First Annual
United States
Environmental Protection Agency
Symposium
on
Solid Waste Testing
and
Quality Assurance
PROCEEDINGS
July 23-26, 1985
Washington, D.C.
Vista International Hotel
Symposium managed by
The American Public Works Association
-------
FIRST ANNUAL
UNITED STATES
ENVIRONMENTAL PROTECTION AGENCY
SYMPOSIUM
on
SOLID WASTE TESTING
AND
QUALITY ASSURANCE
PROCEEDINGS
July 23-24, 1985
Vista International Hotel
Washington, D. C.
Symposium Managed By
The American Public Works Association
Proceedings Edited By
William S. Forester
Director
ISW/ISWA Secretariat
American Public Works Association
Chicago, Illinois
-------
FIRST ANNUAL UNITED STATES ENVIRONMENTAL PROTECTION AGENCY SYMPOSIUM
on
SOLID WASTE TESTING AND QUALITY ASSURANCE
PROGRAM
Vista International Hotel Washington, D.C.
July 24—26, 1985
Wednesday, July 24, 1985
7:00 am — 8:30 am REGI STRATION
8:30 am — 8:50 am OPENING SESSION
Opening presentation given by Dr. John H. Skinner,
Director, Office of Solid Waste, U. S. EPA,
Washington, DC 20460
8:50 am — 9:15 am CONFERENCE OVERVIEW
Overview presented by David Friedman, Manager,
Methods Program, Office of Solid Waste, U. S. EPA,
Washington, DC 20460
9:15 am — Noon SESSION I, ANALYSIS OF INORGANICS
Chairperson: Douglas Gillard, Methods Program,
Office of Solid Waste (WH—562B), U. S. EPA,
Washington, DC 20460
9:15 am — 10:15 am “Comparative Performances of Inductively Coupled
Plasma Atomic Emission Spectroscopy (ICP—AES) and
Atomic Absorption Spectroscopy (AM)”
Speaker: Barton P. Simmons, Hazardous Materials
Laboratory, California Department of Health
Services, 2151 Berkeley Way, Berkeley, CA 94720;
and
“Comparative Performance of Inductively Coupled
Plasma Optical Emission Spectroscopy (ICP) and
Atomic Absorption Spectroscopy (AAS)”
Speaker: Thomas A. Hinners, Hazardous Waste Methods
Evaluation Branch, U.S. EPA—EMSL, P.O. Box 15027,
Las Vegas, NV 89114
10:15 am — 10:30 am BREAK
10:30 am — 11:00 am “Evaluation of SW—846 Methods for Metal
-------
Extraction,’1 isolution of Aqueous, Oil, and Solid
Waste Matrices”
Speaker: Dr. Mirtha Umana, Analytical and
Chemical Sciences, Research Triangle Institute, P.
0. Box 12194, Research Triangle Park, NC 27709
11:00 am — 11:30 am “Interference Reduction Studies Involving Hydride
Generation Arsenic and Selenium Determinations
Utilizing Atomic Absorption and Plasma Emission
Spectrometry”
Speaker: Dr. J. Wilson Hershey, Lancaster
Laboratories, Inc., Lancaster, PA 17601
11:30 am — Noon “Employment of Alkaline Digestion Procedures for
Determination of Metals in Industrial Wastes”
Speaker: Joseph Lowry, Chief, Inorganic
Analytical Section, U. S. EPA—NEIC, Box 25227,
Denver Federal Center, Denver, CO 80225
Noon — 1:30 pm LtJNCH BREP K
1:30 pm — 5:00 pm SESSION II, METHODS FOR IDENTIFYING HAZARDOUS WASTE
CHARACTERI STICS
Chairperson: Todd A. Kimmell, Methods Program,
Office of Solid Waste (WH—562B), U.S. EPA,
Washington, DC 20460
1:30 pm — 2:00 pm “Performance of Ignitable Solids Methods for
Characterizing Hazardous Wastes”
Speaker: Florence Richardson, Office of Solid
Waste (WH—562B), U.S. EPA, Washington, DC 20460
2:00 pin — 2:30 pm “Reactive Sulfides and Cyanides: Test Methods and
Regulatory Threshold Setting Models”
Paul H. Friedman, Studies and Methods Branch,.
Office of Solid Waste (WH—562B), U. S. EPA,
Washington, DC 20460
2:30 pin — 3:00 pin “Mobility of Toxic Compounds from Hazardous Wastes:
Comparison of Three Test Methods to a Lysimeter
Model”
Speaker: Michael Maskarinec, Chemistry Division
(Building 1505), Oak Ridge National Laboratory, P.
0. Box 10, Oak Ridge, TN 37830
3:00 pm — 3:30 pm COFFEE BREAK
3:30 pm — 4:00 pm “Application of the Toxicity Characteristic
Leaching Procedure (TCLP) to Industrial Wastes: A
Single Laboratory Evaluation”
Speaker: L. R. Williams, U. S. EPA—EMSL, BOX
15027, Las Vegas, NV 89114
4:00 pm — 5:00 pm PANEL DISCUSSION — Overview of the EPA Program to
Define the Characteristics of Hazardous Wastes
Chairperson: Todd A. Kimmell, OSW, U. S. EPA
Participants: David Friedman, Paul H. Friedman,
-------
Douglas Gillard, and Florence Richardson, 05W, U.
S. EPA
6:00 pm — 7:30 pm RECEPTION — For All Symposium Speakers and
Attendees
Thursday, July 25, 1985
8:00 am - Noon SESSION III, ANALYZING FOR ORGANICS
Chairperson: Dr. Paul H. Friedman, Methods
Program, Office of Solid Waste (WH—502B), U. S.
EPA, Washington, DC 20460
8:00 am — 8:30 am “Application of SW—846 Methods to Groundwater
Monitoring Programs: Experiences of No Contract
Laboratories”
Speaker: Dr. Denis Lin, ETC, 284 Raritan Center
Parkway, N 3154, Edison, NJ 08818
8:30 All — 9:00 am “Quantitive Analytical Screen for the Determination
of the Appendix VIII Hazardous Constituents”
Speaker: Dr. Mark J. Carter, Rocky Mountain
Analytical Laboratory, 5530 Marshall Street,
Arvada, CO 80002
9:00 am — 9:30 am “The Use of SW—846 Cleanup and Mass Spectroscopy
Methods to Identify and Quantify Compounds in
Complex Industrial Wastes: Petroleum Industry Case
Histories”
Speaker: Alice Boomhower, Radia Corporation, 7655
Old Springhouse Road, McLean, VA 22102
9:30 am — 10:00 am “Methodology for the Analysis of Organic Chemicals
in Petroleum Refining Wastes to Support RcRA Waste
Listing and Delisting and Land Treatment
Demonstration Programs”
Speakers: Dr. Mark J. Carter, Dr. Michael P.
Phillips, and Jerry L. Parr, Rocky Mountain
Analytical Laboratory, 5530 Marshall Street,
Arvada, CO 80002
10:00 am — 10:15 am BREAK
10:15 am — 10:45 am “Development of Groundwater Screening Procedures
for Use in Monitoring Programs: Objectives and
Experiment Progress at Battelle Columbus
Laboratories”
Speaker: S. V. Lucas, Battelle Columbus
Laboratories, 505 King Street, Columbus, OH 43201
10:45 am — 11:30 am EPA/INDUSTRY PANEL DISCUSSION — The Need for
Standard Laboratory Procedures and EPA’S “Methods
for Evaluating Solid Wastes” (SW—846)
Chairperson: Dr. Paul H. Friedman, OSW, U. S. EPA
-------
Participants: Session Speakers (see names above);
Al Verstuyft, Chevron Oil Company, 576 Standard
Avenue, Richmond, C 94802; and Ronald Mitchum, U.
S. EPA—EMSL, P. 0. Box 15027, Las Vegas, NV 89114
11:3Oam—l:OOpm LUNCHBRE K
1:00 pm — 4:30 pm SESSION IV, SAMPLING UNDER RCRA
Chairperson: Martin Myers, Methods Program,
Office of Solid Waste (WH—562B), U. S. EPA,
Washington, DC 20460
1:00 pm — 1:30 pm “Volatile Organic Sampling Trains for Hazardous
Waste Incinerators: Laboratory Validation”
Speaker: Thomas Logan, U. S. EPA—EMSL, Research
Triangle Park, NC 27711
1:30 pm — 2:00 pm “Practical Considerations for Improving Sampling
Accuracy at Groundwater Test Wells”
Speaker: Douglas Richardson, Geo—Research, 2001
Wisconsin Avenue, NW, Suite 200, Washington, DC
20007
2:00 pm — 2:30 pm “The Relationship Between the Design of Wells and
Sampling in Compliance Monitoring for Groundwater
Under RCRA”
Speaker: Roy Murphy, U. S. EPA (wH—527),
Washington, DC 20460
2:30 pm — 2:45 pm BREAK
2:45 pm — 3:15 pm “Sampling Techniques for Risk Management: A Dioxin
Case History”
Speaker: Mark Haulenbeek, Affiliated Engineering
Laboratories, 1095 Aniboy Arena, Edison, NJ 08837
3:15 pm — 3:45 pm “Design and Implementation of Sampling Plans for
RCRA Listing and Delisting Programs”
Speaker: John Nancy, Vice President, ERCO, 205
Alewife Brook Parkway, Cambridge, MA 01238
3:45 pm — 4:15 pm “Practical Statistical Considerations in Designing
a Sampling Plan”
Speaker: John Warren, U. S. EPA (PM—223),
Washington, DC 20460
4:15 pm — 4:30 pm DISCUSSION
7:30 pm — 9:30 pm EVENING WORKSHOP — For Federal and State Agency
Attendees
Chairperson: Michael Barclay, Office of Waste
Program Enforcement, U. S. EPA, Washington, DC
20460
7:30 pm — 8:00 pm “Evaluation of Laboratories Performing Analyses for
the RCRA Program”
Speaker: Alexis Taylor, Technology Applications,
-------
5201 Leesburg Pike, Suite 800, Falls Church, VA
22041
8:00 pm — 9:30 pm “Test Method Support for Enforcement Programs:
Current Applications, Problems, and Experiences”
Participants: Kenneth Jennings, c ’iPE, U. S. EPA;
and Roy Murphy, GWTF, U. S. EPA
Friday, July 26, 1985
8:00 am — Noon SESSION V, QUALITY ASSURANCE ISSUES
Chairperson: Florence Richardson, Quality
Assurance Officer, Methods Program, Office of Solid
Waste, U. S. EPA, washington, DC 20460
8:00 am — 8:30 am “Documenting the Equivalency of Proposed Methods to
Approved Test Methods for Evaluating Solid Waste”
Speaker: L. R. Williams, U. S. EPA—EMSL, p. 0. Box
15027, Las Vegas, NV 89114
8:30 am — 9:00 am “Use of Performance Based Quality Control Criteria
in the Superfund Contract Laboratory Program”
Speaker: Gareth Pearson, U. S. EPA—EMSL, P. 0. Box
15027, Las Vegas, NV 89114
9:00 am — 9:30 am “A National Voluntary Laboratory Accreditation
Program for Environmental Measurements”
Speaker: Peter Unger, National Bureau of
Standards, U. S. Department of Commerce, ADNIN
A—531, Gaithersburg, MD 20899
9:30 am — 9:45 am DISCUSSION
9:45 am — 10:00 am BREAK
10:00 am — 10:30 am “Controlling and Coping with Unwanted Variance in
Groundwater Monitoring Data: Quality Control and
Statistics”
Speaker: Burnell Vincent, Groundwater Program,
Office of Solid Waste (WH—565E), U. S. EPA,
Washington, DC 20460
10:30 am — 11:00 am “Sources and Means of Obtaining Compounds for the
Quality Analysis Materials Bank”
Speaker: Edward Kantor, U. S. EPA—EMSL, P. 0. Box
15027, Las Vegas, NV 89114
11:00 am — Noon PANEL DISCUSSION: Analytical Standards and the
Regulatyed Community
Chairperson: John Winter, U. S. EPA—EMSL, 26 West
St. Clair Street, Cincinnati, OH 45268
Participants: Edward Berg, U. S. EPA—EMSL; Michael
Bolgar, Foxborofl nalabs; Henrie Garie, New Jersey
Department of Environmental Protection; and Thomas
-------
Gills, National Bureau of Standards
Noon ADJOURN
-------
FIRST SESSION
ANALYSIS OF INORGANICS
9:15 am — Noon
Wednesday, July 24, 1985
chairperson: Douglas Gillard
Methods Program
Off ice of Solid Waste
U. S. Environmental
Protection Agency
Washington, D. C.
-------
COMPARATIVE PERFORMANCE OF INDUCTIVELY COUPLED PLASMA AWMIC
EMISSION SPECTROSCOPY (ICP-AES) ND AIOMIC ABSORPTION SPECTROSCOPY
(AAS)
BARTON P. SIMMONS, MILAD S. ISKANDER, AND ROBERT D. STEPHENS,
HAZARDOUS MATERIALS LABORATORY, CALIFORNIA DEPARTMENT OF HEALTH
SERVICES, BERKELEY, CALIFORNIA
ABSTRACT
ICP—AES has rapidly emerged as a powerful tool for the elemental
analysis of hazardous wastes. Comparisons of ICP—AES with the
established AAS techniques has provided data on the relative
precision, accuracy and ruggedness of the technique. Comparative
studies have been completed with contaminated soil, waste oil,
baghouse dust from waste solvent incineration, auto shredder
waste, and fly ash. Good agreement was generally found between
ICP—AES, AAS, and X—ray Fluorescence Spectroscopy (XRF).
Exceptions are discussed. The precision of ICP—AES was determined
and compared with the precision of sampling and extraction.
Particular problems for ICP—AES include the effects of variable
chemical composition and physical properties of hazardous waste
samples. Solutions include the prudent choice of emission lines,
use of an internal standard, choice of digestion technique, and
choice of sample introduction technique.
INTRODUCTION
The growth in concern about hazardous waste has produced a massive
response by both government and the private sector. This response
has been directed at improved methods of management of these
materials as they are generated by today’s modern industry as well
as efforts to identify and correct problems caused by improper
practices of the past.
A principal component of both government and private sector
programs to address hazardous waste problems is to characterize
these complex materials sufficiently to allow for assessment of
public health and environmental impacts, to determine appropriate
approaches for assessing the extent of contaminated land, and to
determine the effectiveness and impacts of a wide variety of
treatment technologies. This requirement for characterization in
light of the considerable complexity of waste and environmental
samples places an unusual demand on the analytical laboratory.
New methods must be developed to determine efficiently and
accurately many analytes on one prepared sample. A necessary
requirement of such multi—parameter analysis must be, in addition,
the ability to give reliable results for a wide variety of
matrices.
Several basic technologies currently are available for
-------
multi—elemental analysis. Among those available are automated,
multi—lamp atomic absorption spectroscopy (MLAA); x—ray
fluorescence spectroscopy (both wavelength and energy dispersive)
(XRF); arid inductively coupled plasma emission spectroscopy (ICF).
Each one of these technologies has its characteristic strengths
and weaknesses. This laboratory has developed a considerable data
base utilizing all three technologies on a wide variety of waste
and environmental samples. This has allowed us to make direct
comparisons between the technologies and to make evaluations as to
the appropriate use of each. As a result of this evaluation, our
laboratory has chosen to rely heavily on ICP as the backbone of
our multi—elemental analysis program. For selected elements, when
very low detection limits are required, atomic absorption is
employed.
This paper presents the consideration of, and the procedures for,
analyzing waste and environmental samples by ICP. In addition,
results of comparative studies are presented for three very
different waste types using XRF, MLAA, and ICP.
PREVIOUS SIUDIES
While MS and ICP have been extensively studied, relatively little
has been published on the comparison on the two techniques in
hazardous waste analysis.
MS and ICP have compared favorably for the analysis of 25
elements in fly ash water extracts (Drenski) and eight elements in
Extraction Procedure (EP) extracts (King, 1984). The latter study
noted poor recovery of silver for both MS and ICP.
ICP and MS have been compared with NAA and XRF in the analysis of
biological materials and soils (Dahlquist, 1978). FM and ICP
compared favorably for the diethyltriaminepentaacetate (DTPA),
exchangeable elements Mn, Fe, Zn, and Cu, and ammonium acetate
exchangeable elements CA, Mg, K, No, and Mn in soil. The authors
concluded that 1CP random and systematic errors were generally
smaller than or equivalent to those for single element techniques
such as MS.
A comparison of ICP and EMS for wear metals in lubricating oils
was reported with good correlation of results from 1CP and FAAS
for oil samples and NBS Standard Reference Materials (SRMs) 1084
and 1085 (Rains, 1985; King, 1984a).
1CP has been evaluated for the screening of hazardous materials
and hazardous waste samples by comparison with MS (Leighty, 1983;
Tzavaras, 1984).
A literature review of multi—elemental analytical techniques for
hazardous waste analysis was recently published (Oppenheimer,
1984). The literature review considered Sb, As, Ba, Be, Cd, Cr,
Cu, Pb, Hg, Ni, Se, Ag, Tl, and Zn analysis by several techniques.
-------
It concluded that ICP, XRF, and Instrumental Neutron Activation
Analysis (1MM) were the most amenable to development as
techniques. The report also recommended additional comparative
studies of the three techniques on samples of hazardous waste.
METHODS AND MhTERIALS
Sample Preparation for ICP Analysis
Many hazardous materials samples contain large amounts of organic
matter for which the sample preparation steps are often long and
involved and sometimes a major source of analytical error. This
includes solid samples such as soil, sludge, auto shredder waste,
and synthetic organic materials; plus liquid samples such as oil,
wastewater and solvent. Nitric acid is used widely for oxidizing
both organic and inorganic substances prior to measurement.
Previous studies have discussed the effects of acid digestion on
1CP performance (for example, McQuaker, 1979). A 20—minute HNO3
digestion was previously compared with more comprehensive
digestion (Tzavaras, 1984).
The HNO3 method (Method A), for digestion and destruction of
organic matter used in this work was a wet digestion in which the
sample was treated with concentrated HNO3, the mixture heated to
100 C—200 C to aid the digestion process, and H202 occasionally
employed in addition to HNO3 as oxidizing agent. Method B was EPA
Method 2050 from Sw—846 (EPA, 1984). Method C was similar to
Method 3050 but used a twofold excess of nitric acid in the
initial digestion step. Wet digestion is much less troubled with
volatilization losses because of the lower temperature compared to
dry ashing methods, particularly for Hg and Se. Using dry ashing
techniques under certain conditions, As, Cd, Cr, Pb, V and Zn have
also been reported to be lost.
The major disadvantage to wet ashing is the possibility of
contamination from the large excess of reagents employed. A
number of precautions can be taken to minimize the difficulties
associated with dissolution in nitric acid. A major problem
associated with HNO3 acid involves samples containing high levels
of total dissolved solids and suspended solids. In such cases,
the dissolved salts in the sample digest is near saturation. This
leads to erratic nebulizer behavior during aspiration of the
solution into the ICP and can eventually lead to total clogging of
the nebulizer due to formation of a deposit inside the glass
capillary of a Meinhard nebulizer. The use of a 0.45 micrometer
filter cartridge with a 20 cc syringe to filter the digest
solution will eliminate most insoluble particles and give a clear
and homogeneous solution. Extending the time for the tip washer
will help to wash any deposits or buildup of the salts on the
nebulizer tip. The use of a peristaltic pump and dilution of the
solution reduces the effect and keeps a constant flow of solution
into the plasma without frequent clogging of the nebulizer.
-------
Figure I shows a comparison of the three digestion methods for a
contaminated soil. Method B gave consistently lower results than
the other two methods. For contaminated soils, of course, none of
these methods are truly total digestion techniques since they are
inefficient at removing elements from silicate minerals. For
regulatory purposes, any of these techniques could be acceptable
provided that they are sufficiently reproducible.
Selection of Analytical Emission Lines for ICP
Routine ICP analysis requires information on the chemical and
physical characteristics of the samples, the lowest quantitative
determinable concentration (detection limit), the analytical
concentration range and the expected potential major spectral
interferences.
Based on this information, the ICP user can consult several
references for selecting the proper spectral lines to be used
under compromise conditions. There are many sources for selecting
the most useful emission lines such as,m “An Atlas of Spectral
Interferences in ICP Spectroscopy,” by M.L. Parsons, et al.
The most widely used wave length tables are commonly referred to
as (NBS) Tables of Spectral Line Intensities and “MIT”.
All wavelengths to be used in the analytical tasks must be stored
for future use. These spectral lines must follow the criteria
used by the majority of published wavelength tables whereby
wavelengths < 200 nm are determined in vacuum and those greater
than 200 nm obtained in air. To adjust for wavelength shift
occurring when lines > 200 run are measured on a vacuum instrument,
each data acquisition program applies the formula by Meggers and
Peters relating the shift to refractive index of air (at 760 mm
pressure and 30 C) to correct each wavelength before performing
analyses.
The ICP—ARL 35000 system contains a special program for wave
length Correction Data (#Elin) and Linearization Correction which
allows the analyst to evaluate the magnitude and rate of change of
correction to be applied within a given segment as shown in Tables
I and II.
Prior to performing the multi—element analysis on hazardous
materials samples, substantial data must be collected to determine
the major interferences.
These interferences were examined when wave length scans for 17
elements were programmed into the system.
The scans were measured around the selected analyte lines on
different complex matrices of samples, standards and blank.
Detailed examination of the scan data used to characterize the
-------
background and off—peak technique was applied by selecting
accordant regions of background on either or both sides of the
spectral lines.
A significantly improved understanding of the sample spectrum caine
about when the major spectral interferences were evaluated from
the scan data and different techniques were applied to eliminate
or minimize the spectral interferences. Selected lines and
comparison between selected lines and major interferences are
shown in Table III.
Effect of Organics
Organic material can often be destroyed prior to analysis by wet
digestion using nitric acid and hydrogen peroxide. The resulting
solution contains the inorganics in acid solution free from
organics and ready for analysis. Any possible effects of residual
organics on the analysis can generally be ignored. Organic
matrices, e.g. solvent extracts, are problematic because the rate
of vaporization cannot be easily controlled with an organic
matrix.
The California waste extraction for hazardous wastes uses a 0.2M
citrate buffer (Calif. Admin. Code). Subsequent metal analysis by
ICP could result in a change in the nebulization process causing
shifts in the slopes of the calibration curve; the organic content
also causes matrix interferences in 1CP from differences in
surface tension and/or viscosity. This can be compensated for by
calibration with standards in the citrate buffer, use of an
internal standard and the use of a peristaltic pump to maintain a
constant flow of sample extract to the plasma. Dilution was used
in some samples to reduce the effect of organics, provided that
resulting detection limits were acceptable.
Instrumentation
Sequential Inductively Coupled Plasma Atomic Emission Spectrometer
Applied Research Laboratories Model 35000:
Grating: Bausch and Lomb diffraction grating
Focal length 1 meter
Blazed at 24001/mm
Computer readout: DPC PDP l1/03L
Rx02 dual floppy disk
Generator: 2.5 KW continuous
Frequency: 27.12 11Hz, quartz stabilized
Power: 1200 watts (3 turn coil)
Torch: Quartz, 18 mm
Argon flow: plasma 0.6 L/min
coolant 12 L/min
carrier 1 L/min
Meinhard nebulizer: 2.1 mL/min flow rate
Observation height: 15 mm above rf coil
-------
SOLID WASTE STUDIES
Auto Shredder Waste
Samples were mixtures of soil and auto shredder waste from a
California facility. Each sample was dried, mixed, ground with
mortar and pestle, and sieved through a 16 mesh sieve. For
“total” analysis 10 g of each sample was digested with
concentrated nitric acid until brown fumes were no longer evolved,
cooled, filtered, and diluted to 100 mLs with 5% nitric acid.
Analysis—Each sample was digested and extracted in triplicate.
Each digest or extract was then analyzed in duplicate, yielding
six measurements for each sample.
The results in Table IV show that the measurement level precision,
as determined by the average CV for replicate measurements, is
consistently better than the overall method precision. Since the
difference is not consistent, i.e., element specific, sample
heterogeneity is the likely major source of variability.
Incinerator Dust
As part of an evaluation of cement kiln incineration of hazardous
wastes, baghouse dust was analyzed for total and extractable lead
by three laboratories. Initial comparison of interlaboratory ICP
and AAS results revealed serious systematic discrepancies.
However, the discrepancy also existed with interlaboratory AA
results, as shown by the initial results in Figure II. The
discrepancy was found to be due to not spectral interferences, but
to the failure of one laboratory to acidify waste extracts prior
to AA analysis. The results after modification of technique are
shown as final results in Figure II.
The comparison of AA and ICP results for routine analysis of
baghouse dust over several months is shown in Figure III.
Surface Soil—XRF vs. ICP
As part of a remedial investigation of a site in Richmond,
California, a comparison of ICP and XRF was made. Samples were
digested with nitric acid according to Method A. XRF analysis was
performed at the Lawrence Berkeley Laboratory. Results are shown
in Figure IV.
Waste Oil Analysis
ICP, AAS, and XRF were used in an evaluation of metals in waste
oil. Samples were digested by nitric acid according to Method A
-------
for analysis by ICP at the Hazardous Materials Laboratory (HML)
and the Air and Industrial Hygiene Laboratory (AIHL). Split
samples were dry ashed and analyzed by ICP at Chevron, Inc.
Results are shown in Figures V—X.
DISCUSSION AND CONCLUSIONS
For the routine analysis of solid waste samples with
concentrations above the upper limit of the range of the method,
the possibility of interelement interference is high. For those
samples, either interelement correction is required or analysis by
MS using method of additions, or both. For the work reported
here, MS was generally used for high level samples. Samples with
a novel matrix should be analyzed by one of the above procedures.
The analysis for arsenic and selenium at trace levels should be
done by GFAA or hydride generation MS.
Mercury analysis should be done by cold vapor MS rather than
ICPAES due to the superior sensitivity and accuracy of the cold
vapor technique.
Whenever feasible, standards should be prepared in the same matrix
as samples, particularly for waste extracts with significant
concentrations of organics.
1CP analysis of hazardous waste samples must be conducted by or
supervised closely by experienced operators who are able to
recognize samples with high potential for spectral overlap or
other chemical or physical interference. QC samples should be
developed for each major matrix which is to be analyzed.
A consensus still needs to be reached on determining method
detection limits and quantification limits for multi—element
analysis. In particular, LODs and LOQs should be developed which
account for variable sample matrix.
Interlaboratory precision and accuracy for ICP are generally
equivalent to or better than those for MS.
As with industrial waste analysis in general, sampling and sample
preparation is generally a source of greater error than the
analytical measurement.
In general, interlaboratory precision and accuracy are highly
variable and dependent on the experience of the participating
laboratories. ICP requires more training of operators than AA for
production of consistently high quality data (Oppenheimer).
Laboratory certification or a similar program would help to
establish standards for laboratories involved in RCRA and CERCLA
work by ICP.
-------
REFERENCES
Dahiquist, R.L. and J.W. Knoll, Applied Spectros. , Vol. 32, No. 1,
pp. 1—29 (1978).
Drenski, T.L., Paxton, C.A., and Varnes, A.W., “Comparison of
Inductively Coupled Plasma and Atomic Absorption Spectroscopies
for Determination of Twenty—Five Elements in Fly Ash Leachate
Solution”, Hazardous Solid Waste Testing: First Conference, ASTM
STP 760 , R.A. Conway and B.C. Nalloy, Eds., American Society for
Testing and Materials, 1981, pp. 225—239.
Environmental Protection Agency, Test Methods for Evaluating Solid
Waste; Physical/Chemical Methods , 2nd edition, revised 1984,
Office of Solid Waste, EPA, Washington, D.C.
Harrison, G.R., “M.I.T. Wavelength Tables”, M.I.T. Press,
Cambridge (1969).
King, Alan D., “Determination of Hazardous Waste Metal Leachates
by ICP Emission Spectroscopy Using RCRA Recommended Procedures”,
Atomic Spectros. , Vol. 5, No. 6, Nov—Dec., 1984, pp. 228—229.
(a) King, Alan D., “Comparison of Results for Determination of
Wear Metals in Used Lubricating Oil by Flame Atomic Absorption
Spectrometry and Inductively Coupled Plasma Atomic Emission
Spectrometry”, Atomic Spectros. , Vol 5, No. 4, July—Aug., 1984,
pp. 181—191.
McQuaker, n.r., David F. Brown, and Paul D. Huckner, Anal Chem ,
Vol. 51, No. 7, pp. 1082—1084 (1979).
Meggers, William F., Charles H. Corliss and Bowedon F. Scribner,
“Tables of Spectral—Line Intensities, NBS”, U.S. Government
Printing Office, Washington, D.C. (1975).
Oppenheimer, J.A., A.D. Eaton, L.Y.C. Leong, and Thomas A.
Hinners, “Multielemental Analytical Techniques for Hazardous Waste
Analysis: the State—of—the—Art”, EMSL, U.S. Environmental
Protection Agency, Las Vegas, EPA—600/4—34—028, April, 1984.
Phelps III, F.M., “M.I.T. Wavelength Tables, Volume 2:
Wavelengths by Element”, M.I.T. Press, Cambridge, 1982.
Tzararas, John, and A.D. Shendrikar, Am Laboratory , Vol. 16, No.
7, pp. 59—62, 64—65 (1984).
-------
Figure I
COMPARISON OF DIGESTION METHODS
As Ba Cd Co Cr Cu Ni Pb Se Zn
METAL
E
z
D
II
F—
z
w
z
D
-J
F-
F—
1000
800
600
400
200
0
-------
Figure II
(AA) —— m 9 /L
INITIAL RESULT
SYMBOL= o
LI NET YPE=
yci+b x
n=3
a0.2930
b=4. 1835
s 1 =1. 5340
FINAL RESULT
SYMBOL=.
LINETYPE=—--—--—
y a+b*x
n=6
a=0. 2342
b=1. 1160
s, 0. 6123
20
EXTRACTABLE
FROM INCINERATOR
LEAD
DUST
/
-J
cn
E
-J
1 /
15
10
I
.
.
-
....
...
:
I
:
I
0
.
:
I
.. .. 1 ’
I
I,
‘V
,/
1 ’
. 1 ’
I/i
#/‘•
1 ’
1
5
:
...
...
:0
1
/
0
• I
I I
I I i
•
I
I
I
t
I I I
s =2. 3790
Sb0. 9200
r=0. 9767
S 0 0. 7893
bb0. 0957
r0. 9856
0
5 10 15 20
LABB
-------
Figure III
EXTRACTABLE LEAD
FROM INCINERATOR DUST
20 _____________________________________— -
/,
/
I
/
I
,/1
1
15 /
-J /1
I
/•
E
I
/
AAvs. ICP
/ SYMBOL=.
LINETYPE=—---—---—
10 yafb*x
I • n=9
‘I.
/ I o=O. 5856 sa=1. 4609
b=1. 1069 Sb0. 1721
I 4892 r0.9248
I
I
/ • o INITIAL RESULT
/•
• FINAL RESULT
• I/I
• / 1
.7,
I
I , I L.... t ‘ I
o 5 10 15 20
LAB A (ICP) —— rn 9 /L
-------
Figure IV
SURFACE SOIL
RICHMOND, CALIFORNIA
250i
/
Zn
I
200 /
F
p
p
F
0 ? 1
150 •Cu XRF vs. 1CP
I SYMBOL=.
E LINETYPE=—--—--—
/ ya+b*x
Pb n=6
/ a —11. 6651 sf6. 6588
I
Li - 100 /‘ b=1. 1212 s 5 =O. 0562
/ 1 s =1O.031B rC.9950
F
I
I
I
50 /
,
I
Cr/
As Ni
• 1
I
0 L?’ I I I I I I I I I i I t i I i
0 50 100 150 200 250
ICR —— mg/K 9
-------
Figure V
LEAD IN WASTE OIL SAMPLES
A IHL—XRF
Maximum ‘a1ue:
Interval
AIHL—XRF
A IHL—AAS
HML—ICP
CHEVRON—I CP
800 ug/g
80 ug/g
802
578.5
578. 6
561.6
I I I I
CHEVRON CP
AIHL—AAS
HML—ICP
-------
Tigure VI
ZINC IN WASTE OIL SAMPLES
AIHL—AAS
Maximum va1u : 1000 ug/g
HMLICP : 8O .5
Incorval : 100 ug/g
AIHL—AAS : 846. 4
CHEVR ONICP : 786.7
VR0N-ICP HML-ICP
-------
Figure VII
CHROMIUM
WASTE OIL SAMPLES
10 ________________________________________—
/
,/1
I/ I
/
8 7’
/ 1
0 ) I
0 ) /1
6 . “ HML vs. AIHL—AAS
SYMBOL=.
LINETYPE=—--—--—
y=a+b*x
/ n7
c=—0. 7602 sa 0. 6005
• 4 / b1. 2808 SbO. 1458
_j /‘ • s, 0. 4769 r0. 9691
• 1’
/
I
1
2 1
I/I
I/I
0 / i I I I I
0 2 4 6 8 10
AIHL—AAS -— ug/g
-------
Figure VIII
LEAD
WASTE OIL SAMPLES
1000 _____________________________________________
/1
/
Boo •/
1
/.
D )
1 ’
3 600 HML vs. CHEVRON
SYMBOL=.
/ LINETYPE —--—--—
/ ya+b x
a..
Li / a-4.9365 s 0 24.5B69
T 400 / b 1. 0390 sbO. 0393
s 26. 5206 r’=O. 9g71
/
• I/I
I/F
200 F
F/F
• I/F
I/I
O I I I I
0 200 400 600 800 1000
CHEVRON—ftP — - u9/9
-------
Figure IX
LEAD
WASTE OIL SAMPLES
1000 ______________________________________________
/
‘V
800
/
c i ) /
600 /
.‘ HML vs. AIHL—XRF
SYMB OL=.
LINETYPE=—--—--—
yo+b*x
a. ‘V n6
Li
F
/ a-22.6291 s 21.O549
b=0. g987 9b0. 0316
400 /‘ s 22. 1830 r=0. 9980
/
/1
/
200
1
/-
/
0” i I I I I
0 200 400 600 800 1000
AIHL—XRF -— U9/9
-------
Figure X
LEAD
WASTE OIL SAMPLES
1000 ________________________________________
1 ’
/,
F
/
800
/
F
F
1 .
0 ) /
600 /‘
HML vs. AIHL—AAS
/
, SYMBOL=.
F/F L!NETYPE=—--—--—
ya+b*x
D— n=5
I
F a-10. 3037 sf20. 8638
b=1.0180 Sb 0. 0324
400 3542 r0. 9980
:I:
/ 1
I/F
200
/
F
F/ I
F/F
F
o/ I I I I
0 200 400 600 800 1000
AIHL—AAS -- u 9 /9
-------
i’ab1e TABLE OF WAVELENGTH CORRECTION DATA #ELIN
NO. WAVE ELEM DELTA 1 DELTA 2 DELTA 3
1 1849.57 HO 7 0.138 0.136 0.129
2 1880.5 AS 12 0.113 0.123 0.127
3 1899.8 SN 12 0.241 0.241 0.246
4 1908.64 TL 8 —0.247 —0.260 —0.265
5 1837.6 AS 12 0.139 0.141 0.147
6 1942.27 HO 8 0.147 0.149 0.158
7 1960.9 SE 11 0.133 0.181 0.190
8 2020.3 MO 4 0.240 0.226 0.239
S 2039.85 SE 12 0.237 0.231 0.231
10 2055.52 CR 7 0.265 0.2G0 0.266
11 2061.91 ZN 5 0.301 0.301 0.302
12 2068.33 SB 8 0.218 0.207 0.211
13 2138.56 ZN 4 0.202 0.196 0.194
14 2144.38 CD 4 0.202 0.201 0.205
15 2169.99 PB 12 0.2O 0.206 0.198
16 2175.9 SB 5 0.109 0.092 0.090
17 2203.53 PB 9 0.165 0.164 0.159
18 2216.5 NI 5 0.140 0.147 0.145
19 2247 C l i 7 0.163 0.153 0.156
20 2265.02 CD 4 0.187 0.169 0.167
21 2283.16 CO 5 0.171 0.161 0.169
22 2289.02 CD 4 0.170 0.167 0.172
23 2316.04 NI 8 0.147 0.141 0.142
24 2335.27 BA 4 0.176 0.170 0.178
25 2348.61 BE 1 0.130 0.135 0.133
26 2378.62 CO 6 0.136 0.128 0.132
27 2382.04 FE 4 0.133 0.117 0.128
28 2539 5 HG 8 0.137 0.139 0.124
29 2576.1 MN 1 0.103 0.097 0.102
30 2599.4 FE 3 0.129 0.125 0.131
31 2677.16 CR 4 0.125 0.120 0.115
32 2767.87 TL 10 0.087 0.080 0.086
33 2795.53 MG 1 0.064 0.062 0.033
34 2816.15 MO 3 0.073 0.085 0.090
35 2824.02 V 5 0.067 0.066 0.067
36 3082.15 AL 8 0.037 0.045 0.049
37 3130.416 BE 1 0.032 0.035 0.045
38 3247.54 CU 3 0.039 0.028 0.035
39 3260.68 AG 4 0.027 0.029 0.007
40 3372.B TI 2 0.002 -0.002 0.002
41 3382.89 AG 4 0.013 0.008 0.023
42 3578.69 CR 5 —0.047 -0.046 —0.045
43 3710.3 V 2 —0.084 —0.085 —0.074
44 3891.78 BA 5 —0.087 —0.078 —0.078
-------
TabJe Ii
TABLE OF LINEARIZATION CORRECUONS IN #COLI
DELTA DELTA
AT fuN. AT MAX.
WAVELENGTH.
MINIMUM
WAVELENGTH
MAXIMUM
INTCPT SLOPE
0.Oo
1849.57
0.000
0.0000725
0.000
0.134
1849.57
1890.50
0.735
—0.0003249
0.134
0.121
189).50
1899.80
—24854
0.0131049
0.121
0.243
1899.80
1908.94
107.654
—0.0565381
0.243
—0.257
1908.64
1937.60
—26.578
0.0137905
—0.257
0.142
1937.60
1942.27
—3.592
0.0019275
0.142
0.151
1942.27
1960.90
—3.709
0.0019874
0.151
0.188
1960.90
2020.30
—1.351
0.0007850
0.188
0.235
2020.30
2039.85
0.433
—0.0000982
0.235
0.233
2039.85
2055.52
—3.770
0.0019625
0.233
0.234
2055.52
2061.91
—11.818
0.0058775
0.264
0.301
2061.91
2068.33
29.056
—0.0139454
0.301
0.212
2068.33
2138.56
0.647
—0.0002102
0.212
0.197
2138.56
2144.38
—1.772
0.0009207
0.197
0.202
2144.38
2169.99
0.179
0.0000108
0.202
0.203
2169.99
2175.90
38.982
—0.0178706
0.203
0.097
2175.90
2203.53
—5.091
0.0023944
0.097
0.163
2203.53
2216.50
3.408
—0.0014726
0.163
0.144
2216.50
2247.00
—0.823
0.0004363
0.144
0.157
2247.00
2265.02
—1.144
0.0005792
0.157
0.168
2265.02
2286.16
0.269
—0.0000449
0.168
0.167
2286.16
2289.02
—3.489
0.0015990
0.167
0.170
2288.02
2318.04
2.298
—0.0009301
0.170
0.143
2316.04
2335.27
—3.534
0.0015879
0.143
0.174
2335.27
2349.61
7.364
—0.0030787
0.174
0.133
2348.81
2378.62
0.197
—0.0000274
0.133
0.132
2378.62
2382.04
4.538
—0.0018523
0.132
0.129
2382.04
2536.50
0.007
0.0000499
0.129
0.134
2536.50
2576.10
2.230
—0.0008266
0.134
0.101
2576.10
2599.40
—2.953
0.0011853
0.101
0.128
2599.40
2677.16
0.408
—0.0001075
0.128
0.120
2677.16
2767.87
1.181
—0.0003963
0.120
0.084
2767.87
2795.53
2.205
—0.0007663
0.084
0.063
2795.53
2819.15
—2.639
0.0009666
0.063
0.083
2816.15
2924.02
0.506
—0.0001502
0.083
0.067
2924.02
3082.15
0.496
—0.0001467
0.067
0.043
3082.15
3130.42
0.431
—0.0001256
0.043
0.037
3130.42
3247.54
0.127
—0.0000288
0.037
0.034
3247.54
3280.68
1.320
—0.0003961
0.034
0.021
3290.68
3372.80
0.745
—0.0002206
0.021
0.001
3372.80
3382.89
—4.661
0.0013821
0.001
0.014
3382.89
3578.69
1.062
—0.0003095
0.014
—0.046
3578.69
3710.30
0.904
—0.0002655
—0.049
—0.081
3710.30
3891.79
—0.084
0.0000008
—0.081
—0.091
3891.78
9999.00
—0.132
0.0000132
—0.081
0.000
-------
NA = Not Analyzed
ICP Spectral Ir ormation and
Comparison
Fligh level of Ca may present line broading interferences between 3900—4000 A
3933.7 and 3968.5.
*I f1gh level of Al presents a high background on Pb line at 2203.5 that should
due to two primary calcitmt ion line at
be corrected.
Table III
id
Element
SW —846
(EPA, 1984)
Wavelength
Note and Major
Interferences
Solid Waste
flML
Uavelength
Note and Major
Interferences
Ant ir 1 onv
206.833
Bad spectral interferences
Cr — Mo
Same
Correction for background shc4
be selected very c refu1lv
ArsenIc
193.696
Very bad background
189.05
—
Clean line— good background
Detection 1i ri 0.035 PP”
Barium
.
455.403
Outside range of ARL—3500
Mo .
233.527
Detection limit 0.0015 PP .I
B rv]lium
313.042
Same
Clean line— good background
Detection l.i’ ’it 0.001 ?P ’1
Cadmium
226.502
Ni — Fe — Mo
214.438
Pt — W — Pd
Detection limit 0.C02 PPM.
Chromium
267.716
Pt — Fe — Mn
205.552
Good detection limit
Detection limit 0.005 PP
Cobalt
228.616
Same
Detection limit 0.004 PPM
Copper
324. 754
Same
Clean line— good background
Detection limit 0.003 PPt
Le2d
*
220. 353
-
Background from Al
Same
Detection limit 0.015 PPM
r c u rv
NA -
XXX CXXXXXXXXXXXXX) XX
NA
c xxxxxxxxxxxxxxxxxx cx: xxx
‘ o1 denum
202.030
Fe
Same
Detection limit 0.008 PPM
Detection limit 0.006 PPM
Nickel —
Selenium
231.604
Same
196.026
Weak Fe interference
Mn — Fe — Co
196.09
Gocd background
Detection 1i’ iit 0.06 PP!
Silver
328.068
Mq
Sa ne
Detection limit 0.002 PPM
Thal1iu’
190.864
No — V
Same
Detection limit 0.19 PPM
Vanadium
292.402
Weak Fe interference
•. Same
Detection limit 0.C05 PPM
Zj 1
213.856
Ni - V
Sa na
Detection limit 0.002 PPM
-------
Table IV Precision of Auto Shredder Analysis
Mean
Conc Method Level Measurement Level
mg/kg Precision, % CV Precision, % CV
Ag 0.0777 53 35.1
As < 0.35 ND ND
Ba 118 4.5 1.0
Be 0.505 2.4 0.90
Cd 4.4 6.8 2.1
Co 8.3 10.8
Cr 44.0 7.7 0.9
Cu 196 37 1.2
Mo 5.8 1.7 1.5
Ni 51.6 16 1.3
Pb 131 6.4 0.9
Sb 3.3 6 6.2
Se < 0.610 ND ND
Ti < 1.B7 ND ND
V 19.4 2 1.0
Zn 694 4.3 0.9
-------
COMPARATIVE PERFORMANCE OF INDUCTIVELY COUPLED PLASMA OPTICAL
EMISSION SPECTROSCOPY (ICP) AND ATOMIC ABSORPTION SPECTROSCOPY (AAS)
IN WASTE ANALYSES
ThOMAS A. HINNEBS, U. S. ENVIRONMENTAL PROTECTION AGENCY - EMSL, LAS
VEGAS, NEVADA
The performance of analytical techniques can vary for different sample
matrices. The analysis of solid samples presents complications of
analyte heterogeneity, digestion inefficiency, and the potential for
more interferences than the analysis of water samples. For example,
only 2 of 10 elements exceeded a relative standard deviation (RSD) of
20% in an MS collaborative study of water by Temperley (New Zealand
3. Sci. 21:557—564, 1978, whereas 8 of 19 elements exceeded a 20% RSD
in an AA interlaboratory study (Amer. Lab. 13:31—35, 1981) on sewage
sludge digested with nitric acid and hydrogen peroxide. Even for
water samples the performance of ICP and MS can differ. Garbarino et
al. at the U.S. Geological Survey recently reported (Applied Spectros.
39(3):535—541, 1985) better correlation coefficients between 2 labs
for cadmium and lead by ICP (0.9997 and 0.9916) than by MS (0.8227 &
0.9495).
In this talk, data obtained by inductively coupled plasma optical
emission spectroscopy (IC?) and atomic absorption spectroscopy (MS)
will be presented for waste digests and for extracts produced by
Method 1310, the official Extraction Procedure (EP) specified in the
Federal Register . The ICP abbreviation will be used in this talk to
Th1 ito ICP/optical emission spectroscopy, and should not be confused
with ICP/atomic fluorescence or ICP/mass spectrometry.
Applying different analytical techniques to the same samples provides
the means to evaluate the accuracy of methods. In a study conducted
for the EPA office in Las Vegas by Dr. Eaton at Montgomery Engineers,
five wastes were analyzed not only by MS and ICP but also by neutron
activation analysis (NM) and x—ray fluorescence (XRF). Since these
latter 2 techniques are applied to undigested solids while the other 2
are typically applied to liquids, sample homogeneity and rigorous
digestion were critical concerns for this multi—method comparison.
The wastes examined were sludges from electroplating, paint production
and a drying—bed process plus 2 soil samples from a disposal site.
The wastes were air dried and ground (with an agate mortar and pestle)
before the fraction passing a 325—urn sieve was distributed for
analysis by the 4 techniques. Particles smaller than 326 urn are
classified as “fine sand” by geologists, and allow even the 0.1—gram
aliquots used for NM to be homogeneous.
The NM and XRF analyses were conducted at the Lawrence Berkeley
Laboratory, and the MS and IC? analyses were conducted at the
Montgomery Engineers’ facility in Pasadena, California. The NM
analyses involved compressing the samples with cellulose into pellets,
a 10—minute irradiation at 11 kW for the short—lived isotopes
(followed by two counting periods on intrinsic germanium detectors)
and an 8—hour irradiation at 1,000 kW for longer—lived isotopes
-------
(followed by three counting periods). The XRF analyses involved
measuring the pulverized samples directly and after dilution with
sulfur powder (to minimize matrix effects) in an energy—dispersive
system designed and built at the Lawrence Berkeley Laboratory. This
XRF system uses a low power tungsten x—ray tube, a molybdenum
secondary target for some analytes (As, Cr, Cu, Hg, Ni, Pb, Se & Zn),
a terbium secondary target for other elements (Ag, Ba, Cd & Sb), a
1,024—channel pulse—height analyzer and a CDC 6600 computer (as
described in Anal. them 49:62—67, 1977).
Since the sample digestion procedure must solubilize all analytes
completely in order for conventional ICP and MS analyses to agree
with XRF and NM analyses, the adequacy of several digestion
procedures were evaluated using two NBS SRMs (River Sediment and Urban
Particulates). The final digestion procedure selected for use on
seven aliquots of each waste sample involved nitric, perchioric and
hydrofluoric acids and a teflon—lined bomb (Table 1). Table 2 shows
results obtained for the NBS SRMS using this digestion procedure.
The elements and analysis wavelengths for the ICP and MS measurements
are shown in Tables 3 & 4. The MS and ICP measurements were obtained
with a Perkin—Elmer 5500 system. Figure 1 shows a graphic comparison
of data for the electroplating waste by the 4 techniques. The ICP
data is indicated by the solid white area and the MS data by the area
with horizontal lines. The average values by MS and ICP are not
statistically different for Cd, Cr, Cu, Ni and Zn in the
electroplating waste, but the MS silver value is 50% lower than the
ICP average. The ICP Ag value is consistent with the values obtained
by XRF and NAA. The low Ag value obtained with MS may have resulted
from the extensive dilution used to bring the Ag concentration into
the linear range for MS analysis. Figure 2 shows arsenic results for
samples where high variability for ICP and NM values is apparent.
Difficulties in measuring arsenic in complex samples by conventional
ICP have induced some investigators, including Dr. Lowry at the
National Enforcement Investigations Center (NEIC) in Denver, to
develop systems to generate and to deliver the gaseous hydride of
inorganic arsenic (and other hydride—forming elements) into ICP
instruments, which serves to separate arsenic from interferences and
to improve the detection limit. Dr. Lowry is scheduled to discuss
sample preparation at the symposium this afternoon. High lead
concentrations in the drying—bed sludge and the paint waste prevented
accurate arsenic determinations by XRF so no XRF values for these
wastes are shown in Figure 2. Figure 3 shows the cadmium results for
the samples and demonstrates an unusually high measurement variability
for the paint waste by NM. This serves to illustrate that
measurement precision can differ drastically for different matrices.
Figure 4 shows the chromium data where a high bias by XRF is indicated
for soil but not for the paint waste. Figure 5 illustrates high
variability for copper measurements by NM, which has a poor
sensitivity for copper. Figure 6 shows agreement among the 4
techniques for nickel except for a high value by XRF for the
drying—bed sludge. This illustrates that accuracy obtained with a
technique on one matrix does not ensure accuracy for a different
matrix. Figures 7 & 8 show the agreement among three of the
-------
techniques for lead. NM is too insensitive for lead measurements
below the percentage level, so no NM lead values are shown in Figures
7 & 8. Agreement between ICP and MS is shown for all samples for
manganese in Figure 9. Figure 10 shows that the ICP Sb data for the
paint waste is consistent with NM and XRF data while the AAS
(furnace) data for Sb is inconsistent with the other 3 techniques.
Loss of Sb by premature volatilization in the MS furnace is probably
involved in the low MS values. Figure 11 shows zinc results for the
samples and demonstrates general agreement. The ICP and MS precision
obtained in this study is generally in the range of 3 to 10% RSD, and
typically half of the RSD value is attributable to the measurement
process while the majority of variation is ascribable to sample
heterogeneity and sample preparation. For example, the RSD is 3% for
repeated 1CP analyses of one of the 7 digests of the electroplating
waste, whereas the RSD for all 7 electroplating digests is 8.4%.
Six of the elements included in this study were not detected in more
than one sample by more than one of the muitieiemental techniques. A
blend of cellulose containing Ag, Cd, Hg, Sb, Se and Ti was mixed with
the waste and 1CP analysis revealed recoveries as shown in Figures 12
& 13. Added mercury could be measured in these matrices at the
l00—ug/g level by ICP. Lead interference prevented reliable thallium
measurements by XRF so no Ti values are shown in Figure 13 for XRF.
In another study conducted under contract at the EPA’s Las Vegas
facility, the official Extraction Procedure (EP) was applied to 6
wastes and the resulting extracts were analyzed by ICP and MS for
barium, cadmium, chromium, lead and silver. Although arsenic, mercury
and selenium are also EP Toxicity elements, they are not currently
included in the elements specified in ICP Method 6010. The inorganic
EP—criterion centrations for classifying a waste as hazardous are
listed in Table 5. MS measurements were obtained with a Perkin—Elmer
Model 5000, and iCp measurements were obtained with a Spectrametrics
echelle spectrometer (Spectraspan lilA). Analytical conditions for
both techniques are indicated in Tables 6 & 7. The wastes examined in
this study are identified in Table 8. The EP extracts of these wastes
were spiked at 20% arid 100% of the EP—toxicity criterion levels in an
effort to provide two measurable concentrations for the AM and ICP
measurements. ICP Method 6010 requires that EP extracts be digested
prior to analysis and that the method of standard additions be used
for quantitation. Measurements were also conducted to evaluate the
need for digestion of the EP extracts and for the use of the method of
standard additions. 1CP and MS results obtained by the method of
additions for the spiked extracts are shown in Tables 9—13. Table 9
shows the chromium data and includes the RPD (relative percent
difference) for the average concentrations by the two techniques for
the six samples. For chromium the 1CP average concentration exceeded
the AM Cr average by more than 10% only for the lower—spike level (1
ppm) for the plating waste and the storage tank wastewater. The
average Cr RSD was 7% by ICP and 9% by MS for these waste extracts.
Data for digested and undigested extracts are included in Tables 9—13,
but are summarized on Tables to be shown later. Table 10 shows that
the average RSD for Cd is 8.3% by ICP and 7.1% by MS and the relative
percent difference (RPD) only exceeded 10% for the sulfuric acid waste
-------
(where the ICP values are 0.2 mg/L higher than the MS values). Table
11 shows several RPD values above 10%, but the MS Ba detection limit
is near 2 mg/li and precision is expected to be poor near the detection
limit. With the exception of the plating and sulfuric acid wastes (#1
& #5 in Table 11), Ba in digested extracts was much higher than the Ba
in the undigested extracts. While digestion of the EP extracts does
not always seem necessary, it does in the case of Ba for many wastes.
As expected, Ba added to the sulfuric acid waste (#5) was lost as
BaSO4, and soluble Ba was below both MS and ICP detection limits. Ba
recovery was low for the 2 mercury wastes (#4 & #6 in Table 11) even
with digestion of the extracts.
Table 12 shows the AAC/ICP comparison for Pb in the waste extracts.
Precision was worse by 1CP than MS and undigested extract
concentrations were frequently much lower than the concentrations in
the digested extracts. Table 13 shows the comparison for Ag and is
complicated by instability in the concentrations in the undigested
extracts (except for wastes *4 & #5).
Although digestion of EP extracts did not always produce a significant
difference, an extensive difference was observed for Ag, Ba & Pb in
several (but not all) of the wastes examined (Tables 14 & 15). While
Ba & Pb concentrations are higher in digested extracts than in
undigested extracts, the opposite pattern was observed for Ag. The
silver concentrations were low but stable in digested extracts, but
unstable in undigested extracts.
Use of the method of standard additions increases the time and cost of
analyses. Measurements conducted to determine the benefit obtained
using the method of additions (Table 16) for these EP extracts
revealed that a difference of more than 6% for Cr and 7% for Ba & Cd
were only obtained for the storage tank wastewater ( 1 of the 6
wastes). Poor precision and instability made conclusions on Ag & Pb
uncertain. Time and cost can be minimized if the method of standard
additions is only used when “recovery” of a spike added to a prepared
sample indicates a significant interference.
-------
Figure 1. Results for the electroplating.waste (mean values and 95%
confidence intervals In ug/g).
I
V// ø 2 )
r —lcis
2 øø
15000
P
P
‘4
5000
-------
Plating Soil • 1 Soil 2 DryIng Pigment Clean Soil
Wast• Sludge Waste
Figure 2. Arsenic results for the samples (mean values and 95%
confidence Intervals in Mg/g).
__‘UP
I 1 S
P
P
‘4
A
S
-------
I iIcP
>
jAAS
P
P
M
C
D
Soil 1 SoIl 2 Drying Bed Pigment Clean Soil
Sludge Waste
Figure 3. Cd results for the samples (mean values and 95% confIdence Intervals In pgfg).
-------
Figure 4. Cr results for selected waste samples (mean values and 95%
confidence Intervals in ixg/g).
F i’cP
I t4 S
P
P
‘I,
C
R
Soil 1 SoIl 2 P l9nent Clean Soil ( .I’))
Waste
-------
Figure 5. Ct! results for selected waste samples (mean values and 95%
confidence Intervals In ug/g).
F 7 /, 3 >
_I s
P
P
M
U
Soil 1 Son 2 P1 ient Clean Soil
Waste
-------
NI results for selected waste samples (mean values and 95%
confidence intervals in ‘gig).
__ 1 icr
E 1 )
I —l S
P
P
P4
N
I
SoIl 1 SoIl 2 Drying Bed Pigment Clean Soil
Sludge Waste
Figure 6.
-------
I j!cP
r i > 3: F-
I
Figure 7.
Pb results for the low concentration samples (mean values and 95%
confidence intervals In pg/g).
P
P
N
p
B
Plating Waste Soil 1 Son 2 Clean Soil
-------
Drying Bed Pigment
Sludge Waste
I
V Y/iA )W
Figure 8. Pb results for the high concentration samples (mean values and 95%
confidence Intervals In i’g/g).
16000
14000
12000
10000
8000
6000
4000
2000
P
P
H
P
B
-------
Plating Soil 1 Soil 2 DryIng Bed Pigment
Waste Sludge Waste
Figure 9.
Mn results by ICP—OES and AAS (mean values and 95%
confidence Intervals in pg/g).
P
P
‘4
‘4
N
.1_iIcP
t .J S
Clean Soil
-------
I j!CP
I
Flgt re 10. Sb results 7or the paint pigment sample (mean values and 95%
confidence Intervals In sJg/g).
P
P
II
S
B
PI €HT bk srE
-------
Figure ii. Zn results for high concentration samples (mean values and 95%
confidence intervals in pg/g).
I 1cP
2 1 ) F
_i ;s
P
P
M
z
ti
plating Waste Drying Bed Sludge Pigment Waste
-------
__JHG-ICP
I: : j HGXRF
k ?$ ’ SB—ICP
LI SB—XRF
hi
>
0
()
hi
“I..
S.’
Plating Soil 1 Soil• 2 Drying Bed Pigment
Waste Sludge Waste
Figure 12. Percent recovery for Hg and Sb added to the wastes.
-------
Plating Soil 1 SoIl 2 DryingUed PI ent
Waste Waste
Sludge
I 1 SEICP
.:1SE .XRF
r> €i TL-ICP
I I TL-XRF
>-
L i i
>
0
0
w
‘-S
Figure 3. Percent recovery for Se and 11 added to the wastes.
-------
Table 1. Sample Digestion Procedure
Teflon—lined Digestion Bomb
500 mg sample
4 mL HF, 2 mL HNO 3 , 2 mL HC1O 4 (Ultrex grade)
1200 C for B hours
Cool 1 hour
Add 1 g H 3 B0 3 (Ultrex grade)
Dilute to 50 mL with D I water
-------
Table 2. ICP—OES Results for NBS SRMS*
Urban-Particulates
Element Set 1 Set 2
(1648)
True
River
Sediment
(16
45)
Set 1 Set
2
Set
3 True
Ag 6.3 6.5 (6) 3 1 1 no value
As 96 98 115 65 63 62 66
Ba 700 720 (737) 360 340 340 no value
Be 5.3 4.2 no value 2 1 1 no value
Cd 68 69 75 10 9.8 9.8 10.2
Cr 400 380 403 2.7% 2.6% 2.6% 2.96%
Cu 620 630 609 110 102 105 109
Fe 3.9% 3.7% 3.9% 10.1% 9.9% 9.7% 11.3%
Hg 6 8 no value 14 5 3 1.1
Mn 750 790 (860) 750 720 710 785
Ni 76 73 82 32 31 30 45.8
Pb 6600 6500 6550 680 670 670 714
Sb 43 41 (45) 28 10 6 (51)
Se 23 20 (24) 25 9 5 no value
Ti 7.1 6.7 no value 5 2 2 1.44
Zn 4600 4520 4760 1710 1680 1650 1720
*Data in ug/g, except where noted otherwise
Parenthese = uncertified NBS value
-------
Table 3. ICP-OES Conditions Used For Analyses Of Waste Samples
Wavelength
Background Measurements
Distance from Analytical
Detection Limit
(nm)
Wavelength (mu)
(ug/g)
Element
Ag 328.07 0.08 Below 1
As 197.20 0.10 Below 9
Ba 455.40 0.09 Above 0.2
Be 313.04 0.06 Below 0.3
Cd 226.50 0.06 Below; 0.10 Above 0.4
Cr 205.55 0.08 Below; 0.15 Above 0.7
Cu 324.75 0.12 Below 0.6
Fe 259.94 0.10 Above 0.7
Hg 194.23 0.11 Above 3
Mn 257.61 0.12 Above 0.2
Ni 231.60 0.06 Above 2
Pb 220.35 0.10 Below; 0.05 Above 6
Sb 206.83 0.05 Below; 0.12 Above 7
Se 190.03 0.14 Below 9
Tl 190.86 0.10 Below 5
Zn 213.86 0.05 Below; 0.05 Above 0.2
Plasma Gas Flow 12.1/mm.
Aux Gas Flow 0.1/mm.
Nebulizer Gas Pressure 36 psi (2.5 kg/cm 2 )
Forward RF Power 1 .2 kW
Viewing Height 16 mm above load coil
-------
Table 4. Analytical Conditions For AAS Analysis Cf Waste Samples
Wavelength
Slit Width
Bkg
Correction
Element r im
nm
(W,
D 2 ,
None) Method
Ag 328.1 0.7 D 2 Flame
As 193.7 0.7 D2 Furnace
Ba 553.6 0.4 W Flame
Be 234.9 0.7 D 2 Flame
Cd 228.9 0.7 D2 Flame
Cr 357.9 0.7 W Flame
Cu 324.8 0.7 Flame
Hg 253.7 0.7 None Cold Vapor
N j. 232.0 0.2 Flame
Pb 283.3 0.7 Flame
Sb 217.6 0.2 D 2 Furnace
Se 196.0 2.0 D 2 Furnace
P1 276.8 0.7 D 2 Furnace
Zn 213.9 0.7 D 2 Flame
D 2 = Deuterium Lamp
W = Tungsten Lamp
-------
Table 5. EP Toxicity Concentrations
Element EP-Criterion Concentration (mg/L)
Ag
5.0
As
5.0
Ba
100
Cd
1.0
Cr
5.0
Hg
0.2
Pb
5.0
Se
1.0
Table 6. ICP Analytical Conditions
Instrument:
Spectrametrics SpectraSpan III A ICP—DCP Emission Spectrometer
capable of operating in a manual sequential single—element
mode or simultaneous multi—element mode. The Monochromator
is a modified Czerny-Turner using an Echelle grating with 30°
prism for order separation.
Entrance slit size: 50 urn by 200 urn
Background Correction: High 8 and Low 6
Source: Argon ICP
Source Power: 1200 W
Plasma Gas Flow: 16 L/min
Auxiliary Gas Flow: 0.2 — 0.4 L/min
Nebulizer: GMK Babington Nebulizer
Nebulizer Gas Flow: 0.3 L/min at 22 psi
Sample Uptake Rate: 1.5 mL/min using a Gi].son Minipuls 2
Peristalic Pump
Plasma Viewing Height: 15 mm above the load coil
-------
Table 7. Instrumental Detections Limits
AAS
ICP
WL
Detection
WL
Detection
Element
(nm)
Limit (ing/L)
(nm)
Limit (mg/L)
Ag
328.1
0.02
328.068
0.02
Ba
553.6
0.2
455.403
0.04
Cd
228.8
0.01
214.438
0.003
Cr
357.9
0.08
267.716
0.008
Pb
283.3
0.3
283.306
0.5
Pb
220.353
0.1
Table 8.
Description of
Hazardous
Waste Samples
Waste Number
Description
1 Plating waste sludge
2 Drying bed solids
3 NBS QA sludge
4 Mercury storage tank wastewater
5 Sulfuric acid scrubber waste solution
6 Mercury settling basin sludge
-------
TABLE 9. COMPARISON OF ICP AND AAS RESULTS FOR Cr
Icrc AASC
Spike RSD RSD
Sample Levela Digestb x ± s ( ) n x ± s (%) RPDd
1 1 1 4 1.05 ± 0.12 11 6 0.88 ± 0.19 22 +18
2 1 4 4.55 ± 0.09 2.0 6 4.27 ± 0.20 4.7 +6.4
2 2 6 4.57 ± 0.19 4.2 6 4.47 ± 0.34 7.6 +2.2
2 1 1 4 0.90 ± 0.10 11 6 0.85 ± 0.14 16 ‘-5.5
2 1 6 4.87 ± 0.72 15 6 5.00 ± 0.37 7.4 -2.6
2 2 6 3.97 ± 0.06 1.5 6 3.93 ± 0.38 9.7 41.0
3 1 1 4 1.01 ± 0.08 8.0 6 0.92 ± 0.13 14 +8.9
2 1 6 4.85 ± 0.45 9.3 6 4.65 ± 0.22 4.7 +4.2
2 2 6 4.55 ± 0.25 5.5 6 4.59 ± 0.31 6.8 -0.9
4 1 1 4 1.14 ± 0.23 20 6 1.01 ± 0.24 24 +12
2 1 6 4.77 ± 0.38 8.0 6 4.87 ± 0.36 7.4 -2.1
2 2 6 3.71 ± 0.22 5.9 6 3.43 ± 0.23 6.7 +7.8
5 1 1 4 388 ± 24 6.2 6 405 ± 29 7.2 -4.3
2 1 6 384 ± 3 0.9 5 422 ± 12 2.8 —9.3
2 2 6 365 ± 11 3.0 6 381 ± 18 4.7 —4.3
6 1 1 6 27.1 ± 0.9 3.3 6 27.1 ± 0.6 2.4 0
2 1 6 30.0 ± 1.4 4.7 6 31.1 ± 1.4 4.5 -3.6
2 2 6 28.6 ± 2.0 7.0 6 28.8 ± 1.6 5.6 -0.7
a Spike 1 = 20% of the EP toxicity concentration.
Spike 2 = 100% of the EP toxicity concentration.
b Digest 1 = digested sample.
Digest 2 = undigested sample.
c Results are expressed in mg/L.
‘ 1CP - XAAS
dgp 9 = xlOO%
nIcp Xj p + AAS XAAS
ICP + GAAS
-------
TABLE 10. COMPARISON OF ICP AND AAS RESULTS FOR Cd
! CP AASC
Spike RS 5 — RSD
Sample Levela n x ± (%) n x ± S (%) RPDd
1 1 1 4 0.226 ± 0.018 8.0 6 0.232 ± 0.037 16 -2.6
2 1 4 0.930 ± 0.039 4.2 6 0.865 ± 0.052 6.0 +7.3
2 2 6 0.952 ± 0.075 7.9 6 0.948 ± 0.022 2.3 +0.42
2 1 1 4 0.239 ± 0.025 10 6 0.237 ± 0.036 15 +0.84
2 1 4 1.039 ± 0.091 8.8 6 1.097 ± 0.061 5.6 -5.4
2 2 6 1.028 ± 0.068 6.6 6 1.003 ± 0.027 2.7 +2.5
3 1 1 4 1.520 ± 0.079 5.2 6 1.53 ± 0.130 8.5 -0.66
2 1 4 2.29 ± 0.270 12 6 2.36 ± 0.040 1.7 -3.0
2 2 6 2.40 ± 0.140 5.8 6 2.335 ± 0.047 2.0 +2.7
4 1 1 4 0.174 ± 0.024 14 6 0.160 ± 0.033 21 +8.4
2 1 4 0.934 ± 0.065 7.0 6 0.948 ± 0.035 3.7 -1.5
2 2 6 0.928 ± 0.063 6.8 6 0.893 ± 0.031 3.5 +3.8
5 1 1 4 0.277 ± 0.018 6.5 6 0.064 ± 0.011 17 +143
2 1 6 1.028 ± 0.079 7.7 6 0.793 ± 0.030 3.8 +26
2 2 6 0.980 ± 0.096 9.8 6 0.782 ± 0.047 6.0 +22
6 1 1 6 0.208 ± 0.024 12 6 0.230 ± 0.012 5.2 -10
2 1 6 0.975 ± 0.060 6.2 6 0.983 ± 0.028 2.8 -0.82
2 2 6 0.910 ± 0.097 11 6 0.933 ± 0.040 4.3 -2.5
a Spike 1 = 20% of the EP toxicity concentration.
Spike 2 = 100% of the EP toxicity concentration.
b Digest 1 = digested sample.
Digest 2 = undigested sample.
C Results are expressed in mg/L.
ICP - MS
dRp 0 = — x I O O%
1CP XICP + AAS XAAS
ICP + AAS
-------
TABLE 11. COMPARISON OF ICP AND AAS RESULTS FOR Ba
ICPC AASC
Spike RSD RSD
Sample Levela Digestb ± s ( ) n ± s (%) RPDd
1 1 1 4 21.9 ± 1.7 7.8 6 24.8 ± 1.3 5.2 —12
2 1 4 88.7 ± 7.0 7.9 6 91.8 ± 1.9 2.1 —3.4
2 2 6 90.6 ± 7.2 8.0 6 86.4 ± 1.7 2.0 +4.8
2 1 1 4 15.1 ± 2.1 14 6 16.5 ± 0.77 4.7 -8.8
2 1 6 95.7 ± 12 12 6100.0 ± 1.9 1.9 -4.4
2 2 6 6.43 ± 0.28 4.4 6 4.12 ± 1.0 27 +44
3 1 1 4 17.58 ± 0.54 3.1 6 17.62 ± 1.3 7.4 -0.23
2 1 5 91.1 ± 12 13 6 93.1 t 1.8 1.9 —2.2
2 2 6 2.65 ± 0.37 14 6 1.23 ± 0.75 61 +73
4 1 1 4 10.7 ± 1.3 12 6 8.00 ± 0.77 9.6 +30
2 1 6 8.22 ± 2.4 29 6 6.30 ± 0.43 6.8 +26
2 2 6 0.73 ± 0.11 15 6 <2 -
1 1 6 (0.4 6 <2
2 1 6 (0.4 6 (2
2 2 6 <0.4 6 <2
6 1 1 6 20.9 ± 1.4 6.7 6 19.4 ± 1.6 8.2 +7.4
2 1 6 52.4 ± 2.0 3.8 6 52.6 ± 1.1 2.1 —0.38
2 2 6 26.3 ± 1.2 4.5 6 11.8 ± 0.95 8.0 +76
a Spike 1 = 20% of the EP toxicity concentration. e Added Ba lost as Ba SO 4 precipitate from
Spike 2 = 100% of the EP toxicity concentration, this sulfuric acid matrix.
b Digest 1 = digested sample.
Digest 2 = undigested sample.
C Results are expressed in mg/L.
XICP -
dRPD xIOO%
nICP XIcp + ‘ 1 1U\S XAAS
1CP + AAS
-------
Spike
Sample Levela Digestb
b Digest 1 = digested sample.
Digest 2 = undigested sample.
C Results are expressed in mg/L.
d RPD =
ICP - XAAS
x 1OO7
n
TABLE 1 2. COMPARISON OF ICP AND AAS RESULTS FOR Pb
1CPc _____
S
RSD
fl
RSD
(%)
RPDd
I
2
3
4
5
6
a Spike
Spike
1
2
2
1
1
2
4
3
4
1.42 ±
4.78 ±
7.69 ±
0.54
2.08
1.5
38
44
21
6
6
5
1.61 ± 0.39
5.33 ± 1.13
5.06 ± 0.97
24
21
19
-12
—11
+30
1
2
2
1
2
2
1
1
2
1
1.
2
7
8
6
4
6
5
64.5 ±
86.4 ±
19.7 ±
34.6 ±
46.5 t
4.52 ±
14.2
16.2
5 6
3.6
6.3
2.2
22
1Y
28
10
14
49
6
6
6
6
6
6
62.77 ± 0.85
78.8 ± 1.5
21.7 ± 0.71
38.83 ± 1.3
43.3 ± 1.5
4.28 ± 1.6
1,4
1.9
3.5
3.3
3.5
37
+2.7
+9.1
—9.7
—11
+7.1
÷5.5
1
2
2
1
1
2
4
2
1
3.49 ±
3.60 ±
1.66
2.0
0.07
58
2.0
6
6
6
1.45 ± 0.21
4.85 ± 1.1
1.65 ± 1.1
14
23
23
+90
—28
+0.6
1
2
2
1
1
2
1
1
2.04
1.13
6
6
6
6.10±0.75
2.42 ± 0.48
2.62 t 1.06
12
20
40
—16
-61
1
2
2
1
1
2
2
2
1.75 ±
4.78 ±
0.64
0.078
36
1.6
6
6
4
1.64 ± 0.32
5.31 ± 1.1
1.05 ± 1.1
20
21
104
÷6.6
-10
1 = 20% of the EP toxicity concentration.
2 = 100% of the EP toxicity concentration.
‘ 1 ICP XICP + AAS XAAS
flICP +
-------
d Spike 1 = 20% of the EP toxicity concentration.
Spike 2 = 100% of the EP toxicity concentration.
b Digest 1 = digested sample.
Digest 2 = undigested sample.
C Results are expressed in mg/L.
ICP ICP + AAS XAAS
n S
RSU
n
RSD
(%)
TABLE 13. COMPARISON OF ICP AND AAS RESULTS FOR Ag
ICPC AASC
Spike
Saj ple Levela Diqestb RPDd
I
2
3
4
5
6
1
2
2
1
1
2
4
3
5
0.926
1.79
0.634
± 0.24
± 0.83
± 0.16
26
46
25
6
6
6
0.877
1.36
3.49
± 0.12
± 0.15
± 0.65
14
11
19
+5.5
+29
-130
1
2
2
1
1
2
4
3
6
0.749
2.04
0.854
± 0.16
± 0.34
± 0.20
21
17
23
6
6
6
0.667
1.97
2.76
± 0.13
± 0.27
± 0.43
20
14
16
+10
+3.5
—105
1
2
2
1
1
2
4
3
6
0.735
1.90
1.65
± 0.088
± 0.20
± 0.22
12
11
13
6
6
6
0.713
1.63
2.64
± 0.068
± 0.17
± 0.23
9.5
10
8.7
+3.0
+16
-46
1
2
2
1
1
2
4
2
6
0.962
3.90
3.90
± 0.24
± 1.1
± 0.30
25
28
7.7
6
6
6
1.005
4.38
4.45
± 0.070
± 0.21
± 0.16
7.0
4.8
3.6
-4.4
-11
—13
1
2
2
1
1
2
3
4
6
0.18
0.30
1.81
± 0.15
± 0.25
± 0.48
83
83
26
5
6
6
0.086
4.47
<0.5
± 0.042
± 0.37
49
8.3
-
+125
-85
1
2
2
1
1
2
6
6
6
0.816
1.36
1.53
± 0.15
± 0.40
± 0.33
18
29
22
6
6
6
0.782
1.26
2.14
± 0.085
± 0.17
± 0.21
11
14
9.8
+4.3
+7.6
-33
d RPD =
•ICP - XAAS
x 100%
niCp + AAS
-------
TABLE 14. COMPARISON OF DIGESTED AND UNDIGESTED ALIQUOTS BY ICP
Digested Aliguot
Element n x ± s
Sample
n
Undi qested Al I quot
x±S
RPD (%)
I
(plating waste
sludge)
Ag
Ba
Cd
Cr
Pb
4
4
4
4
3
1.79 ±
88.7 ±
0.930 ±
4.55 ±
4.8 ±
0.83
7.0
0.039
0.09
2.1
5
6
6
6
1
0.63
90.6
0.952
4.57
6.8
±
±
±
±
±
0.16
7.2
0.079
0.19
1.5
+101
—2.1
-2.3
-0.44
-34
2
(drying bed
solids)
Ag
Ba
Cd
Cr
Pb
3
6
4
6
8
2.04 ±
96 ±
1.039 ±
4.87 ±
86 ±
0.34
12
0.091
0.72
16
6
6
6
6
6
0.85
6.43
1.028
3.97
19.7
±
±
±
±
±
0.20
0.28
0.068
0.06
5.6
+96
+175
+1.1
+20
+115
3
(NBS s1udge)
Ag
Ba
Cd
Cr
Pb
3
5
4
6
6
1.90 ±
91 ±
2.29 ±
4.85 ±
46.5 ±
0.20
12
0.27
0.45
6.3
6
6
6
6
5
1.65
2.64
2.40
4.55
4.5
±
±
±
±
±
0.22
0.37
0.14
0.25
2.4
+14
+206
-4.7
+6.4
+153
4
(Hg storage
tank wastewater)
Ag
Ba
Cd
Cr
Pb
2
6
4
6
2
3.9 ±
8.2 ±
0.934 ±
4.77 ±
3.60 ±
1.1
2.4
0.065
0.38
0.01
6
6
6
6
1
3.90
0.73
0.92
3.71
1.66
±
±
±
±
0.30
0.11
0.052
0.22
0
+167
+0.64
+25
+66
5
(sulfuric acid
scrubber waste
solution)
Ag
Baa
Cd
Cr
Pb
4
6
6
6
1
0.30 ±
<0.4
1.028 ±
384.5 ±
2.04
0.25
0.079
3.4
6
6
6
6
1
1.81
0.980
365
1.13
±
<0.4
±
±
0.48
0.22
4.1
-125
+4.8
+5.2
+57
6
(Hg settling basin
sludge)
Ag
Ba
Cd
Cr
Pb
6
6
6
6
6
1.36 ±
52.4 ±
0.975 ±
30.0 ±
8.8 ±
0.40
2.0
0.060
1.4
5.3
6
6
6
6
5
1.53
26.3
0.910
28.6
8.4
±
±
±
±
±
0.33
1.2
0.097
2.0
4.7
-12
+66
+6.9
+4.8
+4.6
Added Ba lost as Ba SO 4 precipitate from this sulfuric acid matrix.
-------
TABLE 15. COMPARISON OF DIGESTED AND UNDIGEStED ALIQUOTS BY AAS
SamDle Element
Digested Aliguot
n x±S
Undi gested Al I guot
n x±S
RPD (%)
1
Ag
6
1.36 ±
0.15
6
3.49
±
0.65
—88
(plating waste
Ba
6
91.8 ±
1.9
6
86.4
±
1.7
+6.1
sludge)
Cd
Cr
Pb
6
6
6
0.865 ±
4.27 ±
5.3 ±
0.052
0.20
1.1
6
6
5
0.948
4.47
5.06
±
±
±
0.022
0.34
0.097
-9.2
-4.6
+4.6
2
Ag
6
1.97 ±
0.27
6
2.76
±
0.43
-33
(drying bed
Ba
6
100.0 ±
1.9
6
4.1
±
1.0
+184
solids)
Cd
Cr
Pb
6
6
6
1.097 ±
5.00 ±
78.8 ±
0.061
0.37
1.5
6
6
6
1.003
3.93
21.7
±
±
±
0.027
0.38
0.8
+9.0
+24
+114
3
Ag
6
1.63 ±
0.17
6
2.64
±
0.23
-47
(NBS sludge)
Ba
Cd
Cr
Pb
6
6
6
6
93.1 ±
2.360 ±
4.65 ±
43.3 ±
1.8
0.040
0.22
1.5
6
6
6
6
1.23
2.335
4.59
4.28
±
±
±
±
0.75
0.047
0.31
1.6
+195
+1.1
+1.3
+164
4
Ag
6
4.38 ±
0.21
6
4.45
±
0.16
-1.6
(Hg storage
Ba
6
6.30 ±
0.43
6
<2
tank wastewater)
Cd
Cr
Pb
6
6
6
0.948 ±
4.87 ±
4.8 ±
0.035
0.36
1.1
6
6
6
0.893
3.43
1.6
±
±
±
0.031
0.23
1.1
+6.0
+35
*100
5
Ag
6
0.086 ±
0.042
6
4.47
±
0.37
-192
(sulfuric acid
B 8
6
<2
6
(2
scrubber waste
Cd
6
0.793 ±
0.030
6
0.782
±
0.047
+1.4
solution)
Cr
Pb
5
6
42? .
2.42
12
0.48
6
6
381
2.4
±
±
18
1.1
+10
+0.83
6
Ag
6
1.26 ±
0.17
6
2.14
±
0.21
-52
(Hg settling basin
Ba
6
52.6 ±
1.1
6
11.8
±
0.9
+127
sludge)
Cd
Cr
Pb
6
6
6
0.983 ±
31.1 ±
5.3 ±
0.028
1.4
1.1
6
6
4
0.933
28.8
1.1
±
±
±
0.040
1.6
1.1
+5.2
+7.7
+131
a Added Ba inst as Ba, 0 4 precipitate from this sulfuric acid matrix.
-------
TABLE lb. COMPARISON OF STANDARD-ADDITIONS AND DIRECT-CALIBRATION RESULTS FOR Ba BY ICP
Spike
Sample Levela Digestb
SAC Cc
a Spike Level 1 = 20% of the EP toxicity concentration.
Spike Level 2 = 100% of the EP toxicity concentration.
b Digest 1 = digested sample.
Digest 2 = undigested sample.
C SA = results from standard additions.
C = results from direct calibration. Results are expressed in mg/I.
‘ ‘ICP ‘ ICP + AAS XAAS
x 100%
fl
RSD
(%)
n X±S
RSD
(%)
RPDd
1
1
2
2
1
1
2
4
4
6
21.9 ± 1.7
88.7 ± 7.0
90.6 ± 7.2
7.8
7.9
7.9
4
4
6
23.10 ± 0.65
87.5 ± 1.6
89.3 ± 4.6
2.8
1.8
5.2
-5.3
#1.4
+1.4
2
l
2
2
1
1
2
4
6
6
15.1 ± 2.1
95.7 ±12
6.43 ± 0.28
14
12
4.4
4
6
6
15.8 ± 1.1
99.6 ± 8.5
6.44 ± 0.22
7.0
8.5
3.4
—4.5
-4.0
-0.16
3
1
2
2
1
1
2
4
5
6
17.58 ± 0.54
91.1 ±12
2.64 ± 0.37
3.1
13
14
4
6
6
17.58 ± 0.40
92.9 ± 8.9
2.57 ± 0.10
2.3
9.6
3.9
0
-2.0
+2.7
4
1
2
2
1
1
2
4
6
6
10.7 ± 1.3
8.22 ± 2.4
0.73 ± 0.11
12
29
15
4
6
6
9.19 ± 0.51
7.24 ± 0.78
0.621± 0.058
5.5
11
9.3
+15
+13
+16
5
1
2
2
1
1
2
<0.4
<0.4
<0.4
<0.4
<0.4
<0.4
6
1
2
2
1
1
2
6
6
6
20.9 ± 1.4
52.4 ± 2.0
26.3 ± 1.2
6.9
3.8
4.6
6
6
6
20.68 ± 0.98
54.4 ± 3.6
28.2 ± 1.1
4.7
6.6
3.9
+1.1
-3.8
-7.0
d RPD =
XICP - XAAS
ICP + AAS
-------
EVALUATION OF Sw-846 METHODS FOR METAL EXTRACTION/DISOLUTION FROM
AQUEOUS, OIL, AND SOLID WASTE MATRICES
DR. N. LJMANA, D. WHITAKER, R. HANDY, D. NATSCHKE, J. COVINGWN,
W. GIJTKNECHT, S. B. WHITE, B. J. WORK, AND E. D. PELLIZZARI, RESEARCH
TRIANGLE INSTITUTE, RESEARCH TRIANGLE PARK, NORTH CAROLINA
ABSTRACT
The objective of this study was to evaluate two selected digestion
methods (Methods 3010 and 3020) for aqueous wastes, two selected
digestion methods (Methods 3030 and 3040) for oil wastes, and two
selected digestion methods (Methods 3050—FLAA and 3050—GFAA) for
solubilizing metals in solid wastes. This effort was designed to
provide information on several aspects of metal determination,
including accuracy, precision, recovery, matrix effects, and other
methodological parameters. Analyses of the digestates were performed
by flame atomic absorption and graphite furnace atomic absorption
utilizing the method of standard additions for quantitation.
The work was performed in two categories. During category 1
experiments, a variety of waste samples were digested using all
methods. Spiked and unspiked waste samples were analyzed in
duplicate. During category 2 experiments, seven separate aliquots of
one aqueous waste sample, one oil waste sample, and one solid waste
sample were digested using the appropriate methods. Each sample was
prepared unspiked, spiked at level 1, and spiked at level 2. In
addition, procedural blanks and controls were prepared from standard
solutions. They were subjected to the digestion procedure and
analyzed exactly as the waste samples.
The results were analyzed statistically. The usefulness of each
method for solubilizing the metals in a variety of matrices was
evaluated from the results obtained for the category 1 experiments.
Accuracy, as reflected by the recovery of spikes and control samples,
and precision, as reflected by the coefficient of variation, were
determined from the category 2 experiments.
Method 3010 is recommended for the majority of metals tested (except
Ba and Mn). Limited information was obtained for Methods 3020 and
3050—GFAA; they were labor intensive and are recommended only when the
limit of detection of flame atomic absorption is not sufficient for
the particular application.
It is recommended that inductively coupled argon plasma spectroscopy
be considered for evaluation as an alternative analytical technique to
graphite furnace atomic absorption.
Method 3030 is recommended for only a few of the metals tested: Cu,
Ag, Be, Cr, and Hg. Method 3040 was found applicable and is
recommended only for Cu, Cd, Cr, Fe, and Mn. The conclusions obtained
for Methods 3030 and 3040 indicate applicability only to oil—soluble
-------
organometallic compounds. No extrapolations to other inorganic forms
of metals (e.g., emulsions or suspended solids) should be made.
Method 3050—FLAA is recommended for the majority of metals tested.
Exceptions are Ag, V , and Zn.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency’s peer and administrative review policies and
approved for presentation and publication.
INTRODUCTION
The second edition of “Test Methods for Evaluating Solid Waste” [ 1]
contains the procedures that may be used by the regulated community or
others to determine whether a waste is hazardous as defined by the
Resource Conservation and Recovery Act (RCRA). This document
describes methodology for collecting representative samples of the
waste and for determining the ignitability, corrosivity, reactivity,
extraction procedure (EP) toxicity, and composition of the waste.
The manual includes methods approved by the U.S. Environmental
Protection Agency (EPA) and used by the regulated contiminity to support
waste evaluations. The methods are used for listing and delisting
petitions, and for describing the methods that will be used by EPA in
conducting investigations. The manual serves as a compilation of
state—of—the-art methodology for conducting such tests. However, many
of the methods presented in the manual have not been fully evaluated
by EPA using materials characteristic of the wastes regulated under
RCRh.
The objective of this study was to evaluate the metal solubilization
procedures (Methods 3010, 3020, 3030, 3040, and 3050) as described in
the EPA Office of Solid Waste (OSW—EPA) manual. The Research Triangle
Institute (RTI) evaluated these solubilization methods for liquid and
solid hazardous waste materials. The methods were evaluated by
analyzing the resulting solution and comparing precision and accuracy
estimates and other methodological parameters. A variety of aqueous,
oil, and solid wastes were used in the evaluation. This study was
divided into two categories as described below.
In category 1 the purpose was to evaluate the usefulness of the
digestion procedures for a variety of actual wastes and to investigate
possible chemical matrix effects on spike recovery. The focus of the
work in the category 2 was to establish for each element and method,
an estimate of the precision and accuracy for a typical sample matrix
at three concentration levels. In category 1, each sample was
subjected to the digestion procedure in duplicate according to the
methods appropriate to that type of sample (e.g., aqueous samples,
Methods 3010 and 3020; oil and grease samples, Methods 3030 and 3040).
The first experiments performed consisted of the digestion of all the
unspiked wastes using the appropriate methods. A procedural blank and
control were also subjected to the digestion procedure and treated as
the samples. These digestates were then analyzed and, from these
-------
results, the spiking levels were determined (approximately 10 times
the original concentration). The second step was to digest all the
spiked wastes using the appropriate method following both the category
1 and category 2 design. Again, a procedural blank and control were
subjected to the digestion procedure. Finally, all digestates were
analyzed and the data were treated using statistical methodology.
For each element in the SW—846 manual, the sample preparation method
is either specifically referenced or described. In this work only the
referenced elements in each preparation method were evaluated. Two of
the methods, 3010 and 3020, are written specifically for flame (FLAA)
and graphite furnace (GFM) atomic absorption spectroscopy,
respectively, and Method 3050 provides direction for the use of both
techniques. In category 1, all sample preparation and analyses by
FLAA of Methods 3030 and 3040 digestates were completed before GFAA
analyses were attempted. If the concentration of an element in the
sample was within the analytical range of the FLAA calibration curve
and more than 10 times the limit of detection, GFAA analysis of that
sample was not performed. All graphite furnace analyses were
quantitated by the method of standard additions, unless a
single—addition spiking procedure verified that the method of standard
additions was not required. For the method of standard additions, the
digestate was spiked at two different concentrations. The large spike
doubled the apparent original concentration or added an amount equal
to the midpoint of the calibration curve, whichever was the larger.
The second spike was one—half the first. The spiking solution was a
discrete solution for each element because the spiking levels
differed. The sample solution was diluted if the measured
concentration exceeded the linear portion of the analytical
calibration curve.
After the work in category 1 was completed, the data were reviewed to
select a hazardous waste material appropriate for estimating method
precision and accuracy. For this work, one sample was selected for
each method and seven replicates analyzed at three fortified
concentration levels for each element. The criteria used for the
selection of the waste for category 2 experiments were based on (1)
the presence of the metals under study in the original waste, (2) the
availability of the waste, and (3) the origin of the waste.
The three levels spanned the range of the calibration curve. In many
cases, the lowest concentration was near the estimated instrumental
detection limit and the highest concentration was near the upper limit
of the calibration curve. The middle level was a concentration
between the lowest and highest concentrations. If the elemental
concentration in the sample was above 500 ug/mL or 500 ug/g, spiking
was not performed. If the concentration in the sample was appreciably
above the analytical linear range, provisions were made for diluting
the sample before analysis. Since the elemental concentration in the
various samples was different, no one sample material sufficed for all
elements.
In all cases, the analytical technique used after sample preparation
was either FLAA or GFAA spectroscopy. Fourteen different liquid and
-------
solid waste samples were employed. These samples were selected from
the following types of matrices: four aqueous waste samples, four
oil— or grease—containing waste samples, and six solid waste samples.
Statistical analyses were performed and the final data were summarized
as recovery of spikes and controls as well as precision as expressed
by the coefficient of variation.
MATERIALS AND EXPERIMENTAL PKROCEDURES
Materials
All of the chemicals used in this study were reagent grade obtained
from commercial sources. The water used was deionized laboratory
grade (ultrapure, resistivity at 25 C, 18 megohm—cm). Aqueous
Certified Atomic Absorption Standards were used for preparation of
controls, spiking solutions, and calibration solutions for the AA
instrument. Oil—based standards were obtained from Conostan, Inc.
and were used for preparation of controls, spiking solutions, and
calibration solutions; xylene was obtained from Fisher Scientific. A
few highly concentrated standard solutions were prepared at
approximately 10,000 ug/mL in this laboratory from the salts. After
drying, these salts contained an undetermined amount of H20;
therefore, the concentrations were not calculated from gravimetric
measurements. Concentrations were determined by inductively coupled
argon plasma (ICP) spectroscopy analysis. The ICP instrument was
calibrated with the Aqueous Certified AA Standards.
Waste Samples
A variety of liquid and solid wastes were used in this evaluation.
For aqueous samples were used for the evaluation of Method 3010 and
3020. Four samples containing oil or grease were used for the
evaluation of Methods 3030 and 3040, and six solid waste samples were
used for the evaluation of Method 3050. Included in the evaluation
was a used lubricating oil and an oil/water emulsion from a
cold—rolling facility. Of the six solid samples collected, one was a
mixed wastewater treatment sludge, one a solid waste sample from an
electroplating operation, and one a soil contaminated by
metal—containing solid waste. The remaining three solid samples were
sludges selected from other industrial sources.
When possible, samples used in the evaluation were collected from
hazardous waste sources that were regulated for the greatest number of
elements. The hazardous waste classification code [ 2] was consulted
to determine and select the industrial waste sources. This approach
allowed the methods to be evaluated with realistic sample matrices.
A solid sample was supplied by the QA Branch of the EPA Environmental
monitoring Systems Laboratory (Cincinnati, OH). This sample was a
municipal sludge previously analyzed and used as a standard for Method
3050. The remainder of the samples were obtained from local
industries. The description and origin of all samples are listed in
Table I.
-------
In addition, Triangle Resources, Inc., Reidsville, NC, supplied three
types of oil samples. Also, a waste motor oil sample was obtained
from a local service station. This latter sample was composed
primarily of used lubricating oil from car engines operated with
leaded gasoline. The EPA Office of Solid Waste (OSW) supplied soils
contaminated with metals.
METhODS
Method 3010: Acid Digestion Procedure for Flame Atomic Absorption
Spect ros copy
The acid digestion procedure for flame atomic absorption (FLAA)
analysis was used to determine the total amount of metal in the
sample. This digestion procedure is prescribed by the OSW for
preparing aqueous samples, EP and mobility procedure extracts, and
certain nonaqueous wastes for analysis by FL.AA spectroscopy. The FLAA
analytical procedure has been described in the EPA 7000 series.
The metals tested in this study are Sb, Ba, Cd, Cr, Pb, Ni, Be, Cu,
Fe, Mn 1 Na, V, and Zn.
In the acid digestion procedure, the sample is digested with nitric
acid (HNO3) to near dryness in a Griffin beaker. This step is
repeated with additional portions of HNO3 until the digestate is light
in color or until its color has stabilized. After the digestate is
brought to near dryness, it is cooled and taken up in dilute
hydrochloric acid (HC1). The sample is then ready for FLAA analysis.
Method 3020: Acid Digestion Procedure for Furnace Atomic Absorption
Spectroscopy
The acid digestion procedure is prescribed for preparing aqueous
samples, mobility procedure extracts, and certain nonaqueous wastes
for analysis by furnace atomic absorption (GFAA) spectroscopy. It was
used for the following metals in this study: Ba, Be, Cd, Cu, Cr, Fe,
Pb, Mn, Ni, Ti, Ag, V, and Zn.
In the digestion procedure, the sample is digested with HNO3 to near
dryness in a Griffin beaker. This step is repeated with additional
portions of HNO3 until the digestate is light in color or until its
color had stabilized. After the digestate is brought to near dryness,
it is cooled and taken up in dilute HNO3 to a final dilution of 0.5
percent (v/v) HNO3. The samples are then analyzed by GFAA.
Method 3030: Acid Digestion of Oils, Greases , or Waxes
The acid digestion procedure for oils, greases, and waxes is
prescribed for preparing samples that contain substantial amounts of
oil, grease, or wax. In this study, the samples were analyzed for the
total concentration of the following metals: Sb, Ba, Cd, Cr, Pb, Ni,
Ag, As, Se, Hg, Be, Cu, Fe, Mn, Ti, V, Zn, and Na.
-------
To describe the method briefly, a representative sample is placed in a
Kjeldahl or similar flask and digested with sulfuric acid (112S04),
HNO3, and hydrogen peroxide (H202). The digestate is then analyzed
for metal content by either FLAA or GFAA.
Method 3040: Dissolution Procedure for Oils, Greases , or Waxes
Method 3040 is prescribed for preparing samples containing oil,
grease, or wax for the following elements: Sb, Ba, Ca, Cr, Pb, Ni,
Ag, As, Se, Be, Cu, Fe, Mn, Tl, V 1 Zn, and Na. This method may also
be applicable to the analysis of other metals in these matrices.
The method is briefly described as follows. A representative sample
is dissolved in an appropriate solvent (e.g., xylene or methyl
isobutyl ketone). Organometallic standards are prepared using the
same solvent, and the samples and standards are analyzed by either
FLAA or GFAA.
Method 3050: Acid Digestion of Sludges
Method 3050 is an acid digestion procedure used to prepare sludge and
soil samples for analysis by FLAA and GFAA or by ICP spectroscopy.
Samples prepared by Method 3050 were analyzed by atomic absorption
spectroscopy for the following metals: Sb, Ba, Cd, Cr, Pb, Ni, Ag,
Se, Be, Cu, Fe, Mn, Tl, V, Zn, and Na.
To summarize the method, a dried and pulverized sample is digested in
HNO3 and H202. the digestate is refluxed with either HNO3 or HC1.
Hydrochloric acid is used as the final refiux acid for the furnace
analysis of Ag and Sb or the flame analysis of Ag, Sb, Be, Cd, Cr, Ti,
Cu, Pb, Ni, and Zn. Nitric acid is employed as the final refiux acid
for the furnace analysis of As, Be, Cd, Cr, Cu, Pb, Ni, Se, Ti, and
Zn.
DIGESTION PROCEDURES
All the unspiked wastes were digested in duplicate according to the
appropriate method (i.e., aqueous wastes by methods 3010 and 3020, oil
wastes by methods 3030 and 3040, solid wastes by method 3050). The
digestates were analyzed by atomic absorption (AA) spectroscopy and
the results were used to estimate the spiking levels for category 1
and category 2 experiments (samples were spiked at approximately 5 and
10 times the original concentration or 5 and 10 times the LOD). For
those samples in which the original concentration of an element was
very high (>500 ug/mL), additional element was not added.
After the wastes were spiked, they were digested according to the
design for both category 1 (all wastes in duplicate) and category 2
(one waste in seven replicates) experiments. One procedural blank and
one control were digested together with the unspiked samples and
another procedural blank and control were subjected to the digestion
procedure with the spiked samples (categories 1 and 2).
-------
Method 3010
The procedural blank consisted of deionized water, and the control
consisted of a standard mixture of approximately 10 ug/niL for each of
the required metals prepared in deionized water. The 10 ug/mL
concent ration value was chosen because most regulated metals have a
hazardous limit in the 5 to 10 ppm range.
The wastes were homogenized and a representative 100.0 niL aliquot of
each waste was used per digestion. The aliquots were transferred by a
100.0 niL volumetric pipet into 400 mL, Griffin beakers, which had
previously been acid—washed in 5 percent HNO3. For those samples that
required spiking, the appropriate volume of spiking solution was
added. Three milliliters of concentrated HNO3 were added to each
beaker, the beakers covered with a watch glass and placed on the hot
plate. Samples were evaporated to approximately 25 to 50 mL. The
samples were cooled (except for 2258—9—A—3A), and another 3 mL of HNO3
added. This process was continued several times until the digestion
was completed. Since sample 2258—9—A—3A solidified when cooled, acid
was added while the sample was still warm.
Due to the nature of some of the samples being digested, reduction to
near dryness as specified was impossible. The samples were initially
saturated solutions and concentration caused substantial precipitation
and finally solidification upon cooling. For this reason, samples
were removed from the hot plate before solidification occurred.
For category 1, digested samples 2258—9—A—3C and 2258—9—A—4A readily
dissolved in dilute HC1 when they were diluted to a 100.0 mL final
volume. Digested samples 2258—9—A—3A and 2258—9—A—3B, however, did not
dissolve in 100.0 mL and were subsequently diluted to 500.0 niL.
Sample 2258—9—A—3A readily dissolved in the 500.0 niL, but sample
2258—9—A—36 did not. The liquid portion of the latter was decanted
and analyzed.
Sample 2258—9—A—3C was chosen for Method 3010 category 2 digestions
because it contained measurable amounts of most of the elements of
interest.
Method 3020
The wastes were homogenized and a representative 100.0 niL aliquot of
each waste was used per digestion. The aliquots were transferred by
100.0 mL volumetric pipet into 400 niL Griffin beakers, which were
acid—washed in 5 percent HN03. Three milliliters of concentrated HNO3
were added to each beaker, the beakers covered with a watch glass and
placed on the hot plate. Samples were evaporated to approximately 25
to 50 niL.
The samples were cooled (except for 2258—9—A—3A), and another 3 mL
HNO3 added. This process was continued several times until the
digestion was completed. Since sample 2258—9—A—3A solidified when
cooled, acid was added while the sample was still warm. Final
dilution of samples was with dilute HNO3.
-------
The procedural blank was prepared with 100.0 mL deionized water and
was treated the same as the samples. The control was prepared using
100.0 mL distilled, deionized water spiked with metals at
approximately 0.25 ug/znL.
After digestion, sample 2258—9—A—3A solidified to green crystals. The
digestate was diluted to a final volume of 500 niL (rather than 250 niL
used for other samples) to dissolve it completely. A clear emerald
green solution resulted.
Sample 2258—9—A—3B formed a gelatinous material after digestion. In
fact, the first attempt to digest this waste resulted in serious
“bumping” and contaminated the other samples. All these samples were
discarded and digestion was repeated with fresh materials. Sample
2258—9—A—3B (second attempt) dissolved incompletely in 500 niL.
Samples 2258—9—A—4A and 2258—9—A—3C formed dark brown sludge—like
crystals after digestion. All solids readily dissolved in 250 mL
water. All the above described digestates were then prepared for
graphite furnace atomic absorption (GFM).
Sample 2258—9—A—3C was used for the Method 3020 category 2 digestions
since it was also used for the Method 3010 category 2 digestions.
Method 3030
The oil waste samples were homogenized by stirring, aliquoted, and
digested by the procedure described in Method 3030. A minor change
was made in the procedure to allow for a final dilution to 100 mL
instead of 25 niL. This change was necessary to ensure a sufficient
volume of digestate for all of the required elemental analyses. For
each digestion, a sample weight of approximately 8.0 g was used for
the unspiked samples. A procedural blank was prepared using Conostan
base oil, and a control was prepared using the Conostan metal
standards. The final concentrations of the controls were in the 50 to
100 ug/g range.
The exact sample amount was determined by placing an excess of the
sample into a small container and weighing it with a Pasteur pipet and
bulb. The weight was noted and portions were transferred, using the
pipet, to a 300 niL acid—washed Kjeldahl flask. The weight of the
container, sample, and pipet was again noted after the removal of some
sample. The weight difference was the weight of the sample added to
the flask (weight by difference). In cases where an aqueous and an
organic phase were present, the phases were separated prior to
weighing and only the organic phase was digested.
n acid—washed glass bead (6 mm) plus 40 niL concentrated H2S04 were
added to each weighed sample. The flask and the contents were swirled
to ensure complete mixing.
The necks of the Kjeldahl flasks were cooled during the digestion by
directing a stream of ambient air from a heat gun around the middle
-------
portion. The flasks were heated gently until white acid fumes
appeared. At this point, 1 mL concentrated HNO3 was added dropwise to
the hot mixture using a Selectapette pipet. After the HNO3 had boiled
of f and the white fumes reappeared, another addition of HNO3 was made.
If the fumes did not appear in a reasonable amount of time (5—10 mm),
the heat was increased slightly. This process was continued until the
sample was no darker than a straw or amber color. A minimum of 7 mL
concentrated HNO3 (the volume added to the blank and control) was
added to the samples, blanks, and controls.
After the initial additions of HNO3, when the sample was partially
oxidized to a straw color, 0.5 mL additions of 30 percent H202 were
made dropwise using another Selectapette pipet. These H202 additions
were immediately followed by a 1 mL addition of HNO3. Once the white
fumes reappeared, an additional 0.5 mL volume of H202 and a 1.0 mL
volume of O3 were added. These additions continued until the sample
was nearly colorless and clear or until 33 additions had been made.
The temperature was gradually increased over the duration of the
digestion.
After digestion was complete, the flasks were cooled and the contents
quantitatively transferred to 100 mL volumetric flasks. When the
digestion had to be interrupted (e.g., overnight), the flasks were
cooled to room temperature and Parafilm placed over the top until the
digestion could be resumed.
All of the samples were pale yellow and clear at the end of the
digestion. As they cooled they became colorless and, with the
exception of sample 2258—9—0—2B, they remained clear. Sample
2258—9—0—23 was slightly cloudy. Several of the samples produced a
fine white precipitate upon dilution with water. The amount of time
required for the sample digestions was 7 to 10 hours; however, the
blank and control required 27 hours. The blank and control were
digested first and at a lower, more gentle heat, thus they required a
longer time to digest. Since these digestions took an excessive
amount of time, the temperature was increased for the samples. The
samples were generally started on a low setting and the setting was
gradually increased over the duration of the initial digestion with
HNO3. The samples were then kept on a high setting throughout the
H202 treatment.
The oil waste samples were spiked with Conostan organometallic
standards and digested according to the procedure outlined in Method
3030. Approximately 4.0 g of sample were used for digestion. The
amount of spike added was determined by the concentration of the
Conostan standards available and by the need to keep the total oil
spiking volume to S g or less. Sample 2258—9—0—lA was used for the
3030 category 2 digestions.
Method 3040
The unspiked oil samples were prepared for analysis according to
Method 3040. A representative aliquot of approximately 10 g was
weighed directly into a dry 100 mL acid—cleaned volumetric flask.
-------
weighing was performed on a top loading digital balance. The samples
were then diluted to volume using xylene. In the case of sample
2258—9—0—2B, for which several phases were present, only the phases
that would dissolve in xylene were used. Sample phases were separated
with a separatory funnel prior to weighing.
The procedural blank was prepared by weighing Conostan 75 base oil
into a volumetric flask and diluting with xylene. The controls were
prepared at approximately 100 and 170 ug/g by weighing amounts of the
Conostan standards into a volumetric flask and then diluting to volume
with xylene.
All samples dissolved readily in xylene. Samples 2258—9—0—lA,
2258—9—0—2A, and 2258—9—0—2B each left a small amount of light—colored
precipitate after settling.
Oil waste samples were spiked with Conostan organometallic standards
and dissolved in xylene as described in Method 3040. The samples were
weighed directly into dry 100 mL volumetric flasks. A balance capable
of weighing to 0.001 g was used to weigh approximately 5.000 g of each
sample.
Spiking levels were based on the concentration of the Conostan
standards available and also on the need to keep the total ratio of
oil (sample plus standards) to xylene at 1:10. This restriction
limited the total mass of the standards to approximately 5.0 g.
Sample 2258—9—0—lA was chosen for Method 3040 category 2 dissolutions
based on the elements present in the unspiked sample and on the
availability of this particular waste. For the level 1 spiked
samples, approximately 5.0 g of sample were used with 2.5 g of base
oil and approximately 3 g of standards (total standards combined).
The level 2 spike consisted of approximately 5.0 g of sample, no base
oil, and approximately 5.5 g of standards (total standards combined).
The unspiked samples consisted of approximately 5.0 g of sample and
approximately 5.0 g of base oil. All samples were diluted to 100.0 mL
with xylene.
Method 3050
The unspiked solid samples were prepared for FLAA analysis according
to Method 3050. Wastes were homogenized by careful mixing and then
aliquoted.
Sludge samples were first dried in an oven set at 60 C. The oven was
placed inside a fume hood and the drying time varied for each sample.
When the wastes were dry they were pulverized with a mortar and
pestle. The dried wastes were then transferred to glass bottles with
Teflon—lined screw caps.
One gram of each sample was weighed to the nearest 0.1 mg and
quantitatively transferred to a 250 mL acid—washed Phillips beaker.
Ten milliliters of 1:1 HNO3 were added per beaker and each beaker
covered with a watch glass. The samples were heated at approximately
-------
95 C (temperature estimated by comparison to 30 mL water at this
setting). The samples were refluxed at this temperature for
approximately 10 minutes. After cooling, 5.0 mL of concentrated HNO3
were added, the watch glass replaced, and the samples again refluxed
for 30 minutes.
After this second reflux, the samples were cooled to room temperature.
Deionized water (2 mL) and 30 percent H202 (3 mL) were added to each
beaker. The beakers were covered, returned to the hot plate, and
gently warmed. The samples effervesced. Once effervescence had
subsided, the beakers were again cooled and additional 1 mL aliquots
of 30 percent H202 were added. This sequence was repeated a total of
seven times (total volume of 10 mL H202 added).
For FLAA analysis, 10.0 mL deionized water and 5.0 mL 1:1 HC1 were
added and the samples refluxed for 10 minutes. After cooling, the
samples were filtered using Whatman No. 42 filter paper that had been
washed with 0.5 percent HNO3. Samples were filtered directly into 100
mL volumetric flasks and diluted to 100.0 mL final volume.
The control was prepared by adding a spike of 1.00 mL of each aqueous
atomic absorption standard (1,000 ug/mL standard) into a Phillips
beaker and drying at 60 C. This treatment of the control is
consistent with the handling of sludge samples since the evaporation
of the aqueous portion of samples at 60 C is necessary for them as
well. The element concentrations in the control were approximately
1,000 ug/g.
All samples were digested without difficulty. The filtration step,
however, was rather slow (approximately 2 hours) due to occlusion of
the filter by the fine particles left after digestion.
Unspiked solid samples were prepared for GFAA analysis according to
Method 3050. The element concentrations in the controls were
approximately 25 and 1,000 ug/g.
Sample preparation was basically the same as the FLAA method. After
the addition of H202, the samples were left on the hot plate, the
cover glass was removed, and the volume was reduced to approximately 2
mL. The samples were cooled, 10 mL deionized water were added, and
the mixture was rewarmed. The samples were then cooled and filtered.
The procedural blank was prepared in the same manner as the samples.
The control was prepared by adding 25 uL of 1,000 ug/mL aqueous
standards to the 250 mL Phillips beaker, drying at 60 C, and then
digesting. All samples digested with no problems. The filtration
step was somewhat faster (approximately 1 hour) than for flame
preparation due to a smaller final volume.
Solid waste samples were spiked with aqueous standards, dried, and
digested according to the procedure outlined in Method 3050. One set
of samples was digested for flame analysis and another set for furnace
analysis. One gram of each sample weighed to the nearest 0.1 mg was
used for each digestion.
-------
Sample 2258—9—S—5B was used for the Method 3050 category 2 digestions.
ATOMIC ABSORPTION ANALYSIS
Flame Atomic Absorption
An Instrumentation Laboratories, Inc. (IL) Model 357 single—beam
atomic absorption spectrometer equipped with a deuterium arc
background corrector was used for all FLAA analyses. The method of
standard additions was used for the calculation of metal content. The
appropriate 7000 series protocols and the instrumental manual (4) were
consulted for analytical wavelength and other experimental conditions.
The standard that typically corresponded to the top of the linear
portion of the calibration curve was prepared for the element of
interest and run at the start of each analytical day. For this
measurement, the instrument was auto—zeroed with an analytical blank,
which consisted of the acid mixture characteristic of the method under
evaluation. Then, the samples for analysis were diluted so that their
absorbance was equal to or less than one—third the absorbance of the
standard described above.
Based on this treatment, an approximate analyte concentration was
estimated for each sample. The higher standard additions spike added
to the sample digest was designed to double the element concentration
in the sample. The lower standard additions spike was typically
one—half of the higher one.
FLAA analysis was performed on Methods 3010, 3030, and 3050
digestates. Dilutions were made with deionized water or 0.5 percent
HNO3 (if high dilution factors were needed) for Methods 3010 and 3030
digestions, and with 2.5 percent HC1 for 3050—FLM digestions.
Samples from metal extraction method 3040 were also analyzed by FLAA.
The equipment used and general procedure were the same as described
for FLAA analysis of Method 3010 digestates. After the xylene—diluted
spiked samples were prepared, each set was run as a standard addition
analysis. The Method 3040 digestates were not analyzed for Ti (there
is no conunercial source for a Tl—in--oil standard). Analysis for Ba by
FIIA in Method 3040 digestates was impossible; analytical Method 7080
is specific for samples prepared by Method 3040 (dissolution in
xylene) yet it requires the use of potassium chloride (KC1) in water
as an ionization suppressant. This is not compatible with the xylene
solvent. Attempts to analyze without the KC1 were not successful.
Besides the ionization problem noted in Method 7080, there is also a
significant emission problem. Occasional negative absorbances (noted
on the energy meter) lead to overflow conditions and “flex curve”
error messages. Method 7080 is not consistent on the question of
background correction. Section 2.0 requires its use while Sections
3.0 and 7.0, interferences and Procedure, do not mention it.
Background correction was not used for Ba since the deuterium arc
system could not balance with the source lamp at this high a
-------
wavelength (553.6 nm). Comments here concerning background correction
for Ba also apply to Method 7770 for Na. Phenomena such as light
scattering are a function of wavelength, and background correction,
documented in several reports, is not usually considered a requirement
past Cu at 324.7 rim. However, an exhaustive literature search was
outside the scope of this project.
Graphite Furnace Atomic Absorption
The IL instrument described above was used for most of the GFAA
analyses. A second instrument used for a portion of the GFAA work was
a Perkin—Elmer Model 603 AA with the model 2100 graphite furnace, an
AS—i autosampler, a PRS—i0 printer, and a deuterium arc background
correction system. The instrument was used with the following switch
settings: concentration, peak height, background correction, and
expansion. Charring and atomization temperatures were taken from the
manufacturer’s handbook. An injection of 20 uL was used. Peaks were
recorded on a strip—chart recorder with peak height response printed
on the PRS—10. The samples analyzed with the Perkin—Elmer instrument
are noted on the results tables. Care was taken to use the same
calibration standards with both instruments to verify their
equivalency.
GFAA analyses were performed on unspiked samples from extraction
Method 3020 and 3050, and on those elements from Methods 3030 and 3040
that were below the LCD of FLAA.
Analysis was performed by the method of standard additions, in a
manner similar to that for FLAA. The appropriate 7000 series method
was consulted for the required wavelength, background correction
requirements, and matrix modifiers. An aqueous standard was prepared
at a level expected to be at the top of the linear working curve.
This standard was used to confirm satisfactory response under these
experimental conditions. The appropriate dilution factors and spiking
levels were determined by surveying diluted digestates versus this
aqueous standard. Diluted and spiked digestates were then prepared
and analyzed by the method of standard additions. Dilutions were made
with 0.5 percent HNO3 for Methods 3020 and 3050—GFAA digestions;
deionized water or 0.5 percent HNO3 was used for Method 3030
digestions and xylene was used for Method 3040 digestions.
A number of difficulties were encountered that required certain
modifications to the procedures.
For As and Se analysis, a solution of 1,000 ug/mL Ni is required as a
matrix modifier by the 7000 methods. Information obtained from
application chemists with IL indicated that this level of Ni leads to
suppression in their furnace and that 200 ug/mL Ni is more
appropriate. This trend was confirmed experimentally at RTI. Two
hundred micrograms per milliliter Ni was, therefore, used for the GFAA
analysis of 3030 and 3050 digestates for As and Se. An additional
modification was made in that the atomization temperature was set at
2,500 C rather than the manufacturer’s recommended 1,800 C.
-------
Analyses of As and Se in samples prepared by Method 3040 were
attempted unsuccessfully on both the IL and Perkin—Elmer instruments.
Both analytical procedures call for the addition of nickel nitrate as
an aqueous solution. This addition is not possible for preparation
Method 3040 in which the samples are diluted with xylene. Charring
temperatures greater than 1,500 C were necessary to reduce the
background to a correctable level. This temperature resulted in a
loss of As and Se. Char temperatures below 400 C for times greater
than 5 minutes did not combust the matrix.
For Ba and v, the manufacturer’s recommended conditions are pyrolytic
graphite tubes with an atomization temperature of 2,500 C. Even at
2,700 C, no more than a rise in baseline was observed when standard
solutions were used. These two elements are difficult to analyze on
any graphite furnace because of the formation of highly refractory
carbides.
Analysis of Ba was attempted unsuccessfully on both instruments.
Application chemists with IL and Perkin—Elmer were consulted. They
described the analysis as possible but extremely difficult. Carbide
that forms in the furnace is extremely refractory, which makes high
atomization temperatures necessary. In addition, the presence of
strong Ba emission lines close to the absorbance line leads to a high
noise level. Perkin—Elmer stated that the alignment of the furnace
about the light path is critical to this analysis. Two chemists
attempted this alignment. An atomization temperature of 2,800 C (the
highest marking on the furnace power supply) was used without success
on the Perkin—Elmer instrument. IL lists an atomization temperature
of 2,500 C. We attempted 2,700 C in the IL instrument without
success. Higher temperatures were not tried for fear of burning out
the temperature sensor. IL indicates that a fast heating rate
modification of the power supply is beneficial to this analysis. This
modification was not available to this project.
Cold Vapor Atomic Absorption
Mercury was analyzed in oil samples prepared by Method 3030 by the
cold vapor technique. Stannous chloride (0.5M) and the open system
alternatives from Method 7470 were used. The response was recorded on
a strip—chart recorder and peak height was measured electronically.
The method of standard additions was used for calibration, as it had
been throughout the entire study. A desiccant tube was not used.
Past experience in this laboratory has shown that the problems
associated with clogging of the tube cause more trouble than the water
vapor. The reading lamp heater alternative was used instead.
Deuterium arc background correction was used to correct for
nonspecific absorption resulting from the water vapor reaching the
cell. No problems were experienced with the analytical procedure.
Statistical Analysis Procedures
-------
Upon completion of the chemical analysis, the data obtained were keyed
into online computer data sets. category 1 data were collected to
provide recovery information, as a measure of usefulness. The
formulas used in calculating recoveries are:
(Spiked sample measured—procedural blank) —
Percent matrix recovery = ( Unspiked sample measured—procedural blank )
Spiking concentration x 100
( Control measured—procedural blank )
Percent control recovery = (Control concentration) x 100
Note that for statistical analysis purposes all values reported as the
LOD were divided by 2 prior to being used; all data reported as below
the analytical blank were considered 0. While the matrix recovery
incorporates both spiked and unspiked data, the control recovery is
developed separately for spiked data and for unspiked data. There is
a matrix recovery for each method, waste sample, element, and
replicate combination. There is a control recovery for each method,
element, and spiking combination.
The waste—specific matrix recovery means were developed from the two
replicate recoveries for each method and element. The overall matrix
recovery means are based on eight recoveries for Methods 3010, 3020,
3030, and 3040. Since Methods 3050—FLAA and 3050—GFAA involved 6
waste samples each, the overall means are based on 12 recoveries.
Methods 3030 and 3040 are special cases: the unspiked sample data for
these methods are a mixture of FLAA and GFAA data. For each, the GFAA
data, when available, were first incorporated into the data set. The
FLAA data were added to complete the data set. When a FLAA waste
sample measurement was used in a calculation, so were the
corresponding blank and control measurements. Calculations involving
a GFAA waste sample measurement used its corresponding procedural
blank and control measurements.
Category 2 data were collected to provide single—laboratory accuracy
and precision information. For each method, there are seven
replicates for one waste at each of three spiking levels. All seven
replicates were excluded from the statistical analysis when four or
more measurements were at the rOD.
Percent matrix recoveries were calculated as a measure of accuracy.
There is a matrix recovery for each method, element, and spiking
levels 1 and 2.
The procedural blanks used in calculating category 2 recoveries were
the same blanks used for category 1. Since all spiked samples (both
category 1 and category 2) were digested at the same time, only one
procedural blank and control were needed. As noted above for Methods
3030 and 3040, the GFAA data, when available, were first incorporated
into the data set and then FLAA data were added to complete it. The
blanks available from the data set for Methods 3030 and 3040
recoveries calculation were usually GFAA data.
-------
To evaluate precision, the coefficient of variation (CV) was
calculated for each method, element, and spiking level combination.
RESULTS AND DISCUSSION
The waste matrices were digested according to the procedures described
in Methods 3010 and 3020 for aqueous matrices, Methods 3030 and 3040
for oil matrices, and Method 3050 for solid matrices. A total of 262
digestions were performed, and appropriate blanks and controls were
included. Digestates were analyzed by flame atomic absorption (FLM)
and/or graphite furnace atomic absorption (GFAA) according to the
requirements of the methods. A total of 5,050 analyses was performed
using the method of standard additions.
The results of category 1 experiments provided information to evaluate
the usefulness of each method for solubilizing the metals in various
matrices. The performance criteria used for this evaluation were as
follows: recovery + 15 percent of the expected value (100 percent)
was good; recovery between ± 15 percent and 30 percent of the expected
value was fair; and recovery greater than ± 30 percent of 100 percent
is poor. Although the criteria do not constitute a rugged statistical
test, they represent reasonable performance levels in complex
matrices.
The results of category 2 experiments provided single—laboratory
accuracy (as shown by the recovery of spikes) and precision at three
different levels. The precision of the method was estimated from the
category 2 experiments by calculating the coefficient of variation
(CV): a Cv of less than 10 was considered good; a CV of more than 30,
poor; and fair was in between these two limits. A sununary of results
for both categories is listed in Table II.
The limit of detection CLOD) for this study was defined as the
concentration that would yield an absorbance equal to three times the
standard deviation (SD) of a series of procedural blank measurements,
or the concentration that would give, after a series of measurements,
a percent relative standard deviation (% RSD) of 33.3 percent [ 5].
Sample analytes exhibiting an instrument response greater than 10
times the flame LCD were always measured by the preferred FLAA
technique. Moreover, the concentration equivalent to 10 times the
FLAA LCD was defined as the maximum sample concentration appropriate
for GFAA analysis.
Although all matrices were digested and analyzed for all the required
elements, data reported here for Methods 3020 and 3050—GFAA include
only those results within the range of applicability of GFAA.
Because of this limited applicability of GFAA analysis and because
most metals are regulated and considered hazardous by the Resource
Conservation and Recovery Act (RCRA) at a much higher concentration
level (in the 1 to 10 ug/mL range) [ 6,7), the usefulness of Methods
-------
3020 and 3050—GFA1 for waste samples is limited. In the event the
analysis of ordinary samples at levels of environmental interest is
easily satisfied by FLAI , analysis by GFAA is neither necessary nor
recommended. An alternative analytical technique like ICP should be
explored for metals of interest. GFA1 is less cost effective and more
time consuming and labor intensive than FLAA or ICP; therefore, it
should be used only when FLAA or ICP are not applicable.
In all tables showing analytical results, elements that were found at
concentrations below the LCD are reported as less than the LCD (<).
When the absorbance observed was less than the analytical blank, B is
reported in the tables. The LCD is matrix dependent and was
calculated as three times the average noise level determined before
each analytical run.
Because the samples used in this study were of actual wastes, many of
the elements were present in very high concentrations. Therefore many
of the data reported here are above the linear range of FLAA. The
actual measurements, however, were all made within these linear
ranges. Digestates were diluted prior to analysis to bring the
elemental concentrations within range.
C1 TEGORY 1 RESULTS
category 1 experiments consisted of unspiked and spiked duplicate
digestions of four wastes by Methods 3010, 3020, 3030, and 3040 and
the same treatment of six wastes by Methods 3050—FLAA and 3050—GFAA.
The digestates were analyzed by FLAA and/or GFAA to determine
concentrations of the required elements.
Method 3010
The analytical results obtained for duplicate unspiked and spiked
aqueous waste digestates using Method 3010 are shown in Table III.
Due to the originally high concentration of several elements, they
were not added as spikes.
To evaluate the applicability and usefulness of Method 3010, the
recovery of spikes and controls was calculated. This information was
not obtained for elements that were not spiked because of their
originally high concentrations (e.g., Na).
The performance criteria defined earlier were used to classify the
usefulness of Method 3010. Be, Cd, Cr, Cu, Pb, and Zn exhibited good
recovery from both the samples and control, whereas Ba and Mn
exhibited poor recovery from the samples but good recovery from the
control. Matrix composition obviously affected the recovery of these
two elements. This conclusion can also be extended to Fe, Ni, and V
although the effect was less pronounced for them. Sb was the only
element for which control recovery was significantly higher than the
recovery from the sample matrix and was above the expected value. The
217.6 nun line used for Sb analysis by FLAA had an interference with Pb
-------
(which is present in all spiked samples and controls), which resulted
in a false high absorbance value for Sb. All calculated recoveries
were higher than the actual value.
Method 3020
Analytical results for duplicate unspiked and spiked aqueous
digestates obtained by Method 3020 are listed in Table IV.
Measurements were all performed at the linear range of GFAA; to do so,
the digestates were in some cases diluted more than would be
recommended to maintain good performance of the overall procedure.
Therefore, only those elements determined within the range of
applicability of GFAA are reported here. Because of the originally
high concentrations of some elements, they were not added as spikes.
The determination of Be, V, and Ti by GFAA presented difficulties.
Alternative analytical techniques like ICP should be evaluated for
these metals.
Ti analyzed using the IL instrument presented difficulties in unspiked
Method 3020 digestates of sample 2258—9—A—3A. Complete suppression of
Ti response occurred. Further dilution did not improve the assay. No
response was observed from the sample or from Ti spikes added to the
sample. This analysis was attempted twice at different dilution
factors on the IL instrument. It was later analyzed on the
Perkin—Elmer instrument.
The analysis of V by GFAA was not consistent throughout this project.
While analysis of Method 3050 samples proceeded reasonably well, the
attempt to analyze Method 3020 samples failed. Poor peak shape and
inconsistent response characterized the entire run, and eventually
there was no differentiation of response. This lack of success is not
considered related to the samples or the preparation method. Such
random behavior has been observed before in the variations in graphite
tubes. This behavior may be explained by surface area, number of
active sites, and oxygen content of the purge gas as previously
described [ 8]. Deterioration of the graphite tube prevented the
determination of Be in spiked sample 2258—9—A—4A digested by Method
3020.
Quantitative data for the calculation of recoveries from controls were
obtained for only a few elements. The others were either above the
range of applicability of GFAA or were not successfully determined.
Recovery from control samples was obtained for Cr, Cu, Ni, Pb, Mn, P1,
and V. Good recovery was found for Mn and V, acceptable recovery was
obtained for Cu and Ti, and poor recovery was obtained for Cr and Pb.
Method 3030
All samples were analyzed first by FLAA and, for those elements below
the FLAA LIOD, GFAA was performed. The analytical results obtained for
duplicate unspiked and spiked oil waste samples digested by Method
3030 are shown in Table V.
-------
The determination of Ba, As, Se, and Ti presented difficulties that
adversely affected many analyses. Ti analysis of Method 3030 unspiked
digestates required extra dilution to minimize background levels.
Ba was impossible to analyze in Method 3030 digestates because it is
insoluble in sulfuric acid (H2S04). To confirm this insolubility,
equal levels of Ba standards were prepared in 15 percent H2S04 (the
matrix of 3030 samples) and in 0.5 percent nitric acid (11N03). The
absorbance of the standard in H2S04 was initially low relative to the
standard in HNO3 and decreased to near zero over a period of hours.
As was analyzed twice in digestates from Method 3030. The analysis
was first done in a used pyrolitic graphite tube; this method yielded
poor results. The sample was later reanalyzed in a tantalum—coated
tube, and an improvement in the response resulted.
Since there was no commercially available oil—soluble Tl standard, no
measurement of this element was performed.
The determination of Se by GFAA in oil samples prepared by Method 3030
was unsuccessful. Se response was almost totally suppressed in these
samples. A spike of 40 mg Se/mL added to the Method 3050 control gave
a response greater than 50 percent of full scale. The same spike
added to the Method 3030 control yielded only 1 to 2 percent of full
scale. In the case of Method 3030, the sulfate matrix is suspected as
the cause. Separation techniques may well be appropriate in this
case.
The measurement of Pb in oil—matrix samples prepared by Method 3030
was not straightforward. Pb in such samples occurs in a number of
organic and inorganic forms. Organolead compounds may be lost by
volatilization at the H2S04 reflux temperature. In addition, losses
could occur due to the insolubility of lead sulfate (PbSO4). If the
difficulty with determination of Pb after sample preparation by Method
3030 arises from the insolubility of PbSO4, a slight modification to
the method could solve the problem. Working up sulfate samples for Pb
determination in geological samples is given in Reference [ 9]. The
referenced procedure involves dissolution of PbSO4 with ammonium
acetate to form the soluble lead acetate. Barium acetate is also
quite soluble in water. The revised digest method incorporating this
change follows: (1) digesting as described in 3030; (2) transferring
the digestate and one wash to the volumetric flask; (3) boiling a
small amount of 50 percent HC1 in the flask; (4) adding some ammonium
acetate solution and bringing to a second boil; (5) transferring the
mixture to the volumetric flask; and (6) bringing up to volume with
several washes.
Since BaSO4 does not come out of solution until the acid is diluted,
solid precipitation occurs after transfer to the volumetric flask.
Perhaps the first dilution could be made in the Kjeldahl flask with
the first transfer delayed until the BaSO4 has settled out of
solution. Whereas HC1 is the preferred matrix for FLAA, it is not
recommended for GFAA. The substitution of dilute HNO3 should be
-------
investigated.
An alternative analytical technique like ICP should also be explored
for these metals. GFAA is less cost effective and more labor
intensive than FL AA and therefore should be used only when FLAA or ICP
are not suitable.
The cold vapor technique for the determination of Hg yielded good
results. However, a shortened version of the analytical procedure
might benefit samples prepared by Method 3030.
The recovery of elements from spiked and control samples was
calculated to determine the usefulness of the digestion method. A few
elements (Be, Cr, Hg, and Cu) exhibited relatively good recovery for
both samples and control. Several elements showed poor recovery for
both samples and control (As, Na, Pb, Mn, and Zn). Others showed
suppression of the recovery due to matrix effects (Cd, Fe, Se, and V).
As in the case of Method 3010, Sb had very poor (high) recovery that
can be attributed to Pb interference in the analysis. Ag exhibited
good recovery for the spiked samples and poor recovery for the
control; normally Ag has poor solubility in the presence of ions like
S04 (which is present in all digestates). If the wastes contained
chelating agents that complex the Ag ions. However, this would result
in good recovery for the samples and poor recovery for controls. High
recovery of the Ni control could be attributed to laboratory error.
The EPA Office of Solid Waste provided RTI with a preliminary report
by Mr. David Payne, Quality Assurance Coordinator, Region V, EPA, on
the evaluation of Method 3030 using Conostan standard solutions.
Payne’s preliminary results indicated that recoveries for Ba, Pb, Hg,
and Se were unacceptable, but that recoveries of Cd, Cr, and Ag were
good in one case and fair in another. These results are in agreement
with the ones reported here; Hg is the exception.
Method 3040
All samples were analyzed by FLM and for those elements below the
LOD, GFAA was performed. The analytical results obtained for unspiked
and spiked oil waste extracts using Method 3040 are shown in Table VI.
Ba, V, As, and Se could not always be determined by either FLAA or
GFAA, which is consistent with previous results. The measurement of
Pb in these oil matrices had problems. Pb is expected to be present
in oil wastes in several chemical forms, including inorganic compounds
that would be insoluble in xylene, the extracting solvent. This lack
of solubility should result in at least a partial separation during
aspiration.
The literature [ 10] suggests an alternative to extraction Method 3040
for the GFAA determination of As in petroleum products. J.H. Fabec
presents a room temperature, closed—system workup procedure that could
offer some advantages. Some of the data presented suggest equivalent
response for different As species. A good correlation is presented
between this procedure, neutron activation, and X—ray fluorescence. A
-------
slight modification to the procedure allows spiking with aqueous—based
As standards that would avoid any instability of the Conostan
oil—based standard this laboratory has seen in the past. No
information is presented in the manuscript on the fate of other
metals.
There is also a total—metals—in—oils procedure described in R.E.
Kauffman et al. [ 11). Here, the oil sample is treated with a mixture
of concentrated acids at room temperature. The sample is then
emulsified with a product called Neodol 91—6 and MIBK or kerosene.
Reasonable recoveries (>70%) for 13 metals are reported.
To determine the usefulness of extraction Method 3040, the recoveries
of spikes and controls were calculated. Relatively good recoveries
were obtained for samples and controls in the case of Be, Cd, Cr, Cu,
Fe, Mn, Ni, and Sb. Na and Ag appeared to have experienced strong
matrix effects as indicated by the low matrix recovery compared to the
relatively good control recovery. Pb showed good control recovery but
very high matrix recovery. V exhibited good recovery from samples and
very high control recovery, which was attributed to experimental
error. Zn showed higher sample and control recoveries than expected.
This can be attributed in part to the fact that one of the wastes had
very high Zn levels in the unspiked sample; the addition of a
relatively small spike could have led to substantial errors. This was
also true for Pb matrix recovery.
Method 3050—FLAA
The analytical results obtained for duplicate unspiked and spiked
solid waste digestates are shown in Table VII. Due to their
originally high concentrations, several elements were not spiked;
these elements were Na, Fe, Zn and, in some cases, Cu, Ni, and Cr.
Sample 2258—9—S—SB, of known composition, was supplied by the
EPA—Cincinnati Quality Assurance Office. A comparison of results with
previous analyses of this sample show good agreement. A discussion is
presented under the quality assurance/quality control section.
To determine the usefulness of digestion Method 3050—FLAA, the
recovery of spikes from sample matrices and a control solution was
calculated. Most elements (Ba, Be, Cd, Cr, Mn, Na, Ni, and Pb)
exhibited relatively good recoveries from both the samples and the
control. In addition, Fe, Sb, Ti, and Zn demonstrated good recoveries
from the control. As expected, Ag showed a very low control recovery
and only a fair recovery from the sample matrices due to the presence
of chloride. The Ag spiking level of the wastes was significantly
lower (100 ug/mL) than the Ag concentration in the control (1,000
ug/mL); therefore, it was expected that recoveries from the samples
(72%) and control (12%) would be low. A lower Ag concentration had
less of a chance of precipitating silver chloride (AgCl) out of
solution. Cu showed acceptable recoveries from the sample matrices
and control solution. Sb, Ti, and V consistently exhibited very low
matrix recovery from almost all sample matrices.
-------
Method 3050-GFAA
The analytical results obtained for unspiked and spiked solid
digestates are shown in Table ‘1111. The determination of Ba and V
encountered difficulties as explained previously and therefore no
results are reported. Alternative techniques such as ICP should be
evaluated for these metals.
Sample 2258—9—S—5B, of known composition, was supplied by the
EPA—Cincinnati Quality Assurance Office. A comparison of results with
previous analyses of this sample shows good agreement.
Useful quantitative data for the calculation of recoveries of spikes
was obtained for only a few elements: As, Se, and Sb (one matrix
only). Most of the elements that were determined were found to be
above the range of applicability of GFAA, and therefore no recovery
calculation is reported.
The overall recovery of Se from the spiked waste appeared to be very
good. In addition, the recovery of Se from the control was within the
acceptable range. The overall recovery of As from the spiked wastes
was also within the acceptable range although recovery of the control
was poor. Most of the Sb determinations are not reported because the
levels of Sb found were above the limit of applicability of GEM. One
waste (No. 2258—9—3—53), however, contained a relatively low level of
Sb and recovery calculations could be made. In this case, the
recovery of Sb was near zero.
C1 TEGORY 2 RESULTS
Category 2 experiments consisted of seven replicate digestions by all
methods (Methods 3010, 3020, 3030, 3040, 3050—FLAA, and 3050—GFAA).
The samples were digested unspiked, spiked at level 1, and spiked at
level 2. The digestates were then analyzed by FLM or GFAA to
determine concentrations of the elements under evaluation.
The accuracy of the methods was estimated by calculating the percent
recovery of the two spiking levels. The performance criteria for
evaluating method accuracy were the same as those applied to category
1 experiments (recoveries within ±15% of the expected value were good,
between ±15% and ±30% were fair, and greater than ±30% were poor).
Method 3010
The precision of Method 3010 was evaluated by examining the CV by
element. Cd, Cr, Cu, Fe, Ni, Pb, Sb, and Zn had consistently low CVs
(<10) at all spiking levels. Ba, Mn, and V had consistently higher
CVs (>10) at all spiking levels. Be had a high CV in the unspiked
sample and Na had a high CV in the spiked level 2 sample. Method 3010
demonstrated good precision (Cv<10) for the majority of elements.
-------
Because of the intrinsically high concentration of some elements in
the waste, Cr, Cu, Fe, Na, Ni, and Sb were not added as spikes and no
recovery data were obtained for these elements. In addition, the
performance criteria described earlier were applied to the recovery as
a measure of accuracy. Good accuracy was obtained for Cd and Zn and
fair accuracy was obtained for Be, Pb, and V. Poor accuracy was
obtained for Ba and Mn. These results indicate that the accuracy of
Method 3010 is element dependent and that no generalizations can be
made from results obtained by it.
Method 3020
All elements were determined by GFAA. Most of them, however, exceeded
the limit of applicability of GFAA and therefore are not reported
here. Ba, Tl, Be, and V determinations encountered the same
difficulties described for category 1 experiments. Quantitative
information was obtained for Ag and Ti only. The CVs calculated for
these two elements were high, indicating poor precision at the
reported levels.
Method 3030
Elements were determined by FLAA, except for As (determined by GFAA)
and Hg (determined by the cold vapor technique). Ti was not spiked
because no commercial oil standard was available for this element.
The precision of Method 3030 (as shown by the CV) can be examined by
element. Cu exhibited good precision (CV<10) at all spiking levels.
Ag, Cd, Cr, and Na exhibited good precision for the two spiked
samples. As, Be, Fe, Mn, Sb, and V exhibited significantly better
precision for the two spiked samples than at the unspiked level. Hg
and Pb had CVs between 10 and 20 for at least two of the spiking
levels. Ni and Zn exhibited extremely poor precision at one of the
spiking levels. For most elements, the precision of Method 3030 was
better at the higher spiking levels.
In terms of accuracy, the recovery of spikes indicates that Method
3030 demonstrates good accuracy for Ag, Be, Cu, and Hg, and good or
fair accuracy for at least one spiking level for Cr, Mn, and Na. Poor
recovery was obtained at all spiking levels for As, Cd, Fe, Ni, Pb,
Sb, v, and Zn. Thus the accuracy of Method 3030 is element dependent
and the majority of elements tested exhibited poor recovery.
Method 3040
All elements were determined by FLAA. The precision of Method 3040
can be estimated by evaluating the CV of each element. Na, Pb, and Fe
showed good precision at all spiking levels. Ag, Cd, Cr, Cu, Mn, and
V exhibited good precision at the two higher spiking levels. Be and
Sb showed poor precision at the unspiked level. Zn had fair precision
at all levels and Ba had poor precision at all levels. For most
-------
elements studied, the precision of Method 3040 was good at the higher
spiking levels.
The accuracy of Method 3040 can be estimated from recovery data. Cd,
Cu, and Mn exhibited good recovery at all spiking levels. Be, Cd, Fe,
Na, and Sb showed better recovery at the higher spiking level. Ni, V,
and Zn demonstrated poor recoveries. Ag and Pb behaved anomalously;
they had considerably better recovery at the low spiking level than at
the higher spiking level. The behavior of Pb may be attributed to the
intrinsically high initial concentration of Pb in the waste compared
to the added amount of spike (which was part of the Conostan S—21
standard mixture). However, both the Ag and Pb response was more
likely due to experimental error, since the category 1 results for
these two elements do not agree with these data.
Method 3050—FLAA
The precision of Method 3050—FL1 A can be estimated from the CV data.
For Ba, Cd, Cr, Cu, Fe, and Mn, the precision appeared to be good at
all levels of spiking concentrations. Be, Ni, Pb, Sb, Ti, and V
exhibited better precision at the higher levels of spiking. Ag and Zn
had better precision at the lower concentration (unspiked sample). Na
demonstrated relatively good precision independent of spiking level.
Overall, Method 3050—FLAA had relatively good precision for most of
the elements studied.
The accuracy of Method 3050—FLM can be estimated from the recoveries.
Cu, Fe, Na, and Zn were not spiked because of the intrinsically high
concentrations of these elements in the waste. Therefore, no recovery
data were obtained for these elements. Be, Cd, Mn, Ni, and Sb
exhibited good recovery for both levels of spiking. Ba, Cr, Pb, Ti,
and V showed at least fair recovery for one of the spiking levels. Ag
recovery was poor at all levels of spiking; this is not unexpected
since HC1 is present in the matrix and AgCl would not be soluble at
the experimental conditions.
Method 3050-GFAA
All specified elements were determined. However, most of them were
found above the limit of applicability of GFAA and therefore are not
reported here. The precision of Se determination by Method 3050—GFAA
was relatively good especially at the higher spiking levels. As
exhibited acceptable CVs at the higher spiking levels but a rather
poor CV at the unspiked level. The recovery of As was good at all
levels and the recovery of Se was fair at all levels.
CONCLUSIONS AND RECOMMENDATIONS
-------
The overall evaluation of methods was performed based on the results
of category 1 experiments, which yielded recovery information as a
measure of the usefulness of the methods. In addition, category 2
experiments provided recovery information as an estimate of the
accuracy and coefficient of variation as an estimate of precision.
Method 3010
Method 3010 is applicable for the preparation of aqueous wastes for
flame atomic absorption (FLM) analysis, and it yields good precision
and accuracy for Cd, Pb, and Zn. In addition, Method 3010 can be
recommended with caution for Be, Cr, Cu, Fe, Na, Ni, and V. It is not
recommended for Ba and Mn. Data for Sb are inconclusive. It is also
recommended that the written protocol he modified to give the analyst
more guidance when encountering difficult samples.
Method 3020
Because of the limited data obtained on this method, no conclusions
regarding its applicability could be made. However, Method 3020,
which is specific to the preparation of aqueous wastes for graphite
furnace atomic absorption (GFAA) analysis, is more labor intensive (by
a factor of 2) and of limited application. Use of the method, when
applicable, should be restricted only to those cases in which the
limit of detection (LCD) of FLAA is too high to justify using flame
techniques. Further research to evaluate inductively coupled argon
plasma (ICP) spectroscopy as an alternative analytical technique to
GFAA is recommended. As with Method 3010, Method 3020 should be
revised to give the analyst more guidance and to accommodate samples
that show difficult digestions.
Method 3030
Method 3030 is applicable for the preparation of oil wastes for FLAA
or GFAA analyses, and it offers good precision and accuracy for Cu.
Method 3030 can be recommended with caution for Ag, Be, Cr, and Hg.
It is not recommended for As, Ba, Fe, Mn, Ni, Pb, Sb, Se, and Zn. The
results for Cd and V are inconclusive.
The above conclusions and recommendations for Method 3030 apply only
to oil—soluble organometallic compounds and should not be extrapolated
to inorganic forms of metals present as emulsions or suspended solids
in oil wastes. Further research is needed to develop and evaluate
digestion methods for inorganic forms of metals suspended or
emulsified in oil wastes.
Method 3040
Method 3040 is applicable for the extraction of Cu from oil wastes and
has good precision and accuracy for that element. It can be
-------
recommended with caution for Cd, Cr, Fe, and Mn. Method 3040 is not
recommended for Ag, As, Ba, Ni, Se, V, and Zn. Results are
inconclusive for Be, Na, Pb, and Sb.
The recommendations made for Method 3040 apply only to oil—soluble
organometallic compounds and no conclusions or recommendations should
be extrapolated to inorganic forms of metals present in oil wastes as
emulsions or suspended solids. Further research is recommended to
develop and evaluate extraction methods for inorganic forms of metals
suspended or emulsified in oil wastes.
Method 3050—FLAA
Method 3050—FLAA has wide applicability and demonstrates good
precision and accuracy for Ba, Be, Cd, Cr, Mn, and Pb. It is not
recommended for Ag, V, and Zn. Method 3050—FLAA is recommended with
caution for Na, Cu, Ni, and Tl. Results for Sb and Be are
inconclusive.
Further research is recommended to evaluate other analytical
techniques that do not require sample digestion that can modify the
original sample. Possible alternatives are X—ray photoelectron
spectroscopy for qualitative and quantitative analysis of solid wastes
and the determination of valence states on solid samples, which may be
crucial for the toxicity of some metals (e.g., Cr III vs. Cr VI).
Method 3050-GFAh
Limited data were obtained for Method 3050—GFAA. Based on available
data, however, the method is not recommended for Sb and is recommended
with caution for As and Se. As explained for Method 3020, use of
Method 3050 followed by GFM, when applicable, should be restricted to
cases in which the LOD of FL A is too high to justify using flame
techniques. It is recommended that ICP be considered for evaluation
as an alternative analytical technique. ICP is a fast multielement
analytical technique, applicable to the determination of metals in
environmental samples with reported limit of detection in ng,k L for
some elements.
-------
REFERENCES
1. U.S. Environmental Protection Agency. 1982. Test Methods for
Evaluating Solid Wastes, Physical/Chemical Methods, SW—846, 2nd
Ed., U.S. Environmental Protection Agency, Washington, D.C.
2. Title 40—Protection of Environment. 1982. Part 261, pp. 354—
377 (July 1).
3. Federal Register . 1980. Vol. 45, No. 98, Monday, May 19, p.
33123.
4. Sotera, John 3., Bancroft, Martha Fogg, Smith, Stanley B., and
Corum, Timothy L. 1981. Instrumentation Laboratories Atomic
Absorption Methods Manual, Vols. 1 and 2.
5. U.S. Environmental Protection Agency. 1979. Methods for
Chemical Analysis of Water and Wastes, Environmental Monitoring
Systems Laboratory, U.S. Environmental Protection Agency,
Cincinnati, OH.
6. U.S. Environmental Protection Agency. 1979. A Handbook of Key
Federal Regulations and Criteria for Multimedia Environmental
Control, Interagency Energy/Environment R&D Report, EPA—600/
17—79—175, U.S. Environmental Protection Agency, Washington,
D.C. (August).
7. Kohl, Jerome, and Triplett, Brooke. 1982. Hazardous Waste
Management Under RCRA, Industrial Extension Service, School of
Engineering. N.C. State University, Raleigh, N.C.
8. Sturgeon, R.E., Sui, K.W.M., and Berman, S.S. 1984. Oxygen in
the high—temperature furnace, Spectrochimica Acta 39B: (2/3):213.
9. Brandvold, Lynn A. Atomic Absorption Methods for Analysis of
Some Elements in Ores and Concentrates, Circular *142, Bureau of
Mines and Mineral Resources.
10. Faber, J.H. 1982. Direct determination of arsenic in shale oil
and its products by furnace atomic absorption spectrometry with
tetrahydrofuran solvent system, Anal. them . 54:2170.
11. Kauffman, R.E., Saba, C.S., Rhine, W.E., and Eisentrant, K.J.
1982. Quantitative multielement determination of metallic wear
species in lubricating oils and hydraulic fluids, Anal. Chem .
54:975.
-------
TABLE £ WASTE SAMPLES FOB METALS EXTRACTION
RTI
Type/origin
Hazardous
waste No Description
ass igneçsample
pH code
Ott
Samples
NAc Used lubricant motor oil; black NA 2258-9-0—IA
viscous liquid
NA Used diesel oil from shipyard, NA 2258-9-0-28
brown liquid composite of two
wastes
NA Used diesel oil from shipyard, NA 2258-9-0- lA
black, very viscous, semi—solid
EHSL-EPA, Cincinnati NA Oil/water emulsion from cold NA 2258-9-O-5A
rolling facility, representative
of animal fat oils
Aqueous Samples
FOO l Spenl nickel-plating solution from —5 2258-9-A-3A
metal finishing operation, clear
emerald green liquid with solids
F 009 Spent stripping and cleaning ‘ -13-14 22 58-9-A-38
bath solution from metal
finishing operation, foamy/cloudy
yellowish liquid
FOOl Spent chrome plating solution from —O 5 2258-9-A-3C
metal finishing operation, dark
amber liquid with suspended solid
FOOl Chromic acid waste solution with — l 2258-9-A—4A
proprietary catalyst from
electroplating operation, dark
amber liquid with bright orange
suspended solid
Solid Samples
F006 Waste water treatment sludge from —12 2258-9-S-3D
plating opmration, black solids
with moisture
P006 Prectpitsted metals wsste sludge —6 22 58-9-S-48
from electroplating operation,
brown solids with moisture
EflSL-EPA, Cincinnati NA Musiicipsl sludge standard sample; NA 22 589S58
QA Branch dry powder, grey
EMSL-EPA, Cincinnati P006 Sludge resulting from lime NA 2258-9-S-SC
neutralization of wsstewster
from an electroplating facility,
contains bound cyanide and
sulfide
OSW-EPA, Washington NA AT! separator sludge, black NA 22589S6A
solids suspended in liquid,
strong odor
OSW-EPA. Washington NA Soil contaminated with metals NA 2258-9-S-6 9
a reference (3)
boesignates Project No -Sample Type-Origin
cNA = Not svsilshle or not spplicsb le
-------
TABLE II. SUMMARY OF RESULTS
Method/
element
Percent
recovery CVC
b
Category 2 Level 0 Level 1 Level 2
a
Category 1
3010
Ba
Be
Cd
Cr
Cu
Fe
Mn
Na
Ni
Pb
Sb
V
Zn
3020
Ag
Ti
3030
Ag
As
Ba
Be
Cd
Cr
Cu
Fe
Hg
N f l
Na
Ni
Pb
Sb
Se
V
Zn
3040
Ag
As
15
1/—2
10.1
12.6
43.6
89
85/81
35.2
2.0
1.1
93
99/99
3.9
3.0
1.5
90
6.4
1.0
1.3
91
--
3.5
1.1
4.6
84
--
3.7
3.1
3.6
69
27/54
27.9
27.5
15.6
--
--
2.8
5.0
13.2
79
--
7.2
5.5
5.0
91
89/79
2.1
1.6
2.0
81
-—
2.2
4.0
1.3
77
55/76
17.4
20.9
13.2
109
91/92
6.5
6.6
7.3
--
--
37.0
32.0
--
--
--
--
110.0
--
100
88/88
24.7
6.5
2.1
52
26/53
125.0
21.0
43.0
86
100/94
55.8
12.1
19.0
84
26/60
61.0
4.1
5.1
92
75/79
13.3
3.8
4.4
95
101/99
8.1
4.2
3.6
44
567/215
35.7
13.5
9.4
90
99/94
--
18.8
16.6
117
43/79
44.0
14.6
9.8
11
77/85
19.0
2.9
5.1
88
27/39
35.4
137.0
22.2
48
—4/-4
18.2
13.5
39.5
193
167/180
40.5
13.5
14.2
81
58/52
81.9
30.1
20.1
182
-671/931
17.6
23.6
98.4
120
95/56
44.5
3.5
11.7
Ba
96
--
59.2
67.3
52.1
Be
74
69/82
79.4
7.5
14.4
Cd
123
68/84
13.2
8.5
4.1
(continued)
-------
TABLE II. (continued)
305 0-GFAA
As
Sb
Se
aRecovery data from category 1 results represent the mean over all matrices.
bR data from
CCff.i of variation at
Method/
element
Percent
recovery CVC
b
Category 2 Level 0 Level 1 Level 2
Category
1 a
3040 (con.)
Cr
Cu
Fe
Mn
Na
Ni
Pb
Sb
Se
V
Zn
3050-FLA.A
Ag
Ba
Be
Cd
Cr
Cu
Fe
Mn
Na
Ni
Pb
Sb
Ti
V
Zn
102
74
111
130
111
120
98
79
164
130
72
99
93
109
106
126
102
95
89
97
50
70
47
326
75
0
99
110/90
94/89
52/74
99/99
12/72
19/48
93/—25
68/80
885/312
—47 1/47
36/18
72/88
90/9 1
104/99
72/83
93/98
86/ 101
64/92
9 1/91
70/75
69/8 1
102/94
79/79
43.6
11.3
2.8
35.1
7.8
47.4
9.0
46.6
29.6
8.6
3.8
3.9
93.8
2.9
5.0
1.7
4.3
5.3
1.3
98.9
11.3
31.5
91.4
8.6
59
19
4.3
5.5
2.6
12.9
7.5
55.5
9.1
14.2
5.9
10.9
5.9
2.6
2.7
1.8
1.5
2.7
5.4
2.5
10.2
19.0
4.2
11.8
8.9
11.9
19.6
13
8.2
1.2
4.1
10.1
7.3
8.8
5.9
7.8
6.7
8.7
14.6
24.0
2.0
4.8
1.3
2.1
3.7
3.7
3.5
2.8
5.1
1.7
4.3
5.8
1.6
18.8
11
6.4
category 2 results are reported at two levels of spikes.
three concentration levels.
dD not available; see Discussion.
-------
TABLE Ill
RESULTS OF
FLAA
ANALYSIS OF AQUEOUS WASTE SMPLES BY METHOD 3010,
CATEGORY I
Element
Sample No Ba
Be
Cd
Cr
Cu
Fe u n Na
Ni
Pb
Sb
V
Zn
Unapi ked
2258-9-A-3A ‘2 60
<0 960
2258-9-A-3B I 08
CO 600
(0 160
0 130
0 332
0 312
<0 720
(0 130
‘0 0600
‘0 051
< I 60
(2 00
-------
Blank
I
Control
a
LOD. are
b ld on Perkin-Elmer Instrument
Below the analytical blank
dull analyse.
eTbC.C element, were determined above the level of GFAA applicability
t Control v .a prepared at approximately 0 25 pg/m I.
TABLE IV RESULTS
OF GFAA
ANALYSIS OF AQUEOUS
WASTE SAMPLES BY
METHOD 3020,
CATEGORY
I
Element (pg,mL)a
Sample No
Ag
Ba
Be
Cd
Cr
Cu
Fe Mn
Ni
Pb
TI
V
Zn
Unapiked
2258-9-A-3A
0038
(0 0038
d
--
e
t
(0 0021
<0 0045
<
(0 23
0 14
<0 032
t t
? t
t
t
0 0095
0 072
_d
--
--
,
t
2258-9-A-3B
(00038
(00038
--
--
<00026
‘00026
‘00018
<00018
1 S
0 85
0 19
0 13
0 23 0 0036
0 16 <00035
(0 19
<0 19
0 20
0 14
(00032
(00032
--
--
t
t
2258-9-A-3C
0 043
0075
--
—-
t
1
t
1
t
1
I
t
I I
I I
I
I
1
1
0 0053
(00016
--
--
I
I
2258-9-A-4A
0 84
095
--
—-
t
I
t
t
t
I
t
I
t
I I
f
I
1
<0 0016
‘00016
--
--
I
I
Blank
0 00019
--
<0.017
BC
0 0056
0 0021
B 0 00035
B
0 0030
0 00032
--
0 0014
cootroi
0 14
--
t
t
0 19
0 24
1 0.25
0.14
0 16
0 13
--
I
Spiked
2258-9-A-3A
BC
B
--
tb
t
t
t
t
I
t
t
t t
I I
tb
I
t
1
<0 26 b
‘026
1 b
1
1
2258—9-A-3B
B
B
--
——
,
t
t
,
t
I
t I
t
I
t
1
-------
LODa are reported a. leap than ( <)
bAnalyzed by GFAA
CMS yap analyzed uaing cold vapor procedure
analyaju atle pted, see text for explanation
eB Below the analytical blank
TABLE V
RASULTS OF LAA
ANALYSIS OF
OIL WASTE
SAMPLES BY METHOD 303O CATEGORY
I
Elelnent(IJg/g)a
Saeple No A8 Ba Be
Cd
Cr
Cu
Fe
NBC
Ho
Na
Ni
Pb Sb TI V Zn
Unapi ked
2258-9-0—lA
B
B
(0 26
(0 26
--
‘0 07
<0 oB
I
0
07
89
6
4
71
09
19
19
2
4
386
294
B
<0 024
10 6
Il 5
189
174
<2
(4
60
70
27 0
IS I
<21
(22
0
0
B
B
(24 0
(25 0
l 420
668
2258-9-0-2*
0 614
B
(0 26
(0 26
--
--
(0 I
<0 1
I
1
IS
04
<4
(2
10
80
IC
9
6
4
113
95
8
B
<0 024
<0 9
(0 82
13 6
19 5
4
<1
71
70
23 0
21 1
67
65
9
4
B
B
‘34 0
<160
.30 3
26 I
2258-9-0-28
B
B
(0 26
<0 26
--
--
<0 05
(0 05
0
0
42
42
<3
(2
40
00
6
6
80
90
149
175
B
B
CO 82
‘I 30
9 60
10 5
(2
-------
TABLE VI RESULTS OF FLAA NALYSIS OF OIL WASTE SANPLES BY IIETHOD 3040. CATEGORY I
Element
Sample No Ag Bab Be Cd Cr Cu Fe Ho Na Ni Pb Sb Tib V Zn
Unapiked
2258-9-0-IA ‘2 00 -. ‘0 03 Bd (4 50 14 2
<2.20 -- ‘0 04 B (6 90 19 3
2258-9-0—2* <1.40 -- 0 08 B ‘S 60 2 14
<1.40 -- (0.07 B (I 30 (3 60
2258-9-0-2B ‘1.60 -- ‘0 06 B <5 00 1 60
(1.40 -- (0.03 B (2 30 2 46
2258-9-0-5* (0.96 -- ‘0.05 B 7 18 6 72
‘0.3 —— (0 II B 5 75 4 25
Blank (0 74 -- (0 05 ‘0 3 (2 00 (0 95
Control 110 -— 101 157 118 96 6
2258-9-0-IA 174 ‘54 0 151 169 206 205
174 (80 0 136 172 210 196
2258-9-0-2* 38 4 (920 219 201 175 184
45 1 (750 178 211 180 198
2258-9-0-28 174 (96.0 249 209 189 202
176 <140 237 204 189 189
2258-9-0—5* 139 (230 230 218 256 188
152 106 249 226 260 192
Blank <0 78 (13 0 ‘0 36 ‘0 18
-------
TABLE VII RESULTS
OF FLAA ANAlYSIS (IF SOLID WASTE SAMPLES
BY METHOD 3050.
CATEGORY I
Element (/)i
Sample No Ag 8. Be
Cd
Cr
Cu
Fe
No
Ni
Ni
Pb Sb Ti V Zn
Un.piked
2258—9—S—3D 23 6 (180 (I 90 <1 20 10,200 11.900 19.000 133 118.000 87,400 260 (260 (30 0 (760 537
24 4 (120 0 600 (2 40 29.700 10.000 22.400 126 134.000 82.100 148 (130 (30 0 (650 541
2258-9-S-4B 7 00 396
-------
TABLE Viii
RESULTS 01 GFU ANALYSIS OF SOLID WASTE SAIIPLES BY METHOD
3050.
CATEGORY I
Element (IJR/g)’
Sample No Ag A. Ba
Be
Cd
Cr
Cu
Fe
Pin
Ni
Pb
Sb Sr TI V Zn
Unapiked
2258—9—S-3D 1 5 t t t 3 t t 3 t 3 <0.014 CO 033 -— 3
3 1 7 -- <0 075 3 3 3 3 3 t t 3 (0 074 0 84 -- 3
2258—9—S-48 3 7 3 -- <0 075 3 3 1 3 1 (32 3 3 <0 074 (0033 -— I
3 8 3 - - 3 t t I t I (32 t 3 053 (0033 —— I
2258-9-S-SB 3 (0 54 -- 3 3 3 3 3 3 (64 3 6.1 1 0 0.089 —- t
3 071 -— <0075 1 t I t t I I II 21 0.16 —— t
2258-9-S-SC 3 (0 54 -- t t t 1 3 1 3 1 <4 0 <0 15 <0 66 -- I
3 (054 -— 3 1 3 1 3 3 3 3 (40 051 031 -— I
2258-9—S-SD 1 20 -— 3 3 t I 3 3 (32 1 (40 (0074 0051 -— I
3 6.6 -— 1 3 3 t I t ‘32 t (4.0 032 (0.033 -— I
2258-9-S-6A 0 29 0 63 —- 3 1 3 3 1 3 t 3 (4 0 27 (0.033 -- I
1 (0.54 —— 3 1 t I I I 1 (40 20 (0033 —— I
0.0066 0 54 -- (0.0015 0 0090 0 48 0 061 8° 0 035 B 0 041 B (0.14 0 0090 -- I
COntrOl 8 t 13 -— I I I I 3 3 3 1 1 13 3 -— I
Spiked
2258-9-A-3D 1 d 83 b __b,c C tb 3 b t i 30 b t < 1600 b
1 110 -— I I I I I I I 1 3 130 1 <1,600 1
2258-9-S-4B 3 140 -— 3 3 3 3 I 1 1 1 3 100 3 I ,600 I
3 96 -- 1 3 3 1 1 1 1 3 t 110 3 <1.600 3
2258-9-S-SB 3 87 -- 1 3 3 3 1 1 3 3 t 79 3 3 3
1 76 -- 3 1 3 3 1 1 3 t 3 85 3 1 3
2258-9-S-SC t (19 -— I I 1 1 1 1 1 3 (2 5 34 3 <1,600 t
1 (19 -- 3 3 3 3 3 t t 3 32 26 3 (1,600 1
2258—9—S-5D 3 73 -- 1 1 1 1 3 t 3 1 1 140 t I I
3 98 -- 1 3 3 3 1 t 3 3 1 130 3 1 1
2258-9-S-6A t 70 -- I t I t 3 1 t I 1 160 1 1 1
3 71 -— I 1 3 1 1 3 t I 3 ISO t 3 3
Blank 1 0 43 -- I ll 0 67 3 1 B B B <0 25 2 9 (0 78 (1,600 1
Control 1 1 970 -- I I I I I I I 1 1 1,000 1 ‘1.600 1
•LOD. are reported as lena than (C) The IA)lJ is calculated as three average no,.e level determined before each .n Iytical run
bAnalyala periorned with the Perkin- lmer AAS Inbtrumrnt
CUnCuc I.el.sful analysia. see leSt for expl .ndtIo.I
dlIrteruiiiuIi.,l t i l.. .,l,,,v. lii, liv, I i.l t.lAA .1.1.1 Ii ii,, I ,1 .
-------
INTERFERENCE REDUCTION STUDIES INVOLVING HYDRIDE GENERATION ARSENIC
AND SELENIUM D r it MINATIONS UTILIZING ATOMIC ABSORPTION AND PLASMA
EMISSION SPECTROIIETRY
Dr. J. WILSON HERSHEY, LANCASTER LABORATORIES, INC., LANCASTER,
PENNSYLVANIA; and Dr. PETER N. KELIHER, CHEMISTRY DEPARTMENT,
VILLANOVA UNIVERSITY, VILLANOVA, PENNSYLVANIA
HEALTH AND REGULATORY ASPECTS
Regulatory and environmental concerns have made the determination of
trace amounts of arsenic and selenium extremely important. Selenium
is an essential trace nutrient while arsenic is a cumulative poison.
Unlike synthetic organic chemicals, these “heavy metals,” of course,
are not man—made. Environmental pollution problems associated with
these elements, therefore, are the result of redistribution of the
naturally occurring elements by agricultural and industrial processes.
Arsenic, which accounts for approximately 5 x 10—4 percent of the
earth’s crust, is produced as a by—product of the mining and refining
of other minerals and is used commercially, primarily in pesticide
formulations and agricultural applications. Organic arsenicals,
hundreds of which have been synthesized, are widely used as livestock
feed additives and as therapeutic drugs. Inorganic and aliphatic
organic arsenicals are poisonous to all lower animals. The toxicity
varies greatly with formulation and valence state, with the most
dangerous preparations being the trivalent arsenites.
Selenium accounts for 10—5 percent of the earth’s crust with selenides
often being found in sulfide ores. Selenium is obtained from flue
dusts produced in the roasting of sulfide ores and as a by—product of
sulfuric acid manufacture. The largest current use of selenium is as
a food supplement to animal feeds.
Specific Environmental Protection Agency (EPA) regulations now apply
to arsenic and selenium in drinking water (1) and in Extraction
Procedure Toxicity Leachates (2) of sludges and solid wastes. Arsenic
(as As2O3) is regulated in food (3) while Occupational Safety and
Health Administration (OSHA) limits in air have been set to protect
workers’ health (4). These regulations are summarized in Table I.
Finally, many states regulate the levels of arsenic and selenium in
solid waste.
The object of our research was to develop improved hydride generation
methods for the analysis for these and other hydride—forming elements
in difficult matrices. Although hydride generation methods are widely
used, it has been known since the middle 1970’s (5) that these methods
are subject to interelement interferences. Although we studied six
hydride—forming elements to some extent, we will focus here on the
improvements made for arsenic and selenium.
arsenic (7061) and selenium (7741) in solid and liquid samples. The
-------
samples are prepared by a nitric/sulfuric acid digestion before being
converted to a volatile hydride. The hydride gas is swept into an
argon—hydrogen flame located in the optical path of an Atomic
Absorption Spectrophotometer (MS).
However, the currently published methods are approved only for sample
matrices that do not contain high concentrations of chromium, copper,
mercury, nickel, iTver, cobalt and molybdenium. “High level” is
undefined. Cautions are also issued in that certain arsenic and
selenium compounds are volatile and therefore may be lost during the
digestion. The use of spiked samples and relevant reference materials
is necessary to determine the applicability of the methods to a given
waste.
Although many real world samples can be accurately analyzed using the
SW—846 methods, many others contain high levels of interfering
elements. We have recently investigated interelement interference
reduction techniques in hydride generation AAS and Inductively Coupled
Plasma Emission Spectrometry (ICP). These included manipulation of
acid strength and sodium borohydride concentration and the comparison
of multiple hydride generation systems. Extensive studies were also
performed using both Chelex 100 and AG 50W resins as a means of
reducing the interferent to analyte ration in real world samples.
EXPERIMENTAL ASPECTS
Apparatus and Operating Procedures
The I.L. Model 951 dual channel Atomic Absorption Spectrophotometer
fitted with the conventional, single slot burner head was used for
these studies. All analyses were performed using an argon hydrogen
entrained air flame and the peak area integration modes.
The I.L. Model 100 sequential Plasma Emission Spectrometer was
utilized for the ICP hydride determination.
Three commercially available hydride generation devices were used in
this study. The I.L. manual device was designed to produce a
transient MS signal following the NaBH4/saxnple mixing. The P.T.
Analytical (Wilmington, Mass.) device was designed to produce a steady
state signal and was designed to be used with either MS or ICP
systems. The I.L. plasma hydride device (Figure 1) was designed to
produce a steady state signal for ICP determinations. All samples
were acidified to form Se (IV) or reduced with KI to form AS (III).
PRELIMINARY INTERFERENCE STUDY
The interelement interference effects of 50 elements were examined
during the hydride generation MS and ICP determinations of arsenic
and selenium. The 35 elements in Table II resulted in interelement
interferences of less than 10% when present at levels of 1,000 fold
greater than the analyte. The 15 elements in Table III exhibited
-------
interferences of greater than 10% for either As or Se, or both, at
interferent levels of 5 to 1,000 fold in excess of the analyte.
Representative recoveries are presented in Table III.
Figures 2 and 3 show typical interelement interference curves. The
analyte concentration was held constant at 0.100 ug/mi while the
interferent concentration was varied. These interferent levels were
selected to span the range from no interference to nearly complete
signal suppression. Hydride interelement interference was first
identified during the middle 1970’s, but even today only very
inadequate techniques for dealing with these interferences are
available.
INTERFERENCE REDUCTION TECHNIQUES
Effect of Acid Strength
Constant levels of interferents were prepared in 1.2, 2.4, 4.8 and 7.2
M HC1. These solutions also contained 0.100 ug/mi analyte. Thus, in
contrast to the interference study previously discussed, this work
varied the acid level while the interferent concentration remained
constant.
Variation in acid strength caused only small changes in the standard
readings but resulted in very large changes if significant levels of
certain interferents were present.
Higher acid concentrations greatly improved the signal for both
arsenic and selenium when the following interferents were present:
cobalt, copper, lead, molybdenum, nickel, palladium and rhodium. The
signals from standard solutions containing no interferents, while
changing slightly, were not affected to nearly the same extent. The
interference of gold and iridium on selenium was reduced, as was that
of tungsten.
Figures 4 and 5 present typical examples of this improvement. Slight
improvement was noted for the gold interference on arsenic and
selenium and for mercury interference on selenium. No reduction was
noted for the iridium and platinum interferences on arsenic. Higher
acid levels yielded no reduction for the germanium, platinum and
tellurium interferences on selenium.
It is interesting to note that a few interferences were less severe at
moderate acid strength than at either extreme. These include platinum
and tellurium (on selenium).
Only one interferent resulted in much poorer analyte recovery at high
acid strength. Arsenic produced poor selenium recovery at arsenic
levels of 100 ug4nl.
In summary, with the exception of arsenic and germanium, the effect of
all interferents remained essentially constant or was significantly
reduced at high acid concentration for either or both arsenic and
-------
selenium. Thus, the final acid level in real world samples becomes
critical if a hydride determination is to be performed.
Cation Exchange Separation
Several resins were evaluated to provide a means of separating the
interfering elements from the hydride—forming elements. The goal was
to develop a clean—up procedure that was rapid, inexpensive,
relatively independent of pH and applicable to a wide variety of real
world samples.
The following table compares the results obtained following a
minicolumn resin separation.
The elements that had previously been found to interfere most severely
with hydride generation were subjected to the mini—column tretment at
pH 2 on both AG SOW—X16 and Chelex 100. The percent removal is shown
in Table IV.
As can be seen, levels of six of the elements (cobalt, copper, lead,
mercury, nickel and silver) were reduced by at least 96%. Gold and
palladium were removed by greater than 98% on the Chelex resin.
Platinum and rhodium were not removed to a significant extent on
either resin. Bismuth removal was greater than 98% on Chelex 100
while tellurium removal was greater than 75% on both resins.
Arsenic and selenium were not removed to a significant extent. That
is, they both passed through the resin without being retained.
APPLICA TIONS
Reference Materials
These clean—up techniques were applied to various standard reference
materials and to real world samples. Table V provides results
obtained on EPA WP 1178/481 concentrates 1 and 2. Tables 6 and 7
provide data on NBS materials following nitric/perchloric/sulfuric
acid digestions. The pH of the digest was adjusted with dilute
ammonium hydroxide before being subjected to resin treatment. Each
slide provides examples of at least one sample that required the use
of either high acid concentrations or resin treatment in order to
obtain acceptable results.
Leachates
Tables 8 and 9 provide examples of arsenic and selenium recovery data
in leachates from solid waste samples.
Sludge Samples
-------
Municipal sludge samples were obtained from various cities in
Pennsylvania. Following acid digestions, arsenic and selenium were
determined by hydride generation (Tables 10 & 11).
SUMMARY
The hydride generation MS methods found in SW 846 are applicable to
many environmental samples. However, these methods will not provide
acceptable results on samples that are highly contaminated with
interfering elements. Interference reduction techniques such as the
adjustment of acid strength and ion exchange clean—up are necessary
when dealing with difficult sample matrices.
-------
References
1. Code of Federal Regulations, Title 40, 141.11.
2. Federal Register, May 19, 1980, 33127.
3. Code of Federal Regulations, Title 40, 180, 192—196.
4. NIOSH Publication 81—123, Jan. 1981, U.S. Government Printing
Office, Washington, D.C., 20402.
5. A.E. Smith, Analyst, 100, 300 (1975).
-------
2 Channel
hriitajtic Pump
Schematic of 1.1. Plasma Hydride Generator
.11-
• 10-
• 09-
1 •OQ
.07-
.06-
• 05-
• 04-
• 03-
• 02-
.01-
w i r I.1t E DP ’l
• Flgu 2
J — I•—— I I I I
.2 .4 .6 .8 J.O 1.2 1.4
-
1.6 1.8
(I-1c L_. >
Figure I
.0
2
-------
• 09•
S
3
C-
‘ . 4
I-
• 1.
.01
• 0
• 0:
x
NI
-
.11-
.10-
09-
• 08-
• 07-
.06-
• 05-
.04-
• 03-
.02-
• 01-
DP J 5E ‘
c 1Jp.1c - EFFECr C I1 DII E )
_____ —-
u-SI • • .-i.iu_ i . v ‘ -
I__ L_ 1: U II PIT E E t 1
x
Figure 3
• 05
04
• 03
• 02
.01•
K
1.0
I i opic — EF-F-Ec-r
1. 2.0
( DI4C - C III L.)
cw-ar DP .l )
0
0
Figure 4
i-ici._ 1DL., r r -r 1 cr . :t cp.ic_
FIgure 5
‘ N ’ - d’) 1• r . -.
-------
TABLE I
MAXIMUM ALLOWABLE LEVEL
Agency
EPA
EPA
OSHA
World Health Organization
EPA
Matrix
Drinking Water
EP Toxicity Extract
Workplace Air (l)
Drinking Water
Food (2)
Arsenic Selenium
0.05 mg/I. 0.01 mg/I
5.0 mg/i 1.0 mg/i
10. ug/m 3 0.2 mg/rn 3
0.05 mg/i 0.01 mg/i
3.5 ppm
(1) Eight hour time weighted average
(2) As As 2 0 3
-------
TABLE II
NON-INTERFERING ELEMENTS
Aluminum Phosphorus
Barium Potassium
Beryl 1 iuin Scandium
Boron Selenium
Cadmium Silicon
Calcium Sodium
Cerium Strontium
Cesium Tantalum
Chromium Thallium
Gallium Thorium
Iridium Tin
Iron Titanium
Lanthanum Tungsten
Lithium Vanadium
Magnesium Yittrium
Manganese Zinc
Mercury Zirconium
Niobium
Elements not interfering with the I.L. Manual Hydride-AAS
determination of Arsenic, Antimony, Bismuth and Selenium.
Interferent concentrations were 100 ug/mi or greater while
analyte concentrations were 0.100 ug/mi.
-------
I ABLE III
INTEREL MENT 1 ERFERE CES BY AAS
Percent Recov ry
Conc.
Interferent ( ua/mfl Selenium Areenic
Antimony 100 89 159 (1)
Arsenic 100 34
Bismuth 100 75 92
Cobalt 100 28 33
Copper 100 2 101
Germanium 100 20 88
Gold 50 57 88
Lead 100 84 96
Molybdenum 100 74 97
Nickel 10.0 42 45
Palladium 0.5 104 47
Platinum 0.5 97 79
Rhodium 0.5 71 95
Silver 5.0 12 97
Tellurium 100 124 (1) 67
(1) Apparent contamination in standard.
Analyte concentration 0.100 ig/m1
-------
TABLE IV
% REMOVAL DURING
MINI-COLUMN PROCESS AT pH 2
Element AG 50W—X16 Che]ex 100
Antimony 6.2 34.2
Arsenic 0 5.8
Bismuth 12.2 >98
Cobalt >99.9 99.5
Copper >99.9 >99.9
Germanium 0 6.0
Gold 69.4 98.6
Lead >99.9 99.7
Mercury >98 93
Nickel >99.9 99.5
Palladium 42.2 >99.8
Platinum 0 12.2
Rhodium 11.8 18.8
Selenium 0 2.4
Silver >99.9 >99.9
Tellurium 78.4 72.0
12.5 ml of pH adjusted 100 mg/i standard were
subjected to the minicolumn process.
-------
TABLE V
EPA WP11781481 CONCENTRATES 1 AN!) 2
Al]. values are given in micrograms per liter.
X = Average
C.I. = 95% Confidence Interval
vy ’r4 m iiPa1
Sample No. Results X True X
EPA V 1u
C.’.
Mini—Column
Resin
Treatment Element
1
26.5.
26,
26.5,
26
26.3
27
26.3
19.8
—
34.2
No
Arsenic
2
257,
253,
261
247
255
235
234
182
—
286
No
Arsenic
1
10.5,
11.0,
11.5,
10.5
10.9
11
10.3
6.5
—
14.1
No
Selenium
2
27.0,
28.0,
28.0,
26.5
27.4
50
46.7
31.3
—
62.1
No
Selenium
2
44.0,
48.0,
48.0,
54.0
48.5
50
46.7
31.3
—
62.1
Yes
Selenium
-------
TABLE VI
ARSENIC RESULTS ON
NBS STANDARD REFERENCE MATERIALS
SRM ExDerimenta l Certified Value
1566 13.6 13.4 ± 1.9
1568 0.40 0.4]. ± 0.05
1572 3.3 3.1 ± 0.3
1575 0.23 0.21 ± 0.04
1633a 133* 145 ± 15
* In 6 N HC1
All values are given in micrograms per gram. No
resin treatment was necessary.
-------
TA LE VII
SELENIUM RESULTS ON NBS STANDARD REFERENCE MATERIALS
* In 6 H HC1
The AG 50W-X16 mini—column resin process was utilized. All
values are given in micrograms per gram.
Before Resin
Treatment
After Resin
Treatment
1566
1.5
2.1
2.1
±
0.5
1567
0.90
0.99
1.1
±
0.2
1568
0.3.9
0.49
0.4
±
0.1
1577
0.37
1.1
1.1
±
0.1
1577a
0.11
0.73*
0.71
±
0.07
1633a
7.7
10.0*
10.3
±
0.7
-------
TABLE VIII
ARSENIC RECOVERY FOLLOWING MINI-COLUMN RESIN PROCESS
% Spike Recovery
Sam 1e No. Initial H AG 50W-X16 Chelex 100
284286 1 104
284286 4 97 91
284286 7 10]. 88
278030 1 99
278030* 4 10]. 89
278030* 7 104 55
SELENIUM RECOVERY FOLLOWING MINI-COLUMN RESIN PROCESS
% Spike Recovery
Sam 1e No. Initial pH AG 50W-X16 Che].ex 100
284286 1 104
284286 4 102 84
284286 7 99 73
278030 1 105
278030* 4 48 28
278030* 7 44 23
285553 1 90
285553* 4 91 4
285553* 7 89 <4
*Significant precipitate present after pH adjustment to 4 and 7.
-------
TABLE IX
SELENITJ14 RECOVERY IN E.P. TOXICITY LEACHATE EXTRACTS
Sam 1e No ,
333342
333343
337567
337918
337919
339540
343009
344313
344880
344881
344882
344883
344884
344885
345348
345984
346269
Original Sample
Following Resin
Treatment
0.004
<0.004
<0.004
<0.004
<0.004
0.004
<0.004
0.005
<0.004
<0.004
<0.004
<0.004
<0.004
0.004
<0.004
<0.004
<0.004
Sample Spiked,
No Resin
Treatment
0.020
0.017
0.007
0.018
0.006
0.028
0.091
0.034
0.015
0.017
0.015
0.018
0.016
0.014
0.059
0.088
0.058
Sample Spiked,
Then Pasased
Through Resin
0.101
0.102
0.104
0.097
0.106
0.102
0.103
0.107
0.101
0.103
0.103
0.102
0.102
0.103
0.103
0.099
0.101
The resin used was
0.100 mg/i.
AG 50W-X16. The spike concentration was
All values in mg/i.
-------
TABLE X
ARSENIC RECOVERY IN MUNICIPAL SLUDGE DIGESTS
City Number Arsenic % Recovery
1 2.8 96
2 2.5 95
3 2.1 94
4 2.0 96
5 2.6 100
6 1.2 96
7 1.1 92
8 <1 92
9 <1 105
10 <1 99
11 <1 108
12 2.4 102
13 <1 104
14 <1 102
15 1.3 104
16 1.4 101
The arsenic values re given in micrograms
per gram on an as rQceived (wet) basis.
Digests were spiked at the 0.100 mg/i level.
-------
TABLE XI
PERCENT SELENIUM RECOVERY IN MUNICIPAL SLUDGE DIGESTS
No Resin Treatment Ninicolumn Resin Treatment
City 1.2 N HC1+ 6.6 N HC1+ 1.2 N HC1+ 6.6 N HC1+
Number 1.8 N H 2 S0 4 0.9 N H 2 S0 4 1.8 N H 2 S0 4 0.9 N H 2 S0 4
1 31 80 30 86
2 39 77 58 101
3 38 92 40 93
4 32 83 42 101
5 51 64 53 78
6 45 102 52 102
7 27 88 27 89
8 51 95 56 97
9 100 105 102 95
10 95 110 99 100
11 67 102 69 97
12 101 109 100 104
13 51 98 54 96
14 98 109 96 104
15 102 91 102 97
16 56 111 59 104
The values tabulated above are given in percents. The selenium
results in the sample, obtained after resin treatment and the higher
level acidification, ware all below 4 micrograms per gram on an as
received (wet) basis.
-------
EMPLOYMENT OF ALKALINE DIGESTION PROCEDURES FOR DETERMINATION OF
METALS IN INDUSTRIAL WASTES
JOSEPH H. L( ’JRY AND D. S. KENDALL, NATIONAL ENFORFICEMENT
INVESTIGATIONS CENTER, U. S. ENVIRONMENTAL PROTECTION AGENCY, DENVER,
COLORADO
ABSTRACT
For the past five years, NEIC has been involved in the analysis of
samples collected from hazardous waste sites. The waste samples have
included mostly drum samples, but samples of piles, logoons, tankers,
and spills also have been analyzed at NEIC. The diversity and
complexity of the waste compositions encountered demanded widely
applicable methods for the analysis of these samples for trace and
major elemental constituents. Often samples were mixed phases
containing any combination of an aqueous liquid, organic solvent,
solids, pastes, gels, greases and tars. A number of analytical
strategies have been studied and decomposition of the sample matrix
with a molten fusion followed by Inductively Coupled Argon Plasma
Optical Emission Spectroscopic analysis has been found satisfactory
for most sample types and study objectives. This approach has been
employed successfully in the analysis of the matrices found in drum
samples and has also worked well in the analysis of soils.
Data for the analysis of a wide variety of reference materials
demonstrates that highly precise and accurate analyses can be obtained
with this methodology. Comparison of this method’s results to that of
X—Ray Fluorescene Spectroscopy in the analysis of contaminated soils
further verifies the quality of data that can be achieved. Precision
and accuracy data gathered during the use of this method over the past
few years will be present.
The selection and applicability of a method of analysis is based
largely on what is a significant level for a particular study. One
important factor in defining a significant concentration is the
distribution of an element in the sample population. Abundant data
for most environmental media is available. Such data, however, is
scarce for drum samples. To aid in filling this data base void, the
frequency and level of occurence for thirty—two elements from 1,000
samples collected from forty—five sites in thirty—one states will be
presented. These elemental distributions will be used to evaluate the
applicability of a selected method to different study objectives.
-------
SECOND SESSION
METHODS FOR IDENTIFYING
HAZARDOUS WASTE CHARACTERISTICS
1:30 pm — 5:00 pin
Wednesday, July 24, 1985
Chairperson: Todd A. Kimmell
Methods Program
Office of Solid Waste
U. S. Environmental
Protection Agency
Washington, D. C.
-------
PERFORMANCE OF AN I( ITABLE SOLIDS METHOD IN CHARACTERIZING HAZARDOUS
WASTES: A COLLABORATIVE S IUDY
FLORENCE RICHARDSON, OFFICE OF SOLID WASTE, U. S. ENVIRONMENTAL
PROTECTION AGENCY, WASH ING’ION, D.C.
The United States Environmental Protection Agency (EPA) has specified
in the Federal Re9ister that a solid waste exhibits the characteristic
of ignitability if “it is capable of causing fire through friction,
absorption of moisture or spontaneous chemical changes and, when
ignited, burns so vigorously and persistently that it creates a
hazard.” At the present time, there are no suitable, validated
procedures for determining the ignitability characteristic of solid
(non liquid) wastes.
Since current ignitability wastes are designed for liquids, an
evaluation of potential testing methods for solids was performed at
the Research Triangle Institute (RTI). Techniques for ignitability
measurements were developed and tested on a wide variety of waste
materials. Results of this work were presented at the-ASTM D—34
Hazardous Waste Symposium in May, 1984.
The following test procedures were developed: Radiant Heat Test,
Flame Propagation Test, and Water Extinguishability Test. All tests
were performed in a chamber constructed for this study (Figure A). In
addition to reference materials, actual waste samples were utilized in
the evaluation of these test procedures.
The following wastes were tested:
• Wood wool excelsior
• Textile lint
Paint manufacturing waste (3) — putty, fibrous and sandy
texture
Paint wastes (4)
— amorphous, paste (2), and flake consistency
Coconut/tolunce
Styrene polymore/vermiculite
Still bottoms
• Petroleum waste/vermiculite
Coal/Xylene
• Red oak sawdust/kerosene
Paint filter
• Oil pad
• Fuel oil pad
• Sludge bargewaste
• Waste oils
Lighter fluid
Sterile cotton
• Pipe tobacco
• Polymethane foam
• Polystyrene
-------
An interlaboratory study was conducted to evaluate the three
ignitability test methods. The study consisted of two phases. The
first phase involved the replicate measurements of two
well—characterized test materials (coal/xylene and red oak
sawdust/kerosene) and a reference material by nine laboratories.
Interlaboratory test materials included the folloiwing:
Sterile cotton
• Polyurethane foam
Paint waste
Used motor oil
60% used oil/blend 40% sand
• Waste solvent
80% sawdust/20% kerosene
Wood wool excelsior was chosen as a reference material in this study
because of its use by Underwriter’s Laboratories as a standard for
class A fire extinguisher and its low measurement variability noted
during the single laboratory study. The second phase included similar
replicate measurements of seven varied test materials by a final
roster of only five laboratories.
The study results showed that not all sample types are amenable to
measurement by each test. Although all of the procedures showed some
deficiency, the Badiant Heat and Flame Propagation test results
indicated that, with some minor changes in chamber design, they would
be potentially useful routine methods. The Water Extinguishability
test is in need of more major improvements.
RTI has been charged with responding to these needs and is proceeding
to incorporate the recommendations into the existing protocols. We
anticipate having a draft final report ready in early 1986 for review
by the various EPA program offices and the Regions.
Another collaborative study will be conducted; consequently,
participants will be needed. We hope most of the collaborators will
come from this group.
-------
view)
Figure A
lgnitability lest chamber and Controller/Sensor.
-------
REACTIVE SULFIDES AND CYANIDES: TEST METHODS AND REGULATORY THRESHOLD
SETI’ING MODELS
Dr. PAUL H. FRIEDMAN, STUDIES AND METHODS BARNCH, OFFICE OF SOLID
WASTE, U. S. ENVIRONMENTAL PROTECTION AGENCY, WASHINGTON, D. C.
ABSTRACT
Reactivity, with respect to sulfide— and cyanide—bearing wastes refers
to the volatility and the rate of evolution of hydrogen sulfide and
hydrocyanic acid from wastes when these wastes are exposed to an
acidic environment. Determining a hazardous waste as reactive
requires an empirical test and a mathematical model. The test
determines the rate of release of the reactive component from the
waste. The model relates the results of the test to short—term and
short—range exposure of the gas to a receptor.
Both the test and the model are approximations of reality. Both are
described and discussed in terms of their strengths and weaknesses in
pre—dicting evolution rates and down wind concentrations of the toxic
gas. Alternative procedures for determining evolution rates are
discussed and contrasted. Alternative models also are surveyed.
Work in progress is described.
-------
MOBILITY OF TOXIC COMPOUNDS FROM HAZARDOUS WASTES: COMPARISON OF THREE
TEST METHODS TO A LYSIM a MODEL
MICHAEL MASKARINEC AND CHESTER W. FRANCIS, CHEMISTRY DIVISION, OAK
RIDGE NAITIONAL LABORATORY, OAK RIDGE, TENNESSEE
ABSTRACT
In order to assess the environmental impact of the disposal of solid
wastes, it is necessary to develop test methods capable of predicting
this impact. This work has the objective of providing a data base
from which assessments of the mobility of toxic compounds from solid
wastes can be made. The scenario was codisposal of industrial wastes
in a municipal waste landfill at a level of 5%. The experimental
design included the use of large lysimeters to generate municipal
waste leachate, which was then used to leach contaminants from
industrial wastes.
Four lysimeters (lined concrete cylinders 1.8 meters in diameter and
3.6 meters high) were filled with 1.5 Mg of municipal refuse
collected from the Oak Ridge, Tennessee area. Two were used for the
initial phase of this work (Phase I) and two were used for validation
studies (Phase II). Distilled de—ionized water was applied to the
tops of the lysimeters and the leachate generated was used to leach
industrial wastes. The leachates were collected and analyzed over
time in order to construct leaching curves for each contaminant. From
these curves, a target concentration was developed for laboratory
ex—tractions to simulate.
A total of 34 different laboratory extractions were tested for their
ability to match the target concentrations. Factors including medium,
liquid to solid ratio, and extraction techniques were varied. Based
on the Phase I data, two procedures were chosen for further
evaluation, acetate buffer (0.1N pH 5), and distilled de—ionized water
saturated with carbon dioxide, both in a rotary extraction device at a
liquid to solid ratio of 20. The EPA Extraction Procedure was
also tested. A number of statistical procedures were used to rank the
various extraction procedures. Where statistically significant
differences were found, the acetate system was found to best simulate
the leaching data obtained from the field lysimeters.
-------
APPLICATION OF THE TOXICITY CHARACTERISTIC LEACHING PROCEDURE (TCLP)
TO INDUSTRIAL WASTES: A SINGLE LABORATORY EVALUATION
Dr. L. R. WILLLAI’IS, U. S. ENVIRONMENTAL PROTECTION AGENCY—EMSL, LAS
VEGAS, NEVADA
ABSTRACT
A vital step in the process of validating, testing, and measuring
methods for Agency use is the “single—laboratory study.” Such studies
are intended to identify and test critical variables in method
protocols and provide bases for revising the written protocols for
emphasis, clarity, and ease of use. In addition, these studies
develop data on the accuracy, precision, and limits of reliable
measurement for the test method with representative sample materials
in one or more capable laboratories. These studies represent the
vital, and frequently missing, link between the written method and
multi—laboratory collaborative study. The role of the
single—laboratory study in overall strategy of method validatiion was
highlighted using the current evaluation of the Toxicity
Characteristic Leaching Procedure (TCLP) as the primary example.
Study design and preliminary results of the single—laboratory testing
of the TCLP were presented and discussed. Aspects of this method
evaluation were conducted at two different laboratories in a parallel
effort.
-------
THIPD SESSION
ANALYZING FOR ORG NICS
8:00 am — Noon
Thursday, July 25, 1985
Chairperson: Dr. Paul H. Friedman
Methods Program
Of fice of Solid Waste
U. S. Environmental
Protection Agency
Washington, D. C.
-------
APPLICATION OF SW-846 METHODS TO GROUNI ATER MONITORING PROGRAMS:
EXPERIENCES OF ‘IWO CONTRACT LABORATORIES
Dr. DENIS C. K. LIN, DAVID SPEIS, KAREN KOTZ-GEBEL, AND DIANE FOSTER,
ENVIRONMENTAL TESTING AND CERTIFICTION CORPORATION, EDISON, NEW JERSEY
INTRODUCTION
On July 26, 1982, the U.S. Environmental Protection Agency (EPA)
published interim final regulations under the Resource Conservation
and Recovery Act (RCRA), which set permit procedures and operating
standards for hazardous waste land disposal facilities. These
regulations also established groundwater monitoring requirements for
such facilities. Under certain circumstances the regulations required
facility owners to conduct chemical analyses for the compounds listed
in Appendix VIII, Part 261. On September 21, 1982, a notice published
in the Federal Register (Vol. 47, No. 183, at pages 41562 through
41563) amended two sections of the regulations (specifically, in title
40, Code of Federal Regulations, Section 122.20 of the consolidated
permit regulations and Section 260.11 of the hazardous waste
regulations) to incorporate by reference the Second Edition of the EPA
Manual “Test Methods for Evaluating Solid Waste, Physical/Chemical
Methods,” EPA Publication SW—846. More recently EPA proposed on
October 1, 1984 to make SW—846 methods mandatory for all testing and
monitoring activities defined under Subtitle C, as specified in 40 CFR
parts 260—271. Thus, SW—846 has become a key document in the analysis
of groundwater for the 375 Appendix VIII parameters (Table 1).
The Environmental Testing and Certification Corporation (ETC) has been
involved in the analysis of Appendix VIII parameters in groundwater
since the fall of 1982. In May, 1983, ETC reported to Chemical
Manufacturers Association (CMA) the results of a project whose
objective was to evaluate the effectiveness of SW—846 to serve as a
methods manual for the sampling and analysis of groundwater on
leachate for the Appendix VIII parameters. In the summer of 1984, ETC
participated in an “Inter— and Iritra—Laboratory Assessment of SW—846
Methods Manual for Analysis of Appendix VIII Compounds in Groundwater”
sponsored by CMA. (1) In the fall of 1984, ETC performed a study for
EPA to serve as a preliminary investigation in using presently
available methodology to analyze as many Appendix VIII parameters as
possible. Furthermore, over the past two and one—half years ETC has
been developing an analytical scheme to meet the demand for Appendix
VIII analysis. In this article we will report the present status of
Appendix VIII analytical procedures based on our experience.
THE APPENDIX VIII PARAMETERS
To design an analytical scheme for the 375 Appendix VIII parameters in
the Grand Table (Table 1), it is essential to examine the physical and
chemical characteristics of the individual compounds. In the October
1, 1984 Federal Register (Vol. 49, Page 38786) EPA identified several
-------
categories of parameters. It proposed to categorize 13 parameters as
unstable in water; and 10 parameters as “exotic” in nature and for
which no satisfactory analytical methods are available. Furthermore,
for those parameters that have metals in their compositions it is
sufficient to analyze for the metal components only. Fifty—one
parameters were assigned to this category. Using the same rationale
EPA proposed 14 parameters can be testd as cyanides. Since one
exotic, two unstable, and seven cyanide compounds are also classified
as metals, EPA categorization effectively addressed 78 parameters
leaving 297 that need to be addressed in an analytical scheme.
The analytical scheme designed at ETC for the Appendix VIII parameters
follows the rationale suggested by SW—846. Various parameters are
categorized according to their chemical and physical characteristics
and they are grouped together under analytical methods by which they
can be quantified at optimal method detection limits. We agree with
EPA’S approach to the unstable, exotic and metallic compounds. We
focus our attention on the remaining 297 parameters.
There are 11 additional parameters among the 297 that ETC believes
should be classified as “unstable in water” (Table 2). EPA, while
acknowledging that there are other unstable compounds, suggests that
there is some likelihood they can be found in water. Since EPA has
not yet reclassified any of these compounds as unstable, we have
decided to incorporate them into our analytical scheme. However, we
do not set data acceptance criteria for them.
There are another 33 parameters (Table 3) that ETC believes should be
classified as exotic compounds because they require special methods
for their analysis. Ten of these exotic compounds are antineoplastic
agents or other drugs. Five are uncommon compounds where pertinent
information is unavailable. Three are alkaloid poisons or mycotoxins.
No are water soluble dyes. One is a very volatile compound that can
explode easily. These compounds and the remaining 12 compounds
possess unusual chemical and physical characteristics, which in turn
dictate that their analysis at meaningful detection limits would
require individual tailor—made analytical methodology. In most of
these cases, EPA recommends including them in analytical methods in
SW—846. Little, if any, data is available to support such
recommendations. Although ETC includes these parameters in its
current Appendix VIII analytical scheme, we believe the recommended
methodologies are unproven and most likely ineffective. We strongly
urge EPA to re—examine its position. Perhaps only under certain
circumstances should analysis be required for the selected exotic
parameters in Table 3.
Among the 297 parameters there are 28 that are listed as “N.O.S.”
(not otherwise specified). These 28 parameters, along with
aflatoxins, coal tars, cresote, cresols, phthalic acid esters,
tetrachloroethane, and toluenediamine, constitute 35 complex mixtures
or classes of compounds (Table 4). Since some of mixtures and classes
include thousands of individual compounds it is necessary to select
representatives for analysis. As much as possible, we have chosen
“priority pollutants” and commonly available reference standard
-------
compounds to represent the 35 parameters (see Table 4).
In the next section we will detail the present status of the Appendix
VIII analytical scheme in our laboratory.
THE ETC ANALYTICAL SCHEME FOR APPENDIX VIII PARAMETERS
The ETC Analytical Scheme consists of analyses based on the
methodology described in Sw—846 and the parameters are divided into
the following categories:
1. Water Soluble Compounds by Direct Aqueous Injection
GC ?1S. (DAI)
2. volatile Compounds by Purge and Trap GCflIS. (P & T)
3. Extractable Acid/Base/Neutral/Pesticides Compounds
by GC/MS. (A/B/N)
4. Pesticides by GC/EC. (PEST / BERB)
5. Pesticides by GC/FPD. (PEST,4IERB)
6. Herbicides by GC/EC. (PEST/NERB)
7. Polar and Thermally Unstable Compounds by HPLC/UV).
(HPLC)
8. Metallic and Organontetallic Compounds by ICAP, M,
and Cold Vapor PA. (METALS)
9. Conventionals. (CONy)
Within the nine categories of compounds that can be specifically
analyzed, not all the reference stardards are readily available. In
Table 5 we list the number of such parameters in each appropriate
category and the number for which reference standards are missing. In
the GC/14S analyses we are able to search for the specific compounds
that do not have corresponding reference standards by using standard
spectra from the literature. In other cases, such as GC and HPLC,
where retention time characteristics are the only means to identify
the compounds, we are unable to draw any conclusions about those
compounds for which corresponding reference standards are not
available.
In each category we achieve rigorous compliance with the instrumental
requirements and performance criteria of established EPA methods, such
as those in methods 624 and 625, before any sample analysis is
initiated. For GC/MS analysis where standards are available,
identification is performed using relative retention times, the
relative abundance of three characteristic ions and the abundance
ratios. The entire mass spectrum is reviewed to confirm each
-------
identification. Quantitative analysis is performed using an internal
standard with a single characteristic ion. When compounds without
corresponding reference standards are identified, they are quantified
assuming the same response factor as the internal standard.
VOLATILE COMPOuNDS BY PURGE AND ThAP GC/NS
For the analysis of volatile compounds by Purge and Trap, Methods 8240
and 5030 are used. The analysis can be summarized as follows: Helium
is bubbled through a 5 ml water sample contained in a specially
designed purging chamber at ambient temperature. The purgeable
volatile organic compounds are transferred from the aqueous phase to
the vapor phase. The vapor is swept through a sorbent column where
the organic compounds are trapped. After the purge cycle is complete,
the sorbent column is heated and back flushed with helium to desorb
the organic purgeables onto a gas chromatographic column. The gas
chromatograph is temperature programmed to separate the purgeable
mixture. The separated purgeable components are then identified and
quantitated using a computerized mass spectrometer.
WATER SOLUBLE COMPOUNDS BY DIRECT AQUEOUS INJECTION GC/MS
For the analysis of water soluble compounds, 5 ul of aqueous sample is
injected directly into the CC/MS system. The chromatographic column
employed in the procedure is the same column used for purge and trap
analysis. MS scanning was begun prior to sample injection to capture
mass intensity data for early eluting compounds. The GC oven
temperature program used is that specified in procedure 8240.
EXTRACTABLE ACID/BASE,4JEUTRAL AND PESTICIDE COMPOUNDS BY CC/MS
For the analysis of the Acid, Base/Meutral and Pesticide compounds in
water, Methods 3510 and 8270 are used. The analysis can be summarized
as follows: a measured volume of sample, approximately 1 liter, is
extracted with an aliquot of methylene chloride without pH adjustment
and then the sample is adjusted to a pH greater than 11 and extracted
with another aliquot of xnethylene chloride. These two aliquots were
combined. The pH of the sample is then adjusted to a value less than
2 and extracted with another aliquot of methylene chloride. A
separatory funnel or continuous extractor is used to perform the
extractions. The two extracts are dried and concentrated to a 1 ml
final volume. The extracts are then combined just prior to injection
into a GC/MS instrument.
HERBICIDES AND PESTICIDES BY CC
The methods employed in the analysis for herbicides and pesticides are
established EPA methods taken from the “Manual of Analytical Methods
for the Analysis of Pesticides in Humans and Environmental Samples,”
June, 1980 and methods 8080, 8140 and 8150.
-------
The herbicide method can be summarized as follows: A measured volume
of sample, approximately 500—1000 ml, to which sodium sulfate has been
added is acidified and extracted with methylene chloride. The
methylene chloride extract is evaporated to dryness, and the residue
is derivatized with diazomethane and injected into a gas chromatograph
equipped with a 63Ni electron capture detector.
The pesticide method can be summarized as follows: A measured volume
of sample, approximately 500m1, is extracted with methylene chloride.
The extract is dried and concentrated to a final volume of imi and
injected into a gas chroniatograph equipped with a 63Ni electron
capture detector and Flame Photometric detector in phosphorus mode.
METALS
The determination of metals in aqueous samples is performed according
to the methods published by EPA in “Methods for Chemical Analysis of
Water and Wastes.” EPA—600/4—79—020, March, 1979, and the inorganic
methods in SW 846. Arsenic, selenium and thallium are determined by
furnace AA; silver, aluminum, barium, beryllium, boron, cadmium,
calcium, chromium, copper, cobalt, iron, magnesium, manganese,
molybdenum, nickel, lead, sodium, antimony, tin, titanium, vanadium,
and zinc are determined by ICP emission. The determination of mercury
is performed by cold vapor AA.
THERMALLY UNSTABLE AND POLAR COMPOUNDS BY HPLC
The analysis of thermally unstable and polar compounds are based on
the HPLC methods 8320 and 8330. The compounds analyzed fall into two
categories: direct aqueous injection and Base/Neutral extractables.
Twenty ul of the sample or extract is injected into an HPLC equipped
with a reverse phase column. Gradient elution and UV detector at 210
and 250 nm are used.
CONVENTIONAL PARAMETERS
Total cyanide analysis is performed using Method 8010. Sulfide
analysis is performed using Method 9030.
DISCUSSION
Our initial study on Sw 846 sponsored by CMA concluded that SW 846 was
not an acceptable manual but could serve as a base upon which to build
an effective methods manual. The CMA inter— and intra—laboratory
assessment of SW—846 methods manual for analysis of Appendix VIII
compounds in groundwater studied specifically Methods 8240, (P & T,
GC /T4S), 8270 (A/ /N, GC/NS) and 8330 (HPLC). Three prominent
laboratories were asked to analyze a series of groundwater samples for
36 compounds. The study concluded that: “the list of Appendix VIII
-------
compounds reportedly amenable to Methods 8240 and 8270 was found to be
somewhat less than reported by EPA. Method 8330 was found to be
completely inadequate for the detection and quantification of the
three compounds included in the study.” From our experience such
conclusions are not surprising since SW—846 does not provide
sufficient procedural details for all appropriate Appendix VIII
compounds.
Based on our EPA study and our extensive experience using the ETC
analytical scheme, SW—846 can be applied diligently with success for a
substantial number of Appendix VIII compounds. In the first phase of
our EPA study, attempts were made to define which Appendix VIII
compounds could be analyzed under the analytical scheme. To achieve
this, the reagent water was spiked with compounds of interest at
reasonably high levels so that all indications of non—performance for
a parameter would be definitive. The first phase of the study was
suggested adjustments that should be made in the analytical scheme.
With the revised scheme, groundwater sample was spiked at high, medium
and low levels. The high and low levels were analyzed in seven
replicates and the medium level was analyzed in triplicate. Based on
the results, method detection limits were proposed for individual
compounds in the Appendix VIII parameters. However, the study was
limited by the availability of Appendix VIII standard reference
compounds. The data obtained in the spiking experiments were
submitted to EPA. Spike recovery results of selected nonpriority
pollutant Appendix VIII compounds are shown in Table 6.
For the 55 parameters (Table 5) we have included in the purgeable
volatile GC44S (P & T) fraction, six are listed in Table 1 and 2 as
potential candidates as unstable or exotic. We do not expect the
analysis for these six compounds using Method 8240 methodology can be
performed successfully. For the other P & T compounds in this
fraction our investigation showed that the methodology performed
satisfactorily. The ions we select to identify and quantify the
compounds, as well as their relative retention times, have been
submitted to EPA for review. Our investigation also showed
crotonaldehyde (Table 7) could not be analyzed satisfactorily. At 50
ug/l crotonaldehyde could not be recovered using Method 8240
methodology.
enty—two parameters were selected to be analyzed by direct aqueous
injection GC/MS methodology (DAl). All of these 22 compounds are
highly soluble in water. We suspected Method 8240 would not work for
these compounds. The CMA inter— and intralaboratory study indeed
showed unsatisfactory results for isobutanol. (1) In our
investigations using DAI/INJ, eight parameters still yield
unsatisfactory results (Table 7). We intend to try other methods on
these water soluble compounds.
Of the 18 compounds we listed in Table 7 as compounds that require
further investigation, nine originated from the extractable
acid/base/neutral/pesticide compounds by CC/MS fraction (A/B/N). We
suspect these nine compounds can be analyzed at higher method
detection limits. While the Method 8270 methodology works well for
-------
the compounds investigated, the success rate may not be the same for
the 30 parameters for which reference standards were unavailable.
we did not list any parameters in Table 7 for pesticide and herbicide
fractions. The SW—846 GC methods 8080, 8140 and 8150 for pesticides
and herbicides are sound. We investigated all but three of the 28
parameters in these categories and did not discover any problems. We
have just acquired the three compounds as reference standards and we
do not expect any major difficulties in their analyses.
Contrary to the GC methods in SW-846, the HPLC methods appear to be
unsound. In the inter— and intra-laboratory study none of the Method
8330 analytes were successfully analyzed. (1) We have substantially
modified the HPLC methods in order to accommodate the 37 parameters
included in this fraction and thus far have successfully analyzed for
19 parameters. We do not have reference standards for seven
parameters. We are investigating the remaining 11 parameters in the
HPLC analyses.
To summarize, among the 297 organic Appendix VIII parameters that need
to be included in the ETC analytical scheme, 43 reference standards
are unavailable at present. We have 40 newly acquired reference
compounds on hand. Naturally, we intend to add these to our on—going
investigation.
CONCLUSION
Of the total 375 Appendix VIII parameters, we presently have no
spiking experimental data on 83 parameters. Some of these compounds
(14) can be qualitatively searched for since their mass spectra are
documented. Of the remaining 292 parameters, we are not satisfied
with the experimental results of the 198 parameters as shown in Table
7.
Of the SW—846 methods employed for the organic parameters in Appendix
VIII, the GC methods 8080, 8140 and 8150 performed satisfactorily.
The HPLC methods 8320 and 8330 require extensive modifications. The
Methods 8240 and 8270 GC/MS methodologies were found to be suitable
for most of the compounds investigated in those appropriate fractions.
-------
REFERENCE
1. George H. Stanko and Peter E. Fortini, Hazardous Waste and
Hazardous Materials, Vol. 2(1), pp. 67—97, 1985.
-------
E TC ENVIRONMENTAL
TESTING and CERTIFICATION
TABLE 1
GRAND TABLE
ETC
REF NO COMPOUND NAME
Acetonit rile
2 Acetophenone
3 Warfarin
4 2-Acetylaminofluorene
5 Acetyl chloride
6 l-Acetyl-2-thiourea
7 Acrolein
8 Acrylaniide
9 Acrylonitrile
IC Aflatoxins
1OA Aflatoxins, Total
ii Aidrin
12 Allyl alcohol
13 Aluminum phosphide
13A Aluminum
14 4-Aminobiphenyl
15 Mitomycin C
16 5-(Azn inomethyl)-3-isoxazolol
17 Amitrole
18 Aniline
19 Antimony and Compounds, N O.S.
19A Antimony
20 Aramite
21 Arsenic and Compounds, N OS.
21A Arsenic
22 Arsenic acid (Orthoarsenic acid)
23 Arsenic pentoxide (Arsenic (V) oxide)
24 Arsenic trioxide (Arsenic (III) oxide)
25 Auramine
26 Azaserine
27 Barium and Compounds, N O.S
27A Barium
28 Barium cyanide
29 Benz(c]acridine
30 Benz [ a]anthracene
31 Benzene
32 Benzenearsonic acid
33 Dichloromethylbenzene
34 Benzenethiol
35 Benzidine
36 Benzo [ b]fluoranthene
37 Benzo(j]fluoranthene
38 Benzo [ a]pyrene
39 p-Benzoqu inone
40 Benzotrichloride
41 Benzyl chloride
42 Beryllium and Compounds, N.O S.
42A Beryllium
43 bis(2-Chloroethoxy)methane
44 bis(2-Chloroethyl) ether
45 Chiornaphazine
46 bis(2-Chloroisopropyl )ether
47 bis(Chloromethyl)ether
48 bis(2-ethylhexyl)phthalate
49 Bromoacetone
50 Methyl bromide
51 4-Bromophenyl ohenyl ether
52 Brucine
53 2-Butanone peroxide
54 Butyl benzyl Dhthalate
55 2-sec-Butyl-4,6-dinit rophenol
-------
E TC ENVIRONMENTAL
TESTING and CERTIFICATION
TABLE 1
GRAND TABLE
ETC
REF NO. COMPOUND NAME
56 Cadmium and Compounds, N 0 S.
56A Cadmium
51 Calcium chromate(Chromic acid,calcium salt)
57A Calcium
58 Calcium cyanide
59 Carbon disulfide
60 Carbon oxyfluoride
61 Chloral
62 Chlorambucil
63 Chlordane
64 Chlorinated Berizenes, N C S
64A 1 ,2,3-trichlorobenzene
64B 2,4 ,6-trichlorobenzene
64C 1 ,2 ,3,4-tetrachlorobenzene
64D 1 ,2,3,5-tetrachlorobenzene
65 Chlorinated Ethane, N 0 S.
65A Chloroethane
66 Chlorinated Fluorocarbons, N O.S
66A Freon TF
67 Chlorinated Naphthalene, N OS.
67A 1 -ch1orona hthalene
68 Chlorinated Phenol, N 0 S
68A 2 ,3,5,6-tetrachloro henol
68B 2,3,4,5-tetrachiorophenol
69 Chioroacetaldehyde
70 Chloroalkyl Ethers, N O.S.
ii p-Chloroaniline
72 Chlorobenzene
73 Ch lorobenzilate
74 p-Chloro-m-cresol
75 1 -Chloro-2,3-epoxypropane
76 2-Chioroethyl vinyl ether
77 Chloroform
78 Methyl chloride
79 Chloromethyl methyl ether
80 2-Chloronaphthalene
81 2-Chiorophenol
82 1 -(o-Chlorophenyl )thiourea
83 3-Chioropropionit rile
84 Chromium and Compounds, N 0.S
84A Chromium
85 Chrysene
86 Citrus red No 2
87 Coal Tars
87A Acenapthene
87B Acenaphthalene
87C Anthracene
87D Benzo(ghi)perylene
87E Benzo(k)fluoranthene
87F Fluorene
8713 Phenathrene
87H Pyrene
88 Copper cyanide
88A Copper
89 Creosote
89A 2,Nttrophenol
90 Cresols
90A o-Cresol
90B m+p-Cresol
91 Crotonaldehyde
92 Cyanides (soluble salts and cornplexes)N 0 S
92A CyanLde, Total
-------
E TC ENVIRONMENTAL
TESTING an CERTIFICATION
TABLE 1
GRAND TABLE
ETC
REF NO COMPOUND NAME
93 Cyanogen
94 Cyanogen bromide
95 Cyanogen chloride
96 Cycasin
97 2-Cyclohexyl-4 .6-dini trophenol
98 Cyclophospharnide
99 Daunomycin
IOU 4,4-DDO
101 4 .4’-DOE
102 4 .4-DDT
103 Diallate
104 Dibenz [ a,h]acrtdine
105 Dibenz [ a .j]acridine
106 Dibenz(a,h)anthracene
107 7H-Dibenzo [ c ,g)carbazole
108 Dibenzo [ a,e)pyrene
109 Dibenzo [ a,h]pyrene
110 Dibenzo [ a,i]pyrene
Ill I ,2-dibromo-3-chloropropane
112 I .2-Dibromoethane
113 Dibromomethane
114 Di-n-butyl phthalate
115 I .2-Dichlorobenzene
116 I 3-Dichlorobenzene
Ill I ,4-Dichlorobenzene
118 Dichlorobenzene, N 0.S.
119 3.3-Dichlorobenzid ine
120 I .4-Dichloro-2-butene
121 Dichlorodifj.uorometharie
122 1 1 -Dichloroethane
123 1 .2-Dichioroethane
124 I .2-Trans-d ichioroethylene
125 Dichioroethylene, N 0.S.
126 1 , 1 -Dichioroethylene
127 Methylene chloride
128 2,4-Dichiorophenol
129 2,6-Dichloropheno] .
130 2,4-D
131 Dichiorophenylarsine
132 Dichloropropane, N.0 S.
I32A I .3-Dichioropropane
133 1 ,2-Dichloropropane
134 Dichioropropanol, N 0 S.
134A Dichioropropanol
135 2,3-Dichloropropene
136A cis-1 .3-Dichioropropene
136B trans-I ,3-Dichloropropene
137 Dieldrin
138 I .2.34-Diepoxybutane
139 Diethylarsine
140 N ,N ,-Diethylhydrazjne
141 Carbophenothion
142 0,0-Diethylphosphoric acid. O-p-nitrophenyl ester
143 Diethyl phthalate
144 Thionazin
145 Diethyistilbesterol
146 Dihydrosafrole
147 3,4-Dihydroxy-alpha-(methylamino)methyl benzyl alcohol
148 Ditsopropyifluorophosphate
149 Dimethoate
150 3 ,3-Dimethoxybenzidine
151 p-Dimethylaminoa2obenzene
-------
E iC ENVIRONMENTAL ______________
TESTING and CERTIFICATION
TABLE 1
GRA TABLE
ETC
REF NO. COMPOUND NAME
152 7, 12-D irnethylbenz [ a]anthracene
153 3,3’ -Dimethylbenzidine
154 Dimethyl carbamoylchloride
155 1 , I -Dimethyihydrazine
156 1 ,2-Dimethylhydrazine
157 Thiofanox
158 alpha-alpha-Dimethylphenethylamine
159 2,4-Dimethylphenol
160 Dimethyl phthalate
161 Dirnethyl sulfaLe
162 Dinitrobenzene , N O.S
162A m-Dirutrobenzene
163 4 ,6-Din itro-o-cresol
164 2 ,4-DinitroDhenol
165 2 ,4-Dinitrotoluene
166 2 .6-Dinitrotoluene
161 Di-n-octyl phthalate
168 1 ,4-Dioxane
169 DiDhenylamine
170 1 ,2-Diphenylhydrazine
171 N-Nit rosodi-n-propylam ine
172 Disulfoton
173 2,4-Dithiobiuret
174A Endosulfan I
174B Endosulfan II
1 75 Endrin
176 Ethyl carbaniate
171 Ethyl cyanide
178 Ethylenebisdithiocarbamic acid
179 Ethyleneimine
180 Ethylene oxide
181 Ethylenethiourea
182 Ethyl methacrylate
183 Ethyl methanesulfonate
184 Fluoranthene
185 Fluorine
186 2-Fluoroacetamide
187 Fluoroacetic acid
188 Formaldehyde
189 Formic acid
190 Glycidylaldehyde
191 Halomethane, N O.S.
191A Chiorodibrornomethane
191B Dichiorobrornomethane
192 Heptachior
193 Heptachior epoxide
194 Hexachlorobenzene
195 Hexachiorobutadiene
196A Alpha-BHC
196B Beta- BHC
196C Gamma -BHC
196D Delta-BHC
197 Hexachiorocyclopentadiene
198 Hexachioroethane
199 Hexachlorohexahydro-endo. endo-dimethanonaphthalene
200 Hexachlorophene
201 Hexachloropropene
202 Hexaethyltetraphosphate
203 Hydra zine
204 Hydrogen cyanide
205 Hydrofluoric acid
206 Hydrogen sulfide
-------
E TC ENVIRONMENTAL
TESTING and CERTIFICATION
TABLE 1
GRAND TABLE
ETC
REF NO COMPOUND NAME
207 Hydroxydimethylarsine oxide
208 Indeno(l ,2 ,3-cd)pyrene
209 lodornethane
210 Iron dextran
210A Iron
211 Methyl isocyanate
212 Isobutyl alcohol
213 Isosafrole
214 Kepone
215 Lasiocarpine
216 Lead and Compounds, N 0 S
216A Lead
217 Lead acetate (Acetic acid, lead salt)
218 Lead phosphate (Phosphoric acid, lead salt)
219 Lead subacetate
220 Maleic anhydride
221 Maleic hydrazide
222 Malononitrile
223 Melphalan
224 Mercury fulminate
225 Mercury and Compounds, N.0 S.
225A Mercury
226 Methacrylonitrile
227 Methanethiol
228 Methapyrilene
229 Methomyl
230 Methoxychior
231 2-Methylaziridine
232 3-Methyicholanthrene
233 Methyl Chiorocarbonate
234 4,4’ -Methylenebis(2-chloroaniline)
235 Methyl ethyl ketone
236 Methyl hydrazine
237 2-Methyllactonitrile
238 Methyl methacrylate
239 Methyl methanesulfonate
240 Aldicarb
241 N-Methyl-N’ -nit rosoguanidine
242 Methyl parathion
243 Methylthiouracil
244 Mustard gas
245 Naphthalene
246 1 ,4-Naphthoquinone
247 l-Naphthylamine
248 2-Naphthylamine
249 I -Naphthyl-2-thiourea
250 Nickel and Compounds, N.0 S.
250A Nickel
251 Nickel carbonyl (Nickel tetracarbonyl)
252 Nickel cyanide (Nickel (II)cyanide)
253 Nicotinic acid
254 Nitric oxide
255 o-Nitroaniline
256 Nitrobenzene
257 Nitrogen dioxide
258 Nitrogen mustard and hydrochloride salt
259 Nitrogen mustard N-Oxide and hydrochloride salt
260 Nitroglycerin
261 4-Nitrophenol
262 4-Nit ro uinoline-1 -oxide
263 Nitrosamines, N 0 S
263A N-N itrosodiphenylamine
-------
E TC ENVIRONMENTAL ______________
TESTING and CERTIFICATION
TABLE 1
GRAIL TABLE
ETC
REF. NO. COMPOUND NAME
264 N-Nitrosodi-n-butylam ine
265 N-Nitrosodiethanolamine
266 N-Nitrosodiethylamine
261 N-Nitrosodirnethylamine
268 N-Nitroso-N-ethylurea
269 N-Nitrosomethylethylamine
270 N-Nitroso-N-methylurea
271 N-Nitroso-N-methylurethane
272 N-Nitrosoinethylvinylamine
273 N-Nitrosomorpholine
274 N-Nitrosonornicotine
275 N-Nitrosopiperidine
276 N-Nitrosopyrrolidine
277 N-Nitrososarcosine
278 5-Nitro-o-toluidine
279 Octamethy1pyro hosphoramide
280 Osmium tetroxide (Osmium (VIII) oxide)
280A Osmium
281 Endothal
282 Paraldehyde
283 Parathion
284 Pentachlorobenzene
285 Pentachioroethane
286 Pent achioronit robenzene
287 Pentachlorophenol
288 Phenacetin
289 Phenol
290A m-phenylenediamine
290B o-phenylenediamine
290C p-phenylenethamine
291 Phenylmercury acetate
292 N-Phenylthiourea
293 Phosgene
294 Phosphine
295 Phorate
296 Famphur
297 Phthalic acid esters
298 Phthalic anhydride
299 2-Picoline
300 Polychlorinated Biphenyl. N.O.S.
300A Aroclor 1242
3006 Aroclor 1254
300C Aroclor 1260
300D Aroclor 1248
300E Aroclor 1232
300F Aroclor 1221
300G Aroclor 1016
301 Potassium cyanide
301A Potassium
302 Potassium silver cyanide
303 Pronamide
304 1 ,3-Propane sultone
305 n-Propylamine
306 Propylthiouracil
307 2-Propyn-1-ol
308 Pyridirie
309 Reserpine
310 Resorcinol
311 Saccharin and salts
312 Safrole
313 Selentous acid (Selenium dioxide)
314 Selenium and Compounds. N.O.S.
-------
E TC ENVIRONMENTAL
TESTING and CERTIFICATION
TABLE 1
GRAND TABLE
ETC
REF NO COMPOUND NAME
314A Selenium
315 Selenium sulfide (Sulfur selenide)
316 Selenourea
317 Silver and Compounds, N OS.
317A Silver
318 Silver cyanide
319 Sodium cyanide
319A Sodium
320 Streptozotocin
321 Strontium sulfide
321A Strontium
322 Strychnine and salts
323 1 ,2 ,4,5-Tetrachlorobenzene
324 2 ,3 ,7 ,8-TcDO
325 Tetrachloroethane
326 1,1,1 ,2-Tetrachioroethane
327 1,1 ,2,2-Tetrachloroethane
328 Tetrachioroethylene
329 Carbon tetrachioride
330 2 ,3,4,6-Tetrachiorophenol
331 Tetraethyldi thiopyrophosphate
332 Tetraethyl lead
333 Tetraethylpyrophosphate
334 Tetranitromethane
335 Thallium and Compounds, N O.S.
335A Thallium
336 Thallic oxide (Thallium (III) oxide)
337 Thallium (I) acetate (Acetic acid, thallium
338 Thallium (I) carbonate
339 Thallium (I) chloride
340 Thallium (I) nitrate
341 Thallium selenite
342 Thallium (I) sulfate
343 Thioacetamide
344 Thiosemicarbazide
345 Thiourea
346 Thiuram
347 Toluene
348 Toluenediamine
348A Toluene-2,4-Diamine
349 0-Toluidine hydrochloride
350 Toluene diisocyanate
351 Toxaphene
352 Bromoform
353 1 ,2 ,4-Trichlorobenzene
354 1 , 1 ,1 -Trichioroethane
355 1,1 ,2-Trichloroethane
356 Trichioroethylene
357 Trichioromethanethiol
358 Trichlorofluoromethane
359 2,4 ,5-Trichlorophenol
360 2 ,4 ,6-Trichlorophenol
361 2,4 ,5-T
362 2,4 ,5-TP (Silvex)
363 Trichloropropane, N 0 S
363A I • 1 ,2-trichloropropane
363B 1 ,2 ,2-trichloropropane
364 I ,2,3-Trichloropropane
365 0,0,0-Triethyl phosphorothioate
366 sym-Trinitrobenzene
367 Iris (1-Azridinyl) phosphine sulfide
-------
E TC ENVIRONMENTAL _______________
TESTING an CERTIFICATION
TABLE I
GRA1 TABLE
ETC
REF NO. COMPOUND NAME
368 Tris(2,3-dibromopropyl) phosphate
369 Trypan blue
370 Uracil mustard
371 Vanadic acid, ammonium salt
372 Vanadium pentoxide (Vandium CV) oxide)
372A Vanadium
373 Vinyl chloride
374 Zinc cyanide
374A Zinc
375 Zinc phosphide
-------
E TC ENVIRONMENTAL ______________________________
TEST/NC and CERTIFICATION
TABLE 2
PARAMETERS EXPECTED TO BE UNSTABLE IN WATER
40 Berizotrichioride
53 2-Butanorie peroxide
75 I -Chloro-2 ,3-epoxypropane
79 Chioromethyl methyl ether
138 1 ,2,3,4-Diepoxybutane
148 Dl isopropylfluoroohosphate
161 Dimethyl sulfate
220 Maleic anhydride
295 Phorate
298 Phthalic Anhydride
331 Tetrae thyldlthiopyrophosphate
-------
E TC ENVIRONMENTAL ________________________________
TESTING and CERTIFICATION
TABLE 3
“EXOTIC’ COMPOUNDS THAT REQUIRE SPECIAL METHODS
10 Aflatoxins
15 Mitomycin C
25 Auramine
26 Azasertne
52 Brucine
62 Chlorambucil
86 Citrus red No. 2
98 CycloDhosDhamlcie
99 Daunomyciri
103 Diallate
147 3,4-Dihydroxy-alpha-(methylamino)methyl benzyl alcohol
173 2 4-Dithiobiuret
188 Formaldehyde
189 Formic acid
202 Hexaethyl tetraphosphate
241 N-Methyl-N’-nitrosoguanidine
262 4-Nit roquinoline-1-oxide
277 N-Nit rososarcosine
279 Diphosphoramide, octamethyl
281 Endothall
282 Paraldehyde
304 1 ,3-Propane sultone
306 Propylihiouracil
311 Saccharin and salts
320 Streptozotocin
322 Strychnine and salts
324 2,3.7 ,8-T DD
334 Tetranitromethane
343 Thtoacetamide
344 Thiosemicarbazide
367 Tri( I -azridxnyl)phosphine sulfide
369 Trypan blue
370 Uracil mustard
-------
E TC ENVIRONMENTAL
TESTING and CERTIFICATION
TABLE 4
COMPOUND CLASSES
CLASS
REPRESENTATIVE
TAB LE*
Antimony and Compounds, N 0 S
Arsenic and Compounds, N 0 S
Barium and Compounds. N 0 S
Beryllium and Compounds, N 0 S
Cadmium and Compounds, N 0 S
Chromium and Compounds, N 0 S
Lead and Compounds, N 0 S
Mercury and Compounds, N 0 S
Nickel and Compounds, N 0 S
Selenium and Compounds, N 0 S.
Silver and Compounds, N 0 S
Thallium and Compounds, N 0 S
Aflatox ins
Chlorinated Benzenes, N 0 S
Antimony QR29
Arsenic QR29
Barium QR29
Beryllium QR29
Cadmium QR29
Chromium QR29
Lead QR29
Mercury QR29
Nickel QR29
Selenium QR29
Silver QR29
Thallium QR29
Aflatoxins, Total QR26
I ,2 ,3-trichlorobenzene QR28
2,4 ,6-trichlorobenzene QR28
I ,2 ,3,4-tetrachlorobenzene QR28
I ,2,3,5 Let rachlorobenzene QR28
1 ,2-Dichlorobenzene QR28
I ,3-Dichlorobenzene QR28
I ,4-Dichlorobenzene QR28
,2 ,4-Trichlorobenzene QR28
Chloroethane QR27
I , 1 , I -Trichloroethane QR27
I ,1 ,2-Trichloroethane QR27
Freon TF QR27
1 -chloronaphthalene QR28
2,3,5,6—tetrachiorophenol QR28
2,3,4,5-tetrachlorophenol QR28
2,4-Dichiorophenol QR28
2 ,6-Dichlorophenol QR28
2,4 ,5-Trichlorophenol QR28
2,4 ,6-Trichiorophenol QR28
bis(2-chloroethyl) ether QR28
bis(2-chloroisopropyl)etherQR28
2-Chloroethyl vinyl ether QR27
Benzene QR27
Acenapthene QR28
Acenaphthalene QR28
Anthracene QR28
3,4-Benzofluoranthene QR28
Benzo(ghi )perylene QR28
Benzo(k)fluoranthene QR28
Fluorene QR28
Phenathrene QR28
Pyrene QR28
Cresols SEE BEL Yv)
Napthalene QR28
Phenol QR28
Toluene QR27
19
21
27
42
56
84
216
225
250
314
317
335
10
64
65 Chlorinated Ethane, N 0 S
66 Chlorinated
67 Chlorinated
68 Chlorinated
Fluorocarbons, N 0 S
Naphthalene, N 0 S
Phenol, N 0 S
19A
21A
2 7A
4 2A
56A
84A
216A
225A
250A
31 4A
31 7A
335A
1 OA
64A
64 B
64C
64D
115
116
117
353
65A
354
355
66A
6 7A
68A
688
128
1 29
359
360
44
46
76
31
8 7A
878
8 7C
8 7D
87E
8 7F
8 7G
8 7H
871
go
245
289
347
70 Chioroalkyl Ethers, N 0 S
87 Coal Tars
Results table numoer in ETC Appendix VIII technical report
-------
E TC ENVIRONMENTAL __________________
TESTING and CERTIFICATION
TABLE 4 (cont’d)
COMPOUND CLASSES
CLASS REPRESENTATIVE TABLE*
89 Creosote 74 p-Chloro-m-cresol QR28
89A 2,Ni.trophenol QR2B
90A o-Nesol QR28
90B m resol QR26
163 4,6-Dinitro-o-cresol QR28
261 4-Nitrophenol QR28
90 Cresols 90A o-Cresol QR28
90B m+p Cresol QR28
92 Cyanides, N 0 S 92A Cyanide, Total QR29
118 Dichlorobenzene, N OS. 115 1,2-Dichlorobenzene QR28
116 l ,3-Dichlorobenzene QR28
117 1,4 chlorobenzene QR28
125 Dichloroethylene, N Os. 126 1 , Lnch1oro ’’ylene QR27
132 Dichioropropane, N O.S 132A 1,3-Dichior ne QR27
133 1,2-Dichioropropane QR27
134 Dichloropropanol, N O.S. 134A Dichioropropanol QR26
162 Dinitrobenzene, N 0 S. 162A m-Dinitrobenzene QR28
191 H 1oniethane, N 0 S. 191A Ch1orod]br,morn Ihane QR27
191B Dich1’ iobronioethane QR27
263 Nit rosamines, N.O.S. 263A N-Nit rosodiphenylamine QR28
264 N-Nitrosodi-n-butylamine QR28
265 N-Nitrosodiethanolamine QR28
266 N-Nitrosodiethylarnine QR28
267 N-Nitrosodimethylamine QR28
268 N-Nitroso-N-ethylurea QR25
269 N-Nitrosomethylethylamine QR28
270 N-Nitroso-N-methylurea QR25
271 N-Nitroso-N-methylurethane QR28
272 N-Nitrosomethylvinylamine QR28
273 N-Nitrosomorpholine QP28
274 N-Nitrosonornicotine QR28
275 N-Nitrosopiperidine 0R28
276 N-Nitrosopyrrolidine QR28
277 N-Nitrososarcosine QR28
297 Phthalic acid esters 48 bis(2-ethylhexyl)phthalate QR28
54 Butyl benzyl phthalate QR28
114 Di-N-Butyl phihalate QR28
143 Diethyl phthalate QR28
160 Dirnethyl phthalate QR28
167 Di-n-octyl phthalate QR28
300 Polychiorinated Biphenyl, N 0 S 300A Aroclor 1242 QR3O
300B Aroclor 1254 QR3O
300C Aroclor 1260 QR3O
300D Aroclor 1248 QR3O
300E Aroclor 1232 QR3O
300F Aroclor 1221 QR3O
300G Aroclor 1016 QR3O
325 Tetrachioroethane 326 1,1,1,2-Tetrachioroethane QP27
327 1 .1 ,2,2-Tetrachloroetharie QR27
348 Toluenediamine 348A Toluene-2,4-Diamine QR25
363 Trichioropropane, N.O S. 363A 1,1,2-trichioropropane QR27
363B 1,2,2-Trichioropropane QR27
364 1,2,3-Trichioropropane QR27
*Results table number in ETC Appendix VIII technical report
-------
E TC ENVIRONMENTAL
TESTING and CERTIFICATION
TABLE 5
SUF+IARY OF PARAMETERS UNDER THE
ETC ANALYTICAL SCHEME
PARAMETER NUMBER OF NUMBER OF , UMBER OF STD
CLASSIFICATION PARAMETERS UNAVAILABLE STD AND MASS
SPECTRUM UNAVAILABLE
UNSTABLE 13* N/A N/A
EXOTIC 10* N/A N/A
METALS 48 0 N/A
CONV 8 0 N/A
(Cyanide & Sulfide)
P 1 ST 55 5 4**
DAI 22 1 0
A/B/N 149 30 25***
HERB/PEST 28 0 N/A
HPLC 37 7 N/A
CLASS 5* N/A N/A
TOTAL 375 43 29***
* See discussion under Appendix VIII Parameter Section
‘* Compound Class represented by other parameters
This number is included in the no standard number
N/A - Not Applicable
-------
E TC ENVIRONMENTAL ____________
TESTING and CERTIFICATION
TABLE 6
SPIKING DATA OF SOME APPENDIX VIII COMPOUNDS
(N 7)
Concentration Mean Stci
Added Recovery Deviation
Coiiwound u /l
(DAI/INJ)
12 Allyl alcohol 62500 112 12
182 Ethyl methacrylate 62500 121 13
307 2-ProDyn-1-ol 62500 117 15
(P&T)
209 lodomethane 10 109 6
364 1 ,2 ,3-Trichioropropane 10 118 11
(A/B/N)
2 Acetophenone 50 91 6
14 4-Aminobiphenyl 50 104 6
67A I-chloronapthalene 50 128 30
151 p-dimethylamjnoazobenzene 50 106 13
183 Ethyl methanesulfonate 50 65 12
239 Methyl methariesulfonate 50 65 8
247 1-Naphthylamine 50 82 5
248 2-Naphthylamine 50 93 10
255 p-Nitroaniline 300 87 5
275 N-Nitrosopiperidjne 50 78 7
284 Pentachlorobenzene 50 92 10
286 Pentachloron trobenzene 300 60 3
288 Phenacetin 50 120 8
299 2-Picoline 50 57 15
303 Pronamide 50 92 6
(PEST/HERB)
101 4,4-DDE 0 5 107 10
144 Thionazin 2 5 53 12
172 Disulfoton 12 5 69 8
230 Methoxychior 50 76 8
242 Methyl parathion 2 5 119 14
283 Parathion 2 5 125 9
296 Famphur 25 131 ii
333 Tetraethylpyrophospate 125 58 19
361 2,4 ,5-T 10 92 28
362 2 ,4 ,5-TP (Silvex) 10 96 17
-------
E TC ENVIRONMENTAL ______________________________________
TESTING and CERTIFICATION
TABLE 6 (Continued)
Concentration Mean td
Added Recovery viation
Cornoound 72 72
(HPLC)
3 Warfarin 250 53
150 3,3 -Dimethoxybenzidine 250 83 3
249 1-Napthyl-2-thiourea 250 78 18
290B - -phenylenediamine 250 9
292 N-Phenylthiourea 250 I1 33
348A To1uene-2 4-Diamine 250 53 25
8 Acrylarnide 5250 97 10
82 l-(o-Chlorophenyl)thiourea 5000 81 2
292 N-Phenylthiourea 5000 109
-------
E TC ENVIRONMENTAL _________________________________
TESTING and CERTIFICATION
TABLE 7
COMPOUNDS THAT REQUIRE FURTHER INVESTIGATION
Comoounds Fractions
69 Chioroacetaldehyde DAI
83 3-Chioropropionitrile DAI
155 1,1-Dimethyihydrazine DAI
203 Hydrazine DAI
227 Methanethiol DAI
236 Methyl hydrazine DAI
237 2-Methyllactonitrile DAI
305 n-Propylamine DAI
91 Crotonaldehyde P81
16 5-(Aminomethyl)-3-isoxazolol A/B/N
157 Thiofanox A/B/N
158 a-a-Dimethylphenethylamine A/B/N
223 Melphalar A’B,N
265 N-Nitrosodiethanolamirie A/B/N
269 N-Nitrosomethylethylamirie A/B/N
310 Resorcinol A/B/N
346 Ihiuram A/B/N
368 Tris(2,3-dibromopropyl)phosphate A/B/N
-------
THE USE OF SW-846 CLEANUP P ND MASS SPECTROSCOPY METHODS TO IDEZJTIFY
AND QUANTIFY COMPOUNDS IN COMPLEX INDUSTRIAL WASTES: PETROLEUM
INDUSTRY CASE INDUSTRIES
A. E. BOOrIHOMER, RADIAN CORPORATION, MCLEAN, VIRGINIA
ABSTRACT
SW—846 cleanup techniques are designed to address two concerns: the
elimination of matrix interferences and the lowering of detection
limits for constituents of concern. Traditional Sw—846 methods for
semivolatile constituents in oily samples utilize a series of
methylene chloride extractions to isolate the components of interest.
An acid/base extraction may be used to separate constituents based on
their equilibrium acidity constants. Additional cleanup can now be
accomplished by the use of a proposed method that separates
base/neutral constituents by polarity on an alumina column. Many
constituents of concern in petroleum wastes are aromatic in nature,
and their identification and detection is greatly enhanced through
isolation from the aliphatics commonly found in petroleum residuals.
Mass spectral techniques are used to identify and quantity
constituents of concern for all isolated fractions.
INTRODUCTION
This paper presents a proposed SW—846 method designed to eliminate
matrix interferences and lower detection limits for target
constituents in oily wastes. The method is applicable to any sample
containing significant concentrations of aliphatic hydrocarbons or
highly polar compounds which interfere with GC/MS identification and
quantitation of aromatic constituents. To date, the method has been
applied to solid wastes from the petroleum refining, wood preserving,
coke by—products, and coal slurry pipeline industries. It is
applicable to virtually any sample matrix including oily sludges,
emulsions, soils, dry solids, oily liquids, and aqueous liquids.
The work presented here is being funded by EPA’s Office of Solid
Waste, Characterization and Assessment Division. The initial
development effort for the column cleanup method was performed by the
ERCO Division of ENSECO, of Cambridge, Massachusetts, and S—Cubed,
Inc., of La Jolla, California under contract to EPA. Radian
Corporation (Contract No. 68—01—6940) continued the methods
development effort and analyzed numerous petroleum refining samples as
part of EPA’S hazardous waste characterization effort.
The method, Column Cleanup for Oily Wastes, has been designated SW—846
Method 3570. Although only a proposed method, it is now used
routinely by the Office of Solid Waste’s Characterization and
Assessment Division for the analysis of oily wastes. The method is
almost always used in sequence with other SW—846 extraction and
cleanup methods (the 3500 series). Extracts obtained are analyzed by
-------
Gas Chromatography/Mass Spectrometry (GC/14S). In summary, a
base/neutral extract of the waste to be analyzed is applied to a
neutral alumina column. The column is sequentially eluted with
hexane, methylene chloride, and methanol to obtain the aliphatic,
aromatic, and polar constituents of the sample, respectively. The
aliphatic fraction contains most of the interfering compounds and is
generally not analyzed. The aromatic fraction is analyzed by Method
8270, GC/MS analysis for semivolatile organics. Analysis of the polar
fraction has not been performed by Radian.
ANALYSIS OF OILY SAMPLES
Conventional SW—846 methods are designed to extract methylene chloride
soluble constituents from a waste matrix in preparation for analysis
of these constituents by GC/MS. In the analysis of petroleum refining
wastes, it was found that this methodology frequently resulted in the
extraction and identification of primarily aliphatic hydrocarbons.
Further, the abundance and high concentration of these aliphatic
hydrocarbons in petroleum samples interfered with identification and
quantification of the aromatic constituents of interest to EPA.
Constituents of interest include those given in Part 261, Appendix
VIII of RCRA and other compounds believed to present a substantial
threat to human health and the environment if mismanaged. The rather
lengthy list of chemicals given in Appendix VIII was evaluated for
constituents likely to be encountered in petroleum refining wastes. A
target list of chemicals was then developed from this study of the
industry. Many of the target compounds identified are indigenous to
crude oil or are produced from crude during refining processes. Other
chemicals are used in refining operations as process chemicals and end
up in the industry’s wastes. The list of target chemicals is
presented in Table 1 and includes many structurally similar compounds,
principally polynuclear aromatics (PNAS) and substituted aromatics.
EXTRACTION OF SEMIVOLATILE ORGANICS
Solid wastes as defined by RCRA occur in a wide variety of sample
matrices. In the case of petroleum residuals, these matrices include
oily sludges, dry solids, semisolids, emulsions, organic, and aqueous
liquids. SW—846 methods for the analysis of semivolatile constituents
employ Gas Chromatography/Mass Spectrometry (GC,41S) to identify and
quantify constituents. As preparation for GC/74S, sample extraction
and cleanup are performed in order to isolate target constituents.
There are two objectives of extraction and cleanup: (1) to eliminate
matrix interferences, and (2) to obtain adequately low detection
limits for the constituents of concern. In the analysis of petroleum
residuals, a series of extraction/cleanup techniques may be required
to meet these objectives. The analytical scheme for a sample is
designed based upon the known composition of the waste, upon the
target constituents to be identified, nd the detection limit required
for the target constituents.
The column cleanup (Method 3570), is not dependent upon sample matrix.
-------
In virtually all cases, an aliquot of the waste extract is subjected
to column cleanup. The method selected for obtaining this waste
extract does depend, however, upon the matrix of the sample. Five
extraction and cleanup methods (SW—846) are currently used for the
preparation of solid wastes for analysis. Two are specifically
designed for aqueous liquids, and two are designed to be used with
solid, semisolid, or emulsion sample matrices. The fifth method is
used on extracts resulting from the use of one of the four other
methods.
Figure 1 presents a typical series of extraction and cleanup methods
(including a column cleanup step) used on an oily waste slated for
semivolatile analysis via GC/MS. Organic liquids or solids which are
completely soluble in methylene chloride (or hexane) do not require
extraction, although the column cleanup for these samples is
frequently necessary. Other liquids, generally aqueous samples, are
extracted by either Method 3510 or 3520, Separatory Funnel and
Liquid/Liquid Extraction, respectively. These methods extract
semivolatile and nonvolatile constituents with methylene chloride.
Acid/base extraction (discussed below) may be employed during the
methylene chloride extraction for both these methods. Methods 3540
and 3550, Soxhlet and Sonication Extraction, respectively, are used to
extract solids, sludges, semisolids, slurries, and emulsions.
Acid/base extraction may be used to isolate sample constituents as a
function of their acidic, basic, or neutral properties. For petroleum
samples, typical acid compounds include phenolics and thiols, although
other compounds may also be included. Obtaining the acid fraction
allows for the identification of phenolics and cresolics without
necessitating the analysis of the polar fraction. Identification of
phenolics in petroleum samples could be hindered by the presence of
base/neutral compounds. Base/neutrals tend to interfere in the
analysis of acid constituents in three ways: (1) the GC,41S response
for phenolics is relatively low compared to many base/neutral
compounds; (2) high concentrations of base/neutrals eluting at the
same time in the chromatogram with the acid compounds may
significantly raise detection limits; and (3) high concentrations of
base/neutral constituents act to dilute acid as part of compounds in
the extract. Isolation of acid compounds in the polar fraction was
also considered but recoveries for these compounds might be low and
subject to increased error.
COLUMN CLEANUP FOR OILY WASTES
When EPA’s Office of Solid waste began its study of oily wastes from
petroleum refining, it was confronted with the problem of identifying
and quantifying target compounds in the base/neutral extract. The
same interferences encountered with acid compounds (when not separated
from the base/neutrals), were equally intense within the base/neutral
extract for the target constituents. Detection limits were found to
be unacceptably high.
In a typical sample of oily waste from petroleum refining, PNA’s of
-------
interest in a sample range in concentration from 1 to 100 ppm. The
sample might contain 50 percent methylene chloride extractable
material. This amount of extractable material would necessitate an
injection size that overloaded the GCflIS detection system with
aliphatics to achieve a detectable mass for a PNA. To alleviate this
problem, the column cleanup method separates constituents in the
base/neutral extract based upon their relative polarity. The
separation is designed to isolate aliphatic, aromatic and polar
constituents from one another. Chemical principles used in developing
this method have been well—documented (1—3).
SUMMARY OF THE COLUMN CLEANUP METHOD
A discussion of the key elements of the column cleanup method is
presented here. A 100 to 200 mg aliquot of the base/neutral extract
is solvent exchanged into hexane. The final volume of the extract
should be adequate to just solubilize semivolatile organic
constituents. The dry hexane aliquot is applied to a neutral alumina
column (70 to 230 mesh), and eluted with 13 mL of hexane. The hexane
fraction, designated base/neutral aliphatics is collected. The
aliphatic fraction is not generally analyzed but is retained should
further analysis be indicated. The column is next eluted with 100 mL
of methylene chloride. The fraction obtained is collected and labeled
base/neutral aromatics. Finally the column is eluted with 100 mL of
methanol, the fraction collected, and labeled base/neutral polars.
The polar fraction is retained for later analysis if indicated. The
volume of each fraction to be analyzed is reduced to 1 mL.
Ten percent (100 micrometers L) of the base/neutral aromatic fraction
is evaporated to dryness and subjected to gravimetric analysis to
determine the concentration of aromatics in the base/neutral fraction.
Similar gravimetric analyses may be performed on the aliphatic and
polar fractions should this information be desired. The final volume
of the extract may be adjusted to achieve the desired concentration
for injection, and the aromatic fraction is analyzed by Method 8270,
GC/MS Method for Semivolatile Organics. Spectra for all target
constituents are searched for in the chromatogram and the 10 major
components of the aromatic fraction are identified. Figures 2 and 3
present typical chromatograms for the aliphatic and aromatic
fractions. For most petroleum samples, the aliphatic fraction is
quickly recognized because of its characteristic distribution of
normal alkanes. The aromatic fraction does not show the organized
distribution of alkanes present in the aliphatic fraction and may show
a large undifferentiated peak along with the individual constituents.
SEP R TION OF ALIPH TIC PiND AROMATIC CONSTITUENTS
Method 3570’s primary objective is to eliminate interferences caused
by high concentrations of aliphatics in petroleum refining wastes.
Development of the method was directed toward achieving the split
between aliphatics and aromatics. In this pursuit, the amount of
hexane used to elute the aliphatic fraction was carefully considered.
-------
Ultimately, an amount equivalent to one column pore volume was
selected.
Figure 4 presents the single ion and reconstructed ion chromatograms
for a base/neutral aromatic extract. The extracted ion current
profile for the ion mass 57 is shown in the top chromatograin. This
ion is characteristic of saturated hydrocarbons. The reconstructed
ion chromatograin obtained for the same elution, with 13 mL of hexane
used in the chromatographic cleanup, is shown in the lower portion of
the figure. A 1 micrometer L injection of the sample extract was
made, and a broad undifferentiated peak is readily visible in the
lower chromatogram. The top chromatogram shows significant
concentrations of saturated hydrocarbons, particularly in the latter
portions of the run. A close examination of this profile
demonstrates, however, that there is not an overwhelming presence of
hydrocarbons. The normal alkanes (from approximately 20 carbon atoms
to 35 carbon atoms) are indeed present in the sample. However, the
elution of the major mass of the normal alkanes does not coincide with
the peak of eluted material in the undifferentiated peak. The
baseline upon which these alkanes appear is quite flat.
Figure 5 presents the relationship between amount of hexane used as
the solvent for the aliphatic fraction, and the elution of three low
molecular weight PNAS from the aromatic fraction. As shown in the
figure, increasing the amounts of hexane used to elute “aliphatics”
from the column tends to reduce naphthalene concentrations in the
aromatic fraction. Presumably, large volumes of hexane act to elute
naphthenics along with aliphatics. The aliphatic fraction was not
evaluated in this study but has been examined in other evaluations.
It has been found that, indeed, naphthalene is eluted in the hexane
fraction when large volumes of the solvent are used. Table 2 presents
a complete summary of the analysis of the aromatic fraction for the
successive amounts of hexane used to elute the aliphatics. High
molecular weight PNA’s are not significantly affected by increasing
the amount of hexane.
SURROGATES AND SURROGATE RECOVERIES
The column cleanup method employs surrogates that act as column
performance indicators. The surrogates are added at the beginning of
the analyses, preceding extraction, and are used to demonstrate
recoveries for specific types of constituents. Surrogates used for
Method 3570 are 2—fluorobiphenyl, nitrobenzene—d5, pyrene—dlO, and
terphenyl—d14. Phenol—d5 may be added immediately preceding
application of the extract to the column if the polar fraction is to
be analyzed. Good recoveries of 2—fluorobiphenyl, a low molecular
weight diaromatic, indicates that the low molecular weight PNAs are
eluted and recovered in the analysis. Pyrene—dlO and terphenyl—d14
are used to demonstrate recoveries for the higher molecular weight
PNAs. These two surrogates elute in the later portion of the
chromatogram. The surrogate nitrobenzene—d5 is used to demonstrate
the extent to which relatively polar compounds are retained on the
column during the methylene chloride wash. It has been found that
-------
when no sample matrix interferences are present nitrobenzene—d5 is
almost always retained on the column during the methylene chloride
wash. That is, for blanks and other runs when only the surrogates are
spiked, nitrobenzene—dS is found in the polar fraction. When a sample
is spiked with this surrogate and the column cleanup performed,
however, nitrobenzene—d5 is found in the aromatic fraction. Matrix
effects within the sample appear to alter polarities enough to elute
nitrobenzene—d5. Recovery of nitrobenzene—d5 may, when more data have
been obtained, be useful in making correlations between concentrations
of specific types of polar constituents in the sample and elution
patterns.
RECOVERIES OF TARGET CONSTITUENTS
For most studies of petroleum samples, polynuclear aromatics (PNA.s)
are the major constituents of concern. The column cleanup’s
performance for these compounds is within the recovery ranges given in
SW—846. Column performance can be assessed by spiking known amounts
of pure compounds onto the column. The compounds are spiked in a
hexane solvent, and the cleanup and analysis sequences are performed
as specified in the method. For the column performance evaluations
presented here, only the aromatic fraction was analyzed. Table 3
summarizes the recovery of pure compounds spiked onto the column at an
average mass of 500 micrometers g. A more extensive column
performance evaluation was performed by another laboratory and the
results are summarized in Table 4. For this study, each compound was
spiked onto the column at a mass of 200 micrometers g. Recoveries in
both studies were determined by single ion quantification compared to
a calibration standard.
In the first study, recoveries of PNAs spiked at 500 micrometers g
each were within the range of 90 to 110 percent. In the second study,
the recoveries for the constituents spiked at 200 micrometers g were
more variable. High molecular weight PN .S were recovered well, as
they were in the previous study. Lower molecular weight hydrocarbon
aromatics were inadequately recovered. The reasons for the low
recoveries are unclear, since in other evaluations these compounds
have shown good recoveries. This performance evaluation has not been
repeated. It is possible that there is a matrix effect from the
relatively high percentage of polar compounds in the spike or that
some other effect is being seen. When this group of constituents was
spiked into an oily waste sample containing high concentrations of
aliphatics, recoveries for these low molecular weight compounds were
significantly higher, and were within the ranges specified in SW—846,
(discussed in the following subsection). For other sample
constituents, including oxygenated compounds, amines, and
nitroaromatics, the results were variable. The phenyl ethers and
dibenzofuran are of intermediate polarity and showed good recoveries.
Recovery of the most polar constituents was not reproduced, and did
not meet the criteria specified in SW—846. The phthalic acid esters
were recovered in amounts that indicate their appearance could be a
result of contamination rather than recovery of the spike. Amines and
nitroaromatics were not recovered. The phthalates, nitroaromatics,
-------
and amines are expected to be eluted as part of the polar fraction.
This fraction was not analyzed because it presents analytical problems
that have not yet been resolved by Radian.
METHOD PERFORMANCE FOR OILY WASTE S1 MPLES
Spiking pure compounds onto a column is an ideal situation; real
samples can present a more complex analytical challenge. To determine
the effect of the sample matrix, document laboratory repeatability,
and demonstrate recoveries for target constituents, a methods study
was performed. In the first phase of the study, aliquots of the
sample were spiked with only the surrogates. The
extraction—cleanup—analysis sequence (Methods 3540—3530—3570—8270) was
performed on each aliquot, and the mean constituent concentrations are
determined. In the second phase of the methods study, each target
constituent was spiked into aliquots of the sample and the
extraction—cleanup-analysis sequence repeated. Table S presents the
results of the methods study (aromatic fraction) for constituents of
interest spiked into an API Separator Sludge. Each constituent was
spiked at a concentration of 200 micrometers g in a 1 gram aliquot.
This spiking was performed preceding all sample extractions, and four
aliquots of the unspiked sample and four aliquots of the spiked sample
were analyzed. Mean constituent concentrations and standard
deviations for each of the eight aliquots are presented in the table
along with the mean percent recovery for each constituent.
The data show very good recoveries for the PNAS. Other constituents
including the phthalates, amines, and nitroaromatics, are recovered
poorly or not at all. These results are consistent with the results
of the column performance check in Table 4. One interesting result of
the methods study was that the low molecular weight aromatics were
recovered better in this study than in the column performance
evaluation. Additional work will be needed to investigate this
apparent anomaly.
ANALYSIS OF POLAR FRACTION
Limited work has been performed on the polar fractions, data for
target compounds are not available. There are two problems
encountered with this fraction. First, the methanol solvent tends to
promote degradation of the CC column used for analytical separation
even when a bonded phase column is used. Second, it appears that in
some cases methanol extracts material from the alumina, which in turn
interferes with the CC/MS analysis.
SUMMARY
The column cleanup for oily wastes reduces interferences and lowers
detection limits for constituents of concern in petroleum refining
wastes. The method has been shown to perform well in combination with
established SW—846 extraction and analysis methods. It is
-------
particularly useful in reducing interferences from aliphatic
hydrocarbons in the analysis of polynuclear arornatics (PNl s).
Recoveries of PNAS subjected to the column cleanup are within the
ranges established in Sw—846, Method 8270, GCftIS Analysis of
Semivolatile Organics. Other constituents have also been recovered
within acceptable ranges. Highly polar constituents, believed to
elute in the polar fraction, have not been extensively studied.
Analytical problems with the evaluation of the polar fraction have
limited study of recoveries for target constituents in this fraction.
-------
REFERENCES
Heftmann, E. Chromatography Part B: Applications . New York, New
York: Elsemer Scientific Publishiii , 1983. p. B532.
Neumann, Hans—Joachim, Barbara Paczynska—Lahme, and Dieter Sererin.
Composition and Properties of Petroleum . New York, New York: Halsted
Press, 1981. p. 42.
Brown, Ralph A. and Thomas D. Searle. “Analysis of Polynuclear
Aromatic Hydrocarbons” in Chromatography in Petroleum Analyses , Klaus
H. Altgelt and T.H. Gouw (editors). New York, New York: Marcel
Dekker, 1979.f
-------
CC/MS
for
Semivolatiles
(Method 8270)
Extraction Acid/Base
_________ (Methods 3510, Extraction
Semivolatile
3520, 3540, (Methods 3520
Organics
or 3550) or 3530)
CC/MS
for
Column
Semivolatiles
Cleanup ‘ (Method 8270)
(Method 3570)
Aromatic
(and Polar)
Fraction(s)
Figure 1
SCHEMATIC REPRESENTATION OF ANALYSES FOR SEMIVOLATILE
ORCANIC CONSTITUENTS IN PETROLEUM REFINING SAMPLES
-------
Figure 2
TYPICAL RECONSTRUCTED ION CHROMATOGRAN FOR THE ALIPHATIC FRACTION
Fila 3O169 50.0-400.0 omu. RAG OIL 4077-1 AL1O 6OMOBSW. 35310U10.IH1.luL. 6/12/84GAI
TIC
200000
19000ff
1 6000ff
I 4000ff
1 2000ff
100000•
0000ff
10 12
14
16
18
20
22
24
26
28
30
32
34
-------
Fils 3O164 50.0—400.0 amu. RAG OIL 4077-IF AROM8OM085W. 35310SW. IHI. 1.OuL . 6/10/84.
TIC
Figure 3
1501
I
60000
50000
40000
ii
12 14 16 1 20 22 24 26 28 30 32 34
Il -I -i -i
TYPICAL RECONSTRUCTED ION CHROMATOGRAN FOR THE AROMATIC FRACTION
-------
IIIDRIC.IIASS CI 0I’V1T0G [ %TAi 5N 659 *1
85/21/85 1?.Slse C h Ca.T *2
SNfLE. 05W
-
RkVIGEiG 1.3880 L €LsN • 4.8 QLb 4oA L1.OJ B W6EsU2B . 3
SC 15 230 TO 3880
8224.
57.01?
* 0.500
Figure 4
SINGLE ION CHROMATOGRAM SHOWING THE PRESENCE OF
ALIPHATICS IN AN AROMATIC FRACTION
100.
RIC
500 1000 1580 2000 2580 3000 3500 SCAH
0s06 16 12 24s18 32i24 40s31 48*3? 56*43 TIME
-------
h
- If. --_______
20 30 50 60 70
Volume Rexane Added, m l ..
Figure 5
THE RELATIONSHIP BETWEEN HEXANE ELUTION IN THE ALIPHATIC FRACTION
AND NAPHTHALENE CONCENTRATIONS IN THE AROMATIC FRACTION
2-methyl naphthalene
pht ha I ene
40
80
-------
Table 1
SEMIVOLATILE TARGET CONSTITUENTS ANALYZED IN THIS STUDY
kcid x.tractables
Benzenethiol
p—Chloro—m—cresol
2—Chiorophenol
Cresols
2,4—Dimethyl phenol
4, 6—Dinitro—o—creso].
2, 4—Dinitrophenol
Nitrophenols
Pentachiorophenol
Phenol
Trichiorophenols
Base/Neutral Extractable
Aniline
Anthracene
Benz(c)acridine
Benz(a )anthracene
Benzo(b)fluoranthene
Benzo(j )fluoranthene
Benzo(k )fluoranthene
Benzo(a )pyrene
Benzyl chloride
Bis(2—ethylhexyl) phthalate
Butylbenzyl phthalate
2—Chloronaphthalene
Chrysene
Dibenz(a,h)acridine
Dibenz(a,j)acridine
Dibenz (a, h)anthracene
7, 12—Dimethylbenz(a)anthracene
Dibenzo(a, e)pyrene
Dibenzo(a,h)pyrene
Dibenzo(a, i)pyrene
Di-.n—butyl phthalate
Dichlorobenzenes
Diethyl phthalate
Dimethyl phthalate
Dinitrotoluene
Di—n—octyl phthalate
Fluoranthene
Indene
Indeno(1,2,3—c,d)pyrene
Methylbenzo(c )phenanthrene
3—Me thy ichol an thene
Methylchrysene
1—Me thy lnaphthalene
Naphthalene
Naphthylamine
5—Nitroacenaphthene
p—Nitroaniline
Nitrobenzene
Phenanthrene
Pyrene
Quinoline
Styrene
Trichlorobenzenes
Trimethylbenz(a)anthracene
-------
Table 2
CONCENTRATIONS OF TARGET CONSTITUENTS IN THE AROMATIC FRACTION WITH
INCREASING AMOUNTS OF HEXANE TO ELUTE ALIPHATICS
A B C D
13 mL 26 mL 39 mL 78 mL
Compound Hexane Hexane Hexane Hexane
indene 1.7 j.zg/g 1.2 1.1 NF
naphthalene 123. 14 108.3 103.8 8.8 .ig/g
2—methylnaphtha lene 300.2 261.0 252.8 122.7
1—methylnaphthalene 224.6 192.4 181.8 27.0
acenaphthene 16.2 15.0 18.4 12.5
fluorene 32.6 23.0 27.0 311.9
phenanthrene 97.7 65.8 69.6 711.2
anthracene 5.8 5.3 5.9 5.2
dl—n—buty]. phthalate 0.4 0.5 0.3 0.3
fluoranthene 6.5 5.2 5.9 5.3
pyrene 31.5 27.9 27.3 33.7
chrysene 145.7 25.2 23.8 32.2
benzo(a)anthracene 53.7 39.6 143.2 514.2
bls(2—ethylhexyl) phthalate 1.14 5.5 1.2 2.7
benzo(b)fluoranthene 5.14 3.9 14 1 4
benzo(a)pyrene 7.9 6.3 7.8 8.7
dlbenzo(a,h)anthracene 1.7 0.7 5.6 2.0
indeno(1,2,3 - .c,d)pyrene 0.9 1.6 14.5 14.6
dibenzofuran 3.2 5.0 5.3 10.4
benzo(k)fluoranthene NF NF 2.3 1.3
benzo(g,h,i)perylene NF NF 2.5 NF
-------
Table 3
RECOVERY OF PNAs FOR METHOD 3570 COLUMN PERFORMANCE STUDY
Concentration ug
Comoound Spiked Recovered % Recovered
Naphthalene 560 580 1014
Acenaphthylene 630 600 95
Acenaphthene 5140 1490 91
Fluorerie 500 520 1014
Phenanthrene 5140 550 102
Anthracene 570 550 96
Fluoranthene 560 600 107
Pyrene 530 590 111
Benz(a)anthracene 1480 1490 102
Chrysene 260 280 108
Benzo(b)fluoranthene 11140 1480 109
Benzo(k)fluoranthene 1450 1430 95
Benzo(a)pyrene 650 670 103
Dibenzo(a,h)anthracene 1460 1450 98
Indeno(1,2,3—c,d)pyrene ‘450 1450 100
Benzo(g,h,i)perylerie 390 380 97
-------
Table 14
RECOVERY OF TARGET CONSTITUENTS FOR METHOD 3570
COLUMN PERFORMANCE STUDY
Test #1 Test #2
Compounds Observed Blank Relative %
in Standard n /uL Recovery Recovery Differencea
N- .nitrosodimethylamirie NR NR — —
Styrene 2 2 0
aniline NR NR ——
bis(2—chloroethyl)ether NR NR ——
n—dichlorobenzene 3 2 10.0
. —dichlorobenzene 3 14 7.1
benzyl chloride NR MR ——
benzyl alcohol NH NR ——
Q—di chlorobenzene 6 5 14.5
indene 7 6 3.8
bis(2—chloroisopropyl) NH NH ——
e the r
N—nitrosodipropylamine MR NR
hexachioroethane NR NH
nitrobenzene NR NR
isophorone NR NH
bis(2—chloroethoxy) NH NH
methane
1,2, 1 4—trichlorobenzene 3 3 0
naphthalene 36.5 314 214 8.6
14—chioroaniline NR NH ——
hexachlorobutadiene NR NR ——
2—methylnaphthalene 29 21 8.0
hexachiorocyclopenta— NR NR ——
diene
2—chloronaphthalene 35 26 7.4
3—nitroaniline NR NR ——
dimethyl phthalate NR NH
2,6—dinitrotoluene NR NR — —
acenaphthylene 26 18 9.1
1 4—nitroaniline NR NH ——
acenaphthene 31 22 8.5
dibenzofurari 80 67 4.4
2,4—dinitrotoluene NH NR ——
diethyl phthalate 0.14 NR NR ——
fluorene 63 56 2.9
4—chiorophenyl phenyl 63 614 0.4
ether
azobenzene 87 65 7.2
14—bromophenyl phenyl 101 78 6.14
ether
hexachlorobenzene 10 9 2.6
phenanthrene 0.2 76 70 2.1
anthracene 92 73 5.8
-------
Table 4 (Continued)
RECOVERY OF TARGET CONSTITUENTS FOR METHOD 3570
COLUMN PERFORMANCE STUDY
Test #1 Test #2
Compounds Observed Blank I I Relative I
in Standard ng/uL Recovery Recovery Differencea
di—.n—butyl phthalate 0.14 Trace Trace ——
fluoranthene 104 80 6.5
pyrene 0.3 90 74 14.9
butyl benzyl phthalate NR NR ——
chrysene 0.3 104 71 9.14
benzo(a)anthracene 0.5 100 73 7.8
bis(2—ethylhexyl) 1.2 1 1 0
phthalate
di—j —octyl phthalate NR NR ——
benzo(b)fluoranthene 814 52 11.8
benzo(k)fluoranthene 91 6 14 8.7
benzo(a)pyrene 73 59 5•3
benzo(g,h,i)perylene 69 61 3.1
dibenzo(a,h)anthracene 82 60 7.7
indeno(1,2,3—cd)pyrene 93 56 12.14
2—nitroaniline NR NR ——
naphthylamine NH NR
a bi
2(a+b) • 100
NR = Not recovered.
-------
Table 5
RECOVERY OF
TARGET CONSTITUENTS SPIKED IN AN
API SEPARATOR SLUDGE
Conc ntratiôn (t / m)
Samole
Spiked
xa
SamDlec
Average
Percent
Recov ery
tJri piked
xa
1.7
62.’I
110.6
Styrene
——
514.6
19.6
27.3
m—Djchlorobenzene
——
123.2
114.5
61.6
p—Dichlorobenzene
——
1214.0
16.6
62.0
o—Dichlorobenzene
——
122.3
114.7
61.2
Indene
0.3
67.0
25.3
32.6
Hexachioroethane
——
19.6
18.0
9.8
1,2,4—Trichlorobenzene
——
167.8
143.2
83.9
Naphthalene
13.8
163.14
29.2
50.5
Hexachiorobutadiene
——
7.9
10.0
3.9
2—Methylnaphthalene
22.6
223.9
1414.2
61.6
2—Chloronaphthalene
——
1914.14
29.0
97.2
Acenaphthylene
Acenaphthene
—
6.5
——
1.6
1314.5
162.2
26.8
25.6
67.2
77.8
Dibenzofuran
2.8
0.8
189.14
145.1
93.2
Diethyl phthalate
0.3
0.2
0.3
0.1
0
Fluorene
17.7
14.2
189.1
36.8
85.14
1 4—Chlorophenyl phenyl
——
——
163.6
27.5
81.8
ether
Azobenzene
——
——
176.3
28.0
88.2
14—Bromophenyl phenyl
——
——
170.0
18.3
85.0
ether
Hexachlorobenzene
——
——
188.6
35.0
94.3
Phenanthrene
31.9
9.1
226.7
62.7
96.9
Anthracene
3.11
1.3
120.6
13.2
58.6
Di—n—butyl phthalate
0.5
0.2
0.5
0.1
0
Fluoranthene
1.2
1.0
1149.2
37.7
73.6
Pyrene
6.9
1.9
129.5
21.1
61.2
Chrysene
14.1
0.9
187.8
35.5
91.8
Benzo(a)anthracene
9.0
2.0
176.2
29.7
83.4
Bis(2—ethylhexyl)
3.0
1.6
7.0
9.0
2.0
phthalate
Benzo(b)fluorayithene
1.2
0.2
133.9
23.9
66.4
Benzo(k)fluoranthene
0.5
0.2
140.6
23.8
70.0
Benzo(a)pyrene
Benzo(g,h,i)perylene
1. 14
——
0.2
——
138.7
131.1
17.7
27.2
68.7
65.7
Dibenzo(a,h)anthracene
Indeno(1,2,3—c,d)
155.0
155.0
314.14
25.8
77.5
77.5
pyrene
Bis(2—chloroisopropyl)
0.8
1.7
0.4
er
Di—n—octyl phthalate
Butyl benzyl phthalate
0.1
0.14
0.3
0.8
0.4
NR
0.8
NR
0.2
0
-------
Table 5 (Continued)
RECOVERY OF TARGET CONSTITUENTS SPIKED IN AN
API SEPARATOR SLUDGE
Percent Recovery
Unspiked Samp Le Spiked SampLeC
xa xa
Naphthalene—d 8 52.5 10.5 56 11.8
2—Fluorobiphenyl 68.8 16.7 131 3L1.5
Nitrobenzene—d 5 23.2 27.2 NR NR
Terphenyl—d 111 66.2 17.8 69 18
Pyrene—d 10 59.0 13.1 59 12.6
aAverage of four analyses.
bSD = Standard Deviation
NR Not Recovered
Ccompounds spiked but not recovered:
N a ph thy 1 amine
Benzyl alcohol
Benzyl chloride
N—Nitrosodimethylamine
Aniline
Bis(2—chloroethyl) ether
N—Nitrosodipropy lamine
Nitrobenzene
Isophorone
Bis(2—chloroethoxy)methane
Lj..Chloroaniline
3—Nitroaniline
Dimethyl phthalate
2, 6—Dinitrotoluene
q—Nitroaniline
2, 1 1—Dinitrotoluene
N—Nitrosodiphenylamine
-------
QUANTIThTIVE ANALY’rICAL SCREEN FOR THE DETERMINAITION OF THE APPENDIX
VIII HAZARDOUS CONSTITUENTS
Dr. MARK J. CARTER, JERRY L. PARR, AND 0. JOHN LOGSDON, ROCKY MOUNTAIN
ANMJYTI CAL LABORATORY, ARVADA, COLORADO
ABSTRACT
On October 1, 1984, the Environmental Protection Agency proposed new
Subtitle C testing requirements under RCRA. As part of these testing
requirements many companies will be required to analyze groundwater
and waste samples for the 375 Hazardous Constituents listed in
Appendix VIII of 40 CFR Part 26 using methods in SW—846, “Text Methods
for Evaluating Solid Wastes.”
Due to a number of complex issues both in the proposed rule and in
SW—846, RMAL prepared a report entitled “Evaluation of the
Aplicability of the SW—846 Manual to Support All RCRA Subtitle C
Testing.” As part of this effort, RMAL identified, developed, and
proposed an analytical strategy to determine effectively Appendix VIII
consituents using the methods in SW 846.
This analytical strategy, designated as a Quantitative Analytical
Screen (QAS), utilized 17 different methods based on ICP,
AA, GC, AA, GC/MS and HPLC technology. The methods, consituents
measured by each method and the general approach of the QAS will be
discussed.
-------
METHODOLOGY FOR THE ANALYSIS OF ORGANIC CHEMICALS IN PETROLEUM
REFINING WASTES TO SUPPORT RCRA WASTE LISTING AND DELISTING AND LAND
TREATMENT DEMONSTRATION PROGRAMS
DR. MARK 3. CARTER, DR. MICHAEL P. PHILLIPS, AND JERRY L. PARR, ROCKY
MOUNTAIN ANALYTICAL LABORATORY, ARVADA, COLORADO
ABSTRACT
In April, 1983, EPA initiated a study to re—evaluate the basis for
listing the five regulated petroleum refining wastes. Subsequently,
the American Petroleum Institute retained Rocky Mountain Analytical
Laboratory (RMAL) to track this study. The scope of the study
included the analysis of a subset of approximately 90 organic
constituents drawn from the Appendix VIII list. From october 1983 to
October 1984, EPA prepared three methods manuals, and in July, 1985 a
guidance document concerning methods for analysis of petroleum
ref inery wastes was prepared.
The analysis of the volatile and semivolatile organics is based on
SW—846 methods 8240 and 8270 with separation and clean—up steps to
mitigate the interference from the aliphatic hydrocarbons. This paper
will present the results of a study performed by RMAL of the recovery
and detection limits of the “Skinner” organics in petroleum refinery
wastes.
The EPA has recently released preliminary detection limit requirements
for waste delisting. These limits are below the routine capability of
some of the methods developed for the EPA waste listing study.
Modifications of the methods necessary to obtain the lower detection
limits will be reported.
-------
DEVELOPMENT OF GROUNt ’JATER SCREENING PROCEDURES FOR USE IN MONITORING
PROGRAMS: OBJECTIVES AND EXPERIMENT PROGRESS AT BA 1 TrELLE COLUMBUS
LABORATORIES
S. V. LUCAS, BA11 ELLE COLUMBUS LABORATORIES, COLUMBUS, OHIO
ABSTRACT
1\ .zo of the Heirarchical Analysis Protocol (HAP) methods of SW—846 are
currently the subject of laboratory efforts to establish their
efficacy: Method 8610, “Total Aromatics by Ultraviolet Absorption,”
and Method 3560, “Reverse Phase Cartridge Extraction.” A data base of
physical and UV/visable spectroscopic properties has been established
using a personal computer. For the 90 polar compounds in Method 8610,
literature spectroscopic data were available for all but 10 compounds,
and two or more sources of data were found for 70 of these compounds
as well as for all 39 of the nonpolar analytes of the method. Using a
Cary 17D spectrophotometer, it was found that the target detection
threshold of 0.005 au was well within that instrument’s capabilities
of signal—to—noise ratio and spectrum reproducibility.
Evaluation of Method 3560 resulted in important modifications to the
cartridge cleanup and elution procedures. Using reagen water and
commercially available cartridges, a confident detection threshold of
0.02 au was estimated, and quantitative recovery of selected Method
8610 compounds at levels ranging from 15 to 3000 ug/L was obtained.
When groundwater from actual site monitoring wells was used, the
recovery of spiked compounds was apparently quantitative, but a
substantial variability in the recovery of UV absorbing material in
the groundwater itself prevented accurate quantification of spike
recoveries. The groundwater matrix interference problem seemed to be
associated with the presence of mineral particulate in these samples.
Work presently in progress is directed toward solving the Method 3560
problem of variability of recovery of UV absorbing material for
groundwaters, and the status, to date, of this on—going work as well
as the above described completed work will be presented.
This work is being conducted by Battelle Columbus Laboratories for the
U.S. EPA Environmental Monitoring Support Laboratory (EMSL),
Cincinnati, Ohio, under Contract 68—03—1760, Work Assignment 10. Dr.
Fred K. Kawahara is the EPA Project Officer, and Mr. James E.
Longbottom is the EPA Program Manager..
-------
FOURTH SESSION
SAMPLING UNDER RCRA
1:00 pm — 4:30 pm
Thursday, July 25, 1985
Chairperson: Martin Meyers
Methods Program
Office of Solid Waste
U. S. Environmental
Protection Agency
Washington, D. C.
-------
VOLATILE ORGANIC SNIPLING TRAINS FOR HAZARDOUS WASTE INCINERATORS:
LABORATORTY VALIDATION
THOMAS J. LOGAN, ROBERT G. FUERST, AND N. RODNEY MIDGETI’, U. S.
ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK, NORTH
CAROLINA; AND JOHN PROHASKA, PEI ASSOCIATES, INC., CINCINNAIT, OHIO
ABSTRACT
The measurement of volatile organic emissions from a hazardous waste
incinerator is one of the more difficult source testing problems.
Specific compounds called principal organic hazardous constituents
(POHC) are to be identified and quantified at levels of 0.5 to 100 ppb
in hot, wet incinerator exhaust gas, which may also contain high
particulate and acid levels. The protocol, which describes the
practices used by laboratories making these measurements, allows for
several alternative designs and operating procedures. This paper
describes an experimental program to evaluate under controlled
conditions in the laboratory as many of the acceptable practices as
possible. It describes the results of sampling at various
concentration levels using two tube configurations, two moisture
levels and other procedural variations. These results have led to
conclusions about the specific VOST procedures to be followed during
the field validation studies.
INTRODUCTION
The Code of Federal Regulation, Title 40, Part 264, requires that a
destruction and removal efficiency (DRE) of 99.99 percent be achieved
for each principal organic hazardous constituent (POHC) designated in
the Trial Burn Permit (1). The calculation of DEE requires sampling
and analysis to quantify POHCs in the waste feed material and stack
gas effluent. The “Sampling and 2 nalysis Methods for Hazardous Waste
Combustion” manual provides information on methods that are applicable
for collection and analyses of POHCs in process streams of hazardous
waste incinerator units (2). The “Protocol for the Collection and
Analysis of Volatile POHCs using yOST” (VOST Protocol) describes the
Volatile Organic Sampling Train (yOST) used to measure POHCs in the
stack gas effluent (3).
The purpose of this paper is to describe a laboratory study conducted
by EPA to evaluate the VOST Protocol. This work was undertaken prior
to the field validation to be sure that the best set of procedures
would be validated in the field effort. The project was done under
contract by PEI Associates in Cincinnati, Ohio, who developed the
specific equipment and procedures for the laboratory study. The work
was supported by the Quality Assurance Division (QAD), Environmental
Monitoring Systems Laboratory, Research Triangle Park, NC
EXPERIMENTAL
-------
Six volatile organic compounds were selected for use in the VOST
laboratory evaluation. These compounds are listed in Table I. This
table shows the compounds’ boiling points and incinerability. Two of
these compounds, carbon tetrachioride and chloroform, were included
because of their expected frequent designation as principal organic
hazardous constituents (POHC) in hazardous waste incinerator trial
burns. Benzene was included because of the effect its historically
high background levels has had on Tenax, which is the primary sample
collection medium of the yOST. Tetrachioroethylene or
perchioroethylene (commonly referred to as “perc”) was included in the
study to test recovery of a compound with a boiling point (121 C) near
the high range of the protocol (approximately 100 C).
Trichiorofluoromethane (TCFM) was selected because its low boiling
point (24 C) challenged the recovery of the yOST. The final compound
studied in this investigation was vinyl chloride (VC). VC was
included even though its boiling point (—12 C) made it unlikely to be
quantitatively recovered by the yOST. Except for TCFN, mixtures of
these compounds at the ppb level were readily available as Group 1
gases through the QAD gas cylinder audit development program. The
Q1W’s Group 1 gases were developed and certified by the National
Bureau of Standards (NBS) and, therefore, concentrations of these
mixtures could be traced to NBS for use in determining VOST recovery.
A separate cylinder containing only TCFM and nitrogen was obtained,
and the concentration of this cylinder was similarly traced to an NBS
standard.
Test gas atmospheres for the laboratory study were generated by mixing
gases from a cylinder of the five Group 1 gases (approximately 75 ppb
each component, balance nitrogen), a cylinder of approximately 75 ppb
TCFM in nitrogen, and a cylinder of ultra—pure zero air. At the
conclusion of the project, the cylinders containing these compounds
were analyzed independently by Research Triangle Institute (RTI), to
verify labeled cylinder values and the stability of the low—level
gases over the duration of the project. No values differed by more
than 10 percent. The dilution system used to mix the gases was
checked for leaks, levels of contamination and stability of operation
and found to be acceptable before any test samples were collected.
The dilution system and cylinder gases were used to generate test gas
concentrations of approximately 15 and 0.5 ppb. These two
concentrations were considered to be within the range of normal
hazardous waste incinerator POHC concentrations.
A quad (four—train) VOST was used to collect four samples
simultaneously from a manifold purged with test gas of known levels of
target compounds. Figure 1 shows a schematic of the quad VOST
equipment in the laboratory evaluation. Figure 2 shows a more
detailed description of a single yOST. Two types of VOST sample
collection tubes are specified in the VOST Protocol. Figure 3 shows
the two tube configurations in detail. The term “ST ” is used in this
report to designate the suspended tube design, which has numerous
other names since it is the tube conventionally used in ambient air
sampling with Tenax. The other design is designated “ND” for
-------
neck—down tube. The VOST Protocol defines the ND the inside/inside
tube configuration and the ST as the inside/outside tube
configuration. Since one of the goals was to determine which design
to use in field work, the quad VOST experiments were conducted with
two ST—type trains and two ND—type trains.
All sampling activities were conducted in a mobile laboratory separate
from the analysis and sorbent preparation areas.
The sorbent traps were all prepared from one lot of Tenax and one lot
of charcoal and were prepared and conditioned as specified in the
protocol. Preparation and conditioning of the traps were performed in
a separate area of the laboratory in which no solvents are handled or
stored. The exact handling of each trap during preparation and
conditioning activities was documented in a log book.
Each pair of traps was blank—checked as a unit following conditioning
and prior to use. The blanking procedure consisted of thermally
desorbing the traps, concentrating the blank sample cryogenically,
flash—vaporizing the concentrated sample, and analyzing the sample by
gas chromatography/flame ionization detection (GC/FID). Because the
GC/FID procedure gave only a general indication of blank levels, 1 of
each 10 trap pairs was analyzed as a sample by gas chromatography,’ uass
spectrometry (GC/TIS) instead of GC/FID to determine actual blank
levels of all the target compounds. Except for benzene, the blank
levels for all other compounds were less than 10 nanograms per pair of
traps.
All sorbent traps were analyzed in the same manner and in accordance
with the protocol. Each sampled trap was spiked with an internal
standard, thermally desorbed with organic—free helium gas bubbled
through organic—free water and collected on an analytical sorbent
trap. After the sample desorption steps, the analytical sorbent trap
was rapidly heated and the carrier gas flow was reversed so that the
effluent flow from the analytical trap was directed into the GC/MS.
The volatile organics were separated by temperature—programmed gas
chromatography and detected by low—resolution mass spectrometry. The
mass of volatile compounds was calculated by the internal standard
technique.
Unless specifically designated otherwise, all samples were analyzed
within one week of collection.
The front (Tenax) and back (Tenax/charcoal) sorbent traps were
analyzed separately for each sample run to determine in which portion
of the train the target compounds were actually collected and to
assess the potential for breakthrough.
RESULTS
In the course of performing the laboratory evaluation, several
parameters were studied (sorbent tube design, moisture level in the
sample gas and sample holding time).
-------
The effects of tube design and moisture level varied from one
component to the next and have been evaluated on a component basis.
The following discussion describes the individual component data
presented in Table II, which was summarized from eight quad VOST runs.
The sample means are simple averages of paired run data; standard
deviations are pooled from paired run data. The percent relative
standard deviation (RSD) values represent the pooled standard
deviations expressed as a percent of the means. The percent expected
value (EV) is the mean measured concentration expressed as a percent
of the concentration of the target compound in the sampled gas stream.
The statistical significance of the difference between the paired
measurement results for each type of sampling train was determined by
analysis of variance (ANOVA) for the dry test runs. The statistical
significance of any difference between train types for the wet runs
was determined by use of the t—test. Differences between train—types
were determined to be statistically significant; if the respective
test indicated there was less than a 10 percent probability the
difference was due to chance.
VINYL CHLORIDE
An examination of the data at each test condition indicated that the
ST sampling trains were ineffective in all cases. The maximum amount
of vinyl chloride (VC) recovered by use of the ST train was 60 percent
of the expected value for one of the wet runs at 15 ppb. Only the ND
results at the 15 ppb dry test conditions were considered reasonable,
with an average recovery of 91 percent EV. Unfortunately, subsequent
experiments could not reproduce the 91 percent EV for VC. The
presence of approximately 30 percent moisture in the test gas had a
different effect on the two types of trains. The apparent effect on
the ST results was to improve the average recovery. The effect of
moisture was definitely detrimental to the ND train results at 15 ppb
in that the average recovery was decreased from 91 to 48 percent EV.
RSD values were used as estimates of precision. The best and worst
RSD values were for the ST trains at dry test gas conditions. These
values, however, were considered nonrepresentative because of very low
recoveries. Pooled precision values for the ND train type ranged from
7 to 10 percent RSD at 15 ppb.
TRICHLOROFLUOROMETHANE
The statistical tests showed a significant difference between the
results determined with the two types of sampling trains for the high
concentration of test gas (i.e., 15 ppb) for both wet and dry test gas
conditions. In both cases the highest concentration of TCFM was
reported for the ND sampling train. At the 0.5 ppb level, the results
reported for the two types of sampling trains agreed quite well.
Examination of the TCFN data indicated that the presence of moisture
-------
in the test gas affected only the ST trains at the 15 ppb level. The
average recovery by the ST trains was 56 percent EV for the dry test
runs versus 39 percent EN for the wet runs at 15 ppb. Conversely, the
average recovery by the ND trains was approximately 75 percent EN for
both wet and dry runs at 15 ppb.
The precision of the pooled data in each of the four test conditions
ranged from 6 to 14 percent RSD for the ND trains. Precision of the
pooled ST data was much more variable, ranging from 40—60% RSD.
CHLOROFORM
The statistical analyses indicate that the difference between the
measurements made by the two types of sampling trains was significant
only for the 15 ppb wet gas. For this test condition, the average
concentration measured by the ND trains was 73.34 ng/liter compared
with 65.31 ng/liter for the ST trains, for a difference of
approximately 11 to 12 percent between the types of trains. Although
the statistical comparison indicated the probability that this
difference was due to chance was less than 5 percent, a difference of
12 percent does not seem significant when compared to overall results
including other compounds.
An examination of the data sets indicated the presence of moisture in
the test gas did not affect results obtained by the use of either type
of train.
Precision of the pooled data sets at each condition for the ND trains
varied between 4 and 23 percent RSD. Precision of the pooled data
sets at each test condition for the ST trains varied between 3 and 15
percent RSD.
CARBON TETRACHLORIDE
The statistical analyses indicated that the results reported for the
ND trains were significantly higher than those reported for the ST
trains at all of the nominal test conditions except the 15 ppb dry
gas.
An examination of the data sets indicated that the presence of
approximately 30 percent moisture by volume in the test gas had no
effect on the performance of the ND trains but definitely had an
adverse effect on the recovery of the ST trains. The average recovery
for the ST trains at 0.5 ppb was 71 percent EN dry test condition
compared with 51 percent EN wet test condition, and at the 15 ppb test
condition recoveries were 78 and 56 percent EN at dry and wet test
conditions, respectively. By comparison, the average recoveries of
the ND trains ranged from 85 to 92 percent EN for all test conditions.
Precision of the pooled data sets at each test condition for the ND
trains ranged from 1 to 28 percent, RSD.
-------
BENZENE
The statistical analyses determined that the difference between the
results reported for the two types of sampling trains was not
statistically significant at any of the four nominal test conditions.
An examination of the data at each test condition indicated that the
presence of moisture in the test gas may have had some effect on the
results measured by each type of train, but the differences between
the dry and wet data were similar to the differences between
train—types (which were determined not to be significant).
The very high recovery values obtained by both types of trains at the
0.5 ppb test conditions may be related to inherent benzene
contamination, either because of inconsistent pretest conditioning
effects or sorbent degradation after the conditioning procedure. The
overall average recovery at 0.5 ppb conditions was approximately 150
percent EV. Average recoveries at the 15 ppb test conditions were not
as high as those for the 0.5 ppb test gas and ranged froni 100 to 116
percent EV.
The precision of the test results was also worse for the low—level
tests than for the high—level tests. At 0.5 ppb, the precision of the
pooled data sets varied between 2 and 44 percent RSD. For the tests
conducted at 15 ppb, the pooled precision varied between 2 and 27
percent RSD.
T i’k ACHLOROETHYL.ENE
The statistical tests showed that there was no significant difference
between the tetrachloroethylene (perc) concentrations reported for the
two types of sampling trains.
An examination of the data at each condition indicated the presence of
moisture in the test gas did not affect the ability of either type of
sampling train to measure concentrations of perc. The average
recovery values were higher at the 15 ppb test conditions than at the
0.5 ppb test conditions, ranging from 108 to 133 percent EV and from
100 to 109 percent EV, respectively.
Precision values ranged from 1 to 46 percent RSD for the pooled data
sets at 0.5 ppb and from 3 to 25 percent RSD for the pooled data sets
at 15 ppb.
EFFECT OF SAMPLE HOLDING TIME
All samples for quad Runs Q1, Q2, and Q3 were analyzed in the
laboratory within less than one week after sample collection. Samples
from quad Run Q4 were analyzed two weeks after sample collection and
those from quad Run Q5 were analyzed five weeks after collection. All
of these samples were collected during the same week of sampling and
-------
at the same nominal test conditions, 15 ppb concentration of target
compounds in a dry gas stream. The measured concentration data were
analyzed statistically by ANOVA to determine if the difference between
results obtained at the three sample holding times was significant.
For example, the average concentration of VC for Run Q5 (5—week
holding time) was compared with the average concentration of VC for
Run Q4 (2—week holding time) and for Runs Qi, Q2, and Q3 (holding time
less than 1 week). If the difference between the concentration was
significant, the effect of holding time was significant. A summary of
the results of the analysis of the effect of holding time is presented
in Table III.
As can be seen from Table III, there was an overall tendency toward a
decrease in the reported concentration of each target compound as the
time between sample collection and sample analysis was increased. The
value obtained for any compound on either type of trap, except VC on
ST traps, was the highest when analysis was completed within 1 week.
This apparent effect of sample holding time was statistically
significant for vinyl chloride and trichlorofluoromethane, the two
lowest—boiling compounds studied.
For TCFM, a decrease in concentration of approximately 25 percent
within two weeks of collection for the ND trains indicates that
samples should be analyzed as soon as possible after collection.
Again, the ST trains were not very effective because of low recovery
(—56% EV as shown in Table II). The similarity between results at the
2— and 5—week holding times seems to indicate that any significant
loss of TCFM occurs within the first two weeks.
It should be noted that these results are for high—level (—15 ppb),
dry samples, which may be “ideal” samples. The effect of sample
holding time will also be investigated for actual field samples.
CONCLUSIONS
The laboratory evaluation indicates that the precision and accuracy of
the VOST protocol depends on the target compound and on the type of
sorbent trap used. In some cases, the precision and accuracy may also
be affected by the moisture content of the sample stream and the
amount of compound collected on the sorbent trap. The amount
collected is a function of the concentration in the sample stream, the
sampling rate, the sample volume, and the sampling time.
The most obvious conclusion of the laboratory tests is that neither
the ND— nor ST—type of sorbent trap designs produced acceptable
results for vinyl chloride.
The precision and accuracy for the ND and the ST types of sorbent
traps varied for each compound and must, therefore, be described
individually. For example, of the three compounds——chloroform, carbon
tetrachioride and tetrachioroethylene——only carbon tetrachioride
showed differences in precision and accuracy between the ND and ST
traps. For carbon tetrachloride, the ST traps also showed differences
-------
between wet and dry conditions. The ND traps showed no differences
between wet and dry conditions.
Benzene results must be considered separately because of the inherent
problem of benzene presence in the Tenax sorbent. At concentrations
of 15 or 0.5 ppb, no distinguishable effect was observed for trap type
or moisture condition, but the amount collected had a definite effect.
Average benzene recoveries were 108 and 155 percent, with
corresponding precision estimates of 16 and 35 percent, for
collections of approximately 1,000 and 50 ng, respectively. Precision
and accuracy results for benzene were not comparable to those for the
other compounds when less than approximately 50 ng were collected in
the sample.
The ND data for trichlorofluoromethane and carbon tetrachioride and
all the data for chloroform and tetrachioroethylene indicate that
average recoveries for the four compounds ranged from 75 to 121
percent. Precision (expressed as the pooled RSD of the replicate
pairs) varied between 8 and 23 percent at the two different
concentrations under both wet and dry conditions when approximately 50
ng or more of each compound was collected.
With the possible exception of benzene, the estimates of accuracy
obtained during this laboratory study (75 to 121 percent) were better
than that given in the VOST Protocol (50 to 150 percent). The
precision estimates (23 percent or better) are considered good.
Because these estimates were obtained on a clean air matrix, they
cannot be considered typical of actual test site precision and
accuracy.
In general, the results indicate that the ND traps were superior to
the ST traps. The ND traps are less complicated in design and
handling and have a positive forced flow during desorption.
Therefore, their performance should be as good as or superior to that
of the ST traps for other, untested, compounds.
A trend toward lower recoveries with prolonged holding times was
observed for all compounds included in this study. Excluding vinyl
chloride, the only compound for which the lower recovery could be
considered significant was trichlorofluoromethane, which showed
approximately a 30 percent decrease on both types of traps when
samples were analyzed 2 weeks after collection. Again, these results
were obtained on a matrix of clean, dry air; the effect of holding
time may increase under actual field conditions.
RECONMENDA TIONS
The recommended changes to the VOST Protocol have been described in
detail in the PEI Laboratory Report on the yOST. Recommendations for
minor modifications to the VOST Protocol were made in the following
areas:
Calibration procedure for the dry gas meter.
-------
• Field blank handling.
• Substitution of methanol instead of pentane for extraction
of glass wool.
Elimination of methanol extraction of charcoal.
Prescreening of charcoal prior to its use in traps.
Prescreening of each lot of Tenax.
Predrying of Tenax before it is placed in the vacuum oven.
More specific procedures for sorbent trap conditioning to
enhance consistency of blank levels, including longer
conditioning time and maintenance of positive flow during
heating and cooling.
• Specific procedures for handling conditioned traps to
minimize ambient air contamination.
More sensitive and specific calibration for blank checking.
• More specific procedures for blank checking.
Quality control for blank checking, including system blanks
and recovery checks.
• Different sample desorption flow rate.
Development of a procedure for evaluating performance of the
analytical trap in the purge and trap apparatus.
• Addition of a recovery check procedure for the analytical
system.
Development of procedures for determining acceptable analytical
separation to assist in column selection and other CC
conditions.
Development of procedures for determining appropriate
calibration levels.
Overall, the laboratory evaluation accomplished all of its stated
objectives.
-------
REFERENCES
1. Code of Federal Regulations, Title 40, Part 264 (1980).
2. Harris, J.C., et. al, “Sampling and Analysis Methods for Hazardous
Waste Combustion,” EPA 600/8—84—002, February, 1984.
3. Hansen, E.M., “Protocol for the Collection and Analysis of
Volatile POHCs Using yOST,” EPA 600/8—84—007, March, 1984.
-------
CAPS
FOR MOISTURE R1J S
— —,
- _j
TEFLOM 4-PORT
MAUI FOLD
IN USE
TUBING
DILUTION SYSTEM
WiTH GLASS
MIX 1KG CHAMBER
PURGE/LEAK
CHECK PUMP
(&TRA -PURE
ZERO AIR
—‘70 ppb
GROUP 1
Ct IPO1J4DS
‘-90 ppb
TCFM
FOUR METER CONTROl. UNiTS
Figure 1.
Schematic of sampling arrangement for laboratory evaluation.
-------
10 P hL
I0
Of C h A L
IIO ICAIOO
Lit: ( OI(I/hIAP/fLA j iuU All IN two G Ju1 t %
kCILO-O I&St 1AAfl AL g P WS,IIOt0-t I$t IW$.
FIgure 2. Schematic of single volatile organic sampling train (yOST).
T(NAX OR T(NAX/CHARCOAL TL (
(IOc* 1.6 cm GLASS) )LOS
APPROIIMATLLY 2 oramo 5006(01
SUSPENDED-TUBE (SI)
GLASS CAP
VITON 0-RING JOINT
VITON 0-RING BtTVE(N
IINAX TUBE 0.0. AND TRAP 1.0.
TINAX 00 TENAX/CHARCOAI. TUOI
(10 cm 1.6 cm GLASS) HOLDS
APPROIIMATELY 2 cams S000INT
.120 amsh STAINLESS STEEl. SCREEN
ZINC-COATED STEEL INTERNAL
RETAINING RING (C-CLIP)
VITON 0-RING
VITOR 0-RING JOINT
GLASS CAP
1 /4-tn. SWAGEIO% 3 16 -SS
INg *00 CAP
(Su tLTE* N..) PIRRULES)
NECKED-DOWN (Nu)
4 - W A !
urLon VALOE
‘lost
Li st S
00øLcTI I
10 tEAL
nil
mar
L(fI TH
—s In.
Figure 3. Sorbent trap configurations.
-------
TABLE I. LIST OF COMPOUNDS SELECTED FOR
VOST PROTOCOL VALIDATiON
Compound
Boilin 9
point , 0
°C
Inclnerabllltya
ranking
Comments
LilIC
I 99 ,gy, 2 d
Tetrachioroethylene
Benzene
Carbon tetra-
chloride
Chloroform
Trichlorofluoro—
methane
Vinyl chloride
121
80
77
61
24
—12
15
282
4
10
1
58
1
23
2
44
NL
26
Group 1 compound;e potential
recovery problem
Group 1; historical Tenax
blank problem
Group 1; frequently selected
as a POHC
Group 1; common lab solvent.
Lower limit of acceptable
boiling point range
Group 1; potential break-
through problem
A A ranking of 1 is the most difficult to incinerate.
b The general target range in the VOST Protocol is 30° to 100°C.
C Based on the heat of combustion table for 283 RCRA Appendix VIII constituents.
d Based on the temperature required to achieve 99.99 percent destruction at a
residence time of two seconds (list of 5 compounds compiled by J. J. Cudahy
of IT Enviroscience, September 1983).
e Group number refers to the QAD gas cylinder audit program.
Not listed.
-------
TABLE II. SUMMARY OF PRECISION, ACCURACY, AND TRAIN-TYPE COMPARISON RESULISa
lrfc hlnro. Ca.-hon
Vinyl flti ro- t.’r a. Tetr.chlnro.
P ara,u t.r hln. id, rth n Chlnrvfnr,i chtnr vte N.nz,ne ethy l.n.
N i$nal test —— — — ______ ———. ——— -- — — r
co nd4tinn Trap type ND ST ND ST ND ST Nfl ST ND ST ND ST
b ni —
0.5 ppb , dry - . tiq/liter 0. 0 0.17 2.77 2.37 1.Q0 1.91 2.7? 2.28 2.87 2 55 3.69 3.91
e n (liter 0.OQ 0.01 0.31 1.42 0.08 0.11 0.11 fl.?6 1.04 1.13 0.03 1.79
Y R5 I .t 5.9 13.9 55.9 1.2 5.7 5.1 11.4 36.2 11.3 0.8 45.8
I (V 42 I? 78 83 83 85 85 71 175 155 103 109
Signi f teinte
difference’ Y ,; Ito N I Ycs No
0.5 ppb, wet . ngfliter 0.53 0.36 2.74 2.53 2.02 1.84 2.71 1.63 2.48 2.01 3.79 3.55
a, ngfliter 0.13 0.09 0.17 1.0 0.12 0.27 0.l 0.32 0.75 0.03 0.15 0.40
1 850 24.5 25.0 6.? 39.5 5.9 14.7 I I 4 19.6 30.2 1.5 4.0 11.3
I (V 38 26 97 89 89 81 85 51 153 174 107 100
Si gnIficant
difference’ 9 Yes No N i Ye No No
IS ppb. dP,h , rig/liter 38.55 5.25 63.79 46.78 79.73 73.03 88.23 75.55 57.03 53.01 130.2 117.1
a. r.g/flt r 2.89 2.18 3.47 7.93 18.68 2.19 ?4.M 3.10 15.12 2.20 32.02 10.51
V ISO 7.5 11.5 6.3 11.0 23.4 3.0 21.9 4.1 26.5 4.2 24.6 9.0
¶ (V 91 12 76 56 116 106 92 78 hf ’ 101 121 lOP
Signif ica nt,
difference’ Y No Ito Ni, Nc
lSppb. wet 1 , rig/liter 70.32 19.98 64.41 33.84 73.34 65.31 86.16 53.43 52.64 49.35 143.4 139.2
ci, ng/litnr 7.10 3.71 4.55 6.62 2.89 6 12 1.13 2.39 1.10 3.45 4.16 3.51
I P50 10.3 18.6 7.1 19.6 3.9 9.4 1.3 1.5 2.1 1.0 7.9 2.5
SLY 48 47 75 39 107 95 90 56 101 100 133 129
Significant
dfff erencp 9 No Yes Yes No No
_____________ _____________ I ____________ __________ — I I _ I
lncludes data from quad Runs 01 through 03 and 06 thrnugh 08 and paired Rvns V I throuqh 518. No data are corrected for blant values.
bQu.d Runs 06 through Oil, including all data e,cnpt Run Q7, SauVle No. t.5ST11? values for all c ripounds.
CSte ,idlrd avid relative standard deviations are pciol d from paired run data.
of evoected value • (F(V) I OU. wherp IV is the averag. for apolicable runs taken from tt’ suninary table of dilution systMi det
Mid ecwected concentrations in ng/liter.
t Olff.revice btween trim types based on analysis of v,riancr at Ill’ probability he.el.
Rur” VS through 518, Including ll date The niçiir,turP roritcint of ti ’r sae ’1ed oa’. stream was ePD O tmatPty 30Y.
9 Ba’ed on t.te t statistic at 10 probability lpvpl. d ’Iferericc twtwf’rr train types
PurrS ft through flT , IrrCli,dInc all riot. r,r..pt Our 0’. S..earle Nr’. IISNF)12 1 vflI,rc fOr viryT chlnrtdt ,iirrl tri hlcniiOr h i ’.
‘raired Pu’c 511 th ’ r’uo VI. in ’ ltuiir all dat.r The ‘wrr ’ tiur— r nti’rt ni thn sarç lrr’ air. str.’r ’ war. ppro.rriatrly 30.
-------
TABLE III. SUMMARY OF SAMPLE HOLDING TIME RESULTS 8
S.p , l a
holdipg
hip
(wceksJ
Parameter
Vinyl
thlor ldr
MD ST
tr ichlc .rn-
?luororv’thane
“
MD SI
Chloroform
—- -—
Pin SI
Carbnii
tetrachloriii,
—-
Nfl SI
8 enzene
ND T
letrachinri.
ethylnn,
‘ IT t 51
Trap type
.I
7’
nq/liter
o . na/liter
S RS )
ngFliirr
ngFtIter
S 650
3P.55
2 B9
7.5
iS 23
2.11
13.9
5 75
2.16
41.5
7. 8
0 23
3.7
39
6.4?
S 03
76.3
2?
63.79
347
6.1
47.56
4. ?Q
9.0
—25
50.02
0.40
0.6
-2?
46.76
7 93
17 (1
31 36
2 75
7.2
-33
29 60
4 1’
IS 1
.37
79.73
16.61
23 4
70 7?
1) 76
I.)
-Ii
70 17
1.0?
‘ I 5
-1?
73.03
2.19
3 0
67 74
9.19
13.6
.7
66.26
i 20
7.6
-7
61.23
2 64
/1.9
75 55
0 16
0 S
-14
83.66
3.U9
311
7.55
3 10
1.1
72.33
6
1.7
-4
57 03
15.12
26.5
53.04
3.31
6.?
53 DI
2. U
4.?
47.B0
2.15
5.5
130.2
32.02
74.6
116.0
2.19
1.9
117.1
pri.5
9.0
90.5
7.50
7.5
.15
-7
—30
—II
DifferpnCe. %
-61
70.79
.3.?!
1.7
-7
50.68
1.26
2.5
-11
50.20
3.0?
2.0
126.1
29.93
16.6
)
II ?.?
32.30
11.0
59
ng/iiter
o ngl liler
I 650
10.70
5.64
52.7
Difference. 1
-72
Significant h
difference ‘ Yes Yes No No No
- - - - __— --.- --- -
611 saisites collpcted it a r, inal tarq.t compound concpntr.t Inn 0? 35 ppb in a dry qas stream. Al) data were used from Puns 0) through
05 ,vcept for Run 02 Sample No. tISNfl )2 1 values (Or vinyl chloride and tr lrhlnrn!lunro iiWthane. tin data are corrected fol blank values.
between sample collection and analysis. *1) samples were stored in rofriqerated cans containing charcoal.
tRun; 03 tMrouqh Q3.
dStandard end relative standard deviations are pooled from ’ paired run data.
Run 04.
1 Difference between ‘mean concentrations at indicated hold lnq time and initlil values
• sf —i_initial a 300
initial
9 Run 05.
k iq erpnre between holding times based on analysis of varianrr at lfl probehility
4 The effert of sautle holding ti’, is siqnificao’ for tb’ NO train but not for the SI Irain.
-------
PRACTICAL CONSIDERATIONS FOR IMPROVING SAMPLING ACCURACY AT
GROUN ’1ATER TEST WELLS
DR. DOUGLAS RICHARDSON ND DAVE MIODUSZEWSKI, P. E., GEO-RESEARCH,
WASHINGWN, D. C.
INTRODUCTION
The RCRA groundwater monitoring program requires accurate,
representative, and consistent measurement of contaminants at very low
levels. In most existing groundwater monitoring programs, the
sampling component of the monitoring process represents the greatest
existing potential gap in the quality assurance chain. This
presentation discusses ways in which field conditions, operator
variability, sampling procedure, and sampling device design may impact
sample accuracy. Particular emphasis is given to sampling problems
due to cross—well contamination, well head environmental
contamination, operator technique, inadequate equipment cleaning, and
sampling under extreme climatic conditions.
The lack of adequate attention to basic questions of sample quality
control in the past has already resulted in very serious problems of
data accuracy and validity in the nation’s newly developing
groundwater monitoring programs. Many recent EPA and private reports
suggest that problems of sample quality may be so serious within RCRA
monitoring programs that they could jeopardize the successful
implementation of many subsequent phases of the RCRA program which
will depend on the results of this monitoring data. Cl, 2, 3, 4)
That an important, costly new environmental program, such as RCRA,
should founder for lack of adequate attention to a program component
so basic and easily manageable as sample quality control is
unfortunate, for the sampling process is the least expensive step in
the whole data handling train. To do the job right costs no more than
doing it the wrong way, and in either case, the cost of achieving a
good reliable groundwater sample is insignificant when compared to
overall monitoring program costs or to other monitoring components
such as sample analysis or well construction.
Yet despite the fact that good sample quality is instrumental to the
success of RCRA groundwater monitoring efforts, there are as yet few
uniform standards or minimum State requirements to ensure the minimum
levels of accuracy and consistency these sophisticated new programs
demand.
Currently, sampling procedures in most State groundwater monitoring
programs represent a surprisingly haphazard hodgepodge of everything
from high speed production pumps engineered originally to keep
basements dry to slight modifications bf Biblical era methods designed
originally to bring water in a bucket at the end of a rope to the
surface of the ground as a means to the furtherance of livestock
raising. While these techniques were certainly adequate in the past
to their originally intended function, they are hardly appropriate for
-------
the controlled, consistent collection of groundwater samples to be
chemically analyzed reliably for parts—per—billion of volatile
organics or toxic contaminants, as is now required by today’s RCRA
groundwater monitoring programs.
Although we understand that good progress is now underway at EPA to
address the sampling issue, in the past sampling has taken a decidedly
back seat to analytical methods in EPA efforts to build quality
assurance into the groundwater monitoring process. Consequently the
situation which is now emerging in most State groundwater monitoring
programs is this: an extraordinarily expensive and elaborate set of
high—resolution analytical testing procedures are now being required
by the states, but little attention so far has been given to the basic
means by which these carefully analyzed samples are collected. Under
these circumstances, the success of the monitoring process is
jeopardized far more by poor sample quality than by poor analytical
procedure. (4)
One of the most admirable and clearly stated goals of EPA’S RCRA
monitoring program is to ensure that its sampling and analytical
procedures will provide “accurate, consistent, and comparable testing
results—year to year, facility to facility, and region to region.”
The statement of this goal immediately poses three questions for us
regarding sampling: First, is this goal now being achieved in EPA’S
groundwater monitoring programs? We believe that most people would
agree that it is not. Secondly, can this goal be achieved in the
sampling component of the monitorTi process? We believe the answer
is yes, and that it can be achieved with existing technology and
knowledge. Thirdly, if it can be achieved, then what, from a
practical point of view, can state RCRA programs do to achieve
consistency and accuracy in their sampling? In the time we have here,
we would like to discuss, from a practical viewpoint, four
interrelated areas that should receive attention if this goal is to be
met:
• Dedication of sampling equipment
• Sampling device design
Materials
• Minimizing operator variability
DEDICATION OF SAMPLING EQUIPMENT
Accuracy and consistency of groundwater sampling can be greatly
enhanced by dedication of the groundwater sampling equipment to each
well. This offers enormous benefits in terms of sample quality and
minimizes the amount of time and manpower needed to obtain unbiased
samples. By dedicating sampling equipment to each well, most of the
problems of cross well sample contamination, exposure of sampling
equipment to well head environmental contaminants, and the inadequate
cleaning of multi—well sampling equipment can be overcome. In most
-------
field condition situations, it is extremely difficult, as a practical
matter, to avoid the risk of introducing significant contamination
into a sample well when using multi—well equipment. Mud, ice, dust or
dirt are always present at the well head during the sampling process,
in the pick—up truck, and in the shop where equipment is stored. Long
lengths of wet, unwieldy tubing, cables, pumps, etc., easily pick up
contaminants from use and transport in the field. Elaborate and
expensive procedures usually prescribed for cleansing portable
sampling equipment and tubing with repeated acidic baths and rinses of
distilled water prior to use and between well sampling episodes are
frequently only perfunctorily observed, at best, under actual field
conditions. During winter sampling in the northern half of the United
States, such procedures are not only unrealistic to carry out, but the
well contamination problem is further compounded by the instant
accumulation of ice on all downwell portions of nondedicated sampling
systems.
Nondedicated monitoring approaches are also extremely vulnerable to
operator abuse and variability, as they generally require more care
and training on the part of the operator. Problems in training,
employee turnover, and lack of conscientiousness can become
significant sources of sample error with many nondedicated monitoring
approaches. Although the dedicated approach cannot overcome such
potential problems entirely, they can be greatly mitigated.
Health and safety are other good reasons for requiring the dedicated
approach for groundwater monitoring. As the operator is not regularly
handling the down—well components of the sampling equipment on a
dedicated sampling system, the operator’s exposure to potentially
hazardous down—well contaminants is significantly reduced.
SAMPLING DEVICE DESIG
As mentioned, there is a wide range of groundwater sampling devices
being used in the field at this time to collect RCRA groundwater
samples. The most commonly used are:
The bailer
The electric submersible pump
• The positive displacement gas pump
• The suction pump
The positive displacement bladder type “squeeze” pump
A short description of each sampling device and its advantages or
disadvantages follows.
Baile rs
-------
The bailer has traditionally been widely employed for well sampling.
It is simply a long cup—like cylinder with a line attached to it. The
bailer is lowered into the groundwater sampling well, water enters by
means of check valves or through an opening in the top, much like a
bucket, and then the bailer and the water it contains is manually
pulled to the surface. There are several serious risks associated
with the use of a bailer for well purging and collection of samples,
however, and the following precautions should be noted regarding
sampling with bailers:
The bailer, if not used with extreme care, can easily
disturb well contents during the sampling process. The
piston action of the bailer entering and exiting the
well creates a surging action that can stir up sediment,
move fine particles into the well, aerate the sample and!
or strip dissolved gases from the sample.
The bailer contacts the walls of the well during raising
or lowering, potentially scraping precipitated
contaminants or well casing material off the inside of
the casing and into the well or the sample.
Bailer operator technique has been demonstrated to
greatly affect sample precision. (5,6,7,8,9) Two
different trained field personnel sampling the same well
on the same day can obtain significantly different
samples, particularly when analyzing for volatile organic
constituents.
Purging four volumes from a “typical” 100 foot deep well
with 75 feet of standing water requires approximately
200 repetitions of bailer lowering, filling, raising,
and emptying, thereby multiplying the deleterious
effects of bailer use on sample quality.
The bailer lift line is very difficult to keep clean.
The tendency is to either wrap the line around one’s
arm or to coil it on the ground. When the bailer is
raised and lowered repeatedly during the lengthy
purging process, the opportunity for contamination
to be transferred to the well is extremely high.
Cleaning of the bailer lift line from one well to the
next is very difficult. (10) This problem is further
exacerbated by the braided construction of most lift
lines which encourages water to become trapped and
inaccessible to casual cleaning efforts.
The bailer exposes field personnel to contaminants
due to the intimate handling of wetted components
and splashing and dripping of bailer contents.
Care must be exercised during the transfer of a
sample from the bailer to a sample bottle. (7,9) In
particular with volatile organic constituents, the
-------
loss of significant fractions can occur if the sample
is not transferred with extreme care.
It should be a fundamental assumption that personnel
and environmental conditions will not remain constant
throughout the duration of a sampling program. Due to
the overwhelming lack of precision with the bailer, it
is a poor choice of sampling equipment when the
objective is early detection or tracking of contaminants
over an extended period of time. (5)
Electric Submersible Pumps
It is important to separate the conventional electric submersible,
well water delivery type pumps from designs built for sampling small
diameter monitoring wells. In addition to the problems regarding the
need for external power, the presence of many inappropriate materials,
and difficulty with deploying the conventional electric submersible
pump, there are some other difficulties that should be mentioned:
Conventional electric submersible pumps are limited to
4—inch or larger diameter wells that are more
expensive to install than 2—inch wells, less amenable
to hollow stem auger drilling methods, and require
substantially greater purge volumes.
Studies have shown that the turbulence and pressure
change caused by a rotating impeller can influence the
accuracy of volatile organics sampling. The work also
indicates that the tendency to strip volatiles out of
the sample is variable among pump designs. (11,12)
Appropriate flow control of a conventional electric
submersible pump is not available. Attempting to do
so by valving the discharge can result in further
turbulence and volatile organics loss, due to impeller
cavitation and to a spray nozzle type effect from the
sudden pressure drop across the throttling valve. (9)
The difficulty in cleaning these pumps is significant
as it is often impossible to fully access the internal
wetted surfaces and components of the pump in the
field. (7)
Purging at very high flow rates with electric
submersible pumps can disturb local conditions in the
aquifer and alter the quality of water in the well, due
to induced turbidity and flow from other horizons. (6)
The small diameter specialty electric submersible pump
is a recent development that still shares the basic
design problems the conventional electric submersible
has with regards to the loss of volatile organic
-------
parameters due to induced turbulence and use of the
sample for cooling the pump mechanism. Because of this
point it is not recommended that the small diameter
electric submersible be used for volatile organic
sampling. (11)
Positive Gas Displacement Pumps
Due to the use of compressed gas applied directly to the sample, the
positive gas displacement pump is not recommended for sampling. The
application of the gas will directly affect any dissolved
gas—sensitive or volatile components in the sample. (11,12,8)
Altering dissolved gases can also alter p11, which will affect low
level metals analysis.
The positive gas displacement type pump is applicable for well purging
due to the high flow rates available with it, the ability to pump
sediments without harming pump components, the simplicity of the
design, relative ease of cleaning and portability.
Suction Pumps
The suction pump is only useful to purge wells that are shallower than
25 feet It has been demonstrated that the changes of altering both
volatile organices, pH, and metals levels in a sample are high when
sampling with this type of pump. (11,13,8,9,14) Thus, use of the
suction pump should be prohibited as a sampling device.
Positive Displacement Bladder—Type Squeeze Pumps
This is the most highly regarded sampling device in the field today.
It was specifically designed to sample groundwater. Its efficient use
of compressed gas to raise water to the surface also allows it to be
used effectively as a well purging device. The gas operated squeeze
pump is also the only device tested and proven effective in both
purging and sampling.
The gas operated squeeze pump has been very highly rated as a sampling
device in nearly all recent studies for several reasons. These
include:
The squeeze pump is the most accurate way to sample for
volatile organic parameters. (11,12,8,15) This is the
direct result of a design that separates the drive gas
from the sample or purge water. As volatile organics are
among the most difficult constituents to accurately
sample for, using a method adequate for volatile
constituents ensures that all parameters of interest
are also being sampled for in the best way possible.
Its simple design allows it to be economically constructed
-------
of materials most acceptable for groundwater sampling.
Squeeze pumps constructed entirely of TeflonR are available
at modest cost.
It is possible to pump turbid waters without damaging the
pump. In contrast to other pump designs that include
delicate, close—fitting parts, most squeeze pump designs
do not include complex moving parts that can be damaged
by particles associated with turbidity.
It is possible to sample from extreme depths with the
squeeze pump. Standard squeeze pumps have been used
successfully as deep as 1,000 feet.
Adjustable low flow rates are possible with this pump,
even down to the 100 mul suggested as the optimum flow
rate during the sampling procedure by the U.S.G.S.
Filtration is easily accomplished with this pump design
by simply applying more pressure to the pump and
installing an in—line filter on the pump discharge
line. (16,9)
The squeeze pump is very portable when used with a small
compressed gas cylinder.
Since the drive air is separated from the well water,
any drive gas, including ambient air, may be used.
• The squeeze pump may be pumped dry without harm.
• The flow rate on a gas—operated squeeze pump can be
easily controlled by simply changing the drive gas
pressures. This allows easy flow control for slowly
filling sample containers.
• Because the movement of water into a squeeze pump by
the existing hydrostatic head minimizes turbulence, it
is one of the few designs that can accurately sample
from a discrete location in the well. The ability to
sample at a particular depth is valuable for sampling
many parameters; dedicating the squeeze pump to the
well is especially effective in ensuring sampling at a
discrete level.
The squeeze pump’s simplicity, accuracy, reliability,
and cost make it highly advantageous for dedication to
the well, thus also achieving dedication’s concomitant
benefits of consistency of method and elimination of
cross—well contamination. Thus, the positive—
displacement bladder—type squeeze pump represents the
best currently available technology for groundwater
sampling; its use should be encouraged for RCRA program
monitoring applications wherever possible.
-------
Construction Materials for Sampling and Purging Devices
Sufficient information is now available on material properties and
performance for groundwater monitoring to provide specific guidance on
their selection. (11,10,17,18)
The objectives of groundwater monitoring programs demand the usage of
inert materials (e.g., particularly TeflonR, and to a lesser extent,
stainless steel) for sampling equipment; their cost—effectiveness is
supported on the basis of the assurance of representative samples they
provide and their low cost relative to total monitoring program costs.
TeflonR offers the unique advantage of near universal chemical
inerthess to organics and inorganics, along with minimizing sorption
effects. Its use should be strongly encouraged.
At a very minimum, the use of demonstrated detrimental materials, such
as silicone rubber and carbon steel, should not be allowed. This same
prohibition should certainly extend to unranked materials, such as the
numerous commercial variations of plastics formulations in existence,
unless performance equivalency to acceptable materials has been
demonstrated (19).
CONCLUSION
Serious problems with accuracy and comparability of groundwater data
will result if improper sampling procedures are permitted at the
outset of the RCRI groundwater monitoring program. It is estimated
that only a very small percentage of the RCRA sites nationally have
begun the groundwater monitoring they will be required to conduct
under state and national guidelines. During the next year, a large
percentage of the remaining sites will begin their monitoring
programs. Without adequate guidance at this time from state and
federal regulatory authorities regarding sample accuracy,
standardization and quality control, large sums of money may be
pointlessly spent as very expensive groundwater programs get under way
with little consistency or quality control in the sampling process.
The rare opportunity to build meaningful, scientifically valid,
statistically comparable statewide data bases for groundwater quality
will be lost. What a shame it will be if we now undertake this
expensive new groundwater monitoring effort, only to end up with reams
of scientifically useless data that do not permit statistically
meaningful aggregation at the state or national level due to the lack
of attention to accuracy and consistency in the most basic and least
expensive part of the quality assurance chain, the sampling procedure.
When the sampling methods employed during the monitoring programs are
inadequate, whole classes of important contaminants (such as volatile
organic compounds) may go undetected. The groundwater monitoring
program thus fails in its primary purpose. Good sampling practices
-------
should be instituted at the beginning of the program or, if the
program has already started, as soon as possible. Just because bad
practices are “already under way” in a groundwater monitoring program
is no reason why such practices should be maintained throughout the
rest of the program.
-------
REFERENCES
1. Hazardous Materials Intelligence Report (HNIR). Publication of
World Information Systems, Inc. November 2, 1984.
2. FOCUS. Publication of Hazardous Materials Control Research
Institute. November, 1984.
3. Schneider, H. 1985. “Tracking Toxics: How Good is Groundwater
Monitoring?” in Research Exchange, a publication of the Energy
Research Foundation. Columbia, South Carolina, Summer, 1985.
4. McBean, E.A., F.A. Rovers. “Analyses of Variances as Determined
from Replicates versus Successive Sampling.” GWM Review, Sununer,
1985.
5. Keith, S.J., L.G. Wilson, H.R. Fitch, D.N. Esposito. 1983.
“Sources of Spatial — Temporal Variability in Groundwater Quality Data
and Methods Control.” GWMR. Winter, 1983, 3,1,21—32.
6. Wilson, L.C., J.V. Rouse. 1983. “Variations in Water Quality
During Initial Pumping of Monitoring Wells.” GWMR. Winter, 1983,
3,1,103—109.
7. Seanor, A.M., L.K. Brannaka. 1981. “Influence of Sampling
Techniques on Organic Water Quality Analyses.” Proc. of the National
Conference on Management of Uncontrolled Hazardous Waste Sites, HNCRI,
Silver Spring, Maryland, October, 1981.
8. Houghton, R.L., M.E. Berger. 1984. “Effect of Sampling Method on
Apparent Quality of Ground Water.” Proc. of the Fourth National Symp.
and Exposition on Aquifer Restoration and Ground Water Monitoring.
N *1A, Worthington, OH, May 23—25, 1984.
9. Stolzenburg, T.R., D.G. Nichols. 1984. “Preserving the Chemical
Integrity of a Ground Water Sample.” Proc. of the Fourth National
Synip. and Exposition on Aquifer Restoration and Ground Water
Monitoring. NWWA, Worthing, OH, May 23—25, 1984.
10. Nielsen, D,M. “Groundwater Sampling from Small Diameter
Monitoring Wells.” NWWA Publication, Worthington, OH.
11. Barcelona, N.J., J.A. Helfrich, E.E. Garske, J.P. Gipp. 1984. “A
Laboratory Evaluation of Groundwater Sampling Mechanisms.”
Groundwater Monitoring Review (GWMR). spring, 1984, 4,2,32—41.
12. Gillham, R.W., M.J.L. Robin, J.F. Barker, J.A. Cherry. 1983.
“Groundwater Monitoring Sample Bias.” API Publication 4367,
Environmental Affairs Department, American Petroleum Institute, June,
1983, 206 pp.
13. Ho, J.S. — Y. 1983. “Effect of Sampling Variables on Recovery of
Volatile Organics in Water.” Journal of the American Water Works
Association. 75,11,583—586.
-------
14. Unwin, J.P. 1984. “Sampling Ground Water for Volatile Organic
Compounds: The Effects of Sampling Method, Compound volatility and
Concentration.” Proc. of the Fourth National Symp. and Exposition on
Aquifer Restoration and Ground Water Monitoring. NWWA, Worthington,
OH, May 23—25, 1984.
15. Gibb, J.P., M.J. Barcelona. 1984. “Sampling for Organic
Contaminants in Groundwater.” Journal of the American Water Works
Association, May, 1984, pg. 51.
16. Fetter, C.W., Jr. 1983. “Potential Sources of Contamination in
Groundwater Monitoring.” GWMR. Spring, 1983, 3,2,60—64.
17. Barcelona, N.J., J.P. Gibb, R.A. Miller. 1983. “A Guide to the
Selection of Materials for Monitoring Well Construction and Ground
Water Sampling.” Illinois State Water Survey Contract Report #327;
USEPA—RSKERL, EPA—6001, 52—84—024, 78 pp.
18. Curran, C.M., M.B. Tomson. 1983. “Leaching of Trace Organics
into Water from Five Common Plastics.” GWMR. Summer, 1983,
3,3,68—71.
19. Nacht, S.J. 1983. “Monitoring Sampling Protocol Considerations.”
GWNR. Summer, 1983, 3,3,23—29.
-------
THE RELATIONSHIP BEIWEEN THE DESI OF WELLS AND SAMPLING IN
COMPLIANCE MONIIORING FOR GR0UN P.TER UNDER RCRA
ROY MURPHY AND MARK GILBERTSON, U. S. ENVIRONMENTAL PROTECTION AGENCY,
WASHINGTON, D.C.
INTRODUCTION
The goal of the RCRA groundwater monitoring regulations is to ensure
that owner/operators of land disposal units evaluate and monitor the
impact of their facilities on the aquifer(s) underlying their sites.
The objective of groundwater quality monitoring programs, therefore,
is to obtain samples representative of depth discrete, in—situ fluid
for analysis. A number of factors can influence whether or not a
sample is representative. Two of the most crucial elements of the
monitoring program are the design of the wells and the sampling
methodology.
WELL DESI
Two types of wells are of use in a groundwater monitoring program:
observation and monitoring wells. Observation wells are commonly
installed during the characterization phase of the site investigation
to provide geological data as well as data on water level or
potentiometric head. Observation wells can be constructed of PVC or
any other such material since the parameters to be measured are
physical, not chemical, aquifer characteristics. They are temporary
and are not to be used for regulatory sampling. Monitoring wells, on
the other hand, must be designed and constructed to satisfy the
criteria summarized in Table 1.
Table 1. Criteria for Evaluating Monitoring Well Design
Does the design allow the following tests?:
measurement of total depth;
measurement of depth to static water level;
determination of the existence of nonaqueous phase liquids
(e.g. dense and light);
discrete sampling of nonaqueous phase liquids; and
the taking of a water sample with the least amount of
agitation and loss of volatiles.
Is the well?:
constructed of materials suited for monitoring
any or all of the Appendix VIII constituents in a particular
hydrogeologic environment (e.g. non—reactive, non—sorbitive,
-------
non—releasing);
• constructed to function through the active life and post
closure monitoring period for the facility; and
• constructed in a way as to be secure from tampering or
accidental collision.
Monitoring wells are used to evaluate the chemical quality of water
within the aquifer(s). Two factors that must be considered in the
design of monitoring wells are the chemical characteristics of the
liquids in the saturated zone and the physical characteristics of the
formation (i.e. depth) being monitored.
A variety of construction materials have been used for the casings and
screens of wells including, teflon, steel (stainless, black,
galvanized), PVC, polyethylene, epoxy biphenol and polypropylene.
Many of these materials may, however, affect the quality of
groundwater samples and may not have the long—term structural
characterstics required for RCRA monitoring wells. For example, steel
casing deteriorates in corrosive environments; PVC deteriorates when
in contact with ketones, esters and aromatic hydrocarbons;
polyethylene deteriorates in contact with aromatic and halogenated
hydrocarbons; and polypropylene deteriorates in contact with oxidizing
acids, aliphatic hydrocarbons and aromatic hydrocarbons. In addition,
steel, PVC, polyethylene and polypropylene may adsorb and leach
constituents that may affect the quality of ground—water sample
quality. Presently the Agency is advocating the use of either
stainless steel or teflon as construction materials (in the saturated
zone) and for well screens. Teflon is generally inert to chemical
attack, has low leach potential and is relatively non—sorptive. Table
2 provides some insight into the adsorption characteristics of Teflon.
It is a particular good construction material for use in corrosive
situations where inorganic contaminants are of interest.
Table 2. Exposure of Teflon Resins to Solvents,
Acids and Bases
Weight Increase
Solvent
Hydrochloric acid (10%) 0.0
Nitric acid (10%) 0.0
Sulfuric acid (30%) 0.0
Sodium hydroxide (10%) 0.0
Ammonium hydroxide (10%) 0.0
Ethyl alcohol (95%) 0.0
Acetone 0.3
Toluene 0.3
Ethyl acetate 0.5
Carbon tetrachioride 0.6
Notes:
-------
• Exposure temperature 25 C (77 F)
• Exposure time 12 months
• These are essentially equilibrium test values: additional exposure
times would not increase the values significantly.
• Weight changes less than 0.2% are not considered to be
experimentally significant.
• Values are test averages only and are not for specification
purposes.
Stainless steel is also durable corrosion—resistant material. A
number of various types of stainless steel exist. The types used in
casings and screens are listed in Table 3 along with their chemical
make—up. It is particularly useful construction material in
situations where organic leachate will be encountered.
Table 3. Composition of Stainless Steels in Use for Groundwater
Monitoring
Composition , %*
Designation
or type C Mn Si Cr Ni P S Mo Al
304 0.08 2.00 1.00 18.0—20.0 8.0—10.5 0.045 0.03
316 0.08 2.00 1.00 16.0—18.0 10.0—14.0 0.045 0.03 2.0—3.0
405 0.08 1.00 1.00 11.5—14.5 —————— 0.04 0.03 0.10—0.30
410 0.15 1.00 1.00 11.5—13.0 ——————— 0.04 0.03
*Single values are maximum unless otherwise noted.
The Agency does allow for combinations of materials (composite
designs) so long as Teflon or stainless steel casing and well screen
is used in the saturated zone and the different construction materials
in the upper casing (i.e., casing above the saturated zone).
Allowable materials for use as upper casing include steel, PVC,
polyethylene and polypropylene. Where different metals are used an
evaluation, and possible provision, should be made for galvanic
reactions. Plastic pipe sections must be flush threaded or have the
ability to be connected by another mechanical method that will not
introduce contaminants such as glue or solvents into the well. All
well casing and screens must be cleaned prior to emplacement to ensure
that all oils, greases and waxes have been removed.
Once the material for the casing is chosen, the depth of the wells
becomes the determining factor for design. The preference for all
wells is roughly as follows:
Four (4) inch ID casing.
• One screen per well.
-------
• No screen to exceed ten (10) feet, except the upper
screen in the surfical aquifer which can be of
sufficient length to cover fluctuations of the
water table caused by seasonal or pumping affects.
• Lowest screens to sit on top of aquitards or
aquicludes with no gravel pack beneath the screen.
• Gravel packs not to exceed screen length by more
than two feet.
• Gravel or sand packs.
Only rough—grit, pure Bentonite, certified not to
produce a pH change in a leachate, to be used
below the water table to seal the annulus.
A Bentonite—cement mixture is to be used above the
bentonite to the surface.
PVC, black iron, or other reactive mateials may be
used in the zone above the water table. Anti—
glavanic devices required for mental composites.
• Concrete apron at surface.
Bumper guards.
Locking caps.
• AU purging and sampling is to be by dedicated
bailers, with the latter to be bottom ffilling
types.
An example of such a well is provided in Figure 1.
When a larger diameter well is preferred or when the well must be very
deep (i.e., greater than 150 ft.), six (6) inch wells may be installed
with sealed casings and dedicated, down—the—hole, positive
displacement pumps. These pumps may be used to purge the well and
sample disolved constituents. A two (2) inch ID casing will be
mounted against the inside of the six inch casing, extending through
the cap to the bottom of the screen. It is in this casing that
sampling for light phase, heavy phase and if required dissolved
constituents will be made, as well as the determination of the
existence of phases and the water—level reading. This sampling will
be made by bailers. The lower sample to be taken by bottom filling
bailer. An example of this well is provided in Figure 2.
It should be noted that in each figure the design is different from
anything previously presented by the Agency. The primary difference
is the addition of a section of casing 8”—12” in length, below the
screen, which serves as a trap, or collecting cap, for heavy phase
-------
compounds. Ideally the bottom of the screen should be placed at the
top of a less permeable horizon, if one exists. This modification is
particularly important if the facility receives dioxin compound, PCB’s
or wastes known to produce precipitates with a change in pH.
SAMPLING
There is a series of activities to be accomplished in any sampling
effort. First, a protocol for sampling of monitoring wells must be
developed and implemented which fulfills the following criteria:
allows for the removal of stagnant/standing water from the well before
sampling; does not affect the chemical quality of the sample (i.e.,
degassing, absorption); allows for depth discrete sampling (i.e.,
heavy or light immissible phases); and minimizes human error. n
example of sequence of operations that should be used when sampling a
monitoring well is provided in Table 4.
Table 4. Sequence of Operations to be Followed
When Sampling a Monitoring Well
1. Remove the locking cap, and protective cap.
2. Sample the air above the well head with HNU or OVA, record
reading if any.
3. Use an interface probe to determine the presence of any light
and/or dense phase inmiissible.
4. Use a bottom filling bailer(s) to obtain a sample of the
immissible( s).
5. Use a manometer, or acoustical sounder, to measure the static
water level and record measurement.
6. Measure the record the total depth.
7. Calculate the volume of water in the casing.
8. Determine how best to evacuate the well based on volume and
recharge characteristics. This step may be already determined
by dedicated, down—the—hole equipment.
9. Remove three (3) volumes of water from the well from the center
of the screened zone, if possible.
10. Dispose of the purge water in accordance with the RCRA
regulations and operating procedures at the facility, (away
from the well head).
The water found standing in a well prior to sampling may not be
-------
representative of in—situ ground—water quality. Therefore, the
standing water in the well must be removed so that water which is
representative of the formation can replace the standing water. The
sampling and analysis program must include detailed step—by—step
procedures for the well evacuation procedures. Prior to well
evacuation, water level measurements must be taken so the volume of
standing water in the well can be calculated. These measurements must
include depth of standing water and depth to intake bottom
measurements taken to 1/100 of a foot. Each well should have a survey
point from which its water level measurement is taken. The device to
be used for water level measurements and the procedure for water level
measurement should be specified. Before evacuating the well, the
sampling personnel must check if there is material floating on the
surface of the water. If there is, it must be sampled and if
appropriate, its thickness measured. The evacuation procedure should
ensure that all stagnant water is replaced by new formation water upon
completion of the process.
The procedure used for well evacuation depends on the yield of the
well. (The yield also is a factor in determining what type of
dedicated, down—the—hole equipment to use.) When evacuating low yield
wells, the wells should be evacuated to dryness once. As soon as the
well recovers, the plan must require that the first samples removed
are the ones to be tested for pH, oxidation reduction and/or
volatilization sensitive parameters. During long recovery times
volatilization may occur, hence, under no circumstances should these
samples be taken more than three hours into the recovery period (i.e.,
after evacuation of dryness). Whenever full recovery exceeds three
hours, the remaining samples must be extracted in order of their
volatility as soon as sufficient volume is available for a sample for
each parameter. Parameters that are not pH sensitive or subject to
loss through volatilization, (such as nonvolatile or nonreactive
organics) should be drawn last. For higher yielding wells, withdrawal
of 3 casing volumes is required. Detailed field logs of the
evacuation procedure must be kept to ensure that any equipment
malfunction or other problems are described. An example of an
allowable option of purging equipment in the case of a deep, rapidly
recharging well is shown in Figure 3. Devices used for sample
withdrawal must not alter or contaminate the sample during withdrawal.
Devices must be dedicated to a specific well or capable of being fully
disassembled and cleaned between events. Procedures for cleaning the
sampling equipment must be documented.
When used properly, the following are acceptable sampling devices for
all parameters:
• Teflon bladder pumps with adjustable flow control;
Bottom valve bailed (Teflon or stainless steels); and
Syringe bailers (Teflon or stainless steels).
Appropriate operating precautions for each type of sampling device
-------
include, but are not limited to:
• Bladder pumps must be operated in a continuous manner
so that they do not produce pulsating samples that
are aerated in the return tube or upon discharge;
• Check valves must be designed and inspected to assure
that fouling problems do not reduce delivery
capabilities or result in aeration of the sample;
• Sampling equipment (e.g. especially bailers) must
never be dropped into the well because this will
cause degassing of the water upon impact;
Bailers will be extracted from the well at a uniform
speed. (recommend boom and crank);
• The contents of a bailer must be transferred to a
sample container in a way that will minimize agitation
and aeration; and
• clean sampling equipment must not be placed directly
on the ground or other contaminated surfaces.
Detailed field logs of each sampling event must be made to ensure that
prescribed procedures were followed and that unusual events (e.g.,
slow recharge rates, malfunction of equipment, possible contamination
of samples, etc.) are noted.
When sampling for parameters are not sensitive to volatilization,
pressure differentials, or oxidation/reduction reactions, the
following devices are acceptable:
Peristaltic pumps;
Gas lift devices;
Centrifugal pumps; and
Venturi pumps;
Sampling equipment must be constructed of materials that will not
affect the quality of the sample. This applies to all down—hole
equipment including the cable to lower it into the well. Equipment
with neoprene fittings, PVC bailers, tygon tubing, silicon rubber
bladders, neoprene impellers, polyethylene, and vitron are not
acceptable. Equipment must be made of fluorocarbon polymers (i.e.
teflon) or when appropriate, 316 stainless steel, or other tt proven”
inert material. When using bailers, “Teflon” coated wire, single
strand stainless steel wire or monofilament must be used to raise and
lower the bailer. The use of braided cables, polyethylene, nylon or
cotton cords is prohibited.
-------
When dedicated equipment is not used for sampling (or well
evacuation), procedures must be included for disassembly and cleaning
of equipment before each use. If the constituents of interest are
inorganic, the first rinse must be a dilute hydrochloric acid or
nitric acid and the second and subsequent rinses must be distilled
water or deionized water. Dilute hydrochloric acid is generally
preferred to nitric acid when cleaning stainless steel because nitric
acid may oxidize stainless steel. When organics are the constituents
of concern, the equipment must be steam cleaned, then followed by a
rinse with a solvent (e.g. hexane) and finally with rinsed distilled
or deionized water. Acetone is discouraged due to its proven reaction
with Teflon.
When collecting samples for analysis, pumping rates must not exceed
100 m.illilitersAainute. Higher rates can increase the loss of
volatile constituents and can cause fluctuation in pH and pH sensitive
analytes. Sampling plans must specify sampling rates and field logs
must document sampling rates.
-------
REFERENCES
Barcelone, M.J., J.P. Gibb and R.A. Miller, “A Guide to the
Selection of Materials for Monitoring Well Construction and
Ground Water Sampling”, Illinois State Water Survey Contract
Report 327, 1983.
Data from DuPont: Bulletin T—3E, “Chemical Properties TeFlon
FEP Fluorocarbon Film” and “Teiflon in Chemical Service”.
Gilham R.W., M.J.L. Robin, J.F. Barker and J.A. Cherry,
“Ground—water Monitoring and Sample Bias” American Petroleum
Institute Publication 4367, June, 1983.
Nielsen, D.M. and G.L. leates, “A Comparison of Sampling
Mechanisms Available for Small—Diameter Ground Water Monitoring
Wells”, Ground Water Monitoring Review, Spring, 1985, pp. 83—99.
Redmond, J.D. and K.H. Miska, “The Basics of Stainless Steels”,
Chemical Engineering, October 18, 1982, pp. 79—93.
Rolston, LA., “When and How to Select Plastics”, Chemical
Engineering, October 29, 1984, pp. 70—75.
-------
W \\ -
S 1e
S \
V %
Iu •.‘ C
( ‘°
/
k t.
/ i,\\ q jr O’
/1 / ...
COD.St. G..
‘bC...C.. SD ps
— ( ‘IL”%e.t%
(. e* 4. 2.
.IS.l.A.
... . =z.....
I. • • — — — .
t — — — S •
— — — .
, • • Z. :
I — .
———
III
-ci to
4
-------
t te a.
‘l t ; r v *
bS
IkL\.
( . c\’s t c , 4
S e..\ Qr 0 1q .c. 3 qo j C... e
‘0 I
L .J )
C%e. .#% % SS&% 4*.
t
iJ\\ #o o
fc .i. ec% % &
t .Mb %%A
(o, J. O.
S A O J
e . .J
Cac.s . G.
to P%% kt.
f *# # 4 %Mq 2 .
I,.•’.ue_
‘I bt
o 2.
3 . S. .%Ar.
V
Ste, t. k
C A ‘¼’ 4jk
— $ i.. o O*e s iL. ?‘ #sir4 t.
SV %J %C c_ .s?
-------
3.
f \Ot1O 4c, JELL t E StG &,
- P M
• —
S e.\ Q i .cAe C.., 9
.) A) V ”\ ‘c
E 1 .1 : c 0 - c..
t ..i ’)
/
1
/ / O w c r k (o
IN
/
f •0 ’
- ‘ e
/
/ o e.\ o\Q_
a
5
— Ce , & Coo sc. G
S Mb ttO% %%¼E ..
f fl 4%p% ACt 2 as ’
‘ 4 •”.ut.
V p
(vL’ ‘to %i ce.q t
: : \ • t
:
..:. • : 1 ’ t %J%\
I
- — .———-— . *0 tO cM
C.
-------
SAMPLING TECHNIQUES FOR RISK MANAGEMENT: A DIOXIN CASE HISTORY
MARK HAULENBEEK, AFFILIATED ENGINEERING LABORATORIES, EDISON, NEN
JERSEY AND DR. RICHARD SPEAR AND DOUGLAS STOUT, U. S. ENVIRONMENTAL
PROTECTION AGENCY, WASHINGTON, D. C.
ABSTRACT
In late May, 1983, a site with high levels
2,3,7,8—tetrachlorodibenzo—p—dioxin (dioxin) was discovered in the
ironbound section of Newark, New Jersey. Meetings were immediately
convened among the U.S. Environmental Protection Agency (EPA), the
Centers for Disease Control (CDC), New Jersey Department of
Environmental Protection Agency (DEP), the New Jersey Department of
Health (DOH), and representatives from the New Jersey Governor’s
Science, Advisory Committee to address this issue. This body of
scientific experts developed a conceptual framework for monitoring
programs and overall data quality goals required to develop
information for effective risk management. A particular concern was
that data meet criteria established by CDCs analytical specialist for
medical decisioin—making purposes. CDC has previously developed
monitoring program quality assurance requirements from prior
experience working with dioxin problems in Missouri.
This paper will describe procedures used to develop a data base for
dioxin in the Newark area. Preliminary conceptual planning and work
plan development will be described that allowed for a phased approach
to project planning. This phased approach allowed for evaluation of
analytical results as they related to site conditions prior to
development of additional site assessment phases. Topics that will be
described include:
— Development of an environmentally bound
dioxin standard soil reference material,
— preparation of a QA data base by selec-
tively batching samples to be included
in analytical runs,
— development of quality control charts to
track various contractor laboratories,
— data statistical evaluation criteria to
determine false positive and negative
results at a 95 percent confidence level,
— evaluation of introlaboratory bias, and
— data management and data presentation
methods.
The monitoring procedures implemented on over 25 distinct sampling
phases will be described. The end result of this monitoring program
was the stabilization of on—site high level contamination areas, the
remediation of all offsite contamination areas over one part per
-------
billion and the identification and remediation of a nearby satellite
disposal site.
-------
DESIG J AND IMPLENENThTION OF SAMPLING PLANS FOR RCRA LISTING AND
DELISTING PROGRAMS
JOHN ? 1ANEY, VICE PRESIDENT, ERCO, CAMBRIDGE, MASSACHUSETFS
Although the title of this speech specified RCRA programs, in general
the content should be applicable to any sampling effort. It should
also be mentioned that the content of this talk is aimed at a level of
expertise which falls between that of the uninitiated and the
experienced sampler. The first slide presents an outline of the
topics to be discussed. These topics start with the importance of
defining the objectives of the sampling effort; the actual design of a
sampling plan including statistical, waste, and site—specific factors;
the selection of equipment; a brief discussion of quality assurance,
health and safety, and chain—of—custody considerations; and lastly a
discussion of implementation of the sampling plan and the conpositing
of samples.
One can correctly assume that the need for a sampling program would
naturally be defined prior to spending time and money on sampling.
Once the need for a sampling program is established, the next step is
to define the objectives. This step, the defining and understanding
of the objectives of a sampling program, is often not given proper
attention. In fact, it is understood that a substantial number of
unsuccessful sampling efforts fail not because of problems encountered
during sampling but because of a poor understanding of the objectives.
In summary, when management requests someone to sample a waste, this
request establishes a need for sampling but it does not establish the
objectives of a sampling effort. By investigating the reasons that
the sampling effort was requested one will be able to define the
objectives.
The primary objective for waste sampling is to obtain information
which will be used to evaluate a waste. To further define the
objectives one must question which waste is to be sampled:
The waste as generated?
The waste as prior to or after mixing with other wastes
or stabilizing agents?
The waste prior to or after aging in a drying lagoon?
The waste just prior to disposal?
Also, should the sampling effort be restricted to recently generated
waste or should it also include the sludge disposed of over the last
20 years in the onsite landfill?
Mk what parameters the waste samples are to be analyzed for.
Determine why these parameters, and not others, were chosen.
If, after having the previous questions answered, you still don’t know
why you are collecting samples, ask. There are numerous reasons to
-------
collect samples and often the sampling has to be modified to
accommodate these different reasons. This slide lists a few reasons
why sampling may be required.
If you have been chosen to collect samples, it is probably as a result
of your knowledge of the waste; your input may have a substantial
effect on the success of the sampling program. Therefore, before
collecting samples, determine what the objective are and whether the
proposed sampling plan should be modified.
In the opposite situation, where you are delegating sampling
responsibility, make sure that those performing the sampling are well
aware of the sampling objectives.
A few years ago, ENSECO’s laboratories were assisting a chemical
engineer at a latex company to eliminate a certain phthalate from his
waste. Let’s call this undesired phthalate, phthalate B. The
engineer had changed to raw materials which contained phthalate A and
therefore expected that phthalate B would disapear from his waste
stream. However, 24—hour composite samples continuously found
phthalate B present at a lower concentration. The engineer tried to
find an explanation but could riot solve his mass—balance equations.
The engineer’s dilemma was solved months later when the nightshift
started wondering why they had to collect all these darned samples.
It happened that the nightshift was run by some oldtimers who were
very cost conscious. Instead of discarding the old phthalate raw
material, they decided to save the company some money and blended the
old raw material in with the new. This was a situation where
informing the sampling team of the program objectives not only
affected the sampling plan but changed the manufacturing process.
In summary, be sure to define the objectives of a sampling program and
make sure that everybody participating in the sampling effort is aware
of them.
Once the general objectives of a sampling program are defined, then
one must concentrate on specific objectives. Specific sampling,
analytical, arid data objectives will be best defined by studying the
end—use of the generated data base.
For example, let us consider a situation where the primary objectives
are identical——that is, to determine if the concentration of barium is
less than the regulatory threshold; but the specific objectives will
vary and have a substantial effect on the sampling. To present this
situation, two figures have been excerpted from Section One of SW—846,
and lightly modified.
In this first figure limited information has indicated that the
average concentration of barium is 50 ppm. Let us assume that this is
true and that the concentration of barium is normally distributed
around 50 ppm, which is substantially different from the regulatory
threshold of 100 ppm.
-------
In this second case, historical data indicated an average
concentration of 90 ppm. The specific objective for this situation is
to discriminate between 90 and 100 ppm, while in the first case the
specific objective is to discriminate between 50 and 100 ppm. Greater
accuracy and precision will be required to discriminate between 90 and
100 ppm and this will affect the number and size of samples collected
and analyzed. Thus, specific sampling, analytical, and data
objectives must be defined along with primary objectives to ensure the
success of a sampling program.
Following the definition of the program objectives, the second step in
a sampling program can commence. The second step is the design of a
sampling plan.
A sampling plan is usually a written document which describes the
sampling and analytical tasks that will be performed to achieve the
primary and specific objectives of the program.
To ensure that the sampling plan is designed properly, it is wise to
have all aspects of the effort represented. For example, an end—user
of the data should be involved because he will be using the data to
attain the program objectives and thus will be best prepared to ensure
that these objectives are understood and incorporated into the
sampling plan.
An experienced member of the field team that will actually collect the
samples will be able to offer hands—on insight into potential problems
and solutions and, having acquired a comprehensive understanding of
the entire sampling effort during the design phase, will be better
prepared to implement the sampling plan.
Most sampling plans are actually sampling and analytical plans since
the analytical requirements for sampling, preservation, and holding
times will be factors that the sampling plan will be written around.
A sampling effort cannot succeed if an improperly collected or
preserved sample or an inadequate volume of sample is submitted to the
laboratory for chemical, physical, or biological testing. The
appropriate analyst should be consulted on these matters.
If a complex manufacturing process is being sampled, it will be
necessary to have a representative of the appropriate engineering
discipline present to optimize sampling locations and to ensure that
all waste—stream variations are accounted for.
A statistician should be available to review the sampling approach to
verify that the resulting data will be suitable for any required
statistical calculations or decisions.
Lastly, a quality assurance representative should determine the number
of blanks, duplicates, spike samples, and other steps that will be
required to document the accuracy and precision of the resulting data
base.
Preferably, at least one person will be familiar with the site to be
-------
sampled. If not, then a presampling site visit should be arranged to
acquire site—specific information. If no one is familiar with the
site and a presampling site visit cannot be arranged, then the
sampling plan must be written so that it can address contingencies
that may occur.
It is not coincidental that sampling and statistics are discussed
interchangeably. This occurs since the goal of sampling and
statistics is identical, that is, to make inferences about a parent
population based upon information contained in a sample.
Thus it is not surprising that waste sampling relies heavily upon the
highly developed science of statistics and that a sampling/analytical
effort usually contains the same elements as a statistial experiment.
The Harris pollster connects opinions from randomly chosen people
while we collect waste from randomly chosen locations or times. The
pollster analyzes the information into a usable data base;
laboratories analyze our sample and generate data. Then this unbiased
data base is used to make inferences about the entire population
which, for the Harris pollster, may be the voting population of
Yonkers, New York, or, for those involved in waste sampling, the
entire population may mean the entire contents of a hazrdous waste
landfill.
During the implementation of a sampling plan or a statistical
experiment an effort is made to maximize the possibility of making
correct inferences by obtaining samples which are representative of a
population.
The term “representative sample” is commonly used to denote a sample
which (1) has the chemical and physical properties of the population
from which it was obtained and (2) has these properties in the average
proportions that they are found in the population.
In regard to waste sampling, it should be noted that the term
“representative sample” can be misleading unless one is dealing with a
homogeneous waste from which one sample can truly represent the whole
population. In most cases it is best to consider a “representative
data base” generated by an evaluation of more than one sample, which
defines the average properties or composition of a waste.
Statisticians have developed a number of strategies, such as random or
stratified random sampling, to obtain samples which are unbiased and
collectively representative of a waste. These strategies, which are
discussed in SW—846, can help your sampling plan withstand scrutiny
for purposes of quality control or litigation. Thus, prior to
sampling, study Section One of SW—846 and consult your resident
statistician.
The sampling plan must also address other factors in addition to
statistical considerations. Regarding the waste itself, one must
consider the physical state of a waste. Can that sludge waste support
the weight of sampling personnel or will a boat or some other means of
accessing the waste be required?
-------
What is the volume of waste which has to be represented by the samples
you collect? How you approach the sampling of a 40—square—foot lagoon
as compared to a 40—acre lagoon will vary substantially in terms of
logistics and number of samples collected.
Safety and health precautions and methods of sampling and shipping
will very dramatically with the hazardous properties of the waste.
The homogeneity of waste composition and the required degree of
variance information needed to achieve program objectives will affect
the number of samples to be collected.
Facility—specific factors must be considered when designing a sampling
plan. For example, the plan will have to address the accessibility of
waste at the chosen sampling locations.
The waste generation process will have to be understood, and the
sampling plan will have to address inconsistencies in waste production
such as batch processing that may require the sampling to occur at a
specified time or at a number of different times to account for
waste—stream variation.
The sampling plan should address whether waste generated during
startup or maintenance transients should be sampled or avoided.
The sampling plan must also specify climatic and hazard conditions
that must be overcome. Field personnel have to be prepared for
excessive heat or cold as well as sampling in confined areas such as
manholes.
An obvious step in the design of a sampling plan is the choice of
sampling equipment and sample containers. The choice of this
equipment will be very dependent upon the previously described waste
and site considerations.
The analyst will necessarily plan an important role in the selection
of sampling equipment. He will be aware of interactions between
sampling equipment material and the parameters to be measured and
other factors which can affect the integrity of a sample.
By choosing the proper equipment, the analyst will be able to minimize
negative contamination, which is the loss of a parameter of interest
by volatiliation or adsorption onto container or sampling equipment.
The proper choice of equipment will also minimize positive
contamination which occurs when leaching from container walls or
particle fallout or gaseous air contaminants artificially increase the
concentration of a parameter in a waste sample.
The analyst will also assist in choosing sampling equipment that is
easily cleaned or is disposable so that cross contamination between
samples is minimized.
The chosen sampling device will also have to accommodate the
-------
collection of the required sample volume for analysis.
Deciding factors between sampling devices will be the ease of use and
the degree of hazard associated with deployment under the conditions
that will be encountered on—site.
Lastly, the cost of sampling devices and the labor costs associated
with their deployment must be considered.
Quality assurance can be briefly defined as the process for ensuring
that all data and decisions based on these data are technically sound,
statistically valid, and properly documented. Quality control
procedures are the tools employed to measure the degree to which these
quality assurance objectives are met.
Typical quality control procedures that come to mind are:
trip and field blanks to measure sample contamination during
sampling and shipment;
laboratory blanks to determine contamination during analysis;
reagent blanks to determine background levels in acids and
solvents;
spike samples to identify and characterize matrix effects;
field duplicates and spiked duplicates to determine precision;
check standards to verify calibration throughout the analysis
of a sample batch;
the use of standard reference materials to verify the accuracy
of the analytical protocol; and
surrogate and internal standards to account for sample—to—
sample variation.
In addition a good quality assurance program will be responsible for
authoring standard operating procedures which will ensure that an
essential step is not overlooked during implementation. These
standard procedures should cover a sampling effort from the definition
of objectives to the submission of samples to the laboratory, at which
time the samples will be subjected to the laboratory’s standard
operating procedures.
Safety and health must also be considered when implementing a sampling
plan. A comprehensive health and safety plan has three basic
elements: monitoring the health of field personnel, routine safety
procedures, and those procedures that one follows when an emergency
occurs.
At ENSECO, where employees routinely collect samples in the field and
are routinely exposed to chemicals in the laboratory, we have a
medical examination at the initiation of employment and annually
thereafter. This exam is performed and evaluated by a team of medical
doctors that specialize in industrial medicine. The exam consists of
those elements of a rigorous physical exam as well as a comprehensive
blood chemistry analysis to detect any bodily response to chemical
exposure.
-------
Regarding safety procedures, personnel should be instructed in the
proper use of safety equipment such as Draeger tube air samplers to
detect air contamination as well as the proper use of protective
clothing and respiratory equipment. Protocols should also be defined
regarding when safety equipment should be employed and the designation
of safe areas where facilities are available for washing, drinking,
and eating.
Even when the utmost of care is taken, an emergency situation can
occur as a result of an unanticipated explosion, electrical hazard, a
fall, or exposure to a hazardous substance. To minimize the impact of
an emergency, field personnel should be aware of basic first aid and
have immediate access to a first aid kit. Thus immediate attention to
the injured person can occur while waiting for medical assistance.
Phone numbers for both police and the nearest hospital should be
obtained prior to entering the site. Directions to the hospital
should also be obtained so that when someone suffers a minor injury
they can be taken to the hospital for treatment.
Chain—of—custody procedures document the history of samples and
maintain their legal integrity. Unless the sampling is done only for
internal use and there is no chance that someone will later want to
use the data for other purposes, then you should employ
chain—of—custody procedures. Chain—of—custody procedures can be
helpful, even when sampling is done only for internal use, since
proper chain—of—custody protocols will document pertinent sampling and
sample information.
If chain—of—custody procedures are to be followed, then you are
responsible for implementing those procedures while the sample is in
your possession.
When do you possess a sample? You possess a sample when you are
holding it, looking at it after you have been holding it, or you have
it locked or sealed in your Igloo cooler, or you have stored it in a
secured area which has limited access to authorized personnel.
Where do chain—of—custody procedures start? They start during the
design of the sampling plan. The sampling plan must accommodate the
time, materials, and labor effort that will be consumed to comply with
the documentation protocols.
A minimum chain—of—custody effort would start in the field, when field
personnel document in their field notebooks where, why, and how a
particular sample was or is to be sampled.
Next the sampler completes a sample label by describing the sample,
the sample location, the sampler, and the time and date of collection.
The sample should then be secured under lock. If a lock is not
available a seal should be attached to the sample or a cooler in such
a fashion that it will have to be broken if anyone tampers with the
sample.
-------
Chain—of—custody records must be completed, documenting sample
information and all transfers of possession from the time of sample
collection up to and including receipt in the laboratory.
Upon receipt in the laboratory, the sample manager will be responsible
for completing the chain—of—custody record with a signature, date, and
time. He will then inspect the samples for temperature, any
unauthorized tampering, and any breakage or leakage. He will record
this information and also cross—check the sample descriptions on the
chain—of—custody record with those on the individual sample labels.
All of this information will then be properly filed for further
reference. The samples will be logged into the laboratory’s sample
management system and stored in a secured area at the appropriate
temperature. The analysis of samples will also be confined to
restricted access areas.
Now following all that preparation it is time to implement the plan.
This is best started with a mental walk through the sampling plan,
starting with the preparation of equipment, to the time when samples
are received at the laboratory. This mental excursion should be in as
much detail as can be imagined. Consider the little things, because
these are most frequently overlooked.
It can be guaranteed that by employing this technique you will add at
least one more item to your equipment list and may uncover a major
oversight in the process.
Next, take a second mental excursion through your sampling plan and
try to anticipate what can go wrong. Then decide how you will solve
these problems if they do occur. Add to your equipment list the
materials necessary for solving these problems.
The equipment list should be comprehensive and leave nothing to
memory. The categories of materials that should be considered are:
personnel equipment, which will include boots, raingear,
disposable coveralls, face masks and cartridges, gloves,
etc;
safety equipment such as portable eyewash stations and a
first aid kit;
and field test equipment such as pH meters and Draeger
tube samplers.
Regarding the containers, have an ample supply to address the fact
that once in the field you may want to sample 50% more samples than
originally planned, or you may want to collect a liquid sample while
the sampling plan had only specified solids.
Include not only shipping equipment you plan to use, but additional
sampling equipment that may be useful if a problem arises. Also
include a tool kit.
Shipping and office supplies include items such as tape, labels,
-------
shipping forms, chain—of—custody forms and seals, field notebooks,
random number tables, scissors, pens, etc.
After weeks of preparation, the team finally arrives in the field.
After towing a boat and trailer 200 miles the sampling team finds that
the boat is not necessary since the lagoon to be sampled doesn’t
contain 2 feet of water as they were led to believe. Worse yet, the
lagoon sludge which is to be sampled cannot support the weight of the
sampling team.
What can be done? Going home is one alternative but somebody will
probably send them back the following week. However, going home and
coming back the next week may actually be the optimum solution, if in
talking to the facility operators they discover that next week is when
the facility plans to excavate the lagoon contents. In this case the
sampling team could return the following week and randomly sample
buckets of the sludge as it is transferred to trucks.
If the facility is not going to excavate the lagoon, there may be
other alternatives. For example, if in touring the facility a crane
with a clam shell bucket is discovered, the crane could possibly be
employed to collect samples at random locations throughout the lagoon.
If the sampling team determines that the lagoon is not going to be
excavated and a crane is not located, another alternative is to
randomly sample the perimeter of the lagoon.
This latter approach may actually allow one to meet the objectives,
but what you cannot say is that the lagoon has been sampled. The
sampling team only randomly sampled that part of the lagoon which is
adjacent to the perimeter and the resulting data base is only
representative of that sludge. However, if one knows how the lagoon
was filled and the properties of the waste as it was disposed you can
possibly make some assumptions about the remainder of the sludge.
Regarding the compositing of samples, ENSECO has found it very useful
and cost effective to collect component samples in the field and
composite aliquots of each sample later in the laboratory. Then after
reviewing the data, if any questions arise you can recomposite the
samples in a different combination, or analyze each component sample
separately to better determine the variation of waste composition over
time and space or to better determine the precision of an average
number.
-------
SLIDE #1
Definition of Objectives
Sampling Plan Considerations
— Statistical
— Waste
— Site
— Equipment
- QA/QC
— Health and Safety
— Chain of Custody
Sampling Plan Implementation
Compositing
SLIDE *2
F DEFINITION OF OBJECTIVES
Need - Objectives
SLIDE #3
WASTE — As Generated?
— Prior to Mixing?
— Prior to Aging?
— As Disposed?
— Recent vs. Historical?
SLIDE *4
PARAMETERS —
Why?
—
Why Not Others?
-------
SLIDE #5
SLIDE #6
r
Latex + Phthalate A = Latex + Phthalate A + Phthalate B
Regulatory
Threshold (RT)
25 50 75 100
Concentration of Barium (ppm)
Distance of true value from regulatory threshold
requires less accuracy dnd precision.
WHY — Delisting Petition
— Monitoring
— Characterization
— Litigation
SLIDE #7
a,
>
0
U
C
a,
Confidence
iflIeIvdI
Sample Mean True Mean
(U L)
0.1
-------
SLIDE 8
SLIDE #9
SLIDE #10
0.4.
0.3.
c 0.2
C .
0.1
Sample Mean True Medn
Upper
Limit
(U LI
95 100
Reguldtory
Threshold (AT)
Concentration of Barium (ppm)
Proximity of true value from regulatory threshold
requires more accuracy and precision.
80 85 90
Designed by:
SAMPLING PLAN: DOCUMENTATION OF OBJECTIVES AND SAMPLING TASKS
• End—User of Data
• Field Team Member
• Analytical Chemist
• Engineering Representative
• Statistician
• Quality Assurance Representative
STATISTICAL/SAMPLING CONSIDERATIONS
• Data/Waste Collection
• Data/Waste Analysis
• Inferences
-------
SLIDE #11
REPRESENTATIVE SAMPLE/REPRESENTATIVE DATA BASE
• Chemical and Physical Properties
• Same Proportions
SLIDE #12
WASTE CONSIDERATIONS
• Physical State
• Volume
• Toxicity, Ignitability,
Corrosivity, and Reactivity
• Homogeneity Over Time and Space
SLIDE #13
FACILITY/SITE CONSIDERATIONS
• Accessibility
• Type of Waste Generation
• Transitory Events
• Climate
• Hazards
SLIDE #14
SELECTION OF SAMPLING EQUIPMENT
• Negative/Positive/
Cross Contamination
• Sample Size
• Ease of Use
• Safety
• Cost
-------
SLIDE $15
QUALITY ASSURANCE
• Technically Sound
• Stastically Valid
• Properly Documented
SLIDE #16
QUALITY CONTROL PROCEDURES
• Trip/Field/Lab/Reagents Blanks
• Spiked Samples
• Field Duplicates and Spike Duplicates
• Check Standards
• Standard Reference Materials
• Surrogate and Internal Standards
SLIDE #17
STANDARD OPERATING PROCEDURES
• Definition of Ob]ectives
• Design of Sampling Plan
• Preparation of Containers and Equipment
• Maintenance, Calibration, and Cleaning
of Field Equipment
• Sampling Preservation, Packaging, and
Shipping
• Chain—of—Custody
SLIDE #18
ELEMENTS OF A SAFETY AND HEALTH PLAN
• Health Monitoring
• Safety Procedures
• Emergency Procedures
-------
SLIDE *19
SLIDE *20
SLIDE * i
SLIDE *22
ANNUAL MEDICAL EXAMINATION
• Medical History
• Standard Physical Exam
• Pulmonary Functions Screening
• Chest X—Rays
• ERG
• Urinalysis and Blood Chemistry
SAFETY PROCEDURES
• Instruction
• Clothing
• Eye and Respiratory Protection
• Protocol
EMERGENCIES
• First Aid Kit
• First Aid Training
• Emergency Telephone Numbers
• Directions to Hospital
CHAIN OF CUSTODY
• In Your Posses-
sion
• In Your View
• Locked Up
• In Restricted
Area
-------
SLIDE_ *23
CHAIN—OF-CUSTODY SAMPLING PROCEDURES
• Field Notebooks
• Sample Labels
• Seals/Locks
• Chain—of—Custody Forms
SLIDE #24
CHAIN—OP-CUSTODY LABORATORY PROCEDURES
• Receipt and Inspection
• Maintenance and Information
• Sample Logging
• Storage
SLIDE #25
SAMPLING PLAN IMPLEMENTATION
Mental Excursion
SLIDE #26
EQUIPMENT LIST
• Personnel Equipment
• Safety Equipment
• On—Site Test Equipment
• Sample Container
• Sampling Apparatus
• Shipping and Office Supplies
-------
SLIDE $27
CONTINGENCI ES
• Go Rome
• Employ On—Site Equipment
• Perimeter Sample
SLIDE $28
COMPOSITING
• Minimize in the Field
• Maximize in the Lab
-------
PRACTICAL STATISTICAL CONSIDERATIONS IN DESIGNING A SAMPLING PLJN
JOHN WARREN, OFFICE OF POLICY, PL1 NNING, AND EVALUATION, U. S.
ENVIRONMENTAL PROTECTION AGENCY, WASHINGWN, D. C.
INTRODUCTION
The intent of this paper is to show how differing variabilities at
various stages of collecting and analyzing groundwater data can result
in a variety of sampling plans. Slides will be used throughout to help
the audience visualize the points being made.
PROGRAM STAND RD5
Both standards and interim status standards are found in 40 CFR
264—265:
40 CFR 264 Standards for Owners and Operators of Hazardous
Waste Treatment, Storage, and Disposal Facilities
40 CFR 265 Interim Status Standards for Owners and Operators
of Hazardous Waste Treatment, Storage, and Disposal
Facilities
Sampling and statistics appear in subsections:
264.97 General Ground Water Monitoring Requirements
264.98 Detection Monitoring Program
264.99 Compliance Monitoring Program
264.97 Instructs owner/operator to:
Install sufficient wells to characterize the
background and flow through the site,
Document a quality assurance program,
Perform a minimum sampling requirement,
Use a particular statistical test,
Or use an equivalent statistical procedure
(with permission of the Regional Administrator).
264.98 Instructs the owner/operator to:
Determine if a statistically significant change
-------
in ground water parameter has occurred.
• Follow the requirements of 264.97.
• Resample the wells if a significant change detected.
• Go to 264.99 if still significant.
264.99 Instructs the owner/operator to:
• Determine if certain concentration limits have been
exceeded.
• Follow the requirements of 264.97.
Resainple the wells if a significant difference between
sample and concentration limit is found.
Go to a Corrective Action Program if still significant
difference.
But What’s the PROBLEM ?
Misclassifying a site in either the detection monitoring program or
the compliance monitoring program.
False Positive
The site is classified as leaking when in reality it is secure. This
is a key concern of owner/operators as further testing to refute the
false positive result is expensive and time—consuming. There is also
the community—relations problem of disabusing the public of fears of
health hazards from a contaminating site.
False Negative
The site is classified as not leaking when in reality it is leaking.
This is the key concern orE1 e EPA and the environmentalists.
n ideal test procedure (statistical or otherwise) would have zero
false positive/false negative rates. In a less—than—ideal world, a
balance has to be struck between the concern of owner/operators —
false positives, and the concern of environmentalists — false
negatives.
Sources of Variation in Ground Water Monitoring
Spatial
• Temporal
-------
• Well Construction
• Sample Collection and Chain—of—Custody
• Analytical
Spatial Variation
• Geologic properties of the site.
Hydraulic flow of ground water.
Proximity of site to lakes or rivers.
Proximity of site to other waste sites.
• Multiplicity of aquifers at the site.
Temporal Variation
Seasonal effects.
• Long—term cycles and trends.
• Floods, rainstorms, and irrigation effects.
Temperature differences.
Well Construction
• Contamination from drilling.
Variation in casings (teflon, stainless steel, etc.).
Screen construction.
Sample Collection and Chain—of—Custody
Purge and recharge of well—water.
Rate of removal.
• On—site handling and storing of sample.
Field quality assurance.
Analytical Variation
-------
• Method of analysis.
• Laboratory contamination.
• Operators/analyst experience.
• Calibration.
• Minimum detection/quantification limits.
• Data reporting requirements.
• Intra/inter laboratory differences.
How Can Variability Be Reduced ?
Spatial variability — Use multiple wells
* High cost of construction.
Temporal variability — Take readings over time.
* Difficult as historical records
may not exist.
• Well construction — Ensure similarity in well
characteristics.
* Relatively inexpensive, but needs
planning.
Sample Collection — Ensure consistency in quality
assurance plans.
* Inexpensive, but needs planning.
Analytical — Ensure a rigorous laboratory
quality assurance plan.
* Inexpensive, but needs planning.
Short—Term Improvement in Data Quality Reducing Variabilities
Ensure all wells have the same physical characteristics (or
as close as possible the same).
• Consistently collect and handle samples in the same fashion
(i.e., ensure a good field quality assurance).
-------
• Make sure the laboratory has a rigorous quality assurance
plan.
Try to use the same laboratory for all the analyses.
Ask the laboratories to use the Minimum Detection Limit
procedure as outlined by the Mierican Chemical Society (1980).
Check the performance of the laboratory by sending identical
samples to different laboratories and comparing the results
for consistency.
Plot the data on a graph and visually look for trends in the
data.
Background Readings of Barium
(Micrograms Per Liter)
Contractor A
January 84 600 January 85 530
590 480
650 470
620 450
April 84 300 April 85 420
400 400
300 470
390 480
July 84 400 July 85 450
500 480
380 430
450 460
October 84 560 October 85 600
540 580
500 540
540 570
Illustration of How Differing Variables Affect the Characteristization
of Groundwater
For illustration, suppose 32 readings are taken on a site by two
diffe rent contractors:
Contractor A: Takes four samples every 12 weeks from a background
-------
well for 2 years.
Contractor B: Takes two samples every 6 weeks from a background well
for 2 years.
The problem an owner/operator must face is which set of readings
characterizes the background ground water flow best (recall the
requirements under 264.97).
(It is understood the Contractor B will cost more than A due to the
extra expense of visiting a site to collect samples not being
completely off—set by the increased expense of more sample analyses
per visit.)
Long—Term Improvements in Data Quality Reducing Variabilities
• Increase the number of background and monitoring wells (and
therefore reduce spatial variability by better characterizing
the flow of groundwaters).
• Increase the frequency of sampling (and therefore reduce
temporal variability).
Improve laboratory effectiveness by sending multiple samples to
laboratories (this will reduce the inter—laboratory variability).
Improve analytical precision and accuracy by demanding the
laboratories’ use of the Agency’s Data Quality Objective Program
(and so reduce the intra—laboratory variability).
Background Readings of Barium
(Micrograms Per Liter)
Contractor B
January 84 610 January 85 490
630 520
March 84 430 February 85 590
390 620
1 pril 84 370 April 85 400
310 470
May 84 540 June 85 530
570 510
-------
July 84 420 July 85 460
430 480
September 84 480 August 85 510
490 520
October 84 540 October 85 540
500 610
November 84 600 December 85 460
630 440
Comparing the Two Contractors
Contractor A gives more information on laboratory variability than
Contractor B (four readings per sampling date as opposed to two
readings).
• Contractor B gives more information on temporal variability than
Contractor A (sixteen sampling dates as opposed to eight sampling
dates).
The “ Trade—Off ”
• The laboratory variability can be estimated with 24 degrees of
freedom (a statistical term) for Contractor A, and with 16 degrees
of freedom for Contractor B.
• The temporal variability can be established with 7 degrees of
freedom for Contractor A but 15 degrees of freedom for Contractor
B.
Statisticians can show that when estimating variabilities; 24 degrees
of freedom do not yield that much better answers than 16 degrees of
freedom, but 15 degrees of freedom is much more superior than 7
degrees of freedom. The guiding light is that the more variable
something is, the more degrees of freedom it needs to “control it.”
IN SUMMARY
In purely terms of variability, the probable greatest—to—smallest
ordering would be:
• Temporal
Spatial
-------
• Analytical
• Sample Collection and Chain—of—Custody
• Well Construction
Moral : Owner/operators should concentrate on “reducing” variability
by designing sampling plans with the help of a statistician!
-------
FIFTH SESSION
QUALITY ASSURANCE ISSUES
8:00 am — Noon
Friday, July 26, 1985
Chairperson: Florence Richardson
Methods Program
Off ice of Solid Waste
U. S. Environmental
Protection Agency
Washington, D. C.
-------
DOCUMENTING THE EQUIVALENCY OF PROPOSED METHODS W APPROVED TEST
METHODS FOR EVALUATING SOLID WASTE
Dr. L. R. WILLIAMS, U.S. ENVIRONMENTAL PROTECTION AGENCY—EMSL, LAS
VEGAS, NEVADA
ABSTRACT
The U.S. EPA, in its quest for environmental data of known high
quality, has proposed formal processes for validating standardized
measurement methods and for demonstrating the equivalency of
alternative procedures. This paper addressed the method equivalency
process.
Just as method validation procedures are driven by the data quality
objectives of the client office (e.g., the Office of Solid Waste), so
too are the procedures for demonstrating the equivalency of methods
proposed as alternative to those methods currently in “Test Methods
for Evaluating Solid Waste” (SW—846). Several factors govern the
degree of complexity and formalism that an equivalency testing program
shall take. These depend upon the purpose and extent of the intended
use of the method and the amount of existing information that supports
the petitioner’s claim for equivalency. Appropriate experimental
designs and statistical prpocedures are discussed. Minimum data
requirements for establishing equivalency are presented and discussed.
-------
USE OF PERFORMANCE BASED QUALITY CONTROL CRITERIA IN ThE SUPERFUND
CONTRACT LABORATORY PROGRAM
GABETR PEARSON, U. S. ENVIRONMENTAL PROTECTION AGENCY-EMSL, LAS VEGAS,
NEVADA; AND MICHAEL T. HOMSHER AND FOREST C. GARNER, LOCKHEED
ENGINEERING AND MANAGEMENT SERVICES COMPANY, INC., LAS VEGAS, NEVADA
ABSTRACT
Performance based quality control criteria are quality control
criteria developed from the data generated through the actual use of a
specific analytical protocol under actual operational conditions.
When applied to contract analytical laboratories providing analytical
services to the usepa’s superfund contract laboratory program (CLP),
these criteria can be used to assess data quality and evaluate
laboratory and method performance. The rationale, development,
calculation, and use of performance based quality control criteria in
the cip will be presented. A detailed discussion of the development
and use of performance based surrogate recovery criteria for the GC ..41S
analyses of volatile and semivolatile organic compounds in water and
soil/sediment will be presented. The benefits and application to
additional quality control criteria will be discussed.
-------
A NATIONAL VOLUNTARY LABORATORY ACCREDITATION PROGRAM FOR
ENVIRONMENTAL MEASUREMENTS
P r R S. UNGER, NATIONAL BUREAU OF STANDARDS, U. S. DEAPRTMENT OF
COMMERCE, GMTHERSBURG, MARYLAND
ABSTRACT
The State of Indiana Environmental Management Board requested that the
National Bureau of Standards (NBS) establish under its National
Voluntary Laboratory Accreditation Program (NVLAP) a laboratory
accreditation program (LAP) for hazardous waste analysis in accordance
with methods approved by the U.S. Environmental Protection Agency
(EPA). Under the NVLAP procedures, NBS is requesting public comments
on the need for a LAP which may eventually include the whole range of
environmental measurements. A decision will be made later in 1985 on
whether to proceed with development of a LAP.
This paper provides an overview of NVLAP, summarizes the comments on
the State of Indiana request, reviews related accreditation efforts in
the environmental testing field, and indicates the role that NVLAP
might play.
INTRODUCTION
The Department of Commerce, National Bureau of Standards (NBS)
administers the National Voluntary Laboratory Accreditation Program
(NVIIAP). NVLAP was established in 1976 to serve government and
industry needs for evaluating laboratories and accrediting those found
competent to perform specific test methods in certain fields of
testing.
NVLAP’s goals are to:
• Provide national recognition for competent laboratories;
• Provide laboratory management with a quality assurance
check;
Identify competent laboratories for laboratory users; and
• Provide laboratories with guidance from technical experts to
improve their performance.
NVLAP is comprised of a series of laboratory accreditation programs
(LAPs) established based on requests and demonstrated need. More than
160 laboratories are currently participating in the nine LAPs that
have been established. Although NVLAP has primarily been involved
with narrowly—focused product testing laboratory accreditation, NBS is
able to establish broadly—defined LAPs. NBS is not restricted to a
-------
product—by—product approach in the normal definition of that term.
The NVLAP procedures are flexible enough to adjust to government and
industry needs as they may arise. The acoustical testing services LAP
and the commercial products LAP, which are established and
operational, are two examples of broadly—defined LAPs. For the topic
at hand, the LAP could cover the full range of environmental testing
services.
If a LAP is broadly defined, this does not mean that we will offer a
broad—brush, general, or subjective accreditation. We insist on
having a credible, quality program based on documentation and
accreditation to established standards and test methods. Each
laboratory is accredited for its capability to perform one or more
specific test methods offered within each LAP. This approach is the
fundamental basis of our bilateral mutual recognition agreements with
three foreign national laboratory accreditation systems.
NVLAP accreditation is granted based on conformance with generic
criteria published in the Code of Federal Regulations as part of the
NVLAP procedures (15 CFR Part 7). The criteria are similar to those
found in the International Organization for Standardization (ISO)
Guide 25: “General Requirements for the Technical Competence of
Testing Laboratories.” The criteria address a laboratory’s quality
system, staff, facilities and equipment, calibration procedures,
testing procedures, records, and test reports. Specific technical
requirements are developed for each LAP to address the particular
testing technologies involved. Participation in workshops to develop
specific LAP requirements is open to all interested parties.
NVLAP ACCREDITATION PROCESS
Accreditation is granted following successful completion of a process
which includes submission of an application and payment of fees by the
laboratory, on—site assessments (similar to laboratory system audits
conducted by the U.S. Environmental Protection Agency (EPA)),
successful participation in proficiency testing (similar to
performance audits administered by EPA), resolution of any identified
deficiencies, and periodic technical evaluation and administrative
review for initial and renewed accreditation.
For a more detailed description of the NVLAP accreditation process,
see Appendix I.
STATE OF INDtANA REQUEST
Under the NVLAP procedures, a LAP can only be established in response
to a formal written request and based on demonstrated need. In a
letter of August 24, 1984, the State of Indiana Environmental
Management Board requested that NBS establish a LAP for hazardous
waste analysis. After informally consulting with EPA personnel, NBS
decided to publish the request for public comment. The notice
appeared in the Federal Register on February 20, 1985. EPA requested
-------
an extension of the 60—day comment period beyond the date of the Solid
Waste Testing and Quality Assurance Symposium (July 24—26, 1985) to
give symposium participants the opportunity to express and submit
their views. Accordingly, NBS deferred its decision regarding the
need for and development of a LAP until after the symposium.
SUMMARY OF COMMENTS RECEIVED ‘10 DATE
A total of 33 letters (representing 29 organizations) commenting on
the need for a LAP were received as of July 9, 1985. Of the 29
organizational comments, 24 are positive, four are conditionally
positive, and one is negative. The positive respondents cite the role
the LAP could have in improving the consistency of the testing and
resultant test data, integrating the various efforts of the states,
and supporting regulatory decisions. The conditionally positive
respondents agree with the need for a LAP, but have reservations or
suggest alternatives. The one negative respondent states that current
efforts by EPA and the states are in place and work satisfactorily.
Table 1 shows the number of comments by type of organization:
Conditionally
Type of Organization Positive Positive Negative
Prof ./rrade Association 4 3
Testing Laboratories 5 1 1
Private Industry 6
State Government 9
Total 24 4 1
Of particular significance, the nine states that responded all support
establishment of a LAP. Seventeen of the total number of respondents
indicated that there is a need for an accreditation program to cover
all environmental measurements. An identification of the 29
organizations and excerpts from their comment letters are set out in
Appendix II.
RELATED EFFORTS CONCERNING ENVIRONMENTAL LABORAWRY ACCREDITATION
Approving or accrediting laboratories to perform measurements related
to environmental protection is not a new issue. Over the past decade,
questions of both a general and specific nature have been raised,
though they are not limited to the hazardous waste area. In 1975, the
American Public Health Association (APHA) issued a policy statement
expressing concern over the need for the establishment of standards of
performance for environmental laboratories. The policy statement
discussed the various efforts being made to evaluate environmental
laboratories and concluded by saying:
The efforts underway, comparable though they may be, differ
-------
in specific approach, in definition of environmental
laboratories, especially in types of laboratories included,
in voluntary versus compulsory regulation, and in regulatory
agency, that is, professional society or state and federal
government. If a non—integrated approach continues, a
chaotic condition will no doubt result. The history of all
regulatory actions emphasizes the need for a uniform
solution to a common problem.
In the view of some, this prognosis has come to pass.
APHh sponsored a National Conference on the Establishment of Standards
of Performance for Environmental Laboratories on May 1—3, 1978. The
conference participants “overwhelmingly voted approval of the 10
recommendations to establish a mandatory system.” Excerpts from four
of the recommendations are noteworthy:
1. Standards should be developed as a national level in
order to provide uniformity and to eliminate duplication
of standards and inspection.
2. A single federal agency should be responsible for
standards for all environmental laboratories...
4. Implementation and administration of national standards
should be carried out at the state level...
9. The administration of standards should be carried out
in a manner which eliminates duplication of activity
and provides reciprocity among all the states.
The first official EPA recognition that approval must be given to
laboratories to ensure valid results for compliance testing came with
the National Interim Primary Drinking Water Regulations, proposed in
1975. These regulations require laboratories testing for compliance
under the Safe Drinking Water Act be performed by laboratories
approved either by EPA or those states with primary enforcement
responsibility (primacy states). The EPA program is described in the
“Manual for the Certification of Laboratories Analyzing Drinking
Water,” which was first published in May, 1978. This manual serves as
a guide to the states for implementing “laboratory certification
programs” for drinking water. There are no comparable EPA manuals for
other types of environmental testing. The criteria and critical
elements for drinking water laboratory certification in the EPA manual
are similar to what NBS prepares when developing a LAP under NVLAP.
Similar documentation is needed for hazardous waste analyses and other
types of environmental testing.
Several states have established water laboratory approval programs
as early as 1950 in the case of the State of California) and have
established or are developing programs for other types of
environmental testing. A recent example is the rule proposed by the
State of Wisconsin Department of Natural Resources, currently
undergoing public comment, which will establish a “laboratory
-------
certification/registration program...applicable to commercial,
municipal, and industrial laboratories that analyze wastewater,
groundwater, surface water, hazardous substances, hazardous waste and
sediments.” The comment period on the proposed rule is scheduled to
end on July 29, 1985.
The status of reciprocal recognition agreements among the states for
their programs has not been ascertained by this author, but it is a
major issue. The American Council of Independent Laboratories (ACIL)
has indicated that many commercial laboratories need approval from
many of the states in which they offer their services. Each state
program can require its own formal application, fee payments,
proficiency testing, and on—site assessments. ACIL has indicated that
these multiple accreditation programs are a costly and unnecessary
duplication of effort and that a credible national system is needed
that can be accepted by the states in lieu of their own systems.
Two organizations submitted comments to NBS identifying themselves as
alternative laboratory accreditation systems that can operate on the
national level. The National Sanitation Foundation (NSF) indicated
that it is implementing a drinking water laboratory accreditation
program and plans to develop programs in waste water and solid waste.
NSF submitted documentation to NBS on its program procedures and
requirements for drinking water laboratories. NSF staff has indicated
that the necessary documentation for hazardous waste laboratories has
not been developed yet, but have indicated an intent to do so.
The American Association for Laboratory Accreditation (AALA) indicated
that it has accredited laboratories for water quality testing and is
available to offer accreditation in other areas of environmental
testing. However, AALA did not respond to an NBS request for written
documentation on its accreditation program for environmental testing
laboratories so we are unable to assess how its program might meet the
need.
HCXJ A LAP MIGHT OPERATE
NVLAP staff has had several informal discussions with EPA officials
about the State of Indiana request and how best to proceed. A concept
for a joint EPA/NBS effort to develop and administer a LAP is set out
in Table 2. The estimated annual costs for operating a LAP, based on
NBS experience, range between $2,300 to $3,400 per laboratory. Table
2 shows:
I. EPA/NBS Joint Role
Develop technical requirements for accreditation to
ensure that they are consistent with environmental
laws and regulations.
• Determine desired frequency of on—site assessments
(system audits).
Design assessment protocols, checklists, etc.
-------
• Recruit, select, and orient technical experts to
perform assessment and evaluation functions and
schedule assessments.
• Determine frequency of proficiency testing (performance
audits) and pass/fail criteria.
II. EPA Role
• Administer proficiency testing (performance audits).
Evaluate assessment (system audits) reports and
proficiency test results.
Formulate accreditation recommendations.
III. NBS Role
Prepare announcements about the LAP and distribute
application forms, handbooks, and other information
about the LAP.
• Receive and process applications and fees from
participants.
Disburse funds to cover costs of assessment, proficiency
testing, evaluation and other administrative costs of
the LAP.
Receive recommendations and issue accreditation
certificates.
• Publish directories of accreditated laboratories and
supplements.
CONCLUSION
Based on the comments on the State of Indiana request received so far,
there appears to be a need for a national laboratory accreditation
system for hazardous waste analysis. There is also an indication that
such a program should cover all types of environmental testing.
Whether NBS proceeds to fill this need by developing a LAP will depend
on a final analysis of comments received as well as the position that
EPA takes on the matter. If appropriate, NBS is willing to enter into
an interagency agreement with EPA or make suitable arrangements with
the states which are interested in using NVLAP for their own programs.
NBS is prepared to host public workshops to develop the technical
requirements for such a LAP.
The integrated approach to solve the problem of recognizing competent
environmental laboratories, suggested by the American Public Health
Association 10 years ago, has not been realized. Some would argue
that ten years later we are no further along to a “uniform solution to
a common problem” and that we have a “chaotic condition” as a result.
We cannot fully support this view. Our country is a pluralistic
society which enables it to create several solutions to common
problems. We believe that NVLAP can be one of these solutions.
-------
REFERENCES
Procedures for the National Voluntary Laboratory Accreditation
Program, Code of Federal Regulations , Title 15, Part 7.
Request for Comments on Need for Establishing a Laboratory
Accreditation Program, Federal Register , Vol. 50, No. 34, February 20,
1985, pp. 7095—7096.
Letter from the U.S. Environmental Protection Agency signed by Bernard
D. Goldstein, Assistant Administrator for Research and Development,
and Jack W. McGraw, Acting Assistant Administrator for Solid Waste and
Emergency Response, to Stanley I. Warshaw, Director, NBS Office of
Product Standards Policy, April 12, 1985.
Summary of the National Conference on the Establishment of Standards
of Performance for Environmental Laboratories, Williamsburg, Virginia,
sponsored by the American Public Health Association, May 1—3, 1978.
Manual for the Certification of Laboratories Analyzing Drinking Water,
Criteria and Procedures Quality Assurance, U.S. Environmental
Protection Agency, October, 1982, EPA—570/9—82—002.
Proposed Laboratory Certification and Registration Rule, ch. NR149,
Wis. Mm. Code, June 17, 1985.
Letter from the American Council of Independent Laboratories, Inc.,
signed by Allen F. Maxfield, Chairman, Government Relations Committee,
April 16, 1985.
Letter from the National Sanitation Foundation, signed by Gary W.
Sherlaw, Director, Certification Services, to Peter S. Unger, NBS
Associate Manager, Laboratory Accreditation, April 9, 1985.
Letter from the American Association for Laboratory Accreditation,
signed by Foster C. Wilson, chairman of the Board, to Stanley I.
Warshaw, Director, NBS Office of Product Standards Policy, March 20,
1985.
-------
Appendix I
THE NVLAP ACCREDITIVPION PROCESS
BACKGROUND
The U.S. Department of Commerce, National Bureau of Standards (NBS)
administers the National Voluntary Laboratory Accreditation Program
(NVLAP). NVLAP’s function is to accredit public and private testing
laboratories based on evaluation of their technical qualifications and
competence for conducting specific test methods in specified fields of
testing. Accreditation is granted based on conformance with criteria
published in the Code of Federal Regulations as part of the NVLAP
procedures (15 CFR Part 7).
NVIJ P accreditation is available to commercial laboratories,
manufacturer’s in—house laboratories, university laboratories,
Federal, State, and local government laboratories. Foreign—based
laboratories may also be accredited by NBS if they meet the same
requirements as domestic laboratories and pay any additional fees
required.
NVLAP is comprised of a series of laboratory accreditation programs
(LAPs) established based on requests and demonstrated need.
NVLAP accreditation means recognition of a testing laboratory’s
competence to perform specific test methods in specified fields of
testing. It means that the laboratory’s quality system, staff,
facilities and equipment, calibration procedures, test methods and
procedures, records, and test reports, have been evaluated and found
to meet NVLAP criteria. NVLAP accreditation does not mean a guarantee
(certification) of laboratory performance or product test data; it is
a finding of laboratory competence.
For accreditation to be meaningful, it must be granted by a clearly
credible organization. NVLAP provides an unbiased third party
evaluation and recognition of performance as well as expert technical
assistance to upgrade laboratory performance.
OVERVIEW
Accreditation is granted following successful completion of a process
which includes submission of an application and payment of fees by the
laboratory, on—site assessments, resolution of identified
deficiencies, participation in proficiency testing, and periodic
technical evaluation and administrative review for initial and renewed
accreditation.
APPLICATION AND FEES
-------
An Application Package is sent to a laboratory on request. It
includes: General Application Forms, a Fee Calculation Sheet, and one
or more LAP Handbooks, which describe the requirements of the LAPs.
The General Application Form must be completed and signed by an
authorized representative of the laboratory. The authorized
representative is one who can act on behalf of the laboratory and
commit it to fulfill the NVLAP requirements. Before completing and
signing the application, the authorized representative should review
all documents and become totally familiar with NVLAP requirements.
Although other laboratory staff may be designated to perform
activities, such as handling proficiency testing or receiving an
assessor, the authorized representative is the only one who can
authorize a change in the scope or nature of the application.
In general, the accreditation fee is composed of several parts, some
of which are fixed while others depend on the scope of accreditation
desired and the specifics of the LAP. The total accreditation fee
must be paid before accreditation can be granted.
The laboratory will be scheduled for an on—site assessment after
payment of all required fees and will be notified of any additional
information which must be supplied and of any applicable proficiency
testing requirements which must be completed for the technical
evaluation.
APPROVED SIGNATORY
Under NVLAP criteria, an accredited laboratory must have one or more
individuals or laboratory positions designated as having
responsibility for signing “all test reports endorsed with the NVLAP
logo.” This is the person(s) to whom NVLAP, laboratory clients, or
others would go in case of questions or problems with the report.
There is no formal requirement for nomination or approval of persons
or laboratory positions designated as approved signatories. The
laboratory should inform NVLAP of its appointments by completing the
appropriate sections in he application for accreditation. Approved
signatories should be: persons or positions with adequate
responsibility or authority within the organization, with adequate and
appropriate technical capabilities, and without conflict of interest.
Laboratory test reports carrying the NVLAP logo need not be signed
individually by the approved signatory. Test report forms may be
preprinted with the required information. Forms that are
electronically or computer generated may have the information printed
along with the test results.
TECHNICAL EXPERTS
NVLAP uses Technical Experts (TEs) as assessors and evaluators. These
are individuals knowledgeable in the testing field being evaluated.
-------
They may be engineers or scientists currently active in the field,
consultants, college professors or retired persons. They are selected
on the basis of their professional and academic achievements,
experience in the field of testing, management experience, and tact in
dealing with people. Their services are generally contracted as
required; they are not NVLI P staff members.
Assessors are TEs selected to conduct an on—site assessment of a
particular laboratory on the basis of how well their individual
experience matches the type of testing to be assessed, as well as
absence of conflicts of interest. The laboratory has the right to
appeal the assignment of an assessor and may request an alternate.
Evaluators are TEs selected to review the record of the laboratory as
a whole, including the application, assessment report, deficiencies,
corrections to deficiencies, and proficiency test results and, based
on this record, to reconuiiend whether or not a laboratory should be
accredited. The evaluators are matched to the type of testing being
evaluated and are selected to avoid conflicts of interest.
ON-SITE ASSESSMENT
Before initial accreditation and periodically thereafter, an on—site
assessment of each laboratory is conducted to determine compliance
with the NVLAP criteria. The assessment is conducted by one or more
NVLIAP assessors selected on the basis of their expertise in the field
of testing to be reviewed. Assessors use checklists developed by
NVLAP so that each laboratory receives an assessment comparable to
that received by others. However, assessors have considerable
latitude to make judgments about a laboratory’s compliance with the
NVLAP criteria, depending on the assessor’s experience and the unique
circumstances of the laboratory.
Each laboratory will be contacted to arrange a mutually agreeable date
for an assessment. The time needed to conduct an assessment varies,
but two days is the norm. Every effort is made to conduct an
assessment with as little disruption as possible to the normal
operations of the laboratory. During the assessment the assessor will
carry out the following functions:
Meet with management and supervisory personnel responsible
for the laboratory’s activities (for which accreditation is
being sought) to review the assessment process with the
individuals involved and to set the assessment agenda.
Examine the quality assurance system employed by the
laboratory. The assessor may select and trace the history
of one or more samples from receipt to final issuance of test
reports. The assessor will conduct a thorough review of the
laboratory’s quality manual or equivalent, evaluate the
training program, examine notebooks or records pertaining
to the samples, check sample identification and tracking
procedures, determine whether the appropriate environmental
-------
conditions are maintained, and examine copies of completed
test reports.
• Review records of periodic internal audits, use of check
samples or participation in round robin testing or other
similar programs.
• Review personnel records including resumes and job descriptions
of key personnel, competency evaluations for all staff members
who routinely perform the testing for which accreditation is
sought, calibration or verification records for apparatus used,
test reports, and sample control records.
Observe demonstrations of testing techniques and discuss them
with the technical personnel to assure their understanding of
the procedures.
Examine major equipment, apparatus, and facilities.
At the conclusion of the assessment, the assessor will conduct an exit
briefing to discuss his or her observations with appropriate
laboratory staff and call attention to any deficiencies uncovered. A
written summary of any deficiencies discussed will be left at the
laboratory. The assessor will forward the assessment forms and a
written summary to NBS.
If deficiencies have been noted, the laboratory must, within 30 days
of the date of this notification provide NVLAP with documentation or
certification, by the authorized representative, that the specified
deficiencies have been corrected or that specific actions are being
taken to correct the deficiencies.
A laboratory applying for initial accreditation may request an
extension to complete required corrections.
If any deficiencies are noted at laboratories which are currently
accredited, such deficiencies must be corrected within 30 days after
notification or the laboratory may face possible revocation,
suspension, or expiration of its accreditation. Any test equipment
that is identified as out—of—calibration, should not be used until
corrective action has been completed. All deficiencies noted for
corrective action will be subject to thorough review and verification
during subsequent assessments and technical evaluations.
MONITORING VISITS
In addition to regularly scheduled assessments, monitoring visits may
be conducted by assessors or by NBS staff at any time during the
accreditation period. Monitoring visits may occur for cause or on a
random selection basis. These visits serve to verify reported changes
in the laboratory’s personnel, facilities, and operations or to
explore possible reasons for poor performance in proficiency testing.
-------
The scope of a monitoring visit may range from checking a few
designated items to a complete review. Failure to cooperate with
NVLAP assessors will be grounds for initiation of adverse
accreditation action. No additional fee is required for the
monitoring visit.
PROFICIENCY TESTING
Proficiency testing is an integral part of the NVLAP accreditation
process. Demonstration of appropriate facilities, equipment,
personnel, etc. is essential, but may not be sufficient for the
evaluation of laboratory competence. The actual determination of test
data using special proficiency testing samples provides NVLAP with a
way to determine the overall effectiveness of the laboratory.
Proficiency testing is a process for checking actual laboratory
testing performance, usually by means of interlaboratory comparisons.
Each LAP has unique proficiency testing requirements. The data are
analyzed by NVLAP and summary reports of the results are sent back to
the participants.
For many test methods, results from proficiency testing are very good
indicators of a laboratory’s testing capability. Information obtained
from proficiency testing helps to identify problems in a laboratory.
When problems are found, NVLAP staff members work with the laboratory
staff to solve them. If problems with the test method are suspected,
NVLAP provides information to the appropriate standards—writing
bodies.
TECHNICAL EVALUATION
After a laboratory has completed all the technical requirements of a
LAP and is ready for an accreditation action, a final technical
evaluation is conducted by experts chosen for their experience and
knowledge of the pertinent test methods. They review records on each
applicant laboratory and base their evaluation on:
information provided on the application;
on—site assessment reports;
actions taken by the laboratory to correct deficiencies;
results of proficiency testing; and
information from any monitoring visits of the laboratory.
If the technical evaluation reveals additional deficiencies, written
notification describing them will be made to the laboratory. The
laboratory must respond within 30 days of such notification and
-------
provide documentation or certification by the authorized
representative that the specified deficiencies have been corrected.
Clarification of some issues may be requested by telephone. All
deficiencies must be corrected before accreditation can be granted or
renewed.
ADMINISTRF ITIVE REVIEW
After the technical evaluation has been completed, the NVL1 P staff
prepares an administrative recommendation that the laboratory either
be granted or denied accreditation. This recommendation is based on a
review of the technical evaluation and other records to ensure that
all NVLAP technical, financial and administrative obligations have
been satisfied.
ACCREDITATION ACTIONS
Acting for the Director of NBS, the Director of the NBS Office of
Product Standards Policy makes the following decisions:
Accreditation If accreditation is recommended, the recommendation
forms the basis for granting accreditation. A
Certificate of Accreditation will be issued to the
laboratory.
Denial If denial is recommended, the laboratory is notified
of a proposal to deny accreditation and the reason(s)
the refor.
Suspension If a laboratory is found to have violated the terms
of its accreditation, the accreditation can be
suspended. The laboratory will be notified of the
reasons for and conditions of the suspensions and
the action(s) that the laboratory must take to have
accreditation reinstated.
Revocation If a laboratory is found to have violated the terms
of its accreditation, the laboratory is notified of
a proposal to revoke accreditation and the reasons
therefor. The laboratory may be given the option
of voluntarily terminating accreditation. If
accreditation is revoked, the laboratory must return
its Certificate of Accreditation and cease use of
the NVLAP logo on any of its reports, other
correspondence, or advertising.
If denial or revocation has been proposed, the laboratory may request
a hearing, under United States Code 5 U.S.C. 556, within 30 days of
the date of receipt of the notification. If a hearing is not
requested, the action becomes final upon the expiration of that 30—day
period.
-------
After a participant’s accreditation has been terminated, whether
voluntarily or through adverse action, the accreditation certificate
must be returned to NVLAP. If a laboratory elects not to renew or
wishes to voluntarily terminate its accreditation at any time, the
notification of such intention should be forwarded to NBS in writing.
ACCREDITATION PERIOD
Accreditation is granted for a period specified for each LAP (usually
one year). The accreditation period begins on one of four dates:
January 1, April 1, July 1, or October 1. Once a laboratory has been
assigned an accreditation date, it retains that date as long as it
remains in the program. Accreditation expires and is renewed on that
date.
RENEWAL
Each participating laboratory is sent a renewal application package,
well in advance of the expiration date of its accreditation, to allow
sufficient time to complete the renewal process. The renewal
application contains the same forms used for initial application, and
the laboratory need only indicate where changes have occurred from the
previous period in personnel, equipment, facilities, or the scope of
accreditation desired.
With the exception of an initiation fee for new applicants, the
technical requirements and fees are the same as for initial
accreditation. The application and fees must be received by NBS prior
to expiration of the laboratory’s current accreditation to avoid a
lapse in accreditation.
-------
Appendix II
SUMMARY AND EXCERPTS OF COMMENTS
ON
PROPOSED LABORATORY ACCREDITATION PROGRAM FOR HAZARDOUS WASTE ANALYSIS
On February 20, 1985, the National Bureau of Standards published in
the Federal Register (50 FR 7095—7096) a request from the State of
Indiana Environmental Management Board to establish a laboratory
accreditation program (LAP) for hazardous waste testing laboratories.
The notice requested that public comments on the need for the proposed
LAP be sent to NBS by April 22, 1985. The U.S. Environmental
Protection Agency requested that the continent period be extended beyond
the July 24—26 symposium on solid waste testing and quality assurance
to allow symposium participants the opportunity to consider the need
for laboratory accreditation in this area. The comment period was
extended at least one month beyond the symposium.
NBS has received letters from 29 organizations commenting on the need
for the proposed LAP. Twenty—four of the responses are positive, four
are conditionally positive, and one is negative. The positive
responses emphasize the role the LAP will have in improving the
consistency of interpretation of the test technology and resultant
test data, integrating the various efforts currently underway, and
supporting regulatory decisions. The conditionally positive responses
agree with the need for a LAP, but have reservations or suggested
alternatives. The negative response indicates that current efforts by
EPA and the states are in place and work satisfactorily.
The following are excerpts of the 29 responses received so far.
Positive Responses (24)
International Association of Waste Management Professionals
—“supportive of the concept” —— of an accreditation program
Grocery Manufacturers of America, Inc.
—such a LAP would be “important to maintain laboratory
performance, to ensure the interlaboratory consistency of
analytical results, and to assure the public that they can
have confidence in the analytical data available for
regulatory decisions.”
—“The proposal.. .describes an important need for accredited
analytical services. . .Since no system of accreditation
presently exists, NBS should establish the requested LAP. ..“
——“immediate priority should be given to the LAP for
hazardous waste analysis. The scope can be expanded to other
environmental concerns at a later date.”
National Environmental Health Association
——“the general concept of the proposal appears to be consistent
with what this professional society has been recommending in
-------
the way of practices designed to improve the quality of our
environment.”
Wabash Valley Environmental Association, Inc.
— “M environmental LAP developed and administered at a national
level would be a benefit and a service to the regulator, the
regulated community and the public. Administration at the
national level is necessary since environmental analysis is
conducted as interstate commerce.”
—“development of an environmental LAP by NBS will lead to
sounder environmental control and will provide a much needed
assurance. . .that the analytical results on which so many
environmental decisions are based are made on valid analytical
results.”
—“a LAP should be developed first for the area of hazardous
waste. . .The hazardous waste program has at its heart specific
and detailed analytical and sampling protocols on which the
critical decisions to include or exclude a waste in the
hazardous waste system hinge. The liabilities both financial
and health—related incurred by an incorrect decision based on
analysis using incorrect protocols demand the development of a
hazardous waste LAP as a top priority.”
—“A LAP for hazardous waste is also important with the recent
expansion of RCRA to include small quantity generators. These
individuals will undoubtedly have to rely on commercial
laboratories for analytical support.”
EMS Laboratories, Inc.
—“wholeheartedly support the concept of laboratory accreditation
...The need for a greater degree of uniformity and credibility
in all areas of environmental testing manifests itself more each
day.”
—“Many of the test methods for the analysis of solid and
hazardous wastes are similar or identical to those for the
analysis of water and wastewater. It seems appropriate,
therefore, to develop a LAP which could encompass “the whole
range of environmental measurements” as has been proposed...”
—“such a program would save time for State and Federal
regulatory personnel and would save money for users of laboratory
services. Frequently, we are called upon to generate and submit
a variety of ancillary data with our laboratory reports and the
generation and review of this data is time—consuming for everyone
concerned. It also increases the cost.”
National Laboratories, Inc.
—“strongly support the request...”
—“it is entirely conceivable that many violations can be
attributed to unreliable analytical data. A requirement of
certification of analytical data, accompanied by the
certification number of the laboratory could have a salutory
effect on business for those laboratories that are serious in
their efforts to provide reliable, reproducible data.”
——“Whatever the cost, and however it is accomplished, it should
be done by an agency which is independent of the regulatory
-------
agencies and of the private laboratories. Under the present
(non) system, it is easy for hearing officers to ignore any facts
which are in opposition to the regulatory agency’s position.”
Northern Laboratories, Inc.
—“development of an accreditation program for the analysis of
hazardous materials would be of great benefit to the laboratories
performing the analysis and users of these laboratories alike to
assure that reproducible results can be obtained regardless of
where or when the analysis is performed. Laboratories would have
greater confidence in their procedures and users of these
laboratories could be confident that the work they are having
performed would be acceptable to regulatory agencies.”
—“Accurate and consistent test results achieved by approved
laboratories are a necessary first step in the design of a
comprehensive program of hazardous waste site evaluations and
clean—ups.”
Environmental Consultants
——“the request.. .has merit and should be considered.”
——“The need is clearly present to consolidate all external
laboratory evaluations under one program, where all interested
laboratories can participate.”
——“Most important to the commercial laboratory services industry
is a well—defined protocol for becoming an accredited laboratory
in environmental analyses, and it will equally be beneficial to
State environmental programs to be able to identify those
facilities which are qualified.”
——“would benefit the commercial laboratory industry by
eliminating the current confusion over ‘Certified’ reporting.”
Pollution Control Systems Incorporated
——“voice our support...”
—“would standardize the procedures being used for the analysis
of hazardous waste and decrease interlaboratory variations in
procedures and results.”
——“standardize the interpretation of 40 CFR Part 261.”
Chevron U.S.A., Inc.
—“strongly endorse NBS rapidly establishing a LAP for waste
analysis.”
—“would ensure the quality control of local laboratories...”
—“could provide monitoring of waste analysis data quality and
provide a statistical database to develop real precision
statements for the EPA SW—846 methods.”
——“would weed out incompetent laboratories and improve day—to—day
documentation and practices for quality laboratory measurement.”
——“would be an incentive for laboratories to develop and sustain
a high level of performance.”
——“EPA and some states have proficiency testing for RCRA and
CERCLA, however, these programs are not developed for general
waste characterization on a national scale.”
——“NBS is the proper disinterested third party to develop a LAP.”
-------
Cununins Engine Company, Inc.
—“it would be very desirable for both producers and consumers of
analytical services if there were a universally recognized
national accreditation program which would help assure that the
analytical results produced in support of solid waste management
activity are as accurate as possible.”
—“it would be desirable to expand the scope of any accreditation
program beyond the limit of solid waste procedures to include all
aspects of environmental analytical testing.”
Dow Chemical U.S.A.
—“laudable to verify the consistency and equivalence of testing
being performed throughout the country.”
——“encourages use of standards from ASTM D—34.”
Duke Power Company
—“such a program is warranted and would offer many advantages.”
—“would assure that analytical results for hazardous wastes
throughout the country were obtained using consistent and
equivalent methods.”
—“Present variability between laboratories performing these
analyses would be decreased. Interpretation of test methods
found in 40 CFR Part 261 would also become more consistent.”
—“the burden and the subsequent liability of selecting and/or
not selecting a reputable laboratory, would be diminished for
the generator of hazardous wastes.”
SOHIO The Standard Oil Company
—“support expanding NVLAP into the area of environmental
testing.”
—“a real need for this type of program in all of the
environmental testing subgroups (air, wastewater, drinking water,
solid waste, etc.). However, each should be addressed
independently.”
Sylvania Chemicals/Netals
—“would welcome the opportunity for accreditation...”
—“Our interest...would be best served by a modular approach...”
State of Connecticut Department of Environmental Protection
——“supports...proposal to establish an environmental testing
laboratory accreditation program (LAP)...”
—“would prefer that such a program not be limited to hazardous
waste analyses, but rather also include air and wate analyses
for specific toxic compounds...”
—“such a program ‘should’ be at least partially supported by
user fees.”
——“such a LAP would fill an existing nationwide need and would be
a giant step toward consistency in state and federal hazardous
waste and toxic pollutant analyses policies.”
State of Delaware Department of Natural Resources & Environmental
Control
——“fully supports the request...”
-------
—“would serve two valuable purposes: assurance of the accuracy
of data so that the environment may be effectively protected, and
assurance that laboratories will take the safety precautions
appropriate to protect their employees, the general public, and
the environment from the consequences of the testing process
itself.”
State of Florida Department of Environmental Regulation
——“concurs with the. . . request. ..“
—“support the suggestions to expand the laboratory accreditation
program to include air, water, and solid waste, as well as
hazardous waste.”
State of Georgia Department of Natural Resources
——“wishes to add its support to a voluntary laboratory
accreditation program (LAP) to be promulgated by the National
Bureau of Standards. The need is urgent and the problems are
diverse.”
Couunonwealth of Kentucky Natural Resources and Environmental
Protection Cabinet
—“wholeheartedly support a national accreditation program for
hazardous waste laboratories.”
State of Maine Department of Environmental Protection
——“express this Bureau’s support for the laboratory accreditation
program for hazardous waste.. .also urge that consideration be
given to expanding the concept to include accreditation for
analyses needed to comply with or administer other EPA programs
in addition to RCRA.”
——“This Bureau has had undesirable experiences, including non-
productive use of our resources and misunderstandings...
resulting from questionable analyses for hazardous waste
parameters.”
—we are “considering the creation of a State approval system...
there is a widespread perception that some such program is
needed.”
——“Clearly it would be desirable to have a nationally uniform
system of accreditation, even if there were to be significant
state participation. Having uniform criteria, evaluation
instruments, audit sample programs and so forth would better
ensure an effective program and consistency and facilitate
reciprocity.”
State of Maryland Department of Health and Mental Hygiene
——“there is a need to support such a program so proper and
uniform methodologies and procedures would be used nationwide
in the testing of hazardous wastes.”
Missouri Department of Natural Resources
—“An accreditation program encompassing all environmental
measurements as well as hazardous waste would be a welcome
activity from our perspective.. . from the perspective of the
regulated community.”
-------
—“We in the environmental regulatory field depend heavily
on analytical data to make informed decisions. Our ability to
make sound, uniform and fair decisions is no better than the
data on which the decisions are based. Our years of experience
in other environmental control areas has taught us that there
is a wide spectrum of quality found in the laboratory market-
place. In the absence of control, whether it be regulated or
voluntary, the quality of laboratory services tends toward the
lowest coimnon denominator because of the competitive advantage
of reduced overhead which in many cases means reduced quality.”
State of New Mexico Environmental Improvement Division
—“strongly support the concept of a laboratory accreditation
program (LAP) for environmental testing laboratories, to be
developed by the National Bureau of Standards cooperatively
with the Environmental Protection Agency (EPA). The intent
of the quality assurance (QA) programs required by EPA for any
state—delegated program receiving EPA funds would be addressed
in large part of having a nationwide LAP covering chemical
analysis for all environmental parameters. Such a LAP might
follow the format of the existing EPA Manual for the
Certification of Laboratories Analyzing DrinkTi Water . . .but
should be expai ed...”
Conditionally Positive Responses (4)
American Association for Laboratory Accreditation
—“have no objections to NVLAP developing a program in these
fields in competition with APLA. . .“
American Council of Independent Laboratories, Inc. (ACIL)
—“the need for a national program for laboratory accreditation
definitely exists in the environmental (hazardous waste) field.
And as the notice hints that might well be expanded to the areas
of air, water and solid waste.”
——“while the testing methodology to support accreditation
currently exists, the non—technical issues are very complex.
For example: the need for the LAP to be self—supporting.
Another issue is whether EPA will replace its current system
with NVLAP and the time for such a transition. Another issue
concerns the reciprocity needed between states to have a NVLAP
program successful. Finally, the question of whether or not the
American Association for Laboratory Accreditation (AALA) might
not offer a viable altenative to NVLAP, in this environmental
area, needs to be considered.”
National Sanitation Foundation
—“The need is evident, and NSF can provide a viable third—party
alternative to official regulation.”
——“Drinking water is the first area of laboratory accreditation
where NSF will offer services. When the DWLA program is
established, we plan to expand accreditation services to waste—
water laboratories and ultimately to hazardous waste
-------
laboratories.”
Lancy Laboratories
——“in cooperation with the USEPA, develop a program that would
cover water, wastewater, solid and hazardous wastes.”
——“The only negative comment...Unless a large majority of
certifying agencies accept this program in lieu of their own,
this will just be another certification burden that many
laboratories would have to maintain.”
Negative Response (1)
Analyte Laboratories, Inc.
——“do not feel the environmental issue should be addressed by
the NBS.”
—“First, the EPA, through EMSL, has a well—established library
and distribution network for documentation, especially current
methodology. Second, EPA also maintains a repository for
hazardous material and performance evaluation materials. This
program is more than adequate to provide laboratories with
quality guidance measures. Third, as part of normal interaction
with private business, contract laboratories are routinely and
thoroughly inspected for their capabilities. The intensity of
their investigation is noteworthy. The EPA also performs ongoing
research within its laboratories to maintain the above programs.”
——“EPA. . . is more capable of readily instituting or modifying any
type of LAP. . . see no need for NBS to expand its LAP and
unnecessarily duplicate an established program.”
-------
CONTROLLING AND COPING WITh UNWANTED VARIANCE IN GROUNEX 1ATER
MONITORING DATA: QUALflY CONTROL AND STATISTICS
BURNELL W. VINCENT, PROGRAM MANAGER, GROUNT ATER PROGRAM, OFFICE OF
SOLID WASTE, U. S. ENVIRONMENTAL PROTECTION AGENCY, WASHINGTON, D. C.
INTRODUCTION
Establishing an appropriate statistical comparison program in
accordance with the RCRA regulations may be one of the most vexing
tasks facing a permit writer. The purpose of requiring statistical
comparisons is to provide a standard protocol for determining whether
differences between up— and down—gradient ground—water quality values
are due to facility effects. There are many sources of variance in
the data which are not related to the facility. Some are
representative of actual differences in ground—water quality, such as
near by sources of contamination or naturally occurring special
variability; others are due to measurement error or to bias introduced
by sampling methodology. A good statistical protocol will accommodate
anticipated variance and selectively identify the type of variances
which do indicate facility impact.
Many of the problems addressed by other speakers at this conference
contribute variance in the data reported to a facility’s permit
writer: the problems in obtaining standards, in laboratory
certification, and in many of the aspects ordinarily considered in a
quality assurance program. The reason that a statistical comparison
technique is such a vexing problem to the permit writer at this point
in development of the state of the art is that these various sources
are all confounded in the data. “Statistics” or the selection of a
statistical comparison technique has been given a bum rap. It is
blamed for false positives by owners and operators and false negatives
by environmentalists when, in fact, it does just what we ask of it.
Rather, the point of this presentation is that false positives are
much more directly due to violations of the requirement for
representative samples than they are due to faults of the regulation.
Failure to implement adequate quality assurance programs introduces
variance which can be orders of magnitude greater than the
concentrations representing lethal doses of some of the constituents
of concern. Equally important, there are gaps in the technology of
ground—water sampling. Some devices commonly in use have potential
but undocumented influence on sample quality.
The objective of this presentation is to discuss two initiatives at
EPA which promise to bring some short term relief to the permit writer
who must draft the statistical and quality assurance provisions of
permits. The first of these is an effort to improve guidance
directives specifying quality control. It should be familiar to many
of you who have read our draft Technical Enforcement Guidance
Document , the topic of address by Murphy and Gilbertson at yesterday’s
session. The second is an initiative to develop guidance for a
-------
simple, workable statistical package which satisfactorily complies
with the regulation.
THE PROGRM
The regulations generally require owners and operators to prepare a
ground—water monitoring program capable of detecting any influence
from the facility on ground—water quality. This program, as specified
in both interim status and permit requirements, must include
consistent sampling and analysis procedures that are designed to
ensure monitoring results that provide a reliable indication of
ground—water quality. The regulatory intent was to express only the
performance standard in the promulgated text, and to develop the
details for meeting that performance standard in guidance documents.
Unfortunately, the comparison technique specified in the regulations
assumes that replicate values are as independent of each other as the
between—quarter values are.
Typically, there are three levels of guidance materials: the first is
quasi regulatory, such as SW—846, which will be incorporated by
reference into the regs. The second is interpretative or policy
guidance, such as the permit writers’ guidance manuals and the
Technical Enforcement Guidance Document . These manuals contain “rules
of thumb and other specific normative criteria by which regulatory
officials and owners and operators can decide whether particular
methods or techniques are seen by the Agency as satisfying the
performance standards as promulgated. The third level of guidance
doucments are scientific research reports and technical resource
documents describing the state of the art techniques, but not
prescribing sufficiency from a regulatory standpoint.
We are currently seeing the technical resource documents and research
reports come to maturity. In the midst of the ongoing research,
however, it is clear that a large portion of the variance is
correctable simply by standardizing the materials and procedures for
sampling. Therefore, this second body of guidance documents, the
rules—of—thumb, are being redrafted in an effort to foster such
standardization. Upon finalization of these documents, those rules
which are hard and fast and well—supported by science will become
mandatory as requirements in SW—846.
Several technological gaps have been identified over the last several
years. These are just now appearing in the third level of guidance
documents. One such issue for example is the choice of
non—interactive sample contact surfaces in sampling equipment. The
wells and pumps used to bring samples to the surface must be designed
to overcome many physical obstacles. Some of the best engineering
materials, chosen for durability, flexIbility, and cost, are made of
materials which are directly of concern or interactive with
constituents of concern as contaminants. Until very recently, the
agency has not begun to quantify the influence of the methods and
materials of sampling on the resulting ground—water quality data.
-------
In the midst of ongoing research, and for the sake of consistency,
quality control, and fairness to the regulated community, the Agency
is drafting a list of criteria for the methods and materials of
sampling.
SPECIFIC PROBLEMS
The output of detection monitoring is the determination that there is
or there is not a facility—induced difference between the background
and the potentially affected ground water. With the dozens of sources
of variance which are becoming identified, it is becoming increasingly
apparent that the lack of difference between these values could be due
simply to the artifacts and the real concentrations cancelling each
other (a false negative). Alternatively, sharply differing values due
to reinforcing artifacts could result in a false positive.
The two initiatives previously mentioned are a beginning, in the near
term, to provide controls for the most controllable influence and to
accommodate variances which are not controllable. Over the longer
run, as more and more ground water quality data becomes available,
more sophisticated analyses of sources of variance will help our
understanding of variances which by their nature are so confounded as
to appear inseparable. An example is an apparent seasonality of
volatile constituents which in fact is a bias introduced by sampling
error. When samples to be analyzed for volatile constituents are
exposed to the air during the sample acquisition, preservation, and
containment, the reported values may be influenced by sun, humidity,
and temperature of the season in which they are sampled. Separating
the influences due to these two sources of variance is not now
amenable to statistical methods.
The initiative to quantify and control the controllable influences is
coordinated between RCRA enforcement offices and a Task Force, the
Hazardous Waste Ground—Water Task Force, which was established this
year to accomplish the two major goals of determining whether
regulated facilities ground water requirements and to identify and
evaluate causes of poor compliance. That Task Force will evaluate
approximately 60 regulated facilities. The dual objectives of the
Task Force’s effort are to determine whether the facilities are
leaking and to document the difficulties which have made that
determination so elusive, heretofore.
The Office of Waste Programs Enforcement, using panels of experienced
Agency enforcement officials and researchers, is compiling a set of
criteria which will standardize well design and construction and
sample taking techniques for new wells, and set standards for
acceptability of existing wells. The Task Force is fully integrated
with this effort. Task Force protocols for inspection of the 60
facilities to be evaluated are consistent with enforcement findings;
new issues identified by the task force will be incorporated into the
guidance if they are resolvable issues. Non—resolvable issues will be
prioritized for accelerated research.
-------
The Enforcement standards for controlling sampling induced error are
expressed in the Technical Enforcement Guidance Document which is
currently available in draft for public comment. This manual presents
standard techniques designed to resolve a lengthy list of commonly
encountered but easily resolvable sources of error. For instance, it
is not uncommon for enforcement officials to observe that wells have
been sealed using neat cement. Such seals, in proximity to well
intakes, can have a major impact on pH and through pH on the mobility
and solubility of many of the constituents of concern. The manual
identifies this problem and forbids the use of neat cement.
Steel and plastic in corrosive or solvent environments is identified
as a source of error (due to interaction), as well as a loss of
stability in background data (the loss and the replacement of old
wells which have corroded away can destroy continuity of historic
background data).
Typically, drilling techniques have been developed for water supply
wells or petroleum industry use. Commonly, drilling muds or fluids
are injected ahead of drill bits to loosen or dislodge the cuttings.
While introduction of fluids for high volume production wells is of
little concern, these muds can have a devastating impact on a
detection monitoring program. Similarly, wells for residential use
are commonly field slotted. Rather than purchasing expensive screens
uniformly slotted to a size which excludes packing and aquifer
materials, the well driller will simply use a hacksaw to slot the well
casing, resulting in newly cut surfaces of interactive material, and
in wrong—sized openings which allow passage of sediment along with the
water sample. The experts advising the Agency on “rules—of—thumb”
have suggested that when highly turbent samples are encountered and no
information on well construction can be provided, the guidance must
call for the wells to be replaced.
Filter packs are commonly inserted between the aquifer material and
the well screen; they are designed to keep soils and sediment away
from the screen. Improperly sized and/or interactive materials in a
filter pack, of course, can affect values measured at that well.
Guidance will likely reflect these factors.
The draft document identifies allowable drilling methods. A wide
variety of drilling methods have been developed for various purposes
in various geologic settings. Several of these should be prohibited
because of known effects on monitoring objectives. Guidance documents
should present drilling methods which are acceptable according to
geologic settings. Practices and precautions for their use should be
identified in order to further control the effects on the resulting
data.
The problem of interactive sample contact surfaces is one which
requires considerable further research. Meanwhile, for
standardization and to avoid the probability of induced error, the
advisors have recommended that our guidance require teflon or
stainless steel sample—contact surfaces. The use of other materials
may be proposed by owners and operators in consideration of site
-------
specific factors and monitoring objectives. However, the use of well
materials which require glue or solvents during installation should be
expressly prohibited in all sampling wells. Bentonite pellets were
recommended for sealing the annulous.
Detection monitoring requires information about the concentrations of
specific constituent in the ground water as it exists in situ . As the
ground water leaves the porous medium, flows through t1i filter pack,
and the screen, and enters the well, it may be subjected to influences
affecting those concentrations. Stagnant waters standing in the well
column can be expected to have differences in concentrations of key
indicators. The guidance is drafted in attempt to control such
influence. Free flowing aquifers should be prebailed by an amount
equal to three times the casing volume. While standard practice
permits a general range from 1—10 volumes, this amount of variance can
have measurable effects on concentrations.
Free flowing aquifers must be sampled immediately upon removal of the
third volume of prebailing. In tighter formations, well recovery may
be too restricted to allow discarding the three volumes. This
presents additional problems. In these situations, the recharging
water trickles down over the screen in a manner which allows air
stripping of volatiles. To minimize this effect, the enforcement
official should require that the wells which go dry before purging
three volumes be sampled within three hours of recovery of a
sufficient volume for analysis. While this approach will minimize
variance, a bias remains which must be addressed by new technology
development.
Sample taking procedures are specified to reduce the trauma involved
in handling sampling equipment and in equipment/sample interface.
Commonly, field crews are observed dropping the bailer down the well
casing where it splashes into the water to be sampled. The
enforcement official must forbid this practice, as well as the use of
stranded cord, nylon, or other potentially contaminating line for
retrieving the bailer.
To reduce sampling bias, bailers and pumps for use in sampling may not
have neoprene fittings, PVC, tygon tubing, silicon rubber, neoprene
impellers, polyethylene, or vitron in contact with samples. If they
are not dedicated to an individual well, sampling devices must be
steam cleaned between wells. Pumps may not operate at greater than
100 ml per minute extraction rate for the sample; the extraction rate
under which the sample was taken must be recorded. Samples must be
preserved in accordance with SW—846; they must be placed in a sample
container with no head space. They should not be composited for
subsequent division in the laboratory. Quality assurance programs
must specify the use of field blanks, standards, and spikes.
Calibration of lab equipment must be required. Standard detection
limits must be agreed upon (such as, the ACS, 1980 approach).
These are but a sample of the types of standards needed in order to
reduce the controllable sources of variance before the statistician
can be expected to identify the influence of a facility. The ranges
-------
of error we have associated with these sampling irregularities is
indeed large. Field tests have attributed 10—15% error to even the
better sampling devices. Operator error in some devices has been
shown to result in a 50% loss of some volatiles. Studies at the
Illinois Geologic Survey have identified 80% losses of some volatiles
to silicon rubber tubing. Suction type sampling pumps can routinely
be expected to lose 25—30% volatiles.
INTERIM ST1 TISTICAL GUIDANCE
The use of Cochran’s Approximation to the Behrens—Fisher solution has
raised concerns about false positives. This is intuitively obvious
when we stop to consider the assumptions on which that test was based.
Fortunately, by good sense or good luck, the wording in paragraph
264.96 (h) (1) (ii), the alternate statistical test, does not require
the use of replicates. Permit writers should take advantage of this
flexibility by requiring replicates for quality control but averaging
replicates into a single value.
This flexibility supports fertile cooperation between statisticians
and hydrogeologists in designing a sampling protocol well tailored to
the objectives. Trade—of fs between the numbers of wells, the
frequency, and the number of parameters should be judiciously made.
Some randomization is appropriate, but the intuition of the
hydrogeologist is also important.
we are now drafting advice to our permit writers based on the existing
regulations (S264.97(h)(2)(ii). The draft is in keeping with the old
adage, KISS (keep it simple, stupid), recognizing that no one
statistical comparison technique will ever work for all sites, and
further recognizing the difficulties in determining the comparison
technique before sufficient data is amassed. The draft guidance
advises a relatively simple technique to be used with lots of
intuition and visual displays. Graphs and control charts should be
used in conjunction with any specified simple comparison technique in
order to establish confidence that the technique gives sensible
answers.
The alternative statistical test is only available for sites whose
data exhibit a coefficient of variation less than one. With its
freedom from the replicate requirement, the alternative test should
take full advantage and maximize independence of the upgradient
values. To this regard, no fewer than four upgradient wells are
suggested. Four quarters at four wells will yield 16 independent
values avoiding the problem of falsely inflated degrees of freedom.
The power of the statistical package should be viewed as a whole. A
basic understanding of statistics and the requirements for consistent
procedures for sampling should enable the permit writer to balance the
possiblity of false positives and false negatives.
The permit writer’s agreement to allow an alternative statistical test
should be based on the owner or operator’s commitment to bias—free
sampling procedures. He/she should agree to consistently use a single
-------
laboratory, or to split at least three routine sample periods between
new and old labs prior to changing labs. The ACS detection limit
package should be used. The owner or operator should agree to
safeguard against missing values, such as preservation of additional
or redundant sample volumes. He/she should use site—specific
frequency—spacing selection criteria, not necessarily uniformly
requiring the same frequency for all parameters and all wells. And
finally, the owner or operator should insist on the best available
QIVQC.
In recognition of the increase in system capability to detect
contaminants, the permit writer can offer a fixed degree of freedom
which will achieve more statistical power with less sensitive “hair
trigger” upset values. This is the nature of the Bond—Ferranti test
which adjusts the overall alpha by adjusting for individual
comparisons.
CONCLUSIONS
In summary, the Agency now has a two—pronged initiative aimed at
reducing the “noise—to—signal ratio” in ground water monitoring
programs. It involves reducing the controllable variances which are
due to sampling error and accounting for the real seasonal special and
temporal variances by practical, simple statistical procedures. A
simple comparison technique backup by visual and intuitive tests
should result in heightened efficiency, useful until sufficient
information is obtained on data variability and constituent behavior
is available for more rigorous statistical evaluations.
Owners and operators should become familiar with the concepts
presented in the draft regulatory interpretation guidance documents.
While these concepts are presented in draft for comment, each
addresses a potential source of error which could reduce the
effectiveness of a monitoring system. After close of the comment
period, the Agency will prepare final guidance containing those
concepts most positively identified with the larger components of
variance. Combining these controls on variance with a simple and
reasonable alternative statistical test, should result in a marked
improvement in detection house capability.
-------
SOURCES AND MEANS OF OBTAINING COMPOUNDS FOR THE QUALITY ASSURANCE
MATERIALS BANK
KANTON, U. S. ENVIRONMENTAL PROTECTION AGENCY-EMSL, LAS VEGAS,
NEVADA
ABSTRACT
“The quality of all analytical measurements rests utlimately upon the
quality of reference materials employed.” This statement is especially
appropriate in the determination of trace toxic organic residues in
environmental matrices. To support the need for a certified,
quality—controlled, common data base for the analysis of hazardous
waste residues in a variety of substrates, the U.S. Environmental
Protection Agency maintains a repository of analytical grade reference
standards. Known officially as the U.S. EPA quality assurance
materials bank, the program is located at research Triangle Park,
North Carolina, and is both funded and technically directed by the
agency’s environmental monitoring systems laboratory in Las Vegas, NV
(EMSL—LV). Although originally established by the U.S. public Health
Service in 1965, the program, under the auspices of EMSL—LV, has
continued to expand its inventory of toxoc organic and hazardous waste
standards in support of the requirements of the federal insecticide,
fungicide, and rodenticide act (FIFRA), The resource conservation and
Recovery Act (RCRA, 40 cfr, Part 261, Appendix VIII — hazardous
constituents) and the comprehensive environmental response,
Compensation and Liability Act (CERCLA) 40 cfr, Park 302, Appendix A —
List of Hazardous Substances.
The repository replenishes its supply of standards by direct purchase
from chemical supply houses, from other government repository
programs, and by direct request of chemical and pesticide
manufacturers worldwide. This last source supplies the majority of
chemical standards used by the repository. Letters are sent to
chemical and pesticide manufactureres requesting small quantities of
pesticide chemicals for use as standards. Because the chemicals are
provided free, requests for individual compounds are made to numerous
manufacturers in hope that sufficient quantities of each compound can
be maintained in the repository.
The response to these requests has been very positive. Standards are
currently being provided free by 181 chemical manufacturers; 71 of
these companies are either foreign or foreign—owned U.S. subsidiaries.
Foreign companies provide approximately one—third of the standards
that are ultimately distributed by the repository.
After the compounds are received (gratis or purchased), they are
analyzed to verify their identity and purity. When necessary, these
compounds are purified to meet minimum acceptance criteria as
standards for instrument calibration. Once verified, high puri y
organic standards are developed (standard solution or “neat”) and
distributed for use by laboratories for the calibration of iz)struments
and for quality control in sample analysis. The purity,
concentration, stability, and applicability of each standard are
evaluated by the QAMB and by its support contractor.
------- |