PB81-20380M
Evaluation of Procedures for Identification of
Hazardous Wastes. Part 1. Sampling
Extraction, and Inorganic Analytical Procedures
(U.S.) Environmental Monitoring and Support Lab
Las Vegas, NV
Apr 81
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
-------
PB81-203804
EPA-600/4-81-027
April 1981
EVALUATION OF THE PROCEDURES FOR IDENTIFICATION
OF HAZARDOUS WASTES
Part 1 - Sampling, Extraction, and Inorganic
Analytical Procedures
by
L. R. Williams, E. P. Meier, T. A. Hinners,
E. A. Yfantis, and W. F. Beckert
Quality Assurance Division
Environmental Monitoring Systems Laboratory-Las Vegas
and
T. E. Gran
Northrop Services, Inc.
Las Vegas, Nevada
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY. ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE, IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE.
-------
TECHNICAL REPORT DATA
(Pleat read Inttruenons on tk* reverse btfon completing}
EBPA0-»-81-027
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
EVALUATION OF THE PROCEDURES FOR IDENTIFICATION OF
HAZARDOUS WASTES: Part I - Sampling, Extraction,
and Inorganic Analytical Procedures
5. REPORT OAT
PORT DATE
April 1981
8. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
L. R. Williams, E. P. Meier, T. A. Hinners,
E. A. Yfantis, W. F. Beckert, and T. E. Gran
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
P.O. Box 15027
Las Vegas. Nevada 89114
10. PROGRAM ELEMENT NO.
A8FD1D/C8FD1D; A74D2B/874D2B
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency—Las Vegas, NV
Office of Research and Development
Environmental Monitoring Systems Laboratory
.as Vegas, Nevada 89114
13. TYPE OF REPORT AND PERIOD COVERED
5/79 - 7/80
14. SPONSORING AGENCY CODE
EPA/600/07
19. SUPPLEMENTARY NOTES
16. ABSTRACT
A study was performed to evaluate the sampling, extraction, and analytical
>rocedures (inorganic) proposed in the RCRA regulations for identifying wastes as
lazardous by the toxicity characteristic. Twenty-seven different wastes were sampled
and analyzed in accordance with the RCRA regulations. The high degree of heterogeneity
found in many wastes underscores the need for a carefully designed sampling protocol to
reproducibly obtain representative samples from each waste source. A protocol was
developed and tested for obtaining composite samples from waste ponds or lagoons.
Samplers tested, the pond sampler and the COLIWASA (composite liquid waste sampler),
were found to acceptable for sampling hazardous waste, when used in a well-designed
sampling protocol. Reliability and reproducibillty of the EP were evaluated (RSD <15S).
The blade-type rotary extractor (as cited in the proposed regulations), a tumbling-type
extractor, and a wrist-arm-type shaker were compared. These three types yielded similar
EP extracts. The supporting analytical methods (atomic absorption spectrometry) were
round to be highly reproducible for Cr and Pb, and somewhat less for the Ba (RSDs <3.1%;
4.6%; and 16.4%, respectively). Independent analyses of the same waste extracts by two
laboratories were highly reproducible, i.e., the variance from analyses was negligible.
However, differences in the EP extracts produced by the two laboratories show the need
for a detailed and concise protocol for conducting the EP. Problems with sample
contamination from the blade-type extractor (chromium) and the filtration apparatus
'hayinmN uioi-o
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring Systems
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
ii
-------
CONTENTS
Page
Figures iv
Tables v
Acknowledgements viii
Introduction 1
Background 1
Conclusions and Recommendations 3
Methods 5
Selection of Wastes 5
Collection of Samples 6
Shipment and Storage of Samples 10
Sampling Procedure Evaluation 10
Extraction Procedure (EP) Evaluation 15
Analytical Procedures Evaluation 23
Results and Discussion 25
Sampling Procedures Evaluation 25
Extraction Procedure Evaluation 35
Analytical Procedures Evaluation 61
References 72
Appendix 1. Use of the Pond Sampler and the COLIWASA 73
Appendix 2. Protocol for Conducting Extraction Procedure to
Identify Hazardous Waste 77
Appendix 3. Williams/Beckert Drum and Tank (DAT) Sampler 81
iii
-------
FIGURES
Number Page
1 Sampling grid showing site selection using
the random two-digit number scheme for field
composited samples 14
2 Flow chart of the extraction procedure (EP) for the
Identification of a waste as hazardous by the toxldty
characteristic 17
3 Diagram of NBS tumbling-type extractor 21
4 Levels of chromium contamination found in extraction
blanks of distilled water and 0.1N acetic add
which followed hazardous waste samples 56
5 Evaluation of apparatus contamination - Comparison of
filtered and unflltered allquots (splits) of
sequential acidified blanks 57
A-l Pond sampler 74
A-2 Composite liquid waste sampler (COL IUASA) 75
A-3 WilHams/Beckert Drum and Tank (DAT) Sampler 1n drum
sampling configuration 82
1v
-------
TABLES
Number Page
1 Sites Sampled During FY 79: Identification of Sites, Facility
Functions, Waste Streams Sampled, and Sampling Method 7
2 Wastes Sampled Using the Pond Sampler and COLIWASA 13
3 Conditions of the Apparatus Abrasion/Contamination Study 23
4 Municipal Sewage Sludge Samples Tested 24
5 Evaluation of Pond Sampler: Results of pH and Percent Solids
Analyses of Duplicate Aliquots, from Duplicate Samples of Wastes
Taken at Five Locations on Pond 0, Site A 26
6 Pond Sampler: Means and Standard Deviations of Duplicate
Analyses for pH and Percent Solids in Waste Samples from
Pond 0, Site A 26
7 Pond Sampler: Means and Standard Deviations of Data that Reflect
Differences Between Two Waste Samples Taken at Each of Five
Locations on Pond 0, Site A 27
8 Evaluation of Pond Sampler: Results of pH and Percent Solids
Analyses of Triplicate Aliquots, from Duplicate Samples of
Waste Taken at Five Locations on Pond 13, Site A 27
9 Evaluation of Pond Sampler: Means and Standard Deviations of
Triplicate Analyses for pH and Percent Solids in Waste
Samples Taken at Five Locations on Pond 13, Site A 28
10 Pond Sampler: Means and Standard Deviations of Data that
Reflect Differences Between Two Waste Samples Taken at
Each of Five Locations on Pond 13, Site A 28
11 Reproducibility of Pond Sampler: Mathematical Composite of
First and Second Samples from each Location 31
12 Average Relative Standard Deviations (RSD's)
for Various Levels of Sampling 31
13 Evaluation of the Reproducibility of Composite Samples Obtained
with a Pond Sampler - Non-Filterable Solids (%) 32
-------
TABLES (Continued)
Number Page
14 Evaluation of COLIWASA Sampling Method with
Drummed API Separator Waste 33
15 Evaluation of the Reproducibility of Biphasic (Oil/Water)
Liquid Samples Collected with a COLIWASA from
Waste Drums 34
16 ICP Screening Analysis of EP Extracts: Approximate Elemental
Composition of Extracts from Selected Waste Samples 36
17 Evaluation of Extraction Procedure (EP): Means and Standard
Deviations for AA Analyses of EP Extracts for Barium,
Chromium and Lead 37
18a Evaluation of Extraction Procedure (EP): Average Means and
Standard Deviations for AA Analyses of EP Extracts of Wastes
from Ponds 0 and P, Site A 39
18b Relative Standard Deviations (RSD) of Extractions and
Analyses for Selected Metals 39
19 Extractor Comparison - Mean and Standard Deviation of Con-
centrations of Metals Extracted with Three Extractor Types. . . 41
20 Extractor Comparison - Variability Attributable to Sampling
and Analytical Procedures 42
21 Extractor Comparison - "F Values" Calculated from
Analyses of Variance 43
22 Extractor Comparison - Difference Between Techniques,
by Waste Type 44
23 Extractor Comparison - Results of Multiple Comparison
Test to Rank Extraction Techniques 45
24 Levels of Dissolved Metals (mg/1) in Acidified Blank
Samples - Average of Triplicate Determinations 47
25 Levels of Dissolved Metals Found in EP Extracts of Selected
Waste Samples - Average of Triplicate Extractions 48
26 Levels of Dissolved Metals (Total) Found in Digests of Selected
Waste Samples - Average of Triplicate Digests 49
27 Evaluation of Extraction Procedure (EP): Comparison of
Mercury Concentrations (Cold Vapor AA) in Waste Sample
Digests with those Estimated from EP Extracts 50
vi
-------
TABLES (Continued)
Number Page
28 Evaluation of Background Concentrations of Elements from
EP Extractors. Means and Standard Deviations of Triplicate
Atomic Absorption Analyses of Acidified Blank Samples 52
29 Evaluation of Background Concentrations of Elements from
Filtration Apparatus: Mean and Standard Deviations of
Triplicate Atomic Absorption Analyses of Deionized Water
Blank Samples 53
30 Evaluation of Background Concentrations of Elements from EP
Equipment: Means and Standard Deviations of Atomic Absorption
Analyses of Filtered and Extracted Acidified Blank Samples. . . 55
31 Sand Abrasion/Contamination Study - Concentrations of Chromium,
Iron, and Barium Detected in Filtered and Unfiltered Reagent
Blanks and Extracts of Acid-Washed Sand 58
32 Frequency and Magnitude of Contamination of Blanks (0.1N
Acetic Acid) from Extraction and Filtration Apparatus
33
34
35
36
37
38
39
40
41
42
W^W 1 TWVI I'M! 1 II y h/\ ^ 1
Slope of Standards in
Extracts Relative to
Quality
Quality
Quality
Quality
Quality
Quality
Quality
Quality
Suppress
Control
Control
Control
Control
Control
Control
Control
Control
Mon of
Data:
Data:
Data:
Data:
Data:
Data:
Data:
Data:
Atomic
Acetate Buffer and in Hazardous Waste
Slope of Standards in Nitric Acid. . .
Comparison
Compari son
Comparison
Comparison
Comparison
Comparison
Comparison
Comparison
Absorption
of
of
of
of
of
of
of
of
Pb f
Barium Spike Recovery. .
Chromium Spike Recovery.
Lead Spike Recovery. . .
Mercury Spike Recovery .
Silver Spike Recovery. .
Arsenic Spike Recovery .
Cadnium Spike Recovery .
Selenium Spike Recovery.
Response by Undiluted EP
. . 62
. . ,/ 64
1
. . i 65
' 66
. . ' 67
. . 6,7
. . 68
. . 68
. . 69
Extracts of Pond 0 Samples Collected November 1979 70
vii
-------
ACKNOWLEDGEMENTS
We wish to acknowledge the efforts of D. C. Hemphill, L. E. Holboke, and
R. G. Seals in the early phases of this study. A special thanks to G. L.
Bratten, for his extra efforts in providing much of the analytical data upon
which this report is based, and to the Northrop Services Staff for their
collection and extraction of waste samples.
viii
-------
INTRODUCTION
BACKGROUND
The rapid technological advances over the past several decades have
significantly improved the American economy and lifestyle. However, improper
disposal of hazardous wastes generated by industry as a result of these
advances has created a hazard to both human health and the environment.
Congress recognized this problem, and in October 1976 enacted legislation -
the Solid Waste Disposal Act as amended by the Resource Conservation and
Recovery Act (RCRA) of 1976 (and its amendments) - to control the transporta-
tion and management of hazardous waste. Under the authority of this
legislation the U.S. Environmental Protection Agency (EPA) issued proposed
regulations for the identification, transportation, and treatment, storage and
disposal of hazardous waste (Federal Register. Vol. 43, No. 243, Dec. 18,
1978, pp. 58946-59028).
The regulations proposed (and subsequently published, FR 1980) under
Section 3001 of the RCRA provide specific procedures for sampling, extraction,
and analysis of wastes to identify those wastes which are hazardous due to the
presence of Teachable toxic components. Previous studies (in some cases with
wastes of unknown history) have demonstrated the utility and validity of these
methods. However, the EPA felt that additional studies with wastes from known
industrial sources were warranted to better define the reliability and
reproducibility of the proposed procedures. The EPA also recognized that a
strong quality assurance program was required to assure, through
standardization and quality control, that valid and defensible data are
produced in response to the requirements in the regulations. This study was
initiated in April 1979 to support these EPA requirements for the promulgation
and enforcement of the hazardous waste regulations.
The objectives of this study were to:
• Evaluate the sampling, extraction, and analytical
procedures described in the proposed regulations and determine
their reproducibility and, in the case of the analytical
procedures, their accuracy when used for identification of
hazardous waste.
• Evaluate the application of the proposed extraction procedure
to municipal sewage sludge and evaluate candidate extractor types
for use with the proposed extraction procedure.
Efforts during FY 79 and early FY 80 were concentrated in two major task
areas.
-------
Task 1. Evaluation of the Proposed Sampling Procedures. This
effort was designed to evaluate the safety, reliability and
reproducibility of the sampling procedures in the proposed Section
3001 regulations. The COLIWASA and pond sampler methodologies
identified in the EPA report "Samplers and Sampling Procedures for
Hazardous Waste Streams" (de Vera et al., 1980) were used to collect
waste samples at typical waste sites. Any problems encountered were
noted for later review and analysis. The waste samples were analyzed
by physical and chemical methods to determine the reproducibility of
the sampling procedures. The procedures to be evaluated and wastes to
be sampled were determined jointly by the Office of Solid Waste (OSW)
and the Environmental Monitoring Systems Laboratory-Las Vegas
(EMSL/LV). In some cases modifications to the sampling procedures
were developed and evaluated.
Task 2. Evaluation of the Proposed Extraction Procedure (EP) and
Analytical Methods. This effort was designed to evaluate the
reliability and reproducibility of the Extraction Procedure (EP)
described in the proposed Section 3001 regulations (Section
250.13(d)). It was also designed to determine the accuracy and
precision of the analytical methods that were proposed for use with
the EP using a variety of wastes. The wastes used in the evaluations
were determined jointly by the OSW and the EMSL/LV. As an outgrowth
of this work modifications to the EP and analytical methods have been
recommended to the OSW and are reflected in the methods published in
"Test Methods for Evaluating Solid Waste - Physical/Chemical Methods"
(EPA-SW-846, 1980).
-------
CONCLUSIONS AND RECOMMENDATIONS
The following conclusions and recommendations are based upon data
presented and observations made during the study period:
• The method developed and standardized for sampling ponded
wastes provided representative composite samples of the ponded
wastes tested. "Pre-sampling" is recommended to estimate the
heterogeneity of the waste and determine the number of samples or
composites that should be taken.
• The Composite Liquid Waste Sampler (COLIWASA) provided reproducible
samples from drums of the liquid wastes tested. Present design of
the COLIWASA prevents adequate sampling of the bottom-most layer in
drums or tanks. Here again, pre-sampling is recommended where
practical. The alternative sampler design proposed (DAT Sampler)
should be evaluated with liquid wastes in drums, tanks, or vacuum
trucks.
• In intralaboratory studies, the proposed Extraction Procedure was
found to be reproducible (RSD < + 15% for the waste types sampled).
However, inter!aboratory studies indicate that clear, detailed,
step-by-step protocols are needed to eliminate the possibility of
misinterpretation or substitution of non-equivalant procedural
elements. The EP should be evaluated to determine its applicability
to oily or solvent-containing waste samples.
• A problem with contamination of acidified extracts by barium leached
from glass-fiber prefilters was identified. Until prefilters are
identified which do not contribute barium to the filtrate, it is
recommended that a 100-ml portion of 0.1N acetic acid precede each
waste sample through the prefilter and be stored for possible future
use in determining blank correction for the sample. It is
anticipated that such blank correction would only be used in the
event that the barium levels in the samples exceed the criteria
level for hazardous waste identification by the EP Toxicity
Characteristic.
• Intra- and interlaboratory studies indicate that atomic absorption
spectroscopy is an accurate and highly reproducible method for
analysis of most inorganic components of waste extracts.
• Some problems in the analyses of Ba and Hg remain to be resolved.
The method of additions is recommended to provide interference
correction. Extracts very high in dissolved solids concentrations
-------
may cause build-ups of material which alter the flame characteristic
characteristics of the atomic absorption spectrometer. Such samples
should be diluted prior to analysis.
The three extractor types compared, i.e., the blade-type, NBS
tumbler-type, and the wrist-arm shaker, provide comparable waste
extracts when used in the Extraction Procedure (EP).
When possible, EP extracts should be analyzed immediately, as some
waste extracts are not stable over a period of hours or days.
Applicability of the EP toxicity criterion for mercury should be
reevaluated, as low EP recoveries may be misleading with respect to
the toxicity hazard presented by wastes containing high levels of Hg
in temporarily insoluble forms. Inadequate oxidation of mercury-
containing EP extracts should be investigated to explain low mercury
recoveries.
-------
METHODS
SELECTION OF WASTES
The wastes and sites used In this study were selected with the active
assistance of industrial and government facilities that generate a variety of
both hazardous and non-hazardous wastes. The criteria for selecting a waste
were initially determined by the ultimate use of the waste sample. For
evaluating the sampling protocols it was determined that ideally the samples
should have the following characteristics:
1. be heterogeneous,
2. be fluid (pourable at room temperature, 20°C),
3. be accessible to sampling using a liquid core sampler
less than 10 feet long,
4. be available in sufficient quantity to allow at least
20 one-liter samples to be withdrawn without appreciably
decreasing the amount of waste remaining (i.e., at least
800 liters), and
5. contain one or more components which can be used as
indicators of whether or not a series of aliquots of these
samples are equivalent.
For evaluating the EP, it was determined that ideally the samples should have
the following characteristics:
1. be divisible into subsamples of 100-g size
without introducing significant variability due to the
subsampling procedure,
2. contain one or more components which can be used as indicators
to determine whether a series of repetitive extractions, each
one performed on a new subsample of the test material, gives
equivalent indicator concentrations 1n the extracts,
3. contain approximately 25% w/w solids (i.e., separable by
filtration and/or centrifugation), and
4. be available in a quantity sufficient to yield at least
5 kg of solids.
-------
For evaluating the analytical procedures it was determined that the samples
should have the following characteristics:
1. be from typical waste streams that are complex
in nature,
2. contain one or more of the materials listed in 250.13(d)
(1) (43 FR 58956),
3. be divisible into subsamples of 100 g size without
introducing significant variability due to the subsampling
procedures, and be available in quantities of at least 5 kg.
In many cases it was difficult to meet these criteria, either because the
necessary information on the wastes was not available or no wastes with the
desired characteristics were available from the facilities visited. Due to
their concerns about the proprietary nature of the industrial processes that
produced the wastes, many of the facility operators were hesitant to provide
more than minimal information about the waste streams sampled. The wastes
used in the study were selected to represent the most difficult materials for
testing the sampling procedures, the EP, and the analytical procedures (i.e.,
they represented worst-case conditions for each procedure).
COLLECTION OF SAMPLES
During this phase of the study, eleven manufacturing and waste disposal
sites were visited and 27 different wastes were sampled (Table 1). Brief
descriptions of each site and waste follow:
Site A. This waste disposal facility segregates its wastes by type of
industry. The liquid wastes are placed in open ponds where waste volume is
reduced by natural evaporation. On rare occasions, there may be exchange of
waste materials between adjacent ponds. Four ponds were sampled with the pond
sampler at this site. Two of the ponds contained a liquid that the site
operator identified as titanium dioxide process waste. The samples from these
ponds were acidic (pH <1) and contained approximately 1% solids. The samples
were brownish-green and could be separated into layers of an oily aqueous
liquid and a dark-grey fine solid. The third pond was identified by the
operator as an alkaline waste; however, the pH of the samples collected from
different locations on the pond ranged from 2.3 to 7.7. These samples could
be separated into layers of a greenish-yellow aqueous liquid and a light brown
solid (approximately 6% solids). The fourth pond contained a waste identified
by the operator as sulfonation tars. The samples collected could not be
filtered by the proposed filtration procedure; however, they could be
separated by the proposed centrifugation procedure into four layers (a thin
dark-brown oil layer, a non-aqueous liquid layer, an aqueous layer and a very
thin layer of solids). The fifth pond was another alkaline waste (pH > 10 for
all composite samples) with a tea-colored aqueous phase, a brown-orange solid
phase, and a brownish-black tarry layer that appeared in some of the samples.
-------
TABLE 1. SITES SAMPLED DURING FY 79: IDENTIFICATION OF SITES, FACILITY
FUNCTIONS, WASTE STREAMS SAMPLED, AND SAMPLING METHOD
Site Function of Facility
Waste Stream Sampled
Sampler
A Waste disposal
Ponds of Ti02 process waste Pond sampler
Ponds of alkaline waste Pond sampler
Pond of sulfonation tars Pond sampler
B Paint manufacture
C Chemical manufacture
Drum of paint sludge COLIWASA
Drum of laboratory wastewater COLIWASA
Bags of pesticide waste N.A.
D Chemical manufacture
Inlet, grit chamber, and pond Pond sampler
of an API oil separator
Dumpster of chromate oxidation Pond sampler
paste
E Steel manufacture
Waste dust pile from electric Shovel
furnace baghouse
Filter cake waste from blast Gloved hand
furnace scrubber
Waste roll-mill scale pile Plastic bottle
from water treatment plant
Tank truck of lime sludge from Shovel
ammonia still
F Chemical manufacture
Alkyl chloride storage pit Pond sampler
Epichlorohydrin waste sump Pond sampler
Polymerized epichlorohydrin Pond sampler
waste pits
G Chemical manufacture
Filter cake from chlorine/
mercury process stream
Trowel
(continued)
-------
TABLE 1. (Continued)
Site Function of Facility
Waste Stream Sampled
Sampler
Dumpster of asbestos waste, Trowel
clean-up from chlorine process
Dry waste solids from chlorine/ Trowel
mercury process
H Petrochemical manufacture CPI* decant pit
Activated biosludge
Pond sampler
Plastic bottle
I Chlorine manufacture
J Chemical manufacture
Waste chlorine sludge pile Trowel
Industrial sewage filter cake Trowel
from a truck
Tank of chromate reduction Bottle
clarifier underflow
Drum of catalyst fines Glass jar
K Paint removal/
electroplating
Drums of pi ating'waste COLIWASA
identified as tin-lead solution
Drums of alkaline rust remover (red) COLIWASA
Drums of oil/water organic solvent COLIWASA
mixture
* CPI = Chemical Production Industries
Site B. Waste sludge from the solvent recovery operation was sampled at
this paint production facility. A COLIWASA was used to obtain the samples
from a 55 gallon drum. The samples were multicolored and viscous, had an odor
typical of oil-based paint solvents.
Site C. Two wastes were sampled at this chemical manufacturing facility.
One was an acidic (pH =* 1) laboratory waste resulting from COD and other
wastewater analyses performed in the facility's laboratory. The waste sample,
collected from a 55 gallon drum with a COLIWASA, had a brown organic layer and
-------
a dark-orange-brown aqueous layer. The second waste was a solid material
(composite of excess quality control samples contained in plastic bags) from
the production of a urea herbicide.
Site D. Samples were collected with a pond sampler from an API oil/water
separator. The pond sampler was also used to scoop out samples of chromate
oxidation paste from a dumpster at this chemical manufacturing facility. The
samples collected from the API separator contained oily dark-brown solids
mixed with water (pH = 8). The samples of chromate oxidation paste contained
both a clear liquid layer (pH - 7.5) and a layer of brown solids
(approximately 10% solids).
Site E. Four wastes were sampled at this steel manufacturing facility.
(1) A waste-dust pile from an electric furnace baghouse was sampled with a
shovel. The sample was a dark-brown mixture of fine and coarse solid material
that had a light fluffy texture. (2) Filter-cake waste from a blast furnace
scrubber was sampled with a gloved hand. The sample was a black paste that
appeared to contain a small amount of water. (3) A pile of waste roll -mi 11
scale, from one of the facility's water treatment plants; samples were scooped
directly from the surface of the pile with a wide-mouth plastic bottle. The
sample was a mixture of crystalline solids (large and small pieces) that had a
disagreeable odor. (4) Lime sludge from an ammonia still was the final waste
sampled at this site. This sample, taken with a shovel, was a light-brown
mixture of solids in an alkaline liquid (pH = 11.6).
Site F. Three wastes were sampled with a pond sampler at this organic
chemical manufacturing facility. (1) The first sample, collected from an
alkyl chloride storage pit, was a rust-brown liquid (pH = 7) containing
suspended solids. (2) The second sample, collected from an epichlorohydrin
waste sump, contained two layers, a liquid (pH = 9.7) and a gray solid.
According to the plant operator, this waste was a mixture of caustic solids,
phenols, and epichlorohydrin. (3) The remaining samples were taken from each
of two pits of polymerized epichlorohydrin (epoxy resin). These samples
contained sandy white solids in an alkaline aqueous liquid
Site G. Three samples were collected from a chlorine manufacturing plant
which employs the mercury cell process. These samples were obtained using a
small trowel. The filter cake cell sample consisted of two phases, water (pH
= 5.6) and a light-brown solid. The second waste contained asbestos solids in
an alkaline aqueous liquid (pH = 9.8). Chlorine-mercury process-stream waste
(dry) was sampled on a subsequent visit. This sample consisted of chunks of
dry waste with three distinct layers (thin white and tan layers; bulk was gray
in color).
Site H. An activated biosludge and a waste from a CPI (Chemical
Production Industries) decant pit were sampled at this petrochemical facility.
The biosludge sample (taken from a faucet in the pipeline from the settling
tank) was a brown liquid (pH = 6.4) that contained a high concentration of
suspended solids. The decant pit waste was from an adjacent oil refinery
process stream. It was sampled with a pond sampler and yielded a black, oily
liquid sample that had a disagreeable odor. The high concentration of
organics in this sample prevented measurement of its pH.
-------
Site I. A waste pile of chlorine process sludge was sampled with a small
trowel at this chlorine manufacturing facility. The waste was a dark gray,
paste-like solid.
Site J. Three wastes were sampled at this chemical manufacturing
facility. (1) A sample of an industrial sewage filter cake was obtained from
an open truck with a small trowel. (2) A sample (pH = 8.7) of liquid waste
from a chromate reduction clarifier underflow was obtained from a faucet in
the pipeline leading from the settling tank to a treatment plant. (3) The
third waste sample was obtained (with a glass jar used as a scoop) from a drum
of catalyst fines used in a proprietary chemical process.
Site K. Three wastes from a paint stripping and electroplating operation
were sampled at this facility. The wastes were stored in drums and were
awaiting disposal by a commercial disposal company. The samples were obtained
with the COLIWASA. (1) A plating waste identified as a tin-lead solution,
yielded greenish-yellow acidic (pH <1) aqueous samples that contined some
suspended solids. (2) An alkaline rust remover waste, identified as "red"
because of its color, yielded dark red alkaline (pH =12) samples. (3) The
samples obtained from an oil/water/solvent waste had three phases - an oily
surface layer, a tea-colored aqueous layer (pH ^ 6.5) and a solid layer.
The considerable time and effort expended in locating facilities having
suitable wastes and in obtaining approval from those companies to obtain
samples delayed initiation of the laboratory experimental efforts. In several
cases the industrial operators were concerned about the proprietary nature of
their process and required a confidentiality agreement from the EPA contractor
who was actually collecting the samples. Corporate approval of such
agreements resulted in considerable delay and disrupted the sampling schedule.
SHIPMENT AND STORAGE OF SAMPLES
The samples collected in the field were sealed in glass jars (caps had
Teflon inserts). The jars were then sealed in individual plastic bags which
in turn were packed in a specially constructed shipping container. In the
shipping container, the jars were separated by dividers and cushioned with
vermiculite packing material. The containers were locked and shipped (under
chain-of-custody) via either Federal Air Express or truck to the laboratory in
Las Vegas, NV. The samples were stored at room temperature (21° -27°C) until
use. No steps were taken to preserve the samples.
SAMPLING PROCEDURE EVALUATION
Rationale
In designing the sampling procedure evaluations, three basic guidelines
were followed:
10
-------
1. The purpose of the study is to evaluate the"methods and tools
that were specifically developed for this regulatory program (i.e.,
the pond sampler and COLIWASA).
2. The wastes to be used should be among those that are the most
difficult to sample (e.g., a worst-case situation such as a
multiphase waste that contains immiscible liquids and solids of
differing density and particle size).
3. The samples should be obtained by personnel who are not knowledgeable
about the variability or detailed physical or chemical
characteristics of the specific wastes in order to simulate
non-expert use of the sampling methodology.
The sampling methods in Appendix I of the proposed regulations included
both ASTM methods and protocols from the EPA report "Samplers and Sampling
Procedures for Hazardous Uaste Streams" (deVera et al., 1980). It was decided
not to evaluate the ASTM procedures because the Agency has confidence that
those standard procedures have undergone sufficient evaluation through general
use with materials of the type indicated in Appendix I of the proposed
regulations. However, neither the protocols described in the report nor the
pond sampler and COLIWASA had been evaluated for sampling wastes under field
conditions. They were therefore selected for evaluation in this study. The
samplers were used in accordance with the protocols described in the
aforementioned EPA report and are described in Appendix 1 of this report.
This study was designed to determine the ability of these procedures to obtain
representative samples from waste sources. It should be emphasized that the
procedures were not being tested for use in sampling a waste over a period of
time (i.e., to determine if the waste source changes with time).
In developing a method for sampling wastes from pits, ponds, or lagoons,
a number of factors must be considered:
• The method should provide as representative a sample of the
waste as practical, assuming the fact that the waste may be
extremely heterogeneous, present in great volume, or largely
inaccessible to the sampling personnel.
• The method should be simple to use, reproducible in its
application, and statistically sound.
• The method should have broad application in its basic form, but
also be sufficiently flexible to allow for modification when,
in the judgment of the on-site sampling personnel, a waste is
particularly heterogeneous, the distribution of visible waste
components is non-random, or the morphology of the waste site
is complex.
• The method should be so ordered as to ultimately provide the
data analyst with a logical data set readily amenable to
analysis by accepted statistical techniques.
11
-------
As part of this study, a method was developed for collecting represen-
tative samples of ponded wastes. This method combines a systematic sampling
program with the use of the pond sampler. In order to evaluate the ability of
this method to provide samples which are representative of the waste, the
reproducibility of the method was determined using two different ponded
wastes. Since it is impossible to analyze a whole pond so as to obtain a
"true" value for the pond composition, the approach selected used the concept
that if replicate composite samples collected from a pond are the same then
the individual composite samples could be considered to be representative of
the whole.
Experimental Design
The experimental design for the sampler evaluations was developed with
the recognition that it is very difficult to obtain representative samples
from a heterogeneous waste source, especially in the case of large disposal
ponds or pits. A statistical approach must be taken to obtain and analyze the
minimum number of samples that are required to assure, with a stated degree of
confidence, that the sample data are representative of the waste source.
The initial experimental design was based on a one-sided parametric test
that assumed a 4% significant deviation. This design required 39 samples from
the source (i.e., 39 samples/pond) to yield data with 95% confidence of
avoiding Type I and Type II errors. (Type I errors, i.e., rejecting the
hypothesis of no difference between means of sample populations when, in fact,
no difference exists, are minimized by setting the critical probability level
for chance differences very low, e.g., 5% or even 1%. Type II errors, i.e.,
accepting the hypothesis of no difference between means of sample populations
when, in fact, a real difference exists are minimized by increasing the sample
size and hence the discrimination of the test.) This approach was later
modified to provide the appropriate number of samples required for a
hierarchical (nested) analysis of variance (ANOVA) and to define the sources
of variability present in the sequence of sampling and analysis.
Table 2 identifies the waste and regimen selected for evaluation of the
pond sampler and the COLIWASA. The pond sampler was used to sample five
different waste sources at two sites, and the COLIWASA was used to sample five
different waste sources at three sites. The overall strategy of this effort
was to determine the ability of the sampling procedures to collect repro-
ducible samples from a given waste source. No attempt was made to determine
the changes that occur if a waste is sampled over a given period of time (e.g.
comparisons of samples from a single source sampled on different dates).
Percent non-filterable solids, pH, and aqueousrnon-aqueous composition
were initially selected for testing the reproducibility of the sampling
procedures because these parameters are easy to determine and represent sample
characteristics that affect the properties of the waste. For example, while
the elemental concentration in each phase may not change with location, the
concentration observed with the EP will change if the relative quantity of
each phase in the sample changes. Non-filterable solids content was
considered a particularly important measurement parameter for determining the
reproducibility of samples obtained with the pond sampler, because
12
-------
TABLE 2. WASTES SAMPLED USING THE POND SAMPLER AND COLIWASA
Location
Waste
Sampling Regimen
Site A
POND SAMPLER
Ti02 process waste
Alkaline waste
Sulfonation tars
2 ponds, 10 samples/pond
1 pond (revisited), 12
composite samples
1 pond, 10 samples
1 pond, 12 composite
samples
1 pond, 10 samples
Site F
Polymerized epichlorohydrin
waste
2 pits, 20 samples/pit
Site C
Site D
Site K
This waste
COLIWASA
Laboratory wastewater
API separator waste
Plating waste, tin/lead
Alkaline rust remover
*0il /water/sol vent waste
not analyzed during Phase I.
4 drums,
15 drums
3 drums,
3 drums,
3 drums,
3 samples/drum
, 3 samples/drum
3 samples/drum
3 samples/drum
3 samples/drum
considerable operator judgement is required to obtain uniform samples from the
sediment-water interface. However, wet-weight measurements of solids -
routinely used in the EP to determine whether a minimum percent solids level
is exceeded - were not considered precise enough to use in determining sample
uniformity.
The pH of the aqueous phase of each liquid sample was measured with a
laboratory pH meter. The pH meter was calibrated with standard buffers at
pH 4, 7, and 10 just prior to the measurements and rechecked after the
measurements were completed. Percent non-filterable solids were determined in
accordance with the "Non-filterable Residue Method 160.2," Methods for
Chemical Analysis of Water and Wastes (EPA 1979). Duplicate or triplicate
(when sufficient sample volume was available) aliquots of each sample were
analyzed to determine the relative magnitude of variability that can be
attributed to the laboratory analytical procedures.
13
-------
Pond Sampler--
A sampling plan was developed -- for evaluating reproducibility of the
"pond sampler" -- based upon selecting random sampling locations (within reach
from accessible shoreline areas of the waste pond sampled) and combining
samples from five such random locations to form field-composite samples.
Twelve such composite samples from each of two ponds, identified as Pond 0 and
Pond 12 (i.e., a total of 60 sample locations in each pond), were obtained by
the following procedures:
• The accessible shoreline of the pond to be sampled was divided into
six equal stretches (i.e., the shoreline of 300 meters was divided
into six stretches of 50 meters each) and marked with stakes.
• Each of the stretches was marked off into three equal segments (the
resulting segments in our study were 16-2/3 meters in length)
(Figure 1.)
• Three "zones" were established within each segment (nine zones per
stretch — shoreline to 1 meter; 1 meter to 2 meters; and 2 meters
to 3 meters from shore).
• To determine sampling locations, a random series of five two-digit
numbers was assigned for each field-composited sample. Each number
denoted the stretch (first digit = 0 to 5) and zone (second digit =
1 to 9) from which a 600 ml "pond sample" was dipped (see Figure 1).
The five 600-ml subsamples were composited in a clean 1-gallon jar
and sealed with a Teflon-lined screw cap.
Example:
3 meters
2 meters
1 meter
Shoreline
Random number series for
composite sample 1 is 08, 34, (27
55, 39. The number 27 calls
for taking the third sample (of
five to be composited) from
Stretch 2, Zone 7.
Stretch
Figure 1. Sampling grid show ing site selection using the random
two-digit number scheme for field composited samples.
14
-------
• A total of 60 subsamples was collected to produce the 12 composite
samples.
• Using the sampling "by-the-numbers" method, a case of four 1-gallon
jars was filled during each of three circuits around the pond.
The sampling regime could be readily followed by relatively inexperienced
sampling personnel.
COLIWASA--
Reproducibility of sampling from drums with the COLIWASA was evaluated
using two different types of waste. The first waste consisted of 15 drums of
a liquid material collected from an API separator. Three samples were taken
from each of the 15 drums. The percent non-filterable solid content of each
sample was the average obtained by analyzing three aliquots from each sample.
Sample-to-sample variability was used to estimate the reproducibility of the
sampling procedure. The second waste was a biphasic oil/water mixture with no
non-filterable solids. Three samples were taken from each of three drums, and
the oil:water ratio determined on three aliquots from each sample. The
average heights of the oil and water phases were determined and an oil:water
ratio calculated for each aliquot. The average of the three aliquots was used
in the reproducibility determination. The reproducibility of the sampling was
thereby determined by analyzing the sample-to-sample variability of the
oil:water ratio.
EXTRACTION PROCEDURE (EP) EVALUATION
Rationale
The Extraction Procedure (EP) is a key parameter in the screening
mechanism designed to identify those wastes that require special management
because of their potential to cause harm to human health. It must be
emphasized that the EP is not intended to identify the total concentration of
any toxic contaminant in the waste, but rather the concentration of the
toxicant that could occur in groundwater below the disposal site as a
consequence of mismanagement. The approach taken to evaluate the EP was
designed to:
1. determine the reproducibility of the EP described in the proposed
regulations (43 FR 58956),
2. determine if the procedure, as written in the proposed regulations,
is explicit enough for use by non-experienced personnel to produce
valid data,
3. determine the equivalency of a variety of extractors that could be
used with the EP,
15
-------
4. determine if the filtration and centrifugation methods are suitable
alternatives for liquid-solid separation,
5. gain additional experience with the methods as background for
preparation of guidelines manuals to assist those who will use the
EP.
Experimental Design
The Extraction Procedure was used in essential accordance with the
description given in the proposed regulations (see Appendix 2). A flow chart
of the sample treatment followed for the EP is shown in Figure 2. As shown,
triplicate aliquots (minimum of 100 g each) are obtained from a stirred waste
sample and are separated into solid and liquid phases by either filtration or
centrifugation. The liquid phase is stored under refrigeration until the
solid phase has been extracted. The solid phase is then weighed without
drying, placed in a suitable extraction apparatus along with a volume of
deionized water equal to 16 times the weight of the wet solid phase and
agitated. For slurries with an initial pH > 5, the pH of the slurry is
continuously adjusted to 5.0 ± 0.2 with 0.5 N acetic acid during agitation.
However, the maximum amount of acid that is added during the EP is 4 ml per
gram of solid phase, even if the pH of the solution does not reach 5.0 ± 0.2.
For samples with an initial slurry pH of <5, extraction proceeds without
addition of acid. Agitation is continued for 24 hours. The slurry is then
filtered and any solid material is discarded. The filtrate is adjusted with
deionized water to a volume equal to 20 times that occupied by water at 4°C
equal in weight to the solid phase added to the extractor. Note that the
procedure described in the proposed regulations calls for this volume
adjustment prior to filtration. Differences in the resulting volume are
negligible by the two methods (maximum worst-case volume difference is 5%).
This solution is then added to the original liquid phase to produce the
extract. The EP extract is split into two samples; one is acidified to
preserve it for elemental analysis and the other is stored under refrigeration
for organic analysis.
The EP was evaluated with wastes collected from three different ponds.
Each waste was extracted at least once and screened using atomic emission
spectroscopy to identify the major extractable inorganic components that might
be used for evaluation of the EP. Additional wastes were extracted in a study
comparing extractor types (see under Special Studies).
Triplicate aliquots of each sample were extracted as described in
Appendix 2. The extracts were analyzed by standard atomic absorption
spectroscopy methods described in the latest edition of Methods for Chemical
Analysis of Water and Wastes, EPA-600/4-79-020 (this edition supersedes the
1974 edition referenced in the proposed regulations). The results were
averaged to determine a mean and the standard deviation for the triplicate
analyses. The relative standard deviation of the average was then used as a
measure of the reproducibility of the EP.
16
-------
WASTE SAMPLE (stirred)
TRIPLICATE ALIQUOTS
FILTRATION/CENTRIFUGAT10N
SOLID PHASE (weighed)
moist
'/Add deionized H20
116 x weight of non-J
\filterable solids
EXTRACTOR
AGITATION
/,
Adjust and maintain pH at 5.0 t 0.2)
w/ 0.5 N Acetic Acid
Max. acid = 4 ml/g non-filterable
\solids
FILTRATION
SOLID PHASE-
DISPOSAL
—LIQUID PHASE
•LIQUID PHASE
STORE AT 1-5°C
DILUTION
/ Distilled H
(total vol. = — --
\pf non-filterable
^EXTRACTION PROCEDURE (EP) EXTRACT
"2° to \ .'TT
20 x weight) SPLIT
•able solids/
ACIDIFICATION—^-STORAGE (at 1-5'C)
(pH< 2 w/HN03)
ELEMENTAL ANALYSIS ORGANIC ANALYSIS
17
-------
Special Study - Comparison of Three Extractor Types
Rationale--
The rotary stirrer extractor was compared with two other extraction
apparatus to evaluate the variability of these methods with respect to various
waste types. It was anticipated that not all waste types would be adequately
extracted using the proposed blade-type extractor. Tumbling-type extractors
-- provided to us by the National Bureau of Standards -- and wrist-arm shakers
were compared with the blade-type extractions in this special study.
Experimental Design-
Three sample types were extracted with each extraction method. The
samples averaged 100 percent solids, 8.7 percent solids, and 2.5 percent
solids. Twelve composite samples of each waste type were tested. Nine equal
aliquots were taken from each sample; three aliquots were then extracted with
each type of extractor. Thus, 108 separate extractions per waste were
completed. The quantitative analysis of metals was done in duplicate on these
extracts. The statistical program then compared the significance of the
variability among the samples, the extractors, and the analytical
concentrations of the various metals.
The model associated with this experiment is:
Yijkl = " + Si +
Where: u represents a general mean
Si i=l,2,3,....12, represents the effect of "samples". This
factor is a random effect factor. (Instead of having 12
samples, we could have 10, or whatever.)
tj/i J=1.2,3, represents the effect of "techniques". This
factor is of fixed effect because we are specifically
interested in these three (3) techniques.
ak/i/j k=1»2»3» represents the effect of "aliquots". This factor
is of random effect.
ijkl 1=1»2» represents the variation of the analytical
measurements.
The above design for sample and data collection is a hierarchical or "nested"
design. In this design the effect of each one of the factors can be tested
through the F-Test of the ANOVA table. One of the most important factors to
be tested is the techniques. When, from the ANOVA table, a significant
difference between techniques was found, i.e., there was a nonzero effect of
the factor called "techniques," the Student-Newman-Keuls (SNK) multiple
comparison was used to decide which technique was the best and second best.
The SNK procedure is as follows:
18
-------
First, order the averages from each technique from smaller to larger.
Here xj, is the average for the first technique, and xjoi
xj3 are the averages for the second and third techniques,
respectively. To work through an example, sample data are ordered
from smaller to larger:
0.1661; XT2 = 0.2331; xTj = 0.3132
2. Then compare XT, with xj, (smallest with largest) using:
qa,3,12 = *T1 " *T3
Where: MS = mean square (from analysis of variance) for aliquots within
techniques within samples (use value of 0.0107 for the MS in
this sample exercise).
q* = the q statistic, with Studentized range distribution.
*For additional information on the q statistic and use of the SNK procedure,
see Myers 1972.
a = confidence level selected. (Use 0.025 level for this example
to determine whether the probability -- that differences noted
in technique means can be ascribed to chance alone -- is
greater or less than 2.5 percent.
3,72 = degrees of freedom for the MS
therefore:
q = 0.3132 - 0.1661
J
0.0107
q = 0.1471 . . 92
0.0299 *'s*
3. For testing the hypothesis:
HQ: MTi = HJ3 versus Ha:
take from the Studentized range tables the value of q.025,3,72 = 3.79
4. Comparing the two values, the calculated q (4.92) is larger than the
table value (3.79) inferring that wTj > wi3. We can therefore
19
-------
reject the null hypothesis (H0) that there is no difference
between technique means and conclude that one of the techniques is
better than the other.
5. Continuing, H0: vj^ = ^ versus Ha:
q = 0.0801 _ , K7n
0299 - 2'678
q .025, 3,72 • 3-620
6. Since the calculated q (2.678) is less than the table value (3.620),
the H0 hypothesis cannot be rejected at the level 0.025. This
means that there is not enough evidence to say one technique is
better than the other.
7. If the null hypothesis can be accepted for the extreme values of the
ordered means, no additional comparisons need be made.
The handling and treating of samples for this comparison study were
carried out as in our previous extraction studies, with the following
exceptions:
1. The aliquots for the wrist-arm shaker had to be further divided due
to the small flask volume. Larger flasks were found to be
impractical since they tended to break during shaking when filled to
capacity. The addition of acetic acid and monitoring of pH was more
difficult since the aliquot fractions in the small flasks did not
necessarily require equal volumes of acid to maintain the acceptable
pH range.
2. When extracting solids from samples with low solid (e.g., <2%)
contents, the amount of liquid added to the solids from a 100-g
sample was frequently so small as to be adsorbed on the filter paper
or evaporated during the 24-hour extraction time. Therefore, for
several aliquots there was no extraction filtrate to add back to the
original sample filtrate. This evaporation effect was most
pronounced with the blade-type extractor. If final volume
adjustment had been made prior to the final filtration step (per
proposed method) this problem might have been avoided.
3. The combined volume of 100-gram subsample of biosolids — from
high-bulk solid samples -- and the required extractant exceeded the
1/2-gallon capacity of the NBS extractor bottles. Eighty-gram
subsamples proved to be the working maximum with waste of this type
and were therefore adopted for extraction of the remaining eleven
samples with the NBS extractor.
The EMSL/LV Hazardous Waste Laboratory has two NBS tumbling-type extrac-
tors. Each bottle holds four 1.9 liter (1/2 gallon) round plastic Nalgene
bottles. These bottles are held in an aluminum frame by a collar which is
tightened with Allen screws. A low-speed, high-torque motor rotates the frame
through 360 degrees at the rate of 30 revolutions per minute (Figure 3).
20
-------
1/15 hp
Electric Motor
2-liter Plastic or Glass Bottles
Screws for Holding Bottles
Figure 3. Diagram of NBS tumbling-type extractor.
Three standard laboratory wrist-arm shakers — clamped to the lab bench
and holding up to 12 fl-asks each -- were utilized in this study. The
mechanical action simulates human wrist action; it oscillates through an arc
of approximately 40° (20° to each side of vertical) during each cycle. The
rate of agitation was 312 cycles per minute.
Special Study - Background Interferences
Additional experiments were performed to identify any background
interferences that might result from the equipment used in the EP. The
filtration equipment and the extraction apparatus used are made from Type 304
and Type 316 stainless steel, respectively. Since the samples and extracts
are in contact with stainless steel surfaces, there is some concern that use
of such construction materials might result in high background concentrations
of certain metals (especially chromium) in the EP extract. Two groups of
blank samples, one consisting of deionized water and the other of 0.1 N acetic
acid (400 ml of 0.5 N acetic acid to 1600 ml of deionized water), were put
through the EP and analyzed by ICP or AA spectrometry to determine what metals
(and at what levels) might be added to the extracts from the equipment.
Apparatus Abrasion/Contamination Study—
A second series of blanks were run to determine potential contributions
of metals from the extractors under worst case conditions. Conditions for
this study (Table 3) deviated from the standard Extraction Procedure in order
to provide "worst case" conditions. Samples (50 g) of fine, acid-washed
21
-------
"Ottawa Sand" were placed in each of four blade-type stainless steel
extractors and two polyethylene tumbling-type extractors. The polyethylene
extractors served as a blank for the sand since their non-metallic nature
would preclude their contributing measurable metal to the extract.
Fine sand was employed since initial attempts to extract coarse (2-3 mm)
aquarium sand in the blade-type extractors used in our laboratory were
unsuccessful. The extractor blades bound up and could not handle the heavy
load. Even at reduced sample sizes, the blades would catch larger sand grains
and seize. The ability of a particular extractor to handle the specific
materials present in wastes being tested should thus be considered by the
operator when selecting the type of extractor to use.
The fine sand was extracted with dilute nitric acid (pH = 2.5) in two of
the blade-type extractors and one NBS extractor; the sand in two other
blade-type extractors and one other NBS extractor was extracted with a dilute
solution of acetic acid (pH = 5.0). The resulting extracts were split; one-
half of each was filtered, while the other half was not filtered following the
extractions.
The blade-type extractors -- at the specified speed of 40 rpm -- were
unable to suspend the fine sand. Therefore, the blade speed was increased to
80 rpm and the extraction period shortened to 3 hours. The combination of
sand and acidic extractant represented a rigorous test of the ability of the
apparatus to withstand abrasion and dissolution of metallic components.
Special Study - Sewage Sludge Extraction and Analysis
Rationale--
This study was undertaken to evaluate, on sewage sludge samples, the
performance of the toxicity extraction procedure (EP) described in the Federal
Register (Dec. 18, 1978) and the analytical procedures for the determination
of arsenic, barium, cadmium, lead, chromium, and selenium. Samples of the
sludges investigated (Table 4) were sent from the sludge-producing facilities
to both the EMSL/LV and the Municipal Environmental Research Laboratory,
Cincinnati (MERL/Cin) for extraction and analysis. The results of the two
laboratory efforts were then compared to evaluate the interlaboratory
reproducibility of the proposed procedures with sewage sludges. One set of
aliquots of the seven sludge samples was extracted using the EP, and another
set of aliquots was digested. Both the extracts and the digests were analyzed
for the elements of interest.
Experimental Design--
Triplicate aliquots from the seven samples were extracted by the EP.
Another set of triplicate aliquots per sludge was digested with nitric acid
following EPA procedure 4.1.3 in "Methods for Chemical Analysis of Water and
Wastes," EPA-600/4-79-020, March 1979 (EPA 1979).
22
-------
TABLE 3. CONDITIONS OF THE APPARATUS ABRASION/CONTAMINATION STUDY
Material Extracted - Fine "Ottawa" laboratory sand, acid-washed
Extractor Used - Blade-type, stainless steel
- NBS tumbling-type, polyethylene
Extraction Reagents - Nitric acid solution, pH = 2.5
- Acetic acid solution, pH = 5.0
Filtration Apparatus - Nuclepore 142-mm diameter, stainless steel
pressure filtration unit
Extractor Conditions - 3-hour duration at 80 rpm (Blade-type), or
- 3-hour duration at 30 rpm (NBS tumbler)
Extract "splits" - one-half of extract filtered
- one-half of extract unfiltered
Blanks - Extraction reagents, filtered and unfiltered
Analyses - Atomic Absorption Spectrometry, for Ba, Cr and Fe
ANALYTICAL PROCEDURES EVALUATION
Rationale
The analytical procedures proposed for analysis of the EP extracts are
standard methods for analysis of water, wastewater, or industrial effluents in
use by the Agency. Since publication of the proposed regulations in December
1978, the Agency has updated these procedures to take into account the latest
improvements in methodology. While these improved methods have been used for
the analysis of aqueous effluents and liquid wastes, they have not been
extensively tested with samples of industrial "solid waste" or EP extracts of
such waste. DeUalle et al. have reported intra- and interlaboratory data for
several of the methods applied to leachates collected from waste sites.
This phase of the study was therefore designed to determine the accuracy and
reproducibility of these updated methods when used to analyze "solid wastes"
-- which can include aqueous solutions, slurries, sludges, multiphasic waste
materials, etc. — and EP extracts.
23
-------
TABLE 4. MUNICIPAL SEWAGE SLUDGE SAMPLES TESTED
Sample
Number Sample Type Physical Description
1 Combined sludge Homogeneous liquid sludge, dark grey
2 Sludge Moist solid, dark brown
3 Sewage sludge Moist plastic solid, black
4 Primary sludge Heterogeneous sludge, solid and
liquid phases, light grey
5 Digested sludge Homogeneous liquid, black
6 Heat-dried fertilizer Fine granules, dark brown
7 Landfill sludge Solid, large and small lumps,
dark brown
Experimental Design
For the analysis of aqueous wastes and EP extracts, the atomic absorption
(AA) methods (EPA 1979) for arsenic, barium, cadmium, chromium, lead, mercury,
selenium and silver were used. These methods were evaluated -- with standard
solutions and with EP extracts from various wastes containing one or more of
the elements of interest — to determine their accuracy and precision.
Triplicate aliquots of the EP extracts were first analyzed. Aliquots
were then spiked with known concentrations of the elements of interest and re-
analyzed. The mean and standard deviation of the triplicate results obtained
for each spiked sample were used to evaluate the accuracy and reproducibility
of the analytical method. Spike recovery is calculated by dividing the mean
analytical result (less the concentration of that element in the unspiked
sample) by the spike concentration, and multiplying by 100 to obtain percent
recovery. The relative standard deviation is determined by dividing the
standard deviation by the mean; the resulting value is muliplied by 100 to
express as a percentage. Acceptable analytical methods should accurately and
reproducibly indicate the increase in concentrations of the elements in the
spiked samples. Spike recoveries were determined for each element in a number
of sample matrices.
24
-------
RESULTS AND DISCUSSION
SAMPLING PROCEDURES EVALUATION
Pond Sampler
For Initial evaluation of the Pond Sampler, four ponds were sampled at
site A (two samples from each of five locations on each pond). Samples from
two of the ponds, Pond 0 and Pond 13, were analyzed for pH and percent solids.
Four factors contribute to the overall reproducibility observed in the
analytical data obtained with the pond sampler. They are:
1. the homogeneity of the waste in the pond sampled;
2. the effect of the collection of the first sample on subsequent
samples collected at the same location;
3. the reproducibility of the sampling procedure;
4. the precision of the analytical procedure used to determine the
parameter of interest in each sample.
The data obtained with the pond samples (Tables 5 and 8) can be analyzed to
estimate the contribution of these four factors to the overall reproducibility
of the pond sampler. Factor 1, the homogeneity of the waste in the pond, is
reflected in the differences between samples from different locations on the
pond (I.e. differences between the first sample taken at each location -
Tables 6 and 9},. The differences between two samples taken at the same
location (Tables 7 and 10) reflect Factor 3, the reproducibility of the
sampling procedure, and Factor 2, the effect (if any) the collection of the
first sample has on the second sample. The precision of the analytical
procedure, Factor 4, can be estimated by comparing the results from replicate
analyses of the same sample (Tables 6 and 9).
The data for the titanium dioxide process waste from Pond 0 are shown in
Table 5. The procedure for pH required insertion of a pH electrode into the
aqueous phase, whereas the procedure for percent solids required collection of
aliquots from the sample and analysis of those aliquots. Because of the
limited sample volumes, only duplicate aliquots were analyzed for each sample.
The pH of the Pond 0 samples was very low (pH <1), thus the measurements with
the pH meter were not considered reliable enough to identify differences
between locations on the pond. However, the percent solids data can be used
for this purpose. Differences between aliquots from the same sample (Table
5) reflect the reproducibility of the laboratory analytical procedure for
25
-------
TABLE 5. EVALUATION OF POND SAMPLER: RESULTS OF pH AND PERCENT SOLIDS
ANALYSES OF DUPLICATE ALIQUOTS, FROM DUPLICATE SAMPLES OF
WASTES TAKEN AT FIVE LOCATIONS ON POND 0, SITE A
Sample*
1A
IB
2A
2B
3A
PH
0.23
0.23
0.34
0.31
__**
**
0.22
0.17
0.30
0.19
Percent
Solids
1.24
1.26
1.09
1.28
0.82
1.28
2.09
1.28
**
**
Sampl e
3B
4A
4B
5A
SB
pH
0.40
0.46
0.51
0.44
0.37
0.38
0.43
0.43
**
**
Percent
Solids
0.92
1.83
1.54
1.61
__**
1.19
1.74
1.84
1.55
1.80
Number = location on pond; A B = 1st and 2nd sample at that location.
** Data not reported - known discrepancies in analytical procedure.
TABLE 6. POND SAMPLER: MEANS AND STANDARD DEVIATIONS OF DUPLICATE ANALYSES
FOR pH AND PERCENT SOLIDS IN WASTE SAMPLES FROM POND 0, SITE A
PH
First sample
Location
1
2
3
4
5
Average
X
0.23
~
0.25
0.48
0.43
0.35
s(n = 2)
0.00
*
0.08
0.05
0.00
Second sample
X
0.33
0.20
0.43
0.38
--
0.34
s(n = 2)
0.02
0.04
0.04
0.01
__*
Percent
First Sample
X
1.25
1.05
--
1.58
1.79
1.42
s(n = 2)
0.01
0.33
*
0.06
0.07
Solids
Second Sample
X
1.19
1.69
1.38
—
1.68
1.48
s(n = 2)
0.13
0.57
0.64
*
0.18
* Data not reported - known discrepancies in analytical procedure.
26
-------
TABLE 7. POND SAMPLER: MEANS AND STANDARD DEVIATIONS OF DATA THAT
REFLECT DIFFERENCES BETWEEN TWO WASTE SAMPLES TAKEN AT EACH
OF FIVE LOCATIONS ON POND 0, SITE A
Location
1
2
3
4
5
Average
X
0.28
*
0.34
0.43
*
0.35
pH
s(n = 2)
0.07
_-
0.13
0.07
--
Percent
X
1.22
1.37
_-*
_-*
1.74
1.44
Solids
s(n = 2)
0.04
0.45
__
—
0.08
Data not reported - known discrepancies in analytical procedure.
TABLE 8. EVALUATION OF POND SAMPLER: RESULTS OF pH AND PERCENT SOLIDS
ANALYSES OF TRIPLICATE ALIQUOTS, FROM DUPLICATE SAMPLES OF WASTE
TAKEN AT FIVE LOCATIONS ON POND 13, SITE A
Sample*
PH
Solids
Sample
Solids
pH (%}
1A
IB
2A
2B
3A
6.96
7.03
6.99
7.55
7.66
7.57
2.43
2.43
2.43
2.33
2.33
2.34
4.89
4.90
4.91
4.06
4.02
4.31
5.80
5.16
5.98
3.13
3.35
3.26
5.66
5.81
5.77
5.91
6.02
6.05
5.22
3B 5.20
5.76
6.95
4A 7.05
7.04
6.70
4B 6.75
6.70
5.33
5A 5.03
5.40
4.70
5B 4.74
4.67
5.07
4.82
5.03
3.50
3.67
3.20
3.48
3.99
3.79
6.13
5.60
5.76
1.93
2.08
1.75
* Number = location on pond; A and B = 1st and 2nd sample at that location.
27
-------
TABLE 9. EVALUATION OF POND SAMPLER: MEANS AND STANDARD DEVIATIONS OF
TRIPLICATE ANALYSES FOR pH AND PERCENT SOLIDS IN WASTE SAMPLES
TAKEN AT FIVE LOCATIONS ON POND 13, SITE A
pH
First sample
Location
1
2
3
4
5
X
6.99
2.43
4.90
7.01
5.25
s(n
0
0
0
0
0
= 3)
.04
.00
.01
.06
.20
Second sample
X
7.59
2.33
5.23
6.72
4.70
s(n
0
0
0
0
0
= 3)
.06
.01
.03
.03
.04
Percent Solids
First Sample
X
4.13
3.25
5.99
3.46
5.83
s(n
0
0
0
0
0
= 3)
.16
.11
.07
.24
.27
Second Sample
X
5.65
5.75
4.97
3.75
1.92
s(n = 3)
0.43
0.08
0.13
0.26
0.17
Average 5.32 0.06
5.31 0.03
4.53 0.17
4.41 0.21
TABLE 10. POND SAMPLER: MEANS AND STANDARD DEVIATIONS OF DATA THAT
REFLECT DIFFERENCES BETWEEN TWO WASTE SAMPLES TAKEN AT EACH
OF FIVE LOCATIONS ON POND 13, SITE A
PH
Percent Solids
Location
1
2
3
4
5
X
7.29
2.38
5.07
6.87
4.98
s(n = 2)
0.42
0.07
0.23
0.21
0.39
X
4.89
4.50
5.48
3.61
3.88
s(n = 2)
1.07
1.77
0.72
0.21
2.76
Average
5.32
0.26
4.47
1.31
28
-------
determining pH or percent solids. If the standard deviations of the mean for
each location (Table 7) are compared to the standard deviations for the
analyses (Table 6), it appears that the laboratory analytical procedure for
percent solids is largely responsible for the variation in the results. A
hierarchical analysis of variance revealed that the differences between
samples taken at the same location were significant at the 5% level (F10>5 =
6.89). However, the analysis of variance performed on the percent solids data
for Pond 0 indicates no significant differences between locations (F5,^ =
0.95). If the mean value for each field sample (Table 6) is considered as a
single result, the average percent solids for Pond 0 is 1.45 ± 0.27 (n = 8).
Thus, the percent solids results for Pond 0 indicate that the pond sampler can
reproduce samples with a relative standard deviation of ±19%, and that a large
part of the variation in the results is due to the laboratory procedure for
determining percent solids. It is difficult to obtain uniform weights among
waste samples dried under similar conditions, and small weight differences
contributed heavily to the variability among low-solids samples.
The data for the alkaline waste, Pond 13, are shown in Table 8. In this
case, pH and percent solids measurements could both be used to evaluate the
reproducebility of the sampling procedure. Triplicate aliquots of each sample
(two samples from each of five locations on the pond) were analyzed. The pH
measurements were very reproducible for aliquots from the same sample (Table
8) with a relative standard deviation of less than ±4%. The standard devia-
tion for percent solids analyses (Table 9) was similar to that observed for
Pond 0; however, the results are more consistent because of the higher
concentration of solids (relative standard deviation of less than ±8%). When
the standard deviations of the mean for each location (Table 10) are compared
to the standard deviations for the analyses (Table 9), it is evident that the
variation due to analytical procedures is not significant when compared to
variations between two samples taken at the same location. When the mean
value for each field sample (Table 9) is considered as a single result, the
average percent solids for Pond 13 is 4.47 ± 1.38 (n = 10) and the average pH
is 5.31 ± 1.85 (n = 10). The data in Table 10 and the standard deviations for
the average of all Pond 13 samples (relative standard deviations of 31% and
34% for percent solids and pH, respectively) indicate significant differences
between locations on the pond. The differences in pH are especially note-
worthy since they indicate extreme heterogeneity within the pond and also show
that there is little mixing of the aqueous phase of the waste. Data for
percent non-filterable solids (Table 8) also indicate that for this measure-
ment there were significant differences between samples taken at the same
location. This might be expected with a liquid sample that has a significant
solids concentration since the agitation created in taking the first sample
could change the solids concentration observed in the second sample with
little effect on the pH of the aqueous phase.
The results for the samples collected at site A indicate that even with a
very heterogeneous waste the pond sampler has a reproducibility for pH and
percent solids of better than ±35%. The data generally indicate that the
reproducibility of the laboratory procedures for obtaining and analyzing
aliquots from field samples is better than ±8%. However, for samples with low
solids (i.e. < 2%) the reproducibility of the laboratory procedure was not as
good. The standard deviation does not change; however, the relative standard
deviation (standard deviation x 100 * sample mean) changes considerably
29
-------
because of the lower value for percent solids. As expected, the laboratory
analytical procedure for pH was very reproducible (better than ±2%).
Additionally, the data for percent solids indicate that there are differences
between two consecutive samples taken at the same location on each pond.
These differences, especially significant for Pond 13, indicate that for
properties related to the solid/liquid phase composition, the composition of
the second sample can be affected by the collection of the first sample at the
same location. The largest overall source of variability appears to be
between locations on the ponds. Surprisingly, the pH and percent solids data
for Pond 13 showed similar differences between locations on the pond, even
though they had different variabilities between samples taken at the same
location. These results emphasize the fact that waste from sources such as
disposal ponds may be very heterogeneous, and that a number of samples from
different locations on the pond are required to properly represent a waste for
identification (analysis) as hazardous or non-hazardous.
If the two samples collected at each location are treated as independent
samples, two duplicate sets of data can be identified for each pond (i.e. set
of first samples vs. set of second samples). The average value for each set
of data then provides a mathematical composite of the samples for that set.
Comparison of the average (mathematical composite, Table 11) for each set of
samples demonstrates a high degree of overall reproducibility for the pond
sampler. These results indicate that a composite of five samples from dif-
ferent locations on a pond should provide a more reproducible indication of
the pond's composition.
In theory, if replicate composite samples collected from a pond are the
same, then the individual composite samples could be considered to be
representative of the whole.
To determine how well actual composite samples represent the waste in a
pond, the field-composite sampling plan was used to collect 12 composite
samples from each of two ponds, Pond 0 and Pond 12. Pond 0 was selected
because discreet samples from that source had shown large sample-to-sample
variability (Table 12). Pond 12 had not been previously sampled, but was
quite similar to the adjacent Pond P (described previously) in appearance and
function. In fact, it reportedly received periodic overflow from Pond P as
one of its input sources.
Reproducibility
The results of the percent solids determinations (Table 13) for the 216
aliquots (9 aliquots from each of 12 one-gallon samples, from each of 2
ponds) indicate that sample-to-sample variability remains the greatest source
of error in waste analysis. While the cumulative RSD for Pond 12 composite
samples (20.3 %) is lower than we noted with previous discreet pond sample
comparisons, composite samples from the highly heterogeneous Pond 0 still
showed considerable variability (RSD = 53.3 percent). It should be noted that
direct comparisons of sampling variability between Pond 0 discreet samples
(based upon June 1979 sampling) and the composite samples (obtained from the
same pond in November 1979) are misleading. The chemistries of the wastes
30
-------
TABLE 11. REPRODUCIBILITY OF POND SAMPLER: MATHEMATICAL COMPOSITE
OF FIRST AND SECOND SAMPLES FROM EACH LOCATION
Pond 0 Pond 13
Sample 1 Sample 1 Sample 1 Sample 2
Data Composite Composite Composite Composite
Mean pH 0.35
s of mean
RSD of mean
Mean % solids 1.42
s of mean
RSD of mean
Average RSD = 1.7
0.34
0.007
2.02
1.48
0.04
2.76
5.32
0.007
0.13
4.53
0.08
1.90
5.31
4.41
TABLE 12. AVERAGE RELATIVE STANDARD DEVIATIONS (RSD'S)
FOR VARIOUS LEVELS OF SAMPLING
RSD (%)
pH Non-filterable Solids (%)
Sampling Pond 0 Pond 13 Avg. Pond 0 Pond 13 Avg.
Differences between aliquots
of the same sample 10.1 0.9 5.5 17.6 4.7 11.2
Differences between duplicate
samples taken at each location 26.5 4.8 15.4 13.6 30.3 22.0
Differences between locations
on pond 21.5 36.5 29.0 19.9 16.9 18.4
31
-------
TABLE 13. EVALUATION OF THE REPRODUCIBILITY OF COMPOSITE
SAMPLES OBTAINED WITH A POND SAMPLER - NON-FILTERABLE
SOLIDS (%)
Sample Number Pond 12 Pond 0
1
2
3
4
5
6
7
8
9
10
11
12
Avg. Percent Solids -
Standard Deviation -
Relative Standard Deviation -
9.8
9.0
7.4
7.3
11.6
10.3
10.2
6.7
5.7
8.3
7.3
9.5
8.59
1.75
20.3%
2.0
2.7
2.0
3.7
0.7
2.6
2.8
4.3
0.5
3.3
1.0
1.5
2.26
1.20
53.3X
collected on the two dates (discussed in later sections) indicate that indeed
the wastes were different on the two sampling dates. As the waste source is a
disposal facility (rather than, for example, a process effluent sump), it is
not surprising that the wastes cannot be directly compared after an interval
of approximately 5 months. Also, the low average percent solids in the Pond 0
samples result in high RSD's with small actual differences between samples.
In general, sampling precision (RSD's based on percent solids) is better with
samples containing higher solids concentrations (>5%).
COL IWASA
Sample-to-sample reproducibility of the COLIWASA with two different
liquid waste types is presented in Tables 14 and 15. The percent total solids
values presented in Table 14 are averages of three aliquots from each sample.
The reproducibility of the samples within drums (RSDs = 0.53X-5.434) is of the
same high order as that of the aliquots drawn from individual samples. The
homogeneity of this API separator waste is evident from drum-to-drum
comparisons (virtually no detectable differences).
A better test of the COLIWASA's ability to handle complex waste mixtures
was provided by the oil/water biphasic waste (Table 15). Significant drum-to-
drum differences were evident in the oil:water ratios determined for the waste
samples, but the reproducibility of samples within drums was very good (range
of RSDs = 8.3%-19.4%).
32
-------
TABLE 14. EVALUATION OF COLIWASA SAMPLING METHOD WITH
DRUMMED API SEPARATOR WASTE
Non-filterable
Drum Number Solids (%) S (n = 3) RSD (as %)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1.097
1.110
1.090
1.120
1.077
1.077
1.050
1.083
1.077
1.073
1.083
1.077
1.077
1.083
1.067
0.012
0.026
0.010
0.061
0.006
0.006
0.010
0.015
0.006
0.006
0.006
0.006
0.006
0.006
0.005
1.05
2.38
0.92
5.43
0.54
0.54
0.95
1.41
0.54
0.54
0.53
0.54
0.54
0.53
0.54
A caution 1s appropriate relative to sampling multiphasic materials with
the COLIWASA. The bias introduced by the inability of the COLIWASA to sample
clear to the bottom of a drum or tank becomes more significant the lower the
level of waste -- particularly bottommost layers — in the container. The
disporportionality in resulting samples will be further exaggerated with
repeated sampling. Care should be exercised to lower the COLIWASA no more
rapidly than the liquid levels in the sampler and drum can equilibrate. This
is especially important with vtscous liquids. Otherwise, the potential exists
for biasing the resulting samples with materials from the lower portions of
the tank or drum.
Proposed Drum and Tank Sampler Design
To accommodate the need for a drum or tank sampler that can reach to
within 1 cm of the bottom of the containing vessel, a sampler which is
convenient to use (length adjustable for different sampling requirements) and
economical (disposable collection tubes) was designed at EMSL/LV. The sampler
is designated as the Williams/Beckert Drum and Tank (DAT) sampler. The DAT
sampler consists of a support frame of stainless steel tubing. The support
frame, as designed for drum sampling, is 4 feet long. Extension tubes 6 feet
in length give the options of 4-, 6-, or 10-foot sampler lengths as needed.
33
-------
TABLE 15. EVALUATION OF THE REPRODUCIBILITY OF BIPHASIC (OIL/WATER)
LIQUID SAMPLES COLLECTED WITH A COLIWASA FROM WASTE DRUMS
Drum #1
Sample 1
Sample 2
Sample 3
Drum Avg.
Drum #2
Sample 1
Sample 2
Sample 3
Drum Avg.
Drum #3
Sample 1
Sample 2
Sample 3
Drum Avg.
Drum 1
Drum 2
Drum 3
Height of Oil
(mm)
15
17
14
15.3
19
20
22
20.3
54
45
43
47.3
Mean Ratio
0.12
0.34
0.36
Height of H20
(mm)
122
129
124
125.0
59
63
60
60.7
122
137
135
131.3
Std. Dev.
0.01
0.03
0.07
Oil: Water Ratio
0.12
0.13
0.11
0.12
0.32
0.32
0.37
0.34
0.44
0.33
0.32
0.36
RSD (as %)
8.3
8.8
19.4
Avg.
0.273
34
-------
The frame receives and supports standard "3/4-inch Class 125" thin-wall
PVC irrigation pipe — readily available in 10' lengths at very low cost from
garden supply stores — which serves as the disposable collection tube.
Collection tubes of other materials (e.g., glass or stainless steel tubing)
can be substituted for PVC when there is reason to suspect the liquid waste
contains solvents which attack PVC. Ten-foot lengths of the PVC pipe are
easily cut with knife or hacksaw to the lengths desired (generally 4-, 6- and
uncut 10-foot lengths). The DAT sampler and its use are described in more
detail in Appendix 3.
EXTRACTION PROCEDURE EVALUATION
Reproducibility
The extraction procedure (EP) was evaluated to determine its reproduci-
bility when used for the identification of hazardous waste. The EP extracts
were first screened for several elements including arsenic, barium, cadmium,
chromium, and lead by inductively coupled plasma emission spectroscopy (ICP).
The screening analyses were for qualitative identification and were restricted
to those toxic elements (listed in the regulation)- that can be analyzed by the
ICP system at the EMSL-LV. The extracts were then analyzed for each element
of interest by atomic absorption (AA) spectroscopy in accordance with the
proposed regulations.
Table 16 presents the ICP data collected in the Initial phase of this
study. These data indicate very high concentrations of the toxic elements of
interest in several of the EP extracts from samples taken at Site A. Samples
from Pond 0, Site A had to be diluted since the concentrations of As, Cr, and
Pb exceeded the linear range of the analytical method. The EP extracts of
samples from Pond 10, Site A, the pesticide waste from Site C, and the filter
cake from Site G had low or insignificant concentrations of the metals that
could be identified by the screening analysis.
Table 17 gives the atomic absorption (AA) data obtained in the initial
phase of this study. The results confirm the ICP findings and provide
quantitative values for the concentrations of barium, chromium and lead in the
EP extracts. These three elements were selected for AA analysis in the
initial phase because of their relatively high concentrations observed in the
ICP screening analyses,
The samples from Pond 0 and Pond P (Site A) also contained relatively
high concentrations of barium, chromium and/or lead. Triplicate aliquots from
each of two samples from each pond were analyzed using the EP for two reasons:
(1) to obtain additional data on the reproducibility of the EP; and (2) to
point out differences between two pond wastes that, according to the site
operator, came from the same source. The standard deviations presented in
Table 18a reflect the reproducibility.of the analytical procedures performed
on wastes from Ponds 0 and P.
Data in Tables 18a and b indicate the level of extraction reproducibility
observed with the two wastes. High reproducibility was obtained with
35
-------
TABLE 16. ICP SCREENING ANALYSIS OF EP EXTRACTS: APPROXIMATE ELEMENTAL
COMPOSITION OF EXTRACTS FROM SELECTED WASTE SAMPLES
Sampl e
Approximate
(No. of Extracts Analyzed) As
A = 1937.7
Site A, Pond 13, Location 1 (1)
Site A, Pond 0, Location 2 (IS)
Site A, Pond P, Location 2 (7)
Site A, Pond 10,
Sulfonation Tars (2)
Site B, Paint Sludge,
Sampled 4-19-79 (3)
(3)**
Site B, Paint Sludge, •
Sampled 6-13-79 (1)
Site C, Pesticide Waste (2)
Site D, Chromate Oxidation Paste
Site D, API Oil -Water Separator (3)
Site E, Electric Furnace
Baghouse Dust (1)
Site E, Blast Furnace Scrubber Filter
Cake (1)
Site E, Lime Sludge from Ammonia
Still (1)
1.3
168
0.6
<0.4
0.8
<0.4
0.6
<0.4
0.8
<0.4
1.6
0.6
1.6
Site E, Mill Scale from Water Treatment
Plant (1) <0.4
Site G, Filter Cake, Cl/Hg Process
Stream (1)
Site I, Chlorine Process Sludge (1)
<0.4
1.8
Ba
4554.0
0.2
10.8
24.2
0.08
1.13
0.57
18
0.26
0.6
<0.002
0.8
1.3
<0.002
<0.002
<0.002
0.07
Concentration*
Cd
2265.0
0.5
4.2
<0.2
<0.02
0.06
0.05
0.02
<0.02
0.6
<0.02
<0.5
<0.02
<0.02
<0.02
<0.02
--
Cr
2835.6
1.8
1400
124
<0.02
7.1
0.09
4.1
<0.02
4.5
3.6
3.5
0.4
2.3
3.1
<0.02
0.3
(mg/1)
Pb
2203.5
0.5
168
1.2
<0.25
1.2
<0.25
0.25
<0.25
0.4
<0.25
0.5
7
0.4
<0.25
<0.25
0.9
* Chemical, physical and spectral interferences were not minimized. Results
are corrected for dilution. Data for Ponds 0 and P, Site A represent
averages from analyses of extracts from replicate samples; in some cases,
extracts had to be diluted to bring values within the linear range of the
instrument. Emission wavelength (x, in nanometers) is indicated for each
element.
By ASTM procedure with distilled water.
36
-------
TABLE 17. EVALUATION OF EXTRACTION PROCEDURE (EP): MEANS AND STANDARD
DEVIATIONS FOR AA ANALYSES* OF EP EXTRACTS FOR BARIUM,
CHROMIUM AND LEAD
Waste Extracted
Sulfonation Tars
(Site A, Pond 10)
Paint Sludge, Site B
(Collected 4-19-79)
Paint Sludge, Site B
(Collected 6-13-79)
Pesticide Waste, Site C
API Oil Separator Inlet,
Site D
Chromate Oxidation Paste,
Site D
Electric Furnace Baghouse
Dust, Site E
Blast Furnace Scrubber
Filter Cake, Site E
Barium
X
<0.9
<0.9
9.8
5.15
21.1
0.9
<0.9
<0.9
<0.9
<0.9
<0.9
0.80
1.04
0.85
1.06
0.64
0.90
(mg/D
s
—
1.5
0.32
2.7
0.1
—
--
0.06
0.11
0.19
0.08
0.14
0.03
Chromium (mg/1)
x s
<0.02
<0.02
4.1 0.1
1.02 0.13
1.90 0.17
<0.02
9.1 0.1
1.2 0.0
1.0 0.0
7.9 0.2
1.4 0.1
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
Lead
X
0.3
0.3
0.1
0.08
<0.08
<0.08
0.1
0.1
0.1
<0.8
<0.8
0.13
0.13
0.13
15.3
14.4
11.6
(mg/i)
s
• o.i
0.1
0.1
0.1
--
0.1
0.1
0.2
—
0.04
0.03
0.04
0.7
0.7
0.7
(continued)
Flame Atomic Absorption analyses performed in triplicate.
37
-------
TABLE 17. (Continued)
Barium (mg/1) Chromium (mg/1) Lead (mg/1)
Waste Extracted x s x s x s
Mill Scale, Water Treatment
Plant, Site E 0.20 0.03 <0.32 — <0.08
0.19 0.03 <0.32 — <0.08
0.52 0.04 <0.32 -- <0.08
(1) 0.28 0.01 <0.32 -- <0.08
(1) 0.24 0.02 <0.32 — <0.08
(1) 0.22 0.03 <0.32 -- <0.08
Lime Sludge, Ammonia
Still, Site E (2)2.7 1.6 <0.02 -- 0.3 0.1
Filter Cake, Chlorine/Hg
Process Stream, Site G 0.25 0.03 <0.32 — 0.10 0.06
0.14 0.05 <0.32 — <0.08
0.15 0.07 <0.32 -- <0.08
(2) 0.9 -- <0.02 -- <0.08
Chlorine Process Sludge,
Site I
0.48
0.40
0.98
(2) 3.5
0.16
0.46
0.49
1.0
<0.32
<0.32
<0.32
0.1
0.45
0.45
0.48
0.0 0.4
0.04
0.04
0.04
0.1
(1) Results obtained with wrist-arm shaker.
(2) Preliminary and test data.
replicate extractions of the same waste samples when chromium and lead were
used as indicators. The reproducibility of replicate chromium and lead
analyses on a given EP extract was also found to be high. While the
reproducibility of the barium analyses was not as precise as for lead and
chromium, the data suggest that extraction reproducibility for Ba is
reasonable with the wastes tested. Problems with barium contamination
resulting from the filtration apparatus (pre-filter) and problems in barium
analysis are discussed later.
The EP was not evaluated with waste samples containing appreciable
fractions of oil or organic solvents. The applicability of the EP to these
kinds of samples should be evaluated, as a large number of waste types can be
expected to contain such materials.
38
-------
TABLE 18a. EVALUATION OF EXTRACTION PROCEDURE (EP): AVERAGE MEANS (n=3)
AND STANDARD DEVIATIONS FOR AA ANALYSES* OF EP EXTRACTS OF WASTES
FROM PONDS 0 AND P, SITE A
Barium
Sample Extracted
Pond 0
Pond P
2A
2B
2A •
2B
1
1
29
27
x(D
.65
.34
.9
.8
s
0.
0.
4.
3.
(rag/D
RSD
17
05
9
7
10.
3.
16.
13.
3
7
4
3
Chromium (mg/1)
x(l)
1040
943
77.6
82.5
s
17
21
2.4
2.0
RSD
(*)
1.6
2.2
3.1
2.4
Lead (mg/1)
RSD
x(l) s (%}
45.7 0.5 1.1
43.5 2.0 4.6
__
—
* Flame Atomic Absorption analyses performed in triplicate on each of three
aliquots of sample extracts.
RSD = Relative Standard Deviation
Table 18b. RELATIVE STANDARD DEVIATIONS (RSD) OF EXTRACTIONS
AND ANALYSES FOR SELECTED METALS
RSD (J)
Analysis
(Sample source: Ponds 0 and P, Site A) Barium Chromium Lead
Differences between replicate
determinations on a given
EP extract 14.9 1.3 2.0
Differences between replicate
extractions on a given sample
of waste 11.0 1.8 3.0
39
-------
Comparison of Three Extractor Types
Data from preliminary studies (Table 11) had suggested good agreement
between a wrist-arm type shaker and the blade-type extractor described in the
proposed regulation. A study was then undertaken to compare the blade-type
extractor with the NBS tumbling-action extractor and the wrist-arm shaker for
three waste types (Ponds 0 and 12 and biosolids).
Three metals were analyzed for this comparison study. The average
concentrations and standard deviations for each metal by extraction technique
are shown in Table 19. The blade-type extractor tended to yield lower extract
concentrations of the metals extracted than the other two types, especially
from samples with low solids (Pond 0). Of the three extractor types, the
blade-type agitates the samples the least and appears most sensitive to sample
size.
In Table 20, the relative percent contribution to total variance is shown
by sample, technique, aliquot, and analysis. The ranking from highest to
lowest, using the average relative variance, is sample, 63 percent; aliquot,
20 percent; extraction technique, 10 percent; and analysis, 7 percent.
The data in Table 21 are the statistical parameters used to evaluate the
equivalency of the three techniques. These were used in Table 22 to indicate
the significance of differences between samples, techniques, or aliquots. The
Student-Newman-Keuls multiple comparison test (Table 23) was used to show
whether one extraction technique was "better" than another. Only in two cases
did the statistics support one technique over the others. This was for
chromium in the biosolids samples and barium in Pond 0 samples.
The overall performance of the three extractors is perhaps best
summarized by comparing the average RSD for each extractor (average overall
RSD including the variability components associated with sampling, aliquoting
and analysis).
Extractor Overall RSD (as percent)
Pond 0 Pond 12 Biosolids
Blade-type (rotary) 10.8 226.4 12.8
NBS tumbling-type 10.0 129.7 10.8
Wrist-arm shaker 12.3 143.9 10.5
The variability associated with sampling, aliquoting, extracting, and
analyzing the Pond 0 samples for chromium and lead was determined in this
study. Those components contributing to the overall variability associated
with the chromium determination were:
Analyses 0.3% Extraction Techniques 0.1%
Aliquots 5.6% Sampling 94.1%
40
-------
TABLE 19. EXTRACTOR COMPARISON - MEAN AND STANDARD DEVIATION OF CONCENTRATIONS OF METALS
EXTRACTED WITH THREE EXTRACTOR TYPES
Ba
N
Cr
W
N
Pb
N
Pond 0
x=6.21 x=5.40 x=3.72
5=1.25 s=0.72 s=0.84
x=917 x=948 x=907
5=113 s=94.6 5=94.6
x=36.70 x=36.95 x=36.50
5=3.96 5=4.37 5=4.56
Pond 12
NO ANALYSES PERFORMED
x=0.66 x=0.37 x=0.53
5=0.95 s=0.48 s=1.20
x=0.58 x=0.48 x=0.67
5=0.51 s=0.34 5=0.84
Blosolids
x=1.56 x=1.38 x=1.22
s=0.48 s=0.41 s=0.50
x=0.31 x=0.23 x=0.17
5=0.13 5=0.06 5=0.06
x=0.38 x=0.37 x=0.39
5=0.04 5=0.04 5=0.05
W = Wr1st-Arm Shaker
N = NBS Tumbling-Type Extractor
R = Blade-Type Extractor
-------
TABLE 20. EXTRACTOR COMPARISON - VARIABILITY ATTRIBUTABLE TO SAMPLING AND ANALYTICAL PROCEDURES
ro
Relative Sources of Variation (Percent)
Waste Type
Pond 0
Pond 0
Pond 0
Pond 12
Pond 12
Pond 12
Biosolids
Biosolids
Biosolids
Metal
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
°Samples
0
94.0
94.6
50.2
59.5
0
10.6
60.0
"Techniques
61.9
0.1
0.0
NO ANALYSES
0
13.0
49.4
45.2
0
°Aliquots
18.2
5.6
4.3
PERFORMED
47.2
27.2
0
43.6
0
°Analyses
19.9
0.3
1.1
2.6
0.3
50.6
0.6
40.0
°fotal
100.0
10U.O
100.0
100.0
10U.O
100.0
100.0
100. U
-------
TABLE 21. EXTRACTOR COMPARISON - "F VALUES" CALCULATED FKOM ANALYSES OF VAKIANCE
4k
C*>
ANOVA Statistical Significance Criteria
Samples
Waste Type
Pond 0
Pond 0
Pond 0
Pond 12
Pond 12
Pond 12
Biosolids
Biosolids
Biosolids
Metal
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
F(calc)
0.27
37.05
206
10.30
8.87
0.55
1.54
2.36
vl,v2,o
4.07
4.07
4.07
NO
2.21
2.26
2.21
2.21
2.21
Techniques
F(calc)
9.33
1.58
1.40
ANALYSES PERFORMED
0.90
2.42
1.98
4.08
1.55
Fvl,v2,o
2.36
2.36
2.36
1.68
1.69
1.68
1.68
1.68
Levels (A=0.
05)
Aliquots
F(calc)
2.77
11.46
2.27
37.11
186
0.65
158
0.45
vl,v2,a
1.86
1.83
1.83
1.43
1.45
1.43
1.43
1.43
-------
TABLE 22. EXTRACTOR COMPARISON - DIFFERENCES BETWEEN TECHNIQUES, BY WASTE TYPE
Waste
Type
Pond 0
Pond 0
Pond 0
Pond 12
Pond 12
Pond 12
Biosolids
Biosolids
Biosolids
Metal
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
Average
Concentration
(mg/liter)
5.14
924
36.7
0.514
0.562
1.39
0.234
0.380
Total
Number AA
Analyses
69
72
72
216
209
215
216
215
Significant Difference (o=0.05) Between
Samples
No
Yes
Yes
NO ANALYSES
Yes
Yes
No
No
Yes
Techniques
Yes
No
No
PERFORMED
No
Yes
Yes
Yes
No
Aliquots
Yes
Yes
Yes
Yes
Yes
No
Yes
No
Relative Standard
Deviation Between
Techniques (%)
22.1
0.7
<0.1
<0.1
38.2
32.1
31.7
<0.1
-------
TABLE 23. EXTRACTOR COMPARISON - RESULTS OF MULTIPLE COMPARISON TEST TO RANK EXTRACTION TECHNiqUES
U1
Waste
Type
Pond 0
Pond 0
Pond 0
Pond 12
Pond 12
Pond 12
Biosolids
Biosolids
Biosolids
Metal
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
Order of
XR
XR
XR
NO
XN
XN
XR
XR
XN
= 3.72 <
= 907 <
= 36.50
ANALYSES
= 0.37 <
= 0.48 <
= 1.22 <
= 0.17 <
= 0.37 <
Results of Student-Newman-
Keuls Test: Order of
Techniques Sample Means* Technique Population Means
XN = 5.40 < Xy = 6
xw = 917 < XN = 948
< xw = 36.70 < XH =
PERFORMED
XR = 0.53 < Xy = 0
Xy = 0.58 < XR = 0
XjJ = 1.38 < Xy = 1
XN = 0.23 < Xy = 0
Xy = 0.38 < XR = 0
•21 UR < MH = Uy
"R • ^ = VN
36.95 MB = MU = MN
K w N
•66 11 HJ = IIR = My
.67 VM = Wy = yR
.56 PR = MN = PW
.31 Mr> = MM ^ MU
n n n
.39 MM = Mu = M0
n W K
Significant Differ-
ence Between Tech-
niques (a = 0.05}
Yes
No
No
No
No
No
Yes
No
* Subscripts: R refers to blade-type extractor; N refers to NBS tumbling-type extractor;
W refers to wrist-arm shaker.
-------
As determined from the analysis of lead data the variability components were:
Analyses 1.1% Extraction Techniques 0.0%
Aliquots 4.3% Sampling 94.6%
The above data clearly identify sample-to-sample variability as the
greatest component of variability found in analyses of the Pond 0 waste
Samples. Unless sampling personnel have good evidence that a particular waste
or waste stream is physically and chemically homogeneous, it is a recommended
practice to pre-sample the waste. With preliminary sampling, an estimate of
the true sample population mean and variance can be determined. The number of
samples needed to determine the confidence level for estimates of sample
population precision can then be calculated (Cochran, 1963). Even when a
waste is so heterogeneous that it is impractical to obtain enough samples to
achieve high confidence in the precision estimate, some level of confidence
can still be determined.
Sewage Sludge Extraction and Analysis
During a sewage sludge study, performed in conjunction with the Municipal
Environmental Research Labortory (MERL/Cin), we found significant levels of
chromium in reagent blanks run through blade-type extractors 1 and 4. The
dissolved metals analyses on the acetic acid blanks are given in Table 24.
These levels did not impact the classification of the sludge samples tested,
as none exceeded the criterion level for chromium.
The levels of metals found in the EP extract of the sludges tested are
shown in Table 25. Although the precision'was not high (RSD, 33% - 43%) for
chromium levels approaching the lower detection limit, results of chromium
determinations for triplicate sample aliquots extracted in various extractors
were found to be comparable. The highest levels of cadmium were found in
extracts of sludge #2 (0.601 mg/1), sludge #6 (0.116 mg/1), and sludge #7
(0.275 mg/1). These levels would not have resulted in the classification of
these wastes as hazardous. Levels of total metals in the digested sludges are
listed in Table 26.
Although some values agreed between the two performing laboratories,
EMSL/LV and MERL/Cin, a sufficient number of values were different enough to
raise concern. When the measurement units were adjusted to allow direct
comparison, extracted metal concentrations found by MERL/Cin ranged, with few
exceptions, from about 25 percent to 70 percent of the values recorded by
EMSL/LV. The differences appeared to be systematic and suggested that
procedural differences might exist. Discussions did reveal differences in the
procedures followed (e.g., different extractors; use of nitric acid vs.
perchloric acid/nitric acid as digestion medium; and metal values based upon
dried sludge weight vs. sample volume) which may have influenced the
comparative results.
It is evident from the interlaboratory exercise that procedures to be
followed must be precisely spelled out, step-by-step, so that the
46
-------
TABLE 24. LEVELS OF DISSOLVED METALS (mg/1) IN ACIDIFIED BLANK SAMPLES -
AVERAGE OF TRIPLICATE DETERMINATIONS
Source
of Sample
Filtration
Extractor
Extractor
Extractor
Extractor
Apparatus
No.
No.
No.
No.
1
2
3
4
As
<0.02
<0.02
<0.02
<0.02
<0.02
<0
<0
<0
<0
<0
Ba
.45
.45
.45
.45
.45
<0
<0
<0
<0
<0
Cd
.02
.02
.02
.02
.02
Cr
<0.04
0.120 i .050
<0.04
<0.04
0.048 ± .008
<0
<0
<0
<0
<0
Pb
.07
.07
.07
.07
.07
Se
<0.02
<0.02
<0.02
<0.02
<0.02
possibilities of misinterpretations or substitution of non-equivalent
alternate procedural elements are completely eliminated.
Triplicate extracts of sludge #2 were provided to MERL/Cin as a further
check on the analytical reproducibility between the laboratories. The
agreement between average cadmium values -- the greatest concentration
differences recorded in the initial intercomparison were with cadmium -- was
excellent:
EMSL/LV MERL/Cin
Extract A 0.28 ug/1 Cd 0.28 pg/1 Cd
Extract B 1.07 ug/1 Cd 1.05 ug/1 Cd
Extract C 0.45 pg/1 Cd 0.44 pg/1 Cd
Extractability of Mercury
Only two extracts have been analyzed for mercury (Table 27). Although
these samples were expected to contain high mercury concentrations, analysis
of the EP extracts yielded values of less than 4 ug/1* When aliquots of the
waste samples were digested with aqua regia and the digested samples analyzed
for mercury, the results (Table 27) confirmed that the waste samples did
indeed contain high concentrations of mercury (approximately 1-2 mg Hg per
gram of waste). However, the EP extracted only a very small fraction of the
mercury present. The mercury may be in the form of organo-mercury compounds,
it may be tightly bound to the solids in the matrix or it may be present in a
highly insoluable form, e.g., as HgS. If so, the results simply suggest that
47
-------
TABLE 25. LEVELS OF DISSOLVED METALS FOUND IN EP EXTRACTS OF SELECTED WASTE SAMPLES -
AVERAGE OF TRIPLICATE EXTRACTIONS
As Ba Cd Cr Pb be
Samp] e
Designation mg/1 S.D. mg/1 S.D. mg/1 S.D. mg/1 S.D. mg/1 S.D. mg/1 S.D.
Tulsa, OK
combined sludge <0.02 — <0.45 — <0.02 — <0.04 -- <0.07 — <0.02
Fountain Valley, CA
sludge 0.034 0.013 <0.45 — 0.601 0.414 0.07 0.03 <0.07 ~ 0.02 0.01
Cincinnati, OH
sewage sludge 0.031 0.00 0.53 0.18 0.018* 0.002 0.05 0.02 0.08 0.03 O.U2 O.U1
Dallas, TX
primary sludge <0.02 — 0.50 0.16 0.007* 0.003 0.06 0.02 <0.07 ~ 0.03 0.02
Chicago, IL
digested sludge 0.031 0.006 <0.45 — <0.02 -- <0.04 -- <0.07 — <0.02
Chicago, IL
fertilizer 0.030 0.012 <0.45 — 0.116 0.019 0.123 0.010 <0.07 — <0.02
Chicago, IL
landfill sludge <0.02 — <0.45 — 0.275 0.093 <0.04 -- <0.07 — 0.037 0.001
* Single outlier value not included in mean
S.D. = Standard deviation
-------
TABLE 26. LEVELS OF DISSOLVED METALS (TOTAL) FOUND IN DIGESTS OF SELECTED
WASTE SAMPLES - AVERAGE OF TRIPLICATE DIGESTS
As Ba Cd Cr Pb Se
Sampl e
Designation mg/1 S.D. mg/1 S.D. mg/1 S.D. mg/1 S.D. mg/1 S.D. mg/1 S.D.
Tulsa, OK
combined sludge 0.055 0.002 1.27 0.02 0.317 0.004 0.86 0.04 0.756 0.053 0.040 0.002
Fountain Valley, CA
sludge 0.692 0.040 0.722 0.210 2.71 0.12 12.2 1.01 8.97 0.17 0.312 O.OJ4
Cincinnati, OH
sewage sludge 2.72 0.52 4.32 0.64 0.814 0.041 20.7 2.0 31.6 2.2 1.21 0.16
Dallas, TX
primary sludge <0.02 — 0.48 0.16 0.034 0.006 0.35 0.02 0.260 0.008 0.018 O.OU2
Chicago, IL
digested sludge 0.097 0.001 1.09 0.07 0.310 0.007 3.89 0.44 1.13 0.06 0.10 0.01
Chicago, IL
fertilizer 0.466 0.044 3.55 0.77 2.87 0.53 23.5 4.9 3.27 0.51 0.30 0.07
Chicago, IL
landfill sludge 1.08 0.14 2.46 0.14 7.41 0.76 43.0 8.2 17.8 5.6 0.97 0.08
S.D. = Standard deviation
-------
TABLE 27. EVALUATION OF EXTRACTION PROCEDURE (EP): COMPARISON OF
MERCURY CONCENTRATIONS (COLD VAPOR AA) IN WASTE SAMPLE DIGESTS
WITH THOSE ESTIMATED FROM EP EXTRACTS
EP Reconstructed* Digested
Extract (wg/1) Sample (pg/g) Sample (wg/g)
Waste Sample
Filter Cake, Site G <0.2 — <0.0004 - 1970 110
Chlorine Process
Sludge, Site I 3.7 0.2 0.074 0.004 840 130
* Calculated from extract concentration and dry weight of solids in the
waste sample extracted.
the mercury would not leach out from the waste under acidic (pH = 5.0)
conditions. For most elements the fact that only a small fraction of the
total amount present in the waste can be leached out using the EP is of minor
concern. The reason is that non-1eachable materials are not considered to
pose a significant toxicity hazard in a properly managed waste. However, in
the case of mercury the situation is difficult. Mercury compounds are known
to undergo -- in soils and in aquatic environments --changes that affect the
solubility of the mercury species (e.g., HgS-*- HgS04; Hg2*-* Hg°; Hg2+-»-
RHg+-»- R2Hg). Therefore, any waste that contains high levels of mercury or
mercury compounds -- even though the EP extractable part is small -- should be
regarded as a potential time bomb. This indicates that the extraction
procedure or analytical method may be inadequate for identifying mercury
hazards. More work is required to identify the form of mercury present in
these samples and to clarify the questions concerning the use of the
Extraction Procedure for identification of Teachable mercury in waste samples.
Extract Stability
A problem has been experienced with the stability of some EP extracts.
Some of the extracts, especially those with high concentrations of inorganic
and organic materials, formed precipitates over a period of several days.
This problem was observed early in the study and was addressed by adding a
step to the procedure. In this step the EP extract is split into two samples
as it is prepared. One sample is stored in a refrigerator at 4°C until it can
be analyzed for organics; the other sample is acidified to pH < 2 with nitric
acid to preserve the sample for elemental analysis. However, even with this
step, some of the more concentrated samples produced a precipitate within a
few days after preservation. This problem will be investigated in more detail
in future studies.
50
-------
Background Interferences
Given the low but varying levels of contamination found In the blanks
(Tables 28-30), a question arose as to whether the contamination was caused by
the metallic components of the apparatus, or by improper cleaning between
runs. To answer this question, the levels of chromium in the blanks were
compared to the levels in the extracts generated in the corresponding
blade-type extractor immediately prior to the blank. Figure 4 is a graphical
presentation of this data. The data do not represent a special study to
answer this question, but they do suggest that contamination of the equipment
is occurring, and that this is resulting in contamination of the succeeding
extracts. The lack of any correlation between chromium concentration in the
distilled v/ater blank and that in the sample preceeding the blank further
suggests that using distilled water blanks to determine equipment cleanliness
is not adequate. The chromium level in the acidic blanks (0.1N acetic acid)
generally parallels the levels of chromium in the preceding sample.
To further explore this relationship, a number of sequential acidic
blanks were run and data derived from analyzing splits of these "extracts"
were evaluated. These blank extracts were divided into two portions, one of
which was filtered, before analysis. Figure 5 compares the concentration of
barium in the filtered and unfiltered portions. While the barium
concentrations in the unfiltered portions remain below the instrumental
detection limits, the filtered portions consistently contain readily
detectable amounts. This finding, and a direct comparison of barium levels
from extracted (Table 28) and filtered/extracted acidified blanks (Table 30),
suggested to us that some component of the filtration apparatus was
contributing barium to the blanks. It is interesting to note that the data in
Figure 5 correspond to the period in which a transition was made from routine
use of distilled water blanks (or very slightly acidified blanks) to use of
blanks containing the maximum level of acetic acid (0.1N) employed in the
Extraction Procedure. To further define the source of the Ba contribution,
the Nuclepore® prefliter and polycarbonate filter pads were soaked for 50
minutes in 50-ml volumes of dilute nitric acid (pH 2.5) and acetic acid
(0.1N). These extracts were then analyzed for Ba. Only the glass fiber
prefilter pad soaked in dilute nitric acid leached measurable amounts of Ba.
The levels of Ba (approximately 1 mg/1) leached from the prefilter by the
nitric acid reagent are comparable (considering the dilution factor) to those
observed (Table 30) from acidified blanks filtered using the filtration
apparatus. The prefilter used in this test was a Nuclepore® P40 Prefilter
(stock number 211703).
Experiments were then conducted in order to determine whether the barium
could be readily removed by soaking the filter pads in dilute nitric acid.
Three Nuclepore®P40 prefliters were soaked for 40 ninutes in 50-ml volumes of
2N nitric acid. The solutions were then decanted, analyzed, and found to
contain 4.1 mg/1 of Ba (S=0.71). After rinsing three times with distilled
water, they were again soaked in 50 ml volumes of acid, this time for 60
minutes. These solutions average 1.1 mg/1 of Ba (S=0.09). Again the filters
were rinsed three times with distilled water, and extracted with 50 ml of
acid, this time for 75 minutes. The barium concentration in this final
extract was 0.60 mg/1 (S=0.12). The Teachable barium is not fully removed
from the prefilters. A single 100-minute soak in 2N nitric acid was performed
51
-------
TABLE 28. EVALUATION OF BACKGROUND CONCENTRATIONS OF ELEMENTS FROM
EP EXTRACTORS. MEANS AND STANDARD DEVIATIONS OF TRIPLICATE
ATOMIC ABSORPTION ANALYSES OF ACIDIFIED BLANK SAMPLES
Ba (mg/1) Cr (mg/1) Pb (mg/1)
Extractor #
1
2
3
4
4
5
7
1
1
2
2
3
3
4
4
5
5
6
7
7
7
Matrix x s
(*) <0.1 -
(*) <0.1
(*) <0.1 -
(*) <0.1 -
(*) Rerun <0.1
(*) <0.1
(*) <0.1 -
(**) <0.1
(**) <0.03 -
(**) <0.1 -
{**) <0.03 -
(**) <0.1 -
(**) 0.1 0.0
(**) <0.1
\ /
(**) <0.03 -
(**) <0.1 -
(**) <0.03 -
(**) <0.1
(**) <0.1
\ 1 •» » - •
(**) <0.1 -
(**) <0.03 -
x s
<0.31
<0.31
<0.31
<0.31
<0.31
<0.31
<0.31
0.20 0.03
0.22 0.00
0.26 0.02
<0.04
<0.04
0.06 0.01
0.12 0.01
0.06 0.01
<0.04
0.04 0.00
<0.04
<0.04
0.28 0.01
<0.04
x
<0.12
<0.12
<0.12
<0.12
<0.12
<0.12
<0.12
<0.2
<0.09
<0.2
<0.09
0.07
<0.09
0.04
<0.09
0.02
<0.09
<0.02
0.02
<0.02
<0.09
s
_
-
-
-
-
-
-
-
-
-
-
0.03
-
0.02
-
0.01
-
-
0.01
-
-
* Deionized water adjusted to pH 5.0 ± 0.2 with 1.0 M acetic acid. (Total
volume of extract = 2000 ml.)
** Deionized water to which 400 ml acetic acid was added (maximum amount
allowable in the EP). (Total volume of extract = 2000 ml).
52
-------
TABLE 29. EVALUATION OF BACKGROUND CONCENTRATIONS OF ELEMENTS FROM
FILTRATION APPARATUS: MEAN AND STANDARD DEVIATIONS OF TRIPLICATE
ATOMIC ABSORPTION ANALYSES OF DEIONIZED WATER BLANK SAMPLES
Filtration
Apparatus #
2
1
3
3
2
2
3
3
2
2
2
2
2
2
2
2
2
2
2
3
3
3
2
2
2
2
Ba (mg/1 )
x s
<0.9
<0.9
0.10 0.02
.0.12 0.03
<0.06
<0.06
0.11 0.02
0.16 0.03
0.11 0.03
0.09 0.01
<0.06
<0.06
<0.1
<0.1
0.1 0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Cr (mg/1)
x s
<0.02
<0.02
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
0.15 0.01
0.42 0.01
0.20 0.01
<0.31
<0.31
<0.31
<0.31
0.38 0.02
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
Rerun x*
_
-
0.05
0.06
0.11
0.04
0.29
0.02
0.04
0.03
<0.02
<0.02
0.16
0.42
0.17
0.12
<0.02
<0.02
0.09
0.34
-
-
-
-
-
-
Pb (mg/1)
x
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
<0.2
<0.2
<0.2
<0.12
<0.12
<0.12
<0.12
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
s
_
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
_
(continued)
53
-------
TABLE 29. (Continued)
Filtration
Apparatus #
3
3
3
3
3
3
3
3
3
3
Ba
X
<0.1
<0.1
<0.03
<0.03
0.03
0.06
0.03
0.06
0.07
0.12
(mg/i )
s
-
-
-
-
0.01
0.01
0.02
0.00
0.00
0.00
Cr (mg/1)
x s
<0.04
0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
Pb (mg/1)
Rerun x* x s
<0.02
<0.02
<0.09
<0.09
<0.09
<0.09
0.20 0.05
<0.09
<0.09
<0.09
* Blanks rerun on AA at lower (0.02) detection limit for Cr.
on each of two Mi 11ipore® Type AP 15124 prefilters as a comparison to see if
this contamination is to be expected using other brands of prefilters. The
resulting Ba concentration in the nitric acid solution was 1.8 mg/1 (S=0.24).
Barium in the glass matrix of the prefilters tested is considered to be the
source of the barium found in the acid extractants.
While the volume of acid used to extract the prefilters (50 ml) is
relatively small compared to the possible volume of extract in the Extraction
Procedure when applied to actual wastes, it should be noted that wastes with
relatively low solids content (e.g. <20% by weight of waste sample) and which
are already acidic (pH <5.0) would be more affected by the barium contribu-
tions from the prefilter. Until the problem can be resolved, the prefilters
cited here should be used only with caution and full awareness of the nature
and level of potential contamination. It is recommended that an acidic blank
(0.1N acetic acid) precede filtrations and be stored for future barium
analysis (and sample value correction). Such analysis would only be performed
in the event that the corresponding sample extract exceeded the criterion
level for barium.
While the problem of barium contamination was being studied, work was
conducted to determine if the stainless steel extractors or filter apparetus
contributed to the contamination of the extracts. This was of special concern
since close examination of the Nuclepore filtration apparatus suggested a
possible source of metal contamination. Specifically, the underdrain support
appeared to have been attacked by the materials passing through the equipment.
Areas of apparent breakdown at the resistive welds were visible and exposed
ends of the stainless steel rods which compose the underdrain support showed
54
-------
TABLE 30. EVALUATION OF BACKGROUND CONCENTRATIONS OF ELEMENTS FROM EP
EQUIPMENT: MEANS AND STANDARD DEVIATIONS OF ATOMIC ABSORPTION
ANALYSES OF FILTERED AND EXTRACTED ACIDIFIED* BLANK SAMPLES
Ba (mg/1) Cr (mg/1) Pb (mg/1)
Extractor No.** x s x s x s
1
1
2
2
2
3
4
4
4
5
5
5
6
7
7
1.0
0.24
1.0
1.12
(Filter No. 2) 1.2
1.2
2.6
0.6
0.78
1.2
0.8
0.82
1.2
0.5
1.04
0.2
0.01
0.2
0.02
0.1
0.2
0.3
0.3
0.00
0.3
0.3
0.06
0.2
0.4
0.12
0.21
0.18
0.21
0.04
0.13
0.10
0.20
0.05
0.06
0.11
0.08
0.04
0.10
0.20
<0.04
0.02
0.01
0.01
0.01
0.01
0.01
0.01
0.00
0.01
0.01
0.00
0.00
0.01
0.01
~
0.13
0.10
0.13
<0.09
<0.02
0.14
0.14
<0.02
<0.09
0.05
0.03
<0.09
<0.02
<0.02
<0.09
0.03
0.04
0.03
_
-
0.03
0.03
_
-
0.03
0.03
-
-
_
•
* Blank prepared by adding 400 ml of 1.0 M acetic acid (maximum amount
allowable In the EP) to 1600 ml of distilled water.
** Aliquots for extraction were filtered through filtration apparatus No. 3
(Filter No. 3), unless otherwise indicated.
evidence of active corrosion. These areas offer potential for carryover of
prior sample material -- the roughened, pitted surface and weld cracks are not
readily cleaned -- and for direct contribution of metal contamination through
corrosive breakdown and dissolution.
The Nuclepore filtration apparatus used in our studies is made of Type
304 stainless steel. A "sanitary" version is available with Type 316
stainless steel for all sample-contacting surfaces, but this version has not
been tested in our laboratory.
Several distilled water blanks showed levels of chromium above 0.06 mg/1.
About 75 percent of these were filtration blanks -- distilled water run
through the filtration apparatus and analyzed -- while the remainder went
through both the extraction and filtration apparatus. The distilled water
filtration blank which showed the highest level of chromium (0.38 mg/1)
immediately followed the running of a sample containing 33 mg/1 chromium. In
this case, a "memory effect" is likely, i.e., the washing procedure we
employed for the filtration apparatus was inadequate.
55
-------
1.0-
^ 0.3-
o>
{ 0.2H
o
•s
a
c
u
o
U
•g
a
0.1-
0.08
£ 0.04
0.03-
0.02
0.01-
4 values
below
detection
limits
\
Detection Limit
0.1 0.2 0.3 0.4 0.71.0 2.0 3.0 10.0
Chromium Concentrations in Preceding Sample (mg/l)
40 6080
Figure 4. Levels of chromium contamination found in extraction blanks
of distilled water (•) and 0.1N acetic acid (A) which followed
hazardous waste samples.
Nearly 70 percent of the acidified apparatus blanks — run immediately
after waste samples were extracted and routine cleaning performed -- showed
levels of 0.05 mg/l or more of chromium. However, no consistent trends could
be determined from the data to suggest that any one of the blade-type
extractors or filtration apparatus was contributing to the contamination more
than others.
Apparatus Abrasion/Contamination Study--
To assess the potential problem, a worst-case situation was devised to
measure the maximum level of contamination that might be anticipated to occur
using a stainless steel blade-type extractor. Specifically, we extracted
samples of sand, a highly abrasive material, with both strongly acidic (pH
2.5) and mildly acidic (pH 5) solutions, using nitric and acetic acids,
respectively. To insure that non-extractor sources would not be affecting the
results, parallel extractions of the sand were conducted in polyethylene
tumbler-type extractors designed by researchers at the National Bureau of
Standards.
56
-------
10 O-i
Blank ahquou filtered
through apparatus 03
Blank oliquots (unfiltered)
Lower
detection limit
Sequential Blank Number
Figure 5. Evaluation of apparatus contamination - Comparison of filtered
and unfiltered aliquots (splits) of sequential acidified blanks.
The results of the sand abrasion contamination study are presented in
Table 31. No detectable chromium was leached from the acid-washed sand in the
NBS type tumbler, nor were detectable levels found in either filtered or
unfiltered reagent blanks. However, chromium levels up to 0.18 mg/1 were
found in the unfiltered nitric acid extracts from the blade-type extractors.
The levels of chromium in the corresponding filtered samples were approxi-
mately one-half that in the unfiltered samples. Only one of the two
unfiltered acetic acid extracts from blade-type extractors showed any
detectable levels of chromium. After filtration, the chromium concentration
dropped to below the level of detection (0.01 mg/1). These data suggest that
the use of stainless steel extractors for testing strongly acidic, abrasive
wastes may result in contamination of the extracts, and should be avoided.
Similar findings with stainless steel extractors and abrasive wastes were
recently reported (Brown et al. In press).
The data for iron also indicate that stainless steel components might
abrade and contribute metals to highly acidic extracts. While detectable iron
57
-------
TABLE 31. SAND ABRASION/CONTAMINATION STUDY - CONCENTRATIONS OF
CHROMIUM, IRON, AND BARIUM DETECTED IN FILTERED AND UNFILTERED
REAGENT BLANKS AND EXTRACTS OF ACID-WASHED SAND
Chromium (mg/1)
Reagent/Extract Analyzed x s
Reagent Blank, pM 2.5
Unfiltered <0.01
Filtered APParatus 2 <0'01
niterea Apparatus 3 <0>01
Reagent Blank, pH 5.0
Unfiltered <0.01
FiltPrPd Apparatus 2 <0.01
Fi itered Apparatus 3 <0>01
Extract from NBS Extractor, pH 2.5
Unfiltered <0.01
Filtered (Apparatus 3) <0.01
Extract from NBS Extractor, pH 5.0
Unfiltered <0.01
Filtered (Apparatus 2) <0.01
Extract from Blade Extractor 1, pH 2.5
Unfiltered 0.18 0.00
Filtered (Apparatus 3) 0.09 0.00
Extract from Blade Extractor 2, pH 2.5
Unfiltered 0.15 0.01
Filtered (Apparatus 3) 0.08 0.01
Extract from Blade Extractor 4, pH 5.0
Unfiltered 0.06 0.00
Filtered (Apparatus 2) <0.01
Extract from Blade Extractor 5, pH 5.0
Unfiltered 0.01 0.01
Filtered (Apparatus 2) <0.01
Iron
X
<0.06
0.12
0.09
<0.06
<0.06
<0.06
<0.06
<0.06
<0.06
<0.06
3.46
3.53
1.99
1.94
0.54
0.09
0.70
0.44
(mg/1)
s
0.01
0.04
--
--
""
0.03
0.02
0.02
0.02
0.01
0.00
0.01
0.01
Barium
X
<0.07
0.50
0.44
<0.07
<0.07
<0.07
<0.07
0.47
<0.07
<0.07
<0.07
0.47
<0.07
0.41
<0.07
0.09
<0.07
0.12
(mg/1 )
s
0.04
0.01
--
0.04
--
0.04
0.02
0.04
0.04
58
-------
levels were noted in the dilute acetic acid extracts, levels in the nitric
acid extracts were approximately five times as high. Filtration of the nitric
acid extracts did not reduce the iron levels below those of their unfiltered
analogs. However, filtered acetic acid extracts did show substantially lower
iron values than their unfiltered counterparts. One possible mechanism to
explain this phenomenon is that the nitric acid is much better at actually
dissolving the stainless steel fragments abraded from the extractor walls,
and, as soluble iron, none is being lost to filtration.
Detectable levels of iron were also contributed to nitric acid extracts
by both of the filtration apparatus tested. This finding is consistent with
the premise that corrosion and dissolution of the stainless steel filter
apparatus occurs under the more acidic conditions.
Barium was detected only in those extracts (both nitric and acetic) which
passed through the filtration apparatus. This finding is similar to that
cited previously with acetic acid blank splits in which Ba was only detectable
in the extract split which was filtered.
The frequency and magnitude of blank contamination attributable to
extraction or filtration apparatus (either through sample carryover or direct
contribution, e.g. erosion and dissolution) was determined during the conduct
of the extractor comparison study (Table 32). The large numbers of samples
and blanks run in this study provided a good opportunity to examine the
relative background contributions of barium, chromium, and lead. The data in
Table 32 are for the blanks that accompanied the filtration and extraction of
Pond 12, Pond 0, and biosolids samples.
All of the acidic blanks that were exposed to extraction and filtration
apparatus showed levels of barium which exceeded detection limit concentra-
tions. None of the unfiltered blanks contained detectable barium levels.
These data clearly demonstrate the problem of barium contamination from the
filtration apparatus (prefilter) discussed previously. Note the high average
concentration of barium in blanks associated with the Pond 0 sample runs.
Unfortunately, neither "filtered-only" nor "extracted-only" data were
available for the Pond 0 samples to determine if contributions other than
those from the filtration apparatus are reflected in the high blank values
recorded.
All of the blanks that showed detectable chromium levels had been run
through blade-type rotary extractors. No chromium was detected in blanks from
either the tumbling-type extractor or wrist-arm shaker. There is some
evidence -- from the lower frequency of extraction blanks with chromium
detectable after filtering -- which suggests that filtration does remove a
portion of the chromium released from the stainless steel blade-type
extractor.
Although nearly 40 percent of the blanks run in association with Pond 12
samples showed detectable levels of lead, these levels only slightly exceeded
the detection limit and would be unlikely to represent a significant factor in
misclassification of a waste. No lead was detected in blanks associated with
either the biosolids or Pond 0 samples.
59
-------
TABLE 32. FREQUENCY AND MAGNITUDE OF CONTAMINATION OF BLANKS (0.1N ACETIC ACID) FKOM EXTRACTION
AND FILTRATION APPARATUS OBSERVED DURING EXTRACTOR COMPARISON STUDY
cr>
o
Blanks Exceeding Detection Limit Concentrations
Type of Corresjwnding Detection Limit
Process Blank* Wajs^ Run Metal (mg/1)
E pfz
E P12
E P12
EF P12
EF P12
EF P12
E BIO
E BIO
E BIO
EF BIO
EF BIO
EF BIO
EF PO
EF PO
EF PO
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
Ba
Cr
Pb
0.177
0.080
0.091
0.177
0.080
0.091
0.099
0.016
0.034
0.099
0.016
0.034
0.295
0.036
0.049
Average Cone
(mg/1)
0.544
0.122
0.104
0.635
0.141
0.110
NA
0.064
NA
0.672
0.050
NA
1.61
2.39
NA
Standard
Deviation
(mg/i )
0.320
0.060
0.014
0.169
NA
0.010
NA
0.035
NA
0.102
0.024
NA
0.29
1.12
NA
Frequency of
Occurrence
Number/Number Possible
0/26
_ 3/26
4/26
26/26
1/26
15/26
0/67
11/67
0/67
67/67
11/67
0/67
40/40
8/40
0/40
* E - Extraction only; EF = Extraction plus filtration
t Blanks accompanied processing of PO (Pond 0), P12 (Pond 12), and BIO (Biosolids) samples.
-------
ANALYTICAL PROCEDURES EVALUATION
Reproducibility
The proposed regulations require that atomic absorption methods be used
for analysis of the EP extracts for arsenic, barium, cadmium, chromium, lead,
mercury, selenium and silver. As described in the approach, these methods
were evaluated to determine their accuracy and reproducibility for analysis of
the EP extracts. Precision of the atomic absorption analyses of EP extracts
for barium, chromium, and lead is shown in Tables 17, 18a-b, and 21. The
highest relative standard deviations for the extract analyses are 3.1 percent
for chromium, 4.6 percent for lead and 16.4 percent for barium (Tables 17,
18a).
The large variability observed in data from the analysis for barium in EP
extracts is associated with the use of the nitrous oxide-acetylene flame.
This flame is more susceptible to performance variations than the air-
acetylene flame. Solids accumulating on the burner slit (observed when some
extracts were analyzed) caused the variability to be higher in the barium
analyses than in the chromium or lead analyses. Low recoveries for barium
spikes added to some extracts indicate an interference. A method of additions
calibration that incorporates a dilution step should effectively compensate
for a range of interference (such as suppression), although, as mentioned
elsewhere, precision can suffer if the dilution reduces the concentration of
analyte close to its detection limit.
Detection Limit/Sensitivity
Detection limit is used in this report to indicate the minimum
distinguishable concentration of an analyte. While the term sensitivity can
have the same interpretation, confusion can arise since sensitivity in atomic
absorption spectrophotometry refers to a concentration that produces 1 percent
absorption. In addition, the IUPAC defines sensitivity as the slope of the
calibration curve and not as the minimum detectable level.
For both flame atomic absorption and inductively coupled plasma emission,
the conventional procedure involves zeroing the instrument response on some
solution. Typically, the dilute acid used to stabilize the standard solutions
is employed to set the instrument zero level. The instrumental detection
limit for an analysis session is determined from repeated measurements on the
solution used to zero the instrument. Therefore measurement precision affects
the detection limit estimates. A conservative estimate of the detection limit
is obtained by multiplying the standard deviation obtained for the calibration
blank measurements by three, which requires a measurement to exceed the 99
percent confidence level before it is considered detectable. By interspersing
the measurements on the calibration blank solution among measurements on
samples, a more realistic estimate of the instrumental detection limit for an
analysis session is produced than by consecutive measurements.
61
-------
The Instrumental detection limit described above provides an estimate
that can vary depending upon the sample matrices. Samples with high dissolved
solids can produce fluctuations in the flame and plasma behavior that can
persist during measurements on the calibration blank solution.
Accuracy
For accurate analyses, the instrumental response to elements in the
calibrating solutions must be the same as for the elements in the sample
solutions. The comparability of standards prepared with 0.5 percent (v/v)
nitric acid (the calibrating matrix), with acetate buffer (pH 5.0, 0.2M), and
with selected waste extracts was evaluated by linear regression analysis.
Standard solutions of the three matrices -- at three concentration levels
that included the proposed toxic waste criteria levels (Federal Register,
December 18, 1978) — were prepared and analyzed for arsenic, barium,
cadmium, chromium, mercury, lead, selenium and silver using atomic absorption
spectrophotometry. Measurements made four times on each of the standard
solutions and on the appropriate spiked extracts provided data for linear
regression calculations. The values obtained were regressed against those of
the nitric acid standards to compute the regression slopes indicated in Table
33 and a statistical difference was indicated when the 99 percent confidence
intervals for the slopes did not include the value 1.0.
TABLE 33. SLOPE OF STANDARDS IN ACETATE BUFFER AND IN HAZARDOUS WASTE
EXTRACTS RELATIVE TO SLOPE OF STANDARDS IN NITRIC ACID
Standards in Acetate Buffer
Standards in HW Extracts
Element
Slope ± 99% C.I. Difference* Slope ± 99% C.I. Difference*
Arsenic
Ban urn
Cadmi urn
Chromium
Mercury
Lead
Selenium
Silver
0.918
1.041
0.975
0.986
1.003
0.873
1.019
1.016
±0.46
±0.023
±0.009
±0.015
±0.146
±0.028
±0.168
±0.022
-8%
+4%
-3%
No
No
-13%
No
No
0.896
1.014
1.023
1.011
0.976
1.055
0.898
1.016
±0.049
±0.021
±0.045
±0.017
±0.149
±0.023
±0.126
±0.039
-10%
No
No
No
No
+6%
No
No
*Statistical difference in slope relative to standards in nitric acid.
62
-------
For acetate buffer solutions, the slope values were within 8 percent of
1.00 except for Pb which was 13 percent low. Since other investigators have
reported the loss of Pb to container surfaces from solutions with pH >2, the
low slope for Pb in the acetate buffer (pH 5) is not unexpected.
None of the slope values obtained with waste extracts differed more than
10 percent from 1.00 when all data values were included in the regression.
However, quantification of Hg in EP extracts, by the method of standard
additions, could provide more accurate values than quantifications based on
standards prepared in either nitric acid or an acetate buffer. The Hg
responses in the waste extracts appeared to be suppressed at the lower
concentrations (slope was 15 percent low when highest Hg solution was
excluded). Such suppression would not be apparent at high spike
concentrations if the suppressant were "consumed" by mercury at lower levels.
The agreement observed between the Hg in acetate buffer and nitric acid
suggests suppression of instrumental response, rather than loss from the
solution.
Recovery of analyte added to sample extracts can provide an indication of
accuracy. Tables 34-41 show recoveries near 100 percent for many waste
samples, but low recovery for barium and lead has been found with undiluted
Pond 0 extracts. Recovery near 100 percent provides an indication (but no
guarantee) of accuracy. For example, the spike may be added in a form
different from the endogenous form, and consequently it may not reflect the
behavior of the endogenous analyte. Also, a suppressant may be consumed by
the endogenous analyte so that added analyte is not suppressed. Recovery data
will not indicate whether background absorption is contributing inaccuracy.
Use of a background corrector or a non-absorbing wavelength can compensate for
this type of interference. Digestion of EP extracts that contain organic
components might remove the suppression observed with the undiluted Pond 0
extracts. Digestion of EP extracts could also eliminate or minimize the
accumulation of solids on the burner head which can affect data quality and
even force interruption of the analysis (so the burner can be cleaned).
Accuracy will suffer if measurement interference is not detected and
corrected. Undiluted Pond 0 extracts (from second collection) exhibited about
a 2-fold suppression of the lead concentration. Where spike volumes did not
exceed 1 percent of sample volume, lead recovery was only 52 ± 4 percent.
However, dilution by the addition of 10 percent or more volume of spiking
solution obscured this suppression and recoveries approached 100 percent.
Calibration by the method of additions should be used to provide more accurate
data when spiked extract values differ significantly (±20%) from full
recovery. Enhancement as well as suppression of the spike may occur.
Evidence for response enhancement was found for chromium and selenium (Tables
35, 41). The suppression or enhancement of atomic absorption responses can
arise from components in the sample matrix that alter the fluid uptake rate,
nebulization efficiency and/or the atomization efficiency. High dissolved
solids can affect these characteristics; the dissolved solids content for an
undiluted Pond 0 extract was measured at 21.7 ± 0.3 percent (weight/weight).
While addition of lanthanum to samples has been found to minimize
suppression of the barium atomic absorption response for sediment extracts
63
-------
TABLE 34. QUALITY CONTROL DATA: COMPARISON OF BARIUM SPIKE RECOVERY
Sample
Site A, Pond 0
Site A, Pond P
Site B, Paint Sludge
Site 0, Chromate
Oxidation Paste
Site D, API Oil
Separator
Site D, Separator
Sludge
Site E, Filter Cake
Site G, Filter Cake
Site G, Asbestos
Cleanup Stream
Site H, CPI Decant Pit
Site I, Chlorine
Process Sludge
Site J, Catalyst Fines
Site K, Alkali Rust
Remover
Site K, Tin-Lead Waste
Blanks, Filtration
Apparatus, Water (2)
Blank, Extraction
Apparatus, Water
Blanks, Extraction Apparatus
Acetic Acid (2)
Blank, Extraction Apparatus,
Filtered, Acetic Acid
Blank, Sample Container,
Water
Sampl e
Cone.
(mg/1)
1.55
3.10
1.96
0.44
<0.6
0.80
0.80
<0.6
0.4
2.03
0.39
<0.6
2.43
1.02
0.10
0.10
<0.08
0.74
<0.10
Spike
(mg/1 )
2.00
2.00
2.00
2.00
2.00
5.0
10.0
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
Spiked
Cone.
(mg/1)
2.21
4.96
3.98
1.6
2.11
5.96
10.8
1.84
2.67
4.42
2.28
1.72
4.35
3.05
2.06
2.10
2.09
2.85
2.10
Spike
Recovery
(%)
33
93
101
60
105
98
100
92
114
120
94
86
96
101
103
100
104
106
105
RSD*
(Analysis)
(%)
5
11
33
43
31
1
1
7
4
3
23
13
2
2
2
1
3
2
5
* Relative Standard Deviation
64
-------
TABLE 35. QUALITY CONTROL DATA: COMPARISON OF CHROMIUM SPIKE RECOVERY
Sample
Site A, Pond 0
Site A, Pond 0
Site B, Paint
Sludge
Site D, Chr ornate
Oxidation Paste
Site D, API Oils
Separator
Site E, Blast Furnace
Filter Cake
Site E, Mill Scale
Site G, Asbestos
Cleanup Stream
Site G, Asbestos
Cleanup Stream
Site H, CPI Decant Pit
Site J, Industrial
Sewage Filter Cake
Site J, Industrial
Sewage Filter Cake
Site K, Alkali Rust
Remover
Site K, Tin-Lead Waste
Blanks, Filtration
Apparatus, Water (2)
Blank, Extraction Apparatus
Filtered, Acetic Acid
Blank, Extraction Apparatus
Acetic Acid
Blank, Extraction Apparatus
Acetic Acid
Sampl e
Cone.
Ong/1)
7.83
10.5
1.02
0.80
9.12
<0.3
<0.3
<0.3
<0.3
4.53
<0.3
4.25
4.38
7.92
<0.3
' 0.6
'<0.3
'<0.3
Spike
(mg/1 )
2.00
2.20
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.50
2.50
2.00
2.00
2.00
2.00
Spiked
Cone.
(mg/1 )
9.64
12.6
2.15
2.91
11.39
1.85
2.03
1.83
2.76
1.83
2.41
7.14
6.99
10.1
1.98
2.85
2.02
2.03
Spike
Recovery
(»)
90
95
108
106
113
93
102
92
138
92
120
144
104
89
99
112
101
101
RSD*
(Analysis)
(%)
10
8
6
3
9
8
8
0
3
0
4
3
0
1
0
1
1
6
Relative Standard Deviation
65
-------
TABLE 36. QUALITY CONTROL DATA: COMPARISON OF LEAD SPIKE RECOVERY
Sample
Site A, Pond 0
Site A, Pond 0
Site B, Paint Sludge
Site B, Paint Sludge
Site B, Paint Sludge
Site B, Paint Sludge
Site D, Chromate
Oxidation Paste
Site D, API Oil Separator
Site E, Mill Scale
Site G, Asbestos
Cleanup Stream
Site G, Asbestos
Cleanup Stream
Site H, CPI Decant Pit
Site J, Catalyst Fines
Site J, Industrial
Sewage Filter Cake
Site J, Industrial
Sewage Filter Cake
Blank, Extraction
Apparatus, Mater
Blank, Extraction Apparatus
Acetic Acid
Blank, Sample Container,
Water
Sample
Cone.
(mg/D
4.5
0.83
0.05
0.12
0.05
0
<0.02
0.14
<0.08
0.05
0.11
0.21
0.02
0.28
<0.12
<0.09
'<0.09
<0.02
Spike
(mg/i )
2.00
2.00
0.50
1.00
1.00
2.00
1.00
1.00
2.00
2.00
0.50
2.00
2.00
2.00
2.00
2.00
2.50
2.00
Spiked
Cone.
(mg/1)
6.69
2.94
0.62
1.10
1.25
2.33
1.14
1.28
2.05
2.37
0.43
2.47
2.23
2.59
2.13
2.11
2.54
2.00
Spike
Recovery
(*)
110
106
114
98
120
117
114
114
102
116
64
113
110
116
106
106
102
100
RSD*
(Analysis)
(X)
6
3
1
2
7
3
2
5
2
1
1
8
1
3
1
2
1
1
* Relative Standard Deviation
66
-------
TABLE 37. QUALITY CONTROL DATA: COMPARISON OF MERCURY SPIKE RECOVERY
Sample
Site B, Paint
Sludge
Site B, Paint
Sludge
Site B, Paint
Sludge
Sample
Cone.
(iig/D
1.31
1.31
1.31
Spike
(wg/i)
10.00
20.00
50.00
Spiked
Cone.
(ug/D
9.10
18.4
49.9
Spike
Recovery
78
86
97
RSD*
(Analysis}
(%)
1
0
0
* Relative Standard Deviation /
TABLE 38.
Sample
Site B, Paint
Sludge
Site B, Paint
Sludge
Site B, Paint
Sludge
QUALITY CONTROL
DATA:
Sampl e '
Cone. / Spike
(mg/ljX (mg/1)
<0.01
<0.01
<0.01
0.25
0.50
1.00
COMPARISON OF
Spiked
Cone.
(mg/1 }
0.24
0.47
0.98
SILVER SPIKE RECOVERY
Spike
Recovery
96
94
98
RSD*
(Analysis)
W
0
0
0
Relative Standard Deviation
67
-------
TABLE 39. QUALITY CONTROL DATA: COMPARISON OF ARSENIC SPIKE RECOVERY
Sample
Site B, Paint Sludge
Site B, Paint Sludge
Site B, Paint Sludge
Site E, Mill Scale
Site G, Asbestos
Cleanup Stream
Site I, Chlorine
Process Sludge
Site J, Catalyst Fines
* Relative Standard Deviation
TABLE 40. QUALITY CONTROL
Sample
Site B, Paint Sludge
Site B, Paint Sludge
Sample
Cone.
(mg/1 )
0.02
0.02
0.02
0.00
0.02
0.18
0.15
DATA:
Sample
Cone.
(mg/1 )
0.02
0.02
Site G, Mill Scale <0.0
Site 0, Industrial
Sewage Filter Cake
Site J, Industrial
Sewage Filter Cake
Blank, Extraction Apparatus
Blank, Extraction Apparatus
0.07
0.06
0.00
0.00
Spike
(mg/1)
0.20
0.50
1.00
0.50
0.50
0.50
0.50
COMPARISON
Spike
(mg/1 )
0.10
0.50
2.00
2.00
2.00
2.00
2.00
Spiked
Cone.
(mg/1 )
0.26
0.55
1.04
0.52
0.57
0.79
0.69
Spike
Recovery
(«)
115
106
102
104
110
122
108
OF CADMIUM SPIKE
Spiked
Cone.
(mg/1 )
0.12
0.54
1.99
2.02
1.96
1.94
1.99
Spike
Recovery
(%)
100
104
100
98
95
97
100
RSD*
(Analysis)
(*)
1
1
5
7
0
3
4
RECOVERY
RSD*
(Analysis)
(%)
2
1
1
1
1
1
1
* Relative Standard Deviation
68
-------
TABLE 41. QUALITY CONTROL DATA: COMPARISON OF SELENIUM SPIKE RECOVERY
Sample Spiked Spike RSD*
Sample Cone. Spike Cone. Recovery (Analysis)
(mg/1) (rag/1) (mg/1) (%) (%)
Site B, Paint Sludge <0.05 0.05 0.8 160 1
Site B, Paint Sludge <0.05 0.10 0.14 140 2
Site B, Paint Sludge <0.05 0.20 0.24 120 1
Site E, Mill Scale <0.08 0.50 0.68 136 7
Site G, Asbestos
Cleanup Stream 0.11 0.50 0.43 64 1
Site I, Chlorine
Process Sludge
Site J, Catalyst Fines
0.32
0.13
0.50
0.50
0.76
0.32
78
38
12
10
Relative Standard Deviation
(McKown et al., 1978), use of lanthanum with undiluted Pond 0 extracts did not
show a change in the suppression. All barium analyses were conducted with
1000 mg/1 potassium added for ionization control in the flame.
Dilution of EP extracts can minimize interference suppression and the
need for calibration by the method of additions. Pond 0 extracts (second set)
diluted 10-fold before flame AAS analysis showed spike recovery at 97 ± 1
percent. Recovery for lead was also near 100 percent for diluted extracts of
Pond 0 (first set) samples (Table 36). Table 42 shows suppressed and
unsuppressed lead data. Dilution of extracts can also reduce clogging
problems with the analytical instruments. The dilution technique is limited
by the criteria level and the instrument sensitivity.
Contamination can also arise from inadequate cleaning procedures for
equipment. Extractor blanks following some high-chromium wastes revealed some
cross-over contamination (as noted previously). Filters can contribute
contamination as observed for barium.
Accuracy can be jeopardized without the sample itself being contaminated.
When cleanout of the atomic absorption nebulizer and burner system between
solutions is not adequate, falsely high values can be obtained. This effect
is referred to as "memory" because cross-over contamination of the solutions
69
-------
TABLE 42. SUPPRESSION OF ATOMIC ABSORPTION Pb RESPONSE BY UNDILUTED
EP EXTRACTS OF POND 0 SAMPLES COLLECTED NOVEMBER 1979
Pb (mg/1)
Sample ID Undiluted Extracts Diluted Extracts*
10 mg/1 std
30 mg/1 std
163689W
163702N
163706R
163709W
163719N
163724R
10 mg/1 std
30 mg/1 std
9.97
30.09
19.92
18.99
19.38
19.81
18.84
18.34
9.94
29.51
10.13
30.07
40.90 ± 0.53
39.80 t 0.89
35.35 ± 0.45
39.50 ± 1.42
29.85 ± 0.80
29.70 t 0.71
10.20
30.52
* Corrected for 10-fold dilution.
per se is not involved. Waste extracts analyzed (using a Perkin-Elmer Auto
Sampler 200) for lead immediately after high-lead extracts, exhibited the
"memory" effect --near 1 mg/1 -- that would have resulted in falsely
classifying the wastes as hazardous (using the proposed (FR 1978) extract lead
criterion level of 0.5 mg/1). When lead at 5 mg/1 in acetate buffer was
analyzed, the succeeding measurements exhibited "memory" that exceeded the
proposed 0.5 mg/1 criterion level (0.63 and 0.58 mg/1). No such effect was
found when the 5 mg/1 lead was dissolved in nitric acid or an EP extract for
analysis. Although a memory effect of 0.15 mg/1 was observed after 10 mg/1
chromium (in 0.5 percent nitric acid) —with an air-acetylene flame --"memory"
was not observed following an extract containing 39 mg/1 chromium during the
same analytical session. Instrumental interference from sample "memory" can
be eliminated by appropriately scheduled rinses or by prolonging the solution
uptake before the response is measured.
70
-------
Problems
Extracts that are high In dissolved solids can clog the atomic absorption
burner and prevent analyses (the 3-slot air-acetylene burner Is more
resistant to clogging than the single slot burner). Even when insufficient to
prevent measurements, burner clogging can cause data quality to suffer.
Dilution of extracts (with 0.5 percent nitric acid) can reduce clogging
behavior, but may also reduce the concentration below the detection limit.
Use of the furnace atomic absorption technique would be required for such
cases. Digestion of organic-containing extracts could reduce clogging
problems. Chelate-extraction may prove necessary for extracts with high
inorganic salt contents.
Suppression interference has been observed for barium and lead in some
extracts. Use of the method of additions is indicated for such cases.
Dilution can offer an alternative to the method of additions in some cases.
For example, ten-fold dilution of Pond 0 extracts with 0.5 percent nitric acid
essentially eliminated interference suppression for lead but not for barium.
Of course, dilution does require more sample preparation. Mercury analysis was
not a problem, except for evidence that the method of additions may be
required for some extracts. However, low extraction of mercury could be
considered a problem as previously discussed.
Background (or non-analyte) absorption is a potential problem. Soil
digests analyzed for cadmium -- without deuterium lamp background correction
— have exhibited values more than six-fold higher then the corrected values.
71
-------
REFERENCES
Brown, D. K., M. P. Maskarinec, and F. W. Larimer. (Unpublished EPA report).
Comparison of EP Extracts and Landfill Leachates. Submitted by Oak Ridge
National Laboratory, Oak Ridge, Tennessee under EPA-IAG-78-D-X0372.
Municipal Environmental Research Laboratory, Cincinnati, Ohio.
Cochran, W. G. 1963. Sampling Techniques, 2nd Edition, John Wiley and Sons,
New York.
deVera, E. R., B. P. Simmons, R. D. Stephens and D. L. Storm. 1980. Samplers
and Sampling Procedures for Hazardous Waste Streams. EPA-600/2-80-018.
Municipal Environmental Research Laboratory, Cincinnati, Ohio.
DeWalle, F. B., T. Zeisig, J. F. C. Sung, D. M. Norman, J. B. Hatlen, E. S. K.
Chian, M. G. Bissel, K. Hayes, and D. Sanning. (In press).
Analytical Methods Evaluation for Applicability in Leachate Analysis,
Municipal Environmental Research Laboratory, Cincinnati, Ohio.
Environmental Protection Agency. 1979. Methods for Chemical Analysis of
Water and Wastes. EPA-600/4-79-020. Environmental Monitoring and
Support Laboratory, Cincinnati, Ohio.
Environmental Protection Agency. 1980. Test Methods for Evaluating Solid
Waste: Physical/Chemical Methods. SW-646. Office of Water and Waste
Management, Washington, D.C.
Federal Register. 1978. Hazardous Waste Guidelines and Regulations. Volume
43:243 - Monday, December 18, 1978.
Federal Register. 1980. Hazardous Waste and Consolidated Permit Regulations.
Volume 45:98 - Monday, May 19, 1980.
McKown, M. M., C. R. Tschrin and P. P. F. Lee. 1978. Investigation of Matrix
Interferences for AAS Trace Metal Analyses of Sediments.
EPA-600/7-78-085. Environmental Monitoring and Support Laboratory,
Cincinnati, Ohio.
72
-------
APPENDIX 1.
USE OF THE POND SAMPLER AND THE COLIWASA
POND SAMPLER
The pond sampler Is a glass beaker (usually of 250- to 800-ml capacity)
affixed with a clamp to the end of an aluminum handle that is adjustable from
8 to 15 feet in length (Figure A-l). It is used to collect liquids and
semi sol ids from ponds, pits, and lagoons. Two persons are required for
sampling; both must wear appropriate personal safety equipment. Samples can
be taken at or below the surface. The stepwise procedure for use of the pond
sampler is presented below.
• Make sure the beaker is acid washed (and acetone rinsed) and the
sampler is assembled properly. (Nuts and bolts must be tight to
secure the beaker, pole, and clamp.)
• Sample at that depth and distance prescribed in the sampling plan.
To collect a sample, lower the beaker into the pond in an inverted
position. At the prescribed depth, turn the handle to upright the
beaker. Raise the sampler clear of the pond surface and swing it
carefully back over the shore avoiding spillage.
• Pour the sample into the sample container slowly to avoid splashing
sampling personnel or losing sampler contents.
• Clean the sampler after each sample. When subsamples for composites
are being taken from the same pond, cleaning can usually be
accomplished by rinsing the beaker in the pond close to (but not
disturbing) the succeeding sampling site. Between locations
containing different wastes, disassemble the sampler, wipe off oily
or clinging solvent material, flush the handle and clamp with
freshwater and replace the beaker with a fresh acid-washed,
acetone-rinsed beaker or perform the wash and rinse on site.
• Close the container in which the sample is to be shipped, record all
information in the logbook and on appropriate forms after each
sample, and attach the proper labels and seals to the sample
container.
• Clean sampler thoroughly on site if possible, and pack away. If
sampler must be returned to the laboratory for cleaning, dismantle
sampler and pack parts in plastic bags to avoid contamination of
personnel, vehicles, or other equipment.
73
-------
Beaker: Glass or
Polypropylene
(250-800 ml)
Heavy Duty Aluminum Pole
(telescopes from 250 cm to 450 cm length)
Figure A-l. Pond sampler.
Modified from de Vera et al., 1980. Samplers and sampling procedures
for hazardous waste streams. EPA-600/2-80-018, Municipal
Environmental Research Laboratory, Cincinnati, Ohio.
COLIWASA
The COLIWASA (Composite Liquid WAste SAmpler) is a tube-type sampler, 5
feet long and between 1-3/8 and 1-5/8 inches in diameter (ID) (Figure A-2).
It can be fabricated from various materials to sample almost any kind of
liquid waste from drums, barrels, or vacuum trucks. A stepwise procedure for
use of the COLIWASA is presented below.
• The operator must make sure the COLIWASA is clean, functioning
properly, and the stopper fits tightly. Adjust the stopper rod
length if necessary.
• Two persons are required for sampling; both must wear appropriate
personal safety equipment. Company personnel must have already
opened the waste storage vessel.
• Open the T-handle and push it down so it lies on the locking block
forming a T between the handle and sample tube.
74
-------
tapered
stopper
— >||< — 2.86cm(l-I/8")
4
n
II
II
ll
Ji
1
1
1
i
!
II
i
!
•
ji
!
•
|!
!|
!!
n
4*
=1
6
1
I
Locking
-jJ5cm(2-l/2") blocfc >
.52m(S'-0")
1
•tr
i!
i}
i
i
i
i
j
!
i!
i
j
!
i
|
|
M
II
M
II
di
I7.8cm(7")
i
T
Stopper rod, PVC
" 0.95cm(3/B")O.D.
Pipe, PVC, 4.13cm(l-5/8")l.D
* 'i.26cm(l-7/B")O.D.
Stopper, neoprene, 09 with
SAMPLING POSITION
CLOSE POSITION
Figure A-2. Composite liquid waste sampler (COLIWASA).
(Modified from deVera, et al., 1980)
-------
Introduce the sampler carefully into the waste storage vessel,
maintaining the sampler in a vertical attitude. Lower the sampler
slowly to allow liquid levels within and outside the sampling tube
to equilibrate as the sampler approaches the greatest depth of
sampling. This is necessary to allow the taking of representative
samples, especially with more viscous liquids.
When the sampler contacts the bottom of the waste storage vessel, or
when all but 6 inches of the sampler is immersed, pull upward on the
handle and turn it until one end rests firmly on the locking block.
Withdraw the sampler with one hand and carefully wipe the outside of
the tube with a reinforced fiber paper towel in the other hand; the
second person may have to do this.
Place the sampler directly above the sample container and slowly
open the T-handle to release the sample through the bottom and into
the sample container.
Close the sample container, record all information in the logbook
and on appropriate forms, and attach the proper labels and seals to
the sample container.
Immediately clean the sampler. Recommended cleaning precedures
include use of a large, long-handled "bottle brush" which snugly
fits the inside of the COLIWASA tube. In turn, the sampler is
washed with an organic solvent, e.g., hexane, and detergent
solution, and flushed with water. Any oily sample residues should
be wiped off (e.g., with fiber reinforced paper towels) before
initiating the wash/rinse sequence.
76
-------
APPENDIX 2
PROTOCOL FOR CONDUCTING EXTRACTION PROCEDURE TO
IDENTIFY HAZARDOUS WASTE
This is a summary of the protocol followed in the evaluation of the
extraction procedure for identification of hazardous waste. A flow chart for
this protocol is given in Figure 1 in the text.
Safety Note
Laboratory personnel are required to wear safety glasses, rubber or
vinyl gloves, a lab coat or coveralls and safety shoes or rubber
boots when handling hazardous waste samples. Respirators are to be
worn when there is a possible hazard due to toxic gases or vapors
from the sample. All contaminated dry waste materials (excess
samples of dry waste, paper towels, disposable beakers, etc.) are to
be sealed in plastic bags and placed in cardboard boxes for proper
disposal. Used solvents and other contaminated liquid wastes are to
be sealed in metal, glass, or plastic containers, as appropriate,
and stored in a closed hood or sealed drum (in a restricted area)
until disposal. All waste containers are to be labeled "Hazardous
Waste" and must include the type of waste, the date, and the
worker's initials on the label. All hazardous wastes are to be
disposed of by a commercial contractor at a disposal facility
approved for such wastes.
Treatment Prior to Extraction
Triplicate aliquots (100 grams each) of each waste sample are separated
into solid and liquid phases by filtration. If the sample is a liquid but
cannot be filtered through a 0.45 micron filter, it is centrifuged to obtain
phase separation. If neither filtration nor centrifugation will separate the
material into solid and liquid phases, the sample is treated as a solid.
Weight Determinations
The filters to be used for the filtration step are not removed from their
packaging until they are weighed to determine their tare. They are then
stored on clean watch glasses or in petri dishes until they are used for
filtration of the sample. The analytical balances used to weigh the filters
are calibrated monthly with standard weights and are checked with a standard
100-mg weight just before each weighing. The date and results of the
calibration are recorded in the balance logbook. An annual calibration with
77
-------
standard weights traceable to the National Bureau of Standards is performed
when the balances are cleaned and serviced. The EPA property number of the
balance is recorded in the laboratory notebook with the data obtained with
that balance.
Filtration Method
All filtrations are performed in a fume hood to protect the operator from
any toxic vapors that may emanate from the sample. A Nuclepore filter holder
(Nuclepore Corp., Pleasanton, CA 94566) equipped with a 1.5 liter reservoir is
used for the filtration in the following steps:
1. Place a weighed glass fiber pre-filter (124 mm diameter, Millipore AP
25124, Nuclepore P040, or equivalent) and a weighed 0.45 micron filter
membrane (Millipore type HAWP 142, Nuclepore type 112007, or equivalent)
in the filter holder with the pre-filter on top (upstream).
2. Add the sample (known weight) to the reservoir. Seal the reservoir
and pressurize it with argon to a maximum of 75 psi. Continue the
filtration until less than 5 ml of liquid is released during a 30 minute
period. The sample may not require the maximum pressure for
filtration; however, for some dense samples the reservoir must be held at
75 psi before the sample is identified as non-filterable.
3. After liquid flow stops, depressurize and open the top of the
reservoir. Remove the filters and solid sample and place in a petri dish
or other suitable container. Repeat steps 2 and 3 if the sample size
exceeds the capacity of the reservoir.
4. Store the liquid fraction at 1-5°C for later addition to the extract.
5. Weigh the filtered solid sample (filters included) to determine the
weight of the solid material collected (i.e. subtract tare weights of
filters from total sample weight). Extract the filtered sample (solid
material a filters) by the extraction procedure.
6. If the sample does not filter, use the centrifugation method to
separate the solid and liquid phases.
Centrifugation Method
An International Centrifuge, size 2, model K (International Equipment
Company, Boston, Mass.) is used for the centrifugation in the following steps:
1. Centrifuge the sample for 30 minutes at 2300 rpm under controlled
temperature (20 to 40°C).
2. Measure the size of the liquid and solid layers to the nearest mm
(0.04 inch) and calculate the liquid-to-solid ratio.
78
-------
3. Repeat steps 1 and 2 until the liquid-to-solid ratios for two
consecutive 30-minute centrifugations agree within 3 percent.
4. Decant or siphon off the liquid layers and extract the solid by the
extraction procedure. Store the liquid fraction at 1-5°C for use in the
extraction procedure.
Extraction Procedure (EP)
The solid material, obtained by the filtration or centrifugation method
from liquid samples or as an aliquot from solid samples, must be able to pass
through a 9.5 mm (3/8") standard sieve. If the sample will not pass through a
9.5 mm sieve it must be ground to size or must be subjected to the structural
integrity procedure (Federal Register, Vol. 43, No. 243, Dec. 18, 1978).
The extraction procedure is performed in the following steps:
1. Weigh the solid material obtained from the waste sample and place it
in an extractor as identified in the proposed regulation. A suitable
extractor will keep the solids in suspension in the extracting solution
so that sample surfaces are continuously brought into contact with
well-mixed extracting solution. With the exception of special studies,
either the blade or tumbling type extractor was used for this program.
2. Add to the extractor a weight of deionized water equal to 16 times
the weight of solid material added to the extractor.
3. Agitate the sample at 40 rpm and adjust the pH of the solution to
5.0 ± 0.2 with 0.5N acetic acid. Maintain the pH at 5.0 ± 0.2 and
continue agitation for 24 hours. Do not add more than 4 ml of acid for
each gram of solid. If the solution pH is less than 5, do not add any
acid during the extraction. Maintain the temperature of the solution at
20-40°C during the extraction. Follow the procedure for manual pH
adjustment in the proposed regulations.
4. Measure and record the pH at the end of the 24-hour extraction
period.
5. At the end of the 24-hour extraction period separate the liquid and
solid fractions of the extraction material by the filtration method
described above. Adjust the volume of the resulting liquid phase with
deionized water so that its volume is 20 times that occupied by a
quantity of water at 4°C equal in weight to the initial quantity of solid
material placed in the extractor.*
* Note - Beginning with the municipal sludge sample study -- the last study
addressed in this report -- the filtered solid samples (filter cakes) are no
longer dried to constant weight prior to extraction. A separate aliquot is
now used to determine percent filtrable solids content of the waste. The
aliquots to be extracted are simply filtered until all excess moisture is
removed and then either extracted immediately or stored in petri dishes to
maintain their moist condition until extraction.
79
-------
6. Combine this solution with the original liquid phase obtained In the
filtration or centrifugation step. Mix thoroughly and split the combined
solution into two equal samples. Store one sample in glass under
refrigeration at 1-5°C for organic analysis. Preserve the second sample
for elemental analysis by addition of Ultrex nitric acid to reduce the
sample pH to less than 2.
Analysis
The samples obtained by the extraction procedure are analyzed in
accordance with the methods given in the proposed regulations. All samples
should be analyzed as soon as possible after extraction.
80
-------
APPENDIX 3.
WILLIAMS/BECKERT DRUM AND TANK (DAT) SAMPLER
DRUM AND TANK (DAT) SAMPLER
Use of DAT Sampler
The following steps describe the use of the DAT Sampler:
1. Support frame (Figure A-3) is inserted into the drum or tank to
be sampled and slowly lowered to the desired depth (usually
the bottom).
2. After a short delay ~ 10 to 30 seconds, depending on the
viscosity of the material being sampled -- the collection tube
is inserted into the support frame and pushed down slowly until
it slips over the 0-ring in the frame base. The sample thus
trapped is representative of the full liquid column (surface to
bottom or base of sampler) sampled.
3. The support frame, collection tube and sample are removed as a
unit from the drum or tank. Wipe with rag.
4. The bottom of the sampler is next inserted into the mouth of a
wide-mouth collection jar. The contents are then released into
the jar from the bottom of the sampler by carefully lifting the
collection tube free of the 0-ring in the support frame base.
5. Discard disposable collection tube; remove 0-ring; wash base
(still assembled) with detergent solution and rinse with
acetone, then distilled water. Either replace 0-ring with new
one, or wash and rinse (detergent solution, then acetone,
distilled water) old 0-ring and replace on base.
Assembly of DAT Sampler Support Frame
For the 4-foot drum-sampling configuration (Figure A-3) each of the three
4-foot support tubes are threaded securely onto studs in the sampler base.
The "male" end of each tube is inserted through one of three 1/4-inch holes in
the top flange. A stainless cap nut is threaded onto the end of each of the
tubes to secure them to the top flange.
To obtain the 6-foot sampler configuration, directly replace the 4-foot
support tubes with their 6-foot counterparts.
81
-------
Collection Tube
Inserts Into
Support Frame
Here
XaX
O-Ring Seal
(seals fully
inserted
collection tube) Support Frame
(stainless steel)
Type 316
Collection Tube
(disposable PVC,
stainless steel, or
other materials)
Figure A-3. Williams/Beckert Drum and Tank (DAT) Sampler
in drum sampling configuration.
82
-------
To assemble a frame that will support uncut 10-foot PVC collection tubes.
remove the cap nuts from top of the 4-foot support frame (but leave the
existing top flange) and, 1n place of the cap nuts, thread on the 6-foot
extension tubes. Add a new top flange over the male ends of the extension
tubes and secure with the cap nuts removed earlier. If the original top
flange 1s deleted In assembling the 10-foot configuration, the resulting
support frame will not have adequate strength.
Advantages of the DAT Sampler
The design of the DAT Sampler offers the following advantages:
• Support frame Is constructed entirely of corrosion resistant
materials (Type 316 stainless steel).
• Choice of sampling tube materials Includes Inexpensive, disposable
PVC pipe readily obtainable from local sources, and reuseable
stainless steel or other tubing for specific sampling requirements.
• Samples to within approximately 1 cm of the bottom of a waste
container to yield a sample representative of the full liquid column
(top to bottom).
• Extension tubes for support frame -- when used with collection tubes
of various lengths -- give flexibility for sampling from shallow
(drum) and deeper (tanks, vacuum trucks) waste containers.
• Lightweight, take-down design of support frame stores compactly,
cleans and transports easily.
• Accurate, positive closure 1s obtained with 0-r1ng seal and
collection tube guidance system.
• Sample size 1s convenient for compositing representative samples
from multiple waste containers.
• The combination of low cost, disposable collection tube and 0-r1ng
design eliminates cross-contamination of samples and minimizes
on-slte cleaning requirements.
• Virtually any liquid waste that can be safely exposed can be sampled
with some combination of collection tube length and material.
• Triangular support frame configuration 1s more easily passed through
out-of-round or damaged drum openings.
• Support frame can be grounded to steel drums or tanks where
accumulation of explosive vapors 1s suspected.
• Sampler 1s simple 1n design and should be easy to operate even when
suited 1n full protective clothing.
83
-------
The lightweight, thin-walled collection tubes are smooth,
unobstructed, and easy to control and therefore cause minimum
disruption of sample stratification when Inserted.
84
------- |