905R79118
5440
Evaluation of the Procedures for Identification of Hazardous Waste
Interim Report - August 1979
by
Eugene P. Meier, Llewellyn R. Williams, Robert G. Seals, Lawrence E. Holboke
and David C. Hemphill
SUMMARY
This is an interim report for the first four months of the hazardous waste
studies being performed at the Environmental Monitoring Systems Laboratory-
Las Vegas. The majority of this initial phase of the study was spent in
collecting samples and establishing the laboratory analytical operation.
Eleven manufacturing and waste disposal sites were visited and 26 different
wastes were sampled. Although limited analytical data is available for this
report, some tentative observations are presented and should be useful in
preparing the guidelines and regulations for hazardous waste. The initial
data .indicate that the pond sampler gives reproducible results when used in
accordance with the proposed regulations. Data obtained also indicate that
the proposed extraction procedure is reproducible and should be acceptable for
identification of hazardous waste. Some problems were encountered with the
analytical procedures for barium and mercury. These problems are being
investigated and will be discussed in more detail in the final report. The
results and conclusions in this report are interim findings and are subject to
change as more information and data are collected in this study.
INTRODUCTION
Background
The rapid technological advances in industry over the past several decades
have significantly improved the American economy and lifestyle. However, the
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
-------
-------
.transportation and management of hazardous waste. Under the authority of this
leqislation 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 proposed Section 3001 regulations provide specific procedures for
sampling, extraction, and analysis of wastes to identify those wastes which
are hazardous due to the presence of leachable 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.
Objectives
The objectives of this study are 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.
• Initiate a quality assurance program to support the
hazardous waste monitoring efforts that will result from
promulgation of the hazardous waste regulations.
Scope of the Study
The program for FY 79 and FY 80 is being performed in four individual
tasks.
• Task 1. Evaluation of the Proposed Sampling Procedures. This effort
is evaluating the safety, reliability and reproducibility of the
sampling procedures in the proposed Section 3001 regulations. The
COLIWASA and pond sampler methodologies identified in the draft report,
"Sampling Procedures for Hazardous Waste Streams" (EPA Grant No.
R804692010), are being used to collect waste samples at typical waste
sites. Any problems with the sampling procedures when used in the
field are noted and will be reported. The waste samples are then
delivered to the laboratory for analysis by physical and chemical
methods to determine the reproducibility of the sampling procedures.
The procedures to be evaluated and wastes to be sampled are
determined jointly by the Office of Solid Waste (OSW) and the
-------
Environmental Monitoring Systems Laboratory-Las Vegas (EMSL-IV). If
necessary, modifications to the sampling procedures may be developed
and evaluated.
• Task 2. Evaluation of the Proposed Extraction Procedure (EP) and
Analytical Methods. This effort is designed to evaluate the
reliability and reproducibility of the extraction procedure (EP)
described in the proposed Section 3001 regulations (Section 250.13(d)).
It is also designed to evaluate the analytical methods that are
proposed for use with the EP to determine their accuracy and precision
when applied to the EP extracts from a variety of wastes. The number
and priority of wastes to be used in the evaluations are determined
jointly by the OSW and the EMSL-LV. If necessary, modifications to the
EP and/or analytical methods will be recommended to the OSW and may be
developed and evaluated.
• Task 3. Analysis of Specific Wastes from the Proposed Hazardous Waste
List. The objective of this effort, to be performed in FY 80, is to
characterize the wastes that have been placed on the hazardous waste
list, in the proposed regulations. Samples of each waste will be
collected, extracted, when necessary, and analyzed to determine the
identity and concentrations of the hazardous components of these
wastes.
• Task 4. Development of a Quality Assurance Program for Hazardous Waste
Monitoring. The objective of this efort is to develop and coordinate a
national quality assurance program for hazardous waste monitoring.
Efforts in FY 79 and FY 80 will:
1. Provide standard reference materials.
2. Initiate a laboratory intercomparison program.
3. Establish minimum laboratory standards and practices.
4. Establish a laboratory evaluation program.
5. Develop protocols for evaluation of equivalent methods.
APPROACH
Rationale for Waste and Waste Site Selection
The wastes and sites to be sampled were selected with the active
assistance of industrial and government facilities that generate a variety of
both hazardous and nonhazardous wastes. The criteria for selecting a waste to
be sampled 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 nonhomogeneous,
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 1-liter samples to be withdrawn without appreciably
decreasing the amount of waste remaining (i.e., at least
800 liters), and
5. should 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 extraction procedure (EP), it was determined that the
samples should have the following characteristics:
1. be able to be subdivided into subsamples of 100 gm size
without introducing significant variability due to the
subsampling procedure,
2. should 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 in the
extract,
3. should contain 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 solid.
For evaluating the analytical procedures it was determined that the samples
should have the following characteristics:
1. must be from typical waste streams that are complex
in nature,
2. should contain one or more of the materials in 250.13(d)
(1) (43 FR 58956),
3. be able to be subdivided into samples of 100 gm 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 use these criteria, because a priori
information on the wastes was not available or no wastes with the desired
characteristics were available from the facilities visited. Because of
concerns about the proprietary nature of the industrial process 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).
-------
Sites Visited and Wastes Sampled
During the first phase of this study, eleven manufacturing and waste
disposal sites were visited with 26 different wastes being sampled (Table 1).
Brief descriptions of each site and waste-follow:
Site A. This is a waste disposal facility that segregates its wastes by
type of industry. The liquid wastes are placed in open ponds where waste
volume is reduced by evaporation. In some cases there is movement of waste
material from one pond to the next (in reality the ponds are not all
segregated). 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 cou"!d be separated by the proposed centrifugatron
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).
TABLE 1. SITES SAMPLED DURING FY 79: IDENTIFICATION OF SITES,
FACILITY FUNCTIONS, AND WASTE STREAMS SAMPLED
Site Function of Facility Waste Stream Sampled
A Waste, disposal Ponds of Ti02 process waste
Pond of alkaline waste
Pond of sulfonation tars
B Paint manufacture Drum of paint sludge
C Chemical manufacture Drum of laboratory wastewater
Bags of pesticide waste
D Chemical manufacture Inlet, grit chamber, and pond of an
API oil separator
Dumpster of chromate oxidation paste
E Steel manufacture Waste dust pile from electric
furnace baghouse
(continued)
-------
TABLE 1. (Continued)
Site
Function of Facility
Waste Stream Sampled
I
J
K
Chemical manufacture
Chemical manufacture
Petrochemical manufacture
Chlorine manufacture
Chemical manufacture
Paint removal/
electroplating
Filter cake waste from blast
furnace scrubber
Waste roll rrrill scale pile from
water treatment plant
Tank truck of lime sludge from
ammonia still
Alkyl chloride storage pit
Epichlorohydrin waste sump
Polymerized epichlorohydrin
waste pits
Filter cake from chlorine/mercury
process stream
Dumpster of asbestos waste,
clean-up from chlorine process
CPI* decant pit
Activated biosludge
Waste chlorine sludge pile
Industrial sewage filter cake
from a truck
Tank of chrornate reduction
clarifier underflow
Drum of catalyst fines
Drums of plating waste
identified as tin-lead solution
Drums of alkali rust remover (red)
Drums of oil/water organic solvent
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; had a high viscosity;
had the odor of typical oil-based paint solvents; and, could not be filtered
or centrifuged in accordance with the proposed procedures.
Site C. Two wastes were sampled at this chemical manufacturing facility.
One was an acidic (pH =* 1) laboratory waste 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 Tayer 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 0. Samples were collected with a pond sampler from an API oil/water
separator and a dumpster of chromate oxidation paste at this chemical
manufacturing facility. The samples collected from the API separator
contained oily dark-brown solids mixed with water (pH = 8). The chromate
oxidation paste samples contained a clear liquid layer (pH a 7.5) and a layer
of brown solids (approximately 10% solids).
Site E. Four wastes were sampied 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 powdered and 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-mill
scale, from one of the facility's water treatment plants, was sampled with a
plastic bottle. The sample was a mixture of crystalline solids (large and
small pieces) that had a disagreeable odor. (4) A tank truck of lime sludge
from an ammonia still was the final waste sampled at this site. The 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
alky! chloride storage pit, was a rust-brown "liquid (pH = 7) with 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 (pH =* 12).
Site G. Two waste samples were collected with a small trowel from a
chlorine-mercury process stream at this chemical manufacturing facility. The
filter cake waste sample had two phases, water (pH = 5.6) and a light-brown
solid. The second waste sample from this process contained asbestos solids
and water (pH = 9.8).
-------
Site H. An activated biosltflp and a waste from a CPI (Chemical
Production Industries) decant pttwere sampled at this petrochemical facility.
The biosludge sample (taken frcraa faucet in the pipeline from the settling
tank) was a brown liquid (pH = &£} that contained a high concentration of
suspended solids. The decant pttwaste was from an adjacent oil refinery
process stream. It was sampled »th a pond sampler and yielded a black, oily
liquid sample that had a disagreafcle odor. The high organic concentration of
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 sample was a dark-
gray, .paste-like solid.
Site J. Three wastes were sapled at this chemical manufacturing
facility. (1) A sample of an indtetrial sewage filter cake was obtained from
an open truck with a small trowel. (2) A sample (pH = 8.7) of liquid waste
from a. chrcmate reduction clarifiar 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 fwth 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 faint removal and electroplating operation
were sampled at this facility. Be wastes were stored in drums and were
awaiting disposal by a commercial disposal company. The samples were obtained
with the COLIWASA. (1) A platinfwaste identified as a tin-lead solution,
yielded greenish-yellow acidic (p
-------
Sampling Procedure Evaluation
Rationale
In designing the sampling procedure evaluatiae, three general guidelines
were followed:
1. The study is designed to evaluate only tiase methods that were
specifically developed for this regulatory program (e.g., the pond
sampler and COLIWASA).
2. The wastes to be sampled are to be selected from among those materials
that would be most difficult to sample (i-e., a worst case situation
such as a multiphase waste that contains inmisclble liquids and solids
of differing density and particle size).
3. The samples are to be obtained by personis! who are not knowledgeable
in the variability or detailed physical or chemical characteristics of
the specific wastes in order to simulate son-expert use of the
sampling methodology.
The sampling methods in Appendix I cf the prcgosed regulations include:
(1) ASTM methods and (2) protocols from a draft EJR report, "Sampling
Procedures for Hazardous Waste Streams," EPA Grant No. R804692Q10. The ASTM
procedures are not being evaluated because the Agorcy assumes that they are
standard procedures and have undergone sufficient evaluation through general
use with materials of the type indicated in Appeafix I of the proposed
regulation. However, the protocols described inlfce draft report have not
been evaluated for sampling wastes under field coalitions. Since the pond
sampler and COLIWASA had not been evaluated underfield conditions and will
have significant use in supporting the regulation^ they were selected for
evaluation in this study. The samplers are being used in accordance with the
protocols in the draft report and are described mAppendix 1 of this report.
The study is designed to determine the ability of these procedures to obtain a
representative sample from a given waste source. The procedures are not being
tested for use in sampling a waste over a period rf time (i.e., to determine
if the waste source changes with time).
Experimental Design
The experimental design for the sampler evaluations was developed with the
recognition that it is very difficult to obtain representative samples front a
heterogeneous waste source, especially in the caszof large disposal ponds or
pits. In the extreme case, it would be necessary to analyze an entire pond of
waste to determine what combination of samples waffd be required to
characterize the pond's composition. Of course, 16is is impractical; and, a
statistical approach must be taken to obtain and »alyze the minimum number of
samples that are required to assure, with a state*degree of confidence, that
the sample data is representative of the waste soiree. The initial
experimental design was based on a one-sided paraaetric test that assumed a 4%
significant deviation. This design required 39 saples from the source (i.e.,
-------
39 satples/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
miniwzed 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
variaare (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 four
different waste sources at two sites. The COLIWASA was used to sample five
different waste sources at three sites. The overall strategy of this effort
was tedetermine the ability of the sampling procedures to collect repro-
duciffie samples from a given waste source. There was no attempt to determine
the dfcanges that occur if a waste is sampled over a given period of time (i.e.
all sanples from any single source were collected on the same date).
TABLE 2.. WASTES SAMPLED USING THE POND SAMPLER AND COLIWASA
Location
Waste
Sampling Regimen
Site I.
Site f
POND SAMPLER
process waste
Alkaline waste
Sulfonation tars
Polymerized epichlorohydrin
waste
2 ponds, 10 samples/pond
1 pond, 10 samples
1 pond, 10 samples
2 pits, 20 samples/pit
SiteC
Site*
Site I
COLIWASA
Laboratory wastewater
API separator waste
Plating waste, tin/lead
Alkali rust remover
Oil/water/sol vent waste
4 drums, 3 samples/drum
15 drums, 3 samples/drum
3 drums, 3 samples/drum
3 drums, 3 samples/drum
3 drums, 3 samples/drum
10
-------
For the pond sampler, duplicate samples were taken at each of several
locations in the pond or pit. The sampling locations were randomly selected;
however, they were restricted to locations that could be safely accessed by
the sampler operator. The number of samples obtained by the COLIWASA was
restricted by the number of drums of waste available for sampling. In all
cases triplicate samples were obtained from each drum. The samples were
returned to the laboratory where duplicate or triplicate aliquots of the
samples are being analyzed for chemical and physical parameters such as pH and
percent solids (weight of filterable solids/weight of aliquqt).
Percent solids and pH were initially selected for testing the reproduc-
ibility of the sampling procedures because these parameters were easy to
determine and represented sample properties that effect the chemical data
obtained by the EP. In most cases, waste samples are heterogeneous with
respect to such properties as percent solids or aqueous:non-aqueous
composition. 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. When practical, the
percent solids, immiscible phase composition, or some other easily obtained
physical parameter will be used to evaluate the samplers.
The pH of the aqueous phase of each liquid sample is measured with a
laboratory pH meter. The pH meter is calibrated with standard buffers at
pH 4, 7, and 10 just prior to the measurements and rechecked after the
measurements are completed. Percent solids were determined in accordance with
the "Non-filterable Residue Method 160.2," Methods for Chemical Analysis of
Water and Wastes, EPA-600/4-79-020, Environmental Monitoring and Support
Laboratory, Cincinnati, OH 48268, March 1979. Duplicate or triplicate (when
sufficient sample volume is available) aliquots of each sample are analyzed to
determine the relative magnitude of variability that can be attributed to the
laboratory analytical procedures.
Extraction Procedure (EP) Evaluation
Rationale
The proposed extraction procedure (EP) is a key step in the screening
mechanism designed to identify those wastes that are hazardous and require
special management because of their toxic characteristics. It should be noted
that the EP is not intended to identify the total concentration of the toxic
contaminant in the waste, but rather that Teachable concentration 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, i-s
explicit enough for use by non-experienced personnel to obtain valid
data.
3. determine the equivalency of the various extractors that could be used
with the EP.
11
-------
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 is being used in accordance with the proposed
regulations, with minor clarifications of instructions necessary to facilitate
sample extraction (see Appendix 2). A flow chart of the sample treatment
required for the EP is shown in Figure 1. As shown, triplicate aliquots
(minimum of 100 gm each) are obtained front 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 and placed in a suitable
extraction apparatus along with a volume of deionized water equal to 16 times
the weight of the solid phase. For solutions of pH > 5, the pH of the
solution 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
extraction procedure is 4 ml per gram of solid phase, even if the pH of the
solution does not reach 5.0 ± 0.2. Agitation is continued for 24 hours. The
solution is then filtered and any solid material is discarded. The liquid
extract is then 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. This solution is then added to the original liquid phase to
produce the extraction procedure extract. The EP extract is sp5;it into two
samples; one is acidified to preserve it for elemental analysis and the other
is stored under refrigeration for organic analysis.
The EP is being evaluated with as many of the wastes collected (Table 1)
as possible. Each sample is being extracted at least once for screening
analysis by atomic emission spectroscopy to identify the major extractable
toxic components that might be used for evaluation of the EP. Based on the
screening data, the OSW and the EMSL-LV are determining the priority of the
samples to be used for evaluation of the EP.
Triplicate aliquots of each sample are being treated as described in
Figure 1. The extracts are analyzed by standard atomic absorption
spectroscopy methods cited in Section 250.13(d) of the proposed regulations.
The results are then averaged to determine a mean and the standard deviation
for the triplicate analyses. The standard deviation identifies the
reproducibility of the procedure.
If time permits, additional experiments will be performed to compare
various extractors that might be used for performing the EP. Each extractor
will be used to simultaneously extract a minimum of three aliqucts from the
same waste sample (i.e. comparison of two extractors requires a minimum of six
aliquots, three per extractor). The waste sample for this comparison will be
selected from those that have high concentrations of toxic elements. Each
12
-------
WASTE SAMPLE (stirred)
I
TRIPLICATE ALIQUOTS
I
FILTRATION/CENTRIFUSATION-
SOLID PHASE (weighed) -
/Add deionized H20\
^16 x Solid weight/
EXTRACTOR
AGITATION
I,
/Adjust and maintain pK at 5.0 ± 0.2
( w/ 0.5 N Acetic Acid
Max. acid = 4- ml/g solid
V
FILTRATION
SOLID PHASE-
DISPOSAL
-LIQUID PHASE
DILUTION
LIQUID PHASE
STORE AT 1-5°C
1 Distilled H20 to
total vol. = 20 x orig.
sample weight
EXTRACTDN PROCEDURE EXTRACT
SPLIT
ACIDIFICAlON-^-J—"STORAGE (at T-5°C)
(PH< 2
ELEMENTAL fflALYSIS ORGANIC ANALYSIS
Figure 1. Flow Chart" of the Extraction Proca&ire for Identification
of a Hazardous Waste.
13
-------
extract will be analyzed by the procedures cited in the proposed regulations.
The means and standard deviations of the analytical results obtained by each
extractor will be compared. An acceptable extractor should yield a mean that
agrees with the mean obtained with the extractor in the proposed regulations,
and should have an equivalent or lower standard deviation (i.e. equivalent or
better reproducibility).
Additional experiments may be performed to examine other extraction
procedures that have been suggested or used for the identification of
hazardous waste. One waste sample has been extracted with deionized water in
accordance with a procedure suggested by the ASTM. This procedure, while
similar.to the EP, uses distilled or deionized water instead of acetic acid
buffer, does not. subdivide solid samples into smaller units, has a higher
solid to liquid ratio, and uses a less aggressive means of agitation. The OSW
and the EMSL-LV will determine if additional experiments are to be performed
to compare alternative identification procedures. The comparisons will use
the EP and alternate identification procedures simultaneously to characterize
the same waste sample(s). A minimum of three aliquots of the same waste
sample(s) will be extracted in accordance with the instructions for each
procedure (i.e. a minimum of three extractions per procedure). The extracts.
will be analyzed by the analytical methods cited in the proposed regulations.
The means and standard deviations of the results obtained with the alternative
procedures will be compared to that obtained with the EP.
Additional experiments are also being performed to identify any background
interferences that may result from the equipment used for the EP. The
filtration equipment and the extraction apparatus used in this study are made
of stainless steel. Since the samples come in contact with the stainless
steel surfaces, there is some concern that this will cause high background
concentrations of metals, especially chromium in the EP extract- Two groups
of blank samples, one consisting of deionized water and the other Q.13 N
acetic acid (400 ml of 0.5 N acetic acid to 1600 ml of deionized water), are
being tested (in accordance with the EP) to determine what background
concentrations of metals will result from the equipment used to obtain the EP
extract.
Analytical Procedures Evaluation
Rationale
The analytical procedures proposed for analysis of the EP extracts are
standard rrathods for analysis of water, wastewater, or industrial effluents.
They should be acceptable for analysis of liquid wastes and the EP extracts;
however, they have not been extensively tested with samples of industrial
"solid waste" or EP extracts of such waste. The approach taken in this phase
of the study is therefore designed to obtain additional information on the
accuracy and reproducibility of these analytical methods when they are applied
to liquid wastes and the EP extract in accordance with the proposed
regulations. Alternative methods that offer advantages in cost and accuracy
may be identified and evaluated. If necessary, new methods will be developed
and evaluated.
14
-------
Experinental Design
Use proposed regulations require the use of atomic absorption (AA) methods
("Mettods for Chemical Analysis of Water and Wastewater." Environmental
Protection Agency, Office of Technology Transfer, Washington, D.C. 20460,
1974) for the analysis of liquid wastes and the EP extract for arsenic,
barium, cadmium, chromium, lead, mercury, selenium and silver. The AA methods
are being evaluated with EP extracts from various wastes known to contain one
or more of the elements of interest.
Triplicate aliquots of the EP extracts are first analyzed in accordance
with tte proposed regulations. Aliquots are then spiked with known
concentrations of the elements of interest and re-analyzed in accordance with
the prsposed regulations. The mean and standard deviation of the triplicate
results obtained for each spiked sample matrix are used to evaluate the
accuracy of the analytical method. Spike recovery is calculated by dividing
the mean analytical result (less the concentration of that element in the
unspiksd 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 and multiplying by 100. The analytical
methods' should accurately and reproducibly indicate the increase in
concentrations of the elements in the spiked samples. Spike recoveries will
be determined for each analytical method in as many sample matrices as
poss.fjsfe.
Ai£itional experiments are being performed to determine the effect, if
any, uf the acetate buffer matrix of the EP extract on the analytical
proceixires. Solutions of known concentration are prepared in an acetate
buffer matrix and in the standard matrix identified for each element in the
"Metfaads for Chemical Analysis of Water and Wastewater." The standard curves
(constitration vs. absorbance) for each element should be identical if no
matrix problems exist.
A similar approach will be used to evaluate the analytical methods for
endriB, lindane, methoxychlor, toxaphine, 2, 4-d and 2, 4, 5-TP Si 1 vex.
However, these studies will also determine if alternate improved methods are
aval 1*16 for these organic compounds. Because of the low solubilities of
these compounds in water, it is anticipated that these evaluations will be
more difficult and time consuming.
Qua!ity Assurance Program for Hazardous Waste Monitoring
Rationale
It is imperative that the data collected for management of hazardous waste
be of inown quality to ensure that both human health and the environment are
adequately protected and that the regulations are enforceable. A quality
assurance program is being developed to provide a standardized approach to
monitoring hazardous waste operations and to allow continued evaluation of
EPA, state and other laboratories who are responsible for conducting the
measurement and monitoring that is required under the RCRA waste management
systesu. This program is being coordinated by the EMSL-LV and will use quality
assuraace efforts already in place for the air and water programs.
15
-------
Program
A repository for standard reference organic compounds is being developed
jointly by the EMSL-LV and the Environmental Monitoring and Support
Laboratory-Cincinnati (EMSL-CIN). Additional reference materials will be
developed for this program under an interagency agreement with the National
Bureau of Standards (NBS). These reference materials will include soils,
sludges, sediments and "waste-type" samples of known composition that can be
used for methods evaluation, instrument calibration, and laboratory
intercomparison studies.
Laboratory intercomparison studies will be initiated in FY80. These
studies are designed to assure that the participating laboratories can obtain
accurate data with the specific procedures and samples be-ing tested. Samples
of known composition will be distributed as "unknowns" to the participating
laboratories for analysis by the appropriate procedures. These samples may
require application of the EP or may simply require analysis for one or more
components. Each participating laboratory will report its results to the
EMSL-LV. All laboratory results will be combined into a single report (with
laboratories identified by code numbers for confidentiality) which will be
distributed to each laboratory. If an individual laboratory's results differ
significantly from the true value or grand average of the reported results,
that laboratory must identify the source of the discrepancy and take
appropriate action to correct its analytical operations.
The quality assurance program will also include on-site laboratory
evaluations. It is anticipated that on-site evaluations of EPA laboratories
will be coordinated and performed by the EMSL-LV, whereas state and other
laboratories will be evaluated by the appropriate EPA Regional office. These
on-site evaluations are designed to determine if the laboratories that provide
data for the hazardous waste programs have adequate facilities, equipment, and
personnel and are using proper procedures for obtaining the data reported.
Quality assurance guidance will also be provided in the form of minimum
laboratory standards and practices and other guideline documents and manuals.
The EMSL-LV will develop protocols for evaluation of equivalent methods and
will establish an equivalency program to evaluate new methods as they become
available.
RESULTS AND DISCUSSION
Sampling Procedures Evaluation
Pond Sampler
Four ponds were sampled at site A. Two samples were taken at each of five
locations on each pond. Samples from two of the ponds, Pond 0 and Pond 13,
have been analyzed for pH and percent solids. Some difficulty has been
encountered with filtration of the samples from the other two ponds; however,
it is anticipated that the data for those samples will be provided in the
final report. If sufficient sample material is available, the samples from
16
-------
Site A will also be analyzed by the EP to determine the overall
reproducibility of the procedure for identification of a hazardous waste when
applied to waste in ponds.
Four factors contribute to the overall reproducibility observed in the
analytical data obtained with the pond samples. These are:
1. the precision of the analytical procedure used to
determine the parameter of interest in each sample.
2. the effect of the collection of the first sample on
the second sample, when two samples are collected at
the same location
3. the homogeneity of the waste in the pond being sampled.
4. the reproducibility of the sampling procedure.
The data obtained with the pond samples (Tables 3 and 6) can be analyzed to
estimate the contribution of these four factors to the overall reproducibility
of the pond sampler. The precision of the analytical procedure, Factor 1, can
be estimated by comparing the results from replicate analysis of the same
sample (Tables 4 and 7). The differences between two samples taken at the
same location (Tables 5 and 8) reflect the reproducibility of the sampling
procedure, Factor 4, and the effect (if any) of the collection of the first
sample on the second sample, Factor 2. 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 4 and 7), Factor 3.
The data for the titanium dioxide process waste from Pond 0 are shown in
Table 3. Because of the limited sample volumes, only duplicate aliquots were
analyzed for each sample. The pH of the Pond Q samples was very low (pH <1),
thus the measurements with the pH meter were not 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 3) reflect the reproducibility of the laboratory analytical
procedure for determining pH or percent solids. 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. The standard deviations reported in Table 4 provide a
numerical indication of this reproducibility. If the standard deviations of
the mean for each location (Table 5) are compared to the standard deviations
for the analyses (Table 4), 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
(^10,5s 6.89). However, the analysis of variance performed on the percent
solids data for Pond 0 indicates no significant differences between locations
tf = 0.95). If the mean value for each field sample (Table 4) is
17
-------
TABLE 3. 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
18
2A
23
3A
JL. .1 1 -1
PH
0.23
0.23
0.34
0.31
-...**
__**
0.22
0.17
0.30
-Q.19
,
Percent
Solids
1.24
1.26
1.09
1.28
0.82
1.28
2.09
1.28
**
**
. M*J - ft 9 D 1 »
-------
• TABLE 5. 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
pH Percent Solids
Location x s(n = 2) x s(n = 2)
1
2
3
4
5
0.28
—
0.34
0.43
— .
0.07
__*
0.13
0.07
*
1.22
1.37
—
—
1.74
0.04
0.45
— *
— *
0.08
Average 0.35 -- 1.44
* Data not reported - known discrepancies in analytical procedure.
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.
The data for the alkaline waste, Pond 13, are shown in Table 6. In this
case, pH and percent solids measurements could be used to evaluate the
reproducibility of the sampling procedure. Triplicate aTiquots of each sample
(two samples from each of five locations on the pond) were analyzed. The pH
analyses were very reproducible for aliquots from the same sample (Table 7)
with a relative standard deviation of less than ±4%. The standard deviation
for percent solids analyses (Table 7) 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%). If the standard
deviations of the mean for each location (Table 8) are compared to the
standard deviations for the analyses (Table 7), 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. If the mean value
for each field sample (Table 7) 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
5.41 ± 1.85 (n = 10). The data in Table 8 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
19
-------
T18LE 6. 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
Solids Solids
Sample* pH (%) Sample 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
48 6.75
6.70
5.33
5A 5.03
5.40
4.70
58 . 4.74
4.67
5.07
4.82
5.03
3.50
3.67
3.2Q
3.48
3.99
3.79
6.13
5.60
5.76
1.93
2.08
1.75
* Niaber - location on pond; A & B = 1st and 2nd sample.at that location.
significant 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
percsrt solids (Table 8) also indicate that for this measurement there were
significant differences between samples taken at the same location. This
mi gift be expected with a liquid sample that has a significant solids
concsntration 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. Analysis of variance is being
performed on this data and will be presented in the final report.
"^e results for the samples collected at site A indicate that even with a
very heterogeneous waste the pond sampler has a rsproducibility of better than
20
-------
TABLE 7. 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
Percent Solids
Location
First sample
x s(n = 3)
Second sample
x s(n = 3)
First Sample
x s(n = 3}
Second Sample
x s(n = 3)
1
2
3
4
5
6.99
2.43
4.90
7.01
5.25
0.04
0.00
0.01
0.06
0.20
7.59
2.33
5.23
6.72
4.70
0.06
0.01
0.03
0.03
•0.04.
4.13
3.25
5.99
3.46
5.83
0.16
0.11
0.07
0.24
0.27
5.65
5.75
4.97
3.75
1.92
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
TASLE 8. 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
21
-------
±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. < 1%) the
reproducibility of the laboratory procedure was not as good. The standard
deviation does not change; however, the relative standard deviation (standard
deviation -r sample mean x 100%) changes considerably because of the lower
value for percent solids. As expected, the pH was very reproducible (better
than ±2%) for the laboratory analytical procedure. 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 several samples from different locations on the
pond are required for proper identification of hazardous waste.
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 9) 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
different locations on a pond should provide a more reproducible indication of
the pond's composition. Additional data obtained with composite samples is
required to confirm this observation.
Future Efforts. Forty samples (20 samples/pit) have been collected with
the pond sampler from two waste pits at site F. Five wastes (Table 2) were
.sampled from.drums with the COLIWASA (total of 84 samples). These samples are
being analyzed for pH, percent solids, total solids, and other chemical
parameters (as time and sample volume permit) and the results will be included
in the final report. A more detailed statistical analysis and discussion of
the results of the sampler evaluations will be presented in the final report.
Extraction Procedure Evaluation
Progress
The extraction procedure (EP) is being evaluated to determine its
reproducibility when used for the identification of hazardous waste. The EP
extracts are first screened for arsenic, lead, cadmium, barium and chromium by
inductively coupled plasma emission spectroscopy (ICP). The screening
analyses are for qualitative identification and are restricted to those toxic
elements (listed in the regulation) that can be analyzed by the ICP system at
the EMSL-LV. The extracts are then analyzed for each element of interest by
atomic absorption (AA) spectroscopy in accordance with the proposed
regulations.
. 22
-------
TABLE 9. REPROCUCIBILITY OF POND SAMPLER: MATHEMATICAL COMPOSITE
OF FIRST AND SECOND SAMPLES FROM EACH LOCATION
Data
Pond 1
Sample 1 Sample 2
Composite Composite
Pond 2
Sample 1 Sample 2
Composite Composite
Mean pH
of mean
RSD of mean
Mean * solids
of mean
RSD of mean
0.35
1.42
Average RSD =1.7
0.007
2.02
0.04
2.76
0.34
1.48
5.32
4.53
0.007
0.13
0.08
1.90
5.31
4.41
Table 10 gives the ICP data collected in the initial phase of this study.
These data -indicate very high concentrations of the toxic elements in several
of the extracts from samples from 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 extracts from samples from nond 10>, Site
A; the pesticide waste from Site C; and the filter cake frc ' £!';•? G had low or
insignificant concentrations of the metals that could be ict . -,'ried by the
screening analysis.
Table 11 gives the 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 by the EP for two reasons; (1) to
obtain better 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 I*'"/; - 11 reflect
the reproducibility of the analytical procedures performed o-. -.tes from
Ponds 0 and P.
23
-------
TABLE 10. ICP SCREENING ANALYSIS OF EP EXTRACTS: APPROXIMATE ELEMENTAL
COMPOSITION OF EXTRACTS FROM SE13TTED WASTE SAMPLES
Sample (No. of Extracts Analyzed)
Site A, Pond 13, Location 1 (1)
Site A, Pond 0, Location 2 (15)
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 0, 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 (!)
Apprmimate Concentration*
As
1.3
168
0-6
<0.4
0.8
-------
Data presented in Table 12 indicate significant differences in the
elemental composition of the two wastes. However, since the variation of
elemental composition within each pond is not known, additional EP data from
the other samples from Ponds 0 and P are required to confirm this conclusion.
The data fn Tables 12 through 14 also give an indication of the reproduc-
ibility of the EP. For the worst case, barium analysis, the EP yielded a
relative standard deviation of less than ±17%. The initial evaluation
indicates that most of this variability is due to the analytical method
(discussed in the section on analytical methods). This appears to be the
case, since the results for lead and chromium yielded a relative standard
deviation of less than ±5%.
TABLE 11. EVALUATION OF EXTRACTION PROCEDURE (EP): MEANS AND STANDARD
DEVIATIONS FOR AA ANALYSES* OF EP EXTRACTS FOR BARIUM, CHROMIUM AND LEAD
B..arium (mg/1)
Chromium (mg/1) Lead (mg/1)
Waste Extracted
x
.Sulfonation Tars
(Site A, Pond
10)
<0.9
<0.9
<0.02
<0.02
0,3
0.3
0.1
0.1
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
9.8
5.15
21.1
0.9
<0.9
<0.9
<0.9
<0.9
<0.9
1.5
0.32
2.7
0.1
4.1
1.02
1.90
<0.02
9.1
1.2
1.0
7.9
1.4
0.1
0.13
0.17
0.1
0
0
0.2
0.1
0.1
0.08
<0.08
<0.08
0.1
0.1
0.1
<0.8
<0.8
0.1
0.1
0.1
0.1
0.2
* Flame Atomic Absorption .analyses performed in triplicate.
(continued)
25
-------
TABLE 11. fContinued)
Barium
Waste Extracted
Electric Furnace Baghouse
Dust, Site E
Blast Furnace Scrubber
Filter Cake, Site E
Mill Scale, Water Treatment
Plant, Site E
.
(1)
(1)
(1)
Lime Sludge, Ammonia
Still, Site E (2)
Filter Cake, Chlorine/Hg
Process Stream, Site G
(2)
Chlorine Process Sludge,
Site I
(2)
X
0.8
1.04
0.85
1.06
0.64
0.90
G.20
0.19
0.52
0.28
0.24
0.22
2.7
0.25
0.14
0.15
0.9
0.48
0.40
0.98
3.5
(mg/U
s
OJ5
0.2
0.3
G.d
o..»
OJGE
QM
QM
0.01
OJE
0.®
0.&
1.5
0.®
Q.m
O.J3T
—
o,n
o.«
Q.m
i.a
Chromium (mg/1 ) Lead
X
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.02
<0.32
<0.32
<0.32
<0.02
<0.32
<0.32
<0.32
0.1
s x
— . 0.13
0.13
0.13
15.3
14.4
11.6
<0.08
<0.08
<0.08
<0.08
<0.08
<0.08
0.3
0.10
<0.08
<0.08.
<0.08
0.45
0.45
0.48
0 0.4
(mg/1 )
s
0.04
0.03
0.04
0.7
0.7
0.7
— ^
—
—
._
—
—
0.1
0.06
—
—
—
0.04
0.04
0.04
0.1
(1) Results obtained with wrist-arm sh&er.
(2) Preliminary and test data.
-------
TABLE 12. EVALUATION OF EXTRACTION PROCEDURE (EP): MEANS
AND STANDARD DEVIATIONS FOR AA ANALYSES* OF EP EXTRACTS
FROM WASTE IN PONDS 0 AND P, SITE A
Sample Extracted
Pond 0, 2A
Pond 0, 2B
•Pond P, 2A
Pond P, 2B
Barium
X
1.85
1.55
1.57
1.30
1.33
1.39
31.0
24.5
34.2
23.7
28.5
31.1
(mg/1)
s
0.02
0.14
0.34
0.28
0.29
Q. 09
1.4
1.7
10.6
3.8
7.6
3.9
Chromium (mg/1 )
X
1050
1050
1020
950
960
920
78.3
74.9
79.5
84.6
82.5
80.5
s
12
6
12
12
15
25
0.5
0.6
1.0
1.2
1.5
1.3
Lead (mg/1)
X
45.1
46.1
45.9
43.0
45.7
41.8
<0.8
<0.8
<0.8
<0.8
<0.8
<0.8
s
0.6
0.7
1.2
0.8
1.2
0.9
—
—
—
—
—
—
* Flame Atomic Absorption analyses performed in triplicate on each of three
aliquots of sample extracts. For s, n = 3.
27
-------
TABLE 13. AVERAGE RELATIVE STANDARD DEVIATIONS (RSD'S) FOR VARIOUS
LEVELS OF SAMPLING AND ANALYSIS
Sampl i ng
RSD (%)
pH
Percent Solids
Pond 0 Pond 13 Avg. Pond 0 Pond 13 Avg.
Differences between aliquots
of the same sample 10.1
Differences between duplicate
samples taken at each location 26.5
Differences between locations
on pond 21.5
0.9 5.5
4.8 15.4
36.5 29.0.
17.6 4.7 11.2
13.6 30.3 22.0
19.9 16.9 18.4'
Analysis
(Sample source: Ponds 0 and P, Site A) Barium
RSD (%)
Chromi urn
Lead
Differences between replicate
determinations on a given
E? extract
Differences between replicate
extractions on a given sample
of waste
14.9
11.0
1.3
1.8
2.0
3.0
Stability of the Extract
A problem has been experienced with the stability of the EP extract. Sams
of the extracts, expecially those with high concentrations of other materials
(such as inorganic salts and organics as observed by color and density of the
extract), formed precipitates over a period of several days^, even though they
were preserved with acid (pH <2). This problem was observed early in the
study and was addressed by adding a step to the extraction 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. Even with this step, some of the more concentrated
samples produced a precipitate with a few days after preservation. This
problem will be investigated in more detail in future studies.
28
-------
Extractors
Data is also being collected to compare a wrist-arm type shaker to the
extractor described in the proposed regulation. Preliminary data for barium
in the mill-scale sample from Site E (Table 11) indicate good agreement
between the two extraction devices, although the results obtained with the
wrist-arm shaker appear to be less variable than those obtained with the
proposed extractor. More data is required for a definitive statistical
comparison of the two extractors.
Background Interferences
Blank samples (distilled water) identified in Table 15 were used during
routine analyses as controls to check for contamination resulting from
previous samples or from metallic components in the extraction or filtration
systems. Seven extractors (coded 1-7) and three filter systems (coded 1-3)
were used in the study; the results are shown in Table 15. In general, the
data indicate that barium, chromium, and lead are not leached from the
stainless steel components by distilled water. Blanks for extractors 1 and 5
indicated higher than anticipated contamination with Cr and/or Pb. While
stafnTess steel components could contribute to Cr concentrations, it is
unlikely that they would contribute Pb. Therefore, it is highly likely that
the levels of Cr and Pb detected represent contamination from previous use of
the extractors involved.
TABLE H. 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
Barium (mg/1) Chromium (mg/1) Lead (tng/1)
Samp! e
Pond 0
Pond P
Extracted
2A
2B
2A
2B
X
1.65
1.34
29.9
27.8
s -
0.17
0.05
4.9
3.7
RSD
(%)
10.3
3.7
16.4
13.3
X
1040
943
77
82
s
17
21
.6 2.4
.5 2.0
RSD
(*)
1.6
2.2
3.1
2.4
X
45.7
43.5
—
— •
RSD
S (%)
0.5 1.1
2.0 4.6
—
-- —
* Flame Atomic Absorption analyses performed in triplicate on each of three
aliquots of sample extracts.
(1) n = 3
(2) RSD = Relative Standard Deviation
29
-------
TABLE 15. EVALUATION OF BACKGROUND CONCENTRATIONS OF ELEMENTS FROM
EP EQUIPMENT: MEANS AND STANDARD DEVIATIONS OF TRIPLICATE AA
ANALYSES OF OEIONIZED WATER BLANK SAMPLES FROM EXTRACTION
AND FILTRATION APPARATUS
Ba (mg/1) Cr (mg/1) Pb (mg/1)
EP
Equipment Tested
Extractor 1*
Fil
2
3
4
5
' 6
7
tration apparatus
" "
II U
" "
Tl II
II U
II II
II II
II II
II II
II II
II II
1
2
2
2
2
2
2
2
3
3
3
3
Deionized H20**
X
<0.9
<0.9
<0.9
<0.9
<0.9
0.12
0.10
<0.9
^<0.9
<0.06
<0.06
0.11
0.09
<0.06
<0.06
0.11
0.16
0.10
0.12
<0.09
s
„_
—
—
—
—
0.02
0.03
—
—
—
—
0.03
0.01
>_
—
0.02
0.03
0.02
0.03
—
X
0.1
<0.02
<0.02
<0.02
0.1
<0.32
<0.32
<0.02
<0.02
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.32
<0.02
S X S
0 <0.08
-------
The regular analysis of blank samples is recommended to provide quality
co-ntrol and to identify problems with equipment contamination. Additional
data with acidic blank samples (0.13 N acetic acid) that simulate the EP
extract areaeing collected; the results will be presented in the final
report.
Future Efforts
The extraction procedure will be evaluated with as many of the waste
samples collected as possible. The EP will be used to characterize the
samples from Site A to obtain additional information on the feproducibility of
the total pracedure for identification of hazardous waste (sampling through
analysis).
Additional data will be collected to compare the different extractors that
may be used for the extraction procedure. These include the extractor
described in the proposed regulations, a wrist arm shaker available in our
laboratory aad a tumbling-action extractor developed especially for use with
the EP by the National Bureau of Standards.
Analytical Procedures Evaluation
Progress
The prajased regulations require that atomic absorption methods be used
for analysts of the EP extract for arsenic, barium, cadmium, chromium, lead,
mercury, selenium and silver. As described in the approach, these methods are
being evaluated to determine their accuracy and reproducibility for analysis
of the EP esSract. Only limited data has been collected, thus far.
The results from the analysis of the EP extracts for barium, chromium,
lead and mercury are shown in Tables 11, 12, and 16 through 19. Problems were
encountered fn the analyses for barium. The data in Tables 11 and 12 show
relatively high standard deviations for the analytical results for barium.
This was profcably due to a fluctuation in the AA detector response that was
observed during the analyses. The cause of the fluctuation in detector
response vas not positively identified; however, several sources could account
for this and the high standard deviations for the barium analyses:
1. The barium analysis requires a high temperature nitrous oxide flame.
Wher analyzed, the extracts caused beads to form on the burner
heai. These beads cause fluctuation of the flame and can result in
detactor signal fluctuations. This problem was not observed with
the standard solutions and appeared to be due to the extract matrix.
2, The barium lamp used as the source for these analyses may not have
bea functioning properly. However, it was a new lamp and provided
staHe. signals with the barium standards.
3. Low spike recoveries and excessive variability were observed for the
anaiyses (Table 16). When the method of standard additions was used
31
-------
TABLE 16. QUMTY CONTROL DATA: COMPARISON OF BARIUM SPIKE
REOXRY FROM SELECTED SAMPLES (MATRICES)
Sampl e
Site A, Pond P
Site B, Paint Sludge
Site D, Chromate
Oxi.dation Paste
Site D, API Oil
Separator
Site G, Filter Cake
Site I, Chlorine
Process Sludge
Blank, Filtration
Apparatus
Simple
Cone.
ing/l )
3.10
1.96
3.44
0.33
0.15
3.39
e.io
Spike
(mg/1)
2.00
2.00
2.00
2.00
2.00
2.00
2.00
Spiked
Cone.
(mg/1 )
4.96
3.98
1.94
2.11
1.84
2.28
2.10
Spike
Recovery
(%}
93
101
76
89
84
94
100
RSD*
(Analysis)
11
33
18
33
7
23
1
* Relative Standard Cessation
TABLE 17,
Sample
Site A, Pond 0
Site A, Pond 0
Site 8, Paint
Sludge
Site D, Chromate
Oxidation Paste
Site-D, API Oils
Separator
SRLITY
HSDVERY
3ample
Cone.
&g/D
7.83
10.5
1.02
0.80
9.12
CONTROL DATA:
FROM SELECTED
Spike
(mg/1)
2.00
2.20
2.00
2.00
2.00
COMPARISON OF CHROMIUM SPIKE
SAMPLES
Sp-iked
Cone.
(mg/1)
9.64
12.6
2.15
2.91
11.39
(MATRICES)
Spike
Recovery
(X)
90
95
108
106
113
RSD*
(Analysis)
10
8
6
3
9
Site E, Blast Furnace
Filter Cake
Site E, Mill Scale
0
0
2.00
2.00
1.85
2.03
93
102
8
8
* Relative Standard Deviation
32
-------
TABLE 18. QUALITY CONTROL DATA: COMPARISON OF LEAD SPIKE RECOVERY
FROM SELECTED SAMPLES (MATRICES)
Sample
Sample Spiked Spike
Cone. Spike Cone. Recovery RSD*
(mg/1) (mg/1) (mg/1) (%) (Analysis)
Site A, Pond 0
Site A, Pond P
Site B, Paint
Sludge
Site B, Paint
SIudge
Site Q, Chromate
Oxidation Paste
Site D, API Oils
4.5
0.83
0.12
0
0
2.00
2.00
1.00
2.00
1.00
6.69
2.94
1.10
2.33
1.14
1UQ
106
98
117
114
6
3
2
3
2
Separator
Site E, Mill Scale
0.14
0
1.00
2.00
1.28
2.04
114
102
5
3
* Relative Standard Deviation
for barium analysis, higher results were obtained for those samples
that had low spike recoveries. This indicates that the sample
matrix has an interference that suppresses the barium signal, i.e.
causes a low result. The method of standard additions corrects this
problem; however, the precision of the data is reduced, i.e., the
standard deviation for the analyses increases.
No problems were encountered in the analyses of EP extracts for chromium
and lead (Tables 11, 12, 17 and 18). Although limited data is available, the
standard deviations and percent recoveries are those that would be expected
for analyses of water and wastewater. The high relative standard deviations
noted for lead analyses that yield values approaching the lower detection
limit is not unexpected.
Only two extracts have been analyzed for mercury (Table 19). Although
these samples were expected to contain high mercury concentrations, the EP
extracts had less than 4 yg/1. Aliquots of the waste samples were digested
(aqua regia) and the digested samples were analyzed for mercury. The results
(Table 19) show that the waste samples contained high concentrations of
33
-------
TABLE 19. EVAUlflON OF EXTRACTION PROCEDURE (EP): COMPARISON- OF MERCURY
CDNCENTR/5IONS (COLD VAPOR AA) IN WASTE SAMPLE DIGESTS WITH
I8BE ESTIMATED FROM EP EXTRACT CONCENTRATIONS
EP Reconstructed* Digested**
Extract (ug/1) Sample Ug/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 frorasxtract concentration and dry weight of solids in the
waste sample exttacted.
mercury {approximaMy 1 mg Hg/gm waste); however, the proposed procedure
identified only a wry small fraction of the mercury present. The mercury may
be in the form of .«gano-mercury compounds or may be irreversibly bound to the
solids in the matra. If so, the results simply suggest that the mercury
would not leach out from the waste under acidic (pH = 5.0) conditions.
However, if the saa?le does contain high concentrations of elemental or ionic
mercury, the resulismay indicate that the extraction procedure or analytical
method is inadequafefor identifying mercury hazards. More work is required
to identify the few.of mercury present in these samples and to clarify the
questions concemiii. the use of the extraction procedure for identification of
Teachable mercury itwaste samples.
CONCLUSIONS AND RECOMMENDATIONS
It must be empissized that the discussion, conclusions, and
recommendations preBnted in this report are based on only the first four
months of study, "He additional data being collected should provide confident
conclusions concerning the sampling procedures (Pond Sampler and the
COLIWASA), the EP, ard the analytical methods. However, several tentative
conclusions can be »de from these limited data.
(1) Future studies should provide better coordination between the sampling
team and the indusSsy being sampled. If a contractor is collecting the
samples, many of tfeindustrial operators require a confidentiality agreement
from the contractor. Corporate approval of this agreement can take up to four
weeks to complete. Future efforts should also attempt to collect more
information about tte process stream(s) producing the waste. This will
facilitate characterization of the waste; will allow more specific conclusions
about the hazard of the waste stream itself; and may aid in identification of
modifications to the industrial process to reduce or eliminate the hazard.
34
-------
(2} The procedures for identificatiar of hazardous waste should require
composite as well as individual sampfes from the waste source. The number of
samples in the composite would be pr^ortional to the size of the source
and/or a priori knowledge of the wasfe homogeneity. For disposal pits or
ponds, the samples should be taken fran evenly spaced locations (if
accessible) around the pit or pond, further studies are recommended.
(3) The limited data collected for 8e pond sampler were obtained from a
heterogeneous waste and indicate that even under such "worst-case" conditions
the sampler yielded a reproducibilitjrof ±4% or better. Indeed, analysis of
"mathematically composited" data sugpsts that use of composite samples can
yield a reproducibility as good as ^percent. Data from liquid/solid
composition of consecutive samples fran the same location suggest that much of
the variability between these samples can be avoided by strict adherence to
the protocols for use of the pond saapler, i.e. consecutive samples be
sufficiently separated by time and/orspace that removal of the first sample
.does not influence the composition of the second.
(4) The extraction procedure in theproposed regulations yielded data with a
relative standard deviation of lessen ±15% for a heterogeneous waste sample
(i.e., under worst-case conditions).
(5) There is a problem with the stability of the EP extract even with
acidification to pH0.3 mg/1) were
found to be highly reproducible- {RSD- 1.3% and 2.0%, respectively).
(8) The high relative standard devstions for barium analyses (RSD = 14.9%)
suggest problems with the analytical method for barium. This potential
problem should be investigated in more-detail.
(9) The extraction procedure and/or the analytical method for mercury may not
identify the presence of mercury in «ste. Additional work is required to
determine the cause of the low mercusy results found with the two waste
sanoles studied.
35
-------
APPENDIX 1.
USE OF THE POND SAMPLER AND COLIWASA
POND SAMPLER
The pond sampler is simply a 1000 ml glass beaker affixed with a clamp to
the end of an 8-15 ft adjustable aluminum handle (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 personnel must be wearing all the proper personal
safety equipment. Samples can be taken at or below the surface. The stepwise
procedure for use of the pond sampler is presented below.
• The operator must make sure the sampler is clean and put together
properly.
• Sample at the desired depth and distance from the edge. To collect
a sample, the beaker is lowered into the pond in an inverted
position. At the required depth, turn the handle to upright the
beaker. Withdraw the sampler.
0 Pour the sample into the sample container slowly.
• Clean the sampler after each sample. When taking multiple
samples from the same pond, take care to move far enough from
the previous grab sample location to get an undisturbed sample.
• Close the container, 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 and pack away.
36
-------
7arigrip- clarap
• bolt hole
beaker, 1000ml,
glass
telescoping aluminum pole, heavy duty,
2.5 to 4.5 a (8 to 15 ft.)
Figure A-l. Pond Sampler
Modified from draft report by de Vera et al, California Department of Health
Services, for the Municipal Environmental Research Laboratory,
Cincinnati, Ohio. (Grant No. R804692Q10)
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 and is functioning
properly so the stopper fits tightly. Adjust the stopper rod length if
necessary.
* Two persons are required for sampling; both personnel must be wearing
all the proper 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.
» Lower the sampler carefully into the waste storage vessel, maintaining
the sampler in a vertical attitude and making sure visually that the
waste inside the tube is even with the waste outside the tube (this is
not possible with an opaque PVC or a stainless steel COLIWASA}. This
is to maintain a representative sample.
37
-------
o
O «
<*t
U -V
o £
c. u
a
a-*
CO
in
i
S O
-H
Aa ^^
E
- o
Cl
- u cu
c.
O U
4) O
c
•
- CO
b •
O CO
c.
c.:
o ce
•A ^»
to «n
oo
C_J
UJ
_>
0.
o
to
2
en
q
— SO
•3 — O
J O
i _:
H
o
o.
«»•
"«.
T
0
in
2r=
z
O
en
O
c.
_J
Oi
CO
CM
CD
!_
C3
CT>
ul
38
-------
• When the sampler hits the bottom of the waste storage vessel, or until
only 6 inches of the sampler is not immersed, pull upward on the handle
until the handle can be turned so that one end rests firmly on the
locking block.
t 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 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.
39
-------
APPENDIX 2
PROTOCOL: EVALUATION OF THE EXTRACTION PROCEDURE FOR IDENTIFICATION
OF HAZARDOUS WASTE
This is a summary of the protocol followed-tn the evaTuation 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.
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 log book. An annual calibration with
standard weights traceable to the National Bureau of Standards is performed
40
-------
>wfcen 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 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 PQ407-or~ equivalent) and~a~~weTghed 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 fractinrua^-L-^G^e^^=tts^-fn=-the-extr2FCttorr
procedure.
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 and filters) by the extraction procedure.
6. If the sample does not filter, use the centrifugation method to
separate the solid and liquid phases.
Cantrifugatlon 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 - 40°C).
2. Measure the size of the liquid and solid layers to the nearest mm
(0.40 inch) and calculate the liquid to solid ratio.
3. Repeat steps 1 and 2 until the liquid to solid ratios for two
consecutive 30 minute centrifugations agree within 3%.
41
-------
f-
' - 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. It is anticipated that the particle
size of the samples obtained for this study will meet this requirement. 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,
No V 24T,~ Dec. 1B7T97S).
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 not only present stratification of the extraction solution
but will also ensure that all sample surfaces are continuously brought
into contact with well mixed extract!on solution. With the exception of
special studies, the extraclar referenced in the proposed regulation will
be used for this program,
2. Add to the extractor a wight of deionized water equal to 16 times the
weight of solid material added to the extractor.
3. Agitate the sample at 4-0 rprr, and adjust the pH of the solution to
5.0L ±J1^ wdtiv-4^W-^et^-aeWv—Ma4rrtaTir the pH at 5.0 ± 0.2 and
continue agitation for 24- hairs. 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
2Q-4Q°C during the extraction. Follow the procedure for manual pH
adjustment in the proposed regulations.
4. Measure and record the plat 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 tolume is 20 times that occupied by a quantity
of water at 4°C equal in wei^it to the initial quantity of solid material
placed in the extractor.
6. Combine this solution with the original liquid phase obtained in the
filtration or centrifugatioa step. Mix thoroughly and split the combined
solution into two equal samples. Store one sample under refrigeration at
1-5°C for organic analysis. Preserve the second sample for elemental
analysis by addition of Ifltisx® nitric acid to reduce the sample pH to
less than 2.
42
-------
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 they are collected.
------- |