analytical lines used with the number of laboratories using each by
element.
G) Study Design. The study is divided into four parts: quality control,
precision, accuracy, and background. Copies of EPA SW 846 method 3050,
the SCL method, instruction sheets and reporting sheets were sent out
with each set of samples. Each lab was supposed to use only these two
methods. One laboratory , however, used its own modification of method
3050 for instrumentation reasons.
1) Quality Control. To meaningfully compare the accuracy and precision
of two methods through an interlaboratory study, it is important to
control for the variance caused by the inherent differences between the
laboratories. The primary sources of this variance are differing levels
of analyst skill, types of analytical instruments and choice of
spectrographic lines. Three steps were taken to assess the contributions
of these factors to the final results.
a) A method blank was run for each method by each laboratory. This
consists of running the entire digestion procedure using empty glassware.
b) A Practical Quantitation Limit (PQL) was established at 1.0 ug/mL for
all elements except beryllium, which had a limit of 0.2 ug/mL. No value
below this limit was to be reported. This was to provide a control for
the differences in sensitivities of the various analytical instruments.
c) The "Unknown Solution" was also analyzed by every laboratory. Since
this sample was not digested but run directly on the instruments, any
variance from this sample is due entirely to the instrumentation. This
allowed us to quantify the amount of variance contributed by these
differences.
2) Precision (Round I). The objective of this round of study is to
compare the variance or precision of each method. Samples A, B, and C
were distributed to the laboratories. These samples were to be analyzed
in triplicate by each method. The results from each laboratory were
checked to see if the triplicates met the SW 846 precision limits for
duplicates: 20% for each target element. The results from all the
laboratories were averaged and the overall variance between methods was
compared using the statistical methods mentioned above.
3) Accuracy (Round II). The goal of this round is to compare the
1-169
-------
accuracy of each method. Samples D, E, and F were analyzed once by each
laboratory and method. Each sample had a "true" value equal to the
spiked quantities. A comparison between the methods was done by averaging
the results from all laboratories and calculating the means for each
method. These were then compared to the "true" value for the samples.
The %RSD was also compared between methods.
4) Background. A laboratory survey was done using a set of questions
about the state of the laboratory. The subjects included sample load,
division of labor, and staff qualifications. The goal of this survey was
to determine the general state of the participating laboratories.
5) Criteria for Evaluation. Five factors have been identified for
assessing the relative performance of digestion methods: accuracy,
precision, range of elements, linear range of extraction and ease of use.
a) Accuracy. The mean value must fall within EPA SW 846 method 6010
8.5.2 limits. These state that a matrix spike must be recovered within
75% to 125% of the known value.
b) Precision. Replicate values of a method must fall within the limits
set by EPA SW 846 which is that duplicate samples should be within 20%.
c) Range of Elements. Methods that can solubilize more of the target
elements will be of greater use to analysts.
d) Linear Range. The broader the linear range of a method, the better is
its performance.
e) Ease of Use. The simpler a method is, the fewer the mistakes that
will possibly be introduced by the analyst.
RESULTS
Quality control. None of the laboratories showed any significant
contamination in their method blanks and most were able to accurately
quantify the target elements at the PQL.
Listed on Tables II, III and IV are the %RSD for each element from the
"Unknown solution". This number is interpreted as the contribution to
the overall variance of each analyte by the differing instrumentation
employed.
Precision (Round I). Using the F-test to compare variances requires
similar means. As can be seen from table II the high concentrations of
barium in sample A and the antimony in sample C are of significantly
different means and thus cannot be compared. Despite this, the barium in
sample A still had similar %RSDs between methods. Otherwise, all of the
target elements gave statistically similar variances.
Accuracy (Round II). The data from this round is listed on tables III,
IV, and V Antimony and silver had lower recoveries in samples E and F
1-170
-------
by method 3050 than in the SCL method. Barium in sample A was
significantly lower in method 3050 than in the SCL method. All of the
other target elements had means that were well within the SW 846
limitation of 25% of the "true" value using both methods. There were no
significant differences in mean values between methods for any target
element except those listed above.
LIMITATIONS. There were three errors in the execution of this study.
First, beryllium was not run in round one. Second, beryllium and cadmium
were spiked at levels that were too low in sample F of round II. Third,
the sample vials containing the Unknown solution were contaminated with
zinc, so instrumental variance for zinc is not known. These limitations
do not, however, impact the overall conclusions of this study.
DISCUSSION
Accuracy. The accuracy of the mean values in round II for both methods
was well within SW 846 limitations. However, three exceptions: low and
middle concentrations of antimony by method 3050, high concentrations of
barium by both methods, high and middle concentrations of silver by
method 3050, and high concentrations of silver by the SCL method. The
vast majority of mean values were above 90% of the "true" value and all,
with the above exception, were above 80%. Given the large number of
outliers, there is no doubt that the recoveries could have been higher.
Apparently, the filter paper can hold up to 1000 ug of antimony,
depending on how much is present. Method 3050 and the primary digestion
of the SCL method fail to recover this antimony which, at lower absolute
concentrations, is a significant percentage. The hot HC1 wash can
release most of this, so at lower concentrations the HC1 wash is enough
to get accurate results. This can be seen on table VI. As can be seen
in samples E and F, when the concentration of antimony in the soil is
lower, all of the trapped antimony is released in the HC1 wash. As the
concentration increases, the HC1 wash can recover less and less (see
samples C and D). Antimony reacts with nitric acid to form oxides which
are soluble in concentrated HC1 .
Barium and silver are also trapped in the residue and filter paper but
only at higher, not lower concentrations. Silver percipitates as a
chloride when using either method, some of which can be liberated by hot
HC1 . Technique is very important in using this process as evidenced by
the high variability in results. Barium exhibits similar behavior but at
much higher concentrations as its halides are the least soluble of the
group II elements^- .
Precision. As with accuracy, both methods seem to given acceptable
reproducibility for most target elements. One exception to this is the
high levels of barium in samples A and D. In sample A, each laboratory's
triplicates were within 20% of each other; however, the mean values from
each laboratory were wildly different for both methods. In sample D the
variance between laboratories for barium is also extremely high.
1-171
-------
Apparently, at high concentrations, barium is very subject to slight
differences in technique.
At higher concentrations silver is very hard to precisely reproduce. This
is indicated by the large number of laboratories that did not recover any
silver from samples D, E and F. The SCL method has a high %RSD at high
concentrations but all of the laboratories recovered some above the PQL.
The SCL method is more tolerant of slight differences in technique than
is method 3050 for silver.
Likewise, at lower concentrations, antimony does not reproduce well using
method 3050. Several times no antimony was recovered (table III). When
it was recovered, the amounts recovered were highly variable.
Among the other target elements, the main source of variance seems to be
differences in instrumentation rather than differences in methods.
Taking lead as an example, the variability due to instrumentation, as
measured by the %RSD from the "Unknown" solution, was 16%. In round I,
where the sample can not be assumed to be completely homogeneous, the
experimental %RSDs were 26% & 23% for sample A, 21% & 28% for sample B,
and 23% & 27% for sample C by method 3050 and the SCL method
respectively. The variability between methods is only a little higher
than that caused by differences between instruments.
In round II, the samples are very homogeneous. Here the experimental
variability is generally equal to the variability caused by differences
in instrumentation. Again taking lead as the example, the expermimental
variability were 15% & 14% for sample D, 15% & 29% for sample E, and 20%
& 7.1% for sample F. Notably the results from the sets that had %RSDs of
29% and 20% contained outliers. In a clear majority of the cases, when
the %RSD was more than twice the %RSD from the instrumentation, there was
an outlier in the set. In most other cases the %RSD was less than twice
the %RSD of the instrumentation even if outliers were present. Silver
consistently defied this pattern.
From these results it is clear that the SCL method generally yields more
precise data on the target elements of this study.
It is possible to challenge this conclusion because of the large number
of outliers. In round II for example, out of 828 determinations, 68
outliers were reported including less than PQL results for samples with
quantifiable concentrations of target elements. If 11 are eliminated
because they represent elements not quantifiable by method 3050, then
there are 57 outliers which account for 6.7% of the data. These outliers
favor neither method. If it is assumed that the number of outliers
should be about 1%, i.e. the number expected outside three standard
deviations, then we must assume that the number of outliers is
unreasonably high. (A similar number of outliers is obtained by using
control charts with three standard deviation limits where the suspected
outlier is not factored in.)
It can be argued that this may be caused by possible heterogeneity in the
1-172
-------
samples, thus invalidating the study. However, there is no pattern to
suggest this. For example, one laboratory that reported an outlier for
copper, reported chromium close to the true value while an other
laboratory reported the opposite. Rather, it can only be concluded that
an unreasonable number of analytical errors were committed, the majority
of which would seem to be data handling errors as opposed to poor
technique. This is indicated by the number of results that are even
multiples or fractions of the mean and/or "true" value. This view also
supported by the number of simple computational errors committed on the
report sheets which were corrected by the authors.
This conclusion is consistent with the data collected as part of the
Background study to this report. Every participating laboratory has
experienced an explosive growth over last three years. On the average
these laboratories experienced a 210% increase in sample load and a 180%
increase in professional staff. Notably, fewer and fewer positions are
filled by chemists with environmental experience. Only 42% have previous
experience which averaged 3 years. While every laboratory turned in at
least one outlier, the number of outliers was not distributed evenly
among all the laboratories. A few turned in many more than the others.
The high incidence of statistical outliers in this study, we feel, is a
result of the great flux in the environmental field. Since these errors
do not bias the data toward either method, these errors merely increase
the variability of both methods and do not invalidate the data.
Range of Elements. EPA SW 846 method 3050 cannot accurately or precisely
measure antimony; barium, nor silver at common concentrations, while the
SCL method can. Thus the SCL method has a wider range of elements.
Linear Range. For the majority of elements both methods had equal linear
ranges. The linear ranges for antimony, barium, and silver are wider for
the SCL method than for method 3050,
although neither method is as linear with barium and silver as with the
other target elements.
Ease of Use. The method 3050, as written, has at least seven steps,
which are repetitive and have rigid time schedules. The SCL method has
six steps, few of which require timing or careful observation. The SCL
method is hence a simpler and easier method to use. Since the SCL method
has no time limitations and uses aqua regia, which digests more
vigorously, the SCL method the SCL method is a faster method.
SUMMARY AND RECOMMENDATIONS
Judging from the results of this interlaboratory study, it is clear that
the SCL method is superior in all five criteria. It is equally accurate
and precise for the most of the same elements that 3050 claims and is
more accurate and precise for antimony, barium, and silver over a broader
linear range. At the very minimum, we recommend that SW 846 method 3050
should be amended in the method performance section to note that it has
only a limited linear range for antimony, barium, and silver. We
recommend that the SCL method be adopted as an accepted alternative to
1-173
-------
method 3050. Finally, further study is needed on the continuing changes
in the environmental laboratory industry.
ACKNOWLEDGEMENTS
We would like to the thank very much the following laboratories for their
gracious participation in this study: Analytical Technologies, Inc. San
Diego, CA; Brown & Caldwell Laboratories Pasadena, CA; Central Coast
Analytical Services, San Luis Obispo CA; CRL-ENSECO Garden Grove, CA;
Environmental Monitoring and Service, Inc. Camarillo, CA; IT Analytical
Services Cerritos, CA; Montgomery Laboratories Pasadena, CA; S-Cubed San
Diego, CA; West Coast Analytical Service, Inc. Santa Fe Springs, CA.
REFERENCES
1. Test Methods for Evaluating Solid Wastes (EPA SW 846) Method 3050 3rd
Edition. Office of Solid Waste and Emergency Response,
U.S.Environmental Protection Agency:Washington,
D.C., November 1986; Volume 1A
2. Comprehensive Environmental Response Compensation and Liability Act
(CERCLA or "Superfund") sect.101 (14)d Title 40 CFR Part 261
3. Federal Water Pollution Control Act,sect. 307(a)(l) Title 40 CFR
Sub-Chapter D
4. California Administrative Code Title 22. Social Security Division 4
Environmental Health Sect. 66699(b) (Register 85, No.l 12-85) p
1800.77
5. Hinners, T.A., et al.;"Results of an Interlaboratory Study of ICP
Method 6010 combined with digestion Method 3050."; Prodeedings of the
Third Annual USEPA Symposium on Solid Waste Testing and Quality
Assurance, Washington, D.C. July 1987-
6. Kimbrough, D.E. and J.R. Wakakuwa, "Acid Digestion for Sediments,
Sludges, Soils, and Solid Wastes. A Proposed Alternative to EPA SW
8461 method 3050."; Environmental Science and Technology In Press.
7. Test Methods for Evaluating Solid Wastes (EPA SW 846) Method 6010 &
7000 ,3rd Edition. Office of Solid Waste and Emergency Response, U.S.
Environmental Protection Agency: Washington, D.C., November 1986;
Volume 1A
8. Eckschlager, K. Errors, Measurement, and Results, in Chemical
Analysis Van Nostrand Reinhold Co. LTD.: London, 1961 pg. 119
9. Experimental Statistics United States Department of Commerce; National
Bureau of Standards: Washington D.C., 1963 pg T-9
10.Eckschlager, K. Errors, Measurement, and Results, in Chemical
Analysis Van Nostrand Reinhold Co. LTD.: London, 1961 pg. Ill
1-174
-------
11.Bauer, E. L. A Statistical Manual for Chemists, 2nd Ed.; Academic
Press: New York and London, 1971; pg. 22
12.Condensed Chemical Dictionary, 10th Ed.; ed. G.G. Hawley, Van
Nostrand Reinhold Company Inc.:New York , 1981; pg. 79-81.
13.Cotton, F.A. & Wilkenson, G. , Advanced Inorganic Chemistry, 3rd Ed.;
Interscience Publishers:
New York, 1972; pg. 1048
14.ibid. pg. 216
1-175
-------
Table I
Analytical Lines
Analyte
As
Ag
Ba
Be
Cd
Co
Cr
Cu
Mo
Wavelength
189.04
193.76
328.07
233.53
455.40
493.41
543.6
234.8
313.04
214.44
226.50
228.80
228.62
240.7
205.55
267.72
357.87
224.70
324.75
202.03
313.2
# of labs
2
5
8
3
2
1
2
4
4
2
2
4
6
2
2
4
2
1
7
6
2
Analyte
Ni
Pb
Sb
Se
Tl
V
Zn
Wavelength
231.60
323.00
216.99
220.35
280.8
283.31
206.83
217.58
196.03
190.86
276.79
351.92
377.57
292.40
318.4
213.86
# of labs
5
3
1
5
1
1
5
3
7
1
4
2
1
6
2
8
All wavelengths are in nanometers
1-176
-------
Table II
Results from Round I
N = 27 (except for As where N = 24)
ANALYTE
Ag
As
Ba
Cd
Co
Cr
Cu
Mo
Ni
Pb
Sb
Se
Tl
V
Zn
CAMPI F
I.D.
B
C
A
B
C
A
B
A
B
C
A
B
C
C
A
B
C
A
B
C
C
C
C
B
C
A
B
C
UNITS
MEAN VALUES
3050
130 2
920
322
96
120
140
400
2200
5800
320
15000
1700
400
840
290
160
170
2700
3000
120
320
680
800
1200
1400
3500
2400
1600
DOHS
160
900
1600
100
130
130
380
2200
5500
370
15000
1600
420
860
290
160
200
2700
3000
130
830
700
770
1100
1500
3500
2000
1500
ug/g
t-TEST !
2.40
0.50
4.17
0.69
0.83
0.42
1.24
0.14
1.11
1.89
0.36
1.40
0.72
0.19
0.30
0.31
2.21
0.32
0.34
0.36
13.1
0.31
1.49
1.12
1.56
0.25
0.57
1.05
%RSD VALUES
3050
40
21
112
25
33
25
23
23
22
31
37
33
26
49
21
21
26
26
21
23
61
23
14
32
27
25
27
35
DOHS
24
21
98
28
36
25
21
25
26
27
28
27
24
48
26
24
30
23
28
27
13
34
8.7
34
28
29
27
28
UNKNOWN
6.4
16
10
10
10
9.5
8.9
18
18
18
11
11
11
8.3
9.3
9.3
9.3
16
16
16
13
12
7.1
14
14
PERCENT
F-TEST2
0.58
0.97
NA3
1.49
1.37
0.95
1.52
0.83
0.93
0.98
0.78
1.94
0.92
1.00
1.26
1.21
1.87
0.89
1.83
1.52
NA3
2.31
2.60
0.99
1.31
1.15
1.08
1.78
The critical value for t is 2.98 and the critical value for F is 2.58 for alpha = 0.01
2
Silver by method 3050 has only 24 data, as 3 data are less than PQL.
The F-test is not applicable (NA) to barium or antimony because of the differences in
mean values.
The values for t and F are dimensionless.
1-177
-------
TABLE III ROUND II Low Values
N = 9 (except for As and Se where N = 8)
AMAI YTF
/-\li/-\L_ TIC
Ag
As
Ba
Be
Cd
Co
Cr
Cu
Mo
Ni
Pb
Sb
Se
Tl
V
Zn
UNITS
!"TD| IP"
1 rlUC.
VALUE
109
124
197
< 10
< 50
154
128
252
166
204
263
142
155
77
168
148
CAMDI P
OMlVIr LC
I.D.
D
D
F
F
F
D
E
F
E
F
F
E
D
D
E
E
MEAN VALUE
3050
109
127
190
136
117
241
153
195
260
94
139
96
144
153
SCL
128
131
206
138
143
267
160
215
254
157
161
124
162
172
ug/g
t-TEST 1
0.54
0.20
0.98
0.12
1.31
0.56
2.06
1.50
0.30
3.09
0.94
0.56
0.90
1.00
%RSD VALUES
3050
32
32
18
26
21
22
48
13
20
48
27
83
22
16
SCL
28
38
15
20
35
45
39
12
7.1
18
29
83
28
26
UNKNOWN
6.4
16
10
8.9
18
11
8.3
9.3
16
13
12
7.1
14
OUTLIERS
3050
0
1
0
1
0
0
1
1
1
0
0
1
0
0
SCL
0
0
0
1
1
1
1
0
0
0
0
1
0
1
2
TOSO
*J\J \J\J
-------
SOLUBILITY DATA
Table VI
ANTIMONY
BARIUM
SILVER
3050
SCL prim
SCL scdn
SCL total
3050
SCL prim
SCL scdn
SCL total
3050
SCL prim
SCL scdn
SCL total
NO. OF DATA
SAMPLE A
-------
A PERFORMANCE EVALUATION OF THE INORGANIC METHODS USED IN
THE CONTRACT LABORATORY PROGRAM
KEITH ALECKSON AND Y. JOYCE LEE, LOCKHEED ENGINEERING AND SCIENCES CO.,
1050 E. FLAMINGO ROAD, LAS VEGAS, NV 89119; E.J. KANTOR, U.S.E.P.A.,
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY, P.O. BOX 93478, LAS VEGAS, NV
89193-3478.
ABSTRACT. Periodically, Inorganic Method Validation Reports are
generated for the Contract Laboratory Program (CLP) by the
Environmental Monitoring Systems Laboratory in Las Vegas (EMSL-
LV). The reports are an evaluation of each method's performance
over the previous year. The information used to generate the
reports is taken from the results and Quality Control (QC) data
reported from the laboratories on diskettes, from the Quarterly
Blind (QB) performance evaluation standard results, and from the
operating parameters and conditions reported by the laboratories.
Once the information is studied, a report is completed detailing
the precision, accuracy, limits of detection, and linear range of
the methods for soil and water. Other QC results and data
trends, particular to each method, may also be included in the
reports. The reports are of value to the CLP because they
provide a measure of method performance that is useful for
determining if the needs of the program are being met. In the
future, the reports also may be helpful in the identification of
possible problems and the optimization of the methods.
The presentation will include accuracy and precision results for
the inorganic CLP methods from spike, duplicate, laboratory
control sample (LCS), and other EPA performance standard results
taken from the Method Validation Reports for routine data
generated in 1988. Examples of how the method performance can be
compared with operating conditions to optimize method performance
may also be presented.
1-180
-------
STUDIES OF INTELLIGENT AUTOMATION FOR WATER ANALYSIS BY
ICP-AES WITH CLP FRDTOCOL
SUE F. ZHU, ALAN K. MERRICK, SPECTRO ANALYTICAL INSTRUMENTS, INC. 160 AUTHORITY
DR., FITCHBURG, MA 01420, FRANK A. GLODAS, THE WATER WORKS LABORATORIES,
LEOMINSTER, MA
ABSTRACT. Inductively Coupled Plasma - Atomic Emission Spectrometry (ICP-
AES) has been approved as an EPA method for water analysis. ICP offers
excellent detection limits, wide linear working range, and outstanding multi-
element capability. The EPA Contract Laboratory Program (CLP) has recently
become the most widely used protocol for water analysis because of its
stringent quality control procedures, even by laboratories which do not
specifically have EPA contracts. Because environmental laboratories typically
run high numbers of water samples, with 20 or so elements per sample, the ICP
application to the CLP program is ideal.
ICP spectrometers have often been used with automatic sampling devices.
These have typically been the carousel type, with which a sequence of samples
are analyzed, start to finish, with no deviation possible. This presents a conflict
when trying to use this type of automatic sampler with CLP. As required by
CLP, a sequence of check standards must be analyzed periodically with the
decision to run unknowns depending on the results of the check standards. Out
of tolerance results require recalibration, re-analysis of check standards, then
analysis of unknowns. The decision-making requirement during a sample run
thus precluded the use of an auto-sampler, negating a large part of the
advantage of !CP.
In this study, a system was used which employs artificial intelligence to follow
the CLP format with full automation. It has long been recognized that the
simultaneous ICP offers the most efficient means for high sample throughput,
being capable of triplicate analysis and flush-out within two minutes. The more
widely used sequential instruments typically take 5 -10 times longer. The
instrument evaluated in this study uses simultaneous and sequential
spectrometer in combination. The sequential module is employed only as a
contingency basis, such as why unusual elements are requested. The vast
majority of the determinators are done on the simultaneous module.
A program was developed which enabled the computer to control simultaneous
and sequential spectrometers at the same time, as well as, maintaining direct
control of the auto-sampler. Communication between spectrometer and auto-
sampler via fiber-optic cable is also evaluated. The entire program was
designed to support the CLP protocol while operating unattended. This
includes analysis of check standards, and based on these results, the decision
to recalibrate or to proceed with analysis. After analysis of a prescribed number
of samples, the procedure is repeated. When all samples have been analyzed,
the computer will shut down the instrument automatically. Analytical results will
be stored in the computer.
1-181
-------
LABORATORY INFORMATION
MANAGEMENT
-------
CUSTOMIZED LIMS DATA TREATMENT THROUGH
INTERACTION WITH A USER-DEFINABLE SPREADSHEET
RICHARD D. BEATY AND PAUL C. DIFFERDING, TELECATION ASSOCIATES, P.O. BOX 1118,
CONIFER, COLORADO 80433
(303) 838-2088
ABSTRACT
A PC-based LIMS requires a data base management system to handle and track the vast amounts
of information generated in today's analytical laboratory. Yet a data base management system is
not the most convenient tool for handling analytical calculations, which may require the use of
parameters from different records in the data base. For calculations, the spreadsheet has become
an indispensable tool, allowing non-programmers to define a desired manipulation of numerical
data, by simply specifying algebraic formulas to define the needed calculations. This paper will
present an approach to providing straightforward customization of a PC-based LIMS, through
integration of the LIMS data base to a user-definable spreadsheet. Example applications will include
spreadsheets for raw data entry and post analysis data processing for production of QC charts and
management graphs.
INTRODUCTION
A computerized Laboratory Information Management System (LIMS) makes up the heart of
organizational management, for many laboratories. The standard functions contained in a full
functioning LIMS system include: sample log-in and tracking; generation of worksheets, management
reports, results reports, and, if appropriate, invoices; instrument interfacing and automatic data
collection; automated quality control; and archiving of sample results, chain-of-custody, and other
information.
LIMS systems have traditionally been built around mainframe or large mini computers. These large
scale LIMS offerthe ability to handle and organize very large informational data bases, but they are
frequently difficult to set up, requiring extensive work for customized reporting and data treatment,
through expensive and time-consuming custom programming by the manufacturer or by a
laboratory's in-house programming staff.
While the raw computer power of a large LIMS system is required for some operations, recent trends
have seen the emergence of micro computer based LIMS systems. These systems are usually built
around the powerful and fast 80386 hardware, as a file serverfor a network of PC's, distributed about
the laboratory. In addition to a significant initial purchase price advantage, a PC-based LIMS system
can be fully operational in a matter of days, and the PC-based system may offer substantially
enhanced flexibility for easy customization of report formats and data treatment.
Off-the-shelf software for PC's offer a wide variety of sophisticated applications, which are far more
user oriented, than is much of the software for mini and mainframe computers. Some of this software
is well suited for use in the laboratory. Data base management and spreadsheet programs are two
such applications.
1-183
-------
Relational data bases are ideal for storing information and allowing the user to sort and query the
data to retrieve any desired information. This makes a relational data base management system the
obvious choice forthe heart of a PC-based LIMS system. However, many of the functions, which are
required in a full-functioning UMS, are not easily implemented with a data base manager. Complex,
multiple-parameter calculations are just one of these areas. On the other hand, a spreadsheet
program is ideal for handling the areas where the data base management system is weak.
Therefore, the ideal LIMS would consist of a system of interacting data base management and
spreadsheet modules, with each module handling those functions for which it was best suited.
"Integrated" software packages, which provide interacting data management and spreadsheet
modules, are available for the personal computer. Many of these systems are programmable, to
allow creation of dedicated, turn-key applications. TELECATION ASSOCIATES has chosen such an
integrated package for development of its "SMARTLAB" (R) Laboratory Information Management
System. The package chosen was "SmartWare" (R) from Informix Software, Inc. In addition to the
data manager and spreadsheet, SmartWare also provides a word processor, communications
program, and other tools, which will also find use in the laboratory. The application of the data
manager and spreadsheet to the SMARTLAB LIMS system is discussed below.
THE BENEFITS OF A RELATIONAL DATA BASE
The main benefit of a relational data base management system is its ability to maintain large amounts
of information, which can be easily retrieved and tracked by the user. In a LIMS application, the
information to be maintained in the data base file includes, among other things, sample and test
identification fields, with associated analytical information. To provide maximum flexibility in the
tracking of analytical information, each test to be run and tracked occupies its own data base record,
which can be retrieved, as needed, from all the other records in the data base. The test specific
information maintained in the SMARTLAB "Test" data base record is shown in Figure 1.
Test ID:A1 Narr.e: AluEinun
Lab t .... GBE103E
Sar.ple ID: K>;X-2986
Client: Jackson Machinery
Sect :jnetolE
DUE Days: 5 Date:10/30/68
HOLD TIKE Days: 6 Date:}0/25/B8
Method ID: ICP
Spec ID:
Limits: Lov(B)
Instrument ID: ICP-2
High(H)
IDL (11)
0.8000
Result: 5B.2000 ED 1.70000 CV 2.92 (Alt Result: )
weight 1.0000 volume 1.0000 dil 1 Isolids 100.0 decimals (4 max)
1.0000 1.0000 1 100.0000 58.2000
QUALITY
CDI.'TfOL
REPORTED RESULT: SB.2
[ ] ng/L units |
.. Duplicate: 61.1000 (Diff: 4.9
SpiXe: 111.6000 Std: 50.0000 ... i '-eccvery : 1 06 '. B
Contrcl: !, ..ccuracy :
ISOLATE ... Run:10000005 Trend/User:lubeoil QC Reference:
Cor.r.ent: Used lube oil analysis
CHAIN-OF-CUSTODV
H
Container 1 Container Type:
Date Time By Method
PRESERVE D
PREPARED
ANALYZED
APPROVED
00/00/00
00/00/00
10/25/68
10/25/BE
06:OOP
06:04P
CHC
RDB
dil in kerosene
^
Figure 1. SMARTLAB Test' data base record.
1-184
-------
In addition to all the test specific information, there are a number of sample specific parameters,
shown in Figure 2, which must be maintained in a LJMS data base. Since this sample information is
the same for all tests being run on the sample, it is undesirable to record all the sample specific
information in every test record. Therefore, a separate "Samples" data base file is used to store
sample specific information. Relational data base capabilities allow the "Samples" data base to be
linked to the 'Tests" data base, to provide complete information on both sample and test specific
information.
Lab | G881035 LOGIN Time: 04:11P Date: 10/25/68
DUE Days: 5 Date: 10/30/88
STAT: 5/ 5
Complete: *
Sample ID: MXX-2988 Reported:10/2B/88 Invoiced:11/01/88
Account^ Jackson Client Jackson Machinery
Address 399 S. Garrison
City Commerce City State CO Zip 81901
Contact Jim Peters Phone (303) 555-3931
COLLECTION Date: 10/17/88 Time: 8:1SA By: DNP
Location:
PRESERVATION Method:
STORAGE Location:
Keep for: 10 days Discard on: 11/04/88
DISPOSAL Method: return Date: 11/21/88 By: PLN
Comment:
•— — — — — —. — — __ __._ «_ »___ __„ — •_• __.___________•»•- ——•——— —___•__ _._._»_______«_«« —— w —— •-
Purchase Order No. JM-157Q Discount: %
BILL TO: Client Jackson Machinery
Address 300 S. Garrison
City Commerce City State CO Zip 81901
Contact Claude Dillon Phone (303) 555-7339
Figure. 2. SMARTLAB "Samples" data base record.
In order to retrieve desired information from the entire data base of all samples and tests in the
laboratory, it is necessary to isolate selected records (query) or rearrange the order of records (sort)
in the data base. This kind of manipulation of the data base is easy with a flexible relational data base
system.
Forthose predictable manipulations of the data base, subsets, or "indices" of data base records can
be maintained, and these indices may be accessed instantly upon demand, without resorting or
requerying. SMARTLAB maintains predefined sort indices for "lab #", "test ID", "client name" and
more, to allow immediate user access to sorted data base records. Also, predefined isolated
subsets of data base records (e.g., "preparation backlog", "report backlog", "invoice backlog")
allow instant recall of those sample test records meeting selected stages of completion. These
techniques substantially enhance the speed at which selected information can be accessed.
1-185
-------
Another benefit of relational data base systems, is the ability to design custom report formats. Every
laboratory has its own idea of what kind of information should be shown on a report, and where that
information should be placed. Further, special purpose reports which may be very important to one
laboratory, may not even apply in another. Report definition utilities allow generation of report
formats to meet individual needs.
Screen oriented report definition utilities are easy to use, in that the report is laid out on the screen
exactly as it is to appear on paper. Once created, user-defined report formats may be accessed, as
desired, to generate data base reports for any purpose. Predefined SMARTLAB reports include
formal analysis reports, invoices, and a variety of management reports. In addition, the report
definition utility may be used to edit existing reports, or create completely new formats.
THE BENEFITS OF A SPREADSHEET
While the data base manager serves as the heart of the LIMS by maintaining, tracking, retrieving, and
reporting laboratory information, other functions, which are equally important to the overall LIMS
application, are more easily accomplished with a spreadsheet. One of these functions is the
performance of mathematical calculations. While calculated data base fields can be defined, which
automatically show the results of a calculation based on other information in the same data base
record, it is very difficult to incorporate parameters from a different data base record into the
calculation. Inter-record calculations are, on the other hand, very easy with a spreadsheet. An
example of an inter-record calculation, which is easy with a spreadsheet, but difficult to perform from
a data base, is the calculation of the percent difference between duplicate sample runs, which
involves comparing results from two different data base records.
Most spreadsheet programs also provide graphing utilities, which allow data to be presented in a
variety of graphical formats. There are many laboratory applications for this capability. Some of
these will be illustrated later.
When manually entering large amounts of information, it is sometimes easier to enterthat information
into a spreadsheet, rather than a data base. In the two-dimensional row/column spreadsheet layout,
the user can randomly access any cell for data entry purposes, while the field-by-field, record-by-
record forced sequence for entry into a data base, is sometimes tedious.
DATA BASE - SPREADSHEET INTERACTION
Figure 3 lists the basic functions of a LIMS system and identifies the optimum software tool (data
base manager or spreadsheet) which is best suited for implementing each function. Those
functions which primarily involve retrieval of information (sample login, laboratory status review,
reporting, invoicing, and archiving) are clearly data base applications. Calculation intensive
functions (data entry and quality control) are best suited for the spreadsheet. Functions which
require both record retrieval and calculation intensive functions, may best be served by a combination
of data manager and spreadsheet approaches.
1-186
-------
DATA
BASE
Sample Login X
Assigning work X ..
Data entry
Data approval
Lab status
Reporting
Invoicing
Archiving & search
Extended qc, calc
and management
X
X
X
X
X
SPREAD
SHEET
.. X
X
X
X
Areal
DATA
ENTRY
APPLICATIONS
Area 2
POST ANALYSIS
APPLICATIONS
Figure 3. Data base and spreadsheet functions of a LIMS.
The spreadsheet applications fall into two main areas. Area 1 involves the entry of analytical data into
the LIMS system. Since the desired data may be the result of an analytical calculation, the
spreadsheet applies. Area 2 involves the post analysis processing of data for quality control and
other calculation intensive applications.
Even though a spreadsheet may be used for selected LIMS functions, the data base will always be
the permanent "home" of all laboratory information. Therefore, data will have to be moved from the
data base to the spreadsheet, whenever any of the spreadsheet functions are to be performed.
Additionally, for Area 1 (data entry) applications, data will have to be returned from the spreadsheet
to the corresponding data base record, after the spreadsheet function is complete. The required
flow of information between data base and spreadsheet, for data entry applications, is illustrated in
Figure 4.
DATA BASE
SPREADSHEET
(1) ASSIGN TASKS
... selected records sent
to spreadsheet
(3) APPROVE RESULTS
... approved results returned
to data base
Figure 4. Data entry applications of a spreadsheet.
1-187
-------
When samples are logged into the LIMS system, the information is added to a "working samples"
data base consisting of all test records which are awaiting entry of approved test results. During the
"ASSIGN TASKS" function, the records in this data base requiring analysis by a common analytical
procedure may be grouped together and the information contained in these records transferred to
a predefined spreadsheet, called a "Run Log." The Run Log represents an analytical worksheet.
The Run Log can be printed out to provide a hard copy listing of samples requiring a particular
analysis and the corresponding spreadsheet file provides an electronic worksheet, into which the
results of the pending analysis will be entered. Any number of Run Log spreadsheet files, containing
different groups of analyses to be performed, may be created in the ASSIGN TASKS mode, and
saved for subsequent data entry.
When an analytical run of the selected samples has been performed, the "ENTER RESULTS" mode
allows the analyst or data entry clerk to access the Run Log corresponding to the completed run,
and enter the results into the spreadsheet. Alternately, instrument interface options allow automatic
data entry into the Run Log.
Following data entry, security approved personnel may re-access the Run Log in the "APPROVE
RESULTS" mode, for the purpose of evaluating the accuracy of the run. During this mode of
operation, approved results are transferred from the spreadsheet to the LIMS data base records
corresponding to the approved spreadsheet entries.
THE "RUN LOG" WORKSHEET
The spreadsheet is used as a vehicle for analytical data entry for logical, as well as technical reasons.
Creating and saving spreadsheet files, which correspond to an analytical worksheet, maintains,
intact, a complete analytical run. Data is actually acquired in sample/test groupings corresponding
to a single run, and the Run Log provides a data entry worksheet corresponding to that run. The two
dimensional row/column access to the worksheet further simplifies data entry. Additionally, the
quality of analytical data will likely depend on conditions which existed during the run in which the
data was acquired. Therefore, it is proper and in concert with accepted quality control procedures,
to review QC data and approve or disapprove results, on a run by run basis.
While the above benefits of spreadsheet data entry are significant, the most powerful benefit is
derived from the automatic calculation capabilities of the spreadsheet. The Run Log worksheet,
illustrated in Figure 5, is a system defined spreadsheet. The Run Log is a generic spreadsheet,
containing columns of information which apply to almost any kind of chemical analysis which might
be performed. In addition to columns identifying the samples and tests to be run, additional columns
are provided for sample preparation parameters, including sample weight, volume, dilution factor,
and percent solids (for dry weight reporting basis). The "calculated result" column is determined
from the entered result and sample preparation variables, by the following generally applicable
equation:
[Calc Result] = [Init Result]* [volume] *[dil factor]
[weight]*0.01*[%solids]
The sample preparation parameter columns always default to a value of 1.0 (100 for % solids),
making them mathematically innocuous, when they are not used.
1-188
-------
ANALYSIS Run: 10000006 Method: ICP Date: 10/25/88
RUN LOG By: CMC Instr.ID: ICP-2 Time: 06:16P
Lab# 0 TestID Init Result Weight Volume Oil XSolid Spk/Ref Calc Result C Rec/Diff
G881036
G881036 D
G881036 S
G881036
G881036
control C
control C
control C
G881037
G881037
G881037
G881038
G881038
G881038
Mo
Mo
Mo
Cr
Pb
Mo
Cr
Pb
Mo
Cr
Pb
Mo
Cr
Pb
611.0000
624.0000
705.0000
2573.0000
45.0000
767.0000
2767.0000
38.0000
490.0000
3020.0000
0.0530
612.0000
2485.0000
29.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000 100.0000
.0000 100.0000
.0000 100.0000 100.0000
.0000 100.0000
.0000 100.0000
.0000 100.0000 750.0000
.0000 100.0000 2800.0000
.0000 100.0000 35.0000
.0000 100.0000
.0000 100.0000
.0000 100.0000
.0000 100.0000
.0000 100.0000
.0000 100.0000
611.0000
624.0000
705.0000
2573.0000
45.0000
767.0000
2767.0000
38.0000
490.0000 B
3020.0000 H
0.1000 U
612.0000
2485.0000
29.0000
2.1
94.0
102.3
98.8
108.6
102.0
99.4
116.0
Figure 5. Standard "Run Log" spreadsheet.
Other predefined columns provide valuable information on the analytical quality of the run. The "C"
or concentration flag column automatically displays a flag indicating how the result compares to
lower and upper limits which have been predefined for the analysis, and a special detection limit flag
indicates when the result is less than the detection limit for the method. In such a case, the calculated
result is automatically made equivalent to the detection limit, and the detection limit flag is set. The
"Recovery/Difference" column automatically calculates the percent difference between duplicate
runs, the percent recovery for spikes, and the percent accuracy for control samples.
In addition to the above information which is available to all who are authorized to access the Run
Log, a "blind QC" feature will automatically display the percent accuracy on special blind control
samples only to privileged personnel who have been authorized for approving results. Thus, all QC
information is available to the authorized reviewer at the time results are being approved, providing
data evaluation criteria at the time it is needed. Based on the information in the Run Log, the reviewer
may approve results, transferring them from the Run Log spreadsheet back to the LIMS data base
for reporting and archiving, or selected sample analyses may be rejected, requiring them to be
reassigned to a subsequent Run Log for re-analysis.
CUSTOM DATA ENTRY WORKSHEETS
While the Run Log spreadsheet is a useful general purpose data entry worksheet, the user
programmability of spreadsheets makesfeasible an even more powerful applicationforspreadsheet
data entry. The Run Log expects the initial result to be entered in concentration units, to be
compatible with the equation for determining the calculated result. However, many times the
analysis does not provide results in concentration units. Raw results may, for instance, be presented
as "strip chart recorder divisions", "millilitersoftitrant", "absorbance", etc. Results in concentration
units must then be calculated from this raw data.
1-189
-------
In such cases, it would be useful to allow the use of custom spreadsheets, which are definable by
the user, for calculating Run Log "initial results" from the raw data generated by different analytical
methods. Calculating the initial result from raw data not only expedites the routine calculations
which are a part of every laboratory, but also minimizes data entry and human calculation errors prior
to entry of results into LJMS. SMARTLAB allows access to user-defined spreadsheets for data entry,
and automatically reads the computed result from the custom spreadsheet into the "initial result"
cell of the Run Log, whereupon all standard Run Log functions described above are implemented.
An example of a user-defined spreadsheet is shown in Figure 6. This spreadsheet was created to
allow entry of raw millivolt readings from an ion selective electrode measurement of chloride. In this
spreadsheet, the user need only enter the readings for the standards and the samples in the "Enter
Meter Reading" columns, and the concentration is automatically computed for the samples by
comparison to a linear least squares calibration, which was automatically determined from the
standard readings. A single key stroke will transfer the computed sample concentrations into the
"Result" column of the Run Log spreadsheet.
LINEAR
REGRESSION
WORKSHEET
Lab #
Q TestID
===== SAMPLES =====
Enter
CONC. HETER
(mg/L) READING
===== STANDARDS ====
Enter
CONC. METER
(mg/L) READING
1
2
3
4
G881041
G881041
G861041
G881042
D
S
ct
Cl
Cl
Cl
45. 64
44.65
77.12
61.26
3.67
3.59
6.21
4.93
0.00
25.00
50.00
100.00
slope:
int'pt:
r-value:
0.00
2.11
3.85
8.11
12.39
0.15
0.999
Figure 6. Custom spreadsheet for linear calibration.
POST ANALYSIS DATA PROCESSING AND QUALITY CONTROL USING A SPREADSHEET
SMARTLAB's "EXTENDED QC" mode provides access to the spreadsheet for post analysis data
processing, the applications identified as "Area 2" in Figure 3. In the EXTENDED QC mode, the LIMS
data base manager is used to sort and query the data base to select the records intended for
processing. Then the information from the selected records is transferred to the spreadsheet
program, as illustrated in Figure 7. Since for this application, the spreadsheet is used simply as a
processor of information from an already completed data base record, the interaction of the
spreadsheet with the LIMS data base is less involved, than for applications where spreadsheet
results must be returned to the data base. This means that total free use of all spreadsheet functions
is possible.
1-190
-------
DATA BASE
EXTENDED QC
... selected
records sent
to spreadsheet
SPREADSHEET
Figure 7. Post analysis applications of a spreadsheet.
For post analysis data processing, the user again has preprogrammed spreadsheet options, as well
as user-definable options. System defined options include trend analysis and three types of QC
charting. The spreadsheet graphics capabilities are particularly useful for these applications.
Examples of trend and QC charts, generated automatically through the EXTENDED QC mode of
SMARTLAB, are shown in Figures 8 and 9.
400
360-
320
01
-4—*
13
CO
-------
Shewhart QC Chart
5
in
01
o:
IDU-]
140-
oQ -
120-
110-
DO -
90-
80-
70-
60-
50-
/\
V7
O '
Q-^
s\
V1
0
^^—
0
— _Q=^_
0
- X
0
y\
/ \
0
13
0
0
\/
t?
0
0
A
,'S,
0
3333333333
//////////
000011111 1
6789034567
i i i { i I i i i i
999999999 9
Dote
Figure 9. QC chart from SMARTLAB spreadsheet.
CUSTOM SPREADSHEETS FOR POST ANALYSIS APPLICATIONS
As previously identified, calculations involving information contained in multiple data base records
are difficult to implement from within the data base, but easy from a spreadsheet. Therefore, any
type of user application involving processing of information from multiple tests or LIMS data base
records, is a candidate for a user-defined spreadsheet.
One such application is a material balance report for a complete assay of a metal alloy sample. In
this application, all components determined in the sample are summed, and the summation result
is compared with 100 percent, as a measure of analysis accuracy.
In this application, the EXTENDED QC mode is applied to isolate, from the LIMS data base, only
those records which correspond to a test performed on the selected sample. These results are then
sent to the spreadsheet to perform the necessary calculations. The user-defined spreadsheet for
this application is a simple one. The user need only define a calculated cell with the formula to sum
the resutt column from all the tests. An example of a spreadsheet for this application is shown ir
Figure 10.
1-192
-------
ALLOY MATERIAL BALANCE
lab# testid result
G881030 Al 9.8 %
G881030 Cr 15.4 %
G881030 Cu 18.5 %
G881030 Fe 42.2 %
G881030 Zn 11.6 %
G881030 B 2.8 %
TOTAL 100.3 %
Figure 10. Custom spreadsheet for material balance calculation.
Simple custom spreadsheets may be defined on the spot, from within the LIMS EXTENDED QC
mode, or more complex spreadsheet definitions may be preprogrammed with the SmartWare
spreadsheet module, for later routine access. The latter approach makes it possible for data
processing clerks, without spreadsheet experience, to generate complex custom spreadsheet
reports and charts, without leaving the LIMS system.
In addition to applications dealing with analytical data, other information can be extracted from the
LIMS data base and transferred to custom spreadsheets to provide valuable management tools. For
instance, the graph in Figure 11 provides an easy comparison of analyst productivity. And Figure
12 indicates that the analytical load on the laboratory's gas chromatographs is nearing capacity,
and the purchase of an additional instrument should be considered. Similar spreadsheets could be
devised to forecast sample load into the future.
SUMMARY
Integration of a spreadsheet into a LIMS system provides a dramatic enhancement to the usability
of information in the LIMS data base. The spreadsheet's powers of calculation and graphics display
increase the capabilities of the LIMS system well beyond the abilities of a stand-alone relational data
base manager. Additionally, the incorporation of user-definable spreadsheets into the LIMS system,
provides an easy mechanism for user customization of LIMS functions by non-programmers. The
ability to access both predefined and user-defined spreadsheet applications directly from the LIMS
system, without special communication or data transfer programs, makes it possible for these
functions to be utilized by the routine user.
1-193
-------
Telecotion Associates
SMARTLAB CUSTOM LIMS SPREADSHEET
Analyst Productivity Chart
MLG
Figure 11. Management report from SMARTLAB spreadsheet.
240
210
180
-------
EARLY WARNING REPORT: AUTOMATED CHECKING OF QC DATA
Robert Peak, Quality Assurance Director, Paul Duerksen, QA Coordinator,
Kit Wong, Computer Programmer, Ed Szeto, Computer Programmer, Brown and
Caldwell Analytical Laboratories, 373 S. Fair Oaks Avenue, Pasadena,
California 91105
ABSTRACT
Sample log-in, work list generation, data entry, and report production
are data-handling functions common to most Laboratory Information
Management Systems (LIMS's). A newly developed automated review system
now extends those functions to include the use of a LIMS for reviewing
and evaluating quality control (QC) results. This paper delineates the
four major stages involved in creating this automated review system:
concept, preliminary design, programming, and testing. Although
computer programming is typically system-dependent, the logic employed
in QC review should have broad application to many LIMS's. The paper
also describes the daily operation of this new LIMS function on-line in
three networked commercial laboratories with over 140 employees. This
account of the review system's pitfalls, successes, and maintenance
requirements can help other LIMS users achieve a smooth, reliable
installation.
INTRODUCTION
Environmental laboratories are finding increasing incentives to
automate data handling. A Laboratory Information Management System
(LIMS) can process large quantities of data with far less labor than a
corresponding manual system. Many LIMS's however, have a limited
capability for data evaluation. They serve as storehouses and
reporters of results, but do very little to examine those results.
Brown and Caldwell Analytical Laboratories (BCAL) uses a custom-written
LIMS that contains the typical functions of such systems: means for
logging sample data into the computer, printing out work lists for
analysts, entering analytical results, and producing final reports.
With this system as a base, the Quality Assurance Department set out to
add the capability for review and assessment of quality control (QC)
results. The development process required close cooperation among QA
chemists, in-house programmers, and programming consultants. The final
product is our "Early Warning Report" (EWR), a daily report from the
data base that evaluates and reports on QC results from throughout the
laboratory with regard to archived control limits.
The development of the EWR encompassed four stages: (1) choosing the
design concepts; (2) translating those concepts into programming
algorithms; (3) actual programming; and (4) testing. Once developed,
this type of report can take on part of the data evaluation function
required in any analytical laboratory.
1-195
-------
BACKGROUND
The BCAL LIMS is commercially available from Inquiry Computer Systems
of Laguna Beach, California. As ours was the first installation of
this system, our scientists participated in every stage of development.
The computer consultant who did the primary programming, Mr. Ben
Edmonson, also supplied the hardware and operating system.
The PICK operating system (PICK) works on ADDS/Mentor minicomputers.
ADDS is a division of NCR. All three of our laboratories have a
minicomputer, each with 32 to 64 ports. The computers are linked
through leased-line communications equipment via 16-port modems. PICK
is unusually well-suited to LIMS requirements because, unlike most
other operating systems, it is itself a data base manager. No "data
base" software is run on top of the operating system; it is inherently
designed to handle large relational pools of data. Consequently, a
LIMS that runs on PICK is hardware conservative. With a minicomputer
containing only four megabytes of RAM and 500 megabytes of hard disk,
we can keep full records on line and access them for more than a year.
In a typical month, our larger laboratories analyze about 2500 samples,
with all results archived in the LIMS.
PICK does have some drawbacks. Its programming language, called PICK
Basic, is quite similar to other implementations of Basic. Without
access to sophisticated engineering languages like Pascal or Fortran,
PICK'S number-handling capability is limited. Because it is an
uncommon operating system, finding an experienced programmer who is
also familiar with laboratory operations can be quite challenging.
Functions in the BCAL LIMS are typical of such systems. As samples
arrive, they are assigned ID numbers that include a simple sequence; as
samples are logged in to the computer, the LIMS offers the next
sequential number. Log-in creates a record in the "Order" file, much
like the order entry file of a commercial data base. We use specific
client codes to track and cross- reference client information and
determination codes to identify the analyses required and
cross-reference them to a detailed determination file. When log-in is
complete, the computer prints out a description of the job to be done.
A copy of the printout, called the "traveller," is kept in a suspense
file in the laboratory along with the original chain of custody. A
copy marked "acknowledgement" goes to the client for review.
Following log-in, the analytical determinations for each sample are
dispatched within LIMS to the "Work" file. Every night the computer
prints out data from the Work file onto individual worksheets for the
analysts. The worksheets show the client's sample description, the
laboratory ID number, and the samples sorted by determination and due
date.
The LIMS also generates QC work items directly onto the worklists. For
all tests, every tenth work item is a laboratory control standard
(LCS). Because a true value must be declared for data evaluation, the
LCS has two parts: "LC" for the result and "LT" for the true value.
1-196
-------
For inorganic analyses, the computer invokes a pair of duplicates and a
spike for every tenth determination on a particular sample matrix
type. These QC items are designated by suffixes on the sample ID: Rl
is the first replicate, R2 the second, SI the spike result, and T the
true value (or expected result) of the spike. For organics analyses,
the computer alters the pattern slightly, creating Rl and Si plus a
duplicate spike, S2, and T. Figure 1 illustrates a typical worksheet
for a multi-component analysis.
Analysts enter their test results on the worksheets, which then go to a
swing-shift data entry operator for keyboarding. Overnight, the
computer prints another sheet, called a "Work Approval" (Figure 2).
Similar to the original worksheet, the Work Approval displays the
results as well as a field for approval initials. A supervisor or
group leader compares the printed approval sheet against the original
raw data. If results are properly calculated and correctly entered,
the reviewer initials the item as approved. Data review takes place
early in the morning, with the completed approval sheets turned in by
midmorning. Another data entry operator keys in the approval initials.
At this stage, a command is executed that moves the item from the Work
file into the "Samples" file, the final archive location for results.
Corresponding QC items are kept in a parallel file called "Samples,QC".
The program that prints finished reports retrieves the results from the
Samples file. Figure 3 illustrates the customary format for reports.
As with most such systems, the client and project information are
displayed in the header area followed by the individual analytes and
results. With this system in place, BCAL decided in late 1985 to look
into adding QC evaluation to the LIMS.
DEVELOPMENT
As we began the first stage of developing the EWR, we saw that dialog
between the chemists experienced in quality assurance and the computer
programmers would be essential. QA chemists alone cannot implement an
automated system and programmers cannot know the decisions the QA staff
will want made. The project included, at various times, the QA
director, the Pasadena laboratory's QA coordinator, two in-house
programmers, and two outside programmers (including the LIMS designer).
None worked on this development full time.
The first critical design decision was selecting where to intercept the
existing process. Our principal considerations were (1) before work
approval, (2) after work approval, or (3) at final reporting. We
rejected the final report stage as being dangerously late in the
analytical process. If a QC problem occurred on the first analysis
reported but was not discovered until the last item was complete, the
process might be out of control for several days. The next choice was
less clear-cut: Evaluation after work approval would allow simple data
entry blunders to be corrected before appearing on the EWR; evaluation
before work approval would shorten the cycle by one day. The urgency
of correcting a process at the earliest possible time prompted us to
select option (l)--evaluation before approval.
1-197
-------
PfttE
U08KSHEET FQ9 B'JLTIPLC RESULTS
0? IWY 198?
l):0, Uol
6EPT SCCU& OUC
TP CLICK!.
SAItPLE DESCRIPTION.
SANPU LOC sc mn KQTES.
1 CC 09 BOY 23 NAY 14 SO CC.PA'
2 CC 09 BAY 23 BAY 14 SO DC. PP.!
3 CC 09 HAY 23 BAY 14 SO CC.PA!
4 CC 09 BAY 23 BAY 14 SO BC.fHj
5 tC 09 BAY 23 NAY 14 SO CC-PA,
( CC 09 BAY i3 RAY P SO CC.PAi
? CC 05 KAY 23 BAY ;l ID CC.PA
ACCESSI 1C TYPE
:<.R?L[ LOC
UNITS ..
P«t«ctiol LIBII ...
CATCHI.. . . .. . ...
BETHOO .
E8U1P
1 . -DATE XTP Olte Extrjct**
i ,OIL FRC 6ll'/tio« fioUr
) CENZCNE Eeszene . . ..
4 fOL. . lolmne
5 EtDoz EthglbenieM
1 CCnz CUorobenzen* . . .
A Tes
A T«s
ft T«s
A Tss
A T-s
A Tes
A T
0
<
—
t/1
It
"
1-198
-------
BROWN AND CALDWELL LABORATORIES
ANALYTICAL REPORT
373 SOUTH FAIR OAKS AVENUE, PASADENA, CA 91105
(818) 795-7553 (213) 681-4655
FAX: (818) 795-8579
LOG NO: P89-05-001
Received: 09 MAY 89
Reported: 09 MAY 89
Karen Shupe
Brown and Caldwell
150 S. Arroyo Parkway
Pasadena, California 91109
Project: 450-01
REPORT OF ANALYTICAL RESULTS
LOG NO
SAMPLE DESCRIPTION, SOIL SAMPLES
Page 1
DATE SAMPLED
05-001-1 Test 1
05-001-2 Test 2
05-001-3 Test 3
05-001-4 Test 4
05-001-5 Test 5
PARAMETER
Vol.Aromatics (EPA-8020)
Date Extracted
Dilution Factor, Times 1
Chlorobenzene, mg/kg
1 , 2-Dichlorobenzene , mg/kg
1, 3-Dichlorobenzene, mg/kg
1,4-Dichlorobenzene, mg/kg
Benzene , mg/kg
Ethylbenzene, mg/kg
Toluene, mg/kg
05-001-1
05/10/89
1
20
20
20
20
20
20
20
05-001-2 05-001-3
05/09/89 05/10/89
1 1
20 <10
20 <10
20 <10
20 <10
20 <10
20 <10
20 <10
05-001-4
05/10/89
1
30
30
30
30
30
30
30
09 MAY 89
09 MAY 89
09 MAY 89
09 MAY 89
09 MAY 89
05-001-5
05/10/89
1
<10
<10
<10
<10
<10
<10
<10
Jeffrey A. Erion, Laboratory Manager
Figure 3-
Analytical Report
1-199
-------
Next, we had to choose which data to track and what criteria to use for
evaluating that data. Already implemented in LIMS were laboratory
control standards run on a pure reference matrix and duplicate and
spike results run on actual sample matrix types. Thus, we elected to
use LCS results as method accuracy criteria, measured as percent
recovery. Matrix duplicates would serve for precision verification,
measured as relative percent difference (RPD). Both spikes and
duplicate spikes are used for matrix accuracy evalution, again measured
as percent recovery. For evaluation of precision on those tests
subjected to duplicate spikes, we selected the RPD of the spike pair as
the precision variable.
Most laboratory QA programs use a combination of historical and
mandatory acceptance criteria. Many of the newer analytical methods
include guidelines on percent recovery requirements for pure-matrix
samples comparable to our LCS. Where these are available, we have
adopted the stated standards as our control limits. As these are
usually based on something close to three standard deviations of a
series of measurements, we have adopted two-thirds of the range
(approximately two standard deviations) as our warning limits. For
tests without method requirements, we have used a series of at least 20
measurements, gathered during a period when the process is postulated
to be "in control," to calculate warning and control limits.
Very few of the methods include requirements for replicate precision.
Because the order of analysis is not usually the variable of concern,
RPD is expressed without a sign--as absolute value. Consequently, a
series of difference calculations makes up a one-tailed curve with zero
difference as the lower limit value. To calculate the long-term
statistical variation, we used the formula of John K. Taylor of NIST:
2 1 k 2
S = -- 2d
2k 1 i
Where: k = number of sets of duplicates
d = difference of duplicate measurement
i
S has k degrees of freedom
Once this calculation of S squared--the variance—is completed, the
standard deviation is simply S, its square root. We set warning limits
at two standard deviations and control limits at three standard
deviations. In sum, the variables selected and criteria for comparison
were as follows:
Laboratory Control
Standard
(LC and LT)
Method Percent recovery compared to
Accuracy method requirements, if avail-
able. Historical calculation
if not in method.
1-200
-------
Duplicate Samples Sample Relative percent difference,
(Rl and R2) Precision compared with historical
calculation.
Duplicate Spikes Sample Relative percent difference,
(SI and S2) Precision compared with historical
calculation.
Spikes or Sample Percent recovery, compared
Duplicate Spikes Accuracy with historical calculation.
Once we selected these variables and criteria for evaluation, we needed
a way to associate them with certain analytical results. That is, if a
LCS fails acceptance criteria or a duplicate RPD exceeds a control
limit, what sample results are affected? While we were considering
this issue, the EPA published specific QC guidance in the third edition
of SW-846. We found we could meet our own needs and the requirements
of SW-846 by converting our QC testing to a "batch" basis. With this
operational change we would define the beginning and end, for every
analysis, of a closely related group of samples that are run together
at the same time (or sequentially) using the same method, equipment,
reagents, and analyst. Within such a batch, the LCS and the matrix QC
results will serve as the stamp of approval for release of the entire
batch of results. Correspondingly, a QC failure calls into question
all of the batch results, potentially requiring reanalysis of the
entire batch.
Because the LIMS-initiated QC had originally been structured for a
"flow-through" frequency of one-in-ten, some changes were needed to
accommodate batch QC. Reprogramming now allows the analyst to add new
QC items to the work list as needed to keep QC samples in every batch
of client work. We left the automatic generator set at the one-in-ten
frequency to provide a reminder, but the analyst can add to, delete, or
override the computer-initiated QC items. The fundamental operating
rule is that every batch must have QC appropriate to the method being
used.
The second stage of EWR development was the production of decision
algorithms. The end user--in this case, the quality assurance
coordinator--must reduce the judgments to simple yes-or-no decisions if
the computer is to handle them readily. Through dialog with the
programmer, the user can develop a simple statement about what
characteristic is to be compared to what limit. The statement not only
includes the logic of the decision, but also incorporates mathematical
manipulation of specific variables. Once these are clear, the
programmer can translate the algorithms into computer code to create
the QC review program. In consideration of the variables and criteria
previously selected, the following algorithms were worked out between
the QA coordinator and the computer programmer.
1-201
-------
Question
Calculation
Algorithm
Does the lab
control standard
meet acceptance
criteria?
Is duplicate (or
duplicate spike)
precision
acceptable?
Is spike recovery
acceptable?
Was spike range
appropriate for
the sample con-
centration?
Divide LC by LT,
multiply by 100 to
get percent
recovery.
Subtract R1-R2 (or
S1-S2), divide by
mean of the pair,
multiply by 100,
and take absolute
value to get RPD.
If duplicates run,
calculate mean R;
if not, use Rl.
Subtract R from S
and from T; divide
these results to get
spike recovery.
Multiply by 100 for
percent spike
recovery.
Subtract S-R to get
amount spiked.
Calculate a value
10 times archived
detection limit for
this sample matrix;
calculate 0.5 times
and 5 times R.
Compare percent
recovery with upper and
lower warning and
control limits. Return
an error condition for
results out of bounds.
Compare the RPD with
the upper control limit
and the upper warning
limit. Return an error
condition if it exceeds
limits.
Compare percent spike
recovery to upper and
lower warning and
control limits. Return
an error condition for
results outside limits.
Check amount spiked
against 10 times detec-
tion limit. If less,
return an error condi-
tion. Compare amount
spiked to sample native
level. If less than
half or more than 5
times, return an error
condition.
With these algorithms defined, other questions arise. Among these: Is
every error condition equally serious? Should we hold up release of
results in every case? In our system, we recognized two levels of
error. The first, associated with warning limits, is a "non-terminal
error," which warns of QC problems, but does not interrupt data flow.
The second, called a "terminal error," prevents data release until
corrective action is taken. To accommodate changing requirements in QA
programs, the system provides for the future conversion of one level of
error to the other through use of a simple error flag.
Another question concerned the location and structure of files that
contained so many limits. Besides the calculated or method-mandated
acceptance limits, the file must contain matrix-specific detection
limits for every parameter. To archive these new features, we selected
the file already used to control matrix-specific QC frequency. We
1-202
-------
prefer that the QA coordinator update these fields manually rather than
have the computer update them automatically. Although it increases the
labor burden, this practice improves accountability and consistency.
When all parties clearly understand the algorithms, the programmer can
begin direct coding. In this, the third stage of development, the
system moves from concepts and rules into computer instructions that
can actually be executed. Because computer code is so
system-dependent, only the logic outline is presented here.
The block at the top of Figure 4 represents the point of intervention
in the existing LIMS. A program selects results keyed into the the
data base but not yet approved for release. These results are divided
into two groups. In the first group are results vhich do not have
batch numbers or are not subjected to QC requirements. Their character-
istics might include the date a sample was put on hold, flow figures
provided by a wastewater client, or other nonanalytical information.
This group is diverted from the review program and sent directly to
approval sheets.
The second group, which includes work subject to QC review and with a
batch number, goes on for review. This program module carries out the
arithmetic calculations and comparison algorithms. As mentioned
earlier, three results are possible. (1) If no error condition
occurs, the sample results go to the approval sheets without further
action. (2) If there is a QC failure, but it is non-terminal, the
results are printed on both the EWR and the approval sheets, but with
an asterisk to assure careful review. (3) If the QC failure is a
terminal error, the results appear on the EWR but not on the approval
sheets. Correcting errors is one way to move the results onward. The
QA coordinator or a section supervisor can also manually release the
results by issuing an override instruction to the program. After all
items reach the approval sheets, reporting goes forward as in the
original LIMS design.
The fourth major stage of program development is testing. Each of our
computers contains a test account where programs can be executed in a
simulated version of LIMS. When we tested the EWR in this account, we
discovered a few minor errors: some of the alogorithms had not been
fully translated into code, and some of the arithmetic comparisons
suffered from poor numerical precision in the system. Nonetheless, the
difficulties at this stage were minor and relatively easy to correct.
In the second part of testing—provisional implementation--we started
the program on one of our three systems and observed its progress.
Before startup in our Pasadena laboratory, the QA coordinator met with
each analytical group to explain planned changes. The programmers
started each error condition as non-terminal, so no client reports
would be immediately affected by EWR. Training also involved the data
entry operators who would enter the new information being captured in
the data base--primarily the batch number.
1-203
-------
Figure 4. Block Flov Diagram for Early Warning Report
Work Ready
for Approval
Select
Work
For QA
Stamp
Yes
Work With Batch
Number and QA Flag
Work with no
Batch Number or
without QA Flag
Pass
Fail
Terminal
\f
Work Approval
Sheets
Correct error
or manual
release
Fail
non-
Terminal
Hard Copy
EWR Report
V
Final
Report
1-204
-------
The system started up smoothly, but one major flaw and two minor ones
quickly became apparent. In the initial design, the QA Department had
decided that batch numbers could be assigned fresh on each day of
analysis. Thus, for antimony tests run on March 28, the first batch of
the day would be Antimony Batch Number One; the second batch of the day
would be Number Two, and so forth. On March 29 then, the first
antimony batch would also be Antimony Batch Number One. During the
testing, analysts asked how to handle batches on autosamplers set to
run past midnight--the date analyzed for some of the batch was a day
later than other members of the same batch. Clearly, this numbering
scheme was not adequate.
After vigorous discussions, the computer programmers and the QA and
operations staff developed a new scheme. Now the batches are numbered
sequentially for each determination. The counters start at number one
on January first and advance incrementally with each batch reported.
The computer prints on the worksheets the next batch number expected
based on the last batch reported. If analysts run more than one batch
in a day, they look for the next sequential number and write it on the
sheet for the second batch. This system resolved the major problem.
One of the two minor problems concerned holding a batch open for
reporting. The designers had assumed the analyst would report all the
batch results at the same time on the same worksheet. In actual prac-
tice, this is not the case. Rush work, for instance, may be turned in
by itself late in the day while the rest of the batch waits another day
for complete data reduction. If the QC is reported with the rush
results, as it should be, the EWR will indicate an error the next day
when the remaining results are reported without batch QC. Some new
programming has allowed an analyst to hold open a batch for additional
results: The analyst turns in the batch number alone on all associated
work, even those items with pending results. Once the batch number is
entered, the program holds the QC items suspended for continuing com-
parison as each new item is archived. When all members of the batch
have been reported, the QC results are archived as well.
The second minor problem concerned analysis by closely related determi-
nations. For some clients, we employ special subsets of compound
lists. That is, for an ongoing well monitoring program, a client may
request routine reporting of benzene, toluene, and ethyl benzene
only--rather than a full 8020 list. The analyst will run these short-
list tests alongside full 8020 samples in the same batch. Our LIMS
uses distinct determination codes to differentiate the short list from
the full list, so the two determinations have different batch counters.
If the analyst reports QC on one determination, the other appears to
the EWR to have been run without QC--an error condition. Because the
logic of exactly associating related determinations is more complex
than our LIMS can currently handle, we had to settle for a partial
solution. Determination codes for major organics reflect their common
method numbers. The full 8020 list is determination 8020; the short
list adds a period and a suffix, 8020.SHORT. With that consideration,
we readjusted the counting routine so all determinations that match
exactly up to the period are combined in counting registers.
1-205
-------
Vith these three errors corrected, we implemented the EWR in our other
S laboratories, repeating the training cycle and holding -undtable
discussions with staff to consider the merits of batch QC and the EWR.
in eaS implementation, startup vas smoother than in our initial
testing stage. As it now operates in all three laboratories, the EWR
tracks and reports on several key quality control variables.
EWR FEATURES
in its final form, the Early Warning Report is a daily summary of QC
results that require the further attention of an experienced chemist.
Printed out with the worksheets and approval sheets every night, the
EWR is divided by department and distributed every morning. Figure 5
displays one page from a typical report.
Potential QC errors are assigned a flag value of "zero" for non-
terminal and "one" for terminal. If the error is terminal, all the
samples for that batch are printed on the EWR but not on the approval
sheets. If the error is non-terminal, the results appear on both EWR
and approval sheets. Consequently, non-terminal errors may be
overridden by simple approval, whereas terminal errors require the QA
coordinator or a supervisor to intervene and manually move the results
over to approval sheets--a process called "manual release." To ^assure
that non-terminal errors are not overlooked during the EWR review, an
asterisk appears on the approval sheet beside the result for every item
with a non-terminal error.
Terminal errors currently in use, which all relate to the LCS:
Non-numeric results in LC or LT. As the numbers depend upon
percent recovery for validity, they should always contain
numerical values. Non-numeric data indicates a potentially
serious data entry error.
LC or LT missing. No batch may be approved without a laboratory
control standard.
LCS >= Upper Control Limit. The percent recovery exceeds the
archived upper limit for the parameter in question.
LCS <= Lower Control Limit. The result is below the lower
archived level.
Some non-terminal errors that also concern the LCS:
LCS concentration not on file. The QC data file that contains
the limits also contains a customary true concentration for the
LCS. A different one indicates either an error or a deliberate
choice to deviate. As the choice may be valid, the error is
non-terminal.
LCS >= Upper Warning Limit. / LCS <= Lower Warning Limit. These
errors correspond to the control limit errors, but refer to the
1-206
-------
AHO CALDHELL EARLY UfiRHlKG REPORT
AS OF 09 RAY 198?
PACE
HGRK.
DATE.. BATCH* CLIENT.
Analyzed
LOS NO.
8Q2G*PC905002xi*LC
05.0?.89 1
LAB CONTROL 05-002-1
802QxPC905902xi*LT
05.09.89 1
LAB CDKTRQL 05-002-1
8
-J
05.09.89 1
BC.PASA
05-001-1
«ULT DET.
UNITS. RESULT..
8ft OUTLIERS.
LEVEL DESC
Date
Tines
BENZENE
TQL
EtDnz
.CBsz
l,2-&CBnz
1,3-DCBar
1,4-C'CBr.z
Date
Tines
BENZENE
rot.
EtBaz
.C&nz
1,2-DCCnr
1,3-DCBnz
l,4-&CBni
Date
Tines
BENZENE
TOL
EtBiiz
.CBaz
1,2-DCBai
1,3-DCBaz
1,4-DCBaz
ug/L 05/09/89
1
10 *LCS concentration not on file
10 *LC3 concentration not on file
10 »LCS concentration not on file
10 »LCS concentration not on file
10 *LC3 concentration not on file
10 *LC3 concentration not on file
10 *LCS concentration not on file
ug/L 05/09/89
1
15 MLC3 concentration not on file
15 1LC3 concentration not on file
10 *LC3 concentration not on file
15 *LCS concentration not on file
15 aLC3 concentration not on file
15 *LC3 concentration not on file
15 *LCS concentration not on file
ng/kg 05/09/89
1
5 xLCS concentration not oa file
5 *LCS concentration not on file
5 *LCS concentration not on file
5 *LCS concentration not on file
5 *LC3 concentration not on file
5 *LCS concentration not on file
5 *LCS concentration not oa file
O)
-\
*<
0}
73
CD
IQ
C
VJ1
05.09.89
MX*
-------
narrower vindovs of the warning limits. They caution the analyst to do
a close review of the data, but do not prevent reporting.
Matrix precision, which is reviewed for several non-terminal errors:
Rl Result (or R2 result) <= Detection Limit. If no measurable
result was achieved, the RPD calculation will not be valid.
RPD >= Upper Control Limit / RPD >= Upper Warning Limit. This
calculation is used for both Rl, R2 and SI, S2 pairs. If the
difference exceeds the archived limit, this error condition
occurs.
Non-numeric data in Rl or R2. If an RPD calculation is
attempted but encounters non-numeric results, this error occurs.
Matrix accuracy, which considers spike recoveries:
% Recovery >= Upper Control Limit (or Upper Warning Limit).
High spike recoveries are tagged with this message.
% Recovery <= Lower Control Limit (or Lower Warning Limit). The
corresponding low-spike error conditions.
Spike value is < 10 times detection limit. If spike is too
small to be a reasonable measure of recovery, this condition is
noted.
Spike added concentration is < half the sample concentration.
This error condition occurs if the spike result is overshadowed
by the native concentration of the analyte.
By reviewing all of these features daily, we have greatly reduced the
possibility of releasing suspect data. As we gain additional experi-
ence with the program, we will likely convert some of the non-terminal
errors to terminal status. We are also developing a way to put method
blanks on-line as part of each batch record. Some error conditions
will be developed to alert us to contaminated blanks once that process
is complete.
CONCLUSIONS AND RECOMMENDATIONS
The EWR has been well accepted throughout our laboratories. While a
few analysts initially resisted the idea of being checked by a
computer, we have integrated the report into routine data-handling
activities after only a few months of use. Because the system already
ties QC results and associated client samples together, it gave us a
platform on which to create a data retrieval program for preparing a
Batch QC Report for clients on request.
The EVR provides two distinct benefits not otherwise available without
a great expenditure of labor: (1) it reviews every QC result for every
test generated every day; (2) it compares those results to absolutely
firm limits established by the QA coordinator.
1-208
-------
Comparison to rigid limits, especially in the presence of matrix inter-
ference, may be too restrictive. The non-terminal error choice allows
such conditions to be flagged and reviewed without impeding the flow of
approved work.
Any laboratory with a custom-written or adaptable LIMS should be able
to create something like EWR. Key steps in the process are as follows:
1. Select a node in the data process where errors can best be
captured.
2. Choose a single QA person as primary laboratory contact
with the programmer to give a single voice to the
laboratory's needs.
3. Create an environment that allows frequent contact and
thorough dialog between the QA representative and the
programmer.
4. Decide which data will be compared with which criteria
before beginning the programming.
5. Involve analysts thoroughly in the testing stages—and plan
on time for rewrites.
If your LIMS is at least as adaptable as ours--and most are--your
efforts should be well rewarded.
REFERENCES
U.S. Environmental Protection Agency, Handbook for Analytical Quality
Control in Water and Wastewater Laboratories, 1979. EPA-600/4-79-019,
Cincinnati, Ohio.
U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency Response, Test Methods for Evaluating Solid Waste, Third
Edition, 1986. SW-846, Washington, D.C.
Taylor, John K., Quality Assurance of Chemical Measurements, 1985.
National Bureau of Standards, Center for Analytical Chemistry,
Gaithersburg, MD.
1-209
-------
A QUALITY ASSURANCE AND MANAGEMENT SYSTEM
FOR LARGE ENVIRONMENTAL PROJECTS
Joel Karmazyn, Carol Schrenkel, and Sharon Nordstrom, Roy F.
Weston, Inc., West Chester, Pennsylvania 19380
ABSTRACT
Environmental investigation often generate massive quantities
of analytical data. In investigations that deal with physically
large areas, or that are multi-site or multi-media, the
problems normally encountered in tracking, validating, and
managing the data are compounded.
To more efficiently and effectively evaluate and process data
from large projects, particularly those where data/report
turnaround is on a critical schedule, a streamlined data
management and quality assurance system was devised. The data
are tracked from their inception in the field through the
entire review process until the final report is prepared. This
unique system also allows the project team to know at any point
in time the status of any piece of data.
This paper will present the data management process and
tracking system. The quality assurance reviews, performed
concurrently at the laboratory and within the project team,
will also be described.
INTRODUCTION
The management of field and laboratory data has traditionally
been a slow process due to segregation of the project data and
management activities. Interaction among key players such as
laboratory personnel, computer scientists, engineers and
scientists rarely occurs. Traditionally, each discipline
performs appropriate tasks in sequence, creating a lengthy
period of time between sample collection and programmatic data
usage. This process is not cost effective for conducting large
environmental investigations which require ongoing field
activities (RI/FS). An alternative which takes maximum
advantage of a multi-disciplinary approach to data management
and quality assurance (QA) was developed by a team of
scientists and engineers from Roy F. Weston, Inc. (WESTON). The
data management team represents disciplines in chemistry,
geology, engineering and computer science. The two primary
objectives of the team are to streamline various QA functions
and to provide input for data assessment. A key factor in this
program is that all personnel on any given investigation are
intimately involved with project activities and historical
data. This enables team members to note discontinuities during
data validation, electronic data transfer, and data assessment
and plotting. The data management flow scheme is presented in
Figure 1.
1103R2 1-210
-------
COC Distribution
Project
Files
(Master File)
1
(Revised CoC)
Field
(Take Samples and Prepare Copy)
(DCN Must be Submitted Within 72 Hours)
Lab
(Assign Batch and Lab Sample
Identification Numbers)
"
Data Coordinator
Data
Entry
Project Engineer/
Scientist
(Correction
(Make and Distribute Copies)
IDS
Group
Information
DCN-Document Change Notice
TIMS
Coordinator
for DCN)
Data
Review
(DCN if Corrections Required)
FIGURE 1 DATA MANAGEMENT FLOW SCHEME
1-211
-------
SAMPLE GENERATION
Prior to field activities, a sample code is developed to
uniquely identify all samples to be taken. This code is
referred to as the "client ID" and provides site and sample
specific information. It includes the identification of field
and laboratory QC samples such as matrix and blank spikes,
field, trip, and method blanks, field duplicates, etc. The
client ID is also the link to historical data, stored in the
Technical Information Management System (TIMS) database, which
can be accessed at any time to retrieve and manipulate data.
As samples are collected, they are grouped into laboratory
batch-sized lots for shipment (approximately 20 samples)- Upon
receipt at the lab, they are assigned lab batch and sample
numbers (lab sample ID), and logged into the Laboratory
Information Management System (LIMS). Analytical instrumenta-
tion is tied into the LIMS system. As analytical data are
generated, they are collected and stored in LIMS by lab sample
ID. LIMS is used to generate the hard copy data reports and
also directly interfaces with TIMS so that project analytical
data can be directly transferred without the need for manual
re-entry or downloading into an intermediate database. LIMS
also provides the lab project coordinator with a means of
tracking the status of project sample batches as they progress
through the laboratory system.
Immediately upon sample receipt and log-in, a copy of the
chain-of-custody (COC) with the assigned lab sample IDs is sent
to the project team. This allows the site engineer to double-
check for field errors before samples are analyzed and data
reports are generated. Field information is submitted in a
TIMS-ready format, and after a QC check of it and the COC, both
are submitted to the data management group for entry into the
TIMS database. When the analytical data are received from LIMS,
TIMS pairs them with the field data to produce a complete
sample data file.
DATA TRACKING
Receipt of the COC also triggers the project data tracking
system, which is maintained on a personal computer. Figure 2 is
an example of the data tracking system. It is set up so that
data can be sorted by site, lab batch number, analyte, or
sample location (onsite soils, offsite soils, sediment, special
sampling media, groundwater, soil boring, etc.).
The tracking forms are also kept in a notebook, segregated in
alphabetical order by site (or whatever is convenient to the
project). When a COC is received, the batch number, analytes
requested, and sample location/type are entered. This ensures
that the batches are in numerical order for each site. When the
1103R2
1-212
-------
no
CO
Batch No.
8902-L-462
8903-L-571
8903-L-590
Anal
Pest
VOA
BNA
Metal
CN
DIOX
VOA
PCB
Task
Code
GW
ON
OFF
Data
Log
No.
005
010
015
016
190
Master
File
No.
0265
0333
0362
Check-off List
Date
Data
Rec
3/27
3/20
3/25
3/25
4/02
Date
Ret
Rec
3/27
3/21
-
-
4/02
Date
Corr
Rec
4/04
3/28
-
-
4/09
Date
Corr
Rec
3/30
3/30
-
-
4/09
Date
Pkg
App
3/30
3/30
3/26
3/25
4/10
G17-125a
FIGURE 2 SAMPLE DATA TRACKING SYSTEM
-------
status of a specific batch is needed, it can be found easily.
The master file number is assigned at this time so that the
batches are sequential within the project files.
When an analysis for a batch is completed, a hard copy of the
laboratory deliverable is submitted to the project team. The
hard copy data package typically includes a cover page; a copy
of the COG; a chronology detailing lab and client IDs, sampling,
extraction, and analysis dates; a spreadsheet data summary;
case narrative; full sample and standards chromatograms; and
QC. "Short copies" (excluding raw data) are provided for the
project manager and the lead engineer. The log number, which is
sequential within a project task (such as a round of
groundwater sampling), is assigned at this time. The original
full package is filed in the master file, and a copy is
distributed for Project QA.
The site engineer/scientist begins the QA process by initiating
the data review checklist (Figure 3). The client IDs are
checked against map locations, and a document change notice
(DCN) is filled out to change any erroneous IDs. The COC is
checked against the chronology and data summary to be sure that
all samples requesting a specific analysis were analyzed. The
data are also given a "reasonableness check" for any data that
appear anomalous due to sample location (upgradient well,
off site background soil) or a comparison with historical data
that indicate a need for a more in-depth review. Any comments
are written on the checklist, which is signed, dated, and
forwarded to the QA group.
When the data package is received by the QA group, the log
number and date received are filled in on the data tracking
form. The remainder of the checklist is then completed, which
includes a cursory review of chromatograms, standards data,
reporting format, and compliance with methodology and project
protocol. Any data requests from the project engineer are
checked quantitatively.
If the package is correct as submitted, the data review
checklist is dated and signed. A copy is given to the project
manager, and the original is filed in the master file with the
original data package. The date approved is entered on the
tracking form. If corrections are necessary, or clarification
of an interpretation is needed, the questions/comments are
indicated on the checklist, which is telecopied back to the
lab. Revised data pages are returned, or comments are answered
on the checklist. All dates are entered in the tracking system,
allowing approximately one week for corrections. When the
package is approved, the checklist with the questions/answers
is signed and dated, a copy given to the project manager, and
the original with all corrected pages put into the master file.
The approval of the hard copy data package also signals the
authorization to transfer the analytical data electronically to
1103R2
-------
Site Name:.
File #:
Delivery Date; Client.
Reviewed by: Geo:.
. Batch #_
.Log#:_
. State:—
.Eng:
. Analyte:.
.Today's Date:
-Chem:.
D.C.N. Incorporated in Package
Chain of Custody Protocols Followed
Chain of Custody Matches Chronology
Sample Holding Times Met
Case Narrative Flagged Samples
Fvrppding Holding Time
Data Summary Sheet Matches Chronology
QC Data Included
Matrix Blanks ^ean
FiplH/Trip Blanks Clpan
Surrogate Recovery Acceptable
MS/MSD Recovery Acceptable
Appropriate Detection Limits/ Dilutions
J - Values Correct
B - Values Correct
U- Values Correct
Case Narrative Describes
Analytical Diffirultips
Tvnos
Yes
No
NA
Comments
Other Comments:.
Geo:-
Eng:-
Project Manager:.
Date Submitted to P.M.:.
Date:-
Date Submitted for Revisions:
Response Due Date:
Date Received:
Date of Resubmittal:.
Date of Package Approval:.
G17-125
Response Due Date:.
Date Received:
Chem:
FIGURE 3 DATA REVIEW CHECKLIST
1-215
-------
TIMS. Prior to release to TIMS, the LIMS electronic data file
is compared with the "final" hard copy package to ensure that
both are identical and complete. The file is also reviewed by
project data management personnel for ID inconsistancies,
missing or incomplete data records, and discrepancies between
field and laboratory information.
The advantage of this system is that anyone needing to know the
status of any data only has to look at the data tracking pages
for the site in question. This will tell at a glance what
batches have been received at the lab for which analyses,
whether a data report have been generated, whether corrections
have been requested and when the results are expected, and if
the data have been approved. When putting together a site
report, it is critical to be able to tell quickly that all data
have been reviewed and approved.
CONCLUSION
The data management process described embodies four levels of
quality assurance conducted during sample preparation and sam-
ple collection, laboratory analysis and validation, engineering
and scientific project review, and analytical project review.
This multi-disciplinary approach has proven highly successful
at WESTON for nearly two years. Numerous large investigations
with massive databases have been tracked, managed, and assessed
in a fraction of the time normally required. Our success is
principally due to the commitment of the various departments
for developing the process. Several key elements were the
development of TIMS and the LIMS-TIMS tie-in, the code system
which allows continuous tracking, and the close working rela-
tionship developed between the laboratory and a project team.
The process continues to be refined in the areas of computer
data manipulation, computer data validation, advancing LIMS
capabilities, and streamlining the project review process.
Advanced computer data manipulation holds the greatest promise
since other areas have reached practical and economical limits.
ACKNOWLEDGMENTS
We would like to thank Caroline M. Power, Susan Davis, Kay
Adams, Gerry Andrews, and the staff of WESTON's Goshen
Publications Department for their support and assistance in
preparing this paper.
1103R2
1-216
-------
A SMART DATA BASE SYSTEM FOR SELECTING
ANALYTICAL METHODS FOR ENVIRONMENTAL ANALYSIS
CONCEPT AND DESIGN
R. A. Olivero. Senior Scientist, J. L. Boyd, Computer Systems Analyst,
Quality Assurance Department, Lockheed Engineering & Sciences Company,
1050 E. Flamingo Rd. , Las Vegas, Nevada 89119; D. W. Bottrell, Chemist,
Quality Assurance Research Branch, U.S. EPA Environmental Monitoring
Systems Laboratory-Las Vegas, P.O. Box 93478, Las Vegas, Nevada, 89193-
3478; M. T. Homsher, Director of Quality Assurance, National Sanitation
Foundation, P.O. Box 1468, Ann Arbor, Michigan 48106.
ABSTRACT
The expansion of legislative requirements for environmental testing and
the necessity of appropriate analytical procedures have made the
identification of adequate and cost-effective data acquisition,
documentation, and validation alternatives a complex process. Smart
systems designed to sort options based on information need, analytes, and
other criteria represent a consistent, easy-access reference to assist in
selection. The approach described in this work uses a relational data
base, a user-friendly front end and help function, and essentially real-
time data base updates to manage and access information about matrix,
analyte, extent of documentation, time and cost requirements, and data
quality characteristics for indexed analytical methods.
INTRODUCTION
Cost-effective generation of environmental information requires sampling
and analytical techniques that are sufficient as well as efficient to
provide the required data. The elements of risk assessment models expand
as routes of exposure are added, sampling procedures are characterized and
refined, and additional analytical procedures for field and laboratory
become available. The options for choosing techniques, analytes, length
of study period, management risk, and other elements of environmental
studies, are becoming progressively more complex and significant. The
primary goal is the selection of optimal sampling and analysis options to
maximize the information obtained while minimizing cost and delay.
Ultimately, these initial decisions affect the adequacy and cost
effectiveness of the information available for making decisions about
specific sites. Information is available about how sampling and
analytical techniques relate to various matrices, site characteristics,
and information needs. However, few individuals possess the integrated
experience to identify the optimal set of procedures.
1-217
-------
Recent requirements for and development of alternative laboratory and
field procedures appropriate for rapid data generation with different data
quality characteristics continue to complicate the selection of technical
alternatives. In addition, the need for high quality, cost-effective data
that includes sample- and matrix-specific qualifications requires choices
to be made among various methods.
A common demand for analytical methods information arises from the need
to comply with legislative requirements and obtain legally defensible
data. Some of the pieces of environmental legislation that require
monitoring of specific pollutants are the Federal Water Pollution Control
Act (FWPCA), the Safe Drinking Water Act (SDWA), the Resource Conservation
and Recovery Act (RCRA). the Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA or Superfund), the Clean Air Act,
and the Marine Protection, Research and Sanctuaries Act (MPRSA) [1].
The limited access to expert knowledge about how to obtain required
environmental information, the need for rapid identification of
appropriate, cost effective options, and the need for a standardized,
documented rationale for making selections all indicate a problem that may
be appropriate for the application of expert systems. In addition to
meeting immediate needs, expert systems frequently function as tutorials
to assist users in developing subject areas of expertise. The current
requirement for rapid acquisition of appropriate environmental data
suggests the evaluation of expert systems as a practical, cost-effective
approach.
A needs assessment study performed to support the U.S. Environmental
Protection Agency (EPA) Expert System Initiative indicated a high priority
for the development of a Smart Method Index [2]. The EPA Environmental
Monitoring Systems Laboratory-Las Vegas (EMSL-LV) is responsible for
developing this application. In the first phase of development, a
prototype has been proposed to test the feasibility of the preliminary
design. This work describes the conceptual prototype illustrating option
fields, scope, user interface features, and the structural design. Active
participation by potential users in the review, evaluation, and design of
prototype expansion is essential to provide a usable product.
ANALYTICAL CHEMICAL METHODS FOR ENVIRONMENTAL APPLICATIONS
EPA has fostered the development and validation of analytical chemical
methods to measure pollutants of environmental impact included in
regulation lists. The EPA Contract Laboratory Program (CLP) organic and
inorganic methods have wide application and well-characterized performance
and provide legal admissibility of the results. The CLP methods are an
adaptation from some of the methods in 40 CFR 136 [3]. The 40 CFR 136
listing now contains 262 analytes and over 500 test methods.
Many other methods applicable to specific analytes and matrices exist or
are under development. Some non-CLP methods with wide recognition include
1-218
-------
the SW-846 methods [4], Drinking Water methods, and National Pollutant
Discharge Elimination Systems (NPDES) methods. There are various other
organizations involved in developing, validating, and promulgating
analytical methods with environmental applicability. Some of these are
the National Institute of Occupational Safety and Health (NIOSH), the
Association of Official Analytical Chemists (AOAC), the Food and Drug
Administration (FDA), and the American Society for Testing Materials
(ASTM).
The need to standardize and reduce cost by preventing duplication of
effort is well known [5] . For example, a congressional recommendation has
been made for "the establishment of a computerized catalogue of the
availability, applicability and degree of standardization of methods
currently in use in the [U.S. Environmental Protection] Agency" [6]. The
microcomputer-based List of Lists (LISTS) system [7] , a data base
developed by EPA containing over 50 methods, has been proposed for
adaptation to meet that mandate. The implementation of the recommended
data base is expected to facilitate the selection of appropriate methods
and the setting of reasonable data quality objectives for specified
applications; promote method consolidation and identify duplications; and
increase general knowledge of the available environmental monitoring
methods.
Independent attempts have been made to catalogue and to computerize
analytical methods for environmental applications. The Index to EPA Test
Methods compendium contains over 700 air, water, and waste methods [8].
Microcomputer-based prototypes have been developed for water pollutant
analysis methods using Rulemaster® and C language [9] , for SW-846
inorganic analysis methods using Pascal language and for SW-846 organic
analysis methods using dBase® [10], for SW-846 used-oil methods using
Prolog language [11], and for general testing methods using dBase® [12].
The U.S. Department of Defense has prepared a hard-copy compendium of
methods in use with applicability for determination of radioactive
pollutants [13]. Some of the computerized systems mentioned use expert
system techniques in their implementation.
Recently EPA has identified the development, characterization, and
application of field analytical methods as an immediate need. For
example, ongoing research at EMSL-LV encompasses field X-ray fluorescence
[14], soil gas [15], and field gas chromatography [16] techniques. These
methods, implemented by EPA, are described in a computerized catalog, of
field screening methods covering 31 field methods [17]. In addition,
rapid turnaround options for CLP analysis of volatile, PNA, phenol,
pesticide, and PCB organic analytes will soon be available. An increasing
number of analytical methods are being developed or modified as new
technology becomes available and new needs are addressed. An integrated
solution to the problem of determining the existence of adequate methods
and selecting the most appropriate for an application is critical given
the diversity of regulations, method origins, performance needs and
characteristics, method categories, sources of information, and user
requirements. The implementation of a practical source for method
1-219
-------
information will require overcoming some of the major shortcomings of some
existing systems, such as the limitation in size and performance imposed
by most microcomputer environments, the triviality of the functionality
allowed by the system, and the difficulty of learning and use for the
average user when complex searches and operations become necessary. A
time- and cost-saving approach will be one that makes use of the already
assembled data bases, but the problem of compatibility and consistency of
user interfaces needs to be addressed.
SMART METHOD INDEX CONCEPT
There are several levels of information and assistance that a computerized
system can provide. At the lowest level is the plain data base with
general method information. A second level is that type of system which
provides quantitative information to determine the applicability of a
method to a particular use with consideration of performance, quality
control, and resource requirements as they relate to the project data
quality objectives. A third level of assistance would be provided by a
system that, besides making available the necessary information, guides
the user in making a decision about the suitability of methods.
A computerized method index that is both efficient and effective will
require a comprehensive, high-performance data base system and a powerful,
user-friendly interface. Target users for an environmental monitoring
methods index have diverse background and range from scientists to
concerned citizens and from managers to engineers [6]. Systems with
minimal contents and simple functionality could serve as common
denominators to such a diverse group, but would fail to meet the needs of
EPA investigators and the environmental community. Methods with diverse
origin and application should be included. Many separate data bases of
related methods have been assembled already and some have been
computerized. The integration of some of these previous developments to
a more general index could prove to be a cost-effective approach
consistent with agency objectives.
The need to store and manage a large amount of information suggests the
use of a mini or mainframe computer to implement the system. This
centralized approach also presents the most manageable alternative for
system maintenance, data base updates, access, user support, and security
control. This approach also assures the user that the latest information
is always available without depending on the effectiveness and timeliness
of software distribution.
EPA already has in place central and regional networks for computer
access, which make the centralized approach a realistic option. For wider
access, alternative delivery and access channels could be found through
other Government, scientific, and private electronic information networks.
The integration of several data bases in one system would require an
effort to preserve elements of the structure of each data base, since one
include-all data base with a common structure could be very inefficient.
1-220
-------
This modularization presents a serious problem for data base management
and particularly for user operation of the system. An environmental
analytical methods data base with the described characteristics will need
to be "smart" to provide the user with guidance and help to navigate
through the data. Assistance needs to be given to the user to convey the
information need, to find the right data in a modularized system with
multiple structures, and to produce an output that is both relevant and
understandable.
SYSTEM DESIGN
The Smart Method Index prototype development will test the concept and
implementation approaches, provide a way to better define the system
specifications, and produce tools for expanded development. The system
implementation has two major, conceptually separate but operationally
interdependent, components: the data base driver and the user interface
(front end), as depicted in Figure 1. Depending on the details of the
final implementation, a third component for intercommunication may be
necessary. Scientists at EMSL-LV are building a prototype system in a
mainframe computer to test the performance and suitability of different
data base architectures. The system under development initially will
provide information at the first level of assistance, as previously
described, and will be quickly expanded to a second level of assistance.
ENGLISH
QUERY
REPORT
FRONT
END
DATA BASE
QUERY
REPORT
DATA
BASE
USER
MICROCOMPUTER
MINI/
MAINFRAME
COMPUTER
Figure 1. Remote data base access and smart front end concept.
1-221
-------
The choice of software in the available mini or mainframe computers is
limited. The Statistical Analysis System (SAS®) [18] was selected for use
in preparing the prototype because of its flexibility. This software is
one of the most widely used and best supported data processing tools. It
offers the ability to manipulate multiple data sets and provides built-in
functionality for a large number of operations. The SAS® language is a
high-level programming environment that can be used to build very refined
applications, including high-quality graphics. Prompt-driven or menu-
driven user interfaces can be built with SAS®, or alternatively the
program can run in the background with an external interface. Besides the
modularization of the data base to integrate existing method catalogs, the
use of multiple data sets with diverse structures permits the
implementation of relational data base models [19]. Relational data bases
are a flexible and storage-efficient way to organize interrelated data.
Instead of having one large collection of records with a number of fixed
fields (e.g., METHOD NAME, ANALYTE NAME, ANALYTE SYNONYM1, ANALYTE
SYNONYM2, PRECISION) related information is stored together and several
interrelated files are produced. For example, the method name, analyte
name and performance information could be stored together in a file, while
the analyte name field could be used to access the analyte synonyms stored
in a separate file, utilized only when this information is needed. In the
same way, method information (such as source, status, and description)
could be accessed in a third file through the method name field, if
requested. Figure 2 shows an example of a relational data base structure
for storing and accessing chemical analytical method information. The
relational architecture and operation helps avoid giving the user unwanted
information, streamlines searches, facilitates maintenance, and
accommodates different data base structures.
The need to access possibly different data structures suggested the
exploration of application-independent approaches. Prompts and menus
require the user to know how the data base is built and depend on the data
base structure for anything but the simplest applications. Another option
is the increasingly popular Standard Query Language (SQL) approach, which
provides a common query mechanism throughout several data bases; SAS*
Version 5 does not support this feature, although it is planned for
inclusion in the upcoming version [20] . A "natural language" approach was
selected as more user-oriented because it allows the user to ask questions
to the system rather freely by typing them in plain English [21] .
Appropriate tools for implementation of this user interface are not
available in the mainframe computer. Since microcomputers are now in
widespread use and can serve as terminals for a central computer, Prolog
language [22] running in a microcomputer was selected for the user
interface implementation. Prolog is a programming language widely used
in artificial intelligence projects and very suitable for natural language
application development. The resulting front end should be capable of
translating the English-like questions of the user into the corresponding
SAS® command or SAS* program code to query the data bases for the sought
information [23]. The output of the query will be presented to the user
on the computer screen or optionally printed in hard copy. This dialogue-
based operation could allow for progressive questioning where the user
1-222
-------
METHOD FILE
ANALYTE FILE
METHOD NAME
ANALYTICAL TECHNIQUE
METHOD DESCRIPTION
SOURCE NAME
STATUS - APPROVAL
STATUS - DEVELOPMENT
REFERENCE NAMES
ANALYSIS COST
ANALYSIS TIME
DIFFICULTY
QA/QC REQUIREMENTS
SAFETY CONSIDERATIONS
ANALYTE NAME
ANALYTE SYNONYMS
CAS NUMBER
CHEMICAL CLASSES
REGULATION NAMES
REFERENCE NAMES
HAZARD CATEGORIES
SOURCE FILE
SOURCE NAME
ORGANIZATION
PERFORMANCE FILE
ANALYTE NAME
METHOD NAME
MATRIX
ACCURACY
PRECISION
DETECTION LIMIT
QUANTITATION LIMIT
REMARKS
REFERENCE FILE
REFERENCE NAME
CITATION
AVAILABILITY
REGULATION FILE
REGULATION NAME
SOURCE NAME
REFERENCE NAMES
REGULATION DESCRIPTION
Figure 2. Analytical methods relational data base example.
1-223
-------
starts with general questions and, based on the response obtained, refines
the queries to focus on specific areas of interest. In this context-
sensitive operation mode, new questions could indirectly refer to the
context of previous ones. The microcomputer-based component also takes
care of the communication support over data or telephone lines. The
increasing affordability of high-speed modems and improvement in the
quality of communications help to make this remote approach a practical
one. The smart interface capability will provide a vehicle for the future
integration of an expert system in the Smart Method Index to reach
performance at a third level of assistance.
SUMMARY
A mechanism to assist within the selection of methods for measuring
environmental pollutants has been recognized as a high priority need
consistent with the requirement for adequate, cost-effective acquisition
of environmental information. Several alternative approaches have been
initiated which have varying applicability to current needs, future
requirements, and the universe of potential users. The objective of the
task described here is to provide suggestions and considerations, as well
as a proof-of-concept prototype, for the development of a system that is
flexible and appropriate to meet a variety of long-term objectives of EPA
and the environmental community. Primary considerations include the
development of a standardized user interface, the necessity for data base
flexibility in a rapidly expanding domain, and the requirement for the
development of a standardized, multipurpose data base structure. These
considerations are consistent with the original legislative intent to
provide optimal information with minimal duplication of effort. The
ongoing development of a prototype for the data base and the front end,
using SAS® and Prolog, respectively, will allow the testing, refinement,
and validation of the proposed approach.
ACKNOWLEDGEMENTS
The authors want to acknowledge the valuable contribution of Kelly R.
York, Lockheed Engineering & Sciences Company, during the initial testing
of the concept.
NOTICE
Although the research described in this article has been funded wholly or
in part by the United States Environmental Protection Agency through
contract number 68-03-3249 to Lockheed Engineering & Sciences Company, it
has not been subjected to Agency review and therefore does not necessarily
reflect the views of the Agency and no official endorsement should be
inferred. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
1-224
-------
REFERENCES
1. Fisk, J. F. "The Driving Force: Environmental Legislation,"
Environmental Lab, 1(1), pp. 30-35, March 1989.
2. U.S. Environmental Protection Agency. "EPA Expert Systems
Initiative Update," September 1988.
3. Code of Federal Regulations, Title 40, Part 136, Appendix A, 1988.
4. U.S. Environmental Protection Agency. "Test Methods for Evaluating
Solid Waste," 3rd Edition, SW-846, Office of Solid Waste and
Emergency Response, Washington, B.C., 1986.
5. U.S. Environmental Protection Agency. "Availability, Adequacy, and
Comparability of Testing Procedures for the Analysis of Pollutants
Established Under Section 304(h) of the Federal Water Pollution
Control Act" (EPA/600/9-87/030), Environmental Monitoring Systems
Laboratory, Cincinnati, Ohio, 1988.
6. U.S. Environmental Protection Agency. "Computerized Methods
Information System for Agency-Wide Use: Proposal," Internal
Document, Quality Assurance Management Staff, Washington, D.C.,
March 1989.
7. U.S. Environmental Protection Agency. "The 1986 Industrial
Technology Division List of Analytes," Industrial Technology
Division, 1986.
8. U.S. Environmental Protection Agency. "Index to EPA Test Methods"
(EPA 901/3-88-001), EPA Region I, Boston, Massachusetts, 1988.
9. McCarthy, C. A., L. H. Keith, M. T. Johnston, M. D. Ramminger, and
B. J. Hayes. "Methods for Analysis of Water Pollutants," Radian
Corporation, Austin, Texas, 1986.
10. Clamp, P. Dynamac Corporation, private communication, April 1989.
11. Bethke, A. Research Triangle Institute, private communication, 1988
12. Smith, D. U.S. Environmental Protection Agency. Risk Reduction
Engineering Laboratory, Cincinnati, Ohio, private communication,
January 1989.
13. U.S. Department of Energy. The Environmental Survey Manual,
Appendix D: Analytical Methods Radiological, in press.
14. U.S. Environmental Protection Agency. "Evaluation of a Prototype
Field-Portable X-Ray Fluorescence System for Hazardous Waste
Screening," (EPA/600/X-88/127), Environmental Monitoring Systems
Laboratory, Las Vegas, Nevada, 1988.
1-225
-------
15. Marrin, D. L. and H. B. Kerfoot. "Soil-Gas Surveying Techniques,"
Environ. Sci. and Technol., 22(7), pp. 740-745, 1988.
16. Homsher, M. T., V. A. Ecker, M. H. Bartling, L. D. Woods, R. A.
Olivero, D. W. Bottrell, and J. D. Petty. "Development of a
Protocol for the Assessment of Gas Chromatographic Field Screening
Methods," in Proceedings of the First International Symposium on
Field Screening Methods for Hazardous Waste Site Investigations,
U.S. Environmental Protection Agency, Las Vegas, Nevada, pp. 439-
462, 1988.
17. U.S. Environmental Protection Agency. "Field Screening Methods
Catalog User's Guide" (EPA/540/2-88/005), Office of Emergency and
Remedial Response, Washington, B.C., 1988.
18. "SAS® User's Guide: Basics, Version 5 Edition," SAS Institute Inc.,
Gary. North Carolina, 1985.
19. Hamer, R. M. "Using the SAS System with Relational and Hierarchical
Data," in Proceedings of the Twelfth Annual SAS® Users Group
International Conference, SAS Institute Inc., Gary, North Carolina,
1987.
20. Kent, P. and L. Church, Jr. "SQL and the SAS® System: Version 6
and Beyond," in Proceedings of the Thirteenth Annual SAS* Users
Group International Conference, SAS Institute Inc., Gary, North
Carolina, 1988.
21. Feingenbaum, E. A. and A. Barr. "Understanding Natural Language,"
Chapter 4 in The Handbook of Artificial Intelligence, William
Kaufmann, Inc., Los Altos, California, pp. 223-321, 1981.
22. Clocksin, W. F. and C. S. Mellish. "Programming in PROLOG,"
Springer-Verlag Publishing Company, 1981.
23. Winston, T. W. , M. B. Taylor, R. Leeds. "Natural Language Query
Parsing," AI Expert, 4(2), pp. 50-58, 1989.
1-226
-------
INTERFACING OF AN HP GC-MS (5970) WITH A 1000A COMPUTER
SYSTEM TO A VAX COMPUTER AND DEC LIMS
John T. Bychowski, Mike Demorotski, Deborah Hockman, Anne
0'Donnell, Sunil Srivastava, WMI Environmental Monitoring
Laboratories, Inc., Geneva, Illinois 60134; Dennis Couch,
Interlake Material Handling Division, Lisle, Illinois
60532; Mark Hartwig & Mike Rank, York Laboratories,
Schaumburg, Illinois 60195
ABSTRACT
An integral part of the Environmental Monitoring Laborato-
ries (EML) automation strategy has been the development of
instrument management systems. These systems accomodate
the electronic transfer of analytical resultant data
automatically from the analytical instrumentation data
system to the DEC Laboratory Information Management System
(LIMS) resident on our DEC VAX-based computer network.
This paper addresses the interface developed for transfer-
ring resultant data from a Hewlett Packard (HP) 1000A
computer with RTE-A operating system, via Forest Computer's
VAXLINE file transfer software, to DEC'S LIMS/SM residing
on a VAX clustered network.
INTRODUCTION
The EML volatile and semi-volatile organics departments
utilize eight HP 5970B MSD's with 5890A GC's, linked to
four HP 1000A computers with Aquarius software and RTE-A
operating system, for the analysis of organic constituents
in groundwater samples. The resultant data file generated
for a sample is written as an ASCII file to the CI (command
interpreter) directory on the 100OA computer by the Aquar-
ius software. The file containing the "raw data" from the
analysis is also written to the CI directory. The result-
ant data ASCII file contains information about the method
name, sample name, associated raw data file name, instru-
ment operator, time and date stamp, analyte names, analyte
Chemical Abstract Service (CAS) numbers, and analyte con-
centrations (results). The special HP file name characters
(e.g. ", <, >) are translated into alphanumeric characters
(U, G, L) to be consistent with DEC'S file nomenclature.
The Forest Computer file transfer module of the VAXLINE
software resides on the 1000A computer and is automatically
invoked to transfer both the resultant data and raw data CI
directories via DECNET to identically named directories
within the VMS file structure on the IM LAVC (Instrument
Management Local Area VAX Cluster). The raw data files can
1-227
-------
then be archived from the VAX network and the resultant
data files are available for parsing and transfer to
LIMS/SM.
The VAX network hardware involved in the interface effort
comprise two clustered systems. The CI (computer intercon-
nect) cluster utilizes a VAX 8530 and a VAX 8250 computers
to run LIMS/SM (sample management) and LIMS/CS (communi-
cation sub-system) software; six RA82 disk drives provide
3.6 Gb (gigabytes) of primary and shadowed memory for LIMS
and two more RA82 drives are dedicated to raw data storage.
The IM LAVC runs LIMS/CS only on a cluster comprising a VAX
3600 (boot node) and twin uVAX II' s (satellite nodes) with
two RA 82 disk drives providing 1.2 Gb of memory.
When a sample is logged in to LIMS/SM (on the CI cluster) a
test request is generated, and after certain system parame-
ters are met, the test request or requests are sent via
LIMS/CS to the appropriate instrument management (LTP) data
base resident on the IM LAVC. This test request contains
the method name, sample name, analyte names and CAS numbers
applicable to the sample, and serves as a temporary data-
base template awaiting the test resultant data. A custom
parser routine, developed by EML staff, periodically scans
the specified instrument management directory (on the IM
LAVC) for resultant data files. When the parser routine
finds an appropriate data file it parses the file down to
its pertinent information (method name, sample name, ana-
lyte names, CAS numbers and results), matches it with the
waiting test request template in the IM database, then
transfers it via LIMS/CS to LIMS/SM on the CI cluster.
The original copy of the completed test request on the IM
LAVC is then purged.
If an error is detected during the parsing routine (e.g.
incorrect number of analytes, wrong CAS numbers, sample
name typos, missing analyte results, etc.) an electronic
mail error message is sent to the user's account and a
copy of the faulty data file is shipped to an error dir-
ectory on the IM LAVC for inspection and correction. A
successful data transfer initiates an entry to a log file
on the IM LAVC.
SUMMARY
A three-vendor instrument interface system has been devel-
oped for the purpose of automatically and electronically
transferring GC-MS sample results from an HP 1000A compu-
ter to DEC LIMS residing on a VAX clustered network.
1-228
-------
Forest Computer's VAXLINE software and EML custom parsing
software were utilized in developing an instrument manage-
ment system capable of fast and efficient transmission of
the high data output generated by GC-MS analysis of ground-
water samples for organic constituents of environmental
concern.
1-229
-------
MOBILITY METHODS
-------
MIGRATION OF CHLORINATED PHENOL, DIBENZO-P-DIOXINS, AND DIBENZOFURANS IN SOILS
CONTAMINATED WITH WOOD TREATMENT OIL.
DANNY R. JACKSON, SENIOR SCIENTIST; DEBRA L. BISSON, SCIENTIST; AND DOROTHY A.
STEWART, PROGRAM MANAGER. RADIAN CORPORATION, 8501 MO-PAC BLVD., AUSTIN, TX
78720.
ABSTRACT. Soil contamination has resulted at various wood
preserving sites from accidental surface spillage and subsurface
seepage from unlined surface impoundments. Significant concentra-
tions of polychlorinated dibenzo-p-dioxins (PCDDs) and polychlori-
nated dibenzofurans (PCDFs) have been found in soils at wood
preserving plants which use pentachlorophenol (PCP). This study
focused on wood-treatment oil and contaminated soil obtained from
an abandoned site in Montana.
The objectives of this project were to evaluate laboratory methods
to assess the potential release of PCP, PCDDs, and PCDFs and to
determine partition coefficients for these compounds in a water-
soil-oil system. Extraction test parameters evaluated included
the optimum extraction time and the most appropriate method for
liquid/solid separation. Partition coefficients for the organic
compounds were determined by batch extraction methods for soil-
water, oil-water, soil-oil, and water-oil-soil phases.
The optimum duration of soil-water mixing for batch extractions
was 18 hours. The most appropriate method for separation of soil
and liquid phases for the determination of soil partition coeffi-
cients was filtration with a 0.45 urn membrane filter. Filtration
with 0.45-urn membranes produced extracts that were approximately
equal to column leachate in PCDDs, PCDFs, and PCP concentrations.
Centrifugation was found to be the least desirable solid-liquid
phase separation technique.
Measurable soil-water partition coefficients for PCDDs and PCDFs
were in the range of 104 to 105, indicating that these compounds
are highly partitioned onto the soil matrix. However, PCDDs and
PCDFs were highly partitioned into the oil phase in the water-
soil-oil phase system. PCP was found to be highly leachable from
soil, with partition coefficients ranging from 20 to 50. Par-
titioning by PCP was approximately equal between soil and oil in
the three phase system.
Results of this study suggest that oil-phase migration is the most
likely mechanism for off-site subsurface transport of PCDDs,
PCDFs, and PCP at wood treatment sites. Therefore, the discovery
of a free oil phase in the subsurface environment at a wood-
treatment site where PCP was used should be considered to contain
significant concentrations of PCDDs and PCDFs in addition to PCP.
Removal of the free oil phase from the site may eliminate the
greatest potential for off-site migration of these compounds.
1-231
-------
LEACH TESTING OF STABILIZED CONTAMINATED SOILS
Benjamin J. Mason, Soil Scientist, ETHURA, 13785 N. Applegate RD. Grants
Pass, OR 97527. John J. Barich, Environmental Eng., USEPA, Region 10,
1200 Sixth Ave. Seattle, WA 98101. Gretchen L. Hupp, Sr. Scientist,
UNLV, ERG, 4505 S. Maryland Pkwy, Las Vegas, NV 89154. Kenneth W.
Brown, Botanist, USEPA, EMSL, P.O. Box 93478, Las Vegas, NV 89193-3478.
ABSTRACT
Soil stabilization or fixation was evaluated as a remedy of choice at
three Superfund Sites located in the Pacific Northwest. Fixed soil
materials were subjected to a series of leach test procedures designed
to determine the effectiveness of the fixation methods in reducing the
leachability of the contaminants in the soil.
During the first phase of the testing procedures four vendors were asked
to treat samples of soil taken from the Western Processing Site located
in Kent, Washington. Samples of the fixed wastes were subjected to the
Monofiled Waste Extraction Procedure (MWEP), Toxic Characteristic
Leaching Procedure (TCLP), American Nuclear Society (ANS) Test #16.1,
and the Materials Characterization Center (MCC) Test #1 along with a
suite of biotoxicity tests.
Engineering and chemical testing were also carried out on the materials.
The various engineering tests were designed to evaluate the structural
properties of the fixed soils as well as the effects of fixation on per-
meability.
In the second phase two vendors were asked to conduct bench scale
fixation of soils at two NPL sites. One vendor treated the soils from
the United Chrome NPL Site in Corvallis, Oregon and the other treated
soils from the Tacoma Tar Pits NPL Site in Tacoma, Washington. Mate-
rials from both of these tests were subjected to a series of bench-scale
tests similar to those used in the Western Processing Study. The vendor
working on the United Chrome site also carried out a pilot scale
fixation of the wastes. The wastes obtained on site during the pilot
scale test were subjected to the same tests used during the bench scale
study.
DEFINITIONS
The following terms are used in this paper:
Stabilization is a process that alters the form of the chemical
pollutant or detoxifies it by chemically altering the form or species
of the contaminant. Solidification is a process that accomplishes a
reduction in the leachability of the contaminants by improving the
physical characteristics of the waste, decreases the surface area of the
1-232
-------
waste and encapsulates the contaminant. Fixation is a term that is
often used to encompass both solidification and stabilization.
INTRODUCTION
Region 10 of the U. S. Environmental Protection Agency (USEPA) has
carried out a series of demonstration projects designed to evaluate the
efficacy of soil stabilization or solidification as a remedy at
Superfund or RCRA sites. This paper reports the results of testing
carried out during these demonstrations.
Leach testing carried out under current environmental regulations have
not been designed to address wastes that have been fixed. One of the
primary purposes of fixation is to reduce the surface area of the waste
over which leaching can occur. Grinding of the resulting monolith in
order to carry out leaching tests such as the Toxic Characteristic
Leaching Procedure (TCLP) or Monofilled Waste Extraction Procedure
(MWEP) fails to recognize the primary benefit of solidification.
Region 10 working with the Environmental Monitoring Support Laboratory
(EMSL-LV) in Las Vegas, NV and the Center Hill Solid and Hazardous Waste
Research Facility (Center Hill) in Cincinnati, OH has used several
testing protocols in order to evaluate the effectiveness of the
fixation.
STUDIES
This report reviews leaching results from studies carried out on three
of the demonstration project sites used by Region 10. The Western
Processing NPL Site in Kent, WA was used in 1986 to evaluate the
processes used by four vendors. The materials were subjected to the
Extraction Procedure - Toxicity (EP), the TCLP, the Solid Waste Leaching
Procedure (SWLP) as well as a number of physical tests.
One of the vendors was asked to work with Region 10 in evaluating the
efficacy of their product on contaminated soils obtained from the United
Chrome NPL Site located in Corvallis, OR. This study was carried out in
two phases. The first phase was a bench scale study using screened
materials obtained from an abandoned dry well located on the site. The
second phase was a pilot scale fixation carried out on the site using
materials obtained from the dry well area. The TCLP, the MWEP, the
Materials Characterization Center Test No. 1 (MCC-1) and the American
Nuclear Society's Test No. 16.1 (ANS 16.1) were used to leach the mate-
rials fixed during these studies.
The third site was the Tacoma Tar Pits NPL Site located in Tacoma, WA.
A vendor was asked to provide assistance in carrying out the fixation of
materials obtained from the site. The materials on this site provided
and unusual test for the fixation process used to fix the materials.
Not only were there metal contaminants in the soils but also organic
1-233
-------
chemicals that had leached from an abandoned coal tar disposal area.
The soils were mixed with coal tar and with a material known as "auto
fluff". Auto fluff is a mixture of foam, rubber, plastic and non-
ferrous metal left over from the operation of an automobile shredder.
In order for a fixation process to work on this site, it would have to
be able to handle the auto fluff and the tarry materials. The TCLP and
the ANS 16.1 tests were used to leach the materials fixed in this study.
TESTING PROCEDURES
The testing procedures used in all three of the demonstration studies
were designed to provide an assessment of leaching for purposes of
evaluating the risk that might occur from the use of fixation as a
remedy at Superfund sites. The test protocol used at Western Processing
utilized testing methods that are outlined in various environmental
regulations. Two primary tests were used at Western Processing; the
TCLP and the SWLP. Samples were pulverized for both of these tests;
but, a third set of samples were leached as monoliths. Four vendors
took part in this demonstration study. Metals were the contaminants of
interest at this site.
At United Chrome two additional tests were added. Both of these were
patterned after test protocols used by the nuclear industry. The
Materials Characterization Center uses a test (MCC-1) that measures the
concentration of the saturated leachate. Samples are collected from
monoliths that have been leached for different periods of time in site
water. This represents the maximum concentration that should be
expected in the leachate. The second test used is the American Nuclear
Society test (ANS 16.1) for regenerating or flowing leachate. Monolith
samples are leached for set periods of time then transferred to fresh
leachate. The MWEP was used at United Chrome. This is a new name for
the test known as the SWLP; therefore, this is essentially the same
test.
The ANS 16.1 test is used to calculate a diffusion coefficient De that
is calculated by use of Equation 1.
(AF * u / SQ)2= i * (De * At/n) Eq. i.
A leaching index is often calculated using Equation 2.
Le = 1/n [2 (log 1/De)J Eq. 2.
The United Chrome evaluations were made in both the laboratory at a
bench scale and during a pilot scale study carried out in a scaled down
version of an actual fixation operation. The vendor used what became
1-234
-------
known as the mix of record (MOR) to fix the screened samples used in the
bench study. This same mixture was used during the pilot scale study
with some slight modifications. Cr, Pb and Sr were the contaminants of
interest at this site. Manganese was used to provide an evaluation of
leaching from the soil components.
At the Tacoma Tar Pits Site a decision was made to reduce the number of
tests used in order to be able to test the different materials found at
the site. The TCLP and the ANS 16.1 test were the only leaching tests
used during this study. The TCLP provided an evaluation of the toxicity
of the leachate that might be generated from fixed wastes. The ANS 16.1
represents the type of leachate that could be expected from a sample of
undisturbed fixed waste.
RESULTS AND DISCUSSION
The composition of the raw materials used in the fixation studies are
listed in Tables 1 through 3. The soil materials were prepared and
mixed in the field then transported to the laboratory for testing. The
materials from the Western Processing Site were actual site soils;
whereas, the United Chrome "soil" was actually a mixture of dried
plating sludge and contaminated soil taken from an abandoned dry well
used for disposal of liquid wastes from a chrome plating operation.
The TCLP test was used on all three sites. The leachate concentrations
for this test are presented in Table 4. The four processes used at the
Western Processing Site produced mixed results. The soil at Western
Processing failed the TCLP for Ba and Pb. Three of the vendors produced
a product that passed the TCLP for both of these metals but the fourth
vendor's product failed this test. It is interesting to note that a
number of the products exhibited and increase in the amount of material
leached from the fixed waste.
At United Chrome the Cr concentrations in the bench scale TCLP leachate
dropped adequately after fixing to provide a material that passed the
test for toxicity. The pilot scale study material failed the TCLP and
actually showed an increase in lead. This later observation is believed
to be the result of a materials handling problem encountered during the
pilot study. Nodules containing Fe, Pb, and Cr were found in the dry
well material. These were not broken apart during processing. The
chemicals used to reduce the chromium were unable to penetrate the
nodules and therefore both the chromium and lead were available for
leaching from the pulverized material used for the TCLP.
All of the materials except benzene passed the TCLP in the Tacoma Tar
Pits materials. Fixation of organic materials such as the tar causes a
problem with the Portland Cement mixture used in most fixation
operations. There are additives that can improve the retention of the
organics; but, this was probably not done during this operation. Mixing
soil with the tar appears to greatly improve the retention of the
1-235
-------
benzene. The two metals reported during this study were well within the
regulatory limits.
A comparison can be made between the use of the acid leach used in the
TCLP with a distilled water leach used in the SWLP by comparing the
results from the first leach period for the SWLP with the TCLP. Figure
1 shows this comparison for Cr and Figure 2 for Pb. The distilled water
leach removed less Cr than did the acid leach for all vendors products.
The case with Pb was mixed. Vendor PC-l's product showed a marked
difference between the results of the two leaching media. The results
for the other three products were varied.
During the Western Processing study both pulverized and monolithic
wastes were tested with the SWLP. Figure 3 shows the results for Pb.
The pattern was similar for the other metals although there was some
fluctuation especially with the monolith samples. The similarity
between these two curves for Pb suggests that something more than just
leaching is occurring in these samples.
The extensive sampling carried out at the United Chrome site allowed a
number of comparisons to be made that aid in interpreting the data and
in evaluating the effectiveness of the vendors product. Unconfined
compressive strength (UCS) is often used as a test of the fixation
product. Compressive strength results from the set of the Portland
Cement in various mixtures used for fixing wastes. UCS generally
increases with time as the cement cures. Figure 4 shows the results of
curing on UCS with the United Chrome bench study materials.
As the UCS increases there is also a reduction in the permeability of
the waste form. This indicates that a waste that meets a particular
leaching criterion during a short term test should continue to improve
over a period of, time measured in years. Although the data was not
generated to evaluate the effects of permeability on leaching, a set of
data was developed that can be used to infer the effects of decreasing
permeability. Figure 6 shows the effects of increasing UCS upon the
leaching index (Le). As the leaching index increases the amount of pol-
lutant leached decreases; therefore, materials with a high Le have the
smallest leach rate.
The diffusion coefficient (De) which is used to calculate the Le in
Equation 2 tends to decrease logarithmically with leaching time. This
is the result of the loss of readily leached pollutant from the surface
of the waste form and a slower rate of diffusion from the interior of
the waste form. This effect is seen in Figure 7 where four different
waste samples are examined. Wastes marked C were from a pilot study
cell containing a granular mixture that had been compacted. It did not
contain the same amount of cement as did the material in the H cell.
The bench sample was a small diameter cylinder that appears to behave
similar to the compacted granular material. The similarity between the
replicates from the C cell is really quite good considering the
variability of the materials used in the pilot study.
1-236
-------
Table 5 contains the calculated De for both United Chrome and Tacoma Tar
Pits materials. Results of two auxiliary studies are also included in
this table. Samples that were cured at two different temperatures are
compared. There appears to be little effect of the use of the higher
curing temperature. The only benefit was in a quicker set of the cement
based fixing materials.
United Chrome is located in the Willamette Valley of Oregon. This is an
area that goes through an annual wetting and drying cycle; therefore,
the fixed wastes were subjected to the wet/dry stress test outlined in
SW 846. There appears to be no effect of the wet/dry stressing on the
Le.
The De can be used in Equation 1 to estimate the fraction of the mass
that would be leached from the monolith over time. The average De for
the entire testing period was used to make this estimate for Pb for
three of the soil materials. The Tacoma Tar Pits sample showed more
leaching than did the two United Chrome materials. It is interesting to
note that the pilot scale study appears to show less leaching than did
the bench scale study. This is often seen when processes such as
fixation are up-scaled to a field situation such as the pilot study.
The authors recommend that the ANS 16.1 or one of the slight mod-
ifications to this test that have been proposed (Wiles, 1987; Malone and
Jones, 1982; or Ann, 1988) should be used in order to provide an
estimate of the leachability of the wastes. The TCLP or one of the
other regulatory tests will probably have to be carried out. The fact
that these tests require that the waste be pulverized violates the
primary purpose of solidification — reducing the permeability and the
leaching surface of the waste. A positive note on the use of the TCLP
type tests is the fact that if the waste passes the test then there is a
very good possibility that there will be no problems from the waste in a
monolithic form.
The information contained in this paper is covered in detail in a series
of reports submitted to USEPA (Mason, 1987, Rupp, 1989 and Mason, 1989.
REFERENCES
Ahn, Shin H. 1988. Evaluation of Test Protocols for Stabi-
lization/Solidification Technology Demonstrations.
Malone, Philip G. and Larry Jones. 1982. Guide to the Disposal of
Chemically Stabilized and Solidified Waste. SW-872. USEPA. Washington,
DC. 20460.
Mason, Benjamin J. 1987. Statistical Evaluation of Fixed Soils Data from
the Western Processing Site in Kent, WA. Report produced under P0#
5VV0044-NASA for EMSL-LV.
1-237
-------
Mason, Benjamin J. 1989. Fixation of Soils at the United Chrome NPL
Site: Assessment of Data from Bench and Pilot Scale Studies. Report
produced under P0# 8V-0948-NASA for EMSL-LV.
Rupp, Gretchen. 1989. Bench Scale Fixation Study from the Tacoma Tar
Pits Superfund Site: Final Report. Environmental Research Center.
University of Nevada-Las Vegas. Las Vegas, NV. Report produced under
Cooperative Agreement # 814701 for EMSL-LV.
Wiles, Carlton C. 1987. Investigation of Test Methods for Solidified
Waste Characterization (TMSWC). (Draft report). USEPA. HWERL.
Cincinnati, OH.
1-238
-------
TABLE 1
CHEMICAL CONCENTRATIONS IN WESTERN PROCESSING SOIL SAMPLES
OTA 7 T QT T p
Ba Cd
AVERAGE 74 24
S. D. 8 5
C. V. 11.15% 19.06%
CHEMICAL CONCENTRATIONS
MAT^TDT A I QTAT T CTTT1
DRY WELL AVERAGE'
S. D.
C. V.
BACKGROUND AVERAGE
S. D.
C. V.
CHEMICAL CONCENTRATIONS I
CONTAMINANT
As
Pb
Total Phenols
Benzene
Toluene
Xylenes
Pyrene
Benzo( a > anthracene
Benzo( b ) f luoranthene
Benzo( k ) f luoranthene
Benzo( a )pyrene
I ndeno (1,2, 3-cd ) pyrene
D i benz ( a , h ) anthracene
Total PCB's
CONCENTRATION (mg/kg)
Cr Cu Pb Ni Zn
164 100 1113 29 4819
27 15 226 4 730
16.34% 15.01% 20.26% 14.38% 15.15%
TABLE 2
IN UNITED CHROME SOIL SAMPLES
CONCENTRATION (mg/kg)
Cr Mn Pb Sr
71796 345 28862 19563
1966 27 3037 2573
2.74% 7.69% 10.52% 13.15%
64.3 350.8 11.3 59.5
23.0 9.8 5.3 15.8
35.83% 2.79% 46.87% 26.53%
TABLE 3
N MATERIALS FROM TACOMA TAR PITS
CONCENTS AT I ON < mg / kg )
SOIL: FLUFF
C(""lTT TAD —
oU I Li 1 nK
1:1 3:1
137 - 62 141
2490 - 3080 2120
377 201 584 389
0.002 - <.007 <.007
0.008 - <.007 (.007
0.008 - <.007 <.007
8.3 3200 4.2 7.4
3.4 1200 <1.8 <1.7
2.6 350 2.5 2.8
2.3 510 <1.8 1.8
3.0 740 1.4 2.6
1.3 240 1.1 1.4
0.9 200 0.6 0.8
6.2 198 32 13
1-239
-------
TABLE 4
TCLP LEACHATE CONCENTRATIONS anthracene
SOIL
2
17
418U
-
-
-
12U
12U
12U
12U
12U
12U
12U
TAR
2
1U
420
537
1000
674
4
12U
12U
12U
12U
12U
12U
SOIL: TAR
1 • 1
4
1U
1180
59
540
505
5
12U
12U
12U
12U
12U
12U
SOIL:
1:1
8
23
23
_
_
_
13U
13U
13U
13U
13U
13U
13U
FLUFF
3:1
6
11
15
_
_
62U
62U
62U
62U
62U
62U
62U
U = Undetected at this level.
1-240
-------
TABLE 5
AVERAGE LEACHING INDEX CALCULATED FROM ANS 16.1 DATA
UNITED CHROME
MATERIAL AND/OR TEST
CONTAMINANT CURING WET/DRY CYCLE
25 Deg 60 Deg 4th 12 th 20 bh
Chromium 16.2 15.1 16.3 16.4 15.7
Leac3 17.5 17.5 17.8 17.9 17.9
Manganese 17.0 17.1 17.3 17.3 17.4
Strontium 12.7 12.5 13.0 13.1 13.2
TACOMA TAR PITS
MATERIAL
CONTAMINANT
SOIL TAR SOIL;TAR SOILrFLUFF
1:1 1:1
Arsenic >13.6
Copper
Lead >12.2
Phenols 12.4
Pyrene >12.5
Napthalene >9.2
Benzo(a)pyrene >13.1
PCB's >12.1
8.9
>13.2
10.9
>15.2
8.9
>12.7
10.3
>14.7
12.2
14.6
>13.2
11.7
>12.2
>13.4
1-241
-------
D
2
D
D
FIG. 1= COMPARISON OF TCLP AMD SWLP
-1 .5 -
-i -
—2. 5 ~
PC-1
WESTTRN PRQCESSM3 Cr RESULTS
PC-2
C7"71 AQD LE^CH {TWLP}
VENDORS
[V\] Q LEACH {SWLP].
'•.IT
FIG. 2= COMPARISON OF TCLP AM) SWLP
PROCESSES Pb P:E3JLTS
E.OE-O1
4.OC—C'1
i.OE-01
-2.0E-01 -
-EOE-O1
-S..OE-01
-1 .OE+CO
-1 .4E+OO
-1 .BE+OO
•1 £E+C
-------
O.C-4E.
O.G42. -
O.OiS -
O.Q3& -
•s
z
D
Ul
D
S O.COS
$ O.OJ16 -
O.GiS. '
O.O2. •
Q.fl1 E
FIG. 3= COMPARISON OF SAMPLE TYPES
W ESTERM PROCESaWQ Pb RESULTS
1O
33
D FUL'/EFOZED
.55
TIME (hr=)
TO
lvCir43LITH
FIG. 4 EFFECT OF CURWG TIME ON UCS
1-243
-------
FIG. 5= EFFECT OF UCS ON PERMEAHUTY
J'
E
Ci
~^
n
01
3
tt
111
0.
D
El
-5.4 -
-.5.6 -
-5.7 -
-5. & -
—.5.9 -
-B -
-e.1 -
-B.i -
-E.-i -
-a. 4 -
-E. 5 -
-E..6 -
-E..7 -
-B.9 -
-7 -
-7.1 -
-7.1 -
-7.5 -
—7.4 -
Q-
~~^\
^~x.^
^"•--^
~"---.^
~\^
"X
\
\
\
\
x^
X\
\
\
\
\
^
1 1 1 1 1 i :
3 iOO 4CC E-OO M
UCS (p=i>
FIG. 6= EFFECTS OF UCS ON L*
O-FQME: o- ^ts \ E..I RESLILTS
1E.4-
1 50
1 70
2.1 £•
ii«
I-244
-------
FIG. 7= EFFECT OF UEACHWG TME ON De
UNITEE- a-ROMC AIMS 1 S..1 Cr RESULTS
(i
o
D
3
-11
-1.1
-13 -
-14. -
-15 -
-1 B -
-17 -
-ts-i 1 r
I I I 1 1 1 1 1 1 1 1 i 1 1 1
RG. 8= FRACTION LEACHED FROM MOHOUTH
n
ui
i
D
£
U-
D
D
I-
-.5
-4 -
-4.5 -
LOS TIME (yr)
H- UC BENCH
« UC PILOT
1-245
-------
GEOCHEMICAL BASIS FOR PREDICTING LEACHING OF INORGANIC
CONSTITUENTS FROM COAL-COMBUSTION RESIDUES
Ishwar P. Murarka, Program Manager, Electric Power Research
Institute, 3412 Hillview Avenue, Palo Alto, California 94304;
Dhanpat Rai, Staff Scientist, and Calvin C. Ainsworth, Research
Scientist, Battelle, Pacific Northwest Laboratories, P.O. Box 999,
Richland, Washington 99352
ABSTRACT
Pore waters, leachates, and waste-water extracts associated with
electric utilities' wastes and waste disposal ponds exhibit
extremely variable aqueous elemental concentrations. When ionic
strength and aqueous complexation are accounted for, however, the
activities of free species exhibit behavior with pH similar to
those in equilibrium with known solid phases. By comparing
observed ion activity products (IAP) with calculated lAPs from
thermodynamic data for selected solid phases, solubility-
controlling solids may be indirectly identified or suggested for
various elements. To date, A10HS04, A1(OH)3, BaSO/j, 'CaS04,
Cr(OH)3/(Cr,Fe)(OH)3, CuO, Si02, SrS04, ZnO, and several solid
solution/coprecipitates have been suggested to be the solubility-
controlling solid phases for different elements over a wide range
of waste compositions, waste types, an disposal conditions. These
results indicate that the composition of leachates from coal com-
bustion wastes, with respect to several elements, can be estimated
from fundamental reactions involving solubility-controlling solids
of these elements. The results obtained thus far using this
approach are summarized in this paper.
INTRODUCTION
The Resource Conservation and Recovery Act (RCRA) provides the
statutory authority for the U.S. Environmental Protection Agency
(USEPA) to regulate solid waste disposal. One of the major con-
cerns associated with the disposal of solid wastes on land is the
release into groundwater of chemicals with a potential for causing
environmental degradation and possible harm to humans. For the
purpose of defining appropriate management requirements, several
federal regulations have been promulgated to list, ban, or classify
wastes as hazardous or non-hazardous. One of the four criteria
USEPA uses in classifying solid wastes is the "toxicity character-
istic." The USEPA has developed laboratory tests to evaluate the
Teachability and toxicity of the leachates resulting from land dis-
posal. Since 1980, the USEPA has used the Extraction Procedure
1-246
-------
(EP) for this purpose (USEPA 1982). The USEPA is currently in the
process of finalizing a new (USEPA 1986) Toxicity Characteristic
Leaching Procedure (TCLP) to replace the EP. Both the EP and TCLP
use a prescribed leaching fluid in laboratory extraction of waste
constituents. Both procedures are intended to extract soluble
quantities of chemicals from the waste to simulate a waste mis-
management practice of codisposal of many kinds of wastes into a
municipal landfill.
The electric utility industry annually generates over 80 million
tons of solid wastes as byproducts of coal combustion. About 80%
of these wastes are disposed of in landfills and surface impound-
ments at the power plants where they are generated. These land-
fills and impoundments are distributed across the entire United
States and seldom contain the mixture of wastes that occurs in
municipal landfills. Therefore, utilities need a more appropriate
alternative to using the EP or TCLP methods, because these methods
do not accurately simulate the compositions of infiltrating fluids
and the resulting leachates. Furthermore, accurate information on
leachate composition, quantities, and release duration is essential
for reliably predicting the potential for groundwater pollution.
EPRI research completed to date (Ainsworth and Rai 1987; Rai et al.
1988, 1989; Fruchter et al. 1988) indicates that geochemical reac-
tions between coal combustion solids and infiltrating water could
reliably be used to calculate the inorganic chemical composition of
leachates. In this paper, we describe the geochemical basis for
such calculations.
COAL COMBUSTION WASTES
The coal burned in electric power plants contains inorganic con-
stituents (minerals and other impurities) that are left behind as
fly ash or bottom ash that must be disposed of or beneficially
used. In addition, the scrubbing of flue gas to remove sulfur
results in flue gas desulfurization (FGD) sludges that also require
disposal. The total chemical compositions of these wastes vary a
great deal (Table 1), depending on the coal. However, the types of
compounds initially present in the wastes are expected to have
similarities caused by the high temperatures reached during coal-
burning. For example, kaolinites can be expected to convert to
mullite, most aluminosilicate minerals to glass, carbonates to
oxides, and gypsum to anhydride. Among such compounds, those that
have relatively low solubilities and rapid precipitation/
dissolution kinetics, and that are present in large enough amounts
will control the concentrations of their constituent elements in
leachates.
1-247
-------
Table 1. Concentrations of Selected Elements in
Utility Wastes (from Rai et al. 1987)
Flv Ash Bottom Ash FGD Sludge
Element
Aluminum
Calcium
Iron
Silicon
Sulfur
Arsenic
Barium
Boron
Chromium
Copper
Lead
Molybdenum
Selenium
Strontium
Vanadium
Zinc
When such compounds are present, leachate concentrations are pre-
dictable, provided the identity of the solubility-controlling
solid is known and thermodynamic data for the solubility product
and the associated aqueous complexes for the elements involved are
available. In testing this predictive approach in the laboratory
and in field studies, we have found that many elements (e.g., Al,
Ba, Ca Cr, Cu, Fe, S, Si, Sr, Zn) in coal combustion wastes are
most likely controlled by precipitation/dissolution reactions
(Ainsworth and Rai 1987; Rai et al. 1989; Fruchter et al. 1988; Rai
and Szelmeczka 1989; Rai et al. unpublished data). Other
researchers (Roy and Griffin 1984; Talbot et al. 1978; Henry and
Knapp 1980) have also noted that concentrations of some of these
0.1
0.11
1.0
1.02
0.04
2.3
1
10
3.6
14
3
1.2
0.2
30
12
14
20.85
22.30
27.56
31.78
6.44
6,300
13,800
5,000
900
2,200
2,120
236
134
7,600
1,180
3,500
Ma.ior El
3.05
0.22
0.4
5.10
ements (wt%)
18.5
24.10
20.10
- 31.20
<0.04 7.40
Minor El
0.02
109
1.5
<0.2
3.7
0.4
0.84
0.08
170
12
3.8
ements (uq/q)
- 168
9,360
- 513
5,820
932
1,082
443
14
6,440
537
1,796
0.64
0
0.13
0.27
0.08
0.8
<25
42
1.6
9.69
34.50
- 13.80
- 17.70
- 22.80
53.1
2,280
530
180
6 340
0.25
<4.0
<2
70.8
<50
7.7
290
52.6
162
2,990
261
- 612
1-248
-------
elements (e.g., Al, Ca, S) are solubility controlled. Studies are
now under way to confirm the identities and quantities of the
solubility-controlling solids for these and a number of additional
elements.
Completed and ongoing laboratory and field studies under EPRI
sponsorship are relating the aqueous elemental concentrations in
pore waters, leachates, and waste-water extracts to solubility-
controlling solid phases with well-established thermodynamic
equilibrium reactions. These studies are briefly described in the
following section.
LABORATORY STUDIES TO IDENTIFY GEOCHEMICAL BASIS
A total of more than 90 samples of fly ash, bottom ash, FGD sludge,
and oil ash that covered a range of chemical characteristics were
collected and analyzed by Ainsworth and Rai (1987). Hot-water
extracts were among the analyses performed. The quantities of
several elements in the hot-water extracts were extremely variable
from waste type to waste type and also from sample to sample within
a given type of waste. Geochemical modeling of the measured con-
centrations revealed, however, that the large variabilities in
extract compositions could be accounted for by the effects of pH,
aqueous complexation, ionic strength, and the type of the
solubility-controlling solid phases. For example, the aqueous
concentration of Al in extracts from different wastes and waste
types varied over four orders of magnitude (Figure 1), showing a
recognizable amphoteric behavior. However, the A13+ activities,
which correct the concentrations for the differences in aqueous
complexes and ionic strength, vary as a smooth function of pH
(Figure 2). The similarity of these AP+ activities to those in
equilibrium with known Al solid phases suggests that Al concentra-
tions are controlled by A10HS04 at pH values <6 and by Al(OH)3(am)/
Al(OH)3(c) at pH values >5.5. Comparisons of ion activity products
(IAPs) indicated that aqueous concentrations of several other ele-
ments are also controlled by solubility phenomena: Ca and S by
CaS04«2H20/CaS04, Si by Si02(am) and hydrated CaSi03, Mo by
and Ba and Sr by, perhaps, their $64 compounds. No solubility
controls were postulated for trace metals.
Ainsworth and Rai (1987) and.Rai et al. (1987, 1988) showed that
this mechanistic approach to predicting aqueous elemental
concentrations is applicable to fossil fuel wastes. However,
Ainsworth and Rai's (1987) interpretations of solubility-
controlling solids are based on elemental activities as a function
of either pH or a given ligand in extracts from various samples.
Although high temperatures during burning of coal are expected to
1-249
-------
0
-1
-2
-3
-4
en c
o °
-6
-7
-8
0
- D
a
o
a Fly Ash
o Bottom Ash
A FGD Sludge
o Oil Ash
o
8
D
a
Detection Limit
8
pH
10
12
Figure 1. Variation in total aqueous Al concentration in different
wastes as a function of pH and waste type.
(From Ainsworth and Rai 1987).
produce similar solids in all wastes, the conclusions of Ainsworth
and Rai (1987) cannot be relied upon without information on changes
in activities of a given element as a function of pH or ligand con-
centrations from extracts of subsamples of a given waste. In
addition, further information is needed 1) to confirm the nature
and the amounts of the solubility-controlling solids and 2) to
identify solubility-controlling solids of trace elements.
Therefore, studies are continuing with four fly ashes whose
chemical characteristics vary over a wide range (Table 2) These
studies involve size, density, and magnetic fractionation to
enhance the characterization of solid phases by such techniques as
x;ray diffraction. SubsamPTes of fly ashes are equilibrated both
at different pH values and with and without additions of either
1-250
-------
+
§,
O)
o
-5
-10
-15
-20
-25
-30
-35
-40
Figure 2.
a Fly Ash
o Bottom Ash
^ FGD Sludge
8
pH
10
12
14
Variation in A13+ activity, determined from the hot-
water solution concentration data using MINTEQ, as a
function of pH of coal-derived waste samples. Solid
lines represent A13+ activities in equilibrium with
different Al solid phases, calculated from
thermochemical data at an observed average pS04 of 2.45
for pH <6 waste samples. (From Ainsworth and Rai 1987).
known aqueous concentrations or solid phases of a given element.
The purpose of these studies is to identify solubility-
controlling solids indirectly by comparing the observed lAPs with
the lAPs calculated from either the published thermodynamic data or
the estimates that we developed for those solid and aqueous
species for which the data were either unavailable or inaccurate.
The results of these studies have shown that 1) the predictions of
Ainsworth and Rai (1987) are reasonable, 2) direct techniques for
identifying solid phases, such as X-ray diffraction, are not very
successful even when elaborate preparations are made to concentrate
the elements by means of size and density fractionations,
3) indirect techniques are successful in identifying the solid
phases not only for the major elements but also for trace elements,
1-251
-------
Table 2
Aluminum
Calcium
Iron
Silicon
Total Concentrations of Selected Elements in
the Four Fly Ashes
Sample
102
104
112
Ma.ior Elements (wt%)
131
10.3
1.07
17.7
20.2
12.6
0.91
8.2
19.9
9.3
3.34
13.7
23.6
14.0
0.95
6.5
20.8
Minor Elements (wt%)
Chromium
Copper
Strontium
Vanadium
Zinc
0.024
0.011
0.026
0.040
0.048
0.018
0.021
0.111
0.035
0.029
0.022
0.007
0.029
0.019
0.044
0.016
0.022
0.081
0.032
0.017
pH (1:1 paste)
4.1
6.8
13.8
9.1
4) solid solutions/coprecipitates play an important role for
several constituents [e.g., Ba, Cr(III), CrC^, $64, Sr], and 5) in
most cases, necessary thermodynamic data are not yet available. A
few specific findings from these studies are discussed below. The
compounds CuO and ZnO were found to be the solubility-controlling
solids for Cu and Zn. Amounts of CuO and ZnO available for leach-
ing were also determined. In the four fly ashes, Cr was found to
be present entirely as Cr(III), and aqueous Cr(III) concentrations
were found to be controlled at low pH values by (Fe,Cr)(OH)3 and at
higher pH values by (Fe,Cr)(OH)3/Cr(OH)3(am) (Figure 3). The
(Ba,Sr)(S04,Cr04) compounds play an important role in controlling
aqueous concentrations of Ba, Sr, and Cr04- Studies are currently
under way to ascertain the solubility-controlling solid phases of
other trace elements.
Field Studies to Evaluate Applicability of Mechanistic Basis
Although the laboratory results and data analysis indicated that
for several chemicals we can rely on the mechanistic approach for
1-252
-------
-2
-3
—
O
05
O
-6
-7
-8
• W131 ,7 day
O W131 , ~130day
A W131 plus 10-3-49 M Cr(lll) , 7 day
Detection Limit
I
6
pH
10
Figure 3. Observed aqueous Cr(III) activities in pH-adjusted fly
ash are compared with the Cr(III) activities (solid
lines) calculated from the thermodynamic data (From Rai
and Szelmeczka 1989).
defining leachate compositions from coal combustion wastes, ques-
tions immediately arose as to how these findings would apply to
field conditions and how solubility controls might change with
leaching times. To address these and other questions, two field
leaching studies were completed—one in a fly ash test landfill and
another in an FGD sludge pond. In both cases, the leachates gen-
erated by the respective wastes and the rainfall or sluicing waters
were collected and chemically analyzed for several inorganic con-
stituents. Results from both studies are briefly discussed below.
A fly ash landfill, measuring 100 ft x 100 ft at the base and
60 ft x 60 ft at the top, with a height of 10 ft, was constructed
in 1985 by the Pennsylvania Power and Light Company. Natural
rainfall has infiltrated through the landfill since late 1985 (Rehm
et al. 1987). Leachates were collected from the bottom and from
several depths of the test landfill in late 1987 and the chemical
compositions of the leachates were analyzed. The measured aqueous
1-253
-------
concentrations were examined using an equilibrium geochemical
speciation computer code. The results of the analyses show that
the equilibrium solubility controls identified in the laboratory
(e.g., (Ba,Sr)S04 for Ba and Sr, CaS04'2H20/CaS04 for Ca, and CuO
for Cu) are also applicable to the field-scale leachates. The
results also indicated that some of the constituents show a
leaching trend that has not yet been identified as being solubility
controlled (Fruchter et al. 1988).
A full-scale FGD sludge pond, now inactive, was sampled and
analyzed for chemical composition of leachates (Rai et al. 1989).
The FGD sludge disposal unit operated as a pond from 1976 to 1979.
Since 1979, the standing water has been drained, and the pond now
acts as a landfill for most of the year. Sludge cores were taken
at three points in the FGD sludge disposal unit. An elaborate
laboratory set-up was used to separate leachate (or pore water in
this case) from the solids. The Eh and pH of the aqueous phase
leachates were measured in situ and in the laboratory. Because of
very reducing conditions and the presence of sulfides, aqueous
concentrations of Cd, Cu, Ni, Pb, and Zn were found to be at the
detection limits as a result of the solubility controls imposed by
the sulfide solids of these elements (Rai et al. 1989). Additional
elements that were found to be controlled by solubility phenomena
included Ba, Ca, Cr, Fe, S, Si, and Sr. Among the trace elements,
only B and As were present in measurable quantities. As in the
laboratory and the fly ash landfill studies, the results of this
study bear out the applicability of the mechanistic approach.
MODEL TO PREDICT LEACHATE COMPOSITIONS
Already, the empirical data and the mechanistic formulations
developed thus far for the coal combustion wastes have been used to
develop ,an interim Fossil Fuel Combustion Waste Leaching
(FOWLTV ; code (Hostetler et al. 1988). The FOWL™ code is
structured to predict the composition and quantity of leachates
produced as a function of time by solid-waste disposal facilities
containing coal combustion wastes. The current version of FOWL™
runs on an IBM®v^) PC and is available in executable form through
the Electric Power Software Center; the user's manual is available
from the EPRI Research Reports Center. The code contains a geo-
chemical calculation algorithm and a water-balance calculation
routine. The mechanistically based portion of the code deals with
Al, Ba, Ca, Cr, Mo, S, Si, and Sr, which have been found to be
controlled by solubility-limiting solids. The empirically based
portion of the code deals with As, B, Cd, Cu, Fe, Mg, Na, Ni, Se,
and Zn. Improvements in the fundamental basis are being made as
continuing laboratory experiments and field measurements clarify
the solubility controls for the major and trace elements listed
1-254
-------
above, the multitude of solid phases involved, and the role of
kinetics. Improvements in the computer code will follow
accordingly.
SUMMARY
Field sampling and laboratory extractions have yielded many
valuable insights into the geochemical basis of leaching of coal
combustion solid wastes. It is clear that fundamental chemical
reactions between water (leaching fluid) and the solids in the
disposal units govern the chemical compositions of the leachates.
Precipitation/dissolution reactions at equilibrium with a given
solid phase form the basis for leaching. For several inorganic
constituents, the controlling reactions are now known and the
thermochemical data appear to be adequate. However, for several
other constituents, either the controls are not yet known or
adequate thermochemical data are not available. Unfortunately,
because the amounts of controlling solids involved in the quite
complex matrix of coal combustion wastes may be very small, it is
not possible to directly measure the types and amounts of the solid
phases involved. Therefore, many laboratory experiments and field
measurements must be made to identify and define the solubility
controls, the changes in solid phase assemblages, and the short-
term and long-term leaching characteristics. Research of this
nature is in progress, and we expect to complete this work within
the next three years.
NOTES
(1) FOWL™ is a trademark of Electric Power Research Institute,
Palo Alto, California.
(2) IBM® is a registered trademark of International Business
Machines Corporation, Boca Raton, Florida.
REFERENCES
Ainsworth, C. C., and D. Rai. 1987. Chemical Characterization of
Fossil Fuel Combustion Wastes. EA-5321. Electric Power Research
Institute, Palo Alto, California.
Fruchter, J. S., D. Rai, J. M. Zachara, and R. L. Schmidt. 1988.
Leachate Chemistry at the Montour Fly Ash Test Cell. EA-5922.
Electric Power Research Institute, Palo Alto, California.
Henry, W. M., and K. T. Knapp. 1980. "Compound Forms of Fossil
Fuel Fly Ash Emissions." Environ. Sci. Technol. 14(4): 450-456.
1-255
-------
Hosteller, C. S., R. L. Erikson, and D. Rai. 1988. User's Manual
for the Fossil Fuel Waste Leaching (FOWL ) Code: An Interim
Model. EA-5742-CCM. Electric Power Research Institute, Palo Alto,
California.
Rai, Dhanpat, and R. W. Szelmeczka. 1989. "Aqueous Behavior of
Chromium in Coal Fly Ash." Submitted to J. Environ. Qual.
Rai, Dhanpat, C. C. Ainsworth, L. E. Eary, S. V. Mattigod, and
D. R. Jackson. 1987. Inorganic and Organic Constituents in Fossil
Fuel Combustion Residues, Volume 1: A Critical Review. EA-5176,
Vol. 1. Electric Power Research Institute, Palo Alto, California.
Rai, Dhanpat, J. M. Zachara, L. E. Eary, C. C. Ainsworth,
J. E. Amonette, C. E. Cowan, R. W. Szeleczka, C. T. Resch,
R. L. Schmidt, D. C. Girvin, and S. C. Smith. 1988. Chromium
Reactions in Geologic Materials. EA-5741. Electric Power Research
Institute, Palo Alto, California.
Rai, Dhanpat, J. M. Zachara, D. A. Moore, K. M. McFadden, and
C. T. Resch. 1989. Field Investigation of a Flue Gas Desulfuriza-
tion (FGD) Sludge Disposal Site. EA-5923. Electric Power Research
Institute, Palo Alto, California.
Rehm, B. W., B. J. Christel, T. R. Stolzenburg, and D. G. Nichols.
1987. Field Evaluation of Instruments for the Measurement of
Unsaturated Hydraulic Properties of Fly Ash. EA-5011. Electric
Power Research Institute, Palo Alto, California.
Roy, W. R., and R. A. Griffin. 1984. "Illinois Basin Coal Fly
Ashes. 2 Equilibria Relationships and Qualitative Modeling of Ash-
Water Reactions." Environ. Sci. Techno!. 18(10):739-742.
Talbot, R. W., M. A. Anderson, and A. W. Andren. 1978. "Qualita-
tive Model of Heterogeneous Equilibria in a Fly Ash Pond."
Environ. Sci. Technol. 12(9):1056-1062.
U.S. Environmental Protection Agency (USEPA). 1982. Test Methods
for Evaluating Solid Waste: Physical/Chemical Methods. SW-846.
U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency Response, Washington, D. C.
U.S. Environmental Protection Agency (USEPA). 1986. "Hazardous
Waste Management System; Identification and Listing of Hazardous
Waste; Notification Requirements; Reportable Quantity Adjustments;
Proposed Rule." Federal_Reglster 51(114):21648-21693. June 13,
1986.
1-256
-------
EVALUATION OF METHOD 1311 FOR DETERMINING THE RELEASE POTENTIAL OF OILY WASTES
R. S. TRUESDALE, RESEARCH TRIANGLE INSTITUTE, P.O. BOX 12194, RTP, NC 27709;
J. 0. PEIRCE, DEPT. OF CIVIL AND ENVIRONMENTAL ENGINEERING, DUKE UNIVERSITY,
DURHAM, NC 27706; AND G. A. HANSEN, U.S. EPA/OSW, WASHINGTON, D.C.
ABSTRACT. Method 1311, or the Toxicity Characteristic Leaching Procedure, was
designed to model release of contaminants in leachate from a reasonable worst-
case waste mismanagement scenario: codisposal of 5 percent industrial waste
with 95 percent municipal refuse in an unlined sanitary landfill. Previous
work indicated that Method 1311, as currently proposed, is not suitable for
certain oily wastes (e.g., slop oil emulsion, creosote sludge, waste motor
oil) because filter clogging during the initial filtration step can over-
estimate a waste's percent solids, thereby underestimating the release
potential of the primary waste leachate.
Several filtration step modifications were investigated to solve this problem.
For five difficult-to-filter wastes, percent solids results for each
modification were compared with the percent of waste immobile in soil columns
designed to replicate waste release from a reasonable worst-case sanitary
landfill. These experiments indicated that a porous media sintered stainless
steel filter most accurately estimates waste mobility. The steel filter also
showed good reproducibility, with percent relative standard deviation (%RSD)
of percent solids results ranging from 9 to 28 percent for three oily wastes.
Further evaluation of the steel filters demonstrated that the filters are
adequately precise and accurate for use in a modified Method 1311. The %RSD
of the modified procedure ranged from 2 to 16 percent for five volatile
organic waste constituents and from 6 to 32 percent for three inorganic
constituents. Precision and accuracy of the modified procedure also was
determined for semi volatile organic constituents. Accuracy was determined by
comparing the final analyte concentrations determined by the modified method
with concentrations in the leachate from the soil column experiments. For
these determinations, spikes of deuterated polycyclic aromatic hydrocarbons
were added to each waste aliquot prior to the experiments. This provided
convenient marker compounds at known concentrations to follow through the
experiments and to compare in the final extracts and column leachates.
(Complete precision and accuracy data for semi volatile organics are not
available at the time of submission of this abstract).
The results of this study indicate that the modified Method 1311 is a
significant improvement over the existing method for many difficult-to-filter
wastes. The steel filter can filter wastes that clog the glass fiber filter
used in Method 1311, thereby improving the accuracy of this step in estimating
the release of toxic waste constituents associated with the liquid portion of
an oily waste. A full collaborative study of the modified method is necessary
prior to proposal as a standard RCRA test method. Our experience indicates
that careful waste selection, characterization, and aliquoting is critical to
the success of such a study.
1-257
-------
THE LIQUID RELEASE TEST
Carrie Kingsbury, P.E. and Paula Hoffman, Research Triangle Institute, P.O.
Box 12194, Research Triangle Park, NC 27709; Barry Lesnik, U.S. EPA/OSW,
OS-331, Washington, D.C. 20460
Abstract
Under Section 3004 of the Hazardous and Solid Waste Amendments of 1984,
Congress directed the U.S. EPA to promulgate regulations to minimize the
presence of free liquids in containerized hazardous wastes to be disposed in
landfills. The regulations should specifically prohibit landfill disposal
of wastes absorbed in materials that release liquids when compressed, as
might occur in a landfill. As part of the response to this directive, EPA
engaged the Research Triangle Institute (RTI) to develop a suitable test to
determine if liquid is released when an absorbed waste is subjected to a 50
psi load. The objective established by EPA required a test capable of
detecting liquid release from an absorbed waste when the liquid loading
exceeds the saturated concentration by no more than 10 percent.
The result of this method development effort is a Liquid Release Test (LRT)
protocol that is easily performed in the laboratory or field. The LRT
requires a device capable of applying 50 psi continuously to the top of a
confined cylindrical sample. The device consists of a sample holder, the
pressure application device, filter papers to detect released liquid,
supporting screens (to separate the sample from the release detection
filters), and a spacing grid to prevent wicking. The LRT is an attribute
test (i.e., the result is either "release detected" or "no release
detected"), rather than a continuous variable measurement test.
The LRT development entailed extensive testing to evaluate different
equipment designs; release detection techniques; and the effect of test
duration, temperature, and sample size on performance. The LRT was
evaluated in a collaborative study in November 1988 to determine
interlaboratory variability of the test using the device developed by the
Associated Design and Manufacturing Company. The collaborative study also
provided an opportunity for other manufacturers to demonstrate equivalency
of their devices to detect liquid release.
1-258
-------
RESIDUAL FUEL OIL AS POTENTIAL SOURCE OF
GROUNDWATER CONTAMINATION
BEHNAM DAVANIf Organic Department Manager, Analytical
Division, Hall-Kimbrell, 4820 W. 15th. St., Lawrence, Kansas
66049; BILL SANDERS, Associate Chemist, Midwest Research
Institute, 425 Volker Blvd., Kansas City, Missouri 64110 and
GREG JUNGCLAUS, Manager, Peiser Laboratories, Stauffer
Chemical Company, 8410 Manchester, Houston, Texas 77012.
ABSTRACT
As part of the Environmental Protection Agency's (EPA's)
continuing effort to develop the basis for listing certain
refinery wastes as hazardous, residual fuel oil tank bottoms
were characterized for semivolatile organic compounds
including selected polycyclic aromatic hydrocarbons (PAHs).
No. 6 fuel oil samples were collected from four regions in
the United States. Soxhlet extraction with methylene
chloride, followed by fractionation on alumina, were used to
isolate the PAHs in fuel oil samples. The sample extracts
were analyzed by high resolution gas chromatography and
gas chromatography/mass spectrometry. Total selected PAH
concentrations in the fuel oil samples ranged from 71 to
6,560 ug/g. Higher concentrations of other aromatic
compounds including alkylated PAHs and sulfur-containing
aromatic hydrocarbons were also detected in the samples.
Additionally, laboratory-scale soils columns (2 cm ID and
approximately 10 cm long) were used to model the transport
of oily waste in the subsurface and potential migration into
the groundwater. In one set of the experiments, the oil was
introduced on the top of the soil using a disposable
minipipet and was then allowed to disperse freely. In the
second set of experiments, the soil columns were subjected
to a small pressure using pure nitrogen gas. All the studies
were conducted at room temperature and the soil columns were
saturated with water prior to the addition of the fuel oil.
Several oil mobility studies were performed with fine and
coarse soil. The fine soils initially showed no visible
mobility for the fuel oil sample during the 17-day
experimental period. However, additional experiments using
oil/soil ratios of 1:1 and 1:2 (30 g of soil) showed
movement of the fuel oil through the entire length of the
fine soil column (10 cm). The oil moved through the coarse
sandy soil column and also eluted from the coarse column
when the same oil/soil ratios as above were used. The fuel
oil passed through the columns with little adsorption
1-259
-------
(in a matter of minutes or hours) when the columns were
pressurized with nitrogen gas.
In a related study, several Toxicity Characteristic Leaching
Procedure, "TCLP" experiments * ' were conducted to assess
the feasibility of using TCLP for this type of oily waste
and to compare the TCLP results with the corresponding soil
column studies. In light of the results obtained in this set
of experiments, the applicability of TCLP to this type of
oily waste must be cautioned. Furthermore, due to dissimilar
processes and experimental conditions involved in the TCLP
and column studies, it was concluded that comparison of the
two procedures may not be appropriate.
INTRODUCTION
A large volume of crude oil is refined throughout the U.S.
to produce a variety of petroleum-based fuels. For example,
over 60 billion gallons of crude oil were refined by the
petroleum refinery industry in 1984 alone, resulting in the
annual production of approximately 4 billion gallons of
residual fuel oil (2) . Some of the other major products
include motor gasoline, distillate fuel oil (diesel), and
jet fuel.
These petroleum fuels are commonly contained in underground
storage tanks before usage. Leakage of oil from the storage
tanks and potential for groundwater contamination have
recently become a matter of national concern (3,4).
According to one study (5) , 35% of such underground storage
tank systems were estimated to be leaking at the average
rate of 0.32 gal/h.
The Hazardous and Solid Amendment to the Resource
Conservation and Recovery Act (RCRA) require that the EPA
characterize, and regulate if necessary, the wastes
generated by petroleum refineries.
In order to assist the EPA in developing a data base for
listing certain refinery wastes as hazardous, residual fuel
oil tank bottoms were characterized for semivolatile organic
compounds including selected PAHs. Furthermore, the extent
of movement of the fuel oil through soil and potential for
groundwater contamination were investigated by conducting
laboratory-scale column experiments.
A secondary objective of this research was to evaluate the
feasibility of using TCLP to test the leachability of No. 6
fuel oil. Although TCLP is designed and presently is used
1-260
-------
to model the mobility of contaminants in aqueous and
solid wastes (6), the applicability of the procedure to this
type of oily matrix has not been fully investigated.
EXPERIMENTAL
Instrumentation
A varian Model 3700 gas chromatograph was equipped with
flame ionization detector (FID), manual splitless injector
and DB-5, 30-m, 0.25-mm ID fused silica capillary column (J
& W Scientific, Inc., Folsom, California). Conditions for GC
analyses of all the sample extracts were identical and were
as follows: initial temperature, 40°C; initial hold-time, 4
min; final temperature, 280 C; final hold-time, 10 min;
program rate, 10 C/min;Q injector temperature, 280 C;
detector temperature, 300 C; carrier gas helium at 30
mL/min; time for splitless injection, 1 min; chf^t speed,
2.56 cm/min; and attenuation range, from 16 X 10- to 2.56
X 10- amps. A Finnigan Model 5100 gas chromatograph/mass
spectrometer GC/MS) was equipped with manual splitless
injection port, and 30-m DB-5 fused silica capillary column.
Chromatographic conditions were identical for scanning GC/MS
and GC/FID analyses. Mass spectrometry conditions for
scanning analyses were the following: low mass, 35 amu; high
mass, 500 amu; scan speed, 1 scan/s; and electron multiplier
voltage, 1500 V. Sample extract volumes in GC and GC/MS
analyses were 1 uL delivered using a 10-uL syringe (Hamilton
Company).
Extraction and Treatment of Samples
All the fuel oil samples were extracted, treated and
analyzed using the sample procedures except as noted. A
1.0-g aliquot of each of the four fuel oil samples from
different sources was diluted to 10.0 mL using a 1:1 mixture
of hexane/methylene chloride. A 1.0-mL portion of the
diluted solution was spiked with 200 uL of base/neutral
surrogate PAH standard, 500 ug/mL, and adsorbed onto 3.0 g
of neutral alumina (EM Science, West Germany) activated at
150°C overnight. The samples were allowed to air-dry for 1
hour and then were transferred quantitatively to the top of
2 cm ID column which already contained 7 g of neutral
alumina, and 1 g of sodium sulfate packed on the top. The
samples were subsequently eluted with the following
Chromatographic grade solvents: fraction 1, total of 20 mL
of hexane; fraction 2, 50 mL of benzene. The extracts were
then concentrated to approximately 4-10 mL by K-D technique
and further adjusted to an exact 10 mL (for Kansas City
1-261
-------
fuel) and 4 mL for other samples with a gentle stream of dry
nitrogen gas. Similar procedures have successfully been
applied for the separation of PAHs in other high boiling
petroleum distillate (7), synthetic fuel (8), and wastes
from natural gas production (9, 10). The condensed extracts
were then screened by GC/FID. The resulting chromatograms
showed a major removal of the aliphatic hydrocarbon
interferences in the hexane fraction and a complex mixture
of aromatics in the benzene fraction.
A 1 mL aliquot of each of the benzene extracts including one
method blank was analyzed by GC/MS for selected(target) PAHs
as well as approximately 25 major non-target semivolatile
compounds excluding aliphatic hydrocarbons. The analysis was
conducted according to EPA Method 8270 (11), which involves
initial tuning of GC/MS using decafluorotriphenyl phosphine
(DFTP) and use of internal standards, calibration standards,
surrogate standards, system performance check compounds and
calibration check compounds. Calibration standards were
analyzed at six concentrations including 1.0, 2.0, 5.0,
10.0, 50.0 and 100.0 ug/mL under identical conditions with
those for the sample extracts to establish a working curve.
These calibration standards contained target PAHs, surrogate
standards, and calibration check compounds as per Method
8270.
RESULTS AND DISCUSSION
Characterization Of Fuel Oil samples
The residual fuel oil sample was too complex a mixture for
simple solvent extraction and direct analysis, although a
high resolution capillary column was used. However, a rapid
one-step alumina column cleanup procedure was effective in
removing most of the background and hydrocarbon
interferences.
The quality assurance data for surrogate recoveries in the
benzene extracts including the method blank were very good.
The average percent recoveries for 2-fluorobenzene,
D-iQ-pyrene, and D -p-terphenyl were 92.9%, 108%, and 113%,
respectively. D -nitrobenzene surrogate was not detected in
any of the extracts since this compound is frequently
retained by the alumina cleanup column.
A complex mixture of PAHs and other aromatic compounds were
observed for all the samples. However, the fuel oil sample
from Kansas City had the highest abundance of the aromatics
among the samples. A 1:10 dilution of this sample extract
1-262
-------
had to be used as compared to 1:4 dilution for other oil
samples. The identification of these target PAHs was based
on a relative retention time with respect to
corresponding internal standards, mass spectra of the
standards under the same analysis conditions, and also
favorable match to mass spectra in the EPA/NIH data base.
The results for quantitation of target PAHs in these samples
are given in Table 1. Total target PAH concentrations in the
fuel oil samples ranged from 71 ug/g to 6,560 ug/g.
The fuel oil samples including one method blank were also
characterized for some 20 major semivolatile compounds
(SMVs). The majority of these compounds were alkylated PAHs
with the highest concentrations found in the Kansas City
fuel oil sample. These concentrations were less by almost
one order of magnitude in other oil samples (Chevron). The
differences in concentrations between Kansas City fuel oil
and Chevron samples might be due to different origins or
different processes involved in the production of the No. 6
fuel oil. It is important to note that all fuel oil samples
except the one from Kansas City had undergone cracking
processes to provide a wide variety of product mixes and
levels. This was confirmed by the presence of the highest
abundance of large molecular weight aromatic hydrocarbons in
the fuel oil collected from Kansas City. On the other hand,
the Chevron samples showed the presence of low molecular
weight aromatic hydrocarbons. It is believed that the large
aromatic hydrocarbons were decomposed to lighter
hydrocarbons during various cracking processes.
An interesting observation which seems to support the
different origin and/or processes for the fuel oil samples,
was the presence of sulfur-containing aromatic compounds
such as benz- and dibenzothiophene in all Chevron fuel oil
samples. No such class of compounds was detected in fuel oil
samples from Kansas City.
Column Experiments
Several glass mini columns containing fine and coarse sandy
soils were tested to monitor the bulk movement of the fuel
oil. In the first set of experiments, approximately 30 g of
soil was packed uniformly in glass mini columns (2 cm ID)
and about 5 g of the fuel oil sample was introduced on the
top of the soil column using a disposable minipipet. The
fuel oil was then allowed to disperse freely. Two different
sizes of soil, fine (5 urn) and coarse (4Q-ROC, 1.5
mm or smaller) crushed silica, obtained from
Berkely, West Virginia were used in this study. All the
conducted at room temperature (23 C - 25 C), and the soil
1-263
-------
columns were saturated with water prior to the addition of
the fuel oil. The fine soils showed no visible mobility for
the fuel oil sample during the 17-day experimental period.
The coarse sand showed rather low resistance to the oil
transport.
The porosity of the column packing material was estimated
from the volume of water used to saturate the soil column.
Fine-textured soils tend to have greater porosity than the
coarse-textured soils (12,13). Similarly, values for the
porosity in our column studies decreased in the order fine
(Sum, Berkeley) intermediate coarse (1mm, 2Q-ROC) > coarse
(1.5 mm, 4 Q-ROC).
Additional column experiments containing oil/soil ratios of
1:1 and 1:2 (30 g of soil in each experiment) were also
conducted. The objective of this study was to compare the
movement of fuel oil through the column with the previous
studies in which a smaller oil/soil ratio was used (5 g of
oil to 30 g of soil) . The column dimensions and experimental
conditions were kept identical to the previous set of
experiments.
These results are plotted as depth profiles in Figures 1 and
2 (vertical movement of the oil from the top of the column
vs. time) . The weight of the packed soil was kept constant
(30 g) in all the experiments, while the amounts of the oil
were varied as 5, 15 and 30 g. In general, similar trends
were observed for both coarse and fine soil, and the
transport of oil through the column decreased in the
following order: 30 g > 15 g > 5 g.
The oil eluted from the coarse soil columns when higher
oil/soil ratios were used (1:1 and 1:2). The amount of oil
eluted versus time during this study are shown in Figure 3.
Little or no mobility was observed with 5 g of oil using
fine Berkeley soil. These results are consistent with the
fact that a minimum amount of oil (usually greater than the
pore volume of the soil) is needed to observe any
significant movement. As expected, coarse soil showed more
mobility for oil than the fine soil.
Application Of TCLP To Fuel Oil And Comparison Of The
Results With Soil Column Studies
Three TCLP experiments using 30 g of Kansas City fuel oil
alone and a 1:1 mixture of fuel oil and soil (fine and
coarse Berkeley, 30 g each) were conducted. The first
experiment, containing the fuel oil sample only, was
performed to assess the feasibility of using TCLP for this
1-264
-------
type of oily waste. It was suspected that the oily waste
might not filter even during application of pressure.
According to TCLP protocol, "When the pressurizing gas
begins to move through the filter, or when the liquid flow
has ceased at 50 psi (i.e., filtration does not result in
any additional filtrate within a 2-min period), filtration
is stopped. When this procedure was strictly followed, the
amount of filtrate for the fuel oil sample was 1.56 g.,
while no filtrates were obtained using the fuel oil/soil
mixtures. The pressures at which the filtrate was measured,
following TCLP guidelines as outlined above, were 20, 10 and
50 psi for the fuel oil, and mixture of oil/coarse and fine
soil, respectively-
An experiment was then conducted to attempt to increase the
pressure to a maximum of 50 psi during the TCLP experiment
for each sample matrix. The maximum pressures that were able
to be attained for fuel oil and oil/coarse soil mixture were
30 and 35 psi, respectively. In these two experiments, the
N gas immediately started to move through the filter, as
was noticed by a decrease in the applied pressure. However,
under these increased pressures, the amount of the filtrates
obtained for oil and for the oil/coarse soil mixture were
11.5 and 10.5 g, respectively.
The corresponding soil column studies using 1:1 mixture of
fuel oil and soil (30 g each) were also completed. The oil
was allowed to disperse freely, and the transport of the oil
through the column, as well as the amount of the oil eluted
from the column, were monitored with time. The total amount
of oil that eluted from the coarse soil column was 24 g over
a 23-day period. Although the oil moved through the fine
soil and eventually reached the bottom of the column (8.6
cm), no oil eluted from this column during the experimental
period. These results along with the TCLP data are
summarized in Table 2.
In light of the dissimilar processes and experimental
conditions involved in the TCLP and the column studies, the
discrepancies in the results are not surprising. One might
argue that there is a 50% correlation between the two
results, since no amount of filtrate or eluted oil was
obtained during TCLP or the corresponding column study using
the fine soil. However, this comparison cannot differentiate
between "no movement" and "maximum movement" through the
soil, when no elution of the oil from the column occurs. In
fact, the fuel oil moved through the entire length of the
fine soil column in our column studies. There are other
important differences between the two methods which would
1-265
-------
make their comparison more difficult. Specifically, some of
these differences are as follows:
1. "TCLP" and "soil mobility study" are two different
concepts, with two distinct mechanisms. The former is
a separation method based on "high pressure
filtration" while the latter is a transport process
due to "diffusion, accelerated by gravity." While
these two phenomena might be related, there does not
seem to be a simple or linear correlation between
them.
2. The comparison of the two methods does not take into
account the fundamental factor of time in transport
phenomenon. The amount of oil eluted from the coarse
soil column ranged from 1.2 g after approximately 1
day to a maximum of 24 g over a 23-day experimental
period. It is not clear which one of these amounts
should be compared to the corresponding TCLP
experiment. Furthermore, according to TCLP protocol,
the filtration is stopped if it does not result in
any additional filtrate within a 2-min period.
However, the equivalent stopping period for the soil
column study has not been defined.
3. Another parameter, which would substantially affect
the results, is the surface area of soil in contact
with the fuel oil. While the amounts of fuel/soil in
both TCLP and column studies were identical, the
diameters of the filtration device (in TCLP) and the
soil columns were 14.2 and 2 cm, respectively.
Therefore, the extent of contact of oil with the
surface of soil in TCLP is more than the
corresponding column study during the same time
period. However, there are other factors which would
make the comparison of the two methods even more
complex. There is a different experimental period for
TCLP (several minutes) as compared to that for the
soil mobility study (several days). This would
account for the larger amount of oil eluted during
the soil column studies than the amount of filtrate
obtained in TCLP, as was confirmed by our results.
More importantly, the available soil adsorption sites
in the two methods are not identical. The oil was
mixed with the soil to yield a homogenous mixture
prior to TCLP, while the oil was introduced on the
top of the soil for the column studies.
4. The moisture content of the soil is another important
factor, which affects the flow of oil through the soil
and thus contributes to the discrepancies in the two
1-266
-------
results. The infiltration rate of a waste through the
soil is highest when the soil is dry. However, this
rate decreases with increased moisture and approaches
a steady-state value as the soil remains saturated
and porous (12,14). In our column studies, the soil
was completely saturated with water while it was used
as received in the TCLP experiment.
SUMMARY
It may not be feasible to compare TCLP and soil mobility
results due to the intrinsic differences in the mechanisms
of the two methods, as well as so many different
experimental conditions. Additionally, in the TCLP protocol,
the applicability of TCLP to oily wastes has been cautioned.
This caution has to be taken seriously in light of the very
small amount of the oily filtrate produced in TCLP and the
fact that the nitrogen gas almost immediately started to
move through the filter at low pressure (- 20 psi) .
Moreover, all the attempts to increase the pressure to a
maximum of 50 psi failed when using the fuel oil sample. In
each attempt, the increase in applied pressure was offset by
reduction of the pressure due to passage of the nitrogen gas
through the filter in the TCLP filtration device.
AKNOWLEDGEMENT
This study was supported by U.S. Environmental Protection
Agency under EPA Contract No. 68-01-7287. Helpful
discussions with Mr. Ben Smith, Project Officer, at
Characterization and Assessment Division, Office of Solid
Waste, EPA, Washington, D.C.; Dr. Clarence Haile, and Dr.
Andres Romeu at Environmental Chemistry Department, MRI, are
gratefully acknowledged. This research was conducted at MRI
and the GC/MS work was performed by Mrs. Audrey Zoog.
REFERENCES
1. Federal Register, Vol. 51, No. 216, November 7, 1986.
2. Adopted from U.S. Department of Commerce and U.S.
Department of Energy/Energy Information Administration
(DOE/EIA) data in, "Petroleum Refinery Industry:
Structure and Financial Characteristics," by develop-
ment Planning and Research Associates, Inc.
(June 1986).
3. "Development of a Tank Test Method for a National
Survey of Underground Storage Tank,: U.S. EPA Office
of Toxic Substance, EPA-560/5-86-014 (May 1986).
1-267
-------
4. "Proposed Regulations for Underground Storage Tanks:
What's in the Pipeline?" U.S. EPA Office of
Underground Storage Tanks, Publication No. 26A (April
1987) .
5. "Underground Motor Fuel Storage Tanks: A National
Survey,: Volume 1 Technical Report, U.S. EPA Office
of Pesticides and Toxic Substances, EPA 560/5-86-013
(May 1980).
6. Federal Register, Volume 51, No. 9 (January 1986).
7. Hirsch, D.E., R. L. Hopkins, H. J. Coleman, F. O.
Cotton, and J. Thompson, Anal. Chem., 44, 915 (1972).
8. Later, D. W. , M. L. Lee, K. D. Bartle, R. C. King, and
D. L. Vissilaros, Anal. Chem., 53,. 1612-1620 (1981).
9. Eiceman, G. A., B. Davani, and J. Ingram, Envir. Sci.
Technol., 20, 508-514 (1986).
10. B. Davani, K. Lindley, and G. A. Eiceman, Intern. J.
Environ. Anal. Chem., 25, 299-311 (1986).
11. "EPA Test Methods for Evaluating Solid Wastes," Office
of Solid Waste and Emergency Response, Washington,
DC, SW-846, 3rd. edition (September 1986).
12. Eiceman, G. A., J. T. McConnon, M. Zaman, C. Shuey,
and D. Earp, Intern. J. Environ. Anal. Chem., 24,
143-162 (1986).
13. Bower, H., Groundwater Hydrogeology, McGraw-Hill Book
Company, New York, p. 21 (1978).
14. Fryberger, J. S., Groundwater, 15, 155 (1975).
1-268
-------
TABLE 1
TARGET PAH CONCENTRATIONS IN NO. 6 FUEL OIL SAMPLES INCLUDING THE METHOD BLANK
Analytes
Sample
#6560
(From
(ug/g)
Sample
KC) #5630
(ug/g)
Sample
#6060
(ug/g)
Sample
#6963 Method
(ug/g) Blank
Naphthalene
Phenanthrene
Anthracene
Benz(a)
anthracene
Chrysene
Benz(b)+(k)
fluoranthene
Benzo(a)pyrene
Indeno(1,2f3-cd)
pyrene
147
450
TR (49.7)
1,520
3,090
TR (12.5)
TR (38.4)
ND
NO
TR (29.2)
TR
40
ND
TR
80
(2.09)
.0
(29.3)
.1
TR (7.33)
TR (20.6)
ND
NO
TR (35.6)
ND
ND
ND
ND
ND
436
101
ND
ND
ND
ND
ND
ND
ND
ND
ND compound not detected at quantisation limit
TR compound detected, but at a level less than quantisation limit
Quantitation limit for KC fuel oil was 100 ug/g; for others including the method blank was 40 ug/g.
I-269
-------
Figure 1
Plots of Fuel Oil Transport Through
Course (4Q-ROC) Berkeley Soil
O 5g Fuel Oil
A 15g fuel Oil
Q 30g Fuel Oil
Elution of Oil
from the Column
O—OOOOOO
15
20 25
I-270
-------
g
I 5
o
Ǥ
O-O
-o—oooooo
O 5g Fuel Oil
A 15g fuel Oil
D 30g Fuel Oil
Elution of Oil
from the Column
10
Figure 2
_ Plots of Fuel Oil Transport Through
Course (5|im) Berkeley Soil
3 4
Time (days)
10 15 20 25
1-271
-------
2 -
10
19
3- 12
LU'g
6«
— c 14
o»
i£ ,„
•= 18
20
22
24
26
28
30
_L
10 12 14 16
Time (days)
18 20 22 24
Figure 3
Breakthrough Curves for Two Amounts of the
Fuel Oil Using Coarse (4Q-ROC) Berkeley Soil
I-272
-------
TABLE 2.
COMPARISON OF THE AMOUNTS OF OIL ELUTED DURING
COLUMN STUDIES WITH THE FILTRATE DURING TCLP
Type
of
Matrix
Amount of
oil eluted
from the
Column (g)
Wt. of
Filtrate
in TCLP
Wt. of
Filtrate
under Max.
pressure .
attained
Oil/coarse
Berkeley Soil
(1:1 mixture,
30 g each)
24.0
Zero
10.5
Oil/fine
Berkeley Soil
(1:1 mixture,
30 g each)
Zero
Zero
Zero
Measured at pressures of 10 and 50 psi for mixtures of
oil/coarse and fine soil, respectively.
The maximum pressures attained for the
matrices were 35 and 50 psi, respectively-
two above
1-273
-------
LEACHABILITY OF CHEMICALS FROM HAZARDOUS WASTE LAND
TREATMENT SITE SOILS
David C. Erickson, Laura Rogers, and Raymond C. Loehr, Environmental and Water
Resources Engineering Program of the Civil Engineering Department, The University of
Texas at Austin, ECJ 8.6, Austin, Texas 78712.
ABSTRACT
Chemicals in soil can be subjected to environmental factors which may influence their
mobility. Climatic stress can play a significant role in altering soil structure and metal and
organic chemical retaining properties. Local site conditions also may result in soils with
low pH which can influence chemical mobility. To assess mobility of chemicals, the
Toxicity Characteristic Leaching Procedure (TCLP) has been proposed and is being used.
The objectives of this research are to: (a) determine the leachability of specific metals and
organics from soils at hazardous waste land treatment sites using the TCLP, (b) evaluate
the mobility of hazardous constituents as a function of soil depth for these sites, and (c)
determine the effects of weathering cycles on the leachability of these constituents.
This paper presents the results of work in which soils from several hazardous waste land
treatment sites were initially characterized for selected metals and organics and then their
mobility determined using the TCLP. The soils then were subjected to repetitive
freeze/thaw and wet/dry cycles, and retested using the TCLP to determine whether
significant changes in leaching were caused by the weathering action.
Metals analyses for both the weathered and non-weathered samples indicated that, of the
six metals tested, only zinc exceeded background levels consistently. No clear differences
were detected when metal concentrations in TCLP extracts from weathered and non-
weathered samples were compared. Organic compounds were not detected in either of the
two sets of extracts.
INTRODUCTION
The 1984 amendments to RCRA (Resource Conservation and Recovery Act) prohibit the
disposal of hazardous wastes on land beyond specified dates unless the United States
Environmental Protection Agency (EPA) determines on a case-by-case basis that such a
method is protective of human health and the environment. Land treatment is a land
disposal method affected by the amendments. Land treatment is an engineering process in
which a waste is incorporated into the surface soil layer. The native microbial population
aided by chemical and physical soil processes degrade and immobilize the applied wastes.
Land treatment of hazardous materials is permitted if it can be demonstrated that there will
be no migration of the hazardous constituents for as long as the wastes remain hazardous.
The fate of the waste remaining at the site after a land treatment site is closed will depend
upon local site conditions and post-closure management. It is possible that the wastes will
continue to degrade on-site or remain immobilized and pose no threat to the environment.
A second possibility is that the wastes will migrate into the unsaturated and saturated zones
below the treatment zone, thereby posing a threat to the groundwater. Factors affecting the
mobility of chemicals remaining during post-closure at land treatment facilities include local
soil characteristics, weather, and the type of wastes present. For sites having high levels of
relatively impermeable clay and scant rainfall, the chance of chemical leaching out will be
1-274
-------
lower than at sites with sandy soil and having abundant precipitation.
This paper presents the results of a study in which the potential mobility of metals and
organics in soils from land treatment sites was investigated. This work was supported
through EPA Cooperative Agreement CR-814490. The objectives of this research were to:
(a) determine the leachability of specified metals and organics from soils at hazardous waste
land treatment sites using the TCLP, (b) evaluate the mobility of hazardous constituents as
a function of soil depth and site characteristics, and (c) determine the effects of weathering
cycles on the leachability of these constituents.
MATERIALS AND METHODS
For this study, soils were collected and characterized for chemical content. They then were
used in leaching and weathering tests. A flow chart detailing the tests is shown in Figure 1
and details of the laboratory procedures follow.
FIGURE 1
OUTLINE OF TEST PROCEDURES
Initial Characterization of Core Samples
Leaching of Weathering of
Core Samples Core Samples
I
Freeze/Thaw Cycles
TCLP Wet/Dry Cycles
TCLP
.1
Analyze Leachate for Analyze Leachate for
Metals and Organics Metals and Organics
Site Selection and Sample Collection
Soils used in this study were sampled from three existing land treatment sites
having certain characteristics. These sites had been in operation for at least ten years and
had received oil refinery or wood-preserving waste. The sites also were accessible for
1-275
-------
sampling and records of past waste application practices were available. The sites sampled
were:
• oil refinery site in Washington state
• oil refinery site in Oklahoma
* wood-preserving site in Montana
At the individual sites, soil core samples were collected with increasing depth to evaluate
the concentration of constituents as a function of depth. Core samples were taken from
sides of test pits, dug by a backhoe at each site, at 0-6, 6-12, 12-24, 24-36, 36-54, and 54-
72 inch depths below the surface. At each of the six depths, duplicate samples were
collected and placed in separate three liter glass jars capped with teflon-lined screw-top lids.
The jars then were placed in fiberboard containers and shipped to the Environmental and
Water Resources Engineering Laboratory at The University of Texas (EWRE-UT), Austin,
Texas, for the leaching and weathering studies.
Initial Soil Characterization
The core samples were analyzed for total concentrations of organics and metals to aid in
interpreting the results from the leaching tests. Soils found to have high concentrations of
certain chemicals could be expected to show recoverable levels of the same chemicals in the
TCLP extracts. These soils analyses also provided background information for
determining the extent of in-situ migration of chemicals by observing the depth to which the
chemicals were found. The specific analyses performed and the analysis method used are
shown in Table 1.
TABLE 1
TESTS AND METHODS FOR INITIAL SOIL SAMPLE
CHARACTERIZATION
Test
Soil pH
Freon Extractables
Polynuclear Aromatic Hydrocarbons
Total Metals
Microtox
SW 846 Method
(U.S. EPA, 1986)(D
9045
9071
8310
3050
Beckman Microtox Method
1-276
-------
The noted analyses were selected because these categories included the principal hazardous
constituents expected to be in the soil-residue matrix. The pH of the soil was of interest for
determining the correct leaching fluid for subsequent leaching tests and to determine
whether past waste applications had produced either alkaline or acidic conditions which
could influence the migration of chemicals in the soils. Freon extractable analyses provided
a simple means of estimating the nonvolatile hydrocarbon content in the test soils.
The specific metals tested for were: cadmium (Cd), chromium (Cr), copper (Cu), lead
(Pb), nickel (Ni), and zinc (Zn). The polynuclear aromatic hydrocarbons (PAH's) were
the sixteen compounds listed in Table 2. Microtox is a microbial analysis useful for
determining the relative toxicity of the soils analyzed. The toxicity of the soils was
measured to screen for toxic compounds not specifically tested for which might have been
present in the samples.
TABLE 2
POLYNUCLEAR AROMATIC HYDROCARBONS TESTED
Naphthalene
Acenaphthene
Acenaphthylene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Dibenzo(a,h)anthracene
Benzo(g,h,i)perylene
Indeno(l ,2,3-c,d)pyrene
Soil Leaching Experiments
Soil from each of the six levels sampled was extracted using the TCLP^). The basic
procedure is described below.
Twenty-five grams of soil passing through a No. 9 sieve was placed in a Zero Head
Extractor (ZHE). Five hundred mL of Extraction Fluid (TCLP fluid #1) were added to the
vessel and the apparatus was sealed. For each run, four extractors were set up, placed in a
rotary tumbler and rotated at approximately 30 rpm for eighteen hours.
After the eighteen-hour extraction was completed, the vessels were removed from the
tumbler and the contents of each were pressure filtered through an acid-rinsed glass fiber
filter. The individual filtrates were collected in glass bottles and refrigerated pending
analysis.
The TCLP leachates were analyzed for the noted six metals following method 3010 (SW
846, U.S. EPA, 1986), using a Perkin-Elmer 303 flame atomic absorption spectrometer.
The analysis of the same metals in the soils and leachates permitted a comparison of the
concentrations of the metals in the two matrices.
The leachates also were surveyed for the presence of chromatographable organic
1-277
-------
compounds. One hundred mL of the leachate was extracted in methylene chloride (SW 846
Method 3510). The extracts were analyzed on a Hewlett Packard 5890 gas chromatograph.
It was not known what organics might be present in the extracts, so selected TCLP samples
were spiked with known amounts of specific organics to provide a means of evaluating the
relative abundance of organics present in the extracts. The added compounds were 2,4,6-
trichlorophenol, diphenylamine, 2,4-dinitrotoluene, and ortho-cresol. From the GC
analysis, the extracts which showed the presence of organics above background were
tested using gas chromatography mass spectroscopy (GC/MS). Any peaks which appeared
were tentatively identified by library searches of stored reference spectra.
Weathering Procedures
The methods used for the simulated weathering cycles were adapted from ASTM methods
for tests of soil-cement mixtures^3). The ANSI/ASTM D569 and ANSI/ASTM D559
methods were used for the freeze/thaw and wet/dry weathering experiments respectively.
The procedures assume the use of samples in the form of monoliths and for this study 50
grams of loose soil was used. Consequently, modification of the methods was made in
which the samples were "packaged" in a double layer of nylon. This served to contain the
soil and still permit good exposure to the wet/dry and freeze/thaw cycles.
Samples from each of the six depths collected at each of the sites were subjected to the
weathering cycles. Briefly, the procedure for the freeze/thaw cycles began by saturating
the samples with distilled/deionized water (DDW) over a seven-day period. The soils then
were chilled to -10° F for 24 hours. At the end of this period, the samples were warmed to
73° F at 100 percent relative humidity for 23 hours. This two-day period of freezing and
thawing constituted one cycle of the test. The cycle was repeated for a total of 11 cycles.
Samples subjected to the freeze/thaw procedure were weathered further using the wet/dry
procedure in which the soils first were saturated with DDW. They then were placed in a
160° F zero humidity room for 42 hours. The cycle was repeated for a total of 11 cycles.
Leaching of Weathered Samples
The samples which underwent the weathering cycles were leached using the TCLP. The
procedure described in the Soil Leaching Experiments section was followed.
RESULTS AND DISCUSSION
A large number of samples was analyzed for several different chemicals. All of the results
could not be presented here, so, as an alternative, selected data representative of the overall
trends observed are shown. The complete results will be available in 1990 as an EPA
project report.
Soil Core Characterization
The soil characterization provided information regarding the physical characteristics of the
soils as well as the concentration of organic compounds and metals. The pH of the core
samples from each of the four sites was between 6.0 and 8.1. The Microtox assays
determined also that the soils were nontoxic to the test microorganism. These results
indicated that there were no unusual conditions which might inhibit the natural degradation
1-278
-------
of organics in the soil, or facilitate movement of contaminants through the soils, such as a
low pH. The amount of organic waste applied to the four sites varied from site to site and
the freon extractable data provided an estimate of nonvolatile organics remaining in the soils
on the dates sampled. These data indicated that the highest levels of organics remained near
the surface zone of incorporation. Example freon extractable data for the Washington and
Oklahoma sites are shown in Table 3.
TABLE 3
FREON EXTRACTABLE CONCENTRATIONS IN SOIL (mg/kg)
Depth (inches)
-- Washington
0-6
6-12
12-24
24-36
36-54
54-72
Freon
Extractables
61,700
56,600
8,350
< 1,000
1,700
< 1,000
Depth (inches)
-- Oklahoma
0-6
6-12
12-24
24-36
36-54
54-72
Freon
Extractables
5,500
< 1,000
< 1,000
< 1,000
< 1,000
< 1,000
The freon extractable data in Table 3 show the concentrations of such organics decreased
quickly with increasing depth. The concentrations of PAH compounds also decreased as
the depth increased. An example of the PAH data is shown in Table 4. Similar trends in
the Montana and Oklahoma data were observed (data not shown).
TABLE 4
PAH CONCENTRATIONS IN WASHINGTON SITE SOIL (mg/kg)
Depth (inches)
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Dibenzo(a,h)anthracene
Benzo(g,h,i)perylene
Indeno(l,2,3-c,d)pyrene
0-6
<5
<10
<5
8
5
<0.5
<0.5
98
144
<0.5
220
310
66
15
12-24
<5
<10
<5
6
3
2
<0.5
60
32
<0.5
62
47
8
6
24-36
<5
<10
<5
<1
2
<0.5
3.4
2
<1
<0.5
<0.5
<1
<1
<0.5
36-54 Detection Limit
<5
<10
<5
<1
<0.5
<0.5
<0.5
0.6
<1
<0.5
<0.5
<1
<1
<0.5
5
10
5
1
0.5
0.5
0.5
0.5
1
0.5
0.5
1
1
0.5
1-279
-------
For metals, the total concentrations of cadmium, chromium, copper, nickel, lead, and zinc
also were determined as a function of depth. The analyses indicated that the metals were
confined primarily to the top 0-12 inches of soil. Data from the Washington site are shown
in Table 5.
TABLE 5
METAL CONCENTRATIONS IN WASHINGTON SITE SOIL (mg/kg)
Depth Cadmium Chromium Copper
(inches)
Nickel
Zinc Lead
0-6 <1C
6-12 <1C
12-24 <1C
24-36 <1C
36-54 <1C
54-72 <1(
1 320
) 340
) 110
) 160
) 140
) 80
100
130
30
28
27
29
94
96
45
70
63
61
230
250
87
62
65
61
130
230
27
6
6
6
Soil Leaching and Soil Weathering
The TCLP procedure was developed as a means of estimating the potential for wastes
codisposed with municipal refuse to leach inorganic and organic chemicals. In this study,
it was used as a standardized approach to simulate the leaching of chemicals from the land
treatment soils. As described in the MATERIALS AND METHODS section, core samples
and weathered core samples were leached using the test.
Organic compounds at concentrations above detection limits were not recovered from any
of the samples tested using either gas chromatography or gas chromatography mass
spectrometry. Many of the organics present in these soils were high molecular weight
compounds with low solubilities in water. It is not surprising that they were not recovered
from the aqueous leaching solution.
Analysis of the six metals showed that each of the metals except cadmium were found in
varying amounts in the TCLP extracts from each of the sites. The TCLP test method was
published with threshold limits for classified hazardous wastes. None of the extracts
contained metal concentrations in excess of the threshold limits.
Weathering had no appreciable effect on the concentration of metals in the TCLP extracts
from the top soil layers. Where metals were present below the 12-inch layer,
concentrations appeared slightly lower in the TCLP extracts from the weathered samples.
This may have resulted from the weathering having more of an effect on the core samples
than on the surface samples which had already been exposed to natural weathering.
Examples of the TCLP metal analysis results for non-weathered and weathered samples
from the Washington site are presented in Tables 6 and 7, respectively. Also shown in
Table 6 are the TCLP regulatory limits for cadmium, chromium, and lead. Limits for
1-280
-------
copper, nickel, and zinc concentrations in TCLP leachates have not been published.
TABLE 6
METAL CONCENTRATIONS IN WASHINGTON SITE TCLP EXTRACTS
Non-Weathered Samples (mg/L)
Depth Cadmium
0-6 <0.1
6-12 <0.1
12-24 <0.1
24-36 <0.1
36-54 <0.1
54-72 <0.1
Threshold Limit 1.0
Chromium
0.25
<0.1
<0.1
0.3
<0.1
<0.1
5.0
Copper Nickel
<0.1 <0.1
<0.1 0.7
<0. 1 <0. 1
<0. 1 <0. 1
<0. 1 0.2
* *
Zinc
0.2
0.2
0.8
0.3
0.2
*
Lead
<0.1
<0.1
<0.1
<(U
5.0
*Regulatory levels for these elements have not been published.
TABLE 7
METAL CONCENTRATIONS IN WASHINGTON SITE TCLP EXTRACTS
Weathered Samples (mg/L)
Depth (Inches) Cadmium Chromium Copper Nickel
0-6 <0.1 <0.1 <0.1 <0.1
6-12 <0.1 <0.1 <0.1 <0.1
12-24 <0.1 <0.1 <0.1 <0.1
24-36 <0.1 <0.1 <0.1 <0.1
36-54 <0.1 <0.1 <0.1 <0.1
54-72 <0.1 <0.1 <0.1 <0.1
Zinc
0.1
0.2
0.4
0.1
0.1
0.1
Lead
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
SUMMARY AND CONCLUSIONS
This paper presented results from a study in which the mobility of contaminants in soils
was investigated. The wastes which had been applied to the plots sampled consisted of oil
refinery and wood-preserving wastes, and had been present for at least ten years. Soil
samples were collected at the sites from depths down to six feet below the surface. The
PAH compounds, freon extractables, and metals were confined primarily to the upper 12
inches of the soil at the sites.
Soils from the sites were extracted using the TCLP. In addition, soils weathered using
1-281
-------
repetitive freeze/thaw and wet/dry cycles also were extracted with the TCLP. Results
showed that organics were not extracted from either the weathered or non-weathered
samples. The concentrations of metals tested in the leachates from the weathered and non-
weathered soils were below the regulatory levels designated by EPA for the TCLP.
ACKNOWLEDGEMENTS
Support for this research was provided wholly or in part by the United States
Environmental Protection Agency under Cooperative Agreement CR-814490. This paper
has not been subject to EPA review and therefore does not necessarily reflect views of
EPA. No endorsement by EPA should be inferred. The authors wish to thank Ms. Lynn
G. Sanders for her assistance in preparation of the manuscript and Mr. Scott G. Ruling and
Mr. John E. Matthews of EPA for their support and assistance.
REFERENCES
1. U.S. Environmental Protection Agency (1986). Evaluating Solid Waste (Third
Edition). USEPA/SW-846. Washington, DC.
2. Federal Register 57:114. Friday, June 13, 1986.
3. American Society for Testing and Materials (1984). Annual Book of ASTM
Standards: Part 19. Soil and Rock, Building Stones. American Society
for Testing Materials. Philadelphia, PA.
1-282
-------
EVALUATION OF LEACHABILITY OF RADIUM CONTAMINATED SOIL
Thomas F. McNevin, Ph.D.
New Jersey Department of Environmental Protection
Division of Hazardous Site Mitigation
Trenton, New Jersey 08625
ABSTRACT. Composite samples of radium contaminated soil were batch
extracted at pH 5 and pH 3 in a modification of ASTM Method D3987-81 to
evaluate the relative leachability of radium-226 from soil which had been
blended with a non-contaminated soil, in order to achieve approximate
background concentrations of radium in the soil mix. Comparisons were made
between contaminated soil, diluent soil and the resulting amended mix.
Analysis for radium was conducted by deemanation of radon-222. Data from the
extractions were used to evaluate a bench-scale demonstration of soil
blending, which was judged to be a technically feasible disposition
alternative for soil contaminated with relatively low levels of radium.
Soils with radium concentrations of approximately 60 pCi/gm were amended with
soil with a mean radium concentration of 0.7 pCi/gm to yield mixtures with a
mean activity of 2.6 pCi/gm. Extraction yielded mean radium releases of 10.4
pCi/1, 1.2 pCi/1 and 3.2 pCi/1 for the contaminated, diluent, and amended
soils, respectively. Comparison of the contaminated soil extraction value of
radium with the radium concentration detected in monitoring wells placed
within the area of contaminated soil indicated a tendency for the extraction
tests to overestimate the concentration of radium which would actually be
seen in groundwater associated with the soil being evaluated. The mean
amended soil radium extraction value of 3.2 pCi/1, as well as its
corresponding 95% upper bound of 4.1 pCi/1, was then used in a simple
analytical groundwater model which incorporated only transport and
dispersion. A maximum value of 1.5 pCi/1 radium was predicted at a point 500
feet from a hypothetical emplacement site of the amended soil. This value,
which was derived using several worst case assumptions, was well below all
concentration limits deemed acceptable by present regulatory standards.
INTRODUCTION
In recent years, much attention has been focused on the disposition of
contaminated soils. As landfill space has continued to become increasingly
scarce, and correspondingly expensive, emphasis has shifted to treatment
alternatives. Many approaches, such as soil washing and bioremediation have
been extensively studied (1). Dilution of contaminated soil with clean fill
has generally not been pursued. In the case of synthetic lipophilic organics
which can be reconcentrated in the biosphere, this is a sensible
proscription. The same caveat however need not apply when the contamination
present is from naturally occurring inorganic contaminants which with
appropriate treatment can be made to approximate background concentrations.
Land spreading of soils with low levels of radioactivity to approximate
natural background radiation levels has been cited by EPA as a possible
treatment option, while land spreading of radium sludge from drinking water
treatment system has been carried out in Illinois since 1984 (2).
1-283
-------
In 1986, following the loss of access to a preplanned disposal site, the
State of New Jersey was faced with the need to get rid of approximately 4,100
yd of soil which was contaminated with radium-226 at a mean value of
approximately 60 pCi/gram. To facilitate in-State disposal of this material,
a bench-scale demonstration was undertaken to evaluate the technical
feasibility of blending this material to a pseudo-background target
concentration of <3 pCi/gm. Technically, such amended product material,
which would be less than the EPA soil standard of 5 pCi/gm in the first 15
cm, and 15 pCi/g in any subsequent 15 cm increment (40 CFR 192.12a), would
constitute clean fill and would be potentially available for unrestricted use.
The contaminated soil originated from a pilot project in which soil was
excavated from under and around a number of single-family residences that had
been constructed over contaminated fill, thus leading to elevated levels of
radon-222 and progeny in indoor air, as well as elevated gamma levels both
indoors and out. The houses chosen were taken to be representative of a much
larger group of contaminated structures. Removal of the contaminant source
from this subgroup was evaluated in 1985 as a potential remedial alternative
for general implementation. As noted, loss of the originally planned
disposal site produced a need to consider other alternatives.
SAMPLE PREPARATION
Ten sub-samples were gathered from drummed soils representing each of the
three contaminated neighborhoods (Virginia-Franklin, Carteret, Lorraine) and
from the proposed diluent soil site. The sub-samples were then mixed to
yield a large representative composite sample for each location. In the case
of the contaminated samples, mixing was monitored via gamma spectroscopy to
insure that the resulting composites were in the 55-75 pCi/g range, in
correspondence with the estimated weighted site-wide average Ra-226 activity
of the contaminated soils. Each contaminated composite was then amended with
the diluent composite in a cement-type mixer, at a ratio so as to produce an
amended soil of £3 pCi/g Ra-226. Ratios of 30:1, 35:1, and 40:1 were
utilized.
Five 100-gm replicates were drawn from each contaminated composite and each
amended mix, and batch extracted for 48 hours in accord with ASTM D 3987-81
Standard Test Method for Shake Extraction of Solid Waste with Water (3). In
a modification to the standard method, 400 ml solutions of pH 3 (60/40
H2SO^/HN03) and pH 5 (0.1N sodium acetate) were shaken with the soil
replicates for 48 hours. The extracts were paper filtered to remove gross
solids and then passed through a 0.45 urn filter. Three 40 ml aliquots were
drawn from each filtered replicate, placed in tubes, flame-sealed, and set
aside for radon in-growth. Radon was de-emanated (degassed) from the
solutions containing radium and quantitated in a counting cell. A fourth
aliquot was taken and utilized as a matrix spike. Spiked blanks and reagent
blanks were also deemanated in conjunction with each group of composite
replicates at each pH. Parent radium concentration was then determined from
the radon value (4, 5).
1-284
-------
FINDINGS
A. Leachability Testing
Extraction results along with soil radium concentrations are presented in
Table 1. Statistical comparisons of data sets are given in Table 2. All
references to statistical significance refer to the 5% significance level.
Examination of these Tables along with Tables 3a and 3b yields a number of
observations.
1. In all cases, contaminated soil leachate grand means were significantly
greater than those of the amended or diluent soil.
2. For a given soil no significant differences were seen between the
leachate means at pH 3 and 5, with the exception of the contaminated
Lorraine composite (Table 2).
3. The mean values of amended soil leachates ranged from 1.3 pCi/1 to 4.8
pCi/1 (Table 1), and averaged 2.7 pCi/1 (Table 3). All of these values
are below the EPA Primary Drinking Water Standard for radium-226 of 5
pCi/1.
4. The mean values of contaminated soil leachate replicates ranged from 7.4
to 13 pCi/1 (Table 1), with a mean leachate value of 10.6 pCi/1 (Table
3).
5. Under these same experimental conditions, the diluent soil yielded a mean
leachate value of 1.2 pCi/1 (Table 3).
6. The mean percentage radium extracted from the contaminated composited
soils was 0.065%. This is 10 times less than the average percent
extracted from the amended and diluent soils (0.6%, cf. Table 3).
Discussion
The observation noted in item 1 above, proportionality between the amount of
radium in solution and the amount in the soil is consistent with observations
in the literature (6). Implicit in the leachate concentrations observed
however is that strict solubility concerns are not the controlling influence
on radium mobility in these soils. Simple solubility alone would allow
concentrations of aqueous radium to be present at many orders of magnitude
beyond those encountered here. For example, in the presence of equimolar
concentrations of sulfate, radium could be soluble at 7.4 x 10 M
(K RaSO, = 5.5 x 10 )(4). This solubility potential would allow
quantitative extraction to occur at the soil concentrations encountered in
the present work. That this is not seen is clear evidence that additional
mechanisms are at work to limit the aqueous concentration of radium.
Substantial affinity for ion-exchange, adsorption, and co-precipitation
reactions have been noted for radium (5, 7, 8).
The lack of differentiation in radium leachate concentrations with pH for a
given soil indicates that under the conditions studied, with a volume/mass
(V/M) ratio of 4, the soil buffer complex dominates the reaction and levels
the potential effects of extractants of differing pH values. As the amending
1-285
-------
soil displays effervescence, and therefore contains free carbonate, in-situ
buffer capacity is effectively infinite with respect to a dilute extractant.
This same leveling effect would function in the environment as well, i.e.,
infiltrating water would be rapidly buffered to the pH of the soil. The
anomalous behavior of contaminated Lorraine is indicative of differing
leachabilities due to sample heterogeneity rather than a true pH effect, as
all amended samples displayed post-extraction pH values of 7.4+0.2.
To a substantial extent, the results obtained from a leaching test, are a
function of how that test was designed and conducted. In the present case a
low V/M ratio allowed the soil buffer to control the pH of the leaching
process, which occurred during 48 hours of batch agitation. Altering V/M or
agitation time sufficiently would have produced different results, as would
have sequential extraction with fresh leachant, as would have the execution
of a column or field lysimeter study. Variations in sample preparation
leachant composition, temperature, and manner of leachate separation may also
produce varying results (9).
Agitated tests, such as the one here employed, provide a type of acceleration
and thus rapidly achieve a degree of steady-state behavior which is
indicative of the contributions of relatively short term equilibria
phenomena. The chemical properties of the system independent of short term
kinetic limitations may thus be evaluated. Assessment of the ultimate impact
of long term release requires either a field study in real time, or more
practically, more intensive acceleration, e.g., elevated temperature,
stronger leachant, sequential extraction, particle size reduction. Implicit
in the conduction of an accelerated test is that the actual leaching
mechanisms which occur under natural conditions are not altered by the
acceleration process. As this is not necessarily true, caution should be
exercised in the interpretation of results.
When conducting a leaching experiment, it is vital to keep in mind the
ultimate goal of the experiment. Very often, that goal is the prediction of
the behavior of the studied material in the environment. Often however, it
is merely the comparison of the leaching behavior of a substance relative to
another substance, or to a standard. Most tests however presume mimicry of
the environment to some degree. Even regulatory tests such as EP Toxicity
(40 CFR 261) or the newer TCLP (FR 5J., (114), 21648, June 13, 1986), which
contain standards to which the behavior of tested materials is compared,
incorporate within their protocols assumptions indicative of certain
environmental conditions, i.e., leaching by organic acid in conjunction with
presumed co-disposal of the analyzed material with municipal solid waste.
The standardization of this procedure, which may or may not mimic actual
environmental behavior, does implicitly allow the comparison of the various
materials tested relative to one another. When modelling a particular
phenomenon however, e.g., long term releases, the requirements of data
quality and appropriateness are correspondingly greater as the goal is an
accurate prediction of actual environmental behavior. Unfortunately, there
exists no single standardized leach test methodology for producing high
quality enviromimetic data. Cote'', et. al. , have recently compiled a detailed
review of current procedures (9).
1-286
-------
In the present work, the experimental conditions do not necessarily mimic
actual environmental behavior. They do however allow for comparison of
relative behavior of the different materials, which have been tested under
the same conditions. Amended soil leachate means were significantly lower
than their corresponding contaminated soil means (Table 1). These values are
presented as overall means by soil classification in Tables 3a and 3b. The
pH 3 and pH 5 leachate data have been combined. Radium concentrations in the
amended soil averaged 3.0 times greater than that from the diluent composites
(Table 3b). Amended soil leachate means, averaged 2.3 times greater than
those of the leachate from the diluent soil. These ratios are essentially
equivalent. The diluent soil, which is present within the amended soil to
great excess, provided a volumetric solid phase dilution which effectively
dominated the leaching character of the amended soils. The diluted
contaminated soil proportionately released approximately the same amount of
radium as did the diluent soil. Mean percentage radium extracted was rather
consistent between the amended and diluent soils and averaged 0.6%.
Contaminated soils yielded 10-fold less radium averaging 0.065% extracted.
This is also illustrated by the high degree of correspondence between the
apparent distribution coefficients (K* ) of the amended and diluent soils
and their mutual order of magnitude variation from that of the contaminated
soils. The apparent distribution coefficient is so designated as
distribution coefficients, K,s, are typically determined by introducing and
allowing a quantity of aqueous analyte to equilibrate in the presence of a
solid phase (10). The amount taken up on the solid phase is then determined,
as opposed to the present work which measures the amount of analyte released
by the solid phase under the specified conditions.
Thus, while the solid phase radium concentration of contaminated soils is
approximately 100 times greater than the diluent soil, the net radium
leachate contribution from the contaminated soils relative to the diluent
soil is only one order of magnitude greater. On a unit activity basis, the
native diluent soil therefore leached approximately 10 times more radium than
the contaminated soils. Differences in radium distribution between the solid
and aqueous phase related to the physical forms in which it is present, are
clearly indicated. This is plausible in that the contaminated soil is known
to contain unreacted carnotite ore as well as ore processing residues such as
BaSO , both of which would be distinct from the native diluent soil.
BaSO, in particular is known to strongly retain trace quantities of radium
(5).4
As noted, the results achieved from a leaching test are to a large extent a
function of the test that was done. Thus the observation noted in item 3
above, that amended soil leachate means are below the radium Primary Drinking
Water Standard of 5 pCi/1, need not necessarily imply that this assumed
standard would be met in an aquifer that was impacted by leachate emanating
from this material. In order to further assess potential environmental
impacts, validation of the testing protocol is required.
In the present case, the mean contaminated soil leachate concentration of
10.6 + 2.3 pCi/1 may be compared with the mean (1984-86) radium
concentrations detected in shallow monitoring wells which were located in or
immediately downgradient from the areas of known contamination in the
affected neighborhoods, 2.4 + 2.6 pCi/1 (Table 4). While not exact, this
1-287
-------
comparison does allow the inference that the leaching tests tend to
overestimate the concentration of radium which would actually be seen in the
environment under present conditions. This trend is similarly supported by
comparison of the diluent soil leachate mean, 1.2 ± 0.4 pCi/1, to a ground
water sample taken from the vicinity of the diluent soil source material,
0.12 pCi/1. While even less exact than the previous comparison, it does
serve to support the inference that the leaching test tends to overestimate
actual environmental concentrations.
B. Ground Water Transport
In-situ leachate generally does not remain stationary in the environment.
The potential for radium transport in ground water from an emplacement site
containing the amended soils was thus evaluated by utilizing the batch
extraction radium concentrations as input data to a simplified analytical
model (5). Analytical models typically make a number of simplifying
assumptions such as constant recharge and discharge rates, and uniform
aquifer characteristics. Thus, rate of leaching and subsequent transmission
through the unsaturated zone was determined by assuming a constant mass
influx rate over time. Short term seasonal recharge rate variations as well
as long term changes due to the chemical evolution of the amended materials
were not evaluated.
Conceptually, the model assumed high potential solute mobility in the
subsurface, i.e., emplacement siting was assumed to be in a permeable
environmental with little capacity for chemical interaction between aquifer
materials and solutes. Additionally, consistent with the presumption of the
amended material to be available for unrestricted use, no controlling layers
(caps, liners, etc.) were assumed to be present.
Radium would thus exit the emplaced amended soil as leachate through the
unsaturated zone contributing an initial concentration to the saturated zone
ground water as a function of leaching rate, radium concentration within the
leachate, and the retardation factor operating in the unsaturated zone.
Leached radium would then add to the radium already present in the saturated
zone. For conservativeness, retardation was assumed to be negligible. This
new total radium value would then be subject to attenuation via mixing and
dilution through transport and dispersion along the extant downgradient flow
path of the saturated zone. Retardation of the radium plume due to
distribution of the elevated quantities of dissolved radium onto the solid
phase of the downgradient aquifer, as above, was assumed to be negligible
thus further enhancing the conservativeness of the model. That these
retardation assumptions are indeed conservative is illustrated by the
observations of Krishniswami, et. al., whose in-situ studies of Connecticut
ground waters revealed that radium injected into the aqueous phase via alpha
recoil from the adjoining solid surfaces was removed from the aqueous phase
on a time scale of the order of minutes, the net result being that radium did
not migrate through ground waters far from the point of injection (11).
Water budget data was derived from EPA's Hydrological Evaluation of Landfill
Performance (HELP) computer program, based on inputs of New Jersey annualized
climatological data (12). Calculations were done utilizing parametric values
appropriate to a vegetated, loamy top soil overlying the emplaced amended
soil, which was assumed to have sandy characteristics.
1-288
-------
Model calculations were performed assuming an emplacement of amended soil
occupying an area of 900 x 900 feet to a depth of 10 feet, which was then
underlain by a 50 foot thick sand and gravel aquifer. Leachate concentration
of Ra-226 at the bottom of the fill was assumed to be either 3.2 pCi/1 or 4.1
pCi/1, which were the weighted (with respect to soil volumes represented by
the composited samples) mean and 95% upper bound confidence limit for the
amended soils in the batch extraction tests. Three dilution scenarios were
generated based on zero downgradient recharge, a "moderate" rate of recharge,
and the full recharge rate based on the HELP model analysis. Results
displayed in Table 5 showed radium concentrations ranging from 0.8 pCi/1 to
1.5 pCi/1, varying as a function of initial leachate concentration and
distance downgradient from the site (500-5000 feet). All of these values are
well below the Primary Drinking Water Standard for radium.
It should be noted that radium leachate values reported are for total aqueous
radium. No effort was undertaken to distinguish between radium which was
truly dissolved or which was present in a colloidal phase. The absence or
presence of significant quantities of radiocolloids cannot be conclusively
demonstrated here. This question is of some relevance as colloidal particles
are likely to display transport behavior different from that of dissolved
species. Such facilitated transport occurs when colloids move with advective
ground water flow (8). While radium has been reported not to form
homogeneous colloidal particles, it is known to adsorb onto other colloidal
species such as iron (8, 13, 14).
Further studies to confirm the apparent utility of the soil amendment
process would be desirable. A series of well designed laboratory column
and/or field lysimeter studies would provide a much closer simulation of
actual environmental behavior of these materials. As flow through a column
inherently incorporates both dilution and retardation, it would be reasonable
to anticipate radium values somewhat lower in magnitude than those
demonstrated in the current batch studies.
CONCLUSIONS
1. Among amended and diluent soils, no significant differences were seen
between extracts of pH 3 and pH 5, due to both the soil/water ratio used
in the extraction, which allowed the soil buffer to predominate, as well
as the calcareous nature of the diluent soil which provided an
effectively infinite buffer capacity.
2. Amended and diluent soil leachates are significantly lower in radium
concentration than those of the contaminated soils, however as percentage
of total radium extracted, the contaminated soils yield 10 times less
radium than the diluent or amended soils. This is reflected by the
apparent distribution coefficients of 600, 780 and 6190 ml/g for the
diluent, amended, and contaminated soils, respectively.
3. Comparison of mean leachate results from the diluent soil to a regional
groundwater value, along with comparison of the contaminated soil
leachate mean to the mean radium value from monitoring wells in contact
with the contaminated materials suggests that the leaching tests, as
performed, tend to overestimate the actual environmental concentration of
radium that would result from contact with the tested materials.
1-289
-------
Using the experimental amended soil leachate concentrations (3.2 and 4.1
pCi/1, weighted mean and 95% confidence limit, respectively) as inputs to
a simplified analytical groundwater model, results in radium
concentrations ranging from 0.8 pCi/1 to 1.5 pCi/1 varying with initial
input concentration and distance from the emplacement site.
In consideration of the conservativeness contained in the above points
(overestimate of environmental concentrations, permeable amended soil,
sand and gravel aquifer, negation of solute retardation) the exceedance
of the Primary Drinking Water Standard for radium of 5 pCi/1 in an
aquifer in receipt of leachate from this amended soil is unlikely.
The results presented represent a preliminary verification of the
environmental acceptability of amending soil contaminated with low level
radionuclides with a "clean" diluent soil. Further work to support this
conclusion to a higher degree of confidence should focus on colloidal
particles as no attempt was made to differentiate between radium in a
truly dissolved or colloidal state, as well as actual field studies which
can provide a more accurate assessment of the environmental dynamics
extant at an emplacement site.
Acknowledgements are graciously extended to C.J Touhill, Ph.D., E.A.
Rothfus, and G. Gumtz of Baker/TSA, Inc., along with R. Melgard of
TMA/Eberline, M.L. Morris, Ph.D. and A.P. Verma, Ph.D. of NJDEP, and R.A.
Salkie of USEPA (formerly NJDEP), for their diligent contributions to this
project which were meritoriously provided under conditions most challenging.
The opinions and conclusions expressed above are those of the author and
not necessarily that of the New Jersey Department of Environmental Protection.
1-290
-------
REFERENCES
1. Technology Screening Guide for Treatment of CERCLA Soils and Sludges.
EPA/540/2-88/044, Sept. 1988.
2. Technological Approaches to the Cleanup of Radiologically Contaminanted
Superfund Sites. EPA/540/2-88/002, August 1988.
3. Shake Extraction of Solid Waste with Water, D3987-81 (Method A) American
Society for Testing Materials, Annual Book of Standards, Volume 11.04, p.
32, 1984.
4. Research Report on the Extraction of Radium-226 from Soils, TMA/Eberline,
Albequerque, NM, March 1987.
5. Technical Background Information Report for the Soil Blending Program.
Baker/TSA, Inc., Coraopolis, PA, June 1987.
6. Shearer, Jr., S.D. , and G.F. Lee, Leachability of Radium-226 From Uranium
Mill Solids and River Sediments, Health Physics, _10, 217, (1964).
7. Shoesmith, D.W., The Behavior of Radium in Soil and in Uranium Mine
Tailings, Whiteshell Nuclear Research Establishment, Atomic Energy of
Canada Limited, AECL-7818, 1984.
8. Benes, P., M. Obdrzalek, and M. Cejchanova, The Physicochemical Forms of
Trace of Radium in Aqueous Solutions Containing Chlorides, Sulfates and
Carbonates, Radiochem. Radioanal. Letters 50, 4, 227-242, (1982).
9. Cote, P., T. Constable, J. Stegemann, R. Dayal, S. Sawell, R. Caldwell,
J. McLellan, Guide for the Selection of Leaching Tests. USEPA/Ontario
Ministry of the Environment, Draft October 1988.
10. Sheppard, M.I., D.I Beal, D.H. Thibault, P.O'Connor, Soil Nuclide
Distribution Coeffiecients and their Statistical Distributions,
Whiteshell Nuclear Research Esrablishment, Atomic Energy of Canada
Limited, AECL-8364, 1984.
11. Krishniswami, S, W.C. Graustein, K. F. Turekian, Radium, Thorium, and
Radioactive Lead Isotopes in Groundwaters: Application to the in Situ
Determination of Adsorption-Desorption Rate Constants and Retardation
Factors. Water Resources 1J3, 1663, (1982).
12. Schroeder P.R., et al., The Hydrologic Evaluation of Landfill Performance
(HELP) Model, USEPA, Technical Resource Document, EPA/530-SW-84-010.
13. Benes, P., Physicochemical Forms and Migration in Continental Water of
Radium from Uranium Mining and Milling, in "Environmental Migration of
Long-Lived Radionuclides", IAEA-SM-257/84, 1982.
1-291
-------
14. Szabo, Z., and O.S., Zapecza, Relation Between Natural Radionuclide
Activities and Chemical Constituents in Ground Water in the Newark Basin,
New Jersey, in "Radon in Groundwatwater", Lewis Publishers. P-roceedings
of the NWWA Conference April 7-9, 1987, Somerset NJ.
15. Draft Supplemental Feasbility Study for the Montclair/West Orange and
Glen Ridge Radium Sites, Volume 2. USEPA Contract Number 68-01-6939.
Camp, Dresseri and McKee, Inc. Edison, NJ. April 3, 1989.
1-292
-------
Table 1. Results of Radium Batch Extraction. Grand Means of Replicate Determinations.
Radium Phase Soil
Soil (pCi/g) C
Leachate (p Ci/1) C
Soil A
Leachate A
Soil
Leachate
N>
CO
CO
(b)
Virginia/Franklin
63 ± 4.4
11 ± 2.5
2.4 ± 0.4
2.4 ± 0.3
pH 3
(a)
Carteret
58 ± 5.9
12 ± 1.9
2.0 ± 0.2
3.7 + 2.9
Lorraine
74 ± 4.2
8.1 ± 0.8
2.4 ± 0.3
1.3 + 0.1
Diluent
0.7 ± 0.1
1.5 ± 1.2
Soil
Leachate
Soil
Leachate
Soil
Leachate
C
C
A
A
67 ± 19
7.4 ± 3.3
2.3 ± 0.4
2.8 ± 2.2
pH 5
(a)
56 ± 3.2
13 ± 0.9
2.0 ± 0.4
4.8 ± 2.7
75 ± 4.7
12 ± 2.6
1.7 ± 0.1
1.4 + 0.5
0.7 ± 0.1
0.9 ± 0.3
(a) Initial leachant pH
(b) C = Contaminated Soil, A = Amended soil
-------
Table 2
Comparison of Leachate by t-Test
at the 95% Confidence Limit, pH 3 vs.
pH Contaminated
Virginia-Franklin 3
NSD
5
Carteret 3
NSD
5
Lorraine 3
5 +
Diluent
pH 5.
pH
3
5
3
5
3
5
3
5
Amended
NSD
NSD
NSD
NSD
NSD = No Significant Difference
+ = Significantly Greater
= Significantly Lesser
1-294
-------
Table 3a. Mean Radium Values (±SD) in Soils and Leachate, Percent Radium Extracted and Apparent
Distribution Coefficients.
Leachate (pCi/1)
Soil (pCi/g)
% Extracted
Amended
(a)
Contaminated
Diluent(b)
(a)
2.7
10.6
1.2
± 1.4
± 3.3
± 0.4
2.1
65.6
0.7
± 0.3
± 8.0
± 0.0
0.51
0.065
0.69
780
6190
600
&
Ol
(a) n = 6
(b) n = 2
f , K' = C
(c) d
soil (Ra)/ leachate (Ra)
Table 3b. Radium Concentration
Amendeded /Diluent
Contaminated /Diluent
Contaminated /Amended
Ratios
Leachate
2.3
8.8
3.9
Soil
3.0
93.7
31.2
-------
Table 4. Radium Concentrations in Overburden Monitoring Wells Within Areas of Contaminated Soil (15).
Quarterly Ground Water Monitoring Results 8/27.84-6/12/86, pCi/1 ± SD.(a)
Well
M-S-3
M-S-1
M-S-2
G-S-2
iG G-8-1
CO
O>
11.8 ± 0.6 Dry Dry Dry Dry Dry 4. 2 ±0.5 0.6 ±0.3
3.8 ± 0.2 Dry Dry Dry Dry Dry 5. 2 ±0.5 1.0+0.3
N/A 3.3±0.2 1.7+0.1 3.2±0.2 0.5 ± 0.1 0.6+0.2 0.4±0.2 0.3+0.2
N/A 1.8+0.1 3.4±0.2 0.0+0.1 Dry Dry 7. 2 ±0.6 0.9 ±0.3
2.4 ± 0.1 4.0 + 0.2 0.6 + 0.1 1.8 + 0.1 1.5 ± 0.1 0.8 ± 0.4 1.0 + 0.4 1.3 + 0.4
Mean Value = 2.4 + 2.6 pCi/1
n = 26
(a) Only wells whose mean radium values exceeded background are tabulated. Background was assumed
to be <1.0 pCi/gm Ra-226.
-------
Table 5. Ground Water Transport Modeling Results.
R (ft/yr)
L (ft)
C =3.15 pCi/1
Cr = 1.21 pCi/1
= 4.10 pCi/1
=1.45 pCi/1
0
0
0
0.5
0.5
0.5
1.96
1.96
1.96
0
0
0
0.5
0.5
0.5
1.96
1.96
1.96
500
1500
5000
500
1500
5000
500
1500
5000
500
1500
5000
500
1500
5000
500
1500
5000
1.21
1.20
0.99
1.20
1.16
0.91
1.16
1.09
0.77
1.45
1.44
1.19
1.44
1.39
1.09
1.39
1.31
0.93
C = concentration of radium in leachate as it enters the ground water.
C = concentration of radium in ground water after mixing.
R = downgradient recharge rate
L = distance downgradient from emplacement fill boundary
C = radium concentration at distance L.
Constant Values
C, = Background aquifer radium concentration = 0.55 pCi/1 (based on weighted
average of US public drinking water supplies).
Qb = Aquifer flow = 3.6 x 105 1/d
Qr = volumetric flow rate of recharge through the fill = 1.23 x 10 1/d
Z = fill emplacement depth = 50 feet
V = velocity of radium in the aquifer = 296 feet/year
(V = V , aquifer velocity, as retardation is assumed to be zero).
c w J
1-297
-------
INTERIABORATORY COMPARISON OF METHODS 1310, 1311 AND
1312 FOR LEAD IN SOIL
ALVIA GASKUL, JR., ENVIRONMENTAL REFERENCE MATERIALS, INC., RESEARCH TRIANGLE
PARK NC 27709 GAIL A. HANSEN, U.S. ENVIRONMENTAL PROTECTION AGENCY, OFFICE OF
SOLID WASTE, WASHINGTON, DC 20460, ROBERT S. TRUESDALE, WILLIAM B. YEAGER,
RESEARCH TRIANGLE INSTTTUTE, RESEARCH TRIANGLE PARK, NC 27709.
ABSTRACT
The Extraction Procedure Toxlcity Test (EP), Method 1310 and its intended
replacement, the Toxicity Characteristic Leaching Procedure (TCLP), Method
1311 were both designed to simulate leaching of an industrial waste
disposed in a sanitary landfill. The applicability of either one to
assessment of the clean up levels of contaminated soils has been
questioned. The basic objection is that the sanitary landfill codisposal
scenario of Methods 1310 and 1311 does not apply to contaminated soils
where organic acids like the acetic acid used in these methods are not
expected to be present. If Methods 1310 or 1311 were used to assess such
site clean ups, the acetic acid might solubilize certain elements like lead
and incorrectly classify the soil as hazardous when, in fact, no such
mobilization would actually be expected to occur.
To address the need for a predictive leaching method for contaminated soil,
Method 1312, Synthetic Precipitation Leach Test for Soils has been
developed. Method 1312 is designed to determine the mobility of both
organic and inorganic contaminants present in soils. The Method uses the
equipment and conditions of Method 1311, but instead of acetic acid, it
uses an extraction fluid which is intended to model the precipitation of
the region of the country where the soil site is located.
The purpose of this study was to determine the precision of Method 1312.
This was accomplished by conducting an interlaboratory study involving six
soils each containing lead at levels ranging from hundreds of /;g/g to
percent levels. In addition, comparison data were obtained using Methods
1310 and 1311. Four acid soils and two alkaline soils wer° collected and
made as homogeneous with respect to bulk lead as possible without excessive
particle size reduction. Bulk lead was determined to characterize the
soils and to assess the success of the homogenization steps.
1-298
-------
A COMPARISON OF THE TOXICITY CHARACTERISTIC LEACHING
PROCEDURE (TCLP) AND A MODIFIED TCLP IN AN EVALUATION OF
STABILIZED OIL SLUDGE
KATHRYN BRADY TONER. ELAINE D. KEITHAN, PH.D., AND
STEPHEN PANCOSKI, ENVIRONMENTAL SCIENCE LAB, BUCKNELL
UNIVERSITY, LEWISBURG, PENNSYLVANIA 17837
ABSTRACT
An acidic petroleum sludge was stabilized in a related
research project which utilized a variety of products
including cements, organophilic clays, pozzolanic
materials, and proprietary agents. The effectiveness of
the stabilization process was evaluated by using two
procedures, the Toxicity Characteristic Leaching Procedure
(TCLP) and a modification of that procedure. The TCLP
(excluding volatile organics analysis) was performed in
accordance with EPA specifications. The modified procedure
used sulfuric acid (H2SO^) as the extraction fluid instead
of acetic acid, as designated by the TCLP- Replicate
stabilized (treated) sludge samples, as well as
unstabilized (untreated) sludge samples, were evaluated
concurrently using both testing procedures to allow
comparison.
Initial findings indicate that differences exist in the
results obtained from the TCLP and the modified TCLP. In a
comparison of the modified and unmodified TCLP, the results
indicate that more semivolatile organic compounds (in
particular, straight-chain alkanes and polyaromatic
hydrocarbons) are detected in the modified extract.
INTRODUCTION
An investigation is currently underway whereby an acidic
petroleum sludge is stabilized using a variety of materials
including cements, organophilic clays, pozzolanic
materials, and proprietary agents; this methodology has
been published previously (Evans, et al., 1988; Pancoski,
et al., 1988). The effectiveness of this treatment was
evaluated by physical and chemical analyses; however, only
the chemical testing is discussed in this paper. The
chemical testing consists of analyzing the treated and
untreated oil sludge with respect to leaching potential
using the Toxicity Characteristic Leaching Procedure (TCLP)
as specified by the EPA (Federal Register, 1986), and a
modification of that procedure. The modified procedure
used sulfuric acid (H2SO4) as the extraction fluid instead
of acetic acid, as designated by the TCLP. Acetic acid is
intended to simulate landfill leachate. Because the waste
1-299
-------
used in this study will be disposed of on-site, and not in
a sanitary landfill, other acids were considered for the
extraction fluid. H2S04 was chosen for two reasons.
First, the oil sludge was acidified with H2S04 in the
refining process. Secondly, the oil sludge is currently
stored in open pits, constantly exposed to the environment;
thus, the H2SO4 would mimic conditions caused by acidic
precipitation.
METHODS
Replicate stabilized sludge samples were evaluated
concurrently using both testing procedures. The TCLP
(excluding volatile organics analysis) was performed in
accordance with EPA specifications. In the modified TCLP,
a sample of the treated sludge was evaluated by, first,
reducing the particle size using a 9.5 mm sieve, and then
placing a 100 gram sample in a flint glass jar with 1600 ml
of deionized water (16x the sample weight). The pH was
then determined and, if the pH was greater than 5, H2SO4
was added until the pH was 5 or less. The sample was
extracted in a rotary agitator for eighteen hours. The pH
was monitored during this extraction, and adjusted (to pH 5
or less) when necessary. After the agitation period, 400
ml of deionized water (4x the sample weight) was added, and
the sample was filtered under pressure using glass filter
paper. The filtrate obtained was monitored for the
presence of several Priority Pollutant metals, as well as
other metals found in the stabilization additives, using
atomic absorption spectroscopy; in addition, a one liter
subsample of the filtrate was extracted with methylene
chloride following the TCLP base/acid extraction procedure.
An internal standard (dlO-acenaphthene) was then added to
the resulting sample (at a concentration of 85 ppb). A 1
ul portion of this sample was then analyzed for
semivolatile organic compounds, including several Priority
Pollutants and straight-chain hydrocarbons from C10 to C24,
using a Hewlett Packard gas chromatograph/mass selective
detector equipped with HP-UX Chemstation software.
Untreated sludge samples were also evaluated using both the
TCLP and modified TCLP to provide a basis for comparison.
Another parameter investigated was the relative hydrocarbon
concentration, which was a measurement devised to compare
the relative amount of unresolved material under the gas
chromatograph "hump" (area of unresolved peaks) to the area
under the internal standard peak (See Figure 1) .
Specifically, the area of unresolved peaks was determined
using a digitizer, and this number was divided by the area
under the internal standard, reported in digital counts.
1-300
-------
This value was then multiplied by IxlO5 and termed the
relative hydrocarbon concentration.
RESULTS
In this laboratory investigation, the untreated oil sludge
was tested five times: three times using the modified
(sulfuric acid) procedure and two times using the
unmodified (acetic acid) TCLP. It should be noted that the
untreated sludge contained free liquids along with a
viscous petroleum sludge material. For both leaching
procedures, only the viscous sludge material was analyzed.
For the treated, or stabilized, material, thirteen
different samples were analyzed, once with sulfuric acid
and once with acetic acid. Results of these analyses are
presented in Table 1 (metals analysis data and RHC values)
and Tables 2a and 2b (semivolatile organics analysis data).
For the untreated sludge, the test results differed with
regard to the two leaching procedures. In the metals
analysis, higher concentrations were reported for calcium
and sodium when acetic acid was used. In the case of
nickel, higher concentrations were reported for the leach
tests in which sulfuric acid was employed. Regarding the
organic analysis, higher concentrations of dibutyl
phthalate and benzyl alcohol were found in the acetic acid
procedure. For the samples leached in sulfuric acid, the
concentrations of the following semivolatile organic
compounds were one to two orders of magnitude higher than
the same samples leached with acetic acid: decane,
undecane, dodecane, tetradecane, hexadecane, octadecane,
eicosane, tetracosane, naphthalene, fluorene, phenanthrene,
pyrene, methyl naphthalene, and dimethyl naphthalene. The
first eight compounds in the list above are straight-chain
alkanes, and the latter six compounds are polyaromatic
hydrocarbons.
For the thirteen treated sludge samples, some differences
in chemical concentrations between the two leach procedures
existed. With regard to the metals analysis, calcium
concentrations were higher in all thirteen of the samples
tested with sulfuric acid. No apparent differences existed
in overall comparisons among the other metals which were
analyzed. In the organic analysis, dibutyl phthalate
concentrations were higher in eight of the thirteen
samples, and phenol concentrations were higher in ten of
the thirteen samples when acetic acid was used. In seven
of eight samples in which fluorene was detected, sulfuric
acid was the leaching fluid. In general, there were no
readily apparent differences in concentrations for the
1-301
-------
other organics as evaluated by the two different leaching
procedures.
The relative hydrocarbon concentration (RHC) was used as a
general indicator of the amount of hydrocarbon constituents
present but not quantifiable by GC/MS. Nine of the
thirteen samples tested with acetic acid had higher RHC
values than the same samples tested in sulfuric acid.
Therefore, the unmodified TCLP (utilizing acetic acid)
tended to leach greater amounts of hydrocarbon contituents
(unresolvable by gas chromatography) from the stabilized
petroleum sludge.
SUMMARY
Differences exist in the results obtained from the two
leaching procedures, a modified TCLP (sulfuric acid
leaching fluid) and the unmodified TCLP (acetic acid
leaching fluid). In general, the modified procedure, using
sulfuric acid, leached more semivolatile organics (in
particular, straight chain alkanes and polyaromatic
hydrocarbons) from the untreated oil sludge. Also, higher
nickel concentrations were reported in the untreated sludge
tested with sulfuric acid; however, acetic acid produced
higher calcium and sodium concentrations in the untreated
sludge. For the treated sludge samples, higher
concentrations of dibutyl phthalate and phenol were
reported when the acetic acid procedure was used, while use
of the sulfuric acid procedure resulted in higher fluorene
and calcium concentrations. These preliminary results
suggest that acetic acid and sulfuric acid vary in their
ability to leach specific metals and semivolatile organics
from treated and untreated petroleum sludges. Further
research is planned on this topic.
ACKNOWLEDGEMENTS
Funding for this research has been provided by the
Pennsylvania Ben Franklin Partnership and the SUN Refining
and Marketing Company. The authors wish to thank the
following people for assistance in this research project:
Dr. Jeffrey C. Evans, Dr. Michael D. LaGrega, Dr. Arthur
Raymond, Jason Strayer, Jeff Lasselle, Jay Kisslak, Connie
Snyder, and, especially, Mr. Jim Spriggle.
1-302
-------
REFERENCES
Evans J. c.; LaGrega M.D.; Pancoski, S.E.; and Raymond
A., "Methodology for the Laboratory Investigation of the
Stabilization/Solidification of Petroleum Sludges",
Superfund '88 - Proceedings of The 9th National Conference,
pp. 403-408, HMCRI, Silver Spring, MD, 1988.
Federal Register, "Appendix I to Part 268, Toxicity
Characteristic Leaching Procedure, (TCLP)", Part II
Environmental Protection Agency 40 CFR Part 268 et al, pp.
40643-40653, 51, (216), Friday, Nov. 7, 1986.
Pancoski, S.E.; Evans J.C.; LaGrega M.D.; and Raymond
A., "Stabilization of Petrochemical Sludges", Hazardous and
Industrial Waste - Proceedings of the Twentieth Mid-
Atlantic Industrial Waste Conference, pp. 299-316, HMCRI,
Silver Spring, MD, 1988.
1-303
-------
s
FIGURE 1: SAMPLE CHROMATOGRAM ILLUSTRATING
RELATIVE HYDROCARBON CONCENTRATION
-------
TABLE 1: METALS ANALYSIS DATA AND RHC VALUES
extraction calcium sodium lead nickel copper
SAMPLE acid
used
ppm
ppm ppcn ppm
ppm
magnesium zinc cadmium chromium relative
hydrocarbon
ppm ppm ppm ppm concentration
treated sludge 51
treated sludge 51
treated sludge 52
treated sludge 52
treated sludge 67
treated sludge 67
treated sludge 70
treated sludge 70
treated sludge 77
treated sludge 77
treated sludge 78
treated sludge 78
treated sludge 89
treated sludge 89
treated sludge 90
treated sludge 90
treated sludge 99
treated sludge 99
treated sludge 105
treated sludge
treated sludge
treated sludge
treated sludge
treated sludge
treated sludge
treated sludge 108
untreated sludge
untreated sludge
untreated sludge
untreated sludge
untreated sludge
51
51
52
52
67
67
70
70
77
77
78
78
89
89
90
90
99
99
105
105
106
106
107
107
108
108
le
ie
ie
ie
ie
acetic
su If uric
acetic
su If uric
acet i c
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acet i c
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acet i c
sulfuric
acetic
acetic
sulfuric
sulfuric
sulfuric
0.00
5.81
0.00
20.26
0.00
11.10
0.00
24.80
0.00
16.30
0.00
9.90
0.00
8.70
0.00
29.30
0.00
35.20
0.00
7.68
0.00
7.07
0.00
8.93
0.00
7.29
13.90
19.57
12.57
13.07
11.45
71.00
45.00
62.00
55.00
40.00
50.00
180.00
180.00
200.00
180.00
100.00
110.00
100.00
130.00
210.00
230.00
90.00
100.00
40.00
40.00
50.00
40.00
80.00
70.00
50.00
60.00
1320.00
1170.00
21.00
23.00
21.00
0.10
0.00
0.10
0.30
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.37
0.03
0.67
0.64
0.61
0.63
0.65
0.03
0.01
0.02
0.02
0.65
0.59
0.63
0.64
0.62
0.64
0.03
0.03
0.03
0.01
0.04
0.01
0.04
0.02
0.00
0.00
0.18
0.18
0.07
0.01
0.10
0.01
0.50
0.03
0.03
0.02
0.14
0.04
0.03
0.05
0.04
0.08
0.05
0.03
0.08
0.03
0.01
0.04
0.03
0.04
0.01
0.04
0.01
0.04
0.02
0.04
0.05
0.00
0.00
0.00
18.00
111.00
28.00
135.00
2.50
1.80
11.40
1.40
10.00
54.00
10.00
33.00
0.70
5.30
2.50
2.10
1.80
1.80
1.20
2.90
1.30
2.50
0.90
0.80
1.30
1.20
5.00
8.00
8.00
8.00
6.00
0.01
0.65
0.16
0.01
0.04
0.08
0.07
1.00
0.00
0.01
0.00
0.00
0.02
0.00
0.21
0.75
0.09
0.75
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.51
1.00
1.00
1.00
1.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.40
0.00
0.00
0.00
0.50
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.10
0.00
0.05
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.15
0.00
0.00
0.335
0.113
0.242
0.07
0.204
0.103
0.113
0.014
0.185
0.095
0.116
0.413
0.024
0.104
0.034
0.045
0.041
0.022
0.122
0.066
0.061
0.038
0.522
0.025
0.028
0.075
0.148
0.188
0.115
0.241
0.16
-------
TABLE 2a: SEMI VOLATILE ORGANICS ANALYSIS DATA
extraction decane undecane dodecane tetradecane hexadecane octadecane eicosane tetracosane phenol dimethyl methyl
SAMPLE acid phenol phenol
used ppbppbppb ppb ppb ppbppb ppb ppbppbppb
treated sludge 51
treated sludge 51
treated sludge 52
treated sludge 52
treated sludge 67
treated sludge 67
treated sludge 70
treated sludge 70
treated sludge 77
treated sludge 77
treated sludge 73
treated sludge 78
treated sludge 89
treated sludge 89
treated sludge 90
treated sludge 90
treated sludge 99
treated sludge 99
treated sludge 105
treated sludge
treated sludge
treated sludge
treated sludge
treated sludge
treated sludge
treated sludge
untreated sludge
untreated sludge
untreated sludge
untreated sludge
untreated sludge
51
51
52
52
67
67
70
70
77
77
78
78
89
89
90
90
99
99
105
105
106
106
107
107
108
108
je
3e
36
ge
ge
acetic
su If uric
acetic
su If uric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
acetic
sulfuric
sulfuric
sulfuric
ND
ND
ND
NO
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
81.00
177.00
289.00
ND
NO
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
9.00
291.00
435.00
141.00
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
27.40
10.20
553.00
686.00
592.00
ND
ND
ND
ND
ND
6.20
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
14.60
7.10
1923.00
1790.00
1878.00
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.50
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
2.30
ND
19.00
11.80
2616.00
2626.00
2396.00
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
14.30
12.10
2069.00
1843.00
1701.00
ND
ND
NO
ND
ND
ND
ND
ND
NO
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
25.50
17.20
3502.00
2895.00
2725.00
ND
ND
ND
ND
29.60
ND
ND
ND
ND
11.10
ND
ND
6.00
ND
8.40
35.80
ND
ND
ND
ND
ND
ND
ND
ND
NO
ND
131.20
11.00
3829.00
3381.00
2731.00
1297.00
457.20
553.30
218.50
1288.00
1464.00
1855.00
1778.00
3499.00
3389.00
2484.00
2166.00
1916.00
1741.00
2211.00
2479.00
2585.00
2203.00
2256.00
2281.00
2038.00
640.00
2260.00
467.00
2527.00
1364.00
1106.00
823.70
1845.00
3928.00
620.00
119.70
31.70
43.50
19.60
60.30
79.90
39.50
53.30
206.60
156.90
101.50
57.70
166.20
88.00
57.60
58.70
58.30
73.90
202.90
193.20
176.60
185.50
201.10
143.50
241.50
124.80
189.50
129.00
218.00
250.00
186.00
397.90
280.20
71.70
55.10
271.20
364.10
229.70
272.20
757.40
646.30
485.20
355.00
460.10
307.40
303.80
333.50
355.50
372.10
527.00
547.10
475.90
545.60
567.00
408.90
650.70
344.30
242.30
151.60
716.00
242.00
361.00
ND - Not Detected
-------
TABLE 2t>: SEMIVOLATILE ORGANICS ANALYSIS DATA
SAMPLE
treated sludge 51
treated sludge 51
treated sludge 52
treated sludge 52
treated sludge 67
treated sludge 67
treated sludge 70
treated sludge 70
treated sludge 77
treated sludge 77
treated sludge 78
treated sludge 78
treated sludge 89
treated sludge 89
treated sludge 90
treated sludge 90
treated sludge 99
treated sludge 99
treated sludge 105
treated sludge 105
treated sludge 106
treated sludge 106
treated sludge 107
treated sludge 107
treated sludge 108
treated sludge 108
untreated sludge
untreated sludge
untreated sludge
untreated sludge
untreated sludge
extraction
acid
used
acetic
sulfuric
acetic
sulfuric
acet i c
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acetic
sulfuric
acet i c
sulfuric
acetic
acetic
sulfuric
sulfuric
sulfuric
dibutyl
phthalate
ppb
6.20
4.70
7.70
6.20
3.10
2.10
4.50
2.80
2.30
2.10
1.70
ND
2.80
3.50
2.20
2.80
2.10
2.50
2.50
4.90
3.00
1.30
3.60
1.40
4.20
3.30
2.70
12.00
ND
ND
ND
diethylhexyl
phthalate
ppb
174.60
133.70
205.10
143.80
124.00
106.70
44.40
60.70
239.00
296.40
106.40
78.60
1916.00
192.00
34.50
1525.00
44.10
35.30
125.50
199.30
127.90
144.10
1032.00
1156.00
2566.00
170.80
135.20
35.40
ND
ND
68.00
naphthalene
ppb
27.40
21.90
14.00
9.20
17.80
21.20
18.10
13.90
25.30
26.30
15.20
15.80
15.00
20.90
15.30
16.40
11.30
21.60
22.80
24.80
20.70
26.90
22.40
21.40
24.70
16.30
80.20
54.60
162.00
165.00
143.00
f luorene
ppb
ND
ND
ND
ND
ND
2.20
ND
ND
ND
2.20
ND
ND
2.10
ND
ND
ND
ND
1.90
ND
2.50
ND
2.10
ND
1.90
ND
2.10
9.40
8.80
91.00
84.00
72.00
phenanthrene
ppb
3.60
2.20
1.30
0.90
4.40
4.30
1.30
1.30
2.40
4.00
1.50
1.70
3.60
3.80
2.20
2.30
2.50
3.50
4.60
4.30
2.10
3.50
2.20
2.30
3.60
3.60
30.50
22.10
353.00
312.00
243.00
pyrene
ppb
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
5.00
2.20
114.00
111.00
85.00
methyl
naphthalene
ppb
40.00
33.80
20.80
15.20
27.50
35.60
25.00
21.60
38.10
41.50
25.40
23.10
27.00
33.70
26.10
26.20
27.70
36.00
34.20
38.10
28.70
42.20
34.60
32.20
39.10
25.00
172.00
122.50
732.00
702.00
666.00
dimethyl
naphthalene
ppb
11.40
10.80
5.80
4.10
9.90
12.50
ND
6.80
12.10
13.10
10.70
ND
10.50
11.60
9.30
9.60
10.60
12.80
12.20
14.70
10.80
14.20
10.10
9.80
12.60
8.60
82.30
67.20
842.00
762.00
780.00
diethyl
phthalate
Ppb
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.20
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
benzyl
alcohol
ppb
243.60
259.70
13.40
ND
19.70
27.00
ND
12.80
26.00
63.40
13.40
ND
46.70
9.70
865.90
947.20
666.60
749.90
729.70
612.40
18.60
3.00
65.60
14.20
21.40
34.60
70.10
17.50
ND
ND
ND
ND - Not Detected
-------
TCLP EXTRACTION OF REFERENCE WASTE SAMPLES STORED OVER TIME
Susan S. Sorini, Western Research Institute, P.O Box 3395, University Station,
Laramie, Wyoming 82071
ABSTRACT
Four sets of reference samples were prepared and stored at 4°C for 24 months.
The materials used to prepare the samples were a metal plating waste, an API
separator sludge, a creosote-contaminated soil, and a. smelter dust. Immediately
following preparation of each set, a sample was withdrawn at random and tested
using the Toxicity Characteristic Leaching Procedure (TCLP). Periodically
during their storage, additional random samples were selected from each set and
tested by the TCLP. The average elemental concentrations determined in the TCLP
extracts generated over time for each set of samples were compared. Although in
many cases the data are limited due to elemental concentrations being below
analytical detection limits, they provide information on TCLP reproducibility,
sample stability, and sample variability. The results in general suggest that
(I) the TCLP shows good reproducibility for inorganic analytes; (2) based on
TCLP characterization of their inorganic constituents, the samples have remained
stable; and (3) based on the quantity of sample used in the TCLP and the
inorganic data available, there is interbottle homogeneity among samples of the
metal plating waste, creosote-contaminated soil, and smelter dust. However, the
data indicate that the API separator sludge samples are poorly homogenized.
INTRODUCTION
Public Law 94-580, the Resource Conservation and Recovery Act (RCRA), requires
that all wastes destined for land disposal be evaluated for their potential
hazard to the environment. Wastes are deemed hazardous by a series of four
classification tests that assess reactivity, ignitability, corrosivity, and
toxicity.
Introduced in 1980, the Extraction Procedure Toxicity Characteristic (EPTC) is
currently used to assess the toxicity of wastes destined for land disposal (1).
However, in 1984, Congress amended RCRA requiring the EPA to develop a more
accurate leaching test (2). A new test procedure was officially proposed in the
January 1986 Land Disposal Restrictions (3) , in a further modified version as a
proposed replacement to the EPTC in June 1986 (4), and in still another version
in the November 1986 final rule making on the Land Disposal Restrictions (5).
It has not yet appeared in a final version that replaces the EPTC, but is
expected to do so in the near future. The new test is known as the Toxicity
Characteristic Leaching Procedure (TCLP).
In satisfying the requirements of the regulations under RCRA, analyzing wastes
using the TCLP will account for a significant volume of analytical testing by
the regulatory community. As a result, standard reference samples prepared from
wastes for use as performance evaluation and/or quality assurance samples for
the TCLP will be very beneficial. For this reason, a study was undertaken to
prepare reference solid waste samples for the TCLP and to evaluate the stability
of the samples stored over time based on TCLP characterization of their
inorganic constituents. Data generated in this study provide information on
TCLP reproducibility, sample stability, and sample variability.
EXPERIMENTAL
A metal plating waste, an API separator sludge, a creosote-contaminated soil,
-------
and a smelter dust were obtained for preparation of TCLP reference samples.
Each waste material was homogenized in a batch mixer equipped with paddle-style
blades. Homogeneity was determined by visual inspection. Following
homogenization, the material was added to commercially precleaned bottles (6).
Each bottled sample was placed in a poly bag, sealed with a twist tie, and
placed in a metal can, which was then filled with vermiculite and sealed with a
lid. A crimp overseal was placed on the can to prevent accidental opening. A
label was placed on the outside of the can corresponding to the label on the
sample bottle identifying the waste, the container size and type, the position
in the sequence in which it was packaged, and the mass of material in the
container. The sample cans were then placed in cardboard boxes in sequence and
stored at 4'C.
Immediately following preparation of the sets of metal plating waste samples,
creosote-contaminated soil samples, and smelter dust samples, a container was
selected at random from each set and its contents tested using the TCLP (4).
Due to an oversight a sample of the API separator sludge was not tested by the
TCLP prior to storage. The TCLP was applied in duplicate to the metal plating
waste and smelter dust, whereas a single TCLP extraction was performed on the
creosote-contaminated soil.
Periodically during their storage, additional random samples were selected from
each set and tested by the TCLP. During the period between sample preparation
and the first stability testing episode, the November 1986 version of the TCLP
was published in the Federal Register (5). The TCLP as specified in this
publication was followed for all stability testing.
The metal plating waste samples were tested for stability after six months, 18
months, and 24 months of storage. The API separator sludge samples were tested
after five months, 17 months, and 23 months of storage. The creosote-
contaminated soil samples were tested for stability after 4.5 months, 17 months,
and 23 months of storage, and the smelter dust samples were tested after 10
months and 16 months of storage.
For each set of samples, stability testing involved randomly selecting five
sample containers from each set and applying the TCLP to a 100-gram sample of
material from each container. In addition, single alkalinity determinations
were performed on a portion of sample from each bottle. The TCLP was also
applied to a blank consisting of extraction fluid from the same batch used to
extract the five samples. The pH values of the TCLP slurries prior to
filtration and of the TCLP leachates following filtration were recorded. The
leachates were analyzed for inorganic elements, generally including the eight
metals currently regulated under RCRA and three to four additional elements
selected as indicators of matrix stability. Inductively coupled plasma (ICP)
spectroscopy and atomic absorption spectroscopy (AAS) were used to analyze the
extracts (7). Specific methods used and the elements determined for each waste
type are given in the data tables presented in the results section.
RESULTS AND DISCUSSION
Metal Plating Waste
The time zero, six-month, 18-month, and 24-month leachates of the metal plating
waste samples had pH values ranging from 5.6 to 5.9. In addition, the
extraction slurries generated at each of the testing times had pH values ranging
from 5.7 to 5.9.
1-309
-------
Listed in Tables 1 through 4 are the elemental concentrations determined in the
TCLP leachates of the metal plating waste samples. Leachate concentration
values determined for silver, arsenic, cadmium, chromium, lead, selenium, and
mercury are all below instrumental detection limits, whereas barium
concentrations range from approximately 0.2 mg/L to less than 0.10 mg/L. The
concentrations of copper, magnesium, and nickel in the leachates are all above
detection limits, and as a result, can be more closely compared in Tables 5
through 7 .
Listed in these tables are the elemental concentrations in the 15 TCLP leachates
that have been generated over time. The average concentrations for the six-
month, 18-month, and 24-month testing times; the percent standard deviation of
the values for each testing time; and the 90% confidence interval for each data
set are also presented. In addition, the values from the three testing times
have also been treated as a single data set, and the overall average, percent
standard deviation, and 90% confidence interval calculated from the 15 values
are listed.
The data presented in Tables 5 through 7 indicate that the TCLP results for
copper, magnesium, and nickel have been reproducible over 24 months, and based
on the concentrations of these elements, the samples appear to have remained
stable. For each of these elements, the variability between data generated for
each testing time is approximately the same as the overall variability, and the
average elemental concentrations determined at each testing time for each of the
elements are similar. Based on the overall percent standard deviations and the
percent standard deviations at each testing time, the samples are homogeneous in
their distributions of magnesium and nickel; however, they do not appear to be
homogeneous with respect to their copper concentrations. This may be due to the
chemical speciation of copper in the metal plating waste matrix.
API Separator Sludge
The pH values of the five-month, 17-month, and 23-month extraction slurries
ranged from 5.3 to 5.8. The pH values of the five-month, 17-month, and 23-month
final leachates ranged from 5.3 to 6.0.
Tables 8 through 10 list the elemental concentrations determined in the TCLP
leachates of the API separator sludge samples. Leachate concentration values
determined for silver, arsenic, cadmium, lead, selenium, nickel, and mercury are
all below instrumental detection limits, whereas chromium concentrations range
from 0.11 mg/L to less than 0.10 mg/L. The concentrations of barium, iron, and
magnesium are all above detection limits, and as a result, can be compared more
closely in Tables 11 through 13.
These tables present data similar to the data in Tables 5 through 9 for the
metal plating waste; however, interpretation of the API separator sludge data is
more difficult. The percent standard deviations between the five-month
concentration values for all three elements are low, whereas variability between
the 17-month values and variability between the 23-month values for the three
elements are very high. This does not appear to be due to sample instability or
poor TCLP precision. For barium, the percent standard deviation between the
average values calculated at five months, 17 months, and 23 months is 21%, for
iron it is 11%, and for magnesium it is 7%. This indicates that although the
concentration values vary considerably overall and within the 17-month and
23-month testing times, the average values do not.
1-310
-------
The API separator sludge was obtained from a petroleum refining facility. The
material had been withdrawn from a sludge pit that had been used as a collection
area for other sludges, plant slops, debris, and soil. The API separator sludge
was also contaminated with fragments of metal, wire, and cloth.
Based on the nonuniformity of the API separator sludge material used to prepare
the samples and the data presented in Tables 11 through 13, it appears that the
samples were poorly homogenized, but they have remained stable, and their TCLP
results are reproducible.
Creosote-Contaminated Soil
The pH value of the time zero extraction slurry and final leachate was 5.1. In
addition, the pH values of the 4.5-month, 17-month, and 23-month extraction
slurries and final leachates were also 5.1.
Tables 14 through 17 list the elemental concentrations determined in the TCLP
leachates of the creosote-contaminated soil samples. Mercury was not determined
in the 4.5-month leaohate (Table 15) because of its high ICP detection limit.
Iron was added to the list of analytes at the 17-month testing time (Table 16)
to provide additional data because so many of the previously determined
concentration values were below instrumental detection limits. Leachate
concentration values determined for silver, arsenic, cadmium, chromium, lead,
selenium, and mercury are all below instrumental detection limits, whereas
nickel concentrations range from 0.15 mg/L to less than 0.10 mg/L. The
concentrations of barium, iron, and magnesium are all above detection limits,
and as a result, can be compared more closely in Tables 18 through 20.
These tables present the elemental concentrations determined in the TCLP
leachates that have been generated over 23 months. The average concentrations
for the 4.5-month, 17-month, and 23-month testing times; the percent standard
deviation of the values for each testing time; and the 90% confidence interval
for each data set are also listed. In addition, the values from the three
testing times have been treated as a single data set and the overall average,
percent standard deviation, and 90% confidence interval calculated from the
values are listed.
The data presented in Tables 18 and 20 show that the TCLP results for barium and
magnesium are reproducible over 23 months, and based on the concentrations of
these elements, the samples appear to have remained stable. Not only is the
variability between the data generated for each testing time very similar to the
overall variability, but the averages for each of the testing times are also
very similar and many of the confidence intervals overlap. Because the
4.5-month data for iron were not determined, it is difficult to make conclusions
based on the data presented in Table 19; however, it appears that the samples
may not be homogeneous in their distribution of iron.
Smelter Dust
Comparison of the pH values of the extraction slurries and final leachates
generated during each extraction of the smelter dust samples shows close
correlation among the data. The pH values of the time zero extraction slurries
and final leachates were 4.4, whereas the pH values of the 10-month and 16-month
extraction slurries and final leachates ranged from 4.6 to 4.7.
Tables 21 through 23 list the elemental concentrations determined in the TCLP
leachates of the smelter dust samples. Leachate concentration values determined
1-311
-------
for mercury are all below the instrumental detection limit, whereas chromium
values range from 0.15 mg/L to less than 0.10 mg/L, and silver values range from
0.18 mg/L to less than 0.10 mg/L. The concentrations of barium, cadmium,
copper, iron, lead, arsenic, magnesium, and nickel are all above instrumental
detection limits, and as a result, can be compared more closely in Tables 24
through 31. In addition, because selenium was analyzed by furnace AAS in the
10-month and 16-month extracts, concentration values for that element are also
compared in Table 32.
These tables list the elemental concentrations determined in the 10 TCLP
leachates that have been generated over 16 months. Although there are only two
data sets that can be compared for each element, it is advantageous that there
are data for nine elements. The average concentrations for the 10-month and
16-month testing times are listed along with the percent standard deviation of
the values, and the 90% confidence interval for each data set. In addition, the
values from the two testing times are treated as a single data set and the
overall average, percent standard deviation, and 90% confidence interval
calculated from the 10 values are listed.
The data presented for barium, cadmium, copper, lead, magnesium, nickel, and
selenium in Tables 24, 25, 26, 28, 30, 31, and 32, respectively, show similar
averages for the two testing times, similar overall variability compared to the
variability within the two data sets, and overlap of the 90% confidence
intervals calculated from the 10-month and 16-month data. The data indicate
that with respect to these elements, the samples have remained stable over 16
months and that their TCLP results are reproducible. In addition, sample
homogeneity in terms of the distribution of these elements also appears to be
good. However, it should be noted that comparing the time zero values listed
for cadmium in Table 25 with the 10-month and 16-month data shows that the time
zero values are higher. Also, a comparison of the time zero values listed for
lead in Table 28 with the 10-month and 16-month data shows that the time zero
values are lower. It is difficult to make any conclusions based on these
comparisons because the time zero data are limited in that only one sample was
tested. Interpretation of the data for iron in Table 27 is difficult. The time
zero values from the testing of one sample in duplicate are similar to those
values determined at 10-months; however, all of the values determined in the
16-month extracts are much lower. The same is true of the data for arsenic in
Table 29. Possible explanations are that the samples are not homogeneous in
their distributions of iron and arsenic, and by chance, random samples were
selected at each testing time with similar distributions; that error was made in
analyzing the sets of leachates; or that the samples have not remained stable
with respect to iron and arsenic. It is not possible to determine if any of
these explanations are correct based on the limited data available. Additional
testing is required to make any conclusions concerning the concentrations of
these elements in the TCLP leachates of the smelter dust samples.
As previously mentioned, the sample labels contain information concerning sample
position in the preparation sequence. Examining the positions of the smelter
dust samples selected for testing showed no correlation between elemental
concentrations in the TCLP leachates and sample position in the preparation
sequence. This was also true for the metal plating waste samples, API separator
sludge samples, and creosote-contaminated soil samples.
CONCLOSIONS
Recognizing that the data are limited because many elemental concentrations are
below analytical detection limits, the results to date suggest that (1) the TCLP
1-312
-------
shows good reproducibility for inorganic analytes; (2) based on TCLP
characterization of their inorganic constituents, the samples have remained
stable; and (3) based on the quantity of sample used in the TCLP and the
inorganic data available, there is interbottle homogeneity among samples of the
metal plating waste, creosote-contaminated soil, and smelter dust. However, the
data indicate that the API separator sludge samples are poorly homogenized. The
metal plating waste and creosote-contaminated soil could have been better
homogenized prior to sample preparation. This is indicated by the
concentrations of copper determined in the metal plating waste leachates and the
iron concentrations in the creosote-contaminated soil TCLP leachates. The TCLP
data for iron and arsenic in the smelter dust samples are too limited to base
any conclusions concerning sample stability, homogeneity, or TCLP
reproducibility. However, they do emphasize the need for further testing to
determine why the 10-month values vary so much from those determined at 16
months.
Table 1. Metal Plating Waste TCLP Leaohate Analysis Time Ze
Elemental Concentrations,1 mg/L
Duplicate Ag As Ba Cd Cr Pb Se Cu Mg2 Ni
Leachates
1 <.10 <1.0 0.22 <.l
2 <.10 <1.0 0.21 <.l
Ave -^.10 <1.0 0.21 <.l
% Std Dev 4%
<.50 <1.0 73.0 0.64
<.50 <1.0 55.0 0.62
<.50 <1.0 64.0 0.63
8% 4%
Method BIJc ^.10 <1.0 0.06 <. 10 ^.10 ^.50 <1.0 0.9 -i.10
Table 2. Metal Plating Wasto TCLP Leacbate Analysis Six-Month Storage Time
Elemental Concentrations,1 mg/L
Leachate
A
B
C
D
E
Ave
% Std Dev
Method Blk
Ag
<.10
<.10
<.10
<.10
<.10
<.10
<.10
As
<1.0
<1.0
<1.0
<1.0
<1.0
<1 .0
<1.0
Ba
0.22
0.12
0.13
0.13
0.16
0.15
27%
<.10 •
Cd
.10
.10
.10
.10
.10
.10
C.10
cr
<.10
<.10
<.10
<.10
<.10
<.10
<.10
Pb
<.50
<.50
<.50
<.50
<.50
<.50
<.50
Se
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
Cu
17.8
54.5
50.1
58.0
23.6
40.8
46%
<.10
Mg
119
116
120
120
121
119
2%
<.05
Ni
0.45
0.59
0.64
0.50
0.59
0.55
14%
<.10
Table 3. Metal Plating Waste TCLP Leachate Analysis 18-Month Storage Time
Elemental Concentrations, mg/L
Leachate Ag As3 Ba Cd Cr Pb Se3 Cu Mg Ni Hg4
A <.10
B "£ . 10
C *C . 10
D < . 10
E <.10
Ave <.10
% Std Dev — — —
Meth Elk <.10
<.05
<.05
<.05
<.05
<.05
<.05
<.05
0.13
<.10
0.12
0.16
<.10
<.10
<.10
<.10
<.10
<.10
<.10
<.10
<.10
<.10
<-10
<.10
<.10
<.10
<.10
<.10
<.50
<.50
<.50
<.50
<.50
<.50
<.50
<.05
<.05
<.05
<.05
<.05
<.05
<.05
40
59
64
30
42
47
<.':
.9
.0
.9
.7
.0
.5
10
112
109
105
106
109
108
3%
<.05
0.48
0.49
0.42
0.45
0.43
0.45
7%
<.0004
<.0004
<.0004
<.0004
<.0004
<.0004
<..0004
1 Determined by ICP spectroscopy unless otherwise noted
2 No data available
3 Determined by furnace AAS
4 Determined by cold vapor AAS
1-313
-------
Table 4. Metal Plating Waste TCLP Leachate Analysis 24-«onth Storage Time
Elemental Concentrations, mg/L
Lo&chata Ag
A <.10
B <. 10
C <.10
D <.10
E <.10
Ava < . 10
Meth BIX <;.10
AS2
^.10
<.10
<.10
<.10
<.10
<.10
^.10
Ba
0.16
<.10
<.10
<.10
<.10
—
^.10
Cd
<.10
<.10
<.10
<.10
<.10
<.10
<.10
cr
<.10
<.10
<.10
<.10
<.10
<.10
<.10
Pb
"J.50
<;.50
<.50
<:.50
<.50
<.50
<.so
Sa2
<.05
<.05
<.05
<.05
<.05
<.05
<.05
Cu
56.3
52.6
23.7
66.6
57.7
50.8
32%
Mg
112
114
112
111
115
113
1%
<.05
Ni
0.61
0.57
0.43
0.57
0.57
0.55
13%
E
<.
<.
<,
<,
<,
"••
^'
ig3
002
.002
,002
,002
,002
.002
,002
Table 5. copper Concentrations in Matal Plating Waste TCLP Leachatas, mg/L
Time Zero Duplicate Values: 73.0, 55.0
.-17.8
After Six Months 54.5 Ave: 40.8
of Storage 50.1 % std Dev: 46%
58.0 90% CI4: 23.0-58.6
L23.6
p40.9
After 18 Months 59.0 Ave: 47.5
of Storage 64.9 * std Dev: 30%
30.7 90% CI: 34.1-60.y
'-42.0
.- 56.3
After 24 Months 52.6 Ave: 50.8
of storage 23.7 % std Dev: 32%
66.6 90% CI: 35.5-66.1
L 57.7
Overall Ave: 46.4
Overall % std Dev: 34%
Overall 90% CI: 37.7-55.1
Table 6. Magnesium Concentrations in Metal Plating Waste TCLP Leaohates, mg/L
ot determined.
. 119
After Six Months I 116 Ave: 119
of Storage j 120 t Std Dev: 2%
120 90% CI4: 117-121
• 121
. 112
After 18 Months I 109 Ave! 108
of Storage | 105 % Std Dev: 3%
106 90% CI: 105-111
• 109
. 112
After 24 Months I 114 Ava: 113
of Storage | 112 % Std Dev: 1%
111 90% CI: 111-115
• 115
Overall Ave: 113
Overall % std Dev: 4%
Overall 90% CI: 110-116
2 Determined by furnace AAS
Determined by cold vapor AAS
4 90% confidence interval determined using t distribution
1-314
-------
Table 7. Nickel Concentrations in Metal Plating Haste TCLP Leaohates, mg/L
Time Zero Duplicate Values: 0.64, 0.62
.-0.45
After Six Months 0.59 Ave: 0.55
of Storage 0.64 % Std Dev: 14%
0.50 90% CI1: 0.48-0.63
"-0.59
p.0.48
After 18 Months 0.49 Ave: 0.45
of Storage 0.42 % Std Dev: 7%
0.45 90% Oil 0.42-0.48
*-0.43
_0.61
After 24 Months 0.57 Ave: 0.55
of Storage 0.43 % Std Devi 13%
0.57 90% 01: 0.48-0.62
1-0.57
Overall Ave: 0.52
Overall % Std Dev: 14%
Overall 90% CI: 0.48-0.52
Table 8. API Separator Sludge TCLP Leaohate Analysis Five-Month Storage Time
Elemental Concentrations,2 mg/L
Leachate Ag As Ba Cd Cr Pb Se Fe Mg Hi
A <.10 <1.0 0.46 <.10 <.10 <.50 <1.0 175 14.1 <. 10
B <.10 <1.0 0.51 <.10 <.10 <.50 <1.0 188 15.0 <. 10
C <.10 <1.0 0.50 <.10 <.10 <.50 <1.0 219 14.5 <. 10
D <.10 <1.0 0.49 <.10 <.10 <.50 <1.0 211 14.3 <.10
E <-10 <1.0 0.46 <.10 <-10 <.50 <1.0 175 15.4 <.10
Ave <.10 <1.0 0.48 <.10 <.10 <.50 <1.0 194 14.7 <.10
% Std Dev 5% 10% 4%
Method Blk <.10 <1.0 <.10 <.10 <.10 <.50 <1.0 0.15 <.05 <.10
Table 9. API Separator Sludge TCLP Leachate Analysis 17-Month Storage Time
Leachate
A
B
C
D
E
Ave
% Std Dev
Meth Blk
Element;
al Concentrations, mg/L
Ag As3
<. 10 <.
<.10 <.
<.10 <-
<.10 <.
<.10 <.
<.10 <.
<"o <.
05
05
05
05
05
05
05
0
0
0
0
0
0
<
Ba Cd Cr
.12 <.
.25 <.
.52 <.
.66 <.
.33 <.
.38 <.
10 <.10 <
10 0.11 <
10 <.10 <
10 <.10 <
10 <.10 <
10 <
Pb
.50
.50
.50
.50
.50
.50
.50
Se3
<,
<,
<,
<.
<,
<,
^
,05
,05
,05
.05
,05
.05
.05
Fe
220
334
263
202
188
241
24%
0.19
Mg
14.1
18.4
12.2
24.1
15.6
16.9
<.05
Hi
<.10 <,
<.10 <,
<.10 <.
<.10 <,
<.10 <,
<.10 <
<.10 <
Hg4
,002
,002
,002
.002
.002
.002
.002
Table 10. API Separator Sludge TCLP Leachate Analysis 23-Month Storage Time
Elemental Concentrations,2 mg/L
Leachate
A
B
C
D
E
Ave
% Std Dev
Meth Blk
Ag As3
<.10 <.10 0
<. 10 <. 10 0
<. 10 <. 10 0
<. 10 <. 10 0
<.10 <.10 0
<.10 <.10 0
Ba
.55
.69
.86
.12
.69
.58
Cd Cr
< , 10 < . 10 <
•< . 10 ^ . 10 •<
<.10 <.10 <
<.10 <.10 <
<.10 <.10 <
< . 10 < . 10 •*
Pb
.50
.50
.50
.50
.50
.50
<
<
<
<
<
<
Se3
.05
.05
.05
.05
.05
.05
48%
<.10 <.10 <
.10
<.10 <.10 <
.50
<
.05
Fe
162
203
369
175
270
236
36%
<.10
12
13
24
19
13
16
Mg
.7
.8
.0
.1
.3
.6
Hi
H,4
<.10 <.002
< . 10 •<
<.10 <
<.10 <
<.10 <
<.10 <
.002
.002
.002
.002
.002
29%
<.
05
•< . 10 <
.002
^ 90% confidence interval determined using t distribution
2 Determined by ICP spectroscopy unless otherwise noted
3 Determined by furnace AAS
4 Determined by cold vapor AAS
1-315
-------
Table 11. Barium Concentrations in API Separator Sludge TCLP Leachates, mg/L
_0.46
After Five Months 0.51 Ave: 0.48
of Storage I 0.50 % std Dev: 5%
0.49 90% CI1: 0.46-0.50
1-0.46
[0.12
0.25 Ave: 0.38
0.52 % Std Dev: 57%
0.66 90% CI: 0.18-0.58
0.33
_0.55
After 23 Months 0.69 Ave: 0.58
of storage 0.86 % Std D«v: 48%
0.12 90% CI: 0.31-0.85
1-0.69
Overall Ave: 0.48
overall % std Dev: 43%
Overall 90% CI: 0.37-0.59
Table 12. Iron Concentrations in API Separator Sludge TCLP Leachates, mg/L
. 175
After rive Months I 188 Ave: 194
of Storage | 219 % Std Dev: 10%
211 90% CI1: 175-213
• 175
- 220
After 17 Months I 334 Ave: 241
of Storage | 263 % std Dev: 24%
202 90% CI: 185-297
• 188
. 162
After 23 Months I 203 Ave: 236
of Storage | 369 % std Dev: 36%
175 90% CI: 155-317
- 270
Overall Ave: 224
Overall % Std Dev: 27%
overall 90% CI: 191-257
Table 13. Magnesium Concentrations in API Separator Sludge TCLP Leachates, mg/L
.14.1
After rive Months I 15.0 Ave: 14./
of Storage | 14.5 % Std Dev: 4%
14.3 90% CI1: 14.2-15.z
• 15.4
.14.1
After 17 Months I 18.4 AVe: 16.y
of Storage | 12 . 2 % Std Dev: 27%
24.1 90% CI: 12.5-21.J
• 15.6
. 12.7
After 23 Months I 13.8 Ave: 16.6
of Storage | 24.0 % Std Dev: 29%
19.1 90% CI: 12.0-21.^
' 13.3
Overall Ave: 16.0
Overall % std Dev: 23%
Overall 90% CI: 13.9-18.1
1 90% confidence interval determined using t distribution
1-316
-------
Table 14. Creosote-Contaminated Soil TCLP Leaohate Analysis Tlma Zero
Elemental Concentrations,1 mg/L
Leachata Ag As Ba cd Cr Pb So Fa2 Mg2 Hi Hg
1 <.10 <1.0 1.16 <.10 <.10 <.50 <1.0 0.15 <.20
Method Blk <.10 <1.0 <.10 <.10 <.10 <.50 <1.0 <.10 ^.20
Table 15. Creosota-Contaminated Soil TCLP Laaohate Analysis 4.5-Month storaga Tims
Elemental Concentrations,1 mg/L
L4
>achate
A
B
C
D
E
Ave
Method Blk
Ag As
<.10 <1.0
<.10 <1.0
<.10 <1.0
<.10 <1.0
<.10 <1.0
<.10 <1.0
•c.10 <1.0
Ba
1.22
1.26
1.27
1.19
1.22
1.23
Cd
<.10
<.10
<.10
<.10
<.10
<.10
<.10
Cr Pb
<.10 <..50
<.10 <.50
<.10 <.50
<.10 <.50
<.10 <.50
<.10 -,.50
<.10 ^.50
Se Fa2
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
-------
18. Barium Concentrations in Creosote-Contaminated Soil TCLP Leachates, mg/L
Tim« Zero Single Value: JL . 16
1.22
After 4.5 Months 1.26 Aval 1.23
of Storage 1.21 * Std Dev.- 3%
1.19 90% CI1: 1.20-1.26
L 1.22
[1.37
1.25 AveI 1.29
1.32 % Std Dev: 5*
1.22 90* CI: 1.23-1.35
1.27
[1.27
1.29 Ave: 1.23
1.15 % Std Dov: 6*
1.26 90% CI! 1.16-1.30
1.15
Overall Ave: 1.26
Overall % Std Dev: 5%
Overall 90% CI: 1.23-1.29
Table 19. Iron Concentrations in Creosote-Contaminated Soil TCLP Leachates, mg/L
Time zero and t.5-Month data were not determined.
r 1.06
After 17 Months 1.54 Ave: 1.18
of Storage I 1.50 % Std Dev: 26%
0.82 90% CI1: 0.87-1.49
1-0.96
-0.64
After 23 Months 0.65 Ave: 0.75
of Storage 0.72 % Std Dev: 10%
0.78 90% CI: 0.68-0.82
L-0.78
Overall Ave: 0.96
Overall % Std Dev: 32%
Overall 90% CI: 0.75-1.17
Table 20. Magnesium Concentrations in Creosote-Contaminated Soil TCLP Leaohates, mg/L
Time zero data were not determined.
r 20.9
After 4.5 Months 20.9 Ave: 20.0
of Storage 19.6 % Std Dev: 3%
19.7 90* CI1: 19.4-20.0
"- 19.4
[21.4
21.6 Ave:21.6
21.6 % Std Dev: 1%
21.5 90* CI: 21.4-21.8
21.4
[20.2
22.3 Ave: 21.5
20.9 * Std Dev: 4%
21.8 90% CI: 20.6-22.4
22.2
Overall Ave: 21.u
Overall * Std Dev: 4*
Overall 90* CI: 20.5-21.5
1 90* confidence interval determined using t distribution
1-318
-------
Tabl* 21. Smaltor Dust TCLP Loaohato Analysis Time Zero
Duplicate Ag
Leaohates
Elemental Concentrations,1 rag/L
As Ba Cd Cr Pb
Mg
1 0.18 2.70 0.21 7.12 0.15 4.66 <1.0 1580 317 74.9 1.78
2 0.15 1.49 0.20 6.62 <.10 4.57 <1.0 1661 296 66.3 1.82
Ave 0.17 2.09 0.20 6.87 --- 4.61 <1.0 1620 306 71.6 1.80
% Std Dev 11% 41% 3% 5% --- 1% --- 3% 5% 6% 2%
Method Blfc <.10 <1.0 <.10 <.10 <.10 <,50
-------
Table 25. Cadmium Concentrations in Sm«lt«r Dua~t TCLP Leachates, mg/L
Tljae Zero Duplicate Values: 7.12, 0.62
.-6.30
After 10 Months 6.00 Ave: 5.86
of Storage 4.98 * Std Dev: 9%
6.17 90% CI1: 5.36-6.36
L5.87
.-5.67
After 16 Months 5.48 Ave: 5.57
of Storage 5.20 % Std Dev: 6*
5.38 90% CI: 5.24-5.90
1-6.12
Overall Ave: 5.72
Overall % Std Dev: 8%
Overall 90% CI: 5.42-6.02
Table 26. copper Concentrations in Smelter Dust TCLP Leachates, mg/L
Time Zero Duplicate Valuesi 1580, 1661
._ 1850
After 10 Months 1890 Ave: 1778
of storage 1600 % Std Dev: 7%
1720 90% CI1: 1666-1890
L 1830
p 1760
After 16 Months 1750 Ave: 1663
of Storage 1570 % Std Dev: 5%
1593 90% CI: 1579-1747
L- 1641
Overall Ave: 1720
Overall % Std Dev: 7%
Overall 90% CI: 1642-1798
Table 27. Iron Concentrations in Smelter Dust TCLP Leachates, rog/L
Time Zero Duplicate Values: 317, 296
296
After 10 Months I 265 Ave: 269
of Storage 219 % Std Dev: 11%
292 90% CI1: 240-298
L 275
rm
After 16 Months 167 Ave: 184
of Storage 158 % Std Dev: 15%
191 90% CI: 158-210
L 229
Overall Ave: 227
Overall % Std Dev: 23%
Overall 90% CI: 192-262
90* confidence Interval determined using t distribution
1-320
-------
Table 28. Lead Concentrations in Smelter Dust TCLP Leachates, mg/L
Tijne Zero Duplicate Values: 9.66, 4.57
After 10 Months
of storage
After 16 Months
of Storage
.5.94
6.44 Ave: 5.85
5.17 % Std Devi 9%
5.48 90% CI1: 5.35-6.35
• 6.23
- 6.70
6.75 Avel 6.12
6.43 % Std Dev: 12%
5.65 90* Oil 5.43-6.81
- 5.09
Overall Ave: 5.99
Overall % Std Dev: 10%
Overall 90% CI: 5.57-6.40
Table 29. Arsenic concentrations in smelter Dust TCLP Leachates, mg/L
Time Zero
Duplicate Values i
t
After 10 Months
of
Storage
After 16 Months
of
Storage
2
-2
2
2
2
-1
- 0
0
0
0
-o
70,
50
50
10
30
90
65
63
58
72
90
1.49
Ave
% Std Dev
90% CI1
Ave
% Std Dev
90% CI
2.26
9%
2.06-2
0.70
18%
0.58-0
46
82
Overall Ave: 1.48
Overall % Std Dev: 57%
Overall 90% CI: 0.91-2.05
Table 30. Magnesium Concentrations in Smelter Dust TCLP Leachatos, mg/L
Time Zero Duplicate Values :
After 10 Months
of Storage
After 16 Months
of Storage
: 74.9,
r76.1
81.8
65.0
66.9
1-77.9
[74.3
75.0
71.5
68.2
67.2
63.3
Ave:
% Std Dev:
90% CI1:
Ave:
% Std Dev:
90% CI!
73.5
10%
66.6-80.
71.2
5%
67.9-74
.4
.5
Overall Ave: 72.4
Overall % Std Dev: 8%
Overall 90% CI: 68.7-76.1
1 90% confidence Interval determined using t distribution
1-321
-------
Table 31. Nickel Concentrations in smelter Dust TCLP Leachates, mg/L
Time Zero Duplicate Values: 1.78, 1.82
p 1.96
After 10 Months 2.02 Ave: 1.86
of Storage 1.65 % Std Dev: 8%
1.84 90% CI1: 1.73-1.99
•- 1.84
p 1.73
After 16 Months 1.70 Ave: 1.68
of Storage 1.76 % Std Dev: 5%
1.55 90% CI: 1.60-1.76
'-1.64
Overall Ave: 1.77
Overall % Std Dev: 8%
Overall 90% CI: 1.67-1.87
Table 32. Selenium Concentrations in Smelter Dust TCLP Leachates, rog/L
Tljne Zero Duplicate Values!
-------
MODIFICATION OF THE TCLP PROCEDURE TO
ACCOMMODATE MONOLITHIC WASTES
Larry I. Bone. Dow Chemical Company, 3867 Plaza Tower Drive, Baton Rouge,
Louisiana 70816; Mark Bricka, USAE Waterways Exp. Station, P. O. Box 631
WESEE, Vicksburg, Mississippi 39180; Peter Hannak, Canviro Consultants, 180 King
Street S, Suite 600, Waterloo, Ontario N231P8; Sunil I. Shah, Union Carbide
Corporation, P. O. Box 8361, Building 2000/3424, South Charleston, West Virginia
25303; Neil Prange, Monsanto, 800 N. Lindbergh, St. Louis, Missouri 63167; Paul J.
Marsden, S-Cubed, Box 1620, La Jolla, California 92038-1620; 3. E. Waggener,
Resource Consultants, Box 1848, Brentwood, TN 37027; Marvin Miller, Earl
Johnson, Dow Chemical Co., Building 1261, Midland, MI 48667; and Steve 3.
Robuck, ALCOA, Alcoa Center, Pennsylvania 15069.
ABSTRACT
The current EPA Toxicity Characteristic Leaching Procedure (TCLP), SW 846
Method 131l(D, requires that all wastes be milled into smaller particles prior to
being placed into the extractor. A multi-laboratory study has been conducted to
determine if milling is necessary or if the tumbling action of the TCLP procedure
will cause monolithic wastes, which are not strong enough to survive in environ-
ment, to fall apart in the extractor. Multiple plugs of eight different solidified
waste samples of variable strength were prepared. The unconfined compressive
strength of a plug of each sample was measured and the environmental survivabil-
ity was evaluted in two separate laboratories using the ASTM Freeze/Thaw
(D4842-89) and Wet/Dry (D4843-88) testing procedures. Identical plugs of each
sample were tumbled in glass bottles, equipped with a stainless cage as proposed by
EPA.(2) Additional identical plugs were tumbled in unbreakable plastic bottles
following the same protocol used with the stainless steel cage. Five separate
laboratories performed the stainless steel cage experiments in duplicate while
three laboratories performed the plastic cage tumble; also in duplicate. No
chemical analysis was performed; each laboratory simply weighed the amount of
material which would not pass through a 9.5 mm sieve after the 18-hour tumble
was complete. This weight compared to the initial plug weight was compared with
the freeze/thaw, wet/dry and unconfined compressive strength results.
All samples that had low strength and failed the wet/dry tests fell apart when
tumbled in either type of container. Comparison of the cage and plastic bottle
results clearly showed the plastic bottle to be superior. The results better matched
the strength data, the data was more reproducible on both an intra- and
inter-laboratory basis, and a plastic bottle is easier to clean. ASTM D-34 is
drafting a leaching procedure similar to the TCLP which will use a nonbreakable
bottle and will not require particle size reduction of the waste prior to tumbling in
the extractor.
INTRODUCTION
Section 3001 of the Resource Conservation and Recovery Act (RCRA) has charged
the U.S. Environmental Protection Agency (EPA) with the task of developing
1-323
-------
methodology for identifying wastes which may pose a hazard to human health and
the environment. EPA accomplishes this task by identifying RCRA regulated
wastes in two ways. They have developed four lists of wastes from specific and
non-specific sources which must be managed as hazardous by virtue of their listing.
Wastes must also be managed as hazardous if they meet one or more of four
characteristics. The testing procedure required for one of these characteristics,
extraction procedure toxicity, is the subject of this paper.
When RCRA was initially promulgated a procedure called the Extraction Procedure
Toxicity (EP) was the required procedure of testing the characteristic of a waste to
leach toxic constituents at hazardous concentrations. This procedure required that
the waste be leached in an acetate solution (essentially at pH 5) for 24 hours.
Following leaching the resulting solution was tested for fourteen components to
determine if it was to be managed as a RCRA hazardous waste. The test further
required a Structural Integrity Procedure (SIP) to see if the wastes were of
sufficient strength to remain intact in the environment. If they failed they were to
be milled to pass through a 9.5 mm screen prior to being leached.
Subsequent to the EP procedure, the EPA developed a new leaching procedure, the
Toxicity Characteristic Leaching Procedure (TCLP), designed to facilitate
analyzing leachate for a more extensive list of components. To date the TCLP has
only been promulgated as a part of the Land Disposal Restriction Ruled), but not
for its original purpose, i.e., characteristic testing. Since its promulgation for
Land Disposal Restrictions, EPA has proposed^) some modifications to the test.
One of the modifications to the TCLP proposed by EPA involved eliminating
the SIP and the resulting requirement to grind wastes prior to TCLP tumbling. The
modified procedure would simply allow wastes to be tumbled in the extraction
vessel without particle size reduction. It should be noted that wastes which were
to be tested for volatiles in the Zero Headspace Extractor (ZHE) still would require
particle size reduction since the method of leachate expression in the consolidated
device could not accommodate large residual pieces of material. The simplifying
procedure which EPA proposed for wastes tumbled in a bottle was based on an EPA
contractor study by Phillips and Marsden(3' showing that wastes which were not of
sufficient strength to survive in the environment also fell apart during the
tumbling. Although the reverse was not always true, this result is conservative in
that wastes of insufficient strength are tested in a finely divided state while many
wastes which are sufficiently strong to remain intact in the environment are
leached more nearly as they would be in a landfill containing the waste.
The work of Phillips and Marsden used a stainless steel cage placed inside a glass
bottle to keep the bottle from breaking as the monolithic wastes were tumbled. A
subcommittee of ASTM committee D-34 on Waste Disposal reviewed the EPA
proposed modification and agreed to conduct a more extensive multi-laboratory
evaluation of the "cage" modification. During the review of the Phillips and
Marsden results, it was noted that tumbling in the stainless steel cage seemed to be
too severe in that many samples that were of sufficient strength to survive in the
environment were seriously degraded in the cage. Consequently the ASTM D-34
subcommittee decided to test, along with the cage, a tumbling container which
would be less severe; i.e., a plastic bottle. It was deemed that a number of plastic
1-324
-------
materials would be suitable for testing for metals while TFE containers would be
suitable even if the leachate were to be tested for organics. This paper describes
these tests designed to compare the structural integrity of monolithic waste in the
environment to their ability to survive tumbling in plastic bottles or cage modified
glass bottles. If the comparison is good and the procedure is reproducible, neither
the structural integrity testing nor particle size reduction should be required for
TCLP testing. This will in a sense give high strength materials credit for their
strength. It will eliminate the rather foolish practice of solidifying or encapsulat-
ing wastes to reduce their leachability and then testing the effectiveness of the
procedure on the solidified waste after it has been milled.
EXPERIMENTAL PROCEDURE
Multiple plugs of six waste/solidification agent mixture were prepared at the
Waterways Experiment Station for use in this study. Wastes used were RCRA
listed waste F006 (spent solvent) and bag house dust. The wastes were solidified
with either cement, cement and fly ash mixtures, kiln dust and fly ash mixtures or
lime and fly ash mixture. Each waste/agent combination was solidified such that
plugs of widely varying strength resulted; i.e., both high and low strength plugs
were prepared. Table I identifies the waste/solidification agent used for each of
the eight plug types tested.
Unconfined compressive strength tests were run in quadruplicate by Waterways
Experiment Station on each type of plug following ASTM D309. The results are
shown in Table 2.
Freeze/Thaw (ASTM D4842-89) and Wet/Dry (ASTM 4843-88) tests were run in
triplicate (as prescribed by the method) by Waterways Experiment Station and
Alberta Environmental Centre (Peter Hannak). The results of these studies are
shown in Table 3. The table shows the total percent of mass lost in twelve cycles
of the tests and the number of cycles which the waste passed (i.e., specimen has
lost less than 30 percent of its initial weight). The results show that the wet/dry
data correlates best with unconfined compressive strength. The best correlation
between both tests and unconfined compressive strength seems to be the number of
cycles passed rather than the cumulative mass lost.
Identical plugs of each of the solidified wastes were tumbled in a glass bottle
containing a stainless steel cage following the EPA proposed procedure/2) Since
the purpose of this study was only to compare the survivability of the sample to the
tumble compared to its predicted survivability in a landfill, no chemical analysis
was performed on the leachate. After the tumble was complete each laboratory
simply weighed the amount of sample which would not pass through a 9.5 mm sieve.
Experiments were conducted in duplicate at five laboratories; Waterways
Experiment Station, Alberta Environmental Centre, S-Cubed, Alcoa and Resource
Consultants. The results of this data are shown in Table 4.
Much of the inter-laboratory variability in the cage tumbling can be attributed to
the different size of bottles used. All laboratories used cages of the same
diameter (2.5"); the Waterways cage was 6.2" high, the Alberta cage was 8.0" high
and all of the others were 10.5" high. However, Waterways and Alberta used large
1-325
-------
enough bottles to allow the entire 150 g specimen to be tumbled. The other
laboratories were forced to break the plug to get a small enough sample to allow a
20:1 dilution. Breaking the plugs could have weakened them leading to a greater
weight loss when the smaller bottle was used. The larger plug size to cage size
ratio resulting from the use of a larger bottle somewhat restricted the tumbling
which could also have biased the results in the direction observed.
Identical plugs of each of the solidified wastes were also tumbled in two liter wide
mouth plastic bottles following the same procedures used in the cage tumble. All
laboratories had to break the plugs to obtain a small enough sample to allow for the
20:1 dilution. Experiments were conducted in duplicate at three laboratories; Dow,
Monsanto and Union Carbide. The results are shown in Table 5.
CONCLUSIONS AND SUMMARY
Tables 6 and 7 compare the unconfined compressive strength, freeze/thaw and
wet/dry results with the percent of speciment retained by a 9.5 mm sieve after
tumbling in the cage (Table 6) and the plastic bottle (Table 7). The survivability in
both cage and the plastic bottle matched the unconfined compressive strength
reasonably well, although the match with the plastic bottle was better. The
tumbling results did not match the freeze/thaw and wet/dry results as well, but it
is important to note that, with the possible exception of waste E, plugs which do
not have enough structural integrity to survive in the environment also did not
survive the tumble in the TCLP bottle. Consequently it should not be necessary to
grind any wastes prior to running this TCLP procedure. Conversely, some of the
wastes which are probably strong enough to survive in the environment did not
survive the tumble. However, the plastic bottle appears to be less severe and the
results match the strength and environmental survivability better. This is
particularly true if the plastic bottle results (where tumbling is not restricted) are
compared to the cage results where the tumbling was not restricted by the cage;
i.e., the results from Labs 6, 7 and 8.
The results of this study indicate that the plastic bottle is superior to a stainless
steel cage in a glass bottle for testing the leachability of monolithic wastes which
have not been subjected to particle size reduction. The results are less scattered
and they match the strength data better. The plastic bottles are much easier to
clean than the cage and are, of course, unbreakable. Even with use of the cage
several glass bottles were broken during the course of this study.
These results have been reviewed by a subcommittee of ASTM Committee D-34,
leading to the conclusion to ballot a TCLP procedure which would-not require
particle size reduction, strength testing or environmental structural integrity
testing. All wastes would simply be tumbled in a plastic bottle in the structural
form in which they are to be disposed. The procedure would otherwise be identical
to EPA Method 1311. This method would not address TCLP testing for volatiles.
1-326
-------
REFERENCES
1. Federal Register, November 7, 1986, 51., 216, 40572ff.
2. Federal Register, May 2*, 1988, 53, 18792.
3. Phillips, R. B. and Marsden, P. J., S-Cubed, "Modification of TCLP to
Accommodate Solidified Wastes", U.S. EPA Contract No. 68-03-1958, 1987.
1-327
-------
TABLE 1
SOLIDIFIED WASTE SAMPLE IDENTIFICATION
Solidification Relative Percent
Code Waste Agent Strength Moisture
A F006 Cement High 34
B F006 Cement Low 47
C Bag Cement High 10
House Dust
D Bag Kiln dust/fly ash Low 17
House Dust
E F006 Lime/fly ash High 35
F F006 Lime/fly ash Low 35
G F006 Cement/fly ash High 35
H F006 Cement/fly ash Low 32
1-328
-------
TABLE 2
SOLIDIFIED WASTE UNCONFINED COMPRESSIVE STRENGTH
Waste Age Avg. UCS*
Code (days) (psi)
A 7.00 37.0
A 14.00 62.0
A 20.96 77.0
A 28.04 54.7
A 35.04 71.1
B 7.00 8.1
B 13.96 8.1
B 21.00 14.8
B 27.96 25.3
C 6.92 303.4
C 14.04 334.1
C 20.96 354.9
C 27.96 383.6
C 35.00 468.8
D 6.88 919.7
D 13.96 1,177.1
D 20.88 1,261.9
D 27.88 1,759.2
D 35.00 1,312.3
E 7.04 58.4
E 13.96 413.4
E 20.96 636.8
E 28.00 998.6
F 7.00 6.1
F 13.96 86.5
F 21.00 251.4
F 28.00 299.7
G 6.00 106.3
G 13.04 290.3
G 20.00 411.9
G 26.96 485.3
G 34.04 439.9
H 6.00 27.5
H 13.04 99.5
H 20.00 129.5
H 26.96 172.6
H 33.92 186.6
* Unconfined Compressive Strength, ASTM Method
D309; average of four broken cubes
1-329
-------
TABLE 3
PREDICTION OF WASTE SURVIVABILITY IN THE ENVIRONMENT
FREEZE/THAW:WET/DRY RESULTS
Freeze/Thaw Wet/Dry
Waterways Alberta Env Waterways Alberta Env
Code L C L _C_ L _C_ L _C_
A 95 1.7 100 2 98 8 100 2
B 33 5.7 100 6 M 12 100 5
C 3.2 12 100 4 3.1 12 2.9 12 380
D 2.8 12 21 12 2.6 12 1.7 12 1760
E 100 7.3 100 3 53 12 96 6 1000
F 78 12 100 9 26 12 31 12 300
G 97 12 85 8 25 12 33 12 490
H 100 5 100 3 100 2 100 4 173
L = Percent of the mass loss in the Wet/Dry or Freeze/Thaw test.
C = Number of cycles passed
UCS = Unconfined compressive strength at 28 days
1-330
-------
TABLE 4
SAMPLE SURVIVABILITY IN STAINLESS STEEL CAGE TUMBLE
Code
Laboratory
B
D
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
Waterways
Alberta
S-Cubed
Alcoa
Resource Consultants
% of Specimen
Retained on a
9.5 mm sieve*
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
19.75/21.60
17.41/32.46
24.08/22.28
22.46/22.53
32.44/31.01
44.31/43.71
113.94/115.27
12.02/13.48
10.97/11.23
15.03/14.72
52.85/52.65
95.99/96.61
4.83/5.12
5.98/7.66
5.91/25.76
37.89/38.1
89.34/91.16
4.22/4.19
5.59/5.75
5.56/7.08
1-331
-------
% of Specimen
Retained on a
Code Laboratory 9.5 mm sieve*
~~H~~ Waterways 0.00/0.00
Alberta 0.00/0.00
S-Cubed 0.00/0.00
Alcoa 0.00/0.00
Resource Consultants 0.00/0.00
Samples run in duplicate.
1-332
-------
TABLE 5
SAMPLE SURVIVABILITY IN PLASTIC BOTTLE TUMBLE
Code
B
D
H
Laboratory
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
Dow
Monsanto
Union Carbide
% of Specimen
Retained on a
9.5 mm sieve*
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
0.00/0.00
51.80/52.28
50.10/51.99
40.63/43.62
27.96/29.11
26.70/33.41
26.96/27.28
11.37/13.88
15.87/12.76
8.67/9.96
10.78/7.02
7.12/9.02
5.86/12.33
0.00/0.00
0.00/0.00
0.00/0.00
Samples run in duplicate.
1-333
-------
TABLE 6
SURVIVABILITY OF SAMPLES OF SOLIDIFIED WASTE
STRENGTH VS. CAGE TUMBLE
Freeze/Thaw Wet/Dry Unconfined % of Specimen Retained on a
Sample
No.
A-h
B-l
C-l
D-h
E-h
F-l
G-h
H-l
Cycles
Passed*
2
6
8
12
5
11
10
4
Cycles
Passed*
5
9
12
12
9
12
12
3
Comp
Strength
54
25
470
1800
1000
300
490
173
9.5 mm Sieve**
Lab 4
0.00
0.00
0.00
20.7
44.0
52.8
38.0
0.00
Lab 5
0.00
0.00
0.00
24.9
115
96.3
90.2
0.00
Lab 6
0.00
0.00
0.00
23.2
12.8
5.0
4.2
0.00
Lab 7
0.00
0.00
0.00
22.5
11.1
6.8
5.7
0.00
Lab 8
0.00
0.00
0.00
31.8
14.9
15.8
6.3
0.00
h = high strength
i = low strength
* average of two laboratories
** average of duplicate samples
1-334
-------
TABLE 7
SURVIVABILITY OF SAMPLES OF SOLIDIFIED WASTE
STRENGTH VS. PLASTIC BOTTLE TUMBLE
Sample
No.
A-h
B-l
C-l
D-h
E-h
F-l
G-h
H-l
Freeze/Thaw
Cycles
Passed*
2
6
8
12
5
11
10
4
Wet/Dry
Cycles
Passed*
5
9
12
12
9
12
12
3
Unconfined
Comp
Strength
54
25
470
1800
1000
300
490
173
% of Sample Retained
on a 9.5 mm Sieve**
Lab 1
0.00
0.00
0.00
52.0
28.5
12.6
8.9
0.00
Lab 2
0.00
0.00
0.00
51.0
30.1
14.3
8.1
0.00
Lab 3
0.00
0.00
0.00
42.1
27.1
9.3
9.1
0.00
h = high strength
1 = low strength
* average of two laboratories
** average of duplicate samples
1-335
-------
Precision and Ruggedness Evaluation of Method 1312. D. Miller, P. Marsden
TECHNICAL PAPER UNAVAILABLE
1-336
-------
THE PACIFIC BASIN CONSORTIUM FOR HAZARDOUS WASTE RESEARCH
HAZARDOUS MATERIALS LEACHATE DATABASE
Eugene A. Burns. Division Vice President, Larry E. Michalec, Scientist, Environmental
Technology Group, S-CUBED Division of Maxwell Laboratories, Inc., 3398 Carmel
Mountain Road, San Diego, California 92121; Gail A. Hansen, Chemist, Office of Solid
Waste, U.S. Environmental Protection Agency, 401 M Street, S.W., Washington, D.C.
20460
BACKGROUND
The countries of the Pacific Basin are in various stages of industrial development—from
well-developed industrial countries, to developing countries, and countries with
essentially no development. As the industrial revolution continues or begins in these
countries, clear-cut actions will be necessary to contain the hazardous wastes generated as
by-products of this industrial growth. The Pacific Basin countries will be experiencing
the same kinds of issues for containment of hazardous waste as are and were experienced
in the more industrialized countries. Groundwater contamination will become an ever
larger issue as development in the area continues, especially because significant amounts
of drinking water in all of the countries comes from groundwater.
Most of the industrialized countries have been disposing of industrial wastes in landfills.
Much of the information and technology needed for assuring that hazardous constituents
are not leached from these wastes and enter into the groundwater has been determined by
the more advanced countries. In some cases, the waste under question may be unique to
the Pacific Basin country and there would be no prior experience on which to base
corrective actions.
In 1987, scientists and engineers from several organizations established the Pacific Basin
Consortium for Hazardous Waste Research (PBCHWR). The membership of the
PBCHWR now comprises 49 member organizations from 14 different countries
represented, including Australia, Canada, China, Hong Kong, Indonesia, Japan, Korea,
Malaysia, Mexico, New Zealand, Philippines, Taiwan, Thailand, and the United States.
The purpose of the Consortium is to address the special needs of the Pacific area with
respect to hazardous wastes. The objectives of the PBCHWR are as follows:
• Identify and assess hazardous waste management problems common to
the Pacific Basin,
• Foster and, where appropriate and possible, initiate research, engineering
development, and pilot-scale testing of hazardous waste treatment equip-
ment and processes,
1-337
-------
• Create a forum for information exchange and technology transfer, and
• Provide training opportunities for technical personnel.
Other information regarding the Consortium is presented at the end of this paper.
One of the goals of the PBCHWR is to develop a systematic database which (1) contains
prior information, (2) is easily accessible by researchers, and (3) can serve as a starting
point for new definitive research with a specific waste in mind. To that end, the Methods
Development Section of the Office of Solid Waste has assisted the PBCHWR in initiating
the development of a hazardous materials leachate database. The development of a model
for leaching of hazardous constituents from landfills is a goal of both EPA and PBCHWR.
One of the essential steps in establishing such a model is to test the validity of the model
compared to existing experimental results.
Over the years, a considerable amount of experimental information regarding leaching
from landfills and in simulated laboratory studies has been generated. Unfortunately,
much of this information is not readily retrievable. The objective of this project is aimed
at developing a comprehensive leachability database to make the information readily
available for (1) validating future landfill leaching models as well as (2) serving as a
database in solving landfill disposal problems. It is vital to provide some facile means for
the developed countries to share their environmental knowledge with developing
countries. Such sharing will provide to the developing countries the benefit of solutions
to early mistakes discovered by the developed countries.
DATABASE DEVELOPMENT
Many of the early studies were performed for a variety of purposes and incomplete
information was reported. A key element of this project was to establish a recommended
reporting format (1) to aid in documenting and characterizing leaching studies and (2) to
be used in future leachability studies. This reporting format covers details of the
conditions of the experimental studies regarding the nature of the wastes, location and
type of sampling system, soil type, periods of sampling, leachate flow rate, analytes and
constituents measured, temperature of waste and sampling system, analytical chemistry
procedures used, and other pertinent parameters. The name of the study, geographic
location, and researchers are also documented.
Because such information is now currently being generated in both the United States and
in Pacific Basin countries, it is important that this information be reported in a consistent,
uniform, recommended format and included in the database. This project provides an
excellent mechanism for collaborative activities among all Pacific Basin countries.
1-338
-------
The initial candidate reporting format shown in Figure 1 was distributed to EPA con-
tractors, cognizant EPA officials, PBCHWR members, and ASTM Committee D34 on
Waste Disposal participants for their comments and suggestions for modifications/
improvement. A key part of this information transmittal was a request for user interest
and anticipated participation to provide leaching study results for inclusion in the
database.
A separate leaching database was constructed using a dBASEm+ compatible structure on
a MS-DOS-based personal computer. This structure was selected because it is as close to
an industry standard database as can be found. The revised/modified recommended
reporting format was converted to a user-input form for implementation of the database.
Selected leachate studies performed under EPA contract were incorporated into the
database. The initial data loading was used to test the report formats (which were
designed to meet both consortium member's and EPA's needs). Hard copies of the
reports were evaluated by project participants for ease of use, appropriateness, and value.
Attendees at this symposium may participate in a leachate database demonstration.
Interest in the use and future participation by symposium attendees will be solicited.
Future phases of this project consist of transferring database software and input forms to
EPA and Consortium participants that desire to use the database. Participation by many
users will facilitate the development of a comprehensive leachate database. A PC-based
database data access system will be set up in later phases of the project. A bulletin-board
type system will be installed for use by participants to access the database information. A
program will be developed which will automatically call the database system and log each
individual user onto the system. The user may then leave mail to others, upload or
download database information, or chat with users on line. Ultimately, it may be
necessary to transfer the database to a mainframe environment, but that step can be
delayed by adding extra PCs to the system as required by increase in usage.
It is possible that eventually the volume of information will be too great for a personal
computer-based system to handle. If that happens, the mainframe host could then satisfy
the requirements. Development, initial testing, and initial implementation of the concept
is less expensive when performed on a personal computer. The database developed and
the code written will be easily transported to the mainframe system if required. Users will
be able to dial up either the PC- or mainframe-based system from any telephone and
upload and download data as their needs dictate.
This project has established a database which will serve both the EPA and all countries of
the Pacific Basin. The easily available information will help prevent the duplication of
research efforts in the leachate data field by making the results of research already
conducted readily available. The information will also assist the developing countries by
making available the experience of the developed countries which would otherwise be
1-339
-------
HAZARDOUS WASTE LEACHING REPORT INFORMATION
Project Title _
Investigators).
Date
Reference
Institution
Waste Identity
Source
Pre-Test Characterization of Waste
Leaching Configuration
Waste/Leachate Contact Method _
Leaching Fluid: Composition
Testing Duration
Leaching Test Method.
Leachate Flow Rate
Analysis Procedures Used _
Other Information
Soil Type
Leaching Location: Lab U Field Site
Waste Weight
Volume
Sampling Period _
Initial pH _
Test Temp.
Leachate Sampling Method.
&
Leachate (1): Date/Time Sampled
Composition
Leachate (2): Date/Time Sampled_
Composition
Etc.
Leachate Volume
Leachate pH
Leachate Volume
Leachate pH _
5
Post-Test Characterization (Observations) of Waste
Figure 1. Candidate Reporting Format
1-340
-------
unavailable. This is especially important because the environment and its protection must
be high on the list of priorities for developing nations. The information will be available
simply by making a telephone call. The user cost will be little more than the cost of a
telephone call. The use of the personal computer as a smart terminal facilities
downloading and use of the information. Hard copies of the various reports will also be
available to system users who do not have access to a PC.
PACIFIC BASIN CONSORTIUM FOR HAZARDOUS WASTE RESEARCH
Role of the Consortium
Production and use of hazardous materials are increasing rapidly throughout the Pacific
Basin. These hazardous materials include wastes but also many useful, though toxic,
chemicals. The more industrialized Pacific Basin nations have begun to assess their
hazardous waste problems and to improve their site remediation and waste management
practices. Because the basin is being industrialized more rapidly than any other part of the
world, treatment and safe disposal of hazardous wastes are especially acute problems.
Scientists and engineers from several organizations have established the Pacific Basin
Consortium for Hazardous Waste Research. The purpose of the Consortium is to address
the special needs of the Pacific area with respect to hazardous waste. Its objectives are as
follows:
• Identify and assess hazardous waste management problems common to
the Pacific Basin.
• Foster and, where appropriate and possible, initiate research, engineering
development, and pilot-scale testing of hazardous waste treatment
equipment and processes.
• Create a forum for information exchange and technology transfer.
• Provide training opportunities for technical personnel.
The Consortium augments, rather than replaces or duplicates, the hazardous waste
research, development, and training currently conducted by its member organizations.
Because it facilitates cooperative research programs and the sharing of non-proprietary
research results, the Consortium helps speed the performance and reduce the high costs of
hazardous waste research for all its members, including those now at the forefront of this
field.
1-341
-------
The Consortium has the following four areas of activity:
• Information Exchange. The Consortium promotes information exchange
through newsletters, symposia, annual technical conferences, and
workshops.
• Training. The Consortium provides training opportunities by conducting
courses, collaborating in the training activities of other organizations, and
promoting the exchange of staff among member institutions for training
purposes.
• Professional Network. The Consortium maintains contacts with a broad
spectrum of professionals in the hazardous waste research community.
• Collaborative Research. The Consortium promotes collaborative
research on hazardous waste problems in the Pacific Basin. The
collaborative efforts may take the form of staff exchange among member
institutions; coordinated research programs, and special research teams.
Membership in the Consortium is open to organizations located in areas either on the rim
or within the basin of the Pacific Ocean. The Consortium is not an intergovernmental
activity; there is no official governmental representation. Member institutions represent
their own perspectives only. For further information contact:
Richard Cirillo, Executive Secretary
Pacific Basin Consortium for Hazardous Waste Research
c/o East-West Center
Environmental and Policy Institute
1777 East-West Road
Honolulu, Hawaii 96848
(808) 944-7555
1-342
------- |