PB8-196 001
INTERLABORATORY EVALUATION OF SW-846 METHODS
7470 AND 7471 FOR THE DETERMINATION OF
MERCURY IN ENVIRONMENTAL SAMPLES
U.S. Environmental Protection Agency
Las Vegas, Nevada
Ap 88
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
NIIS
-------
PB88-196001
EPA/600/4-88/011
April 1988
INTERLABORATORY EVALUATION OF SW-846 METHODS
7470 AND 7471 FOR THE DETERMINATION OF
MERCURY IN ENVIRONMENTAL SAMPLES
by
J. E. Gebhart, J. D. Messman, and G. F. Wallace
Battelle Columbus Division
Columbus, Ohio 43201-2693
Contract Number 68-03-3226
Work Assignment 1-12
Project Officer
WERNER F. BECKERT
QUALITY ASSURANCE AND METHODS DEVELOPMENT DIVISION
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
LAS VEGAS, NEVADA 89193-3478
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193-3478
-------
TECHNICAL REPORT DATA
(Please read Instiuctions on the evene before completing)
1 REPORT NO 2
EPA/600/4-88/O11 I
3 REcIPIENT’S ACCESSION NO
R8 1 9 6 0 0 .1 14S
4 TITLE AND SUBTITLE
5 REPORT OATE
INTERLABORATORY EVALUATION OF Sw—846 METHODS 7470 )
7471 FOR THE DETERMINATION OF MERCURY IN ENVIRONMENTAL
April 1988
6.PERFORMINGORGANIZATIONCODE
SAMPLES
7 AUTHOR(S)
I. PERFORMING ORGANIZATION REPORT NO
J.E. Gebhart, J.D. Messman and C.F. Wallace
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle Columbus Division
10. PROGRAM ELEMENT NO
11 CONTRACT/GRANT NO
Columbus, OH 43201—2693
Contract No. 68—03—3226
12 SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring Systems Laboratory — LV, NV
Office of Research and Development
13 TYPE OF REPORT AND PERIOD COVERED
14 SPONSORING AGENCY CODE
U.S. Environmental Protection Agency
Las_Vegas,_NV__89193—3478
EPA/600/O7
15 SUPPLEMENTARY NOTES
16 ABSTRACT
The EPA protocols for SW-846 Methods 7470 and 7471 are cold-vapor atomic
absorption spectrometric (CV-AAS) methods for the determination of mercury in
aqueous and solid environmental samples. In continuation of a previous single-
laboratory study in which a revised more sensitive mercury CV-AAS method for
environmental analyses was evaluated, the revised CV-AAS method has been
subjected to an interlaboratory study. The revised CV-AAS system, operated
in an open configuration, incorporates a dedicated gas sparging bottle for
reduction-aeration and an on-line amalgamation/thermal desorption step. With
these modifications, the CV-AAS method provides increased sensitivity and
also alleviates nonspecific background absorption interferences so that
instrumental background correction is not required. The analytical results
reported by the collaborating laboratories were statistically examined. In
general, the results indicate that the amalgamation CV-AAS method currently
is not sufficiently rugged for routine use but, when properly implemented by
proficient laboratory personnel, may serve as an alternative approach to the
recirculating CV-AAS method described in the current EPA protocols.
17 KEY WORDS AND DOCUMENT ANALYSIS
a DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
C. COSATI Field/Group
18. DISTRIBUTION STATEMENT
19 SECURITY CLASS (This Report)
21 NO OF PAGES
RELEASE TO PUBLIC
TTNCTAgcTrTyn
20 SECURITY CLASS (ThiJ page)
6
22 PRICE
UNCLASSIFIED
EPA Form 2220—1 (R. . 4—77) PREVIOUS EDITION IS OBSOLETE
1
-------
NOTICE
The information in this document has been funded wholly or in part by
the U.S. Environmental Protection Agency under Contract Number 68-03-3226
(Work Assignment 1-12) to the Battelle Memorial Institute, Battelle Columbus
Division, Columbus, Ohio 43201-2693. It has been subject to the Agency’s
peer and administrative review, and it has been approved for publication as
an Environmental Protection Agency document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for
use.
•11
-------
ABSTRACT
The EPA protocols for SW-846 Methods 7470 and 7471 are cold-vapor
atomic absorption spectrometric (CV-AAS) methods for the determination of
mercury in aqueous and solid environmental samples, respectively. In
continuation of a previous single-laboratory study in which a more sensitive
mercury CV-AAS method for environmental analyses was evaluated, the revised
CV-AAS method has been subjected to an interlaboratory study. The revised
CV-AAS system, operated in an open configuration, incorporates a dedicated
gas sparging bottle for reduction-aeration and an on-line
amalgamation/thermal desorption step. With these modifications, the CV-AAS
method provides increased sensitivity and also alleviates nonspecific
background absorption interferences so that instrumental background
correction is not required.
Silver-wool amalgamation cells, mercury stock standard and spiking
solutions, a deionized water sample with spiking instructions, a coal fly
ash reference material, and instructions for analysis by the amalgamation
CV-AAS method were sent to 18 participating laboratories having prior
experience with the current EPA protocols for Methods 7470 and 7471. Ten of
the 18 laboratories were then invited to participate in a more rigorous
collaborative study. To evaluate the revised Method 7470, three aqueous
sample types were analyzed: ground water, waste water, and dilute nitric
acid. To evaluate the revised Method 7471, three solid sample types were
analyzed: marine sediment, incinerator fly ash, and municipal sewage sludge.
Some of the samples were designated for spiking with inorganic or organic
mercury and also with copper, a potential interferent.
The analytical results reported by the collaborating laboratories were
statistically examined in an attempt to characterize the overall accuracy,
precision, and ruggedness of the amalgamation CV-AAS method. In general,
the interlaboratory results indicate that the amalgamation CV-AAS method
currently is not sufficiently rugged for routine use but, when properly
implemented by proficient laboratory personnel, may serve as an alternative
approach to the recirculating CV-AAS method described in the current EPA
protocols. To fully realize the analytical benefits of the amalgamation CV-
AAS method and to obtain accurate and precise data, a complete appreciation
of the integral factors for successful trace analyses is essential, and a
high level of sophisticated operation and operator skill is required.
111
-------
TABLE OF CONTENTS
1
3
5
6
6
8
8
11
11
14
15
15
18
24
32
33
33
41
50
51
52
Notice
Abstract
Figures
Tables
Acknowledgements
1. Introduction
2. Conclusions
3. Recommendations
4. Materials and Methods
Instrumentation
Amalgamation Cells
Reagents
Standards
Environmental Samples..
5. Experimental Procedures
6. Results and Discussion.
• . 11
111
V
V
Vi
Phase I - Laboratory testing of the silver-wool
amalgamation cells and characterization of
the interlaboratory samples
Phase II - Preliminary evaluation of the technical
capabilities of the participating
laboratories
Phase III - Interlaboratory evaluation of the
silver-wool amalgamation CV-AAS method
References
Appendices
A. Method 7470 (Proposed Revision)
B. Method 7471 (Proposed Revision)
C. Phase III - Option A sample list
D. Phase III - Option B sample list
E. Interlaboratory data for Revised Methods 7470 and 7471..
iv
-------
FIGURES
Number Page
1 Schematic diagram of the amalgamation CV-AAS system 7
2 Schematic diagram of the silver-wool amalgamation cell 9
TABLES
Number Page
1 Summary of mercury concentrations in deionized water spiked
with inorganic mercury and in coal fly ash 20
2 Summary of percent recoveries of target mercury concentrations
in spiked deionized water and coal fly ash 22
3 Statistical summary of mercury measurements on spiked
deionized water and coal fly ash 23
4 Summary of laboratories participating in Phase III 24
5 Statistical summary of mercury measurements using
revised Method 7470 27
6 Statistical summary of mercury measurements using
revised Method 7471 28
E-1 Summary of calibration and reagent blank values for
Phase II analyses 52
E-2 Summary of calibration and reagent blank values for
Phase III analyses 53
E-3 Summary of mercury concentrations in ground water 54
E-4 Summary of mercury concentrations in waste water spiked with
inorganic mercury 55
E-5 Summary of mercury concentrations in dilute nitric acid
spiked with inorganic mercury 56
E-6 Summary of mercury concentrations in marine sediment MESS-i 57
E-7 Summary of mercury concentrations in incinerator fly ash 58
E-8 Summary of mercury concentrations in incinerator fly ash
spiked with inorganic mercury 59
E-9 Summary of mercury concentrations in municipal sewage sludge.... 60
E-10 Summary of mercury concentrations in municipal sewage sludge
spiked with organic mercury 61
V
-------
ACKNOWL EDGEMENTS
The authors acknowledge the support of Dr. Werner F. Beckert of the
U.S. EPA Environmental Monitoring Systems Laboratory, Las Vegas, for his
valuable technical input and guidance during the course of this project, and
for his review of this report. Our appreciation is also expressed to
Ms. Heidi deShon of the Columbus (Ohio) Water Testing and Surveillance
Laboratory for graciously furnishing some of the environmental samples used
in this study. The technical contributions of Mr. Douglas M. Jennings,
Mr. Robert L. Livingston, and Mr. Bernard G. Snyder for the laboratory
testing of the silver-wool amalgamation cells and environmental samples, and
for the packaging and shipment of the amalgamation cells and environmental
samples to the participating laboratories are greatly acknowledged. We
acknowledge the contributions of Dr. Steven W. Rust and his associates for
the statistical analyses of the interlaboratory data. The contributions of
Ms. Ramona Mayer and her associates who provided the quality assurance
reviews, Ms. Jennifer Patton and her assistants who provided the graphical
illustrations, and Ms. Julie Teeters who provided secretarial support for
this report are greatly appreciated. We also thank the individual
laboratories for their participation in this collaborative study.
vi
-------
SECTION 1
INTRODUCTION
Publication SW-846, entitled “Test Methods for Evaluating Solid Waste,”
by the Office of Solid Waste and Emergency Response of the U. S.
Environmental Protection Agency provides chemical and physical methods for
the characterization of waste materials regulated under the Resource
Conservation and Recovery Act (RCRA).’ Included in the chemical methods are
two analytical protocols for the determination of mercury in environmental
waste samples by cold-vapor atomic absorption spectrometry (CV-AAS). The
EPA protocols for SW-846 Methods 7470 and 7471 are intended for mercury
determinations in aqueous and solid waste samples, respectively.
Because of the concern for mercury as a potential hazardous pollutant
in environmental wastes, the EPA prot ocols for mercury are under review. In
a previous single-laboratory study,’ the EPA protocols for SW-846 Methods
7470 and 7471 were evaluated and revised to improve analytical performance.
The methods were evaluated using aqueous and solid environmental samples of
homogeneous and known compositions in order to assess accuracies and
precisions of the methods without introducing uncertainties due to sample
heterogeneities.
The results of the single-laboratory study 2 ’ 3 indicated, in general,
that the digestion procedures were satisfactory for the samples analyzed;
only minor revisions that would improve, clarify, or increase the
flexibility of the digestion procedures were recommended. The recirculating
CV-AAS method described in the current EPA protocols was found to be
adequate only for mercury determinations in samples of relatively high
mercury concentrations. However, the recirculating CV-AAS method was
inadequate for mercury determinations in samples of low mercury levels
because of its relatively poor sensitivity. Moreover, interferences which
were due to nonspecific absorption of primary mercury radiation by vapors of
organic compounds (benzene and methyl ethyl ketone) degraded the accuracy of
the recirculating CV-AAS method.
Significant modifications of the cold-vapor apparatus were recommended
and evaluated to overcome the analytical limitations of the recirculating
CV-AAS method. Instrument detectability was improved 10-fold by
collectively using a gas sparging bottle as a dedicated reduction-aeration
vessel and silver-wool amalgamation in the CV-AAS system operated in an open
configuration. The modified CV-AAS method, with the incorporation of these
three instrument modifications, was designated amalgamation CV-AAS.
Although the individual contributions of the three instrument modifications
were not determined, the amalgamation step provided much of the overall
increase in sensitivity. The on-line amalgamation/thermal desorption
process of the modified CV-AAS system also effectively eliminated
interfering water and organic matrix vapors prior to the mercury absorption
1
-------
measurement so that an instrument without dynamic background correction
capabilities could be used. Good accuracy and precision were obtained with
the amalgamation CV-A.AS system for the analyses of four reference sediment
materials.
The objective of this project was to conduct an interlaboratory
evaluation of the revised EPA protocols for the determination of mercury in
environmental samples. The mercury protocols to be evaluated included minor
revisions in the digestion procedures of Methods 7470 and 7471, and
quantif c tion of mercury in the sample digests by the amalgamation CV-AAS
method. ” The interlaboratory study was conducted in three phases: (1)
Phase I - fabrication and testing of the silver-wool amalgamation cells and
characterization of selected interlaboratory study samples by the lead
laboratory, (2) Phase II - preliminary evaluation of the technical
capabilities of the participating laboratories to utilize the amalgamation
CV-AAS method for determining mercury in a coal fly ash reference material
and in deionized water spiked with inorganic mercury, and (3) Phase III -
evaluation by the qualifying laboratories of the amalgamation CV-AAS method
for the determination of mercury in a marine sediment reference material and
in representative aqueous and solid waste samples. The test samples
contained concentrations of endogenous mercury or were spiked with
concentrations of inorganic or organic mercury that were in the optimal
range of the amalgamation CV-AAS method. A statistical analysis of the
mercury concentration data submitted to Battelle was conducted in an attempt
to characterize the interlaboratory accuracy, precision, and ruggedness of
the amalgamation CV-AAS methods.
2
-------
SECTION 2
CONCLUS IONS
The results of this interlaboratory study indicate that only a few of
the collaborating laboratories demonstrated proficiency in incorporating the
silver-wool amalgamation apparatus into their CV-AAS systems within a
relatively short learning period. These laboratories reported calibration
slopes in an acceptable range based on Battelle’s experience with the
amalgamation CV-AAS method. Only a few of the laboratories were able to
perform accurate and precise measurements on both aqueous and solid
environmental samples.
A statistical treatment of the interlaboratory results indicates that
the amalgamation CV-AAS method currently lacks sufficient ruggedness to be
successfully used by analytical laboratories for routine testing but, when
properly implemented by proficient laboratory personnel, it may serve as an
alternative to the recirculating CV-AAS method described in the current EPA
protocols of Methods 7470 and 7471. However, to fully realize the
analytical benefits of the amalgamation CV-AAS method and to obtain accurate
and precise data, a complete appreciation and awareness of the critical
factors affecting successful trace analyses are essential, and a high level
of sophisticated operation and operator skill is required. Moreover,
because of the increased sensitivity of the amalgamation CV-AAS method,
improved analytical techniques and the use of high-purity reagents are
required to minimize mercury contamination that would not be a problem when
using the recirculating CV-AAS method with samples of higher mercury
concentrations.
To achieve better results, many of the laboratories reported that they
had to modify the operating parameters of the silver-wool amalgamation
apparatus from those specifically described in the revised protocols. This
may have been necessary when the silver-wool amalgamation apparatus was not
assembled and configured properly according to the directions in the
protocols, or when the amalgamation apparatus had to be reconfigured to fit
a specific commercial or customized CV-A.AS system. The different operating
conditions used by some laboratories may have contributed to the high
dispersion in the amalgamation CV-AAS data submitted to Battelle.
Most of the comments from the collaborating laboratories addressed the
increased sensitivity and the longer analysis time of the amalgamation CV-
AAS method. Many laboratories recognized the need for higher operator skill
and the importance of the need for high-purity reagents to minimize mercury
contamination. The benefit of an amalgamation CV-AAS method with increased
sensitivity but at the expense of longer analysis time was challenged by
some laboratories from a business economics viewpoint; the longer analysis
time would increase sample analysis costs to the laboratory that would have
to incur the additional costs or charge higher analysis fees to the
customer. Although the faster recirculating CV-AAS method was considered
adequate for the majority of their samples, other laboratories expressed
special needs for the increased sensitivity of the amalgamation CV-AAS
3
-------
method and indicated that it would be valuable in research and for difficult
analysis problems. Some laboratories expressed enthusiasm about the
amalgamation CV-MS method and indicated that they plan to incorporate this
method into their laboratory operations for additional flexibility. The
extent of time and cost constraints for adapting the amalgamation CV-AAS
method to a specific sample workload may be influenced by whether the
emphasis of the analytical laboratory is on commercial testing or on
research, development and specialty problems in a support function of a
larger organization.
4
-------
SECTION 3
RECOMMENDATIONS
The results of this interlaboratory study, in conjunction with the
results of the previous single-laboratory study, indicate that the
recirculating CV-AAS method and the amalgamation CV-AAS method have distinct
analytical advantages for the determination of mercury in environmental
waste samples. However, until its ruggedness is improved, the amalgamation
CV-AAS method is not suitable for routine testing by most analytical
laboratories. Battelle recommends that the amalgamation CV-AAS method would
best serve the analytical community as an alternative method for the
determination of trace mercury concentrations in those complex sample types
which can not be routinely analyzed by the current recirculating CV-AAS
method. Specifically, Battelle recommends that the CV-AAS methods be used
to characterize waste samples for mercury as described below:
(1) The current EPA protocols for the recirculating CV-AAS
method should be used for analyses when the threshold
concentration value for mercury is relatively high and
when the samples do not contain volatile organic
compounds that could cause nonspecific absorption
interferences.
(2) The amalgamation CV-AAS method should be used for the
analyses of samples containing low mercury
concentrations which can not be easily determined by the
recirculating CV-AAS method.
(3) The amalgamation CV-AAS method should be used for the
analyses of samples containing volatile organic
compounds which will cause nonspecific absorption
interferences and inaccurate results with the
recirculating CV-AAS method.
Because of the greater sensitivity and the more sophisticated apparatus
of the amalgamation CV-AAS system, Battelle emphasizes that the laboratory
analysts be properly trained and that they demonstrate proficiency with the
amalgamation method before initiating regulatory analyses for mercury.
5
-------
SECTION 4
MATERIALS AND METHODS
INSTRUMENTATION
The amalgamation CV-AAS system used in Phase I i identical to the
system previously used in the single-laboratory study.” 3 The schematic
diagram of the amalgamation CV-AAS system is shown in Figure 1. The
operating procedures and instrument parameters are the same as those used in
the single-laboratory study except for the primary mercury radiation source
used in the atomic absorption spectrophotometer. In the present study, a
mercury electrodeless discharge lamp (EDL), supplied with 5 watts of radio-
frequency power, was employed instead of the mercury hollow cathode lamp
(HCL). The greater spectral intensity of the 253.7-nm resonance line from
the EDL allows a lower gain setting to be used for the photomultiplier tube
(PMT) detector resulting in a higher signal-to-noise ratio. The relative
sensitivities, or slopes, of the calibration curves for the two mercury
spectral-line sources are similar.
The amalgamation cell is positioned in the CV-AAS system as shown in
Figure 1. Mercury vapor evolved from the reduction-aeration cell is trapped
and concentrated by amalgamation on the silver-wool plug positioned in the
center of the amalgamation cell. After the reduction-aeration and the on-
line amalgamation steps are completed, mercury is desorbed from the silver
wool by resistance heating to approximately 500°C. Heat is supplied by
applying approximately 20 volts of power to the Chromel A wire windings.
The mercury vapor is transported via a nitrogen-gas purge into the
absorption cell, and the maximum peak absorbance is recorded. The geometry
of the absorption cell used in the single-laboratory and interlaboratory’
studies was satisfactory although no attempts were made to optimize it for
sensitivity. By using the parameters and experimental conditions described
in the revised protocols, the appearance time of the mercury peak on the
strip-chart recorder with a 1-second time constant is approximately 17
seconds; the maximum peak absorbance occurs between 25 and 30 seconds after
heating of the silver-wool plug is initiated.
Four operating parameters of the amalgamation CV-AAS system w r
optimized and tested for ruggedness in the single-laboratory study.”
These parameters included the nitrogen purge-gas flow rate, the time
required for the silver-wool concentration of mercury, the Variac voltage
setting for thermal desorption, and the mass of the silver wool. The
optimized values for these parameters were used in interlaboratory study.
The timing of the amalgamation and desorption events are especially
critical for maximum precision. Based on Battelle’s experience, the timing
should be reproducible to within a few seconds. A stopwatch is conveniently
used to consistently monitor all timing events in the method to insure
maximum reproducibility for within-day and day-to-day analyses.
6
-------
Charcoal
Needle Valve Trap
Nitrogen
Cylinder
Purging Cylinder
(Reduction-Aeration
Sample Cell)
V
Glass Stopcock
—4
Glass
Absorption
Cell
Ty go n - to - Glass
Connections
Silver Wool
Flow Meter
Chrom-Alumel
Resistance Heating
Winding
Figure 1. Schematic diagram of the amalgamation CV-AAS system.
-------
In addition to residual vapors resulting from matrix components that
survive the digestion procedure, water vapor evolved during the reduction-
aeration step may absorb the primary mercury radiation and produce
nonspecific absorption interferences. The amalgamation/thermal desorption
process has been demonstrated to successfully eliminate interfering water
and organic vapors prior to the mercury a b orption measurement without
compromising the instrument detection limit.” Although continuum-source
background correction should also provide accurate compensation for such
vapor interferences, the increased noise level in most of the older atomic
absorption systems when operated in the background-correction mode degrades
the det c ion limit. This effect has been previously discussed in greater
detail.” An additional benefit of amalgamation CV-AAS is that the method
can be successfully implemented even by using an atomic absorption
instrument that is not equipped with a dynamic background-corrector
accessory.
AMALGAMATION CELLS
To minimize experimental variability, all of the mercury amalgamation
cells used in this study were manufactured, assembled, and tested by
Battelle staff. The amalgamation cell with ball-joint connection adapters
allowed it to be fitted to commercial cold-vapor mercury accessories.
Twenty-three amalgamation cells were constructed, according to the design
and specifications in Figure 2, by Battelle glassblowers. The amalgamation
cell, with an overall length of 105 mm, consists of ST 12/5 male and female
ball joints annealed to opposite ends of Pyrex glass tubing (5 mm i.d.). A
silver-wool plug (Fisher Scientific Company, Fair Lawn, NJ) of 0.7 g, with a
mass uncertainty of 0.05 g, is inserted into the Pyrex tube and then
compacted into a cylinder having approximate dimensions of 5 mm diameter and
20 mm length. This is facilitated by using two narrow stainless steel rods
which are directed through the two ends of the Pyrex tube; the silver wool
is compacted within the centermost part of the tube length so that it is
snug within the tube and so it will not easily be dislodged. A 92-cm length
of 22-gauge Chromel A wire (Fisher Scientific Company, Fair Lawn, NJ) is
wrapped around the Pyrex tubing, providing 30 windings and a resistance of
22 ohms/ft. This wire-wrapping provides uniform heating over a region
approximately three times longer than the silver-wool plug. The O.7-g plug
of silver wool compacted into a cylinder under these conditions
quantitatively amalgamates mercury vapor at a nitrogen carrier-gas flow rate
of 0.55 L/rnin with no apparent buildup of leak-inducing backpressure.
REAGENTS
Deionized water with a minimum electrical resistivity of 14 megohm-cm
(The Barnstead Company, Division of Sybron Corporation, Boston, MA) or
equivalent; deionized water with a minimum electrical resistivity of 1
megaohm-cm (Peck Water Systems Corporation, Canton, OH) served as the inlet
source for the Sybron/Barnstead deionization unit.
Hydrochloric acid, HC1, concentrated (37 percent, 12 Normal), FW 36.5,
“Baker Ultrex” grade (J. 1. Baker Chemical Company, Phillipsburg, NJ) or
equivalent.
8
-------
Pyrex Glass
105 mm
—I
Ball
15-mm Hole
Chromel A
Wire
12/1
Silver-Wool Plug
l
Figure 2. Schematic diagram of the silver-wool amalgamation cell.
-------
Nitric acid, HNO 3 , concentrated (70 percent, 16 Normal), FW 63.0,
“Baker Ultrex” grade (J. 1. Baker Chemical Company, Phillipsburg, NJ) or
equivalent.
Nitric acid, 1-percent (v/v) solution: Prepared by carefully mixing 1
volume of concentrated nitric acid with 99 volumes of deionized water.
Sulfuric acid, H 2 S0 4 , concentrated (96 percent, 36 Normal), FW 98.08,
analytical reagent grade (Mallinckrodt, Inc., Paris, KY) or equivalent.
Aqua regia: Prepared immediately before use by carefully adding three
volumes of concentrated hydrochloric acid to one volume of concentrated
nitric acid.
Hydroxylamine hydrochloride, NH 2 OHHC1, FW 69.49, “Baker Analyzed”
reagent grade (J. T. Baker Chemical Company, Phillipsburg, NJ) or
equivalent.
Sodium chloride, NaC1, FW 58.44, “Baker Analyzed” reagent grade (J. T.
Baker Chemical Company, Phillipsburg, NJ) or equivalent.
Sodium chioride-hydroxylamine hydrochloride solution: Prepared by
dissolving 60 g sodium chloride and 60 g hydroxylamine hydrochloride in
deionized water in a 500-mL volumetric flask, and by diluting to calibrated
volume with deionized water.
Potassium permanganate, KMnO 4 , FW 158.04, “Baker Analyzed” ACS reagent
grade (J. 1. Baker Chemical Company, Phillipsburg, NJ) or equivalent.
Potassium permanganate, 5-percent (w/v) solution: Prepared by
dissolving 50 g potassium permanganate in approximately 500 mL deionized
water in a 1-L volumetric flask, and by diluting to calibrated volume with
deionized water.
Potassium persulfate, K 2 S 2 0 8 , FW 270.32, “Baker Instra-Analyzed” ACS
reagent grade (J. T. Baker Chemical Company, Phillipsburg, NJ) or
equivalent.
Potassium persulfate, 5-percent (w/v) solution: Prepared by dissolving
12.5 g potassium persulfate in deionized water in a 250-mL volumetric flask,
and by diluting to calibrated volume with deionized water.
Stannous chloride, SnC1 2 2H 2 O, FW 225.63, “Baker Analyzed” ACS reagent
grade (J. T. Baker Chemical C ompany, Phillipsburg, NJ) or equivalent.
Stannous chloride, 15-percent, (w/v) solution: Prepared by dissolving
150 g stannous chloride in 300 mL concentrated hydrochloric acid and by
diluting to 1 L with deionized water.
Copper nitrate, 1000 mg/L, cupric nitrate in dilute nitric acid, Fisher
atomic absorption standard (Fisher Scientific Company, Fair Lawn, NJ) or
equivalent.
10
-------
Methyl mercuric chloride, CH 3 HgC1, FW 251; Pfaltz and Bauer, Inc.,
Waterbury, CT, or equivalent.
STANDARDS
Mercury stock solution, 1000 mg/L, mercuric chloride in distilled
water; Fisher atomic absorption standard (Fisher Scientific Company, Fair
Lawn, NJ) or equivalent.
Mercury intermediate standard, 10 mg/I: Prepared by transferring 1.0 ml
of 1000-mg/L stock solution to a 100-nil volumetric flask containing 90 ml
deionized water and 1 mL concentrated nitric acid, and by diluting to
calibrated volume with deionized water.
Mercury intermediate standards, 1 mg/L and 0.1 mg/L: Prepared by
transferring 10.0 mL and 1.0 ml, respectively, of 1O-mg/L intermediate
standard to individual 100-mL volumetric flasks containing 90 mL deionized
water and 1 mL concentrated nitric acid, and by diluting to calibrated
volume with deionized water.
Methyl mercuric chloride stock solution, 900 mg/I Hg nominal
concentration: Prepared by transferring 0.11-g CH 3 HgC1 into a 100-mI
volumetric flask; the weighed aliquot was dissolved in deionized water and
then diluted to calibrated volume with deionized water.
Methyl mercuric chloride intermediate standard, 9 mg/I Hg nominal
concentration: Prepared by transferring 1.0-niL of the methyl mercuric
chloride stock solution to a 100-niL volumetric flask containing 90 ml
deionized water and 1 niL concentrated nitric acid, and by diluting to
calibrated volume with deionized water.
Methyl mercuric chloride intermediate standards, 0.9 mg/L and 0.09
mg/I Hg nominal concentrations: Prepared by transferring 10.0 ml and 1.0 ml,
respectively, of the nominal 9-n ig/L intermediate standard to individual 100-
ml volumetric flasks containing deionized water and 1 niL concentrated nitric
acid, and by diluting to calibrated volume with deionized water.
ENVIRONMENTAL SAMPLES
NBS-SRM 1633a (Coal Fly Ash )
This Standard Reference Material (SRM), purchased from the National
Bureau of Standards (NBS), Gaithersburg, MD, is a representative fly ash
collected from coal-fired power plants; this fly ash material is generated
from Pennsylvania and West Virginia coals. The certified concentration for
mercury in NBS-SRM 1633a is 0.16 ig/g with an uncertainty of 0.01 g/g. The
estimated uncertainty is based on judgment and represents an evaluation of
the combined effects of method imprecision, possible systematic errors among
methods, and material variability for sample sizes of at least 250 mg.
11
-------
The certified values for the principal matrix components include:
silicon - 22.8 percent, iron - 9.40 percent, potassium - 1.88 percent, and
calcium - 1.11 percent. Although not certified, a concentration value of 14
percent aluminum is provided for information use.
MESS-i (Marine Sediment )
This reference material, purchased from the National Research Council
of Canada (NRCC), Ottawa, Canada, is a freeze-dried sediment originally
obtained from the Gulf of St. Lawrence (Miramichi River estuary). The
reference value for mercury in MESS-i is 0.171 pg/g with an uncertainty of
0.014 pg/g. The uncertainty represents 95-percent tolerance limits for an
individual sub-sample of 500 mg or greater. -
The principal matrix components and their reference concentrations
include: silicon dioxide - 67.5 percent, aluminum oxide - 11.03 percent,
iron(1I) oxide - 4.36 percent, sodium oxide - 2.50 percent, potassium oxide
- 2.24 percent, and magnesium oxide - 1.44 percent. The major and minor
inorganic matrix components represent approximately 95 weight percent of the
total sediment material.
Incinerator Fly Ash
A composite inorganic ash sample was prepared from incinerator ash
obtained from four municipal refuse incinerator sites located in Florida,
Massachusetts, New Jersey, and Oregon. The original composite sample was
comprised of gray ash particles ranging in size from fine dust to
approximately 2.5 cm in diameter. A representative sample of this bulk
material was air-dried and pulverized with a clean rubber mallet. The large
chunks of ash were removed by passing the sample through a 9.5-mm sieve.
The remaining ash material was then consecutively passed through 40-mesh
(420 micrometers) and 200-mesh (74 micrometer) sieves, placed in a plastic
bag and homogenized by tumbling. Approximately 60 g of the sieved ash
material were heated in a muffle furnace at a temperature of 600°C for
approximately 24 hours to volatilize any endogenous mercury. No detectable
mercury was measured in this treated composite ash sample, as characterized
in Phase I by the amalgamation CV-AAS method.
Municipal Sewage Sludge
A composite sample of dried municipal sewage sludge was obtained from
the Water Surveillance Laboratory of Columbus, Ohio. The sample was air-
dried and then sieved to 100 mesh. The sieved material was homogenized by
tumbling, and was then stored in a plastic sample bag. The mercury
concentration of the composite sample, characterized in Phase I by the
amalgamation CV-AAS method, was found to be approximately 1.4 pg/g.
Ground Water
A ground-water sample was obtained from a farm well west of Columbus,
Ohio. The sample was filtered through 0.45-micron Millipore filter paper
12
-------
and acidified to 1 percent (v/v) with nitric acid. The ground-water sample
did not contain any detectable mercury, as characterized in Phase I by the
amalgamation CV-MS method.
Waste Water
A liquid waste sample was obtained from the Water Surveillance
Laboratory of Colunibus, Ohio. The waste water sample did not contain any
detectable mercury.
Deionized Water
Laboratory deionized water was used as the aqueous sample in Phase II.
Laboratory deionized water acidified to one percent with nitric acid was
used in Phase III. The deionized water samples did not contain .any
detectable mercury.
13
-------
SECTION 5
EXPERIMENTAL PROCEDURES
The environmental samples were prepared for analysis by using the
digestion procedures described in revised SW-846 Methods 7470 and 7471 which
are included as Appendices A and B. All sample digests were diluted to
volume in 100-mL volumetric flasks with deionized water following the
addition of the sodium chioride-hydroxylamine hydrochloride reagent solution
to reduce manganese dioxide and excess permanganate to the soluble divalent
manganese form. This procedure permitted sampling of multiple aliquots and
further dilutions as needed. The sample aliquot added to the reduction-
aeration vessel was diluted to 100 mL with Type II water.
All mercury concentrations were calculated by the use of 4-point
calibration curves. The four points were comprised of a reagent blank and
three mercury standards within the linear range of the amalgamation CV-AAS
method. In most cases, the calibration standards contained 10, 50, and 100
ng of mercury, respectively.
The mercury concentration data submitted to Battelle for each analysis
option and for each sample type in Phase III were statistically analyzed.
The accuracies of the revised Methods 7470 and 7471 were assessed by
comparing the measured mercury concentrations reported by the laboratories
with the target mercury concentrations for each sample. The precisions of
the methods were estimated for each sample in Options A and B from the
standard deviations calculated for duplicate and triplicate analyses,
respectively. The ruggedness of the methods was qualitatively derived from
the relative accuracies and precisions.
14
-------
SECTION 6
RESULTS AND DISCUSSION
PHASE I - LABORATORY TESTING OF THE SILVER-WOOL AMALGAMATION
CELLS AND CHARACTERIZATION OF THE INTERLABORATORY
STUDY SAMPLES
Testing of Amalgamation Cells
To facilitate interpretation of the data from the participating
laboratories, the variability in instrument response for different silver-
wool amalgamation cells was characterized. To minimize variability in the
instrument responses, the silver-wool amalgamation cell must possess the
following two characteristics: (1) quantitative amalgamation of at least 100
ng of mercury, and (2) a uniform rate of thermal desorption of amalgamated
mercury. The second characteristic is particularly important for the
present method because mercury is quantified by peak-height measurement.
Differences in desorption rates among amalgamation cells could lead to
different absorbance peak heights, thus adversely affecting sensitivity. If
peak-area integration is used for quantification, a uniform rate of thermal
desorption may be less important.
The uniformity of performance for the 23 silver-wool amalgamation cells
was verified before the cells were distributed to the participating
laboratories. The parameter used to assess uniformity of performance among
the amalgamation cells was the slope of their calibration curves. If the
amalgamation cells provide uniform performance in terms of amalgamation
efficiency and thermal-desorption rates, the slopes and intercepts of the
calibration curves obtained for the respective amalgamation cells should be
similar.
Each amalgamation cell to be tested was positioned within the
instrument configuration; the inlet was connected to the cold-vapor
apparatus but the outlet was disconnected from the absorption cell. With
nitrogen purge gas flowing through the apparatus, the Variac autotransformer
set to 20 volts output was turned on, and each amalgamation cell was pre-
conditioned by heating for 2.5 minutes to clean the silver-wool plug before
conducting the sensitivity check.
Four-point calibration curves were constructed for each of the first 8
amalgamation cells using a reagent blank and aqueous standards containing
10, 50, and 100 ng of mercury. The average value of the slope of the
calibration curve obtained for these cells was 0.0058 absorbance units/ng of
mercury. The standard deviation of these measurements was 0.0006 absorbance
units/ng, which yields a relative standard deviation of 10 percent. The
slopes of the calibration curves for the 8 amalgamation cells ranged from
0.0048 to 0.0065 absorbance units/ng of mercury, and the intercepts ranged
from -0.003 to 0.006 absorbance units. The average intercept, 0.002
absorbance unit, is small and sufficiently close to zero.
15
-------
The relative sensitivities of the remaining 15 amalgamation cells were
verified in a more expeditious manner. After calibrating the instrument
with a single 4-point curve by using the aqueous mercury standards described
above, the performance of each of these amalgamation cells was checked with
a single 50-ng Hg standard. The calibration curve used to assess the cell
performances had a slope of 0.0052 abs/ng of mercury and an intercept of
0.003 absorbance unit; the slope and intercept were within the experimental
ranges of those values determined for the first 8 cells. The average
recovery of the 50-ng Hg standard for each of these 15 cells was 98 percent,
with an experimental uncertainty of 6 percent. The range of recoveries was
89 to 108 percent. These results indicate that the instrument responses for
50 ng Hg for the 15 amalgamation cells are within ± 10 percent of the target
concentration which, based on Battelle’s experience, constitutes a
sati sfactory performance.
The overall testing results indicate uniformity in performance among
the 23 amalgamation cells with respect to amalgamation efficiency and
thermal -desorption characteristics. The differences in instrument responses
for the different amalgamation cells are expected to be small compared to
the overall measurement variabilities between the participating
laboratories.
Characterization of Interlaboratorv Study SamDles
The aqueous and solid environmental samples to be used in the
interlaboratory study were characterized for endogenous mercury content by
using the revised Methods 7470 and 7471 presented in Appendices A and B,
respectively. Replicate 0.2-g portions of solid samples and 20-mL or 50-mL
portions of liquid samples were used for the analyses.
NBS-SRM 1633a (Coal Fly Ash)--
NBS-SRM 1633a (Coal Fly Ash) was selected as a representative solid
waste sample to be used in a preliminary evaluation of the revised Method
7471 in Phase II. This sample was selected because it had a certified
mercury value in the concentration range useful for evaluating the silver-
wool amalgamation CV-AAS method, and the homogenous material would not
introduce errors due to sample heterogeneity.
The coal fly ash sample was analyzed by Battelle staff according to the
revised Method 7471. The mercury results for analyses of duplicate 0.2-g
portions were 0.145 pg/g and 0.160 pg/g. These results are in agreement
with the certified mercury concentration of 0.16 g/g. The recoveries of
inorganic mercury predigestion spikes added to the fly ash samples were 130
percent and 130 percent for duplicate 10-ng mercury spikes, and 112 percent
and 104 percent for duplicate 50-ng mercury spikes.
Incinerator Ash--
A composite sample of municipal refuse incinerator ash was used as one
of the solid waste samples for the evaluation of the revised Method 7471 in
Phase III. The incinerator ash sample was analyzed by Battelle staff
according to the revised Method 7471 to determine the endogenous mercury
16
-------
content that would be used as the target mercury concentration for the
participating laboratories. No detectable mercury was found in triplicate
O.2-g portions of the composite ash sample. No inorganic or organic mercury
predigestion spiking studies were performed.
Municipal Sewage Sludge--
A municipal sewage sludge was one of the solid waste samples used for
the evaluation of the revised Method 7471 in Phase III. Replicate portions
of the sludge sample were analyzed on different days by Battelle staff
according to the revised Method 7471 to determine the endogenous mercury
content that would be used as the target mercury concentration for the
participating laboratories. The average mercury concentration determined in
8 replicate 0.2-9 portions of the sewage sludge sample was 1.4 ig/g; the
compiled results ranged from approximately 1.2 ig/g to 1.4 g/g of mercury.
This average mercury value agrees with data previously obtaineçl from
analyses conducted by the Columbus Water Surveillance Laboratory.’+ The
results indicate that the mercury content of the municipal sewage sludge is
homogeneous to approximately ± 0.1 mg/kg. Of these 8 replicate
measurements, some of the portions were analyzed by using inorganic mercury
standards for calibration while the other portions were analyzed by using
organic mercury calibration standards. The average recoveries of duplicate
90-ng and 180-ng Hg predigestion spikes (as methyl mercuric chloride) were
75 percent and 90 percent, respectively, and the average recoveries of
duplicate 100-ng and 200-ng Hg predigestion spikes (as inorganic mercuric
chloride) were 108 percent and 106 percent, respectively. The overall
recoveries of organic and inorganic mercury added as predigestion spikes are
within ± 25 percent of full recovery.
Ground Water- -
A ground-water sample was used as one of the aqueous environmental
samples for the evaluation of the revised Method 7470 in Phase III. The
ground-water sample was analyzed by Battelle staff according to the revised
Method 7470 to determine the endogenous mercury content that would be used
as the target mercury concentration for the participating laboratories. No
detectable mercury was found in four replicate 50-mL portions of the ground-
water sample. The average inorganic spike recoveries of four sample
replicates were 106 percent for 50-ng mercury spikes and 105 percent for
100-ng mercury spikes.
Waste Water- -
A waste-water sample was one of the aqueous samples used for the
evaluation of the revised Method 7470 in Phase III. Although not analyzed
at Battelle, the waste-water sample did not contain any detectable
endogenous mercury based on the results oX analyses conducted by the Water
Surveillance Laboratory of Columbus, Ohio.
17
-------
PHASE II - PRELIMINARY EVALUATION OF THE TECHNICAL CAPABILITIES
OF THE PARTICIPATING LABORATORIES
Selection of the Laboratories
The initial goal of Phase II was to identify laboratories that would
want to become acquainted with the silver-wool amalgamation accessory and
the experimental procedures for successfully implementing the CV-AAS method
for trace mercury determinations in environmental and waste samples.
Requests for proposals to participate in the interlaboratory study were sent
to 90 laboratories; copies of the revised Methods 7470 and 7471 were also
enclosed to provide the laboratories a detailed technical description of the
silver-wool amalgamation step. Twenty-two laboratories responded with
proposals which included equipment descriptions, qualifications of
personnel, previous relevant experience, and cost bids. Contracts to
participate in Phase II were sent to the 22 laboratories. Eighteen
laboratories responded with signed contracts. A package containing the
necessary materials was shipped to each of these 18 laboratories that had
signed contracts, and to one additional laboratory that participated at no
cost. Each package contained: (1) a tested silver-wool amalgamation cell,
(2) instructions for the installation and use of the amalgamation cell, (3)
two samples and appropriate standard solutions, (4) instructions for the
preparation and analysis of the samples, and (5) instructions for reporting
the analytical results to Battelle. One laboratory with a signed contract
withdrew at the beginning of Phase II after receiving the package, leaving a
total of 18 participating laboratories. Throughout this report, these
laboratories are identified by numbers.
Samples and Instructions for Analysis
The two samples shipped to the participating laboratories consisted of
deionized water for the evaluation of revised Method 7470 and NBS-SRM 1633a
(Coal Fly Ash) for the evaluation of revised Method 7471. A spiking
solution, containing mercury at a concentration of 1.5 mg/L, was included
for spiking the deionized water. The participants were instructed to add 50
L of the 1.5-mg/L spiking solution (equivalent to 75 ng Hg) to a 20-mL
aliquot of the water sample as a predigestion spike. The laboratories were
also provided a stock solution containing mercury at a concentration of 1000
mg/L to be used for preparing intermediate and final calibration standards.
For the analyses of the samples, the participants were instructed to:
(1) construct a 4-point calibration curve comprised of the absorbances for a
reagent blank, 10-ng, 50-ng, and 100-ng Hg standards, (2) analyze the spiked
water sample according to the revised Method 7470, (3) analyze the coal fly
ash sample according to the revised Method 7471, and (4) prepare and analyze
one reagent blank each for the revised Methods 7470 and 7471. For the
reporting of the analytical results, the participants were instructed to:
(1) report the slopes of the calibration curves for both methods in units of
absorbance/nanogram of mercury, (2) report the measured concentration of
mercury in the spiked water sample in units of mg/L, (3) report the measured
concentration of mercury in the coal fly ash sample in units of mg/kg, and
(4) report the absorbance values for the reagent blanks for both methods.
18
-------
Laboratory Results
The slopes of the calibration curves and the mercury absorbance values
for the reagent blanks obtained from the 18 laboratories are summarized in
Table E-1 of Appendix E. These data were not statistically examined for
outliers or as criteria for selecting specific laboratories to participate
in Phase III. Five laboratories were unable to report data for at least one
of the revised methods because of analytical and/or instrumental
difficulties.
The slopes of the calibration curves for the two revised methods ranged
from 0.0001 to 0.0056 absorbance units/nanogram. Ten of the 18 laboratories
reported slopes between 0.003 and 0.006 absorbance units/nanogram for at
least one of the revised methods. Battelle researchers have found that this
is a typical range for the slope of the calibration curve obtained by using
the amalgamation CV-AAS system according to the specific instructions
described in revised Methods 7470 and 7471.
The mercury absorbance values for the reagent blanks ranged from 0.001
to 0.644 absorbance units. Ten laboratories reported reagent blank values
less than 0.03 absorbance unit. Battelle researchers have consistently
measured absorbance values between 0.02 and 0.03 absorbance units for
reagent blanks analyzed according to the revised Methods 7470 and 7471.
Based on a typical instrument calibration slope of 0.005 absorbance
units/nanogram, reagent blank absorbances in this range correspond to
approximately 5 ng Hg. Reagent blanks for aqueous mercury calibration
standards, consisting only of the stannous chloride reductant and deionized
water, generally produce 0.001 absorbance unit; this is negligible relative
to the absorbance values for the reagent blanks carried through the
digestions. Some of the apparent low reagent blank values reported by the
laboratories may be suppressed in those cases in which the reported
calibration slopes were lower than the typical range of values.
The results for the measured mercury concentrations in the spiked water
sample and the coal fly ash sample are summarized in Table 1. The target
mercury values are 0.00375 mg/L for the spiked water sample and 0.16 mg/kg
for the coal fly ash sample. For the specified sample sizes, these values
correspond to absolute mercury target values of 75 ng for the spiked water
sample and 32 ng for the coal fly ash sample that should be measured by the
instrument; both of these mercury values fall within the optimal measurement
range of the amalgamation CV-MS method.
As shown in Table 1, two laboratories did not submit mercury
concentration results for the spiked deionized water sample, and five
laboratories did not submit mercury concentration results for the coal fly
ash sample. These laboratories experienced analytical and/or instrumental
problems with these preliminary sample analyses. For those laboratories
that reported results, the measured mercury concentrations for the spiked
water sample ranged from 0.00076 to 0.076 mg/L, and for the coal fly ash
sample from 0.04 to 48 mg/kg.
19
-------
TABLE 1. SUMMARY OF MERCURY CONCENTRATIONS IN DEIONIZED WATER
SPIKED WITH INORGANIC MERCURY AND IN COAL FLY ASH
Laboratory
Identi fication
Method 7470
(Water Sample)
Method 7471
(Coal Fly Ash SamDle)
Mercury
concentrationb
Mercury
Concentrati 0 a
Number
(mg/L)
(mg/kg)
949 0.0041
950 0.00076 0.22
951
952 0.0030 0.14
953 0.0030 0.14
954 0.0074 0.40
955 0.0046 0.71
956 0.0023 0.04
957
958 0.0034
959 0.00091 0.41
960 0.0016 0.68
961 0.0054
962 0.0046 0.23
963 0.076 48
964 0.0074 1.1
965 0.0035 0.15
966 0.0036 0.17
a Target mercury concentration = 0.00375 mg/L.
b NBS-SRM 1633a; certified mercury concentration = 0.16 ± 0.01 mg/kg.
20
-------
The percent recoveries of the target mercury concentrations in the
spiked water sample and the coal fly ash sample are summarized in Table 2.
The percent recovery is calculated as 100 times the measured concentration
divided by the target concentration. Excluding the results from Laboratory
No. 963 for spiked deionized water, the mercury recoveries ranged from 20
percent (a 5-fold negative bias) to 200 percent (a 2-fold positive bias) for
all laboratories as well as for just the Phase-Ill laboratories. Excluding
also the results from Laboratory No. 963 for the coal fly ash, the mercury
recoveries ranged from 25 percent (a 4-fold negative bias) to 690 percent (a
7-fold positive bias) for all laboratories, and from 88 to 690 percent for
the Phase Ill-laboratories.
A summary of the statistics across all laboratories for the mercury
measurements on the spiked deionized water and coal fly ash samples is
presented in Table 3. The means and standard deviations, reported as
percent recoveries, were calculated for all laboratories and also for only
those eventually selected for Phase III. In Table 3 and in all subsequent
tables containing statistical data, the means and standard deviations are
reported to two significant figures. 5 The results from Laboratory No. 963
were excluded as outliers from the statistical treatment. After these data
values were removed, the remaining data were smoothly dispersed, and no
additional outliers were obvious.
Although the measured mercury concentrations from Laboratory No. 963
are in appreciable error, the reported slopes of their calibration curves
are within the expected range. However, very high reagent blank values were
reported. High reagent blank values consistent with the positive biases
obtained for the sample analyses indicate that reagent contamination, and
possibly sample contamination, were out of control. Although not verified,
inaccurate calculations by the laboratory staff may also have contributed to
the magnitude of the erroneous data submitted to Battelle.
As shown in Table 3, the grand mean of mercury recovery for the spiked
water sample was 99 percent for all laboratories and 110 percent for the
Phase-Ill laboratories. The standard deviations of the percent mercury
recoveries were 53 percent and 57 percent, respectively. The grand means
for the spiked water sample correspond to nearly complete recoveries of the
target mercury concentration by using the revised Method 7470. The grand
mean of mercury recovery for the coal fly ash sample was 230 percent for all
laboratories as well as for just the Phase-Ill laboratories; the standard
deviations of the mercury recoveries were 200 percent and 210 percent,
respectively. The grand means for the coal fly ash sample correspond
approximately to a 2-fold positive bias by using the revised Method 7471;
moreover, the recovery data are considerably more dispersed than the
recovery data for the spiked water sample. This may be due to the lower
mercury concentration measured at the instrument for the fly ash sample, and
to weighing errors and other uncertainties when manipulating solid samples
in preparation for analysis. In general, the high dispersion of data
between laboratories for both samples indicates a lack of ruggedness for the
amalgamation CV-AAS method in this testing phase of the study.
21
-------
TABLE 2. SUMMARY OF PERCENT RECOVERIES OF TARGET MERCURY CONCENTRATIONS
IN SPIKED DEIONIZED WATER AND COAL FLY ASH
Laboratory
Identification
Number
Percent Recovery
Spiked Deionized
Water Coal Fly Ash
949 110
950 a 20 140
951 - --
952 a 80 88
953 a 80 88
200 250
955 a 120 440
956 61 25
957 a
958 91
959 24 260
960 43 420
961 140
952 a 120 140
963 2030 b 30000 b
964 a 200 690
965 a 93 94
966 a 96 110
a Laboratories selected for Phase III.
b Recoveries for Laboratory Number 963 were not used when calculating
means and standard deviations.
22
-------
TABLE 3. STATISTICAL SUMMARY OF MERCURY MEASUREMENTS
ON SPIKED DEIONIZED WATER AND COAL FLY ASH
Number
of
Laboratories
Percent Recoveries
Spiked
Deionized
Water
Coal
Fly
Ash
All Labsa
Mean
99
230
Standard
Deviation
53
200
Phase III Labs
110
230
Mean
Standard
Deviation
57
210
a Recoveries for Laboratory Number 963 were not used when calculating
means and standard deviations.
The data reported by some of the participating laboratories for these
preliminary sample analyses are in appreciable error. Possible contributing
factors to the wide range of results, especially for the coal fly ash
sample, include high absorbance values for the reagent blanks resulting from
sample digestions, a propagation of errors in the calculated data submitted
to Battelle, and the general lack of operator experience with the silver-
wool amalgamation technique. Some laboratories measured reagent blank
values as high as 0.6 absorbance unit, which is equivalent to approximately
120 ng of mercury based on a calibration slope of 0.005 absorbance unit/ng;
such a reagent blank value exceeds the linear range of the amalgamation CV-
AAS method. Battelle staff discussed the results with many of the
laboratories; recommendations for improvements in controlling mercury
contamination of the reagent blank and test sample, as well as suggestions
for implementation of the amalgamation CV-AAS system, were provided.
The results of the preliminary sample analyses in Phase II were used as a
qualitative guideline but not as absolute criteria for the selection of
laboratories for participation in Phase III. Five laboratories obtained
results which were in close agreement with the target mercury values for
both samples. Five other laboratories obtained less satisfactory results
but exhibited sufficient interest and enthusiasm in the silver-wool
amalgamation technique; these 5 laboratories were judged to be capable
participants in Phase III. The remaining 8 laboratories did not exhibit
23
-------
sufficient technical proficiency or enthusiastic interest to be invited to
participate in Phase III.
PHASE III - INTERLABORATORY EVALUATION OF THE SILVER-WOOL
AMALGAMATION CV-AAS METHOD
Ten of the 18 participating laboratories in Phase II were invited to
participate in the interlaboratory study in Phase III. These 10
laboratories, listed alphabetically in Table 4, are identified in the
subsequent tables from Phase III by numbers which are not correlated with
the alphabetical listing in Table 4.
TABLE 4. SUMMARY OF LABORATORIES PARTICIPATING IN PHASE III
Laboratory
Alliance Technologies Corporation (Bedford, MA)
Anatech Laboratories, Inc. (Santa Rosa, CA)
Conoco, Inc. (Ponca City, OK)
Environmental Monitoring & Services, Inc. (Camarillo, CA)
Harris Laboratories, Inc. (Lincoln, NE)
Northern Laboratories and Engineering, Inc. (Valparaiso, IN)
Pittsburgh Applied Research Corporation (Pittsburgh, PA)
Raba-Kistner Consultants, Inc. (San Antonio, TX)
Radian Corporation (Sacramento, CA)
University of Iowa Hygienic Laboratory (Iowa City, IA)
The Phase III study consisted of two options which differed in the
number of samples to be analyzed; Option A was comprised of 16 samples and
Option B was comprised of 24 samples. Five of the 10 laboratories were
selected for each option. The designations of the laboratories for the
specified options were based on their previous results in Phase II as well
as on budgetary constraints.
24
-------
Interlaboratory Study Samoles
Two different sample sets comprised of three solid and three aqueous
samples were shipped to the participants in Phase III. Some of the samples
required additions of a predigestion spike. The sample sets, denoted as
Option A and Option B, are listed in Appendices C and 0, respectively.
Options A and B differed principally in the number of replicate samples
designated for each sample type; Option A samples were supplied in duplicate
and Option B samples were supplied in triplicate. The samples were shipped
unspiked to avoid potential losses of trace mercury spikes added to samples
with a reducing matrix. Predigestion spiking solutions containing organic
and inorganic mercury were sent to the participants with spiking
instructions. One solid sample in Option B was also to be spiked with
copper, a potential interferent in CV-AAS analyses for mercury.
Instructions for Analysis and Reporting
In addition to the study samples and the appropriate spiking solutions,
the participating laboratories were provided a coded sample list and a set
of instructions for their designated option. The instructions included a
list comprised of the coded samples, the amount of sample to use, and the
kind and volume of predigestion spiking solution to add. The spiking
solutions were labeled to indicate for which sample they were to be used.
For the reporting of the analytical results, the participants were
instructed to: (1) report the slopes of the calibration curves for samples
analyzed by the revised Methods 7470 and 7471, both in units of absorbance
units/nanogram of mercury, (2) report all sample analysis data both in units
of absorbance and total nanograms of mercury, and (3) report the data for
the reagent blanks for the revised Methods 7470 and 7471 both in units of
absorbance and total nanograms of mercury.
Laboratory Results - Calibration Data
The slopes of the calibration curves and the mercury absorbance values
for the reagent blanks obtained by the 10 laboratories are summarized in
Table E-2 of Appendix E. The slopes of the calibration curves ranged from
0.0010 to 0.0065 absorbance units/nanogram of mercury. Six of the 10
laboratories reported slopes within the range experienced by Battelle staff
(0.003 and 0.006 absorbance units/nanogram). The absorbance values for the
reagent blanks ranged from -0.004 to 0.075 absorbance units. Eight of the
10 laboratories reported reagent blank values less than 0.03 absorbance
unit. Based on a typical calibration slope of 0.005 absorbance
unit/nanogram, a reagent blank value of 0.03 absorbance unit corresponds to
a maximum concentration of 6 ng Hg in the reagent blank.
In general, the 10 laboratories in Phase III reported a lower
dispersion and a narrower range of results for calibration and reagent blank
data than the 18 laboratories in Phase II. The 10 Phase-Ill laboratories
represent a more consistent range of personnel skills and demonstrated
competence in trace analysis. The lower dispersion in the data may also be
25
-------
a direct result of the acquired experience and familiarity with the
silver-wool amalgamation technique by the 10 laboratories from their
participation in the Phase II study.
LaboratorY Results - Sample Analyses
The results of the sample analyses by the 10 laboratories are compiled
by sample type for the revised Methods 7470 and 7471. The measured
concentrations for the individual replicates and statistical information
(mean and standard deviation) are provided for each laboratory and for each
sample type. In an attempt to characterize the interlaboratory accuracy and
precision of the amalgamation CV-AAS methods, a statistical summary of the
mercury concentration data for each analysis option and for each sample type
has also been compiled. The statistical summaries of the interlaboratory
data for the revised Methods 7470 and 7471 are presented in Tables 5 and 6,
respectively. In all statistical analyses, the means and standard
deviations are reported in concentration units for those environmental
samples originally having nondetectable endogenous mercury concentrations,
and in percent recoveries for those samples having detectable target mercury
concentrations. The percent recoveries are calculated as 100 times the
measured concentration divided by the target concentration.
Outlier tests were performed on the laboratory means for each optio
within each sample type by using the maximum normed residual test.
Laboratory No. 964 was identified as an outlier for the following sample
types: marine sediment MESS-i, incinerator fly ash, ground water, waste
water spiked with inorganic mercury, and dilute nitric acid spiked with
inorganic mercury. In addition, Laboratory No. 966 was found to be an
outlier for ground water, as was Laboratory No. 952 for waste water spiked
with inorganic mercury. Because the concentration data reported by
Laboratory No. 964 in Option A are statistical outliers in the majority of
the sample types, these data have been uniformly excluded from statistical
treatment for all sample types.
An explanation of the statistical terms in Tables 5 and 6 is provided
below. The average mercury concentration or percent recovery over all
laboratories in a given option is reported in the third column. The fourth
column contains the pooled within-laboratory standard deviations of the
mercury concentrations or percent recoveries. The within-laboratory
variances are simply calculated as the averages of the variances for the
individual laboratories. The between-laboratory standard deviations of the
mercury concentrations or percent recoveries are reported in the fifth
column. This statistic characterizes the variabilities in the true average
mercury concentration or percent recovery values across the laboratories.
The between-laboratory variances are calculated according to the following
formula:
S = (S 2 - S )
where is the variance of the average percent recoveries for an individual
laborat ry, n is the number of sample results used to calculate the average
26
-------
TABLE 5. STATISTICAL SUMMARY OF MERCURY MEASUREMENTSa
USING REVISED METHOD 7470
Sample
Pooled
Option Mean Std. Dev.
Between
Std. 0ev.
Total
Std. 0ev.
Samples Having
Nondetectable Target Concentrationsb
1.4
2.3
Ground Water
A 4 1.8
Ground WaterC
B 6 1.5
2.1
2.6
Samples Having
Detectable Target Concentrationsd
45
48
Waste Water,
Spiked
A 140 17
Waste Water,e
Spiked
B 150 15
39
42
Dilute Nitric
Acid, Spiked
A 82 27
27
38
Dilute Nitric
Acid, Spiked
B 82 31
17
36
a Results for Laboratory Number 964 in Option A were not used in the
statistical analyses for any sample.
b Results are in absolute concentration units (ng).
C Results for Laboratory Number 966 were not used in the statistical
analyses for this sample.
d Results are percent recovery.
e Results for Laboratory Number 952 were not used in the statistical
analyses for this sample.
27
-------
TABLE 6. STATISTICAL SUMMARY OF MERCURY MEASUREMENTSa
USING REVISED METHOD 7471
Poo
Sample Option Mean Std.
led
Dev.
Between
Std. Dev.
Total
Std. 0ev.
Samples Having Nondetectable Target Concentrationsb
Incinerator
Fly Ash A 0.021 0.0090 0.0076 0.012
Incinerator
Fly Ash B 0.074 0.021 0.041 0.046
Sam 1es Having Detectable Target Concentrationsc
Marine Sediment
MESS-i A 110 6.5 10 12
Marine Sediment
MESS-i B 110 48 0 48
Incinerator Fly
Ash, Spiked A 74 8.6 32 34
Incinerator Fly
Ash, Spiked B 39 15 11 19
Municipal Sewage
Sludge A 90 14 18 22
Municipal Sewage
Sludge B 88 20 21 30
Municipal Sewage
Sludge, Spiked A 97 26 25 36
Municipal Sewage
Sludge, Spiked B 90 12 26 28
a Results for Laboratory Number 964 in Option A were not used in the
statistical analyses for any sample.
b Results are in relative concentration units (mg/kg).
C Results are in percent recovery.
28
-------
percent recovery 2 for each labor tory, an S 2 is the within-laboratory
variance. If S-. is less than S , then SB i’ set to zero. The total
standard deviatidSis are reported i the last column. The total variance is
simply the sum of the within-laboratory and between-laboratory variances.
Revised Method 7470--
Ground Water--The results for the measured concentrations reported by
the individual laboratories are summarized in Table E-3 of Appendix E. In
addition to the data from Laboratory No. 964 in Option A, the data from
Laboratory No. 966 in Option B have been excluded as outliers for the
statistical analyses. The grand means for both options in Table 5 are less
than 10 ng Hg which is expected for a water sample originally having a
nondetectable target mercury concentration. Absolute mercury concentrations
less than 10 ng are definitively imprecise and are very close to the
baseline noise level of the amalgamation CV-AAS method. Even taking into
account the spreads indicated by the total standard deviations for the two
options, the grand means did not exceed 10 ng Hg.
Spiked Waste Water--This sample was to have been spiked with inorganic
mercury to provide target mercury concentrations of 50 ng for Option A and
25 ng for Option B. The results for the measured mercury concentrations
reported by the individual laboratories are summarized in Table E-4 of
Appendix E. In addition to the data from Laboratory No. 964 in Option A,
the data from Laboratory No. 952 in Option B have been excluded as outliers
for the statistical analyses. The mean percent recoveries for the 8
laboratories ranged from 88 to 220 percent in Option A and from 89 to 240
percent in Option B. The standard deviations ranged from 6 to 21 percent in
Option A and from 0 to 28 percent in Option B. As shown in Table 5, the
grand mean recoveries for both analysis options are not within 15 percent of
full recovery of the spiked mercury concentrations. These results represent
a potential positive bias of 40 to 50 percent for this sample.
Spiked Dilute Nitric Acid--This sample was to have been spiked with
inorganic mercury to provide a target mercury concentration of 25 ng for
Options A and B. The results for the measured mercury concentrations
reported by the individual laboratories are summarized in Table E-5 of
Appendix E. The mean percent recoveries ranged from 18 to 130 percent in
Option A and from 9 to 120 percent in Option B. The standard deviations
ranged from 0 to 40 percent in Option A and from 2 to 47 percent in Option
B. As shown in Table 5, the grand mean recoveries for both analysis options
are not within 15 percent of the spiked mercury concentration by only narrow
margins. These results represent a potential negative bias for this sample.
Revised Method 7471--
Marine Sediment MESS-i--The results reported by the individual
laboratories for the measured mercury concentrations in the marine reference
sediment are summarized in Table E-6 of Appendix E. The mean percent
recoveries ranged from 100 to 140 percent for Option A and from 66 to 150
percent for Option B. The standard deviations ranged from 4 to 8 percent
for Option A and from 3 to 97 percent for Option B. As shown in Table 6,
the grand mean recoveries for both analysis options in Table 6 are within 15
percent of full recovery of the target mercury concentration.
29
-------
Incinerator Fly Ash--The results reported by the individual
laboratories for the measured mercury concentrations in the incinerator fly
ash are summarized in Table E-7 of Appendix E. The mean concentrations
ranged from <0.005 to 0.038 mg/kg for Option A and from 0.019 to 0.20 mg/kg
for Option B. Most laboratories reported measurable mercury concentrations
although the sample contained no detectable mercury when it was packaged and
shipped to the laboratories. These data suggest that the sample became
contaminated, measurement inaccuracies were introduced from uncompensated
reagent blank absorbances, or some other unidentified measurement bias
existed for the laboratory’s instrument configuration. Based on a
quantification limit of 10 ng and on a sample weight of 0.2 g, mercury
concentrations less than 0.05 mg/kg were expected for such a solid sample
not having a detectable mercury concentration. As shown in Table 6, the
grand mean for Option A is less than 0.05 mg/kg. However, the grand mean
for Option B is greater than 0.05 mg/kg which indicates a possible positive
bias.
Spiked Incinerator Fly Ash--This sample was to have been spiked with
inorganic mercury for both options and with copper for Option B. The
results reported by the individual laboratories for the measured mercury
concentrations in the spiked incinerator fly ash are summarized in Table E-8
of Appendix E. The mean percent recoveries ranged from 8 to 120 percent for
Option A and from 13 to 72 percent for Option B. With the exception of
Laboratory No. 957, the percent recoveries for the laboratories in both
options were low by as much as ten fold. With the exception of Laboratory
No. 966, the standard deviations for within-laboratory analyses in both
options were less than 11 percent. As shown in Table 6, the grand mean
recovery for Option A analyses in Table 6 is within approximately 25 percent
of full recovery of the spiked mercury concentration. However, the grand
mean recovery for Option B analyses is low by more than two fold.
The low recoveries of spiked inorganic mercury in Option B suggest that
1 mg copper, in general, interferes with the amalgamation CV-AAS analyses.
The results of this interlaboratory study contradict the results of the
single-laboratory study which revealed that copper concentrations as high as
1 mg per O.2-g test portion of NBS-SRM 1646 (Estuarine Sediment) d id not
interfere with the recovery of the certified mercury concentration. 2 ’ Such
inconsistency between the single-laboratory and the interlaboratory results
for the copper interference test indicates a relative lack of ruggedness of
the revised Method 7471.
Municipal Sewage Sludge--The results reported by the individual
laboratories for the measured mercury concentrations in the municipal sewage
sludge are summarized in Table E-9 of Appendix E. The mean percent
recoveries ranged from 61 to 130 percent for Option A and from 13 to 110
percent for Option B. With the exception of Laboratory No. 953, the mean
concentrations for the laboratories in Option B were within 10 percent of
the target mercury concentration. The standard deviations ranged from 0.5
to 27 percent in Option A and from 5 to 32 percent in Option B. As shown in
Table 6, the grand mean recoveries for both Option A and Option B are within
15 percent of full recovery of the target mercury concentration.
30
-------
Spiked Municipal Sewage Sludge--This sample was to have been spiked
with mercury as methyl mercuric chloride to investigate the recoveries of
organic mercury. The results reported by the individual laboratories for
the measured mercury concentrations in the spiked municipal sewage sludge
are summarized in Table E-1O of Appendix E. The mean percent recoveries
ranged from 54 to 160 percent in Option A and from 12 to 130 percent in
Option B. The standard deviations ranged from 2 to 48 percent in Option A
and from 3 to 25 percent in Option B. As shown in Table 6, the grand mean
recoveries for both analysis options are within 15 percent of full recovery
of the target mercury concentrations.
Summary of the InterlaboratorY Results
The high dispersion of the interlaboratory data precludes any
statistical treatment of the results for the revised Methods 7470 and 7471
that is more rigorous than the analyses conducted and summarized in Tables 5
and 6, respectively. As shown in Table 5 for the revised Method 7470, the
total standard deviations for Options A and B are similar for each of the 3
test samples. However, as shown in Table 6 for the revised Method 7471,
neither option demonstrates a consistently lower total standard deviation
for all samples. For example, the total standard deviations of the results
in Option A are 4-fold lower than those in Option B for the incinerator fly
ash and the marine reference sediment samples. However, the total standard
deviation in Option B is approximately 2-fold lower than that in Option A
for the spiked incinerator fly ash sample. The total standard deviations of
the results in both options are similar for the unspiked and the spiked
municipal sewage sludge sample. The relative differences between the total
standard deviations in Options A and B by the revised Method 7471 for the 5
samples appear to be random.
With the exception of the spiked waste-water sample in Table 5 and of
the spiked incinerator fly-ash sample in Table 6, the grand means using the
revised Methods 7470 and 7471, respectively, are generally within 25 percent
of the target values. However, the total standard deviations for all sample
types represent relative standard deviations as high as 50 percent. Because
of such highly dispersed results, in addition to the contradictory results
of the copper interference test, the revised Methods 7470 and 7471 presently
lack the necessary ruggedness to be successfully used for routine testing by
analytical laboratories without additional analyst training and
demonstration of proficiency.
31
-------
REFERENCES
1. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-
846, 2nd Edition, July 1982, U. S. Environmental Protection Agency,
Office of Solid Waste and Emergency Response, Washington, D.C.
2. Messnian, J. 0., M. E. Churchwell, R. L. Livingston, and D. L. Sgontz.
1986. Evaluation and Testing of EPA Protocols for SW-846 Methods 7470
and 7471. Final Report. U.S. Environmental Protection Agency.
3. Churchwell, M. E., R. L. Livingston, D. L. Sgontz, J. 0. Messman, and W.
F. Beckert. 1988 (in press). Environment International.
4. deShon, H. Private communication. 1987.
5. Youden, W. J., and E. H. Steiner. Statistical Manual of the AOAC.
Association of Official Analytical Chemists, Arlington, Virginia, 1975.
88 pp.
6. Snedecor, G. W., and W. G. Cochran. Statistical Methods, 7th Edition.
The Iowa State University Press, Ames, Iowa, 1980. pp. 252-253.
32
-------
APPENDIX A
METHOD 7470 (Proposed Revision)
MERCURY (MANUAL AMALGAMATION COLD-VAPOR TECHNIQUE )
1.0 SCOPE AND APPLICATION
1.1 Method 7470 is a cold-vapor atomic absorption procedure
recommended for determining low concentrations of mercury in mobility
procedure extracts, aqueous wastes, and groundwaters. All samples must be
subjected to an appropriate digestion step prior to analysis.
2.0 SUMMARY OF METHOD
2.1 Prior to analysis, the samples must be prepared according to the
procedure discussed in this method.
2.2 This method, a cold-vapor atomic absorption technique, is based on
the absorption of radiation at 253.7 nm by mercury vapor. The mercury is
reduced to the elemental state, aerated from solution by a nitrogen purge,
and concentrated by amalgamation on silver wool. The amalgamated mercury is
thermally volatilized from the silver wool by resistance heating of a
nichrome wire wrapped around Pyrex tubing containing the silver wool and is
then carried by the nitrogen purge gas into the glass absorption cell for
the atomic absorption measurement. The analytical calibration curve is
based on the maximum absorbance (peak height) determined for each
measurement.
2.3 The typical instrument detection limit for this method is
approximately 1 ng (0.02 pg/L for a 50-mL sample aliquot).
3.0 INTERFERENCES
3.1 Potassium permanganate is added to eliminate possible interference
from sulfide. In a single-laboratory study, concentrations as high as 2 mg
of sulfide as sodium sulfide per sample aliquot did not interfere with the
recovery of inorganic mercury from a simulated aqueous waste.
3.2 Copper has been reported to interfere; however, in a single-
laboratory study, copper concentrations as high as 1 mg per sample aliquot
had no significant effect on the recovery of inorganic mercury from a
simulated aqueous waste.
3.3 High levels of chloride in seawaters, brines, and industrial
effluents have been reported to interfere; however, in a single-laboratory
study, concentrations as high as 3.4 g sodium chloride per sample aliquot
did not interfere with the recovery of inorganic mercury from a simulated
aqueous waste. Care must be taken to ensure that free chlorine is absent
before the mercury is reduced and swept into the absorption cell. This may
be accomplished by using an excess of hydroxylaniine sulfate or hydroxylamine
hydrochloride reagent (25 mL). In addition, the dead air space in the
33
-------
reduction-aeration sample cell must be purged before adding stannous sulfate
or stannous chloride. Both inorganic and organic mercury spikes have been
quantitatively recovered from seawater by using this digestion method in
previous studies.
3.4 Certain volatile organic materials may absorb at the 253.7 nm
wavelength; however, no nonspecific absorption interferences were observed
for a simulated aqueous waste spiked with benzene and methyl ethyl ketone.
4.0 APPARATUS AND MATERIALS
4.1 Atomic absorption spectrophotometer or equivalent: Any atomic
absorption unit having an open sample atomization area in which to mount
the absorption cell is suitable. Instrument settings recommended by the
particular manufacturer should be followed. Instruments designed
specifically for the spectrophotometric measurement of mercury by using the
cold-vapor technique are commercially available and may be substituted for
the atomic absorption spectrophotometer.
4.2 Mercury hollow cathode lamp or electrodeless discharge lamp.
4.3 Recorder: Any multirange, variable-speed recorder with a one-
second time constant that is compatible with the photometric detection
system is suitable.
4.4 Absorption cell: A cylindrical glass cell having entrance and exit
ports on the side and quartz end-windows to transmit ultraviolet radiation
is suitable. A cylindrical cell, 1 inch in diameter (o.d.) and 6 inches in
length, has been found to be an appropriate geometry. The absorption cell
is strapped to a burner head assembly for convenient positioning and for
two-dimensional alignment in the optical path’ of the sample atomization area
of the atomic absorption spectrophotometer. The alignment is facilitated by
the use of two 2-inch x 2-inch white cards with 1-inch diameter holes cut in
the center of each card. The cards are placed over each end of the
absorption cell, and the cell is then aligned to provide maximum
transmittance of the primary radiation from the mercury spectral lamp.
4.5 Compressed nitrogen gas of commercial grade or better.
4.6 Needle valve for primary control of nitrogen gas flow.
4.7 Flowmeter: Any calibrated flowneter capable of measuring a
nitrogen flow of 0.5 L/min is suitable.
4.8 Charcoal filter trap: A cylindrical tube, 1 inch in diameter
(o.d.) and 6 inches in length, containing activated charcoal for adsorbing
mercury and organics in the nitrogen gas stream is suitable.
4.9 Reduction-aeration sample cell: A 125-mL gas sparging bottle and a
straight glass bubbling tube with an end-frit of a coarse porosity.
34
-------
4.10 Variac autotransformer,
or equivalent,
capable of 20 volts
output.
4.11 Silver wool, micro.
4.12 Chromel-A wire, 22-gauge, 3 feet.
4.13 Aeration tubing: Glass tubing is used for the passage of the
mercury vapor from the gas sparging sample bottle jo the silver-wool
collection tube and then to the absorption cell. Tygon T ’ tubing is used for
glass-to-glass connections as necessary.
4.15 Because mercury vapor j toxic, precaution must be taken to avoid
its inhalation. Therefore, Tygon’ tubing connected to the exit port on the
side of the absorption cell is used to vent mercury vapor into an exhaust
hood.
5.0 REAGENTS
5.1
impurities.
ASTM Type II water (ASTM 01193): The water should be monitored for
5.2 Sulfuric acid (concentrated), H 2 S0 4 , reagent grade.
5.3 Sulfuric acid (0.5N), H 2 S0 4 : Prepare by carefully diluting 14.0
mL concentrated sulfuric acid to 1.0 L.
(concentrated), HNO 3 , reagent grade of low mercury
reagent blank is obtained, it may be necessary to
acid (concentrated), HC1, reagent grade of
5.6 Stannous sulfate, SnSO 4 , reagent grade.
5.7 Stannous sulfate solution: Add 100 g stannous sulfate to 1 L of
0.5N sulfuric acid. This mixture is a suspension and should be stirred
continuously during use. Stannous chloride may be used in place of stannous
sulfate.
5.8 Stannous chloride, SnCl 2 2H 2 0, reagent grade.
5.9 Stannous chloride solution:
of concentrated hydrochloric acid.
dissolved, dilute to 1 L with Type II
Add 150 g stannous chloride to 300 mL
After the stannous chloride has
water.
4.14 The cold-vapor/amalgamation apparatus is assembled as shown via
the schematic diagram in Figure 1A and the block diagram in Figure lB. The
cold-vapor/amalgamation system shown in Figure 1 is an open system where
mercury vapor is passed through the absorption cell only once.
5.4 Nitric acid
content. If a high
distill the nitric acid.
5.5 Hydrochloric
mercury content.
low
35
-------
5.10 Sodium chloride, NaC1, reagent grade.
5.11 Hydroxylamine sulfate, (NH 2 OH) 2 2H 2 S0 4 , reagent grade.
5.12 Sodium chioride-hydroxylamine sulfate solution: Dissolve 120 g
sodium chloride and 120 g hydroxylamine sulfate in Type II water and dilute
to 1 L. Hydroxylamine hydrochloride may be used in place of hydroxylamine
sul fate.
5.13 Hydroxylamine hydrochloride, NH 2 OH HCl, reagent grade.
5.14 Sodium chioride-hydroxylamine hydrochloride solution: Dissolve
120 g sodium chloride and 120 g hydroxylamine hydrochloride in Type II water
and dilute to 1 L.
5.15 Potassium permanganate, KMnO 4 , reagent grade.
5.16 Potassium permanganate, 5-percent (w/v) solution: Dissolve 50 g
potassium permanganate in 1 L Type II water.
5.17 Potassium persulfate, K 2 S 2 0 8 , reagent grade.
5.18 Potassium persulfate, 5-percent (w/v) solution: Dissolve 50 g
potassium persulfate in 1 L Type II water.
5.19 Mercuric chloride, HgCl 2 , reagent grade.
5.20 Mercury stock solution: Dissolve 0.1354 g mercuric chloride in 75
mL Type II water. Add 10 mL concentrated nitric acid and adjust the volume
to 100.0 mL with Type II water (1 mL = 1 mg Hg). (A commercial 1000-mg/L
atomic absorption standard may also be used as the mercury stock solution.)
5.21 Mercury working standards: Make successive dilutions of the
mercury stock solution to obtain a working standard containing 0.01 pg/mL.
This working standard and the dilutions of the mercury stock solution should
be prepared fresh daily. Acidity of the working standards should be
maintained at 0.15 percent nitric acid. This acid should be added to the
flask before addition of the aliquot.
6.0 SAMPLE COLLECTION, PRESERVATION, AND HANDLING
6.1 All samples must have been collected by making use of a sampling
plan that addresses the considerations discussed in Section One of this
manual.
6.2 All sample containers must be prewashed with detergents, acids,
and Type II water. Plastic and glass containers are both suitable.
6.3 Aqueous samples must be acidified to a pH of less than 2 with
nitric acid. The suggested maximum holding times for these samples are 38
days in glass containers and 13 days in plastic containers.
36
-------
7.0 PROCEDURE
7.1 Sample preparation (steam-bath digestion):
7.1.1 All glassware for sample digestions, sample dilutions, and
standard preparations must be prewashed sequentially with an aqueous
detergent solution, mineral acids, and Type II water.
7.1.2 Transfer a 50-mL sample aliquot to a tared, 125-mL
Erlennieyer flask. Add 5 mL concentrated sulfuric acid and 2.5 mL
concentrated nitric acid. Thoroughly mix by hand-swirling after each
addition. Add 15 mL of potassium permanganate solution. Thoroughly mix by
hand-swirling. Sewage samples may require additional permanganate. While
mixing by hand-swirling, add additional portions of potassium permanganate
solution, if necessary, until the purple color persists for at least 15
minutes. Add 8 mL potassium persulfate solution and mix by hand-swirling.
Cover the top of the flask with a watch glass. Heat for 2 hours on a steam
bath maintained at a temperature of approximately 95°C. After cooling and
immediately prior to analysis, add 6 mL sodium chioride-hydroxylamine
hydrochloride solution or sodium chloride-hydroxylamine sulfate solution to
reduce excess permanganate in solution and the manganese dioxide particles
adhering to the flask wall. Quantitatively transfer the sample solution to
a 100-mL volumetric flask and dilute to calibrated volume with Type IL
water. Mix thoroughly and agitate by hand-swirling.
7.2 Standard preparation: Transfer appropriate aliquots of the
mercury working standards to a series of 125-mL Erlenmeyer flasks. The
aliquots must be selected to provide suitable calibration standards within
the linear dynamic range of the instrument. Add Type II water to each flask
to adjust the total volume to 50 mL and thoroughly mix by hand-swirling.
Add 5 mL co’ncentrated sulfuric acid and 2.5 mL concentrated nitric acid to
each flask, mixing by hand-swirling after each addition. Add 15 mL
potassium permanganate solution to each flask and allow to stand at least 15
minutes. Add 8 mL potassium persulfate solution to each flask and heat for
2 hours on the steam bath maintained at a temperature of approximately 95°C.
After cooling and immediately prior to analysis, add 6 mL sodium chloride-
hydroxylamine hydrochloride solution or sodium chioride-hydroxylamine
sulfate solution to reduce excess pernianganate in solution and the manganese
dioxide particles adhering to the flask wall. Quantitatively transfer the
standard solutions to 100-mL volumetric flasks and dilute to calibrated
volume with Type II water. Mix thoroughly and agitate by hand-swirling.
7.3 Analysis: Quantitatively transfer an appropriate aliquot of
sample, standard, or reagent blank solution from the 100-mL volumetric flask
to the 125-mL sparging bottle. Add Type II water to adjust the total volume
to 100 niL. Add 5 mL stannous sulfate or stannous chloride solution. Ensure
that the nitrogen purge gas flow is set between 500 and 600 mL/min and that
the purge gas flow is directed through the bypass valve instead of through
the aeration bubbler tube of the cold-vapor/amalgamation apparatus. Also,
ensure that the exit end of the silver-wool collection tube is disconnected
from the absorption cell. Connect the sparging bottle to the cold-vapor
apparatus, and redirect the nitrogen purge gas flow through the aeration
37
-------
apparatus, and redirect the nitrogen purge gas flow through the aeration
bubbler tube for exactly 2 minutes. After the 2-minute concentration time,
redirect the nitrogen purge gas flow through the bypass valve, and connect
the exit end of the silver-wool collection tube to the absorption cell.
After the nitrogen purge gas has been directed through the bypass valve for
1 minute, turn on the Variac autotransformer and strip-chart recorder.
Initiate an atomic absorption measurement (maximum peak height) of 60
seconds duration; this measurement time may be pre-programmed with some
spectrophotometers to provide a digital display of the maximum absorbance
during this time interval. The mercury atomic absorption peak maximum
should appear on the strip-chart within 45 seconds. When the atomic
absorption peak has receded to 0.010 absorbance unit, turn off the Variac
autotransformer and strip-chart recorder, and cool the outside of the
silver-wool collection tube by a stream of air. When the mercury value
exceeds the linear range of the calibration curve, the analysis must be
repeated with smaller sample aliquots.
7.4 Construct a calibration curve by plotting the absorbance of a
series of mercury standards versus micrograms of mercury. Determine the
absorbance peak height of the unknown sample from the strip-chart or from
the digital display of the atomic absorption spectrophotometer. Determine
the mercury concentration value from the standard curve.
7.5 Analyze, by the method of standard additions, all EP extracts, all
samples analyzed as part of a delisting petition, and all samples that
suffer from matrix interferences.
7.6 Duplicates, spiked samples, and check standards should be
routinely analyzed.
7.7 Calculate mercury concentrations by (1) the method of standard
additions or (2) from a calibration curve or (3) directly from the
calibrated concentration readout display of the spectrophotometer. All
dilution or concentration factors must be taken into account.
8.0 QUALITY CONTROL
8.1 All quality control data should be maintained and available for
easy reference or inspection.
8.2 Calibration curves must be composed of a minimum of a reagent
blank and three standards.
8.3 Dilute samples if their mercury concentrations are higher than
that of the highest standard or if their mercury concentrations fall on the
plateau of the calibration curve.
8.4 Analyze a minimum of one reagent blank per sample batch to
determine if contamination or memory effects are occurring.
8.5 Analyze check standards after approximately every 6 samples.
38
-------
8.6 Analyze one duplicate sample for every 10 samples.
8.7 Spiked samples or standard reference materials shall be
periodically analyzed to ensure that correct procedures are being followed
and that all equipment is operating properly.
8.8 The method of standard additions shall be used for the analyses of
all EP extracts, for all analyses submitted as part of a delisting petition,
and whenever a new sample matrix is being analyzed.
9.0 METHOD PERFORMANCE
9.1 Because of the lack of ruggedness demonstrated in an
interlaboratory study, the accuracy and precision of the method has not been
statistically determined.
10.0 REFERENCES
<>Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-
846, 2nd Edition, July 1982, U. S. Environmental Protection Agency, Office
of Solid Waste and Emergency Response, Washington, D.C.
<>Messman, J. 0., M. E. Churchwell, R. L. Livingston, and 0. L. Sgontz.
1986. Evaluation and Testing of EPA Protocols for SW-846 Methods 7470 and
7471. Final Report. U.S. Environmental Protection Agency.
<>Churchwell, M. E., R. L. Livingston, D. L. Sgontz, J. D. Messman, and
W. F. Beckert. 1988 (in press). Environment International.
<>Youden, W. J., and E. H. Steiner. Statistical Manual of the AOAC.
Association of Official Analytical Chemists, Arlington, Virginia, 1975. 88
pp.
<>Gebhart, J. E., J. D. Messman, and G. F. Wallace. 1987.
Interlaboratory Evaluation of SW-846 Methods 7470 and 7471 for the
Determination of Mercury in Environmental Samples. Final Report. U.S.
Environmental Protection Agency.
39
-------
Charcoal
Absorption
Cell
Silver Wool
Chrom-Alumel
Resistance Heating
Winding
Nitrogen.
Cylinder
Purging Cylinder
(Reduction-Aeration
Sample Cell)
Figure 1 A. Schematic diagram of the amalgamation CV-AAS system.
Figure 1 B Block diagram of the amalgamation CV-AAS system
40
-------
APPENDIX B
METHOD 7471 (Proposed Revision)
MERCURY (MANUAL AMALGAMATION COLD-VAPOR TECHNIOUE )
1.0 SCOPE AND APPLICATION
1.1 Method 7471 is a cold-vapor atomic absorption procedure
recommended for measuring low concentrations of total mercury (organic and
inorganic) in soils, sediments, bottom deposits, and sludge-type materials.
All samples must be subjected to an appropriate digestion step prior to
analysis.
2.0 SUMMARY OF METHOD
2.1 Prior to analysis, the samples must be prepared according to one
of the procedures discussed in this method.
2.2 This method, a cold-vapor atomic absorption technique, is based on
the absorption of radiation at 253.7 nm by mercury vapor. The mercury is
reduced to the elemental state, aerated from solution by a nitrogen purge,
and concentrated by amalgamation on silver wool. The amalgamated mercury is
thermally volatilized from the silver wool by resistance heating of a
nichrome wire wrapped around Pyrex tubing containing the silver wool and is
then carried by the nitrogen purge gas into the glass absorption cell for
the atomic absorption measurement. The analytical calibration curve is
based on the maximum absorbance (peak height) determined for each
measurement.
2.3 The typical instrument detection limit for this method is
approximately 1 ng (5 ng/g based on a 0.2-g sample weight).
3.0 INTERFERENCES
3.1 Potassium permanganate is added to eliminate possible interference
from sulfide. In a single-laboratory study, concentrations as high as 2 mg
of sulfide as sodium sulfide per O.2-g sample did not interfere with the
recovery of mercury from NBS-SRM 1646 (Estuarine sediment).
3.2 Copper has been reported to interfere. In a single-laboratory
study, copper concentrations as high as 1 mg per 0.2-g sample generally had
no effect on the recovery of mercury from NBS-SRM 1646 (Estuarine sediment).
However, in an interlaboratory study, low recoveries of spiked inorganic
mercury in a composite incinerator fly ash sample containing 1 mg of copper
were obtained.
3.3 High levels of chloride in seawaters, brines, and industrial
effluents have been reported to interfere; however, in a single-laboratory
study, concentrations as high as 30 mg sodium chloride per O.2-g sample had
no effect on recovery of mercury from NBS-SRM 1646 (Estuarine sediment).
Care must be taken to ensure that free chlorine is absent before the mercury
41
-------
is reduced and swept into the absorption cell. This may be accomplished by
using an excess of hydroxylamine sulfate or hydroxylamine hydrochloride
reagent (25 mL). In addition, the dead air space in the reduction-aeration
sample cell must be purged before adding stannous sulfate or stannous
chloride. Both inorganic and organic mercury spikes have been
quantitatively recovered from seawater by using this digestion method in
previous studies.
3.4 Certain volatile organic materials may absorb at the 253.7 nm
wavelength; however, no nonspecific absorption interferences were observed
for NBS-SRM 1645 (River Sediment) spiked with benzene and methyl ethyl
ketone.
4.0 APPARATUS AND MATERIALS
4.1 Atomic absorption spectrophotonieter or equivalent: Any atomic
absorption unit having an open sample atomization area in which to mount the
absorption cell is suitable. Instrument settings recommended by the
particular manufacturer should be followed. Instruments designed
specifically for the spectrophotonietric measurement of mercury using the
cold-vapor technique are commercially available and may be substituted for
the atomic absorption spectrophotometer.
4.2 Mercury hollow cathode lamp or electrodeless discharge lamp.
4.3 Recorder: Any multirange, variable-speed recorder with a one-
second time constant that is compatible with the photometric detection
system is suitable.
4.4 Absorption cell: A cylindrical glass cell having entrance and exit
ports on the side and quartz end-windows to transmit ultraviolet radiation
is suitable. A cylindrical cell, 1 inch in diameter (o.d.) and 6 inches in
length, has been found to be an appropriate geometry. The absorption cell
is strapped to a burner head assembly for convenient positioning and for
two-dimensional alignment in the optical path of the sample atomization area
of the atomic absorption spectrophotometer. The alignment is facilitated by
the use of two 2-inch x 2-inch white cards with 1-inch diameter holes cut in
the center of each card. The cards are placed over each end of the
absorption cell, and the cell is then aligned to provide maximum
transmittance of the primary radiation from the mercury spectral lamp.
4.5 Compressed nitrogen gas of commercial grade or better.
4.6 Needle valve for primary control of nitrogen gas flow.
4.7 Flowmeter: Any calibrated flowmeter capable of measuring a
nitrogen flow of 0.5 L/min is suitable.
4.8 Charcoal filter trap: A cylindrical tube, 1 inch in diameter
(o.d.) and 6 inches in length, containing activated charcoal for adsorbing
mercury and organics in the nitrogen gas stream is suitable.
42
-------
4.9 Reduction-aeration sample cell: A 125-mL gas sparging bottle and a
straight glass bubbling tube with an end-frit of a coarse porosity.
4.10 Variac autotransformer, or equivalent, capable of 20 volts
output.
4.11 Silver wool, micro.
4.12 Chromel-A wire, 22-gauge, 3 feet.
4.13 Aeration tubing: Glass tubing is used for the passage of the
mercury vapor from the gas sparging sample bottle jo the silver-wool
collection tube and then to the absorption cell. TygonT tubing is used for
glass-to-glass connections as necessary.
4.14 The cold-vapor/amalgamation apparatus is assembled as shown via
the schematic diagram in Figure 1A and the block diagram in Figure lB. The
cold-vapor/amalgamation system shown in Figure 1 is an open system where
mercury vapor is passed through the absorption cell only once.
4.15 Because mercury vapor j toxic, precaution must be taken to avoid
its inhalation. Therefore, Tygon’ tubing connected to the exit port on the
side of the absorption cell is used to vent mercury vapor into an exhaust
hood.
5.0- REAGENTS
5.1 ASTM Type II water (ASTM 01193): The water should be monitored for
impurities.
5.2 Nitric acid (concentrated), HNO 3 , reagent grade.
5.3 Hydrochloric acid (concentrated), HC1, reagent grade.
5.4 Aqua regia: Prepare immediately before use by carefully adding
three volumes of concentrated hydrochloric acid to one volume of
concentrated nitric acid.
5.5 Sulfuric acid (concentrated), H 2 S0 4 , reagent grade.
5.6 Sulfuric acid (O.5N), H S0 4 : Prepare by carefully diluting 14.0 mL
concentrated sulfuric acid to 1 liter.
5.7 Stannous sulfate, SnSO 4 , reagent grade.
5.8 Stannous sulfate solution: Add 25 g stannous sulfate to 250 mL of
0.5N sulfuric acid. This mixture is a suspension and should be stirred
continuously during use. Stannous chloride may be used in place of stannous
sulfate.
5.9 Stannous chloride, SnC1 2 2H 2 0, reagent grade.
43
-------
5.10 Stannous chloride solution:
of concentrated hydrochloric acid.
dissolved, dilute to 1 L with Type II water.
5.11 Sodium chloride, NaC1, reagent grade.
5.12 Hydroxylamine sulfate, (NH 2 0H) 2 2H 2 SO 4 , reagent grade.
5.13 Sodium chioride-hydroxylamine sulfate solution: Dissolve 120 g
sodium chloride and 120 g hydroxylamine sulfate in Type II water and dilute
to 1 L. Hydroxylamine hydrochloride may be used in place of hydroxylamine
sulfate.
5.14 Hydroxylamine hydrochloride, NH 2 OH HCl, reagent grade.
5.15 Sodium chioride-hydroxylamine hydrochloride solution: Dissolve
120 g sodium chloride and 120 g hydroxylamine hydrochloride in Type II water
and dilute to 1 L.
5.16 Potassium permanganate, KMnO 4 , reagent grade.
5.17 Potassium permanganate, 5-percent solution (w/v): Dissolve 50 g
potassium permanganate in 1 L Type II water.
5.18 Mercuric chloride, HgC1 2 , reagent grade.
5.19 Mercury stock solution: Dissolve 0.1354
in 75 mL Type II water. Add 10 mL concentrated nitric
volume to 100.0 mL with Type II water (1 mL = 1 mg Hg).
mg/L atomic absorption standard may also be used as
solution.)
5.20 Mercury working standards: Make successive dilutions of the stock
mercury solution to obtain a working standard containing 0.01 pg/mL. This
working standard and the dilutions of the mercury stock solution should be
prepared fresh daily. Acidity of the working standards should be maintained
at 0.15 percent nitric acid. This acid should be added to the flask before
adding the aliquot.
6.0 SAMPLE COLLECTION, PRESERVATION, AND HANDLING
6.1 All samples must have been collected by
that addresses the considerations discussed
6.2 All sample containers must be prewashed with detergents, acids,
and Type II water. Plastic and glass containers are both suitable.
6.3 Nonaqueous samples shall be refrigerated and analyzed as soon as
possible.
Add 150 g stannous chloride to 300 mL
After the stannous chloride has
g mercuric chloride
acid and adjust the
(A commercial 1000-
the mercury stock
plan
manual
making use of a sampling
in Section One of this
44
-------
7.0 PROCEDURE
7.1 Sample preparation (steam-bath digestion):
7.1.1 All glassware for sample digestions, sample dilutions, and
standard preparations must be prewashed sequentially with an aqueous
detergent solution, mineral acids, and Type II water.
7.1.2 Weigh a 0.2-g portion of wet sample into a tared, 125-mL
Erlenmeyer flask. Add 5 mL Type II water and 5 mL aqua regia. Mix by
hand-swirling. Cover the top of the flask with a watch glass. Heat
for 2 minutes on a steam bath maintained at a temperature of
approximately 95°C. After cooling, add 50 mL Type II water and 15 ml
potassium permanganate solution. Mix thoroughly by hand-swirling.
Heat for 30 minutes on the steam bath maintained at a temperature of
approximately 95°C. After cooling and immediately prior to analysis,
add 6 mL of the sodium chloride-hydroxylamine hydrochloride solution or
sodium chloride-hydroxylamine sulfate solution to reduce the excess
permanganate in solution and the manganese dioxide particles adhering
to the flask wall. Quantitatively transfer the sample solution to a
100-mL volumetric flask and dilute to calibrated volume with Type II
water. Mix thoroughly and agitate by hand-swirling.
7.1.3 Standard preparation: Transfer appropriate aliquots of the
mercury working standards to a series of 125-mL Erlenmeyer flasks. The
aliquots must be selected to provide suitable calibration standards
within the linear dynamic range of the instrument. Add Type II water
to each flask to adjust the total volume to 10 mL. Add 5 mL aqua regia
and thoroughly mix by hand-swirling. Cover the top of the flask with a
watch glass. Heat for 2 minutes on the steam bath maintained at a
temperature of 95°C. After cooling, add 50 mL Type II water and 15 mL
potassium permanganate solution. Mix thoroughly by hand-swirling.
Heat on the steam bath for 30 minutes. After cooling and immediately
prior to analysis, add 6 mL of the sodium chloride-hydroxylamine
hydrochloride solution or sodium chloride-hydroxylamine sulfate
solution to reduce excess permanganate in solution and the manganese
dioxide particles adhering to the flask wall. Quantitatively transfer
the sample solution to a 100-mL volumetric flask and dilute to
calibrated volume with Type II water. Thoroughly mix and agitate by
hand-swirl ing.
7.2 Alternate sample preparation (autoclave digestion):
7.2.1 All glassware for sample digestions, sample dilutions, and
standard preparations must be prewashed sequentially with a detergent
solution, mineral acids, and then Type II water.
7.2.2 Weigh a 0.2-g portion of wet sample into a tared, 125-mL
Erlenmeyer flask. Add 5 niL Type II water, 5 mL concentrated sulfuric
acid, 2 mL concentrated nitric acid, and 5 mL potassium permanganate
solution. Mix thoroughly by hand-swirling. Cover the top of the flask
with a piece of aluminum foil. Autoclave the sample at 121°C and 15
45
-------
psi for 15 minutes. After cooling and immediately prior to analysis,
add 6 mL of the sodium chioride-hydroxylamine hydrochloride solution or
sodium chloride-hydroxylamine sulfate solution to reduce excess
permanganate in solution and the manganese dioxide particles adhering
to the flask wall. Quantitatively transfer the sample solution to a
100-mL volumetric flask and dilute to calibrated volume with Type II
water. Mix thoroughly and agitate by hand-swirling.
7.2.3 Standard preparation: Transfer appropriate aliquots of the
mercury working standards to a series of 125-mL Erlenmeyer flasks. The
aliquots must be selected to provide suitable calibration standards
within the linear dynamic range of the instrument. Add Type II water
to adjust the total volume to 10 ml in each flask. Add 5 mL Type II
water, 5 mL concentrated sulfuric acid, 2 mL concentrated nitric acid,
and 5 mL potassium permanganate solution. Thoroughly mix by hand-
swirling. Cover the top of each flask with a piece of aluminum foil.
Autoclave the standards at 121°C and 15 psi for 15 minutes. After
cooling and immediately prior to analysis, add 6 mL sodium chloride-
hydroxylamine hydrochloride solution to reduce excess permanganate in
solution and the manganese dioxide particles adhering to the flask
wall. Quantitatively transfer the standard solutions to 100-mL
volumetric flasks and dilute to calibrated volume with Type II water.
Mix thoroughly and agitate by hand-swirling.
7.3 Analysis: Quantitatively transfer an appropriate aliquot of
sample, standard, or reagent blank solution from the 100-mL volumetric flask
to the 125-mL sparging bottle. Add Type II water to adjust the total volume
to 100 mL. Add 5 mL stannous sulfate or stannous chloride solution. Ensure
that the nitrogen purge gas flow is set between 500 and 600 mL/min and that
the purge gas flow is directed through the bypass valve instead of through
the aeration bubbler tube of the cold-vapor/amalgamation apparatus. Also,
ensure that the exit end of the silver-wool collection tube is disconnected
from the absorption cell. Connect the sparging bottle to the cold-vapor
apparatus, and redirect the nitrogen purge gas flow through the aeration
bubbler tube for exactly 2 minutes. After the 2-minute concentration time,
redirect the nitrogen purge gas flow through the bypass valve, and connect
the exit end of the silver-wool collection tube to the absorption cell.
After the nitrogen purge gas has been directed through the bypass valve for
1 minute, turn on the Variac autotransformer and strip-chart recorder.
Initiate an atomic absorption measurement (maximum peak height) of 60
seconds duration; this measurement time may be pre-programmed with some
spectrophotometers to provide a digital display of the maximum absorbance
during this time interval. The mercury atomic absorption peak maximum
should appear on the strip-chart within 45 seconds. When the atomic
absorption peak has receded to 0.010 absorbance unit, turn off the Variac
autotransformer and strip-chart recorder, and cool the outside of the
silver-wool collection tube by a stream of air. When the mercury value
exceeds the linear range of the calibration curve, the analysis must be
repeated with smaller sample aliquots.
7.4 Construct a calibration curve by plotting the absorbance of a
series of mercury standards versus micrograms of mercury. Determine the
46
-------
absorbance peak height of the unknown sample from the strip-chart or from
the digital display of the atomic absorption spectrophotometer. Determine
the mercury concentration value from the standard curve.
7.5 Analyze, by the method of standard additions, all samples analyzed
as part of a delisting petition and all samples that suffer from matrix
interferences.
7.6 Duplicates, spiked samples, and check standards should be
routinely analyzed.
7.7 Calculate mercury concentrations by (1) the method of standard
additions or (2) from a calibration curve or (3) directly from the
calibrated concentration readout display of the spectrophotometer. All
dilution or concentration factors must be taken into account.
Concentrations reported for multiphased or wet samples must be appropriately
qualified (e.g., 5 pg/g dry weight).
8.0 QUALITY CONTROL
8.1 All quality control data should be maintained and available for
easy reference or inspection.
8.2 Calibration curves must be composed of a minimum of a reagent
blank and three standards.
8.3 Dilute sample digests if their mercury concentrations are higher
than that of the highest standard or if their mercury concentrations fall on
the plateau of the calibration curve.
8.4 Analyze a minimum of one reagent blank per sample batch to
determine if contamination or memory effects are occurring.
8.5 Analyze check standards after approximately every 6 samples.
8.6 Analyze one duplicate sample for every 10 samples.
8.7 Spiked samples or standard reference materials shall be
periodically analyzed to ensure that correct procedures are being followed
and that all equipment is operating properly.
8.8 The method of standard additions shall be used for all analyses
submitted as part of a delisting petition, and whenever a new sample matrix
is being analyzed.
9.0 METHOD PERFORMANCE
9.1 Because of the lack of ruggedness demonstrated in an
interlaboratory study, the accuracy and precision of the method has not been
statistically determined.
47
-------
10.0 REFERENCES
<>Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-
846, 2nd Edition, July 1982, U. S. Environmental Protection Agency, Office
of Solid Waste and Emergency Response, Washington, D.C.
<>Messman, J. 0., M. E. Churchwell, R. L. Livingston, and D. L. Sgontz.
1986. Evaluation and Testing of EPA Protocols for SW-846 Methods 7470 and
7471. Final Report. U.S. Environmental Protection Agency.
<>Churchwell, M. E., R. L. Livingston, 0. L. Sgontz, J. 0. Messman, and
W. F. Beckert. 1988 (in press). Environment International.
<>Youden, W. J., and E. H. Steiner. Statistical Manual of the AOAC.
Association of Official Analytical Chemists, Arlington, Virginia, 1975. 88
pp.
<>Gebhart, J. E., J. 0. Messman, and G. F. Wallace. 1987.
Interlaboratory Evaluation of SW-846 Methods 7470 and 7471 for the
Determination of Mercury in Environmental Samples. Final Report. U.S.
Environmental Protection Agency.
48
-------
Charcoal
Glass Stopcock
Absorption
Cell
Glass y
Needle Valve
Chrom-Alu
Resistance Heating
Winding
Flow Meter
Figure 1 A. Schematic diagram of the amalgamation CV-AAS system
Figure 1 B Block diagram of the amalgamation CV-AAS system
49
-------
APPENDIX C
PHASE III - OPTION A SAMPLE LIST
Option A of Phase III
two solid waste materials,
materials and two of the
predigestion spikes. Based
mix represents 16 samples.
#1
#2
#3
#4
#5
#6
#7
#8
#9
#10
#11
#12
#13
#14
#15
#16
specifies a sample mix of a reference sediment,
and three aqueous samples. The two solid waste
aqueous samples are designated for mercury
on duplicates of each sample, the total sample
These 16 samples are listed below:
Canadian reference sediment MESS-i
Duplicate of #1
Incinerator fly ash
Duplicate of #3
Incinerator fly ash spiked with 100 ng of inorganic mercury
Duplicate of #5
Municipal sewage sludge
Duplicate of #7
Municipal sewage sludge spiked with 100 ng of organic mercury
Duplicate of #9
Ground water
Duplicate of #11
Waste water spiked with 50 ng of inorganic mercury
Duplicate of #13
One percent (v/v) HNO 3 spiked with 25 ng of inorganic mercury
Duplicate of #15
50
-------
APPENDIX 0
PHASE III - OPTION B SAMPLE LIST
Option B of Phase III specifies a sample mix of a reference sediment,
two solid waste materials, and three aqueous samples. The two solid waste
materials and two of the aqueous samples are designated for mercury
predigestion spikes. Based on three replicates of each sample, the total
sample mix represents 24 samples. These 24 samples are listed below:
#1 Canadian reference sediment MESS-i
#2 Duplicate of #1
#3 Triplicate of #1
#4 Incinerator fly ash
#5 Duplicate of #4
#6 Triplicate of #4
#7 Incinerator fly ash spiked with 37.5 ng of inorganic mercury
plus 1 mg copper
#8 Duplicate of #7
#9 Triplicate of #7
#10 Municipal sewage sludge
#11 Duplicate of #10
#12 Triplicate of #10
#13 Municipal sewage sludge spiked with 50 ng of organic mercury
#14 Duplicate of #13
#15 Triplicate of #13
#16 Ground water
#17 Duplicate of #16
#18 Triplicate of #16
#19 Waste water spiked with 25 ng of inorganic mercury
#20 Duplicate of #19
#21 Triplicate of #19
#22 One percent (v/v) HNO 3 spiked with 25 ng of inorganic mercury
#23 Duplicate of #22
#24 Triplicate of #22
51
-------
APPENDIX E
INTERLABORATORY DATA FOR REVISED METHODS 7470 AND 7471
a Slope of calibration
0.005 abs/ng).
curve (typical slope for Battelle testing is
TABLE E-1.
SUMMARY OF CALIBRATION
FOR PHASE II ANALYSES
AND REAGENT BLANK VALUES
Laboratory
Identification
Method
7470
Method
7471
Slopea
Reagent
Blank
Slopea
Reagent
Blank
Number
(abs/ng)
(abs)
(abs/ng)
(abs)
949
0.0035
0.040
950
0.0037
0.007
0.0048
0.011
951
952
0.0038
0.004
0.0038
0.006
953
0.0041
0.013
0.0041
0.019
954
0.0015
0.006
0.0015
0.006
955
0.0011
0.120
0.0006
0.086
956
0.0008
0.008
0.0008
0.003
957
958
0.0032
0.054
959
0.0056
0.014
0.0025
0.011
960
0.0010
0.001
0.0010
0.001
961
0.003
0.005
962
0.0018
0.004
0.0016
0.001
963
0.0047
0.117
0.0033
0.602
964
0.0036
0.122
0.0036
0.122
965
0.0036
0.258
0.0001
0.644
966
0.0011
0.003
0.0011
0.003
52
-------
TABLE E-2. SUMMARY OF CALIBRATION AND REAGENT BLANK VALUES
FOR PHASE III ANALYSES
Laboratory
Identification
Number
Method
7470
Method
7471
Slopea
(abs/ng)
Reagent
Blank
(abs)
Slopea
(abs/ng)
Reagent
Blank
(abs)
Option A
955
0.0036
0.021
0.0030
0.025
957
0.0021
0.027
0.0017
0.020
962
0.0018
0.010
0.0023
0.012
964
0.004
0.014
0.004
0.O04
965
0.0029
0.046
0.0035
0.054
Option B
950
0.0036
0.006
0.0034
0.008
952
0.0056
0.008
0.0028
0.008
953
0.0036
-0.004
0.0065
0.003
954
0.0019
0.014
0.0024
0.075
966
0.0010
0.002
0.0010
0.002
a Slope of calibration curve (typical slope for Battelle testing is
0.005 abs/ng).
53
-------
TABLE E-3. SUMMARY OF MERCURY CONCENTRATIONS IN GROUND WATERa
Laboratory
Identification
Number
Mercury
Concentration,
ng
Mean
Std. 0ev.
Option A
#11
#12
955
2
3
2
0.71
957
1
<1
11
- -
962
4
8
6
2.8
964
7150
8700
-
- -
965
5
8
6
2.1
Option B
#16
#17
#18
950
<10
<10
<10
<10
- -
952
6
11
6
8
2.9
953
5
4
4
4
0.58
954
<1
<1
1
1
- -
966
51
44
42
-
--
a Endogenous mercury concentration - not detectable (<1 ng); based on
20-mL sample aliquot.
54
-------
TABLE E-4. SUMMARY OF MERCURY CONCENTRATIONS IN WASTE WATERa
SPIKED WITH INORGANIC MERCURY
Laboratory
Identification
Number
Mercury
Concentration,
ng
Percent
Recovery
Mean
Std. Dev.
Option A
#13
#14
955
37
51
88
20
957
80
84
160
5.7
962
115
100
220
21
964
9850
8600
- -
- -
965
32
45
77
18
Option B
#19
#20
#21
950
59
62
62
240
6.9
952
346000
300
150
- -
- -
953
23
22
22
89
2.3
954
48
42
34
160
28
966
27
27
27
110
0
a Endogenous mercury concentration - not detectable (<1 ng); based on
20-niL sample aliquot.
Option A spike: 50 ng Hg; target mercury concentration = 50 ng.
Option B spike: 25 ng Hg; target mercury concentration = 25 ng.
55
-------
TABLE E-5. SUMMARY OF MERCURY CONCENTRATIONS IN DILUTE NITRIC ACIDa
SPIKED WITH INORGANIC MERCURY
Laboratory
Identification
Number
Mercury
Concentration,
ng
Percent
Recovery
Mean
Std. Dev.
Option A
#15
#16
955
4
5
18
2.8
957
22
22
88
0
962
40
26
130
40
964
8650
5950
- -
- -
965
29
16
90
37
Option B
#22
#23
#24
950
42
29
20
120
44
952
40
22
18
110
47
953
2
2
3
9
2.3
954
19
14
27
80
26
966
24
21
26
95
10
a Endogenous mercury concentration - not detectable (<1 ng); based on
20-mL sample aliquot.
Option A and B spikes: 25 ng Hg; target mercury concentration = 25 ng.
56
-------
TABLE E-6. SUMMARY OF MERCURY CONCENTRATIONS IN MARINE SEDIMENT MESS 1a
Laboratory
Identification
Number
Mercury
Concentration,
mg/kg
Percent
Recovery
Mean
Std. 0ev.
Option A
#1
#2
955
0.19
0.21
120
8.3
957
0.18
0.17
100
4.1
962
0.17
0.18
100
4.1
964
0.77
0.28
- -
- -
965
0.22
0.24
140
8.3
Option B
#1
#2
#3
950
0.31
0.18
0.23
140
41
952
0.12
0.12
0.093
66
6.8
953
0.14
0.20
0.14
94
20
954
0.09
0.26
0.42
150
97
966
0.19
0.18
0.18
110
3.4
a Reference mercury concentration = 0.171 ± 0.014 mg/kg.
57
-------
TABLE E-7. SUMMARY OF MERCURY CONCENTRATIONS IN INCINERATOR FLY ASHa
Laboratory
Identi fi cation
Number
Mercury
Concentration,
mg/kg
Mean
Std. Dev.
Option A
#3
#4
955
0.037
0.014
0.026
0.016
957
<0.005
<0.005
<0.005
- - -
962
0.013
0.018
0.016
0.0035
964
0.43
0.14
- - -
- - -
965
0.033
0.043
0.038
0.0071
Option B
#4
#5
#6
950
<0.05
<0.05
0.075
0.058
0.014
952
0.017
0.025
0.029
0.024
0.0061
953
0.079
0.10
0.026
0.068
0.038
954
0.22
0.18
0.20
0.20
0.020
966
0.014
0.014
0.028
0.019
0.0081
a Endogenous mercury concentration - not detectable (<0.005 mg/kg).
58
-------
TABLE E-8. SUMMARY OF MERCURY CONCENTRATIONS IN INCINERATOR FLY ASHa
SPIKED WITH INORGANIC MERCURY
Laboratory
Identification
Number
Mercury
Concentration,
mg/kg
Percent
Recovery
Mean
Std.
0ev.
Option A
#5
#6
955
<0.005
0.075
8
10
957
0.55
0.62
120
10
962
0.40
0.41
81
1.4
964
2.62
1.09
- -
- -
965
0.49
0.42
91
10
Option B
#7
#8
#9
950
<0.05
<0.05
<0.05
<26
- -
952
0.019
0.029
0.028
13
2.9
953
0.062
0.050
0.033
25
7.7
954
0.16
0.13
0.12
72
11
966
0.058
0.17
0.094
56
30
a Option A spike: 100 ng Hg; target mercury concentration = 0.50 mg/kg.
Option B spikes: 37.5 ng Hg + 1 mg Cu; target mercury concentration =
0.19 mg/kg.
59
-------
TABLE E-9. SUMMARY OF MERCURY CONCENTRATIONS IN MUNICIPAL SEWAGE SLUDGEa
Laboratory
Identification
Number
Mercury
Concentration,
mg/kg
Percent
Recovery
Mean
Std. 0ev.
Option A
#7
#8
955
1.40
1.30
96
5.1
957
2.04
1.51
130
27
962
1.08
1.07
77
0.51
964
2.26
3.42
- -
- -
965
0.80
0.90
61
5.1
Option B
#10
#11
#12
950
1.64
1.49
1.45
110
7.2
952
1.23
1.21
2.02
110
32
953
0.20
0.24
0.11
13
4.8
954
1.55
1.70
1.30
110
14
966
1.14
1.41
1.91
110
28
a Target mercury concentration = 1.4 mg/kg.
60
-------
TABLE E-1O. SUMMARY OF MERCURY CONCENTRATIONS IN MUNICIPAL SEWAGE SLUDGEa
SPIKED WITH ORGANIC MERCURY
Laboratory
Identification
Number
Mercury
Concentration,
mg/kg
Percent
Recovery
Mean
Std. Dev.
Option A
#9
#10
955
2.22
0.92
83
48
957
3.19
2.76
160
16
962
1.90
1.72
95
6.7
964
2.53
3.60
- -
- -
965
1.00
1.05
54
1.9
Option B
#13
#14
#15
950
1.76
1.74
1.66
100
3.2
952
1.39
1.65
1.52
92
7.6
953
0.17
0.25
0.15
12
3.2
954
2.30
1.75
2.55
130
25
966
1.70
1.87
1.77
110
5.2
a Option A spike: 100 ng Hg; target mercury concentration = 1.9 mg/kg.
Option B spike: 50 ng Hg; target mercury concentration = 1.65 mg/kg.
61
------- |